SCALING IRRIGATION AND MALARIA RISK IN MALAWI By April N icole Frake A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Geography Doctor of Philosophy 2019 ABSTRACT SCALING I RRIGATION AND MALARIA RISK IN MALAWI By April Nicole Frake A primary means of increasing agricultural productivity to reduce global food insecurity is through intensification, including scaling irrigation measures (USAID, 2014). Aid agencies have strongly promoted agricultural intensification efforts to increase productivity and improve overall livelihoods. While scaling of irrigated agriculture has demonstrated significant boost s in productivity ( ADB, 2013; Melaine & Nonvide, 2017 ), agrarian transformatio n of the landscape through irrigated agricultural practices is associated with a number of water - related diseases (Hunter et al., 1993; Lacey & Lacey, 1990; Mather & That, 1984) including malaria (Ghebreyesus et al., 1999; Koudou et al., 2005; Oomen, De Wo lf, & Jobin, 1988). The government of Malawi established through the Green Belt Initiative (GBI) a long - term program aimed at land use modifications for the development of small and large - scale irrigation. While Malawians continue to face challenges direct ly related to food insecurity, the country is simultaneously holoendemic for malaria, wherein the disease is found in essentially all members of the population. This research investiga tes changes in disease dynamics of seasonal malaria cycles as a result of land transformation for irrigated agriculture using remote sensing and spatial analytical approaches. It is conducted against a backdrop of scaling up irrigated agricultural solutions across varying sectors, and myriad actors. To that end, the meaning o Research and Development (R&D) institutions and a conceptual framework of scaling up was constructed to promote ontological agreement of scaling up from defining programs through to final evaluation of success. Three scena rios for e stimated s patio - temporal distribution of suitable area for mosquito breeding pool formation and persistence were produced for the Bwanje Valley Irrigation Scheme (BVIS) using remotely sensed and field - based data. In addition, an estimation of hab itat suitability during the dry season was produced for the 8 - km area surrounding BVIS, the Bwanje Valley. P otential malaria transmission at the national scale driven by the GBI is presented through analysis of the current extent of irrigated agriculture, proposed expansion, and historic malaria prevalence data assessed by the 2012 , 2014 , and 2017 Demographic Health Survey s (DHS) in combination with the results of a habitat suitability model generated in Google Earth Engine . The conclusions from this study provide a strong foundation for agricultural land use decision making with respect to malaria transmission across Malawi. For Adam. ACKNOWLEDGEMENTS I am grateful to a number of people who were instrumental in ensuring this projec t was a success. First, to my advisor Dr. Joseph (Joe) Messina for his guidance, support, and innumerable conversations wherein he graciously pulled me off research rabbit - trail s and set me back on track. I am deeply grateful for having had the opportunit y to journey through this process with him and am a far better researcher as a result . A special amount of thanks is also extended to Dr. Edward (Ned) Walker for his mentorship and guidance both for research applicable to the work herein and well beyond. W orking alongside Dr. Sue Grady was a pleasure, I appreciated her encouragement, careful review of materials, and ever - present willingness to discuss and work through developing more stream - lined workflows. I am thankful to Dr. Leo Zulu for challeng ing my thinking , t eaching me to , and on a lighter note, making me laugh out loud both in and out of the field. Finally, I am grateful to Dr. Pouyan Nedjahashemi for his helpful feedback , patienc e as I stumbled my way through learning about agricultural engineering, and collaboration. I would also like to acknowledge my field assistants in Malawi: Stanley Phiri, Mayamiko Kakwera, and most especially, Willy Namaona. In addition, my undergraduate r esearch assistant, Celia Hallan who meticulously cataloged photographs and helped with data cleaning. I have had the joy of working alongside incredible lab mate s: Brad Peter, Leah Mungai, and the late Dr. Shengpan Lin. I am deeply grateful for the encou ragement and support I have felt from ed new languages (coding and spoken), and argued about everything from map projections to barbeque. To the members of Supporting Women in Geography (SWIG) , thank you for all the ways you supported me along the way. Most importantly, I am grateful to my husband, Adam and children, Ezra and Levi. This immensely grate the entirety of my academic career. vii TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ............... x LIST OF FIGURES ................................ ................................ ................................ ........... xi CHAPTER 1 ................................ ................................ ................................ ........................ 1 1.1 Introduction ................................ ................................ ................................ ................................ .......... 1 1.1.1 Food Insecurity ................................ ................................ ................................ ........................... 1 1.1.2 Agricultural Intensification ................................ ................................ ................................ ........ 2 1.1.3 Irrigation ................................ ................................ ................................ ................................ ....... 3 1.1.3.1 Irrigated Agriculture & Parasitic Disease ................................ ................................ ...... 4 1.1.4 Malaria ................................ ................................ ................................ ................................ .......... 5 1.1.5 Malaria Transmission ................................ ................................ ................................ ................. 7 1.1.5.1 Vector ................................ ................................ ................................ ................................ . 7 1.1.5.2 Environment ................................ ................................ ................................ ..................... 9 1.1.5.3 Humans ................................ ................................ ................................ ............................ 11 1.2 Research Design ................................ ................................ ................................ ................................ . 12 1.2.1 Study Country ................................ ................................ ................................ ............................ 12 1.2.2 Study Site ................................ ................................ ................................ ................................ .... 13 1.2.3 Study Area ................................ ................................ ................................ ................................ .. 13 1.2.4 Specific Aims ................................ ................................ ................................ ............................. 14 1.3 Outline of the Dissertation ................................ ................................ ................................ ............... 15 CHAPTER 2 ................................ ................................ ................................ ...................... 16 2.1 Introduction ................................ ................................ ................................ ................................ ........ 16 2.2 Scaling ................................ ................................ ................................ ................................ ................... 18 2.3 What is Scale? ................................ ................................ ................................ ................................ ...... 18 2.4 Scaling Up ................................ ................................ ................................ ................................ ............ 19 2.4.1 Pathways to Scaling ................................ ................................ ................................ .................. 20 2.5 A Visual Analysis of Scaling Up ................................ ................................ ................................ ....... 21 2.6 Discussion ................................ ................................ ................................ ................................ ............ 25 2.7 Conclusions ................................ ................................ ................................ ................................ ......... 29 CHAPTER 3 ................................ ................................ ................................ ...................... 32 3.1 Introduction ................................ ................................ ................................ ................................ ........ 32 3.2 Methods ................................ ................................ ................................ ................................ ............... 32 3.2.1 Stu dy Site ................................ ................................ ................................ ................................ .... 32 3.2.2 Study Design ................................ ................................ ................................ .............................. 33 3.2.3 Environmental Characteristics ................................ ................................ ................................ 35 3.2.3.1 Soils ................................ ................................ ................................ ................................ ... 35 3.2.3.2 Land Cover ................................ ................................ ................................ ...................... 36 3.2.3.2.1 Data preparation ................................ ................................ ................................ ... 36 3.2.3.2.2 Land Cover Data Collection ................................ ................................ ............... 36 3.2.3.2.3 Land Cover Classification ................................ ................................ .................... 37 3.2.3.2.3.1 BVIS: Rainy Season ................................ ................................ .................... 37 viii 3.2.3.2.3.2 BVIS: Dry Season ................................ ................................ ....................... 43 3.2.3.2.4 Validation of Classification Results ................................ ................................ .... 45 3.2.4 Anthropogenic Influence ................................ ................................ ................................ ......... 48 3.2.4.1 BVIS Structure and Irrigation ................................ ................................ ....................... 48 3.2.4.2 Drainage ................................ ................................ ................................ ........................... 5 3 3.2.4.3 BVIS Operations and Governance ................................ ................................ .............. 55 3.3 Results ................................ ................................ ................................ ................................ .................. 57 3.3.1 Breeding Distribution Scenarios ................................ ................................ ............................. 57 3.3.1.1 Mid - Rainy Season ................................ ................................ ................................ ........... 58 3.3.1.2 Dry Season: Limited Water Resources ................................ ................................ ........ 58 3.3.1.3 Dry Season: Abundant Water Resources ................................ ................................ .... 63 3.4 Discussion and Conclusions ................................ ................................ ................................ ............. 64 CHAPTER 4 ................................ ................................ ................................ ...................... 67 4.1 Introduction ................................ ................................ ................................ ................................ ........ 67 4.2 Bwanje Valley and the Bwanje Valley Irrigation Scheme ................................ ............................. 67 4.3 Methods ................................ ................................ ................................ ................................ ............... 69 4.4 Land Use and Land Cover ................................ ................................ ................................ ................ 71 4.4.1 Data Preparation ................................ ................................ ................................ ....................... 71 4.4.2 Land Cover Data Collection ................................ ................................ ................................ ... 72 4.4.3 Land Cover Classification ................................ ................................ ................................ ........ 72 4.4.3.1 Bwanje Valley Irrigation Scheme ................................ ................................ .................. 73 4.4.3.2 Bwanje Valley ................................ ................................ ................................ .................. 73 4.4.4 Validation of Classification Results .. 78 4.5 Soils ................................ ................................ ................................ ................................ ....................... 79 4.5.1 BVIS ................................ ................................ ................................ ................................ ........... 80 4.5.2 Bwanje Valley ................................ ................................ ................................ ............................ 80 4.6 Modeling Breeding Pool Suitability ................................ ................................ ................................ .. 82 4.7 Resul ts ................................ ................................ ................................ ................................ .................. 84 4.7.1 BVIS Breeding Suitability ................................ ................................ ................................ ........ 84 4.7.2 Bwanje Valley Suitability ................................ ................................ ................................ .......... 84 4.8 Discussion and Conclusions ................................ ................................ ................................ ............. 87 CHAPTER 5 ................................ ................................ ................................ ...................... 90 5.1 Introduction ................................ ................................ ................................ ................................ ........ 90 5.2 Methods ................................ ................................ ................................ ................................ ............... 91 5.2.1 Irrigated Agriculture in Malawi ................................ ................................ ............................... 91 5.2.1.1 Irrigation Development and National Policy Frameworks ................................ ...... 91 5.2.1.2 Irrigation Development in Malawi ................................ ................................ ............... 96 5.2.1.3 Spatial Extent of Irrigated Agriculture in Malawi (2015) ................................ .......... 98 5.2.1.4 Expansion of Irrigated Agriculture ................................ ................................ ........... 102 5.2.2 Modeling Habitat Suitability for Mosquitoes ................................ ................................ ..... 108 5.2.2.1 GEE ................................ ................................ ................................ ............................... 109 5.2.2.2 Habitat Suitability Modeling Methods ................................ ................................ ...... 109 5.2.2.3 Model Construction ................................ ................................ ................................ ..... 109 5.2.2.3.1 Predictor Variables ................................ ................................ ............................. 110 5.2.2.3. 1. 1 Climate ................................ ................................ ................................ ....... 111 5.2.2.3. 1. 2 Land ................................ ................................ ................................ ........... 111 ix 5.2.2.3. 1. 3 Water ................................ ................................ ................................ .......... 115 5.2.2.4 Model Outputs ................................ ................................ ................................ ............. 117 5.2.3 Malaria in Malawi ................................ ................................ ................................ ................... 121 5.2.3.1 DHS MIS ................................ ................................ ................................ ...................... 121 5.2.3.2 DHS MIS Data and Processing ................................ ................................ ................. 121 5.2.3.3 Malaria Prevalence in Malawi ................................ ................................ ..................... 122 5.3 Results ................................ ................................ ................................ ................................ ................ 129 5.3.1 Habitat Suitability and Malaria Prevalence ................................ ................................ ......... 129 5.3.2 Scaling Irrigation and Malaria Risk in Malawi ................................ ................................ ... 133 5.4 Discussion and Conclusions ................................ ................................ ................................ .......... 135 CHAPTER 6 ................................ ................................ ................................ .................... 142 6.1 Introduction ................................ ................................ ................................ ................................ ...... 142 6.2 Overall Contributions ................................ ................................ ................................ ...................... 143 6.2.1 Toward a Common Ontology o f Scaling Up in Development ................................ ....... 143 6.2.2 LULCC for Irrigated Agriculture and its Impact on Mosquito Distribution ............... 144 6.2.3 Malaria Vuln erability and Dynamic Changes to Irrigated Agriculture ........................... 146 6.2.4 Measuring agreement of predictive mosquito habitat ................................ ...................... 148 6.3 Future R esearch ................................ ................................ ................................ ................................ 149 6.3.1 Interpretations of Scaling and Their Influence on Development ................................ .. 149 6.3.2 Heterogeneity in Irrigated Landscapes ................................ ................................ ............... 150 6.3.3 Spatial Structure of Active Agriculture ................................ ................................ ............... 152 6.3.4 Surface Wetness ................................ ................................ ................................ ..................... 152 6.3.5 Spatio - Temporal Expansion of Irrigated Agriculture ................................ ...................... 153 6.3.6 Habitat Suitability and Malaria Prevalence ................................ ................................ ......... 153 A PPENDICES ................................ ................................ ................................ ................. 155 A PPENDIX A: Rainy Season LULC Data: BVIS ................................ ................................ ...... 156 APPENDIX B: Dry Season LULC Data: BVIS ................................ ................................ ......... 206 APPENDIX C: Dry Season LULC Data: Bwanje Valley ................................ .......................... 289 BIBLIOGRAPHY ................................ ................................ ................................ ............ 331 x LIST OF TABLES Table 3.1: Rainy season land cover class descriptions for BVIS ................................ ............................... 39 Table 3.2: Dry season land cover class descriptions for BVIS ................................ ................................ .. 40 Table 3.3: accuracy; E.O., omission errors; C.O., commission errors . ................................ ................................ .... 47 Table 3.4: Dry season cla ssification accuracy assessment accuracy; E.O., omission errors; C.O., commission errors ................................ ................................ ........ 48 Tab le 4.1: Land cover classifications for the Bwanje Valley ................................ ................................ ...... 74 Table 4.2 : accuracy; E.O., omission errors; C .O., commission errors . ................................ ................................ .... 79 Table 4.3 : accuracy; U.ac, user accuracy; E.O., omission errors; C.O., commissi on errors ................................ .. 79 Table 5.1: Summary of existing schemes by typology and hectarage ................................ .................... 101 Table 5.2: Existin g, considered, and new irrigation schemes under the IMP (Data Source: Irrigation Master Plan and Investment Framework, 2015) ................................ ................................ ...................... 103 Table 5.3: Predictor variables, data sources, resolutions, and t hreshold variables for model construction ................................ ................................ ................................ ................................ ................... 110 Table 5.4: MODIS Type 1 LULC classes and their suitability for An. gambiae s.s. .............................. 113 Table 5.5: Percentage of children ages 6 - 59 months according to microscopy for Ma lawi DHS MIS Surveys 2012, 2014, and 2017 ................................ ................................ ................................ ..................... 123 Table 5.6: Percent area of habitat suitability for proposed GBI si tes under a maximum estimation 138 Table 5.7: R anked categorization of GBI sites based on their likelihood to increase malaria risk through production of larval habitat ................................ ................................ ................................ .......... 139 xi LIST OF FIGURES Figure 1.1: The Bwanje Valley Irrigation Scheme (BVIS) located in the central region Malawi and the study area, defined as the 8km surrounding BVIS. ................................ ................................ ..................... 13 Figure 2.1: Visual analysis of 'Scaling Up' definitions across CG Consortium, USAID, and IFAD. 22 Figure 2.2: A framework for conceptually defining scaling and putting it into practice for development activities ................................ ................................ ................................ ................................ ..... 27 Figure 3.1: Conceptual framework for the study. Soils data are taken from JICA (1994) .................... 34 Figure 3.2: Soil types at BVIS. Data and soil ch aracteristics are taken from JICA (1994) .................... 35 Figure 3.3: Rainy season land cover classification for BVIS including field photos depicting varied stages of rice growth within the scheme. ................................ ................................ ................................ ...... 42 Figure 3.4: Dry season land cover classification for BVIS including field photos of varied land covers surveyed. ................................ ................................ ................................ ................................ ............... 46 Figure 3.5: Variations between BVIS farmer cooperative intended distribution of crop types in relationship to where crops were sampled during field surveys. ................................ ............................... 47 Figure 3.6: Bwanje Valley Irrigation Scheme (BVIS) 2017 surveyed perimeter (green). Digitized perimeter of BVIS from JICA's Basic Design and Study Report (2005) and visual inspection of Google Earth v. 7.3.1 imagery (black) initially used for data analysis and field collection. .................. 51 Figure 3.7: Irrigation water control structures at BVIS ................................ ................................ .............. 52 Figure 3.8: Mounded grass and debris are placed in front of a branching canal slide gate at BVIS to pro hibit water flow. ................................ ................................ ................................ ................................ ......... 53 maintenance area of the canal ................................ ................................ ................................ ........................ 53 Figure 3.10: Examples of improper drain management capture during field surveys. Surface drains are meant to be free of weeds and grasses in order to facilitate the flow and redirection of water to other portions of the scheme. Drain maintena nce is the shared responsibility of farmers and the BVIS cooperative ................................ ................................ ................................ ................................ ............. 55 Figure 3.11: Rainy season breeding scenario under abundant water resource conditions at BVIS. .... 60 Figure 3.12: Breeding scenario under limited water resource conditions during the dry season ......... 61 Figure 3.13: Breeding scenario under abundant water resource conditions during the dry season as a result of the construction of the Bwanje dam ................................ ................................ ............................. 62 xii Figure 4.1: The Bwanje Valley and Bwanje Valley Irrigation Scheme (BVIS) locate d in Dedza district, central Malawi. The Bwanje Valley is defined as the 8km area surrounding BVIS. ................ 69 Figure 4.2: Conceptual Framework ................................ ................................ ................................ ............... 70 Figure 4.3: Combined LULC Dry Season classification produced for the Bwanje Valley and BVIS. 75 Figure 4.4: Sample photographs depicting classified dry season land cover types of the Bwan je Valley. ................................ ................................ ................................ ................................ ................................ 76 Figure 4.5 : The Namikokwe River basin east of the BVIS headworks during the dry season. Limited water resources lead to the stagnation of water often resulting in conversion of the area to dimba gardens along the channel. ................................ ................................ ................................ .............................. 78 Figure 4.6 : Soil types of the Bwanje Valley and their potential for ponding ................................ .......... 81 Figure 4 .7 : Soil drainage of the Bwanje Valley ................................ ................................ ............................ 82 Figure 4.8 : Dry season breeding scenario for the Bwanje Valley ................................ .............................. 86 Figure 5.1: Conceptual framework ................................ ................................ ................................ ................ 93 ameworks associated with irrigation expansion within the Agriculture, Economic, and Irrigation sectors. ................................ ................................ ................................ ................. 95 Figure 5.3: Timeline of expansion of irrigated agriculture in Malawi (1949 - 1979) ................................ 98 Figure 5.4: Spatial Extent of Irrigated Agriculture in Malawi (2015) ................................ .................... 102 Figure 5.5: Malawi's proposed irrigation schemes, by type. ................................ ................................ .... 105 Figure 5.6: Proposed Green Belt Initiative Sites including intended crop types. Total area for proposed irrigation across GBI sites is nearly 1,000,000ha. ................................ ................................ ... 107 Figure 5.7: Anopheles gambiae s.s. habitat suitability in Malawi, 2012 ................................ ....................... 118 Figure 5.8: Anopheles gambiae s.s. habitat suitability in Malawi, 2014 ................................ ....................... 119 Figure 5.9: Anopheles gambiae s.s . habitat suitability in Malawi, 2017 ................................ ....................... 120 Figure 5.10: Percentage of Children Age 5 - 59 Months Microscopy Positive f or Malaria Infection by DHS Cluster (2012). ................................ ................................ ................................ ................................ ..... 126 Figure 5.11: Percentage of Children Age 5 - 59 Months Microscopy Positive for Malaria Infection by DHS Cluster (2014). ................................ ................................ ................................ ................................ ..... 127 Figure 5.12: Percentage of Children Age 5 - 59 Months Microscopy Positive for Malaria Infection by DHS Cluster (2017). ................................ ................................ ................................ ................................ ..... 128 xiii Figure 5.13: Habitat suitability dri vers by MIS cluster, 2012 ................................ ................................ .. 130 Figure 5.14: Habitat suitability drivers by MIS cluster, 2014 ................................ ................................ .. 131 Figure 5.15: Habitat suitability dri vers by MIS cluster, 2017 ................................ ................................ .. 132 Figure 5.16: Change in distribution of habitat suitability for An. gambiae s.s. after LULCC for GBI sites under a maximum estimation ................................ ................................ ................................ ............. 134 1 CHAPTER 1 1. 1 Introduction 1. 1.1 Food Insecurity Food insecurity is a global phenomenon with a distinctly spatial character. Prior to 2016, the prevalence of undernourishment was declining, from 18.6% of the global population in 1990 - 1992 to 10.9% recorded from 2014 - 2016 (FAO, IFAD, & WFP, 2015) . Recent reports show a potential reversal of trends: In 2018, 821 million people were undernourished (FAO, IFAD, UNICEF, WFP, and WHO 2018) ; an increase over the 777 million estimated in 2015 (FAO, IFAD, UNICEF, WFP, 2017) . The Food and Agriculture Organization of the United Nations (FAO) defines food security as, and nutritious food which meets thei r dietary needs and food preferences for an active and healthy (FAO, 2003, p.29) (FAO, 2003, p.29) . While access to sufficient food is considered a basic human right (United Nations, 1948) , many remain hungry. Spatial trends of food ins ecurity reveal that populations with the greatest vulnerability to food insecurity live in developing countries (United Nations, 2016) . At the regional scale, substantially larger shares of undernourishment are experienced in Southern Asia and sub - Saharan Africa (FAO et al., 2015) report, the prevalence of undernourishment in sub - Saharan Africa was 23.2% in 2017; and 14.8% in Southern Asia (FAO, IFAD, UNICEF, WFP, 2018) . By December 2018, FAO reported that across Africa 31 countries required e xternal assistance for food: 11 countries were experiencing widespread lack of access to food, 19 faced severe localized food insecurity, and the Central African Republic had excepti onal shortfalls in aggregate food production and supplies (FAO, 2018) . A related point to food insecurity pressures is population expansion: more people translates to the need for more food. B y mid - 2017 the global population totaled nearly 7.6 billion, with a 2 projected increase to 9.7 billion by 2050 according to medium - variant projections (UN, 2017) . Challenges to feeding the expanding population are considerable. While sa tisfying global food needs will require an increase in food production (FAO, 2017; Rockström et al., 2017) , the demand for food is spatially uneven. Important shifts in population dyn amics have occurred in the proceeding decades, including rural to urban migration and a growing number of live births. By mid - century, it is estimated that two - thirds of the global population will live in urban areas (FAO, 2017) . Further, across regions, asymmetrical population growth is expected to continue: Africa and Asia are ant icipated to be the largest contributors to population expansion between 2017 and 2050, increasing their populations by 1.7 billion and 750 million respectively (United Nations, 2017) . Coupled with an expanding population, demand for food is further exacerbated by sig nificant gains in life expectancy across regions. Estimated projections suggest that the global life expectancy at birth will rise from 71 years in 2010 - 2015 to 77 years by 2045 - 2050 (United Nations, 2017) . 1 .1 .2 Agricultural Intensification To meet the demands of a n increasing population projected to live longer, food systems must continue public investment in scientific research designed for agricultural development l ed to a revolution in significant increases in calories produced per acre. Between 1960 and 2015, agricultural production has more than tripled (FAO, 2017) . Yet, despite these tremendous strides, widespread hunger and malnutrition remain pervasive, exacerbating the need for continued innovation. In a speech delivered by FAO Director - usual would mean a huge and simultaneous increase in the need for food, ener gy, and water in the (FAO, 2015) . 3 Addressing the challenge of feeding more people requires a multi - faceted approach. Only 11% of global land surfac e area is suitable for agriculture; Of this, poor natural resource management has already compromised 38% (USAID, 2017) . With finite land resources, and an inability to expand, emphasis on doing more with less is paramount. Ag agricultural production per unit of inputs (which may be labor, land, time, fertilizer, seed, feed, or (FAO, 2004, p.3) . A related response to agricultural intensification is sustainable intensification, where more food i s produced from the same land area, but with the added emphases of reducing environmental impacts and increasing natural capital and flow of environmental services (FAO, 2011; Godfray et al., 2010) . By design, intensification should occur over time, achieving increasing yields while subs equently using fewer resources. Approaches to intensification are considerable including, use of high - yield crop varieties, chemical pesticides and fertilizers, irrigation, and mechanization (Ringler et al., 2013) . 1. 1.3 Irrigation Agriculture is the single largest driver of environmental change (Rockström et al., 2017) and irrigation has been shown to have substantial environmental impact (Dougherty & Hall, 1995) . In irrigated systems, water is artificially applied to soil to assist with crop, tree, or pasture produc tion (FAO, 2014) . Surface irrigation involves the application of water by gravity flow to the field surface; sprinkler irrigation systems are typified by spraying water on to crops; and in drip irrigation systems, water is applied directly t o the soil through emitters wetting the immediate root zone of each plant (Brouwer, Prins, Kay, & Heibloem, 1988) . Globally, 40% of crop production is under irrigation (AQUASTAT, 2014) and i n Africa, total land area used for irrigation is expected to increase from 11.9 million hectares (ha) in 1990 to 15.9 million ha in 2020 (Rosengrant & Perez, 2000). Approximately 70% of t he African population are small holder famers (AGRA, 2017) who depend on rain - fed, sta ple crop production (Burney & Naylor, 2012) . Subject to seasonal weath er 4 fluctuations, production constraints, and characteristically low yields, small - scale irrigation has been promoted as means to mitigate the effects of climate variability (Mango, Makate, Tamene, Mponela, & Ndengu, 2018) , increase crop productivity (Kamwamba - Mtethiwa, Weatherhead, & Knox, 2015) , and serve as a poverty alleviation tool (Burney & Naylor, 2012) . 1. 1.3.1 Irrigated Agriculture & Parasitic Disease While increases in food production can be achieved through intensifying agricultural practices, the process of land conversion for intensified agriculture collectively alters natural biotic interactions within ecosystems. Previous literature has shown agrarian transformation of the landscape for irrigation is ass ociated with a number of water related parasitic diseases, including schistosomiasis, filariasis, onchocerciasis, and malaria (Boelee & Madsen, 2006; Ijumba & Lindsay, 2001; J.M.Hunter, L.Re y, K.Y.Chu, E.O.Adekolu - John, 1993; Patz, Graczyk, Geller, & Vittor, 2000) . W ater is a requisite for mosquito development. As such, land use and land cover changes that alter the distribution and flow of water across the landscape can have profound impac ts on the epidemiology of malaria. Keiser et al. (2005) highlight that a s much as 9 0% of the global malaria problem can be attributed to environmental factors including the establishment of irrigated schemes. Irrigation for crop production can encourage pathogen transmission through a number of pathways. First, through the development of vector habitat and the production of adult stage mosquitoes (Van Der Hoek 2004; Mutero et al. 2004). Intensification of agriculture involves a significant change to the natural landscape occurring across areas, altering vegetation and expanding surface w ater availability. Likewise, expanding irrigation can promote vector longevity by significantly increasing relative humidity over large areas (Secretariat & WHO, 1996) . Collectively, land scape modifications for irrigated agriculture have the potential to both promote diversity of breeding sites and reduce predation of vectors (Sutherst, 2004) . Further, environmental and ecological changes for irrigated 5 agriculture can increase the frequency of human - vector contact thereby encouraging transmission (Secretariat & WHO, 1996) . The association between irrig ated agriculture a nd malaria is well documented in the literature. In some studies, malaria prevalence increases (Yaw Asare Afrane et al., 2004; Ghebreyesus et al., 1999; Guthmann, Llanos - Cuentas, Palacios, & Hall, 2002; Jaleta et al., 2013; Keiser, Caldas, et al., 2005; Kibret et al., 2010; Kobayashi et al., 2000) . Contrastingly, other studies have shown a decrease or no change in prevalence of infection (Assi et al., 2013; Diakité et al., 2015; Faye et al., 1995; Ijumba, Mosha, & Lindsay, 2002; Klinkenberg, Van Der Hoek, & Amerasinghe , 2004; Mutero et al., 2004) . The contradictory nature of such studies suggests the necessity for further investigation on the impact of irrigated agriculture on malaria transmission particularly in light of continued emphasis on expansion throughout mala ria endemic areas to meet crop production goals. 1. 1. 4 Malaria Discovered by Laveran in Constantine, Algeria, in 1880, the malarial agent is a parasitic protozoan of the genus Plasmodium and the class Sporozoa (May, 1961) . Only four type s of Plasmodium infect humans - Plasmodium malariae, P. ovale, P. vivax, and P. falciparum. P. falciparum and P. vivax are the two most widespread malaria parasites (NIH, 2016) . In sub - Saharan Africa, P. falciparum was responsible for an estimated 99% of in fections in 2016 (WHO, 2017) . Roughly 64% of malaria cases in the WHO Region of the Americas, over 30% in th e WHO South East Asia, and 40% in the Eastern Mediterranean Region are attributable to P. vivax (WHO, 2017) . The life cycle of the malaria parasite occurs in three stages - sporozoites, merozoites, and gametocytes, and involves two hosts: humans and female Anopheles mosquitoes (CDC, 2016) . In humans, transmission occurs during a blood meal when sporozoites are injected into the human host replicating in the liver cells (Klein, 2013) and maturing into schizonts (CDC, 2016) : the human liver 6 stage. After 6 - 15 days (Klein, 2013) , the liver schizonts rupture, releasing merozoit es into the bloodstream to invade red blood cells and beginning the human blood stages (CDC, 2016) . Within the red blood cells, merozoites replicate asexually and progress through ring and trophozoite stages (Klein, 2013) to develop either a blood - stage schizont, or gametocyte (Crutcher & Hoffman, 1996) . Blood - stage schizonts will eventually rupture, releasing on average 16 daughter merozoites into the blood stream to further infect red blood cells (Klein, 2013) . Clinical manifestations of malaria including the characteristic fever cycle, are attributable to these blood stage parasites (CDC, 2016; Klein, 2013) . The sexua l stage of the plasmodium, the gametocyte (Crutcher & Hoffman, 1996) in either male (microgame tocytes) or female (macrogametocytes) form is ingested by Anopheles mosquitoes during blood meal (CDC, 2016) . Gametocytes will mate within the gut of the mosquito, producing zygotes that penetrate the midgut wall of the mosquito and form an oocyst (CDC, 2016; Klein, 2013) . As the oocyst develops, it produces thousands of sporozoites (K lein, 2013) . These sporozoites are preparation for perpetuating the malaria life cycle (CDC, 2016) . Populations that typicall y reside in malaria endemic areas often experience mal nourishment in part due to lower socio - economic status. The relationship between malaria and malnutrition is complex (Almeida et al., 2015) . Fillo l et al. ( 2009) suggested that malnutrition may facilitate the development of protective anti - malarial immune response . Contrastingly, Friedman et al. ( 2005) demonstrated that stunting was significantly associated with increased odds of conc urrent malaria and that wasting increased severe malaria anemia risk. Likewise, Pérez - Escamilla et al. ( 2009) showed that in Haitian children <5 years old , severe household food insecurity was associated with malaria risk , even after controlling for Body Mass Index (BMI). It is important to note that malnutrition is not simply a lack of sufficient calories, but a lso a lack or imbalance of essential vitamins and minerals (i.e. 7 micronutrients). Micronutrient deficiencies , in particular protein - energy malnutrition, vitamin A, zinc, and folate play a critical role in malaria morbidity and mortality (Caulfield, Richard, & Black, 2004; Shankar, 2000) . Shankar ( 2000) provides a critical review of the relationship between malaria and nutritional status including vitamin A, B, and E, zinc, iron, folate, unsaturated fatty acids, amino acids, and selenium. Findings suggest that improvement in dietary intake may significan tly reduce malaria morbidity and mortality, highlighting the need for integration of effective nutrient - based interventions to existing malaria intervention programs. 1. 1.5 Malaria Transmission Despite a long history of efforts to curb infection through emphasis on vector, parasite, and habitat modification, malaria remains one of the most significant challenges to global public health (CDC, 2018) . In 2016, the World Health Organization (WHO) estimated that 216 million cases of malaria occurred worldwide; 445,000 cases resulted in death (WHO, 2017) . These statistics reflect an increase in the overall number of cases in 2015 (211 million), but a slight reduction in the number of fatalities (446,000) (WHO, 2017) . The WHO African Region carries the greatest disease burden accounting for 90% of all malaria cases (WHO, 2017) . Malaria transmission and intensity are a byproduct of the complex interplay of a wide variety of biotic and abiotic factors. Herein specific conside ration is given to factors related to the vector, environment, and humans. 1.1.5.1 Vector Human malaria is caused by plasmodium parasites transmitted by mosquitoes of the genus Anopheles . Of the 475 formally recognized and 50 unnamed members of the specie s complex (Mosquito Taxonomic Inventory, 2018) , approximately 70 have the capacity to vector malaria parasites (Service & Townson, 2002; Sinka et al., 2012) . Sinka et al. (2012) characterize 41 of these species a s dominant vector species/species complexes (DVS) based on their transmission capability and public health concern. In Africa, An. gambiae, An. arabiensis, and An. funestus are the primary vectors responsible for malaria transmission (Tonnang, Kangalawe, & Yanda, 2010) . 8 The capacity of a given mosquito species to be an effective malarial vector is tied to a number of defining characteristics including: host selection preference, patterns of feeding and resting, longevity, biting rate, and vector competence. Mosquitoes more likely to transmit malaria are those whose b ehavior facilitate greater vector - human contact. If reasonably available, many species exhibit a predisposition for specific host types: human (anthropophilic) or animal (zoophilic) (Mcdonald, 1957) . Both An. gambiae and An. funestus demonstrate strong anthropophilic feeding characteristics (CDC, 2015) ; An. arabiensis , while still exhibiting a preference for biting humans, is described as less anthropophilic (Foster & Walker, 2009 ; Sinka et al., 2010) . Mosquito feeding patterns are characterized by timing (nocturnal, diurnal, or crepuscular) and location of biting, indoors (endophagic) or outdoors (exophagic). While only female mosquitoes feed on vertebrate blood, both male and female mosquitoes regularly feed on plant sugars for nutrition and energy (Foster & Walker, 2009 ) . Mosquito resting behavior has important implications for indoor residual spraying. Endophilic mosquitoes are females that preferentially rest indoors from the time of blood - feeding to the onset of searching for an oviposition s ite (Pates & Curtis, 2005) . Exophilic mosquitoes rest outdoors after blood feeding (CDC, 2017) . Malaria transmission intensity is ch (Garrett - Jones, 1964; D. L. Smith et al., 2012) . Smith et al. (2012) present a comprehensive synthesis of the historical d evelopment of vectorial capacity and its significance to public health. The from all the mosquitoes that would bite a single fully infectious person on a (Smith et al., 2012, pg. 6) and is comput ed as : (1) 9 where m is the ratio of adult mosquitoes to humans (vector density), a is human biting rates, n is the parasites Extrinsic Incubation Period (EIP), and p is daily probability of mosquito survival. While VC principally influe nces the transmission of malaria, vector competence too is essential (Cohuet, Harris, Robert, & Fontenille, 2010) . Vector competence refers to the ability of an organism to become infected, maintain and replicate the pathogen, then transmit it (Fortuna et al., 2015) . Thus, mosquito longevity is a critical function of malaria transmission: mosqui toes must survive long enough to become infected, get through the extrinsic incubation period, then deliver infectious bites (Cohuet et al., 2010; D. L. Smith et al., 2012) . A related point to consider is abundance, both human a nd vector. For transmission to occur, mosquito species populations must be high enough to ensure an encounter with an infectious human carrier of the parasite. Spatio - temporal fluctuations in vector capacity are often a byproduct of differences in mosquito densities across space and time (D. L. Smith et al., 2012) . In addition, each mosquito must carry enough malaria parasites within their saliv ary glands to ensure parasitic transmission. 1.1.5.2 Environment Environmental factors that influence malaria transmission are Land Use and Land Cover (LULC), weather and climate. The association between LULC and malaria transmission is well established, in part because availability of breeding habitat strongly influences mosquito activity. As such, LULC can encourage or restrict vector populations through availability of breeding sites. Many mosquito species exhibit specific preferences in breeding habita ts (Secretariat & WHO, 1996) . For instance, An. gambiae s.s. prefer small, temporary, sunlit pools generally free from organic matter (Gimnig, Ombok, Kamau, & Hawley, 2001; Sinka et al., 2010) whereas An. funestus s.s. prefer breeding sites with emergent vegetation and large, permanent or semi - permanent fresh water bodies (Sinka et al., 2010 ) . A related point is the influence of phenology and availability of sugar sources associated with various land cover types on mosqu ito population dynamics. (Gu et al., 2011) . 10 The influence of land cover on microclimatic conditions is considerable given that temperature is an important determinant of malaria risk (K. P. Paaijmans, Read, & Thomas, 2009) . Mosq uito lifecycles are temperature sensitive. Warmer temperatures can promote faster development of some species, leading to smaller adults (Christiansen - Jucht, Parham, Saddler, Koella, & Basáñez, 2014) . This is an important consideration as adult mosquito size influences longevity, biting rate, size of bloodmeal, and gonotrophic cycle (i.e., the length of time between a bloodmeal and oviposition of eggs) (Christiansen - Jucht et al., 2014) . Further, Briegel (1990) showed that a positive correlation exists between body size and fecundity, which may feedback through density dependent competition and mortality in aquatic stages of development (M. T. White et al., 2011) . Beyond body size, higher ambient temperatures are also associated with faster bloodmeal digestion, subsequently shortening the gonotrophic cycle, and increasing biting frequency (Yaw A Afrane, Lawson, Githeko, & Yan, 2005) Temperature is also an important determinant of parasite biology (Blanford et al., 2013) . Extrinsic Incubation Period (EIP) refers to the amount of time between a mo squito taking an infectious bloodmeal and becoming infected. Models to predict the EIP of malaria parasites are based on assumptions of malaria parasite development in relationship to accumulated degree days above a lower temperature threshold for developm ent (e.g., Detinova, Bertram, & Organizati on, 1962; Moshkovsky, 1946) . For P. falciparum , as temperature decreases the number of days necessary for parasitic development increases (M. T. White et al., 2011) In the classic Detinova model, 16°C is the lowest developmental thresh old for P. falciparum (Detinova et al., 1962) . Two additional meteorological variables besides temperature play an important role in vector population dynamics: precipitation and humidity. The availabilit y of larval habitats is largely influenced by the frequency, duration, and intensity of rainfall. While precipitation works to create or expand breeding sites, rainfall has also been shown to contribute to high larval mortality rates through flooding and s ubsequent flushing out larvae (Krijn P Paaijmans, Wandago, Githeko, & Takken, 2007) . 11 Precipitation may also indirectly influence malaria transmission through its effect on relative humidity. Relative humidity levels are associated with mosquito longevity; higher humidity levels enhance survival (Yaman a & Eltahir, 2013) . Likewise, Anopheles become active shortly after dusk, a behavior associated with higher humidity and minimized risk of desiccation Reece, 2016) . It is important to note the critical distinction between weather and climate as they relate to malaria transmission. Weather is short - term and can have direct, immediate effects on vector distribution and the timing and intensity of outbreaks (Mayer & Pizer, 2011) . Long - term climate conditions influence the spatial and seasonal limits of transmission. 1 .1.5.3 Humans Human immunity and behaviors influence malaria transmission both at the individual and community levels. Humans may possess three types of acquired or adaptive immunity to plasmodia: (1) antidisease immunity, providing protection against clinical disease; (2) an tiparasite immunity, protection against parasitemia; and (3) premonition, the state of maintaining a low - grade and generally asymptomatic parasitemia that protects against new infections (Doolan, Dobaño, & Baird, 2009) . A related point is human genetic resistance to malar ia. For instance, structural variants of hemoglobin and the sickle - cell allele are associated with conferring resistance (Hedrick, 2011) . For those living in malaria endemic regions, most do not experience overt disease as a result of Naturally Acquired Immunity (NAI) to falciparum infection. Exceptions include infants and young children, along with pregnant woman whose NAI is compromised during pregnancy (Doolan et al., 2009) . Human behavior contributes to the epidemiology of malaria substantially; social, political, and cultural factors collectively influence transmis sion, prevention, and control. From a spatial perspective, human movement increases plasmodium dispersal beyond what would be possible for mosquitoes alone (Wesolowski et al., 2012) . Likewise, the movement of people from malaria - endemic to malaria - eradicated areas can lead to resurge nce of disease (Martens & Hall, 2000) . Myriad efforts to 12 prevent and control malaria have e merged over time (Gachelin, Garner, Ferroni, Verhav e, & Opinel, 2018; Packard, 2007) , notably insecticide treated bed nets, antimalarial drugs, and insecticides. An additional vector management approach is environmental modification, specifically the use of engineering or water - management activities to pr event the spread of malaria (i.e., Utzinger, Tozan, & Singer, 2001) . Ironically, environmental modification of landscapes is also associated with the proliferation of vectors and change in malaria disease dynamics through multiple path ways including, land use changes (Paul, Kangalawe, & Mboera, 2018) , deforestation (Yasuoka & Levins, 2007; Yomiko Vittor et al., 2006) , agricultural development (P ackard, 1986) , and, as is considered directly in this body of work, irrigated agriculture. 1.2 Research Design The primary goal of this research is to investigate change in temporal disease dynamics of malaria driven by irrigated agriculture and subsequ ent land use modifications. Given the complexity of this relationship, this work draws on several theoretical frameworks namely, medical geography, ecological and species distribution modeling, the triangle of human ecology, and political ecology. It is th e intention that study outcomes advance the theory of disease systems. The study design for this body of research considers the relationship between land transformation for irrigated agriculture and malaria risk explored at varying scales of analysis: coun try, site, and area. 1.2 .1 Study Country Malawi is a landlocked country in southern Africa with a strongly agrarian society and a long - standing history of chronic food insecurity. From October 2018 to March 2019, FAO ( 2018) estimated that 3.3 million people were food insecure. Simultaneously, Malawi faces a significant malaria burden despite more than a decade of increased intervention implementation in rural area s (Roca - Feltrer et al., 2012; Wils on, Walker, Mzilahowa, Mathanga, & Taylor, 2012) . A total of 4,901,344 confirmed cases of malaria were reported to the WHO in 2017; 3,613 deaths were reported (WHO, 2018 ) . In 13 Malawi, P. falciparum accounts for 100% of infection with An. arabiensis, An. funestus, and An. gambiae as the dominant vectors of transmission (WHO, 2016) . 1.2.2 Study Site The Bwanje Valley Irrigation Scheme is located in the central region of Malawi within the Dedza district and Mtakataka Extension Planning Area ( Figure 1.1 ). Situated along the Lake Malawi -scale irrigation scheme and benefits more than 2,000 smallholder farmers from 14 surrounding villages (GoM, 2015). Figure 1.1 : The Bwanje Valley Irrigation Scheme (BVIS) located in the central region Malawi and the study area, defined as the 8km surrounding BVIS. 1.2. 3 Study Area The study area is defined as the 8km area surrounding BVIS. An 8km value was selected in order to consider the influence of mosquitoes and their immediate progenies developing through the aquatic stage at BVIS, and those produced outside. The flight distance for the Anopheles gambiae s.s. mosquito, one of the primary vectors of the study area, has been documented from less than 1.0km (Costantini et al., 1996) to a maximum flight distance of 1.7km (Thomas, Nall, Cross, & Bøgh, 2013) . Given the 14 uncertainty of flight distance, the decision was made to conservatively estimate the 4.0km surrounding area to BVIS as the area w here mosquitoes and their immediate progenies produced at BVIS would dwell. By doubling the 4km distance to 8km, the study aimed to adequately estimate the land use and land cover besides that attributed to BVIS that may contribute to mosquito development within the area. 1.2. 4 Specific Aims Aim 1: Address LULC decisions and their impact on the spatio - temporal structure of agricultural growth and mosquito development by: 1. Describing LULC of BVIS and the Bwanje Valley and its impact on breeding pool for mation through development of a land classification system for irrigated agriculture in rainy and dry seasons Aim 2: Demonstrate the influence of irrigation schemes for agriculture on mosquito breeding pool formation and persistence by: 1. Modeling breedi ng pool scenarios based on spatio - temporal, environmental and anthropogenic characteristics at BVIS under three scenarios 2. Describing the association between LULC and breeding potential at BVIS contrasted with the 8 - km area surrounding the scheme Aim 3: Assess the impact of dynamic changes in irrigated agriculture on malaria vulnerability in Malawi by: 1. Examining spatio - temporal changes in irrigated agriculture at the national scale 2. Describe habitat suitability for Anopheles gambiae s.s . mosqui toes in Malawi through construction of a habitat suitability model in Google Earth Engine 3 . Addressing the historic and plausible future impact on malaria vulnerability driven by the expansion of irrigated agriculture 15 1.3 Outline of the Dissertation Chap ter 2 literature . Findings suggest that definitions fall into three distinct categories: Interventions, ed in two fashions: as a noun (outcome) and verb (process). A conceptual framework for scaling up is presented that gives greater emphasis on separating the noun scale, from the verb, to scale . Chapter 3 reports the results of a characterization study at BVIS to explore the influence of anthropogenic influences pertinent to breeding distribution are used to generate three spatio - temporal breeding scenarios acros s the scheme. Results illustrate how perturbations to irrigated systems in the form of water availability, water management, and crop cover can influence the distribution of aggregated water bodies and thereby influence disease ecology for the local area. Chapter 4 considers the association between LULC and mosquito breeding potential within the Bwanje Valley Irrigation Scheme, contrasted with the 8km area surrounding the scheme during epidemiology and the impact of LULCC for irrigated agriculture on spatio - temporal disease dynamics of malaria. Chapter 5 addresses the impact of expansion of irrigated agriculture on malaria vulnerability at the national scale in Malawi through examina tion of spatio - temporal changes of irrigated agriculture in relationship to historical malaria prevalence . Chapter 6 summarizes the key findings and future research as a byproduct of this body of work . 16 CHAPTER 2 2.1 Introduction Food security is a complex mix of production and supply constraints as well as access to nutritious food. With a distinctly spatial character, food security solutions often take ungeneralizable forms. Scaling up development interventions within the agricultural sector specifically targeting food security challenges is frequently proposed as a solution to the global hunger crisis (e.g. Linn, 2012)). While the tutions and in Natural Resource Management (NRM) literature, experience has shown that the term lacks ontological agreement (Hartmann & Linn, 2007; Menter, Kaaria, Johnson, & Ashby, 2004; Uvin, 1995). Further, scaling up is often used broadly to refer to a variety of processes (Menter et al., 2004), or occurs concurrently with discussions on innovation, particularly agricultural innovation, and concepts related to spatial diffusion when there are in fact important distinctions. Approaches and viewpoints on scaling exist across a range of disciplines (Wu & Li, 2006). Further, interpretations of scaling are often driven by perspective and perceptual bias (Levin, 1992). One view of scaling is results-based, increasing impact to reach a greater number of people (e.g. Linn, 2012) . How impact is achieved involves additional perspectives on scaling up including the expansion of programs, technologies, or projects from pilot experiences to larger enterprises. To deliver multiplier impacts, scaling up investment too, is critical. An added dimension relates to policy and governance: what is appropriate at one level may not be suitable at another (Veldkamp, Polman, Reinhard, & Slingerland, 2011; Wu & Li, 2006). Spatially-based perspectives often involve the expansion of a c reach (e.g. Noordin, Niang, Jama, & Nyasimi, 2001), or estimating impact at larger scales from small, field or plot sized experiments (e.g. Zhang et al., 2007) . Wigboldus et al. (2016) present a considerable literature review on scaling perspectives exercised Citation Information: Frake, A.N. & Messina, J.P. (2018). Toward a Common Ontology of Scaling Up in Development. Sustainability, 10(3), 835. ht tps://doi.org/10.3390/su10030835 17 through a wide range of approaches, notably: agricultural systems, interdisciplinary, transdisciplinary, research, innovation systems, value chain, landscape, socio - ecological systems, transitions to sustainab ility, and the multi - level perspective on socio - technical transitions. Other relevant perspectives are those held directly by the observers or actors involved in the scaling up process: developers, donors, extension agents, and farmers. In this analysis, w e approach scaling up from a development perspective giving particular consideration to scaling up interventions targeting food security challenges. We accept that while scaling up is not always a positive force, or the only pathway for development (Pitt & Jones, 2016) , effective scaling up of development i nterventions is often cited as a measure of success in reducing food insecurity. Working towards food security solutions involves a wide range of actors embedded within myriad social and environmental systems. Given this inherent complexity, the number of actors, and often necessity for collaboration between development partners to achieve sustainable impacts, a clear, ontological understanding of what scaling up means is essential. Herein, ontological disagreement refers to the varied meanings of scaling up inherent within institutional definitions. Imprecision of definitions across institutions and actors creates ambiguity in defining and measuring outcomes of scaling programs. In turn, uncertainty of scaling up from the onset of development programs cont ributes not only to inflated reports of success, but failure of programs to actually scale as either a product or process. application, we argue for precis ion of definitions where scale is considered both as a function of outcome (noun) and process (verb). To explore the varied meanings of scaling up across institutions, a text analysis is presented of adopted definitions across institutions, pointing to a c onflation of scaling up operating as an Intervention, Mechanism, or Outcome. The article concludes with the introduction 18 outcome or process, along with the ne cessary role of monitoring and evaluation on both innovation and development scaling up efforts. 2 . 2 Scaling In recent years aid agencies have increasingly begun to recognize the the importance of scaling up to achieve widespread impacts (e.g. Hartmann et al., 2013) . For instance, the International Fund for Agricultural Development (IFAD) has decl embedding it throughout corporate strategy (Hartmann et al., 2013) . Given the signifi cance of scaling 2 .3 What is Scale? definitions. Even in Geograp hy, a discipline where scale is intrinsic to all inquiry, across the sub disciplines meanings vary (Sheppard & McMaster, 2008) . W hen describing scale, Goodchild (Goodchild, 2001) but four distinct meanings of scale: level of spatial detail, representative fraction, spatial extent, and process scale. Likewise, scale is applied in varying contexts including geographic, temporal, and spectral (Goodchild, 2001; Quattrochi & Goodchild, 1997) . Further, Smelser and Baltes, (Smelser & Baltes, 2001) contend that in science, scale takes three distinct forms: cartographic, analysis, and phenomenon scale. Cartographic scale refer relationship to actual size. Analysis scale is the extent of a given study area, and phenomenon scale describes the size at which human or physical structures or processes exist (Smelser & Baltes, 2001) . For geographers, scale in diffusion research typically is conducted as functional, which reflects decisions made by varying aggregatio ns of individuals, or spatial, directly reflecting the manifestations of these decisions within a spatial context (Brown, 1981) . Beer (1968) describes the differentiation between scales, regularly conducted in scientific investigation, as the Cones of Resolution problem. Beerian Cones of Resolution examine spatial processes beginning at the micro scale and gradually 19 work up towards a larger, macro scale perspective. Beer argues that since complex systems comprise a wide variety of subsystems, each operating with their own distinct attributes, a pivotal mark in scientific research is identifying meaningful scales of analysis to properly address the research question at hand. Extending this idea, Manson (Manson, 2008) questions whether a single definition of scale actually exists. Rather, he presents a scale continuum for human-environment systems to assist in framing researching methodologies. Given the existence of such a continuum, rather than searching for one widely accepted theory of scale it becomes prudent to understand of how epistemological contexts work to define scale. 2.4 Scaling Up There exists a long literature on the uncertainty of scaling terminology. In April 2000, participants of the Consultative Group on International Agricultural Research (CGIAR)-nongovernmental organization (NGO) committee met in the Philippines and defined the objective of scaling up as, (Menter et al., 2004). As illustrated by Menter et al. (Menter et al., 2004) , a number of issues surround this definition beginning with it defining the objective rather than the definition of scaling up itself. According to this definition, scaling up reflects two critical, impact- centered factors: extent and quality (Menter et al., 2004). Menter et al. (2004) go on to introduce (Gonsalves, 2000). Vertical scaling up is institutional, it involves institutions accepting and internalizing the fundamental principles of an innovation and allowing them to guide practice. Horizontal scaling up refers to geographical spread, whereby more people are impacted through replication and adaptation (Menter et al., 2004). Uvin ( 1995) provides a rich literature review that engenders consideration of not only the variety of definitions of the term across the literature, but a suggestion that there are several forms of 20 scaling up: Quantitative, Functional, Political, and Organizational. Yet, Menter et al. (2004) highlight that this reference to scaling up as a catch-all term for a variety of processes has in part led to the confusion of its meaning. Adding to the conversation, Hartmann and Linn (2008) in working to develop a framework for effective scaling up of development interventions, begin their analysis by (2008) proposed definition of scaling up, adapted from the one used by the World Bank (Mundial, 2004) considers quality of impact, scale, and sustainability across projects, programs, and policies. Further, Wigboldus and Brouwers (2016) posit whether the term scaling itself is in fact the appropriate terminology where alternative verbs or their related nouns, including institutionalism, mainstreaming, expansion, or spreading would provide more clarity. 2.4.1 Pathways to Scaling Beyond the varying interpretations of scaling up are the number of interpretations of how direct, positive impacts in the agricultural sector can be effectively scaled. The historical approach to scaling methods was top-down: researchers influencing extension agents who then directed farmers on the adoption of new practices. Over time this model evolved to one that emphasizes a more circular flow of information between each of the parties involved in the extension process (Chester, 2005). This new method allowed for a feedback mechanism for each of the stakeholders and actors involved, facilitating greater communication with one another. Since 2003, a number of new efforts have Driven Development (CDD) Plan (Chester, 2005). The CDD model highlighted both strengths and priorities for taking programs to scale through three distinct stages: initiation, scaling up, and consolidation (Binswanger & Aiyar, 2003). Extending the ideas of CDD, the World Agroforestry Center (ICRAF) and the Food and Agriculture Organization (FAO) later worked to build scaling up 21 initiatives where farmers became central figures in the extension process, and where education institutions and service providers were tasked with the role of meeting farmers (Chester, 2005) . Understanding t he need for implementers to be provided with a mechanism for achieving scale, the Academy for Educational Development developed the System - wide Collaborative Action for Livelihoods and the Environment (SCALE) program. SCALE emphasized the necessity for inc reasing the number of individual and group stakeholders along with the linkages between them (Chester, 2005) . Today, two of the most common approaches to scaling up within the international development community are the Management Systems International (MSI) Framework and IFAD Framework (Cooley & Linn, 2014) . The MSI Framework provides practitioners with specific tools geared at designing effective management frameworks. In contrast, the IFAD framework aims at providing high - level policy and direction on scaling up (Cooley & Linn, 2014) . Cooley and Linn (2014) present a review of both frameworks. Most recently, a five - fold strategy for achieving impact at scale was presented by CGIAR as a part of the CGIAR Strategy and Results Framework 2016 - 2030 (CGIAR, 2015) . This strategy includes 1) Deliberate prioritization of research efforts; 2) Close alignment of efforts by centers and center research programs in selec ted areas; 3) Coordinated planning with implementation partners; 4) Commitments from clients and national partners to make complementary investments and policy reforms where CGIAR is investing; and 5) Institutionalization of a culture of regular monitoring and evaluation. 2 .5 A Visual Analysis of Scaling Up To explore the varying interpretations of scaling up in development literature, definitions (when available) from the 15 CGIAR Centers, United States Agency for International Development (USAID), and IF text. Definitions included in this analysis from the CG Consortium were from the following 22 institutions: International Center for Agricultural Research in the Dry Areas (ICARDA), International Center for Tropical Agriculture (CIAT), International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), International Food Policy Research Institute (IFPRI), International Livestock Research Institute (ILRI), International Maize and Wheat Improvement Center (CIMMYT), and World Agroforestry Center (ICRAF). , all definitions were included in display was set to twenty-five; minimum frequency was set to 1. The result is presented in Figure 2.1 where for each word in the word cloud, its size is proportional to the frequency it is found in The word cloud contained twenty-four terms. One te most frequently applied term, followed by nine terms of relative importance given their similar size: Figure 2. 1: Visual analysis of 'Scaling Up' definitions across CG Consortium, USAID, and IFAD . 23 terms in the remainder of our analysis. Across the institutions considered in this study, analysis of the ten words most commonly used in the definitions of sc aling up reveal a distinct categorization of terms emphasizing individually or a combination of, Interventions, Mechanisms, or Outcomes. We arrived at this categorization by considering the terms through the lens of the formalized components to scaling: in novation, scaling, and monitoring an evaluation; or more directly, what is being scaled, how is scaling occurring, and the end goal of developing sustai nable food security solutions? Interventions often take the form of new or adapted, existing innovations. Here, definitions across institutions point to technologies, policies, or programs; interventions which are suspected to have long - standing, positive impacts. The up means expanding, replicating, adapting, and sustaining successful policies, programs, or projects to ( Linn, 2012) . The United States Agency for International Development (USAID) echoes this sentiment in their emphasis on scaling referring to access and ing up] means that more poor farmers (Linn, 2014) . It is worth noting the profound difference between an explicit measure of numbers and broadly defined access. Finally, IFAD, in choosing to define scaling up adopts similar verbiage to IFPRI where scaling again means, nd sustaining successful policies, programs, or projects in (IFAD, 2011) . Mechanisms refers oth on and accomplishing a scaling up program. Uvin (1995) introduced the concept of scalin g as a 24 (Menter et al., 2004) dissention, is a term still frequently applied in the literature. The International Livestock Research Institute (ILRI) gives particular attention to (Hendrickx, Ballantyne, Duncan, Teufel, & Ravichandran, 2015) . Expanding on the idea of scaling up as a process, most recently Wigboldus and Brouwers (2016) moved beyond the processes described by Uvin to further include scaling up as quantitative, spatial, kinematic, or physical. The advantage of calling scaling up a process is that it is binary - a yes or no proposition. From a development perspective you can declare success by meeting an indicator of process rather than outcome. It sets a low bar. An added dimension is management. Writing for the International Center for Tropical Agricul Beyond the Interventions and Mechanisms are the Outcomes of scaling up: what is the end result ? Here we come to the most often applied term in the description of scaling up, sustainable. Closely related is impact. Both terms not only apply to end results but to intentions. The World Agroforestry Center (ICRAF) places particular emphasis on outcomes (Simons et al., 2011) . For many, at the center of scaling up development int erventions is the idea of reducing food security challenges for smallholder farmers; impactful programs are successful at working towards or meeting this end goal. Likewise, sustainable programs are those that work at maintaining success in reducing the bu rden of hunger over time. A related point to consider is resiliency; the ability to recover after disaster or unforeseen circumstances. The mission of the CG Consortium emphasizes improvement in resiliency directly, however when considering the visual anal ysis of scaling definitions presented, it is worth noting that necessary to meeting long term food security challenges. These two terms are not interchang eable, 25 nor should the idea of resiliency be absorbed by sustainability entirely; the approaches to sustainable versus resilient programs and their scaling can have distinct differences. Many of the definitions included in this analysis emphasize not one, b ut a combination of Interventions, Mechanisms, or Outcomes in their definition of scaling up, leading to ambiguity in their agricultural technologies and Outcomes: that more poor farmers have access to and effectively utilize aforementioned technologies. The International Maize and Wheat Improvement Center (CIMMYT) implement succ essful interventions and expand, adapt, and sustain them in different ways over time (Ro . The International Food Policy Research replicating, adapting, and sustaining successful policies, programs, or projects to reach a greater number of people (Linn, 2012 - is scaling up about policies, programs, or projects, their expansion, replication, or adaptation, or their sustainability in reaching a wider audience? 2 . 6 Discussion literature. Yet, as several studies have highlighted, they lack ontological agreement. Uvin (1995 , p. 928) We disagree as ultimately, the a is operationalized (ie: the intervention and/or the pathways to scaling), influences M&E, and affects program success. Likewise, how scaling up is defined will influence how funding is a llocated, and in turn how project development progresses - which projects are made a priority and which are neglected. Given the varying definitions and interpretations of scaling up, Wigboldus and Leeuwis (2013 , p.6) 26 of scaling nature of scaling up the authors contend does little to aid in knowing what scaling measures may best apply in a particular situation. As we h ave highlighted in this paper, how scaling up is defined is rooted in the interpretation of scale itself. When considering the categorization of terms presented in our earlier analysis (Interventions, Mechanisms, Outcomes), it becomes apparent that scale i s applied in two forms: as a verb and a noun. Where scale functions as a verb is demonstrated directly in the mechanics of scaling up with emphasis on adaptation and expansion of innovations. Innovations are a critical component to achieving widespread imp acts in reducing food insecurity, either through new innovation or scaling up existing, successful innovations. As a noun, an innovation taken to scale implies meeting a specified end goal; project outcomes. Likewise, in the evaluation of project results i n relation to its predefined, intended outcomes that were determined at the onset. Existing definitions of scaling up often conflate ideas related to interventions (innovations), the process of their expansion or replication, or their intended outcome; we contend this occurs in part due to the application of scale both as a noun and verb. To illustrate the application of scale as either (or both) a noun and verb , we return to IFPRI, Expanding, replicating , adapting, and sustaining successful policies, programs, or projects to reach a greater number of people - IFPRI (Linn, 2012 , p.46) USAID (Linn, 2014) increasing the size, amount, or importance of something, usually an or ganization or process ILRI (Hendrickx et al., 2015) Bring more benefits to more people, more quickly and more lastingly ICRAF (Simons et al., 2011) 27 To separate scaling up and its associated actions based on the process (verb) and outcomes (noun), we present a conceptual framework for defining scaling up and putting it into practice for development activities (Figure 2.2). Further, we emphasize the necessary role of Monitoring and Evaluation (M&E) on both the innovation and scaling up efforts. Figure 2.2: A framework for conceptually defining scaling and putting it into practice for development activities During the initial stage, scale functions as a noun: a specific purpose or program goal is agreed upon, followed by defining its measures of success and the overall timeframe for the project or program completion. Measures of success are quantifiable outcomes of the innovation or scaling that 28 serve to evaluate program performance. How to achieve these outcomes varies by either developing a new innovation or scaling an existing innovation. Here too, the question of whether scaling an existing innovation is even appropriate to the context should be addressed (Hartmann & Linn, 2007) . It is at this stage that scale begins to functio n as a verb. After development, new innovations follow a pathway through piloting, monitoring and evaluation, then, adaptation. It should be noted that it is often this necessity for adaptation, coupled with imprecise definitions of scaling up that work to declare every program a success. Scaling up existing innovations starts with careful consideration of which type of scaling is appropriate to meet the end goal. Vertical scaling up is institutional; innovations are meant to guide principles of practice. Horizontal scaling up is geographic, where the spatial reach of an innovation expands. Scaling up efforts do not occur in isolation; Vertical scaling up efforts spillover to geographic diffusion across space, likewise geographic expansion can influence up take of institutional practice. Often scaling up requires an integrated approach (Hartmann & Linn, 2007) . Giving consideration to an integral pathway aims at sc aling up along both an institutional and geographic pathway from the onset. An integrated approach can be either sequential or simultaneous depending on the context. Regular M&E of scaling up pathways provides important feedback and creates opportunities for adaptation from onset to completion. Rather than evaluating what went right or wrong at the conclusion of a project, effective M&E strategies attempt to gauge performance through a series of (Crawford & Bryce, 2003) . Monitoring and Evaluation is a requirement for effective scaling up (Mansuri & Rao, 2004) . Yet historically, scaling up efforts and evaluation were often viewed as conflicting objectives for most international development agencies (Duflo, 2004) , despite the value of reliable program evaluations at every stage of scaling up efforts. In this model, the M&E process occurs both for new and existing innovations; likewise, on the scaling up efforts to bring an innovation to scale. M&E specific to scaling up efforts is conducted during 29 several stages: 1) The choice of existing innovation to be brought to scale; 2) The type of scaling up pathway selected- vertical, horizontal, or integrated; 3) On horizontal, vertical, or integral scaling up efforts; and 4) On the declaration of success or choice to adapt. The final stage returns to scale, the noun. It is at this point that measurement between the stated indicators for success and actual outcomes are evaluated. Here, overall program performance is analyzed, and through careful consideration of program shortcomings, new initiatives developed to meet remaining unmet needs of communities. 2.7 Conclusions It is trite to call for improved definitions, particularly given the outcomes of this paper; here, we argue for the careful consideration of the precision of definitions used. The imprecision of definitions, in part the product of uncertainty, contributes not only to the reported regular success of development programs, but also the failure of these programs to scale as both product and process. Literature is replete with examples of attempts to address the scaling up debate, many of which are highlighted herein. Often these discussions include an attempt to redefine or reinvent the terminology to better describe the meaning of scaling up to fit a particular development program rather than stressing the precision of terms already in use. Yet, regular redefinition only leads to the perpetuation of uncertainty particularly where adoption of these improved definitions are asynchronous across institutions. Given the importance of scaling up to development, ontological agreement of scaling up across institutions is vital not only to measurement but to meeting the needs addressed through sustainable development interventions. Ontological ambiguity devalues scaling up; by defining a clear pathway for success, value judgements regarding development as outcome can be examined. Our analysis on the interpretations of scaling up showed that across the different agencies considered, definitions were dissimilar despite some commonalities in etiology and occasionally authorship. The issue here is not alternative phrasing, but rather the lack of ontological agreement 30 among definitions. The categorization of terms highlighted when analyzing the descriptions of scaling up are a byproduct of the varied meanings behind the definitions themselves. In some cases, emphasis is placed primarily on the innovation being scaled (Interventions). Other interpretations give priority to the structure of the scaling up process itself i ncluding institutional or geographical expansion (Mechanisms). Still others underscore the end results (Outcomes) and notably the sustainability, with strikingly no mention of resiliency, of the product or process brought to scale. Finally, many definition s in our analysis revealed an emphasis on not one, but a combination of Interventions, Mechanisms, or Outcomes, leading to further ambiguity. In light of the literature and above categorization of terms, we contend that the continued uncertainty of scali ng up is in part often related to the conflation of the noun, scale, and verb, to scale. Interpretation of scaling up that stresses product or process success, or outcomes is a function of the noun, scale. By contrast, where scaling up emphases reside in a daptation, expansion, geographical spread, or process, these descriptions are rooted in the meaning of the verb, to scale. Working to emphasize and separate the critical functions of scaling, we have presented a conceptual framework that operates in three stages: 1) Defining objectives and creating indicators; 2) Scaling efforts either by a new or existing innovation; and 3) Final measurement of outcomes. The novelty of this model lies in both its separation of scaling up and its associated actions based on the process (verb) and outcomes (noun). Where scaling up has traditionally occurred irrespective of M&E, our model works to showcase the critical role of M&E for both the innovation and development scaling up efforts. Scaling up and M&E are inextricably l inked. Given this relationship and the evolution of M&E, future work could consider critically analyzing the variation in meaning of scale, by institution over time. Further, presenting commonalities and discrepancies in meaning between institutions and co nsideration on how these conceptions of scale influence development interventions. 31 Scaling up product or processes in targeting food security challenges is a vital component to developing sustainable solutions to the global hunger crisis across geographical scales. As such, a consensus on the ontological meaning of scaling up across institutions working towards these solutions is critical. Our aim in developing this model is to engender further consideration on the precision of scaling up terminology when working to bring a product or process to scale. Uncertainty on the meaning of scaling up should not be a barrier to meeting the critical needs being addressed through development interventions. Where there is ontological agreement on scaling up within and across R&D and NRM institutions, there is a higher likelihood for project success. 32 CHAPTER 3 3 .1 Introduction The association between irrigation, mosquito production, and malaria transmission is well documented in the literature (see e.g. Ijumba & Lindsay, 2001; Kibret et al., 2010) However, irrigated schemes are treated as homogenous spatio - temporal units with little consideration for how breeding potential varies across the space and time. Irrigated schemes may change seasonally in crop production and distribution and not every irrigated scheme receives sufficient water resources to operate on an annual, but rather only seasonal basis. In addition, some land covers found in irrigated schemes are not agricultural, but engineered structures including concrete canal networks. The heterogeneity of irrigated schemes spatio - temporal distribution r esults in asymmetrical breeding risk. The structure of heterogeneity is the product of environmental and anthropogenic factors, including, but not limited to, soil type, timing and intensity of irrigation, drainage, and crops type(s) being cultivated. Each of these factors, independently and in combination, influence the amount and duration of pooled surface water available to mosquitoes for breeding. This chapter reports the results of a characterization study at the Bwanje Valley Irrigation Scheme (BVIS) formation. Environmental characteristics and anthropogenic influences pertinent to breeding distribution are used to generate three spatio - temporal breeding scenarios across th e scheme. Results illustrate how perturbations to irrigated systems in the form of water availability, water management, and crop cover can influence the distribution of aggregated water bodies and thereby influence disease ecology for the local area. 3 .2 Methods 3 .2.1 Study Site The Bwanje Valley Irrigation Scheme is an example of a typical smallholder irrigation schem e in Malawi. Established in 2000 through cooperation of the Japanese and Malawi governments, the BVIS 33 is a gravity - fed scheme located in the Shire watershed of central Malawi where annual rainfall is approximately 867mm and temperatures range from 17.2°C to 27.3°C (Adhikari & Nejadhashemi, 2016) . Spanning 800 hectares, the scheme diverts water from the Namikokwe River to ~30X30 meter agricultural plots via a series of concrete, main and branching canals, and earthen, tertiary canals. Grant a id of $15 million USD was provided by Japan, mediated through the Japanese International Cooperation Agency (JICA) (Veldwisch, Bolding, & Wester, 2009) . During the 2016 growing season, 2067 farmers participated at BVIS from 14 surrounding villages; 1089 farmers were women (M. Tarsizio, personal communication, March 17, 2016) . Veldwisch et al. (2009 ) provide a comprehensive 3 .2.2 Study Design Land use and land cover decisions and their impact on the spatio - temporal structure of agricultural growth and mosquito development were assessed using four sources of information: (1) Satellite imagery from the SPOT - 6 sensor at two time periods; (2) Spatial structure and land cover information s alley smallho (JICA, 1994) and (4) Onsite interviews with BVIS scheme personnel conducted in 2016, 2017, and 2019. The 2016 interview was conducted with the scheme manager, who was often present during field surveys to provide insight and commentary. In 2017, an interview wi th the BVIS board chairman was performed in the absence of a permanent scheme manager. The 2019 interview was conducted with the former BVIS manager interviewed in 2016. For survey questions see (Frake, 2019a, 2019b, 2019c) . The con ceptual framework for this study is presented in Figure 3.1 . 34 Figure 3.1: Conceptual framework for the study. Soils data are taken from JICA (1994) 35 3.2.3 Environmental Characteristics 3.2.3.1 Soils There are five primary soils at BVIS whose parent materials are fluvial, colluvial, and/or lacustrine sediments. Top soil (0-30cm) types are sandy loam, sand clay loam, and sandy clay loam to clay. A notable consideration for our study is the drainage capacity for each soil. The dominant soil units at BVIS are A1f5, A1f2, and A1f4 . Data from JICA (1994) does not include specific bulk density measurements (i.e. indicators of soil compaction) for each soil type. However, drainage and ponding potential are available and were utilized for this study. Soil units at BVIS are typified by poor to imperfect drainage, and moderate to severe ponding potential (Figure 3.2). Figure 3. 2: Soil types at BVIS. Data and soil characteristics are taken from JICA (1994) 36 3 .2.3.2 Land Cover 3 .2.3.2.1 Data preparation Multispectral SPOT - 6 images of the study area were acquired on December 4, 2013 and April 1, 2014 at a spatial resolution of 6.0m for four spectral bands: Blue (0.455 m 0.525 m), Green (0.530 m 0.590 m ), Red (0.625 m 0.695 m), and Near - Infrared (0.760 m 0.890 m). Geometric correction to ensure positional accuracy of the imagery was performed using ground control points prior to analysis and a 30m DEM from the Shuttle Radar Topography Mission (SR TM). Sensor calibration and conversion of digital numbers (DNs) to radiance was performed following a solar correction to top of atmosphere (TOA) refle ctance using . Image classification was performed using an unsupervised ISODATA class ification where 255 classes were selected at .98 convergence (see e.g, Messina et al. 1998). Signature evaluation was conducted using the Transformed Divergence measure with separability markers of >1975 deemed acceptable. Using the edited signature set, a Maximum Likelihood supervised classification was Field sample sites were selected by stratified random sampling of land cover classes from the supervised classifi cation for each image: 11 classes in the dry season and 9 in the rainy season at BVIS. A total of 242 points were selected for sampling during the dry season and 72 during the wet season. Fewer sample sites were selected during the rainy season given that there were fewer classes to sample and previous fieldwork had shown that rice was the predominant crop grown across the scheme. 3 .2.3.2.2 Land Cover Data Collection Field based surveys of LULC were carried out during the 2016 dry season and 2017 rainy s eason. A total of 235 dry season samples were collected over a fourteen - day period in mid - August 2016 by a field team of three researchers. These samples included 185 of the sites selected from the random with classification . Rainy season samples included 68 sample sites and 36 accuracy assessment sites 37 sampled over a nine-day period in early April 2017 by a field team of two researchers. At each site, th e Monterra GPS unit. In addition, field notes describing the LULC at each sample site along with descriptions of LULC in all directions were recorded. Where land cover of surrounding areas differed from that recorded at the sample site, the direction of these areas along with approximate distances were included within field notes. Field note transcriptions and associated GPS data collected at each sample site were combined for classification analysis. 3.2.3.2.3 Land Cover Classification Classification of LULC followed a two-step process: (1) Sites were categorized by land use, land cover type, and feature (e.g. maize, rice, bare earth); (2) Field notes were used to include information on stage of agricultural growth, appearance of soils, density of plantings, presence of water within irrigation canals, and locations of trees relative to agricultural growth for each feature identified in step 1.Descriptions of LULC including the number of samples within each class are presented in Tables 3.1 and 3.2. Point shapefiles of all sample sites, by season, and their respective LULC descriptions imagery. 3.2.3.2.3.1 BVIS: Rainy Season Maximum Likelihood classification for the rainy season yielded 9 classes of land cover, of which 7 - 3.3). The number of supervised classes a sampling. While rice plants in some plots were in panicle formation stage, other plants had begun flowering, or were mature. Rice on other plots had been harvested, releasing the plots to renewed planting. The Non-Agricultural class included irrigation canals, concrete and earthen, along with roads and pathways. One particular challenge of the rainy season classification was presence of maize grown in single or double rows, in the narrow spaces immediately adjacent to irrigation canals. These areas are 38 situated at a slightly higher elevation (~2ft) than the plots, owing to the gravity-fed design of the scheme. The proximity of maize in relationship to the structures within - the cause for the mixing of pixels observed in areas immediately adjacent to roadways and irrigation canals. The combination of water availability from seasonal rains and the Namikokwe River during the rainy season allows for the predominant cultivation of rice at BVIS occurring across 87.9% of the scheme. The primary rice varieties are kilombero, a long-grain aromatic rice, and faya (M. Mafosha, personal communication, March 27, 2019). BVIS personnel direct farmers to begin planting in mid- December, with harvesting around mid-February (M. Mafosha, personal communication, August 19, 2016). However, the varied stages of rice growth observed during land cover analysis from this study suggests that farmers are able to practice Besides growing in areas immediately adjacent to irrigation canals, two concentrated areas of maize are cultivated during the rainy season along the southern boundary of the scheme. Rehabilitation plans for BVIS were prepared in 2005 by JICA and included fine leveling for plots (JICA, 2005; Veldwisch et al., 2009). However, according to our interviews with BVIS management, the plots in these areas were not leveled as well as in other areas. As a result, these areas are situated at a slightly higher elevation than those immediately surrounding them preventing efficient water flow and cultivation of rice (M. Mafosha, personal communication, March 27, 2019). 39 Table 3.1 : Rainy season land cover class descriptions for BVIS BVIS TOTAL Class Descriptions Sample AA I.AGRICULTURAL LAND I.ACTIVE Rice Flooded Field 2 0 2 Flooded Field w/ Dried Grasses 2 0 2 Early Growth w/ Visible Water & Grasses 2 1 3 Early Growth w/ Visible Light Standing Water 1 0 1 Early Growth w/ Visible Dark Standing Water 0 2 2 Mid-Growth w/ Visible Light Standing Water 1 0 1 Mid-Growth w/ Visible Dark Standing Water 0 2 2 Panicle 14 8 22 Panicle w/ Dried Grasses 3 1 4 Panicle & Grasses: Tal l 8 1 9 Varied Growth Stages 2 1 3 Maize Mature 7 0 7 Past Maturity 5 2 7 Sweet Potato Mature 1 1 2 Young 2 0 2 Sweet Potato & Grasses 0 1 1 Sweet Potato & Emergent Shrubbery 1 0 1 Pumpkin Intermediate Growth Stage & Grasses 1 0 1 Mixed Cropping Sweet Potato: Young & Panicle Rice 1 1 1 Sweet Potato: Mature, Maize: Mature, & Rice: Panicle 0 1 1 Rice: Panicle & Maize: Mature 3 2 5 II.FALLOW Grasses Tall 3 1 4 Tall and Standing Water 1 1 2 Human Influence Road 1 0 1 Straw Structure on Bare Soil 1 1 2 III.IRRIGATION INFRASTRUCTURE Concrete Canal w/ Flowing Water, Bare Earth & Dried Grasses 1 0 1 Concrete Canal w/ Flowing Water & Emergent Grasses 2 1 2 Concrete Canal w/ Flowing Water, Emerg ent Grasses & Maize 0 1 1 Concrete Canal w/ Flowing Water, Emergent Grasses & Dried Grasses 0 3 3 Concrete Bridge Over Dirt Canal w/ Emergent Grasses 0 1 1 Dirt Canal w/ Standing Water & Dried Grasses 0 1 1 Dirt Canal w/ Standing Water & Panicle Rice 0 1 1 IV.TREES Banana Trees 1 1 2 Green foliage on flooded rice plot 1 0 1 Trees: Roadway Adjacent 1 0 1 40 Table 3. 2 : Dry season land cover class descriptions for BVIS BVIS TOTAL Class Descriptions Sample AA I. AGRICULTURAL LAND I. ACTIVE Maize Mature 9 7 16 Young 4 0 4 Young & Dense Shrubbery 3 0 3 Beans Mature 2 2 4 Mature & Emergent Weeds 2 0 2 Young 5 0 5 Cowpea Mature 2 1 3 Cowpea & Shrubbery 2 1 3 Young 3 0 3 General Agriculture Intercrop: Young Maize & Beans 1 1 2 Mature Maize & Beans 1 0 1 Mustard Greens 0 1 1 II. FALLOW Bare Earth Dark Soil 10 0 10 Dark Soil w/ Straw 12 2 14 Dark Soil w/ Straw & Ridging 0 2 2 Light Soil 2 1 3 Li ght Soil w/ Straw & Ridging 3 0 3 Medium Soil w/ Straw & Ridging 3 0 3 Charred Ground Charred Ground w/ Beans 1 0 1 Charred Ground w/ Cowpea 2 0 2 Charred Ground w/ Emergent Weeds 3 2 5 Charred Ground w/ Shrubbery 1 0 1 Charred Ground on B are Earth 1 4 5 Dried Fields Dried Grass 4 1 5 Dried Grass & Emergent c3 Vegetation 6 0 6 Vertical Rice Straw w/ Little Vegetation 4 0 4 Vertical Rice Straw w/ Emergent Weeds 7 10 17 Vertical Rice Straw w/ Green Tops 1 1 2 Horizontal Rice Straw w/ Emergent Weeds 30 0 30 Straw & Dried Grass 6 2 8 Shrubbery Dense 28 5 33 Sparse 14 2 16 Dense & Emergent Red Weeds 2 2 4 Blackjack w/ Emergent Weeds 1 1 2 Dense Shrubbery w/ Medium Soil & Straw 1 0 1 III. AGRICULTURAL INFRASTRU CTURE Irrigation Canal Concrete 1 2 3 Dirt 0 2 2 Tertiary canal w/ Shrubbery 1 0 1 41 Table 3.2 (con ) Human Influence Road 0 3 3 IV.TREES Banana Trees & Dried Grass 2 0 2 Banana Trees & Dense Shrubbery 0 1 1 Green Foliage on Ac tive Agricultural Land 1 0 1 Green Foliage on Fallow Field 3 1 4 Green Foliage on Fallow Land w/ Green Vegetation 0 1 1 42 Figure 3.3 : Rainy season land cover classification for BVIS including field photos depicting varied stages of rice growth within the scheme. 43 3 .2.3.2.3.2 BVIS: Dry Season To assist in differentiating between active and fallow vegetation in the dry season at BVIS, the Normalized Difference Vegetation Index (NDVI) was used to assess agricultural growth. T he NDVI index is sensitive to live, green plants, elucidating difference between the near - infrared and visible red - infrared light (NASA, 2000) . Of the 207 sample sites across BVIS, we sampled active agriculture at 46. These 46 points were combined with the NDVI scene, their range of values (0.14 to 0.48) assessed, then, visually inspected for variations of NDVI in relation to the supervised land cover classes. This analysis in combination with consideration of feature descrip tions of each sample point determined meaningful class assignment of LULC. Developing a classification system for the dry season presented a number of challenges: (1) Cultivation was widely dispersed throughout the scheme; (2) Bare fields spectrally resem bled earthen roadways and tertiary canals; and (3) Maize and other unmanaged grasses were observed growing immediately adjacent to and often overhanging irrigation canals. The maximum likelihood classification analysis showed eleven significant classes of land cover of which, 5 were classified as - Dry season land cover was also more diverse than initially expected. The appearance of fallow fields varied significantly, attributable t o farmer decision - making on the method of field clearing: cha - roadways, concrete and earthen canals. However, pixels in t his class are also located within agricultural fields. These mixed pixel effects are attributable to the complexity of mapping land use versus land cover 44 in an irrigated scheme. In this case, even within an over- classif ication, a portion of the land functions as agricultural plots while the remainder serves as infrastructure for agricultural growth. Dirt roadways and earthen canals at BVIS are constructed from native soils. Thus, spectrally they resemble bare fields where farmers have cleared their plots post- harvest. cases, intercropping of maize and beans. The majority of observations of maize sampled were mature (72%). Where young and mature. Additional information on the specific crop varieties and cropping calendar for these varieties would assist in further characterizing the spatio-tempo ral structure of agricultural growth throughout the scheme during the dry season. The extent of dry season cultivation at BVIS is determined by water availability; typical cultivation occurs over ~300 hectares (ha) (M. Tarsizio, personal communication, March 17, 2016). The crop types and hectarage cultivated during typical dry seasons are: Maize (200-250ha), Beans (50- 60ha), Cowpea (5-10ha), and Soya (50ha) (M. Tarsizio, personal communication, March 17, 2016). Management reported that water scarcity in 2016 limited cultivation to 180ha: 50ha were allocated to maize, 120ha to cowpea, and 10ha to beans. The decision to forego growing soya and expand cowpea cultivation during the 2016 season was in an effort to mitigate the impacts of drought (M. Tarsizio, personal communication, March 17, 2016). The intended areas of the scheme for cultivation of cowpea, maize, and beans by scheme management for the 2016 dry season are presented in Figure 3.5. However, field sampling showed that actual occurrence of these crops varied considerably, also revealed sporadic, and rare, plantings of sweet potato along the southern border of the scheme in plots not allocated or under the governance of the board. These findings have considerable impact 45 as the intended spatial arrangement of crop types by the farmer cooperative directly influences irrigation distribution and scheduling throughout the scheme during the dry season. 3. 2.3.2.4 Validation of Classification Results 10.5.1. Confusion matrices are a cross tabulation procedure that reflect the agreement between the produced land cover raster and ground truth (Foody, 2002) . Test pixels were evenly distributed across the study areas. The number of test pixels for each class were selected to ensure that at leas t 10 - times the number of test pixels were selected per class, as there were classes (see e.g. Jensen, 2005) . Four kappa coefficient; and (4) over all accuracy. multivariate technique, commonly used for asses estimation of the agreement between the classification map and reference (test pixels) data (Jensen, 2005) as: Landis and Koch (1977) consider values >.80 as strong agreement, values between .40 and .80 as m classification, an indication of satisfactory agreement between classified imagery and reality. The dry rong agreement. The overall accuracy is the product of the correctly classified test pixels in each class, divided by the total number of test pixels. The overall accuracy for the rainy season classification is 80%; the dry season classification is 93%. R esults of classification accuracy assessments are summarized in tables 3.3 and 3.4. (2) 46 Figure 3.4 : Dry season land cover classification for BVIS including field photos of varied land covers surveyed. 47 Figure 3.5 : Variations between BVIS farmer cooperative intended distribution of crop types in relationship to where crops were sampled during field surveys. Table 3.3: Rainy season classification accuracy assessment. P.ac., pr accuracy; E.O., omission errors; C.O., commission errors Rice Maize Non-Agricultural Total P.ac (%) O.E. (%) Rice 29 3 4 36 96.6 3.3 Maize 1 21 4 26 70 19.23 Non-Agricultural 0 6 22 28 73.3 26.6 Total 30 30 30 90 U.ac (%) 80.5 80.76 78.57 C.O. (%) 19.4 19.23 21.42 Overall Accuracy .80 Kappa .70 48 Table 3.4: Dry season classification accuracy assessment. P.ac, p s accuracy; E.O., omission errors; C.O., commission errors Fallow Vegetated Non-Vegetated Total P.ac (%) O.E. (%) Fallow 38 0 4 42 95 5 Vegetated 0 37 0 37 92.5 7.5 Bare Earth 2 3 36 41 90 10 Total 40 40 40 120 U.ac (%) 90.5 100 87.8 C.O. (%) 9.5 0 12.2 Overall Ac curacy 93% Kappa .88 3.2.4 Anthropogenic Influence 3.2.4.1 BVIS Structure and Irrigation ing canals, and water distribution control structures were mapped throughout the scheme. Throughout the literature, the actual perimeter of BVIS is ambiguous, particularly in the southeastern portion of the scheme (e.g. Chidanti- Malunga, 2009; JICA, n.d., 1994, 2005; Johnstone, 2011). Analysis was initially based off a perimeter georeferenced from (JICA, 2005) and verified through visual inspection of Google Earth Pro v. 7.3.1 imagery. However, it became increasingly apparent during field surveys conducted in 2016 that this d in March 2017. The perimeter was mapped by walking along the operational boundary of the scheme with GPS -second intervals. Surveying took approximately 2.5 days. T accuracy, a member of BVIS management was present for the duration of the survey. Results of this survey in relationship to the original georeferenced perimeter are presented in Figure 3.6. 49 The Bwanje Valley Irrigation Sc heme is water from the Namikokwe River and channels it through a series of main, branching, and tertiary canals. The main intake weir has a maximum discharge rate of 1.14 m ³ /s (JICA, 2005) . Since the scheme was established in 1999, irrigation canals have undergone significant rehabilitation due in part to catastrophic flooding in 2002 and 2003 (Veldwisch et al., 2009) . Prior to rehabilitation the scheme was serviced by 1 main and 3 branching canals (Johnstone, 2011) . At present, two main canals, six branching canals and 132 tertiary cana ls service the scheme (Figure 3.7) . Bifurcation structures allow for the flow of water fr om main to secondary canals; 2 structures are located on main canal 1, and 3 on main canal 2 (Johnstone, 2011) . There are five drainage canals located throughout the scheme (Chidanti - Malunga, 2009) totaling 17.3km (JICA, 2005) . Ten drainage canals f acilitate the collection and redirection of excess water to either other parts of the scheme or the outside area. The location of main and branching canals were recorded in March 2016 over the course of two consecutive. Locations were mapped by driving th e length of service roads immediate adjacent to record location at 3 - second intervals. When necessary, the locations of canals were tracked on foot. In both cases , the estimated offset distance of the canals to the position of the GPS unit was recorded. These data were uploaded to ArcMap 10.5.1 and the estimated offset distance were used to adjust the locations of canal transects. Results were verified in two ways: (1) by visual inspection Google Earth The engineering of BVIS includes turnouts that manage water distribution from branc hing to tertiary canals . Turnouts are equipped with manually operated, steel slide gates that control the amount of water permitted to move through the branching canal. Operation decks allow for access and control of slide gates. Turnouts are also equipped with hand gates that afford movement of water from the branching to tertiary canals located on either one, or both sides of the division structure. Surveying of 50 slide and hand gates, and thereby the beginning of each tertiary canal, was performed in March 2016 in conjunction with the survey of irrigation canals. Locations of gates were recorded from the operation deck. A total of 142 hand gates and 90 slide gates were recorded across the scheme. The location of culverts were unmeasured. Irrigation scheduling is the responsibility of the Water Use Association (WUA) at BVIS (M. Tarsizio, personal communication, March 17, 2016). During the rainy season, scheme management practices a 3-day irrigation schedule: water is directed along one branch canal for three days, then redirected to another branch canal. No specific irrigation schedule is followed during the dry season; water is directed along branching canals based on the appearance of crop stress (M. Tarsizio, personal communication, August 19, 2016). Regulatory precision and proportional water distribution at each turnout are limited in two ways. First, hand gates were constructed at a fixed width rather than a width proportional to the service area. In addition, these gates can only be opened at 5, 10, or 15cm. (Veldwisch et al., 2009). It was observed during surveying that water flow was further restricted by mounded grass placed across slide and hand gates in an effort to maximize water flow to tertiary canals upstream (Figure 3.8). In addition, two farmers were observed during sampling re-routing tertiary canal flow direction by removing silt preventing water access to their field, only to mound it further down the canal in order to halt the flow of water. Previous research has highlighted conflict between scheme management and farmers on the basis of water regulation including the appointment of water guards who control the allocation of water to each branch canal, and annual water fees (Chidanti- Malunga, 2009) . 51 Figure 3.6 : Bwanje Valley Irrigation Scheme (BVIS) 2017 surveyed perimeter (green). Digitized perimeter of BVIS from JICA's Basic Design and Study Report (2005) and visual inspection of Google Earth v. 7.3.1 imagery (black) initially used for data analysis and field collection. 52 Figure 3. 7 : Irrigation water control structures at B VI S 53 Figure 3.8 : Mounded grass and debris are placed in front of a branching canal slide gate at BVIS to prohibit water flow. Figure 3.9 : A farmer's name written along a branching canal denoting th maintenance area of the canal 3.2.4.2 Drainage While attempts were made to map the drainage system at BVIS during field surveys, many drainage canals were indiscernible as a result of either unmanaged grasses or rice plants allowed to grow within the drainage areas. As such, drain locations were georeferenced 54 (JICA, 2005) . Surface drains facilitate the movement of excess water caused by either rainfall or the application of too much water (Brouwer, Goffeau, & Heibloem, 1985). Improperly functioning surface drains can lead to waterlogging, allowing for pooling of water at the soil surface (Brouwer et al., 1985). Drainage at BVIS occurs in two distinct fashions: tail water is either collected, then redirected through a series of surface drains from one area of the scheme to another or is uncontrolled scheme has four main drains that total 17.3km (JICA, 200 5). These trapezoidal earth canals have a maximum allowable velocity of 0.75 m/s, with an allowable unit area of drainage discharge of 7.64 l/s/ha (JICA, 2005) . Drains at BVIS are susceptible to many of the same ongoing maintenance issues as those found in tertiary canals: portions of drains were choked with weeds and silt during surveying. De- silting and weed management of drains (and branching canals) are shared responsibilities among the farmers and WUA. Farmer names are written along a portion of the branching canal adjacent to their plot(s) at which point drain and branching canal maintenance for this area becomes their responsibility (Figure 3.9). Failure to maintain these areas may result in a fine from the WUA, though scheme management admits there is often a lack of enforcement resulting in untidy irrigation structures (M. Mafosha, personal communication, March 27, 2019). Figure 3.10 shows a section of a main drainage canal that has been cleared in the foreground. Yet, poor management practices on the part of the farmer operating the adjoining plot has allowed for the proliferation of tall grasses to grow. These grasses restrict water flow and redistribution of resources to other parts of the scheme. The variability of drainage maintenance has considerable influence on the area(s) of BVIS that favor mosquito breeding pool formation. 55 Figure 3. 10 : Examples of improper drain management capture during field surveys. Surface drains are meant to be free of weeds and grasses in or der to facilitate the flow and redirection of water to other portions of the scheme. Drain maintenance is the shared responsibility of farmers and the BVIS cooperative 3 .2.4.3 BVIS Operations and Governance managed by the Malawi government, but by 2003 power dynamics shifted and farmers organized into a cooperative. In 2004, the Cooperative was registered with the Ministry of Industry and Trade (Johnstone, 2011) . The capital to form the program that aimed to develop one local product for trade on both domestic and local markets (Johnstone, 2011) . At BVIS, this product is rice. Rice is produced largely by smallholders in Malawi and is grown as a secondary cereal crop to maize (AICC, 2016) 15.2 million tons with farmers averaging 1500 - 2000 kilograms per hectare (AICC, 2017) . The BVIS Cooperative includes thirty - six farmers elected by scheme participants who serve on either the executive committee (27 members), general committee (9 members), or subcommittees. 56 Twelve farmers are elected from each of the three branch canals to ensure equal representation across the scheme (M. Tarsizio, personal communication, March 17, 2016). Subcommittee membership is held by members of the general committee and includes discipline, health, auditing, finance, marketing, and production. A scheme manager works in cooperation with the farmer cooperative to oversee daily scheme operation. The types, quantity, and location of agricultural crops grown at BVIS are governed by the farmer cooperative. Farmers apply seasonally to grow specific crop types. Across the scheme, farmers practice monocropping; intercropping is discouraged (M. Tarsizio, personal communication, August 19, 2016). Each season, farmers pay a water fee of Malawi Kwacha (MK) 1000 (~1.20USD) per ~30X30m plot of land. In addition, farmers are required to pay an annual K5000 participation fee (M. Mafosha, personal communication, March 27, 2019). Farmers cultivate an average of five plots per season (M. Mafosha, personal communication, March 27, 2019). Though rare, farmers have the option are provided with a plot of land at the standard participation fee rate. However, they are not provided with water access in exchange for their ability to exercise total governance over the crop type(s) under cultivation. In these situations, tertiary canal direction is rerouted to inhibit water access. If found to have tampered with the system so as to obtain water on an unallocated plot, farmers face a fine of MK5000 (M. Tarsizio, personal communication, March 17, 2016) . During dry season field surveys, very few unallocated plots were observed, and all were located along the southwestern border of the scheme adjacent to the Namikokwe River. The most obvious Namikokwe River that allows farmers to more easily cultivate them in an unallocated manner. Irrigation is conducted by watering can, a widely practiced method of irrigation for smallholder farmers (Smith et al., 2014) . While watering can irrigation is a simple and effective means of irrigating, 57 carrying water is labor intensive and regular watering is required, limiting areas that can be effectively irrigated. 3 .3 Results 3 .3.1 Breeding Distribution Scenarios The spatial distribution o f water bodies at BVIS is influenced by seasonality, soil properties, timing and intensity of irrigation, drainage, land cover, crop water requirements, and management. It would be misleading to present a single maximum estimate of breeding potential at BV IS given the variable nature of the factors influencing their spatio - temporal distribution. Rather, projected distributions of breeding under three scenarios are presented: rainy season, dry season with limited water resources, and a dry season with abunda nt water resources. For each scenario, projected distributions represent the seasonal peak period. To assist in this analysis, two breeding pool suitability models (rainy season and dry season) were developed by categorizing individual pixels relative to v ery l ow, low, moderate, or h igh) and seasonal land cover produced from this study. Soil properties were taken (JICA, 1994) . Categorical ranking of breeding area was done by first considering the likelihood of ponding based on soil properties, then land cover. In each model, less breeding will occur in Non - Vegetated areas given their makeup: roadways and irrigation canals. In the dry season, active agriculture presents a gr eater likelihood for breeding over fallow areas given the intermittent presence of water either by traditional irrigation or alternative methods (i.e. watering can). Modeled outputs consider how the availability of water resources, drainage, and crop water requirements would affect the distribution of breeding pool formation and persistence. The influence of drainage in all scenarios was approached was divided into three Water Service Areas (WSA) based on the design of irrigation canals, their flow direction, and the location of drains across the scheme. 58 3 .3.1.1 Mid - Rainy Season In this scenario, ubiquitous pooling of surface water occurs across BVIS as a prod uct of: 1) abundant 3.11 ). The dominant crop is rice, at varied stages of d evelopment, consistent with field survey results. In this scenario, pervasive breeding within plots is expected across the scheme. Modeled environmental characteristics (land cover and soil ponding) show that maximum suitability occurs within the northeast ern section of WSA3. Here the ponding potential in combination with rice cultivation and regular flooding create an environment conducive for mosquito breeding. Yet, environmental characteristics alone do not effectively describe water distribution and per sistence as a function of irrigation management for BVIS. In fact, WSA3 receives less water than WSA1 and WSA2 by virtue of 1) its situation to the headworks (water intake) of the scheme and the necessary diversion of water upstream to branching canals; an d 2) backlogging of water in WSA1 & WSA2 as a result of inefficient drainage and re - direction of tail water. It is both the combination of regular irrigation and inefficiency of tail water drainage that makes the western most portion of BVIS the area of hi ghest breeding potential during the rainy season. 3 .3.1.2 Dry Season: Limited Water Resources The opportunity for breeding during the dry season at BVIS is directly affected by insufficient water resources. This is evident in the reduction of cultivated land area from 800ha in the rainy season to <300ha in the dry. Crop types include horticultural crops, namely maize, beans, and cowpea. In this scenario it is expected that breeding opportunity is limited only to active agricultural areas given the intermi ttent presence of water either through traditional irrigation or alternative methods including watering cans and treadle pumps; one instance of treadle pump use was recorded during the 2017 dry season survey. The influence of drainage and redirection of ta il water has less of an influence on the 59 aggregation of water bodies as often water resources are limited to the point that entire tertiary canals are unable to be serviced with water. Agricultural crop types and their distributions are governed by the BV IS Cooperative. When water resources are limited, distinct spatial arrangements of crop types are valuable for irrigation planning and allocation of water resources. During the dry season irrigation scheduling is governed by the appearance of crop stress, a function of the crops fundamental water requirements. Crop types that require more water are irrigated more frequently than others. In our scenario, crop types under cultivation are maize, common bean, and cowpeas. Maize and common be an require on averag e 500 - 800mm and 300 - 500mm of water per growing period, respectively (FAO, 1991) . Cowpeas are most often grown under dryland, not irrigated conditions given their ability to withstand drought conditions. An nual rainfall for geographical areas producing cowpeas averages 400 - 750mm (Republic of South Africa, 2011) . Therefore, areas of greatest concentration of breeding potential are expected to be in the western most area of WSA1 based on the historical spatial arrangement o f crop types dictated by the BVIS Cooperative (Figure 3.12 ). This area is historically allocated to maize cultivation. resources, irrigation water is su pplied to these tertiary canals. This projection assumes that the majority of farmers adhere to the cooperatives crop distribution guidelines. 60 Figure 3. 11: Rainy season breeding scenario under abundant water resource conditions at BVIS. 61 Figure 3. 12: Breeding scenario under limited water resource conditions during the dry season 62 Figure 3. 13: Breeding scenario under abundant water resource conditions during the dry season as a result of the construction of the Bwanje dam 63 3 .3.1.3 Dry Season: Abundant Water Resources A considerable limitation to the success of BVIS is the availability of water resources, either rain fed or from the Namikokwe River. As designed, the Japan ese estimated that BVIS could only support roughly 150 hectares of dry season crops (Veldwisch et al., 2009) . Though, Veldwisch et al. (2009) reports that the information that BVIS was not meant to supply water year - round came as a surprise construct a dam in an effort to improve water availability and expand dry season cultivation at the scheme. Funded by the European Union , the project began in June 2016 (Moyo - Mana, 2018) . The rockfill dam is the largest in the South African Development Community at 40m high and approximately 150m long with a storage capacity of 500 million cubic liters . Construction was completed in October 2018 and the dam later launched in December 2018 (News, 2018) . The dam is expected to provide BVIS with sufficient water resources to cultivate 600ha of rice during the dry season and additional horticultural crops (maize, common bean, cowpea, and soya). Future plans include expanding irrigable are to 2000ha (M. Mafosha, personal communication, March 27, 2019). Given the significant hydrological changes expected at BVIS, the impact to breeding pool distribution during the dry season where abundant water resources are present, coupled with a primary change to rice agriculture as intended is considerable (Figure 3.13 ). The projected 600ha of irrigable area at BVIS will be located closest to the schemes headworks. As such, ri ce cultivation will be limited to these areas. The remaining 200ha area is expected to be cultivated by horticultural crops given the presence of residual soil moisture and sloping topography owing to the gravity - fed design of the scheme. Irrigation schedu ling will follow a 3 - day schedule and drainage should function in a similar fashion to that of typical rainy seasons. Under these conditions, pervasive breeding is expected throughout the 600ha area of irrigated rice with areas of greatest concentration lo cated in the western 64 portion of the scheme, as a function of water availability, ponding potential of soils, and inefficient drainage. Rather than limited breeding opportunity as a function of water resources, the availability of continuous irrigation from the Bwanje dam has the potential to change breeding distribution from that expected during a prototypical dry season, to the expected distribution more typical of rainy seasons. The consequences could be severe. Where breeding is restricted to less than 5 0% of the overall scheme area during a typical dry season, the availability of water resources from the Bwanje dam increases the temporal breeding landscape, changing not only the distribution of breeding but potential seasonal variation in malaria transmi ssion. 3 .4 Discussion and Conclusions The potential spatio - temporal distribution of breeding sites at BVIS is a product of myriad factors. In estimating their probable locations, our models have significant limitations in defining relative breeding poten tial. First, there are no existing measurements for water volume at BVIS; either diverted from Namikokwe, passing through branching or tertiary canals, or ultimately that passes into applied to each field, in combination with information on irrigation scheduling and crop types would assist in developing a dynamic series of models that estimate breeding potential as a function of total water volume and estimated root water uptake. These data could be coupled with local weather information to further refine models for estimations of water loss through evaporation. A related point is the absence of specific crop variety information at BVIS. This information would assist in better character izing the scheme in relationship to the growing periods and specific root water uptake characteristics of crop varieties. An added consideration to both the rainy season and dry season: abundant water resources scenario is the effect of stage of rice grow th. Non - consecutive planting times will affect the spatio - temporal distribution of larva species. An. arabiensis preferentially breed in open, sun lit pool s (Sinka 65 et al., 2010) , characteristic of plots in preparation for, or just after transplanting. As rice plants begin to grow and water surfaces are shaded by vegetation, abundance of An. arabiensis declines (Sinka et al., 2010) . An. funestus s.s ., however prefers breeding in areas with emergent vegetation and large, permanent or sem i - permanent fresh water bodies (Sinka et al., 2010) . Marrama et al (1995) showed that later, grain head formation and maturation stages of rice growth are associated with An. funestus breeding. Diuk - Wasser et al (2007) demonstrated a significant relationship between land use, including stages of rice, and abundance of An. gambiae . Vectorial capacities of mosquito species are not uniform. Because vector abundance is a critical factor to malaria transmission, cultivation practices will influence the disease ecology of the local area. The noted association between irrigated agriculture and proliferation of mosquitoes is no t novel. Yet, treating irrigated areas as homogenous spatial units leads to inaccurate conclusions on breeding pool formation, persistence, and concomitant exposure risk. It is one thing to assert that irrigation encourages mosquito production, it is quite another to answer where to expect breeding to occur. Particularly for irrigated areas that practice seasonal crop rotation, mixed cropping, or intercropping, the spatio - temporal distribution of crop cover can have profound impacts on the distribution of w ater resources across irrigated areas and thereby, breeding potential. The analysis at BVIS showed that the risk potential for mosquito breeding is seasonally asymmetrical across the scheme owing to environmental and anthropogenic factors. The three scen arios presented illustrate how perturbations to the irrigated system in the form of water availability, water management, and crop cover can influence the distribution of aggregated water bodies. It is prudent to consider that female Anopheles mosquitoes a re most likely to take their of breeding potential across irrigated spaces can have profound implications for the distribution of malaria risk for those l iving in close proximity to irrigated agriculture. Given the importance of 66 irrigation to resolving food insecurity, it is necessary to continue considering how to provide crops with the water resources necessary for adequate production without exacerbating malaria risk. 67 CHAPTER 4 4 .1 Introduction There is a strong association between Land Use and Land Cover (LULC) and malaria transmission (Lindblade, Walker, Onapa, Katungu, & Wilson, 2000; Olson, Gangnon, Silveira, & Patz, 2010; Patz & Olson, 2006; Vittor et al., 2009) , including Land Use and Land Cover Change (LULCC) for irrigated agriculture (see: Afrane et al., 2004; Ijumba, Mosha, & Lindsay, 2002; Keiser et al., 2005) This chapter considers the association between LULC and Anopheles mosquito breeding potential within the Bwanje Valley Irrigation Scheme (BVIS), and the 8km area su season. Eight kilometers was selected to give consideration to the average flight distance of Anopheles mosquitoes. The flight range of Anopheles gambiae s.s. is uncertain: Costantini et al. (1996) reports less than 1.0km, while Thomas, Cross, & Bøgh (2013) estimated a maximum distance of 1.7km. An 8km distance provides ample consideration to all distances a female mosquito reared at BVIS, and their immediate progeny, would fly in search of a blood meal. Further, thi s distance provides an adequate estimation of land use and land cover attributable to BVIS and the Bwanje Valley that may promote mosquito development within the area. 4 .2 Bwanje Valley and the Bwanje Valley Irrigation Scheme For the purposes of this stu BVIS (Figure 4. 1). The Bwanje Valley is characterized by an escarpment to the west dominated by the Dedza - Salima Forest Reserve. To the east, the landscape steadily descends into L ake Malawi. Etched across the valley are the Namikokwe, Nadzipokwe, Livulezi, Chikonbe, and Nadzipulu Rivers. The course since 2002 (JICA, 2005) creating a braided stream network of interweaving channels moving eastward toward the Livulezi River. Twenty - seven villages are located within the Bwanje Valley, each with villagers who participate in BVIS (M. Mafosha, personal communication, August 19, 2016 ). The primary et hnic group of the area are the Chewa, descendants of the Maravi who migrated to Malawi 68 in the 13 th century (DeCapua, 2009). The Chewa are a matrilineal, bantu-speaking people. The Chewa language, Chichewa, is spoken by more than 15 million (Boucher, 2012). A significant cultural institution for the Chewa people is the Gule wamkulu , or the to communicate with the people (Boucher, 2012). The dance fuses history, religion, and culture and is performed by masked Nyau dancers for significant events and rituals. For a review of the Chewa people, the Gule wamkulu , and a description of G ule wamkulu characters see Boucher (2012). (dambo) agriculture at the site of the BVIS (Johnstone, 2011; Veldwisch et al., 2009) and widespread, rain fed agriculture in the surrounding refers to seasonally waterlogged areas within floodplains, or along streams (Turner, 1986). The dambo areas at BVIS were traditionally used for rice and sugar cane (Johnstone, 2011), though the prevailing existing 130ha irrigation scheme, the Mtandamula Irrigation Scheme, was absorbed by BVIS; farmers at Mtandamula predominately grew rice (M. Mafosha, personal communication, March 27, 2019). Other crops produced within the Bwanje Valley were groundnuts, pigeon peas, beans, soya beans, pulses, and cotton (JICA 1994). A brief discussion of BVIS is presented in Chapter 1, Section 1.2.2. For a complete discussion -temporal distribution of LULC see Chapter 3. 69 Figure 4.1 : The Bwanje Valley and Bwanje Valley Irrigation Scheme (BVIS) located in Dedza district, central Malawi. The Bwanje Valley is defined as the 8km area surrounding BVIS. 4.3 Methods Dry season variations in LULC for BVIS and the Bwanje Valley and their impact on the distribution of Anopheles breeding potential were determined from: (1) SPOT-6 satellite imagery from December 2013; (2) field surveys of LULC at BVIS and the Bwanje Valley during August 2016; (3) soils data from JICA (1994) ; and (4) onsite interviews with BVIS personnel in 2016, 2017, and 2019. The purpose of sampling only during the dry season was to assess the availability of surface water for mosquito breeding as a function of irrigation in contrast with the surrounding landscape during a time 70 period when water resources are limited. The conceptual framework for this study is presented in Figure 4.2. The conceptual model for this research consists of three primary components: LULC was determined through field sampling of the Bwanje Valley and BVIS, soils data were assessed for ponding potential, and a suitability model for mosquito breeding pool formation and persistence is constructed that considers LULC, soil ponding characteristics, and the presence of irrigation. Figure 4.2 : Conceptual Framework 71 4.4 Land Use and Land Cover 4.4.1 Data Preparation A multispectral SPOT-6 image of the study area was acquired on December 4, 2013 at a spatial resolution of 6.0m for four spectral bands: Blue (0.455 m 0.525 m), Green (0.530 m 0.590 m), Red (0.625 m 0.695 m), and Near-Infrared (0.760 m 0.890 m). Geometric correction to ensure positional accuracy of the imagery was performed using ground control points prior to analysis and a 30m DEM from the Shuttle Radar Topography Mission (SRTM). Sensor calibration and conversion of digital numbers (DNs) to radiance was performed following a solar correction to top of atmosphere (TOA) reflectance using . Classification was performed on two separate images generated from the corrected December 2013 scene: BVIS and the Bwanje Valley. The BVIS image included only those pixels located within the scheme. For the Bwanje Valley image, the BVIS boundary was used to omit pixels from the scene so that image statistics and classification for the Bwanje Valley were not skewed by the presence of pixel values attributable to the irrigated landscape. For both images, an unsupervised ISODATA classification was performed where 255 classes were selected with a .98 convergence threshold (see e.g., Messina et al. 1998). Signature evaluation of the ISODATA classification was conducted using the Transformed Divergence measure where separability markers of >1975 were acceptable. Using the edited signature set, a Maximum Likelihood supervised classification was performed. The BVIS image had eleven significant classes of land cover while the Bwanje Valley image had 15. Data e sites were selected by stratified random sampling of land cover classes from supervised classifications for each image. At BVIS, 242 sample points were selected; 183 sample points were chosen within the Bwanje Valley. 72 4.4.2 Land Cover Data Collection Field based surveys of LULC were carried out in August 2016. A total of 235 sites at BVIS were surveyed over an eight-day period by a field team of three researchers. These samples included 185 of the sites selected from the random stratified sample, and collected to assist with classification. Most often the selection of accuracy assessment points was made when land cover differed considerably from the surrounding area. Following the BVIS survey, the Bwanje Valley was surveyed over a 6-day period by a field team of two researchers. A total of 51 sites were sampled within the Bwanje Valley from the stratified sample and 53 additional accuracy assessment sites. For both surveys, at each site, the location and elevation, along with geotagged photographs each sample site along with descriptions of LULC in all directions were recorded. Where land cover of surrounding areas differed from that recorded at the sample site, the direction of these areas along with approximate distances were included within field notes. Field note transcriptions and associated GPS data collected at each sample site were combined for LULC map development. 4.4.3 Land Cover Classification Classification of LULC followed a two-step process: (1) sites were categorized by land use, land cover type, and feature (e.g. maize, rice, bare earth); (2) field notes were used to include information on stage of agricultural growth, appearance of soils, density of plantings, and locations of trees relative to agricultural growth for each feature identified in step 1. Point files of all sample sites and their 1, then overlain with supervised classification imagery. Descriptions of LULC including the number of samples within each class for the Bwanje Valley are presented in Table 4.1. See Chapter 3 , Table 3.2 for dry season LULC descriptions for BVIS. Information on engineered structures at BVIS including main, tertiary, and branching canals assisted in differentiating between vegetated and non-vegetated areas. 73 4 .4.3.1 Bwanje Valley Irrigation Scheme See Chapter 3 , section 4.2.3.2.3.2 for presentation of LULC f indings at BVIS. 4 .4.3.2 Bwanje Valley The maximum likelihood classification analysis showed fifteen significant classes of land cover for the 4. 3). Dry season land use for the Bwanje Valley is categorized - vegetation. The dominant land cover type was Fallow Agriculture, in cluding sites categorized as charred ground, dried fields, and shrubbery. The Mixed Foliage classification is characterized by shrubs, grasses, and scattered trees. Mixed Forest areas were predominately made up of trees and located in the western portion of the Bwanje Valley along the escarpment. Bare Earth areas are devoid of vegetation. Sample field photographs for each class as presented in Figure 4 .4 . 74 Table 4.1: Land cover classifications for the Bwanje Valley Study Are a TOTAL Class Descriptions Sample AA I.AGRICULTURAL LAND II.ACTIVE Maize Young 1 0 1 Cowpea Cowpea & Shrubbery 2 0 2 Cotton Cotton 0 1 1 II.FALLOW Bare Earth Dark Soil 2 0 2 Dark Soil w/ Straw 0 1 1 Dark Soil w/ Straw & Rid ging 0 1 1 Light Soil 1 0 1 Light Soil w/ Straw 1 1 2 Light Soil w/ Straw & Ridging 8 4 12 Light Soil w/ Ridging 2 5 7 Medium Soil w/ Ridging & Emergent Weeds 3 0 3 Charred Ground Charred Ground on Bare Earth 1 1 2 Dried Fields Vertical Rice Straw w/ Emergent Weeds 1 4 5 Horizontal Rice Straw w/ Emergent Weeds 0 1 1 Dried Grass 1 0 1 Long Term Fallow Medium Soil w/ Ridging & Dried Vegetation 1 1 2 Shrubbery Dense 2 3 5 IV.TREES Green Foliage on Bare Soil 1 0 1 Green Foliage on Dry Harvested Agricultural Land w Ridging 3 2 5 II.NON -AGRICULTURAL LAND Bare Earth Dark Soil w/ Straw 2 1 3 Light Soil 5 1 6 Light Soil w/ Green Vegetation 3 4 7 Charred Ground Charred Ground w/ Dry Vegetation 3 1 4 Dried Gr ass Dried Grass 0 5 5 Trees Green Foliage & Dry Undergrowth 2 1 3 Green Foliage on Bare Soil 3 1 4 Little Green Foliage w/ Dry Undergrowth 4 7 11 Mostly Dried Leaves 0 3 3 Dried Trees on Bare Rock 1 0 1 Human Influence Road 1 0 1 Human Dwellings 1 1 2 Riverbed Riverbed 0 3 3 75 Figure 4.3 : Combined LULC Dry Season classification produced for the Bwanje Valley and BVIS. 76 Figure 4. 4 : Sample photographs depicting classified dry season land cover types of the Bwanje Valley. Developing a classification for the Bwanje Valley presented a number of challenges. Differentiating between agricultural and non - agricultural land was often difficult. In some areas, land had been left fal low for extended periods, losing features characteristic of agricultural landscapes 77 including distinct crop rows. When available, local villagers assisted with providing information of land uses for areas of uncertainty. In addition, some non - agricultural lands including dried riverbeds were temporally converted to agricultural land to take advantage of residual moisture during the dry season. One example is east of the BVIS headworks along the Namikokwe River. During the dry season, available water resourc es past the headworks are severely limited. The characteristic slowing or stagnating of water in the river channels has historically resulted in land transformation of the riverbed to dimba gardens for cultivation of tomatoes, pigeon pea, maize, and rice ( Figure 4. 5). Our classification includes only one instance of sampled human dwellings. In many cases construction materials for home in the Bwanje Valley include sun baked bricks made from local soils and dried grasses for roofing materials. The materials used for construction of homes across the area make differentiating between homes and naturally occurring environmental features challenging. While some homes have corrugated metal roofs, the small size of many homes often precluded them from being registe red as spectrally different than the surrounding landscape. Visual inspection of classification maps and Google Earth Pro 7.3.2.5776 showed that homes were predominately located in areas classified as Bare Earth. Finally, active agriculture in the Bwanje V alley during dry season field surveys was rarely found or sampled outside of BVIS. Only four instances were sampled: (1) Maize, (2) Cowpea and Shrubbery, and (1) Cotton. Cotton is an industrial export crop, often grown by smallholders and the most commonly grown cash crop along escarpments in Malawi (GoM, 2015b) . T he area of cotton sampled for this study was ~30X30m and was a cash crop grown adjacent to the 78 Figure 4. 5 : The Namikokwe River basin east of the BVIS headworks during the dry season. Limited water resources lead to the stagnation of water often resulting in conversion of the area to dimba gardens along the channel. 4 .4.4 Validation of Classification Results Classification accuracies for both maps were assessed using confusion matrices generated in . The numbers of test pixels for each class were selected to ensure that at least 10 - times the number of test pixels were selected per class, as there were classes (see Jensen, 2005) . Test pixels were distributed across all classes. Four accuracy tests w ere applied for each classification: (1) 4.2 & 4 .3 oefficient value for BVIS is .88; overall accuracy is 93%. These values show strong agreement between the classified, thematic map and reference data for both classifications according to the Landis and Koch (1977) >.80 criteria. 79 Table 4. 2 accurac y; E.O., omission errors; C.O., commission errors Fallow Mixed Forest Mixed Foliage Bare Earth Total P.ac (%) O.E. (%) Fallow 37 0 3 0 40 92.5 7.5 Mixed Forest 0 40 1 0 41 100 0 Mixed Foliage 1 0 35 0 36 87.5 12.5 Bare Earth 2 0 1 40 43 100 0 Total 40 40 40 40 160 U.ac (%) 92.5 97.5 97.2 93 C.O. (%) 7.5 2.4 2.7 6.9 Overall Accuracy 95% Kappa .93 Table 4. 3 : accuracy; U.ac, user accuracy; E.O., omission errors; C.O., commission errors Fallow Vegetated Non - Vegetated Total P.ac (%) O.E. (%) Fallow 38 0 4 42 95 5 Vegetated 0 37 0 37 92.5 7.5 Bare Earth 2 3 36 41 90 10 Total 40 40 40 120 U.ac (%) 90.5 1 00 87.8 C.O. (%) 9.5 0 12.2 Overall Accuracy 93% Kappa .88 4 .5 Soils Soils data for the Bwanje Valley were digitized from soils maps produced by JICA as a part of the (JICA, 1994) . To assess soil types, JICA representatives verified existing soils data by field surveying the area. Thirty assessment of pH, texture, organic matter content, total P, total C, and total N (JICA, 1994) . Includ ed with these data are composition and characteristics of each soil type (landform, altitude, drainage, 80 flooding, ponding, erosion, soil depth, top soil, subsoil, pH, CE, CEC, NPK, surface rockiness, land use and vegetation). 4 .5.1 BVIS See Chapter 3 , 3 .2. 3.1 for description of soil types and their associated properties at BVIS. 4 .5.2 Bwanje Valley The Bwanje Valley is made up of thirty - nine soil unit areas comprising seventeen primary soil types. Greater than 40% of unit areas contain two or more soil ty pes. Parent materials for twelve of the soil types found within the Bwanje Valley are fluvial, colluvial and/or lacustrine sediments; the remaining five are felsic and intermediate igneous and metamorphic rocks (JICA, 1994) . Top soil types (0 - 30cm) include clay, loamy sand, sandy loam, and sandy c 4. 6). Soils along the escarpment are typified by little to no ponding . The expectation for ponding increases moving eastward toward Lake Malawi as a result of poorer drainage (Figure 4. 7) (JICA, 1994) . 81 Figure 4.6 : Soil types of the Bwanje Valley and their potential for ponding 82 Figure 4.7 : Soil drainage of the Bwanje Valley 4.6 Modeling Breeding Pool Suitability The distribution of surface water available for mosquito breeding is a function of precipitation, soil drainage, LULC, and anthropogenic influence on the landscape. To estimate the spatial distribution of breeding potential during the dry season, a breeding pool suitability model was created for the Bwanje Valley, including BVIS. Soil features, land cover types, and irrigated areas data were used for 83 model construction. Soil features were assigned values of relative likelihood for ponding from one to six; one represented areas of no ponding and six represented areas of severe ponding potential. These polygon features were rasterized at a 6m resolution to match LULC classification data. Land cover types are Active Agriculture, Mixed Foliage, Mi xed Forest, Fallow Agriculture, and Bare Earth; assigned values of 1 (most likely to support breeding) 5 (least likely), respectively. Ranking of land cover types was determined using two criteria, Anopheles breeding site preference and the likelihood of persistent water bodies. Anopheles breeding preference was determined through a review of the literature for An. funestus , An. gambiae , and An. arabiensis mosquitoes, the primary mosquito vectors of malaria in Malawi (WHO, 2016) . Preferred larval habitats for An. funestus include areas of emergent vegetation, along with permanent water bodies in savanna envi ronments (Sinka et al., 2010) . Likewise, An. arabiensis show preference for dried savanna landscapes along with spars e woodlands. An. gambiae larval habitats contain little to no vegetation (Sinka et al., 2010) . To that end, Mixed Fol iage areas were considered more suitable than forested areas given the preferences of these mosquitoes for sparsely vegetated landscapes for oviposition sites. Bare Earth and Fallow Agriculture land cover types will support the aggregation of water bodies in small depressions, but the loss of water from either infiltration or evaporation renders these areas as unsuitable for mosquito development. Fallow Agriculture was considered more suitable than Bare Earth given the presence of micro depressions within fields as a result of cropping, human and animal footprints. In addition, areas classified as Bare Earth do not possess sugar sources necessary for mosquito survival. Active Agricultural areas are the most suitable areas for breeding given both the persi stence of water for oviposition and their demonstrated association with vector breeding (Sinka et al., 2010) . Irriga ted areas for this study are known areas where water is supplied regularly through surface irrigation during the dry season. Areas of localized irrigation (bringing water directly to a plant from a water source, i.e. watering can) were not considered, nor were irrigation schemes only operational 84 schemes: BVIS and the Nambuona Irrigation Scheme. Only BVIS met the stated criteria for inclusion, however dry seaso n cultivation at BVIS is limited to ~300ha due to insufficient water resources during the dry season (M. Tarsizio, personal communication, March 17, 2016). This ~300ha area and the remainder of the Bwanje Valley were digitized, reclassified to a binary sys tem (irrigated or non - irrigated), and rasterized to 6m to match the other model inputs. Input grids were combined in ArcMap 10.5.1 to reveal all possible combinations of breeding suitability, then categorically ranked according to presence of irrigation, l and cover type, followed by likelihood of ponding. 4 .7 Results 4 .7.1 BVIS Breeding Suitability Presentation of findings specific to dry season BVIS breeding suitability, including discussion of anthropogenic influence through water management during the dry season can be found in Chapter 3 , section 3 .3.1.2. 4 .7.2 Bwanje Valley Suitability Areas of suitable breeding for Anopheles mosquitoes throughout the Bwanje Valley have a distinct spatial structure. Ranking of all possible combinations of suitability and the percent of land area for each category is presented in Figure 4. 8. Categorization of suitability combinations is based on presence of irrigation and land cover characteristics. Supraoptimal breeding area is most prominent within the irrigated port availability via irrigation, where surface water persists despite the absence of consistent rainfall during the dry season. Irrigated area occupies 1.78% of the Bwanje Valley. Whi le previous literature has demonstrated an association between irrigated agriculture and the proliferation of adult stage vectors, irrigated landscapes are treated as homogeneous spatial units. Yet, even within BVIS, the map reveals that the southern porti suitable area than other parts of the scheme. These are areas of active agriculture whose situation is a 85 direct result of water management: a greater proportion of water is directed to and subsequently confined within this area as a result of improper drainage. Non-irrigated, active agricultural areas occupy 1.1% of the Bwanje Valley. In the absence of irrigation, these would be considered the most favorable lands for dry-season mosquito breeding. Results show that roughly 25% of the Bwanje Valley is satisfactorily suitable for Anopheles breeding concentrated in the central portion of the area. Nearly thirty-six percent of the area is suboptimal; the highest proportion of any classified area. Twenty-five percent is unsuitable and primarily concentrated north of BVIS. For both the suboptimal and unsuitable classifications rapid loss of soil moisture due to evaporation restricts breeding potential. 86 Figure 4. 8: Dry season breeding scenario for the Bwanje Valley 87 4 .8 Discussion and Conclusions Irrig ation plays a pivotal role in the distribution of suitable breeding sites for Anopheles mosquitoes. While environmental suitability models illustrate the extent of landscape conducive to breeding pool formation and persistence, mosquito presence and their overall fitness are also associated with the availability of plant sugar and, for females, vertebrate blood. Host specificity and preference vary widely among species of mosquitoes and may include such invertebrates as mammals, amphibians, and birds (Foster & Walker, 2009 ) . Malaria transmission and intensity present a complex interplay of biotic and abiotic factors. As surface water availabili ty increases across the landscape for irrigated agriculture, so too does the opportunity for female mosquitoes to lay their eggs and larvae to survive through to adult stage. For humans living in close (<1km) proximity to irrigated schemes, the density of adult Anopheles mosquitoes in houses will increase as females seek out human dwellings for blood feeding and resting sites. This relationship will decline with increasing distance from irrigated areas. In irrigated agricultural areas, the consistent availa bility of s u rface water for breeding during the dry season changes t emporal malaria disease dy na mics. For those living in close proximity to irrigated schemes, risk of malaria transmission is higher than those living further away. I t is important to note t hat increases in mosquito populations do not necessarily increase malaria risk. There are myriad factors that affect malaria transmission including the stability of malaria transmission in areas where irrigation is introduced (Ijumba & Lindsay, 2001) , housing quality, availability of anti - malarial drugs (N. J. White, 2008) , economic status (Collins & Paskewitz, 1995) , and the use of insecticide - treated bed nets (Mutuku et al., 2011) . In the Bwanje Valley , supraoptimal conditions for breeding during the dry season are concentrated within BVIS. H uman dwellings within 1 - km of BVIS were identified using Google Earth Pro 7.3.2.57 76. Constantini et al. (1996) demonstrated that maximum flight distance for Anopheles gambiae s.s. was < 1.0 - km while Thomas et al. (2013) reported 1.7km . A total of 320 human dwellings 88 are located within the area. It is expected that temporal disease dyna mics for malaria will differ for distances from the scheme. For anthropophilic mosquitoes, including members of the Anopheles gambiae complex (Sinka et al., 2010) , mosquito dispersal is short in areas of high human population densities (Carter, Mendis, & Roberts, 2000) ; mosquitoes will fly no further than necessary for a blood meal. In rainfed agricultural systems, s easonal peaks in malaria transmission most often occur during the late rainy season or i mmediately after its conclusion. Dry season malaria transmission is limited by the redu ction of inundat ion , or likely to become inundated areas. Malaria transmission for those 320 households within 1 - km of BVIS will not experience the same level of decline in transmission as those beyond 1 - km from the scheme. During the dry season, optimal b reeding conditions are expected in active agricultural areas given the presence of either formal or informal irrigation measures. A notable limitation of this study is the absence of active agricultural areas located outside of BVIS. Further research on dr y season modeling of breeding distribution should prioritize identifying active agricultural areas through remotely sensed or other field - based data. I rrigated agriculture impact on spatio - temporal malaria disease dynamics is considerable , particularly as it expands across sub - Saharan Africa to mitigate food insecurity. Agricultural growth through irrigation is often cited as a critical means for the reduction of rural poverty in Africa (You et al., 2010) . As such, many African governments including Malawi have adopted policies that specifically target increasing irrigation measures (see: GoM, 2016) . As intensification of irrigation continues, research is necessary to fully estimate the impact of scaling irrigated agriculture on malaria risk, particularly during dry seasons. Further, predictive modeling for mosquito suitability should not rely on assumptions that agricultural areas during the dry season are predominately fallow, thus inh ibiting mosquito breeding pool formation. The addition of spatially defined irrigated spaces to predictive 89 mosquito models for sub - Saharan Africa is imperative in light of widespread LULCC for irrigation. This is challenging in the absence of up - to - date, s patio - temporal data on irrigated lands at scale. ( Thenkabail et al., 2009 ; FAO, 2016 ) . The findings of this research demonstrate the asymmetrical breeding potential for Anopheles season. T he introduction of irrigation to landscapes not only changes the geography of mosquito breeding, but malaria risk potential for those living in close proximity to irrigated schemes. 90 C HAPTER 5 5.1 Introduction The expansion of irrigated agriculture is es sential for mitigating food insecurity through increased crop production (Sajidu, Monjerezi, Ngongoro, & Namangale, 2013) . While scaling irrigated agriculture has dem onstrated significant boosts to crop productivity (ADB, 2013; Melaine & Nonvide, 2017) ag rarian transformation of the landscape for irrigated agriculture is associated with encouraging the production of adult malaria mosquito vectors. Water is a requisite for mosquito development. As such, land use and land cover changes (LULCC) that alter the distribution and flow of water across the landscape can have profound impacts on the epidemiology of malaria. Keiser et al. (2005) highlight that as much as 90% of the global malaria problem can be attributed to environmental factors including the establi shment of irrigated schemes. Irrigation for crop production encourages the development of significant populations of malaria disease vectors (Sissoko e t al., 2004) and pathogen transmission through a number of pathways. First, through the development of vector habitat and the production of adult stage mosquitoes (Van Der Hoek 2004; Mutero et al. 2004). Further, intensification of agriculture involves a significant change to the natural landscape occurring across areas, altering vegetation and extending the spatial distribution of surface water across the landscape. Likewise, irrigation can promote vector longevity by significantly increasing relative hum idity over large areas (Secretariat & WHO, 1996) . Collectively, landscape modifications for irrigated agriculture have the potential to both promote diversity of breeding sites and reduce predation of vectors (Sutherst, 2004) . Further, environmental and e cological changes for irrigated agriculture can increase the frequency of human - vector contact thereby encouraging transmission (Secretariat & WHO, 1996) . Studies on relationship between irrigated agriculture and malaria are well documented in the liter ature and show divergent conclusions. In some studies, malaria incidence has increased (Afrane et al., 2004; Ghebreyesus et al., 1999; Guthmann, Llanos - Cuentas, Palacios, & Hall, 2002; Jaleta et al., 2013; Keiser, Caldas, et al., 2005; Kibret et al., 2010; Kobayashi et al., 2000; Urama, 91 2005) . Contrastingly, other studies have shown a decrease or no change of infection (Assi et al., 2013; Diakité et al., 2015; Faye et al., 1995; Ijumba, Mosha, & Lindsay, 2002; Klinkenberg, Van Der Hoek, & Amerasinghe, 2004 ; Mutero et al., 2004) . The contradictory nature of such studies suggests the necessity for further investigation on the impact of irrigated agriculture on malaria transmission particularly in light of continued emphasis on expansion throughout malaria end emic areas to meet crop production demands. In this chapter, probable changes to the spatial epidemiology of malaria in Malawi are described through analysis of habitat suitability for Anopheles gambiae s.s. mosquitoes, the extent of malaria prevalence, a nd proposed spatial expansion of irrigated sites Belt Initiative . 5.2 Methods The impact of scaling irrigated agriculture on the spatial distribution of malaria risk potential was determined from: (1) Historical examination of irrig spatial extent, and intended extent of irrigated sites in Malawi; (2) Habitat suitability for Anopheles gambiae s.s . mosquitoes assessed through creation of a Habitat Suitability Model (HSM) in Google Earth En gine (GEE); and (3) Malaria prevalence data from the Demographic and Health Survey s (DHS) Malaria Indicator Survey (MIS) at the cluster level for 2012, 2014, and 2017. The conceptual framework for this study is presented in Figure 5. 1. 5.2.1 Irrigated Agr iculture in Malawi 5.2.1.1 Irrigation Development and National Policy Frameworks Expansion of irrigated agriculture in Malawi has occurred against the backdrop of national policy frameworks often tied to strategies to increase agricultural productivity as a means of poverty reduction and economic growth. Historically, Malawi has experienced considerable oscillations in economic growth due in part to external shocks and policy implementation (IMF, 2017) . A strongly agrarian society, agriculture is fundamental to economic performance and contributes nearly 30% of annual GDP (Giertz, Caballero, Dileva, Galperin, & Background, 2015) . In addition, nearly 80% of 92 the total workforce is employed by the agricultural sector (GoM, 2010) . It is unsurprising then that historical development strategies to improve the socio - economic status of the country have emphasized increasing agricultural productivity incl uding through scaling irrigation measures. Shortly after independence in 1964, the Government of Malawi (GoM) began setting out sectoral strategies and objectives for economic growth through 10 - I) (IMF, 2017; Record, 2007) . Together, the DEVPOLS aimed at achieving an 8% annual economic growth rate through increasing agricultural productivity, shifting economic activity to the central region, increasing local partic ipation in the economy, and eliminating foreign aid dependence (Record, 2007) . Economic growth eventually faltered in the early 1980s aft er a series of external shocks (National Economic Council, 2000) . As a counter measure, Malawi entered into structural adjustment loan negotiations with the IMF and World Bank (IMF, 2017) ultimately implementing structural adjustment programs wherein the Policy Framework Paper (PFP) was designed for executing medium - term economic policies (National Economic Council, 2000) . By January 1996, the GoM began developing Vision 2020 in response mounting concerns over a need for long - term strategy for development management (National Economic Council, 2000) . Vision 2020 serves as the overarching framework for formulation, implementation, and evaluation of short and medium - term plans to achieve Malawians long - By the year 2020 , Malawi, as a God - fearing nation, will be secure, democratically mature, environmentally sustainable, self - reliant with equal opportunities for and active participation by all, having social services, vibrant cultural and religious values and a technologically driven middle - income economy (National Economic Council, 2000, pg 27) . The Vision 2020 was launched in 2000. Four medium - term national development strategies were formulated to attain the Vision 2020: the Malawi Poverty Reduction Strategy Paper (MPRSP) and the Malawi Growth and Development Strategies (MGDS I; MGDS II; MGDS III) (GoM, 2017) . 93 Figure 5.1 : Conceptual framework 94 Strategies to increase agricultural productivity through promotion of irrigated agriculture are outlined in each including, drainage of marshlands for irrigation, construction of small- medium- and large- scale irrigation schemes, construction of multi-purpose dams, rehabilitation of existing irrigation schemes and dams, establishing piped water systems, and developing areas with irrigation potential (GoM, 2002; 2006; 2011; 2017) . Aligned with Vision 2020, the GoM produced the National Irrigation Policy and Development Strategy (NIPDS) in June 2000 as a comprehensive policy to guide irrigation development (GoM, 2000a) . The document outlined both policy and development objectives for the irrigation sector. Policy objectives emphasized poverty alleviation, increasing and enhancing food security, creating enabling environments for irrigated agriculture through private sector investment, optimizing government investment in irrigated agriculture, facilitating effective research in irrigation technology, and a focus on competitive financing for irrigation projects along with improvement of marketing systems. To meet these policies, eight development strategies were outlined, notably increasing land area under irrigation so that up to 15% of irrigable land was being effectively utilized. Further, under the NIPDS, government support of the existing sixteen government-run smallholder irrigation schemes was meant to be transferred to organizations (Ferguson & Mulwafu, 2005). Overall, the goal for irrigation development was to increase incomes and commercialization of the irrigation sector (NIPDS, 2000). By 2016 irrigation potential in Malawi remained largely unexploited (NIP 2016; GoM 2015). In response, the GoM revised the NIPDS. The National Irrigation Policy (NIP) 2016 was formulated as an extension of the NIPDS that includes policies, plans, and monitoring and evaluation systems to ensure sustainable economic growth based on potential for the irrigation sector (NIP, 2016). Objectives are aligned with the MGDS II, Comprehensive African Agriculture Development Program (CAADP), and Sustainable Development Goals (NIP, 2016). 95 An additional policy effort to meet agricultural growth and poverty reduction goals under the MGDS II is the Agricultural Sector Wide Approach (ASWAp). The ASWAp was developed in 2010 as a priority investment program intended to support activities in the agricultural sector from 2011- 2015 (GoM, 2010). Strategies for increasing agricultural productivity are grouped by focus area and -process (GoM, 2010). Th efficiency and expanding irrigated agriculture through the Green Belt Initiative (GoM, 2010). Figure Agriculture, Economic, a Vision 2020. Figure 5.2 Economic, and Irrigation sectors. 96 5.2.1.2 Irrigation Development in Malawi Irrigation development in Malawi is conducted by the public and private sectors, though historically has been predominately spearheaded by the ministry of Agriculture, Irrigation, and Water Development (GoM, 2012b) . Public irrigation development generally targets smallholder farmers and focuses on irrigating food security crops, namely rice (World Bank, 2011). Private sector schemes operate at larger scales producing cash crops for the export market (FAO, 2005). Donor financing for irrigation has been provided through a variety of investors incl uding the World Bank, the African Development Bank, the Japanese International Cooperation Agency (JICA), the United States Agency for International Development (USAID), the Arab Bank for Economic Development, and the Government of India (GoM, 2012b) . colonial rule as a means of promoting irrigation farming and modernizing peasant agriculture (Gwiyani - Nkhoma, 2005) . Prior to this period irrigation farming was limited. Peasant farmers practiced flood cropping and dimba irrigation for vegetables, maize, and rice (Gwiyani - Nkhoma, 2005) . Dimba gardens are small plots bordering rivers that are cultivated using residual moisture. White settlers practiced irrigation for the production of tobacco (Gwiyani - Nkhoma, 2005) . The - holder irrigation schemes was the result of several factors including the 1948/9 drought and famine. Irrigated agriculture practices were viewed as adaptive strategies for mitigating such disasters. In addition there was an increasing desire to promote rice production (Gwiyani - Nkhoma, 2005) . Established in 1949, the Limphasa Rice Irrigation Scheme was the earliest colonial irrigated scheme in Malawi (FAO, 2005; Gwiyani - Nkhoma, 2005) . Limphasa spanned 700 acres and was located in the Limphasa Dambo within the Nkahata Bay District (Nkahoma, 2005). Additional irrigation projects undertaken during the colonial period were the Shire Valley Proje ct (1952 - 1979), 97 Phalombe - Chilwa Development Project (1952), and the Njala Rice Scheme (1957) (Gwiyani - Nkhoma, 2005) . During the post - colonial period of 1967 and 1982, 16 schemes were constructed (Chirwa, 2002; Gwiyani - Nkhoma, 2011) with a total irrigable land area of 3600ha (FAO, 2005) (Figure 5.3 ) . Designed to increase rice production and train farmers in irrigation farming, the schemes are located along the shores of Lake Malawi , (Karonga, Nkhata Bay, Nkhota Kota) the Lake Chilwa Basin , and the Lower Shire (FAO, 2005; Gwiyani - Nkhoma, 2011) . - Saharan Africa (Woodhouse et al., 2017) . Gwiyani - Nkhoma (2011) reports the failure irrigation farming during this time period stemmed from overdependence on donor funding, lack of local ownership of resources and land alienation, community displacement, failure of the government to consider local context and circumstance during development, and the adoption of structural l the late 2000s that bilateral and other international funders began reinvesting in irrigation and water management (Woodhouse et al., 2017) . Expansion of irrigation for 2 (Gwiyani - Nkhoma, 2005) to 8,255ha in 2002 (Nyondo, 2016) . By 2005, irrigated area for smallholder agriculture nearly doubled to 15,988ha (GoM, 2012a) . Contrastingly, expansion of irrigation for estate schemes faltered, only increasingly by 225ha during the same period to a total o f 48,360ha in 2005 (GoM, 2012a; Nyondo, 2016) . Since 2005, irrigation development has been steadily increasing annually for smallholders to 47,611 ha reported in 2015 (GoM, 2015b) . The statistics for area under irrigation for estate schemes show only a slight increase from 48,360 ha in 2005 to 52,498 ha (GoM, 2015b) . 98 Figure 5.3 : Timeline of expansion of irrigated agriculture in Malawi (1949-1979) 5.2.1.3 Spatial Extent of Irrigated Agriculture in Malawi (2015) The GoM released the Irrigation Master Plan and Investment Framework (IMP) as a comprehensive framework to assist stakeholders in sustainable development and expansion of irrigation in 2015 (2015-2035). Funded by the World Bank, the IMP was developed for the Department of Irrigation as a part of the Irrigation, Rural Livelihoods and Agricultural Development Project (2005-2015) (IEG, 2015). Preparation of the IMP was conducted from November 2013-December 2104 (IMP, 2015). 99 irrigation master plan including a comprehensive inventory of irrigation schemes (IEG, 2015) . Previously, discrepancies in reporting of total land area under irrigation have been reported (World Bank, 2010) schemes. - step process owing to pr evious data inconsistencies (GoM, 2015b) . Scheme locations f inconsistencies in location and scheme size, each district supplied one representative to work alongside a GIS special ist to verify location and size of schemes using Google Earth (GoM, 2015b) . The IMP reported land area for irrigation totaled 104,298 ha; 47,611 ha private estate and 56,687 for smallholders by the end of 2014. Appendix 5 of the IMP provides an inventory of existing schemes allocated by type, formal or informal. F ormal schemes include estate schemes and those whose development involved some form of engineering design and construction and in some cases schemes that evolved though farmer - constructed diversion structures (GoM, 2015b) . Formal schemes are further divided by size (mini, small, medium, and large) and operation (fa rmer organization or private). Table 5. 1 presents a summary of existing schemes. The agricultural sector in Malawi consists of two primary sub - sectors, estate and smallholder, which contribute 30 and 70 percent re spectively to national AgGDP (Go M 2016). E state schemes are large - scale farming operations whose products are produced almost exclusively for the export market (FAO, 2005) . The primary estate - grown crops are sugar, tea, coffee, and tobacco (GoM, 2015c) . Fifty - seven estate schemes are located within thirteen districts in Malawi. The largest proportion of total r (43,414 ha). Roughly 75% of estate schemes utilize sprinkler irrigation. Other technologies include pumping (8%), central pivot (7%), dams (3%), or drip irrigation (1%) alone or in combination. 100 Between 2006 and 2014, irrigated land for smallholder farming increased by almost 300% across Malawi (GoM, 2015c) predominately by gravity-fed (56%) or treadle pump (29%) irrigation. Other technologies include watering cans (7%) and motorized pumping systems (8%) (GoM, 2015b). Small-scale irrigation schemes predominately cultivate green maize, rice, and an assortment of horticultural crops including sweet potato, leafy vegetables, tomato, and onion (GoM, 2015c) . Cropping patterns vary by farmer and include intercropping, mono-cropping, and raised beds. Data for formal and informal schemes were reviewed for duplicate entries and accurate spatial reference information. One hundred-six duplicate entries were removed from the dataset, predominately within the Mzimba district. Data tabulation yielded 1596 irrigation scheme records; 1370 informal schemes and 226 formal schemes. Spatial reference information was missing for 324 schemes including the entire Nkhotakota district. To account for these schemes locations, the IMPs district maps were georeferenced. All georeferencing and visualization was conducted in ArcMap hemes by Irrigation Service District (ISD) are presented in Figure 5.4. 101 Table 5. 1 : Summary of existing schemes by typology and hectarage Irrigation Service District District Informal schemes Formal schemes Informal s chemes (ha) Formal schemes (ha) Total scheme (ha) Chikwawa Chikwawa 74 3 4103 26451 30554 Nsanje 14 3 542 1100 1642 Blantyre Blantyre 19 0 259 0 259 Chiradzulu 3 0 26 0 26 Mulanje 37 13 318 1032 1350 Mwanza 15 0 145 0 145 Neno 5 1 27 150 177 Phalombe 19 1 461 20 481 Thyolo 36 9 680 1223 1903 Balaka 52 4 409 2283 2692 Machinga Machinga 56 7 2086 867 2953 Mangochi 62 6 3436 713 4149 Zomba 67 6 3189 2086 5275 Lilongwe Lilongwe 190 76 1905 9568 11473 Ntcheu 33 12 2851 994 3845 Dedza 93 48 1384 1285 2669 Kasungu Kasungu 25 0 177 0 177 Dowa 89 2 3488 165 3653 Ntchisi 27 1 670 60 730 Mchinji 19 5 108 26 134 Salima Salima 5 2 59 112 171 Nkhotakota 24 15 397 17127 17524 Mzimba Mzimba 196 0 6223 0 6223 Nkhata Bay 43 2 598 535 1 133 Rumphi 88 6 1330 439 1769 Karonga Karonga 26 4 927 989 1916 Chitipa 53 0 1276 0 1276 1370 226 37074 67225 104299 102 Figure 5.4 : Spatial Extent of Irrigated Agriculture in Malawi (2015) 5.2.1.4 Expansion of Irrigated Agriculture The possibilities for irrigation development in Malawi are considerable: potential irrigable land is roughly 400,000 hectares of which only 104,298ha have been developed (GoM, 2015b, 2015a) . The 103 IMP aims to increase irrigable area to 220,000 hectares by 2035 through a combination of expanding existing schemes, working to develop schemes previously identified a nd at various stages of feasibility, and identification of additional schemes (GoM, 2015b) . Statistics for existing, considered, and potential irrigation schemes are presented Table 5. 2. These gains will be in initiated in three phases: (1) 20,000 hectares; (2) 28,500 hectares; and (3) 67,500 hectares. The total co st of the IMP is projected to be roughly $2.0 billion USD over 20 years. Given the financial challenge this presents to the GoM, financing options include: (1) the GoM development budget; (2) development partners; (3) international investment banks and equ ity funds; (4) private agribusiness companies; and (5) individual farmers. Table 5. 2 : Existing, considered, and new irrigation schemes under the IMP (Data Source: Irrigation Master Plan and Investment Framework, 2015) Scheme Ty pe Area (ha) Potential Increase (ha) Future Increase (ha) Existing Schemes Estate 47,5000 22,500 70,000 Smallholder 56,500 23,500 80,000 Sub - Total 104,000 46,000 150,000 Considered Schemes Shire Valley 22,000 0 22,000 Commercial estates 8,500 200 8,700 Chikwawa (GBI) 6,300 0 6,300 On - going DOI Schemes 6,000 0 6,000 PRIDE Schemes 4,000 0 4,000 Songwe River 3,000 0 3,000 Sub - Total 49,800 200 50,000 New Potential Schemes Dambo 41,700 20,3000 62,000 Other new schemes 24,500 0 61,000 Fu ture Lake Pumping 0 62,000 62,000 Sub - Total 66,200 118,800 185,000 Total 220,000 165,000 385,000 104 Potential irrigation schemes are divided into five types: Irrigation Master Plan Potential Irrigation Schemes (IMPPIS), Considered Schemes, Rural Infrastr ucture Development Program (RIDP) II, or Irrigation, Rural Livelihoods and Agricultural Development Project (IRLADP), and Green Belt Initiative (GBI) Schemes (Figure 5.5 ). The IMPPIS were selected using Multi - Criteria Decision Analysis (MCDA) assessed at t he Water Resource Area (WRA) level. Ranking parameters for the MCDA included geophysical suitability, market orientation and linkages, economic viability, environmental acceptability, stakeholder support, and land tenure systems. There are 85 IMPPIS and f our Considered Schemes. Considered Schemes we re previously selected by the Department of I rrigation prior to the IMP. There are 10 RIDP II schemes previously identified under the RIDP Project. The RIDP Project aims to reduce dependency on rain fed agricult ure, diversify cropping, mitigate vulnerability to drought, and enhance rural income and food security. The IRLADP schemes include 12 schemes previously identified by the Irrigation Rehabilitation and Development and Catchment Conservation (GoM, 2015b) . The GBI was introduced in 2010 in direct response to critici sms of the Farm Input Subsidy Program (FISP) (Chinsinga, 2017) . While FISP has been (Denning et al., 2009) many international donors were skeptical that FISPs achievements were largely circumstantial given favorable climatic patterns (Chinsinga, 2017; Chinsinga & Chasukwa, 2012) . Further, questions arose on efficiency and effectiveness of inputs along with long - term affordability (Chinsinga & Chasukwa, 2012) resources to increase agricultural production, productivity, incomes and food security at both household and national levels, and to spu r economic growth and development through development of small and large scale irrigation and maximization of rain - (Chinsinga & Chasukwa, 2012) . The seven major components of the GBI are: (1) Crops, Livestock, and Fisheries 105 Figure 5.5 : Malawi's proposed irrigation schemes, by type. Development; (2) Infrastructure Development and Rehabilitation; (3) Land Administration; (4) Environmental Management; (5) Technology Development and Dissemination; (6) Institutional 106 Development and Capacity Building; and (7) Agro - processing and Market Development (GoM, 2015b) . Under the GBI, large tracts of land all within 20k en perennial rivers have been offered to local and foreign investors in an effort to increase irrigable area to 1 million hectares ( Chinsinga & Chasukwa, 2012; GoM, 2015b) . The 25 conceptua l GBI site locations are presented in Figure 5.6 ; many of whose area overlap s with those proposed sites presented political will, and land acquisition iss ues (Mkwanda, 2017) . By 2015, only four potential scheme locations had been thoroughly assessed owing to financial restrictions. One in Chikwawa, Salima, Mangochi, and Karonga (GoM, 2015b) . 107 Figure 5.6 : Proposed Green Belt Initiative Sites including intended crop types. Total area for proposed irrigation across GBI sites is nearly 1,000,000ha. 108 5.2.2 Modeling Habitat Suitability for Mosquitoes Ecological Niche Theory (Grinnell 1917) environmental condi (Hirzel & Le Lay, 2008) where predictions of a target species likelihood of spatio- temporal occurrence are determined based on environmental characteristics (Hirzel & Le Lay, 2008) . Predictive models of species distribution provide valuable, cost effective information to end users for myriad applications including environmental management, risk awareness, and conservation. Likewise, for testing the effect of climate change on species distributions (Dueri, Bopp, & Maury, 2014). Predictions of human risk exposure to infectious disease vectors using HSMs have been performed for several vector species including ticks (Johnson et al., 2016), triatomines (Sarkar et al., 2010), and mosquitoes (Ayala et al., 2009) . In an era of big data, the availability of satellite-derive global climate, terrain, and land cover datasets presents an opportunity for modeling the suitability of malaria disease vectors across geographies and time scales. Leveraging Google Earth Engine, a raster-based dynamic mosquito suitability model was constructed at a 250-m spatial resolution. The assist in identification of predicted locations of mosquito species for spatially targeted control efforts; and 2) produce an open-source, agile model for end users to model distributions of any mosquito species, across any geography, with or without extensive knowledge of working with geospatial datasets . For the purposes of this study, the model is parameterized for the malaria vector, Anopheles gambiae s.s . in Malawi for 2012, 2014, and 2017 to coincide with Demographic and Health Surveys (DHS) Malaria Indicator Survey (MIS) years. Malaria is a considerable public health issue in Malawi (DHS, 2017) and is vectored by three principle mosquito species: An. gambiae s.s. , An. arabiensis and An. funestus (Chavasse, 2002) . An. gambiae s.s. is the most efficient of the three vectors in transmitting malaria (Coetzee, 2004) and belongs 109 to the Anopheles ga mbiae sensu lato (s.l.) complex. The An. gambiae s.l. complex comprise s seven morphologically, indistinguishable sibling species: An. gambiae sensu stricto (s.s.) , An arabiensis , An. quadriannulatus species A, An. quadriannulatus species B, An. melas , An. merus , and An. bwambae (Bass, Williamson, Wilding, Donnelly, & Field, 2007) . The geographical distribution of An. gambiae s.s . is widespread across sub - Saharan Africa, though predominately concentrated along 10 ° N latitude and between 10 - 20 ° S latitude in Tanzania, Malawi, Mozambique, and Zambia (Wiebe et al., 2017) . Over time, the An. gambiae s.s. speci es diverged into two strains, Mopti (M) and Savannah (S), but is often considered a singular species in the literature (Becker et al., 2003) . For this reason , An. Gambiae s.s. is modeled as a single species in this study , rather than the specific M and S forms. 5.2.2.1 GEE Google Earth Engine (GEE) is an open - source, cloud based platform designed for users to access and process their own private data , or dat - petabyte geospatial catalog (Gorelick et al., 2017) . Users access GEE through an app lication programming interface and associated web - based interactiv e develo pment environment. While traditional methods of data storage and analysis may preclude users from storing, managing, and processing very large geospatial datasets, GEE removes these barriers, allowing users to more readily process data and disseminate thei r results (Gorelick et al., 2017) . For a complete review of the GEE platform, including system architecture and data distribution models see Gorelick et al., (2017) . The GEE platform was selected for model construction to ensure that the widest possible audience of end users could access and use the model wi thout limitations related to data storage or computational processing power. 5.2.2.2 Habitat Suitability Modeling Methods 5.2.2.3 Model Construction To determine the spatio - temporal distribution of An. gambiae s.s. mosquitos, a raster - based dynamic specie s distribution model was constructed in GEE that uses abiotic and biotic variables specific to the species biological requirements (Table 5.3 ). Suitable areas are d efined as those which 110 encourage the creation and persistence of breeding sites for oviposition . Parameter thresholds for each of the input variables were selected on the basis of literature review and are described in the following section. Given the importance of model flexibility for end-users, parameter thresholds are adjustable based on users Customizable variable thresholds are available for start and end year, temperature minimum (Tmin) and maximum (Tmax), NDVI, precipitation percentile, and flow accumulation percentile. Beyond identifying suitable areas, the model provides fine-resolution explicit information on the drivers of suitability. This function allows users to investigate not only the geography of suitable areas, but what variables, or combinations of variables encourage or restrict the likelihood of the species inhabitation across space . To do so, predictor variables are partitioned into three categories: climate, land, and water. 5.2.2.3.1 Predictor Variables Table 5.3: Predictor variables, data sources, resolutions, and threshold variables for model construction Data Sources Product Spatial Resolution Temporal Resolution Threshold Value Temperature NASA MODIS MOD11A2.006 1-km8-day Min: 18°C Max: 32°C NDVI NASA MODIS MOD13Q1.006 250-m16-day >=.30 Land Cover NASA MODIS MCD12Q1.006 MCD12Q1.051 500-mAnnual See table 5.3 Precipitation UCSB Climate Hazards Group CHIRPS Pentad ~5-km Pentad >50% Max: 3200mm Flow Accumulation WWF HydroSHEDS 15 arc seconds ->=25% Water Bodies JRC GSW1_0 30m -111 5.2.2.3.1.1 Climate Temperature Temperature is critical to mosquito life-history (Beck-Johnson et al., 2013; Lyons, Coetzee, & Chown, 2013; Shapiro, Whitehead, & Thomas, 2017). During aquatic life stages, higher ambient temperatures encourage faster development, but have also been shown to cause declining larval survivorship. Likewise, rising environmental temperature significantly increases adult mortality (Christiansen-Jucht, et al. 2014). Bayoh & Lindsay (2004) demonstrated that the upper and lower temeprature thresholds for An. gambaie s.s. larval development were 18°C and 32°C. At higher (38- 40°C) or lower (10- 12°C) temperatures, larval survivorship was shortened significantly. Throughout most of sub-Saharan Africa, larvae may regularly experience high temperatures, particularly during dry seasons, though in some cases only for a limited number of hours during the day. To cope, larvae will dive down away from the water surface (Hauf &Burgess, 1956) or move into shaded areas of breeding pools (Foley et al 2002). Temperature thresholds for TMin and TMax are set to 18°C and 32°C, respectively (Bayoh & Lindsay, 2004). Data were acquired from the Moderate Resolution Imaging Spectroradiometer (MODIS) Terra Land Surface Temperature and Emissivity 8-Day Global dataset (MOD11A2) V006 at a 1-km spatial resolution. These data are a composite of the corresponding daily MODIIAI Land Surface Temperature (LST) (Google, 2018). Data are available from 03- 05-2000 through present day (05-2019), providing an extensive, historical dataset for modeling. 5.2.2. 3.1.2 Land Land Use and Land Cover There is a significant correlation between Land Use and Land Cover (LULC) and distributions of mosquito species (Munga et al., 2009). The geography of vector abundance is a product of numerous factors including the availability and productivity of aquatic habitats, proximity of larval habitats to sugar and blood meal sources, and species-specific site preferences. An. gambiae s.s. larval habitats are 112 characterized as open, sunlit pools (Becker et al., 2003) with little to no vegetation (Sinka et al., 2010 ) . Munga et al. (2009) demonstrated that p referre d breeding habitats for An. gambiae included deforested, cultivated or natural swamps, and cow hoof prints. Further, Sinka et al. (2010) describes An. gambiae habitats as those often associated with human activity including rice fields, wheel ruts, or areas of rice cultivation. The environmental niche of the M - and S - forms of An. gambiae were assessed by Simard et al. (2009) in Cameroon. Results showed that habitat suitability for S - form mosquitoes included dry savannah, areas of higher evapotra nspiration and lower water vapor pressure, and spaces highly degraded by human activity; S - forest (Simard et al., 2009) . M - form mosquitoes preferred areas of with higher frequency of forested area, greater sunlight exposure, higher water vapor pressure, and lower temperatures and evapotranspiration. Data for LULC are taken from MODIS Land Cover Type Yearly Global (MCD12Q1) V051 and V006 Type 1 product at a 500 - m resolution and 1 - year cadence. Data availability ranges from Jan 1, 2001 to December 31, 2016 an d were reduced to a single LULC layer by calculating mode. To determine whether a class of LULC was suitable, class descriptions from the LULC data product were compared to the aforementioned habitat requirements (Table 5. 4 ). 113 Table 5.4: MODIS Type 1 LULC classes and their suitability for An. gambiae s.s. Class ID Class Description Suitable An. gambiae s.s. Land Cover 1 Evergreen Needleleaf Forests: dominated by evergreen conifer trees (canopy >2m). Tree cover >60%. No 2 Eve rgreen Broadleaf Forests: dominated by evergreen broadleaf and palmate trees (canopy >2m). Tree cover >60% No 3 Deciduous Needleleaf Forests: dominated by deciduous needleleaf (larch) trees (canopy >2m). Tree cover >60%. No 4 Deciduous Broadleaf Forests: dominated by deciduous broadleaf trees (canopy >2m). Tree cover >60%. No 5 Mixed Forests: dominated by neither deciduous nor evergreen (40 -60% of each) tree type (canopy >2m). Tree cover >60%. No 6 Closed Shrublands: dominated by woody perennials (1 -2m height) >60% cover. No 7 Open Shrublands: dominated by woody perennials (1 -2m height) 10-60% cover. No 8 Woody Savannas: tree cover 30 -60% (canopy >2m). Yes 9 Savannas: tree cover 10 -30% (canopy >2m). Yes 10 Grasslands: dominated by herbaceous annuals (<2m). Yes 11 Permanent Wetlands: permanently inundated lands with 30 -60% water cover and >10% vegetated cover. Yes 12 Croplands: at least 60% of area is cultivated cropland. Yes 13 Urban and Built -up Lands: at least 30% impervious surface area includin g building materials, asphalt and vehicles. Yes 14 Cropland/Natural Vegetation Mosaics: mosaics of small -scale cultivation 40 -60% with natural tree, shrub, or herbaceous vegetation. Yes 15 Permanent Snow and Ice: at least 60% of area is covered by snow and ice for at least 10 months of the year. No 16 Barren: at least 60% of area is non -vegetated barren (sand, rock, soil) areas with less than 10% vegetation. No 17 Water Bodies: at least 60% of area is covered by permanent water bodies. Yes 114 NDVI The Normalized Difference Vegetation Index (NDVI) is a measure of vegetation presence and health (Jensen, 2005) and is calculated as a ratio of the Red and Near - infrared (NIR) spectral bands: Higher values of NDVI are associated with healthier vegetation wher eas lower values typically signal poor vegetative health, or little to no vegetation present. The NDVI measure is used here as an identifier for suitable land areas for larval breeding sites and mosquito development. Vegetation has several functions during and nutrition (Foster & Walker, 2009) . Vegetated areas also provide natural resting sites, particularly for exophagic mosquito species. During resting - periods , shade provided by vegetated cover may in (Debebe, Hill, Tekie, Ignell, & Hopkins, 2018) . V egetative cover is an important factor in the distribution of larval habitats. For example, in contrast to preference for sunlit pools, An. flavirostris is characterized - (Foley, Torres, & Mueller, 2002) . Further, shade from overhanging plants may reduce the risk of predation on mosq uito larvae and provides protecting from surface disturbance (Metzger, 2004) . Studies on the relationship between mosquito species and NDVI are not novel (e.x., Dambach et al., 2012; Juri et al., 2014; Lourenço et al., 2011) . Notably, Kelly - Hope, Hemingway, & McKenzie (2009) described the rela tionship between An. gambiae s.s., An. arabiensis, and An. funestus and NDVI along with other environmental factors in Kenya. Findings showed mean NDVI was significantly correlated with each of the three species; An. funestus was positively correlated with NDVI while An. gambiae s.s., An. arabiensis were negatively correlated with NDVI. Mean NDVI values measured between the three species ranged from .46 . 52. (3) 115 Beyond an association with mosquito distributions, there is a well - established relationship betwe en NDVI and malaria (e.g., Haque et al., 2010; Hay, S.I., Snow, R.W., Rogers, 1998; Sewe, Ahlm, & Rocklöv, 2016) . Hay, Snow, & Rogers (1998) demonstrated that malaria infection was associated with a minimum NDVI threshold of .30 .40 at three sites in Kenya. These findings wer e corroborated by the work of Sewe, Ahlm, & Rocklöv (2016) who showed that values >0.40 were negatively associated with mortality. A conservative >=.30 value is adopted herein to define suitable areas. NDVI data were acquired from the 19 - year MODIS Terra Vegetation Indices 16 - day Global archi ve (MOD13Q1) at a resolution of 250 - m. 5.2.2. 3. 1. 3 Water Precipitation Water is necessary for mosquito dev elopment and survival. To estimate inundated, or likely to become inundated areas that would support breeding, annual mean precipitation was calculated from the Climate Hazards Group InfraRed Precipitation with Station Data (CHIRPS), version 2.0 product at a resolution of 0.05 arc degrees (~5km) . Data are available from Jan 1, 1981 April 26, 2019. Estimating an accurate precipitation range that provides adequate water resources for breeding pool formation to support larval development is challenging. Mosq uito eggs are laid either on or in water, or in areas likely to pool; only a film of water is necessary to support mosquito development through the larval and pupal periods (Foster & Walker, 2009 ) . Lindsay et al. (1998) examined the relationship between climate variables (including precipitation) and the geographic ranges of An. gambiae s.s. and An. arabiensis throughout Africa. Results showed that total annual precipitation n ecessary for An. gambiae s.s. was 330 - 3224mm . A nalysis showed that precipitation values render all of Malawi suitable for An. Gambiae habitat at a =>300mm threshold. Everywhere is suitable at a sub - optimal level. In order to locate the most suitable areas, an annual precipitation amount threshold of >50% (relative to Malawi) was adopted , while still abiding by the annual maximum threshold described by Lindsay et al. ( 1998) (3200mm) . 116 The production of larval habitats should expand and proliferate during periods of seasonal rainfall and accumulated precipitation. Flow Accumulation Flow accumulation (FA) assists in identifying areas prone to ponding based on drainage characteristics of the landscape. These data assist in identifying streams, water channels , and dambo areas, and are necessary for calculating other hydrologic indices used in previous mosquito distribution models (e.g., Topographic Wetness Index (TWI)) (e.g. Alimi et al., 2015; Mccann et al., 2014). In Kenya, McCann et al. (2014) showed that TWI along with distance to nearest stream were the two most important environmental variables for predicting larval habitats for An. gambiae s.l .. FA and TWI differ only slightly; TWI is calculated as a function of FA and slope of a landscape. FA was used here to maximize potential suitability around flow accumulation areas so that gently sloping areas would not be excluded. To delineate probable breeding habitat as a function of FA, the World Wildlife Fund (WWF) HydroSHEDS Flow Accumulation mapping product was used at a spatial resolution of 15 arc-seconds (~500m) opography Mission (SRTM) from Feb 11, 2000 Feb 22, 2000 (Lehner, Verdin, & Jarvis, 2008) . To calculate FA, Digital Elevation Models (DEMs) are used to determine the natural drainage from a given pixel to adjacent, downslope pixels. Based on flow direction, the accumulated flow to each pixel is calculated and returned. (ESRI, 2018). This model assumes that pixels with a greater number of cells flowing into them represent areas of high er likelihood of breeding potential. To that end, suitable areas were determined by setting a threshold of >=25% to identify those pixels of highest FA, which corresponds well with visual interpretations of satellite images in Malawi, 117 Water Bodies A water bodies layer was produced to capture areas along permanent water body (e.g., rivers, streams) margins likely to pool and support vector breeding. Water bodies were identified using the JRC Global Surface Water Bodies Mapping Layer, v1.0 data product. These data were acquired from a composite of >3 million Landsat 5/7/ 8 scenes acquired between March 16, 1984 and October 10, 2015 at a spatial resolution of 30-meters (Pekel, Cott am, Gorelick, & Belward, 2016). Water bodies were buffered by 250-m to locate water-rich soils within proximity to permanent water areas. A 250- m buffer was selected in order to be consistent with the spatial resolution of the flow accumulation layer. Wat er body data were subsequently masked to remove rivers, streams, and lakeshores where larvae habitat is unsupported due to regular disturbance to the water surface via wind and waves. 5.2.2.4 Model Outputs Variable thresholds are used to create binary, suitable (1) vs unsuitable (0) maps for each predictor during the defined time-period based on the mean. All predictor variable maps are then combined using Boolean logic to produce suitability maps for the target species. Results are displayed in two fashions: suitability ranking (low high) and combined suitability by type (Climate, Land, Water) to elucidate suitability drivers (Figures 5.7 5.9). The 2012 product shows that approximately 4. 45% of Anopheles gambiae s.s ., approximately 39.79% for water plus another factor, and 6.19% is supraoptimal, meeting suitability thresholds for water presence, terrain characteristics, and climatic conditions. The 2014 product differs only slightly from the 2012 product: approximately 4.89% of land area meets suitable water conditions to support vector breeding, approximately 34.23% is suitable according to water and climate or water and land criteria, and 6.21% is supraoptimal. Of the three models produced, the 2017 product shows the supraoptimal ( 6.86%). Approximately 5% of Malawi exhibits suitable water conditions, and roughly 35.46% are suitable for water and another factor. In 118 all models, areas that exhibit suitable conditions for land cover characteristics alone are most common: 44.50% (2012), 51.47% (2014), and 46.70% (2017). Figure 5.7 : Anopheles gambiae s.s. habitat suitability in Malawi, 2012 119 Figure 5.8 : Anopheles gambiae s.s. habitat suitability in Malawi, 2014 120 Figure 5.9 : Anopheles gambiae s.s . habitat suitability in Malawi, 2017 121 5.2.3 Malaria in Malawi 5.2.3 .1 DHS MIS Reliable estimations of historical spatio - temp oral distribution of malaria prevalence in Malawi is challenging in the absence of long - standing standardized measures of the burden of disease (Mathanga et al., 2012) . Malaria prevalence data for thi s study was taken from the Demographic and Health Surveys (DHS) Malaria Indicator Surveys (MIS) in Malawi. The DHS Program is the largest sample survey program in history (Zuehlke, 2009) with a mission to provide accurate, nationally representative data on population, health, and nutrition (USAID, 2019b) . Since its inception in 1984, the DHS program has taken part in more than 400 surveys in 90 countries (USAID, 2019b) . Funding for the DHS Program is provided by USAID and is implement ed by ICF (formerly, Inner City Fund) (USAID, 2019a) . Since 2000, the standard DHS survey has included a malaria - related questions including ownership and use of bed nets, prevalence of fever in children, and intermittent preventative treatment of pregnant women (DHS, 2019a) . The MIS is a standalone survey specific to assessing c ore malaria indicators at the national, regional, and provincial levels. MIS surveys were developed by the Monitoring and Evaluation Working Group (MERG) of Roll Back Malaria and were first conducted in 2011 (DHS, 2018) . Data are collected on household own ership and use of Insecticide - Treated N ets (ITNs), diagnostic blood testing via Rapid Diagnostic Testing (RDT) and Microscopy for children under five, indoor residual spraying, treatment of fever in children under five, and intermittent preventative treatm ent for pregnant women (DHS, 2019b) . Three MIS surveys have been conducted in Malawi: MIS 2012, MIS 2014, and MIS 2017. 5.2.3 .2 DHS MIS Data and Processing The DHS MIS data ( https://www.dhs program.com/data/available - datasets.cfm ) and associated GPS data were used to analyze the distribution of malaria prevalence in Malawi for years 2012, 2014, and 2017. Data for each MIS survey year was requested and approved using the registration process 122 ( http://dhsprogram.com/data/new - user - registration.cfm ). Malaria prevalence for all MIS surveys is observed through microscopic detection of malaria parasites (microscopy) and SD BIOLINE M alaria Ag. P.f/Pan (HRP - P.falciparum - specific, histidine - rich protein - 2 (HRP - 2), not the malaria parasite (DHS, 2017) . Moody (2002) showed that HRP - 2 remains in the blood for up to a month despite adherence to an antimalarial regimen. In areas that experience high rates of P. falciparum, this can lead to inflated p revalence rates via RDT detection (DHS, 2017) . For this reason, this analysis defines prevalence as parasitemia (via microscopy) in children ages 6 - 59 months given the holoendemic nature malaria in Malawi . Data are analyzed at the regional and cluster level. Malawi contains three DHS regions: northern, central, and southern. The DHS clusters are geo - referenced groupings of household s that participated in each survey. To protect confidentiality of survey participants, clusters are displaced (Burgert, Colston, Roy, & Zachary, 2013) according to their classification, rural or urban (Perez - Heydrich & Emch, 2013) . Rural locations are displaced 0 5km (with 1%, or every 100 th point displaced 0 10km) and urban locations are displaced 0 2km. Displace ment is random in both direction and distance (Perez - Heydrich & Emch, 2013) . Analysis at the cluster level provides the finest resolution data available for survey years. The location of each cluster varies by survey. Malawi contained 140 clusters (96 Rural, 44 Urban) in 2012, 140 (9 0 Rural, 50 Urban) in 2014, and 150 (90 Rural and 60 urban) in 2017. Prevalence rates for each survey cluster were calculated by using the weighted sum of children ages 6 - 59 months who tested positive for parasitemia via microscopy divided by the total num ber of children tested. All statistical data processing was conducted in STATA /IC 15.1 then exported to ArcMap 10.5.1 . 5.2.3 .3 Malaria Prevalence in Malawi Since the DHS MIS Survey in 2012, Malawi has made progress in controlling Malaria according to M IS reports. Overall prevalence in 2017 was 24.3%, down from 27.7% in 2012. Malaria prevalence 123 data for Malawi DHS MIS Surveys 2012, 2014, and 2017 are provided in Table 5.5. Subsequent sections discuss statistics for each survey year in greater detail. Table 5. 5 : Percentage of children ages 6 - 59 months according to microscopy for Malawi DHS MIS Surveys 2012, 2014, and 2017 Malaria Prevalence according to microscopy (2012) Malaria Prevalence according to microscopy (2014) Malar ia Prevalence according to microscopy (2017) Background Characteristics Microscopy positive Number of C hildren Microscopy positive Number of Children Microscopy positive Number of C hildren Overall Prevalence 27.7 2167 33.2 2023 24.3 2485 Region Northern 19.8 306 28.5 387 11.2 268 Central 34.4 916 36.1 774 26.0 1065 Southern 23.9 944 32.6 861 25.7 1152 Age (in months) 6 - 8 21.1 105 20.6 112 7.6 141 9 - 11 18.4 106 30.4 105 7.3 120 12 - 17 20.0 266 19.0 237 14.8 2 78 18 - 23 22.4 243 28.0 269 26.1 227 24 - 35 30.9 478 35.5 458 22.5 517 36 - 47 30.0 488 37.5 401 29.4 528 48 - 59 32.7 480 41.1 440 31.5 674 Sex Female 28.3 1157 31.7 1009 24.3 1238 Male 27.0 1010 34.6 1014 24.3 1246 Residence Urban 9.4 280 10.8 275 4.0 339 Rural 30.4 1887 36.7 1748 27.5 2145 2012 The 2012 DHS MIS Survey was conducted from in April and May 2012 and included 3500 households across 140 clusters (DHS, 2015) (Figure 5.10 ). A total of 2167 children were sampled for malaria infection via blood smear for microscopy identification of malaria parasites. Of these, 1566 (72.3%) 124 were negative and 601 were positive (27.7%) for infection; overall prevalence was 27.7%. A larger percentage of children in rural areas (30.4%) experienced malaria infection than children residin g in urban areas (9.4%). The largest percentage of positive malaria infection recorded, was 81% and located ~30 - k ilometers southwest of Lilongwe. Of the remaining nearly 18% of clusters with prevalence rates =>50%, all were concentrated in the Southern and Central regions. R egional analysis showed that the Northern region experienced the lowest overall rates at 19.8%. The Southern and Central regions overall w ere 23.9% and 34.4%, respectively. 2014 The DHS MIS survey was conducted from May 2 June 10, 2014 and included 3500 households (DHS, 2015 ) (Figure 5.11 ) . Overall malari a prevalence in children ages 6 59 months declined by 10% from the 2010 recorded statistics (43%) (DHS, 2015) . There were 2023 children that participated in the 2014 Malawi DHS MIS. Of these, 1352 (67%) were negative and 671 were positive (33%) for infection. The number of ch ildren positive for malaria infection was higher in rural areas (37%) than urban (11%). Prevalence rates ranged from 0 to 100% among the 140 clusters sampled. Overall prevalence for 2014 was 33.2%. Regional analysis shows that the Central region had the hi ghest infection rate (36.1%), followed by the Southern (32.6%), and Northern (28.5%) regions. Roughly 15% of survey clusters had prevalence rates >=50%. Five clusters had prevalence rates >80%; two were located along Lake Malombe in the Southern region, 1 near Nkhotakota on the Lake Malawi Lakeshore, and the remaining two west of Mzimba along the Zambian border. 2017 Data collection for the 2017 DHS MIS Survey was conducted from April 15 June 16, 2017 among 3750 households (DHS, 2017) (Figure 5.1 2 ) . The total number of children that participated in this analysis was 2485. Microscopy results showed that 1592 children were negativ e (64%) and 892 (36%) were positive for malaria infection; P. falciparum accounted for 95.3% of infection. Consistent with 125 previous surveys, a larger percentage of children were positive for malaria in rural as opposed to urban areas. There were 150 cluste rs analyzed with a minimum prevalence of 0 and maximum of 83.3% . Overall prevalence was 24.3%, a considerable decrease from 43% in 2010 (DHS, 2017) . Clusters number 3 and 135 were missing geographic information and were subsequently were removed from the dataset prior to export to ArcMap 10.5.1 for visualization. Regional analysis showed that the Southern and Central regions had ne arly identical malaria prevalence rates, 25.7% and 26%, respectively. The percentage of children who tested positive in the Northern region was 11.2%, a decline of more than 17% from the previous 2014 MIS survey. Roughly 7% of survey clusters had prevalen ce rates >=50%; more than a 50% decline from 2014 (16%). Of these, half are located along the Lake Malawi l akeshore, with one of the remaining clusters near Mt. Mulanje in the southern region , % of clusters where malaria Blantyre. 126 Figure 5. 10: Percentage of Children Age 5-59 Months Microscopy Positive for Malaria Infection by DHS Cluster (2012). 127 Figure 5. 11: Percentage of Children Age 5-59 Months Microscopy Positive for Malaria Infection by DHS Cluster (2014). 128 Figure 5. 12: Percentage of Children Age 5-59 Months Microscopy Positive for Malaria Infection by DHS Cluster (2017). 129 5 .3 Results 5 .3.1 Habitat Suitability and Malaria Prevalence Malaria prevalence in Malawi has a distinctly spatial character , in part a result of environmental characteristics that def ine suitable ecological zones for malaria mosquitoes. To examine these relationships, malaria prevalence data were overlaid for each DHS MIS year with the associated years produced HSM. Suitability drivers associated with each DHS cluster were assessed by buffering clusters to 5 - km for consistency of results per the recommendations of Perez - Heydrich & Emch (2013) . The 2012 pro duct shows that among DHS clusters, habitat suitability is most often defined as land areas exhibiting suita ble land characteristics ( 44.28 % Rural; 55.89 % Urban), followed by areas suitable for land and water characteristics ( 42.05 % Rural; 30.90 % Urban) (Figure 5.13 ) . Among clusters surveyed for the 2012 year, areas exhibiting habitat suitability based on climat e alone were least likely (0. 22 % Rural; 0. 40 % Urban). This is unsurprising given the overall percentage of land area suitable based only on climate is low (. 52 %). Supraoptimal (Land, Climate and Water) areas represented 7.14 % of land area within rural clus ters and 5.39 % of land area with urban clusters respectively. Habitat suitability drivers among 2014 DHS clusters show similar findings to the 2012 results (Figure 5.14 ) . Suitable land areas as defined by land characteristics cover the largest areas : 5 2. 14 % in rural areas and 68.54 % in urban. Suitable areas based on land and water characteristics alone represent 33.40% of rural areas and 21.02% of urban. Clusters that include supraoptimal areas are nearly four times greater in rural areas (8. 8 0%) as oppos ed to urban ( 3.69 %). The 2017 product shows supraoptimal areas for rural and urban areas as 7.90% and 6.28%, respectively (Figure 5.15 ). Consistent with previous years analysis suitable land areas according to land characteristics are greatest for both rural and urban areas ( 53.18 % rural; 68.93 % urban), followed by suitable areas defined by land and water characteristics ( 31.85 % rural and 18.08 % urban). 130 Figure 5. 13: Habitat suitability drivers by MIS cluster, 2012 131 Figure 5.14: Habitat suitability drivers by MIS cluster, 2014 132 Figure 5. 15: Habitat suitability drivers by MIS cluster, 2017 133 5 .3.2 Scaling Irrigation and Malaria Risk in Malawi To examine the expansion of Anopheles gambiae s.s . niche as a result of LULC C for irrigated agriculture, a n estimat io n of maximum extent of habitat suitability was produced using the results of the 2017 habitat suitability model and potential GBI area data (Figure 5.16 ) . The GBI area pixels were re - classified to represent where suitable land and water characteristics for breeding would occur upon LULCC for irrigated agriculture . At the national scale, r esults show that while the 2017 product demonstrated that approximately 6.86% (~6 .7 million km ² ) of Malawi was suprao ptimal for breeding , this increases to 7.06% (~ 6.9 million km ² ) under the GBI scenario. Notable changes also involve areas was suitable based on la nd and water characteristics . T his number increases to 40.66% with the expansion of GBI sites. Under the GBI scenario, the percent area for habitat suitability and total maximally suitable area in square kilometers was assessed for each of the 25 GBI pro posed sites (Table 5.6 ). Results show that the Limphasa site will possess the largest percentage of total area maximally suitable according to water, climate, and land characteristics for An. gambiae s.s. breeding (78%) w ith a maximally suitable area of 35 9 km ² . Other notable sites according to total maximally suitable land area are Karonga (43.7%; 123.94 km ² ) and Bua/Loz i (55.5%; 104.29 km ² ). Eight sites demonstrate no maximally suitable area: Chitipa, Mpherembe, Upper Dwangwa, Kantungu, Bua Dambos, Lisunw i, Phwadzi, and Lingadzi. 134 Figure 5. 16: Change in distribution of habitat suitability for An. gambiae s.s. after LULCC for GBI sites under a maximum estimation 135 5 .4 Discussion and Conclusions As has been highlighted in this chapter, estimated habitat suitability under the GBI show s an increase of nearly 200,000 sq km in supraoptimal breeding area for An. gambiae s.s. mosquitoes across Malawi. Analysis of GBI sites showed that while eight sites would not possess any supraoptim al land area, the remain der of sites would possess 0.01 359 km ² of supaoptimal land area for breeding. In light of the findings of Chapter 3, a notable consideration is that irrigated spaces are not homogenous; risk of breeding will be asymmetrical based o n site characteristics including timing and intensity of irrigation, soil properties, management, and crop type. Likewise, it would an oversimplification to conclude that simply an increase in the likelihood of aggregated surface water for breeding would i ncrease malaria transmission in the local area without site specific information including distance of the irrigated scheme to human habitation and the stability of malaria transmission in the local area. Using data from this study on GBI sites including proposed crop types, habitat suitability under and estimated GBI scenario, and malaria prevalence data for 2017 where clusters were <5km from proposed schemes , a ranked cate gorization of GBI sites based on their lik e lihood to increase malaria risk through production of larval habit at is presented (Table 5.7 ). Recommendations were determined through a three - step process. First, by considering total supraoptimal land area for each site, followed by assessment of proposed crop type . Proposed crop types includ e cereal grains (wheat, maize ), rice, sugar, and cassava. Other GBI sites are intended for ranching, rather than cultivation. Crop types were ranked according to the ir water needs and traditional methods of irrigation (1 5, rice, sugar cane, maize, wheat, cassava; most likel y to encourage vector breeding to least likely ) . Ranching operations were ranked least likely to encourage vector breeding due in part to limited aggregated water bodies on site but also the breeding preferences of An. g ambaiae s.s. mosq uitoes for open, sun lit pools generally free of organic matter ; these conditions are likely not readily met on ranc hing operations. It is assumed tha t areas with crop types such as rice that are irrigated through 136 consistent flood irrigation would encourage greater vector production than crop types sensitive to water logging. Where more than one crop type was listed per site, the crop type with the greatest water needs was considered in the absence of data for percent area or seasonality of cultivation for each crop. The final consideration in ranking re commended sites to develop was the prevalence of malaria within 5-km of the proposed site. Sites with no MIS clusters that meet the <5km criteria are listed separately. An alternative method for gathering data on malaria prevalence for these areas would use . The GBI site whose LULCC for irrigated agriculture is least likely to substantially impact malaria transmission is Chitipa. Chitipa possess no maximally suitable area for breeding and the area is intended to cultivate wheat. In contrast to a crop such as rice where flood irrigation is often practiced, wheat is sensitive to waterlogging (Herzog, Striker, Colmer, & Pedersen, 2016). It is expected that aggregated pools of water suitable for breeding would not exist long enough for mosquitoes to develop from aquatic to adult stages under these conditions. Further, two 2017 MIS clusters were located <5km of the Chitipa site with no detectable malaria parasitemia via microscopy. Other sites that demonstrate very little or no maximally suitable land area for breeding are Mpherembe, Upper Dwangwa, Kantungu, Linggadzi, Bua Dambos, Lisunwi, Phwadzi, and South Rukuru. Each of these sites proposed crop types are cereal grains (maize or wheat), or are intended as ranching operations limiting the availability of consistent surface water available for breeding. The number of MIS clusters <5km from each of these sites ranges from 0 6; mean percentage of malaria prevalence among these clusters ranges from 0 41.2%. Five sites had no MIS clusters <5km: Phwadzi, Lingadzi, Lisangadzi, Lisunwi, Kantungu, and Liviridzi. The GBI site most likely to increase malaria risk for those residing in closest proximity to the scheme is Limphasa. Total maximally suitable area for An. gambiae s.s. is approximately triple that of remaining sites (359 km ²) and land area is intended for rice cultivation. According to 2017 MIS surveys, 137 mean malaria prevalence is 22.4% for the four clusters within 5-km of the scheme. These data indicate that malaria mosquitoes are clearly present within the area. It is expected that the increase in surface water availability from Limphasa would encourage the proliferation of mosquitoes and in turn change the epidemiology of malaria for the local area. Other GBI sites where greater consideration should be given to development in light of their impact on malaria risk are Karonga/Kaporo, Bua/Luzi, Muona/Ruo, Phalombe Plain, Lake Malombe, Lingadzi/Lipimbi, Hara, and Ntchalo. Beyond the amount of land area maximally suitable for breeding potential, each of these sites are intended for rice to some measure. Further information on the percentage of land area within each site intended for rice cultivation and the seasonality of planting would assist in further defining risk potential for the area. The number of MIS clusters for these sites is 0 5; prevalence rates range from 0 27.2%. Two sites require further investigation on malaria disease dynamics for the local area in the absence of MIS data: Chia and Nyachipere. 138 Table 5. 6 : Percent area of habitat suitability for proposed GBI sites under a maximum estimation % Area Habitat Suitability GBI Site Proposed Crop Type Proposed Area (km²) No Data Unsui table Water Climate Land Water +Climate Water + Land Climate + Land Water + Climate + Land Total Supraoptimal Area (km²) Chitipa Wheat, Maize 412.282 10.9 - - - 0.007 - 89.12 - - 0 Karonga/Kaporo Rice 283.617 6.0 0.03 - - - - 50.2 - 43.7 123.94 Hara Ri ce 275.268 6.7 - - 0.01 0.1 - 80.4 0.02 12.8 35.23 Bolero/Kazuni Maize 178.943 0.002 0.002 - - 0.1 - 98.6 - 1.3 2.33 Mpherembe Ranch 331.593 0.005 0.005 - - 0.1 - 99.9 - - 0 Limphasa Rice 460.258 13.5 - - - - 0.002 8.56 - 78.0 359.00 South Rukuru Maize 2028.11 4.8 - - - 0.03 - 95.13 - 0.0005 0.01 Upper Dwangwa Wheat, Maize 526.367 0.02 - - - 0.1 - 99.9 - - 0 Bua/Lozi Rice, Maize, Cassava 187.912 27.5 - - - - - 17.0 - 55.5 104.29 Chia Rice, Maize, Cassava 129.32 32.0 - - - 0.03 - 20.43 0.01 47.5 61.43 Kantungu Wheat 206.575 - - - - 0.1 - 99.9 - - 0 Lingadzi/Lipimbi Rice 525.127 18.32 - - - .03 - 75.94 .0007 5.71 29.98 Bua Dambos Wheat, Maize 2353.69 0.42 - - - 0.02 - 99.6 - - 0 Bwanje/Malembo Rice 471.85 14.22 - - 0.0004 0.06 - 82.6 .001 3.13 14.77 Lisangadzi Sugar 153.641 - - - - 0.07 - 86.6 0.02 13.27 20.39 Liviridzi Wheat 109.909 - - - - 0.11 - 96.1 0.02 3.81 4.19 Lake Malombe Rice 88.4879 10.52 - - 0.01 0.06 - 31.5 0.03 57.9 51.23 Lisunwi Wheat 117.766 0.22 - - - 0.11 - 99.7 - - 0 Lake Chil wa/Chiuta Rice, Sugar 884.768 50.2 - - - 0.02 - 48.3 0.002 1.48 13.09 Phalombe Plain Rice, Maize, Sugar 679.408 33.5 - 0.001 - 0.03 - 58.63 0.0006 7.81 53.06 Phwadzi Ranch 87.463 0.02 - - - 0.10 - 99.9 - - 0 Ntchalo Rice, Maize, Sugar 108.087 2.92 - - - 0.12 - 88.93 0.01 8.02 8.67 Muona/Ruo Rice, Maize 307.162 16.63 - - - 0.03 - 54.3 - 29.10 89.38 Nyachipere Rice, Maize 127.416 11.45 - .0007 - 0.04 0.006 80.65 - 7.90 10.07 Lingadzi Ranch 45.7031 - - - - 0.17 - 99.83 - - 0 139 Table 5. 7 : R anked categorization of GBI sites based on their likelihood to increase malaria risk through production of larval habitat Habitat Suitability Malaria Crop Type Rank GBI Site Proposed Area (km²) % Max Suitable Area Total Max Suitable Area (k m²) Number of MIS Clusters <5km Mean % MIS Cluster Prevalence <5km Proposed Crop Type Sites to Develop: Least l ikely to increase malaria risk 1 Chitipa 412.28 0 0 2 0 Wheat 2 Mpherembe 331.59 0 0 1 5.3 Ranch 3 South Rukuru 2028.11 0.0005 0.01 4 18.0 Maize 4 Bua Dambos 2353.69 0 0 6 19.4 Wheat, Maize 5 Bolero/Kazuni 178.94 1.3 2.33 3 9.1 Maize 6 Upper Dwangwa 526.37 0 0 1 41.2 Wheat, Maize Malaria prevalence information needed - Phwadzi 87.46 0 0 0 n/a Ranch - Lingadzi 45.703 0 0 0 n/a Ranch - L isangadzi 153.64 13.27 20.39 0 n/a Sugar - Lisunwi 117.766 0 0 0 n/a Wheat - Kantungu 206.58 0 0 0 n/a Wheat - Liviridzi 109.91 3.81 4.19 0 n/a Wheat More consideration need: Most likely to increase malaria risk 1 Limphasa 460.26 78.0 359.00 4 22.4 R ice 2 Karonga/Kaporo 283.62 43.7 123.94 5 13.3 Rice 3 Bua/Lozi 187.91 55.5 104.29 2 27.2 Rice, Maize, Cassava 4 Muona/Ruo 307.16 29.10 89.38 1 0 Rice, Maize 5 Phalombe Plain 679.41 7.81 53.06 1 14.3 Rice, Maize, Sugar 6 Lake Malombe 88.488 57.9 51.23 1 0 Rice 7 Lingadzi/Lipimbi 525.13 5.71 29.98 3 25.3 Rice 8 Hara 275.27 12.8 35.23 1 5.0 Rice 9 Ntchalo 108.09 8.02 8.67 1 10 Rice, Maize, Sugar Malaria prevalence information needed - Chia 129.32 47.5 61.43 0 n/a Rice, Maize, Cassava - Nyachipere 12 7.42 7.90 10.07 0 n/a Rice, Maize Special Consideration - Bwanje/Malembo 471.85 3.13 14.77 0 n/a Rice - Lake Chilwa/Chiuta 884.77 1.48 13.09 0 n/a Rice, Sugar 140 There are t wo sites in need of special consideration . The Bwanje/Malembo site i nvolves a pl a nned extension of the Bwanje Valley Irrigation Scheme (BVIS) that is currently undergoing significant changes to its agroecology . In late 2018 constr u ction of the Bwanje Dam was completed (Warm Heart News, 2018) with the intention of supplying sufficient water resources to 6 00 - ha of the 800 - ha scheme throughout the dry season (M. Mafosha, personal communication, March 27, 2019 ). Traditio na lly farmers at BVIS ha ve grown rice during the rainy season and maize, cowpea, and beans over only 300 - ha during the dry season due to limi ted water resources. It is anticipated that during the 2019 dry season that 600 - ha of the scheme will be cultivated with rice, while the remaining 200 - ha will include conve n tional dry season crops (M. Mafosha, personal communication, March 27, 2019) . These changes will impact malaria disease dy na mics of the area. Further, while no MIS cl usters were located within 5 - km during the 2017 surveys, cross - sectional mal a riometric surveys of fourteen villages within 6 - km of BVIS were conducted in April 2016 and 2017 by Mangani et . al (unpublished data). Result s show that household distance to BVIS is a significant predictor of malaria risk. Prevalence of infection across participants (N = 5489 ) within 3 - km of the scheme in 2016 was 33. 3 %; households 3 6kms was 2 5.9 % . The intended changes to cropping practice at BVIS afforded by the Bwanje Dam present a n opport unity for investigating the immediate changes to vector breeding and subseque nt malaria transmission within the area prior to intended expansion under the GBI. The other GBI site in need of special consideration prior to development is the Lake Chilwa. During analysis it was noted that many of the s ite boundaries digitized from the Irrigation Master Plan data either did not line up exactl y with the Malawi boundar y in GEE or portions of the sites were not included in the GEE analysis due to their being classified as water bodies from the JRC Global Surface Water Bodies data and were subsequently masked out during processing . For this reason, many of the GBI sites include For the Lake Chilwa site, >50% of total area is Further analysis should 1) seek to impro ve the accuracy to of the Lake Chilwa 141 site relative to Lake Chilwa; and 2) examine the JRC water bodies data layer and digitize if necessary a water bod ies layer that is a better representation of water body margins. The Government of Malawi (GoM) has a long - standing history of developing policies and initiatives aimed at increasing irrigation not only to miti gate food insecurity concerns, but improve overall rural livelihoods and boost economic growth (e.g., GoM, 2000b, 2015b, 2016) . Through the GBI, Malawi has adopted the most aggressive expansion plan for irrigated agriculture to date; ne arly infection. As irrigation continues to expand across Malawi, geographically informed analysis of GBI sites in relationship to habitat suitability and historical malaria prevalence can assist in estimating to what extent the epidemiology of malaria will change with LULCC for irrigated agriculture . These data can elucidate s ites in need of further consideration and provisioning in an effort to improve food security through irrigation without exacerbating malaria risk. 142 C HAPTER 6 6.1 Introduction Development of irrigated agriculture will not occur wit hout at least some sort of implicit scaling. Yet, this dissertation up research and development and natural resou rce management literature lacks a common ontology. Working toward irrigated agricultural solutions that miti gate food insecurity while simultaneously inhibiting the production of adult stage mosquito vectors will involve a wide range of actors embedded within myriad social and environmental systems. U ncertainty on what it means to scale up either a product or p rocess for irrigated agriculture should not be a barrier to meeting the critical health and food security needs being addressed through irrigation interventions. In this regard, this dissertation not only provides a conceptual framework for defining scali ng up and putting it in to pra ctice for development , but also contributes new knowledge on the impact of scaling irrigated agriculture on spatio - temporal malaria disease dynamics . Primary contributions include : (1) Designed a conceptual framework to aid in precision of terminology for development activities; (2) Demonstrated marked differences in malaria risk for those residing in close proximity to irrigated schemes as a product of the heterogeneity of irrigated spaces; (3) Showed that irrigated agricultur e is a driver of spa t io - temporal change in the geography of malaria risk that occurs during dry seasons; (4) Demonstrated that irrigated agricultural drivers of LULCC are associate d with mosquito production; (5) LULCC for irrigated agriculture will change the geography of suitable Anopheles gambiae s.s habitat; (6) Scale matters for malaria risk exposure; (7) Developed an open - source, dynamic habitat suitability model in Google Earth Engine (GEE) for end - users to model habitat suitability for any mosquito s pecies; and (8) Developed a policy protocol to target irrigation implementations while simultaneously mitigating malaria risk. 143 6.2 Overall Contributions 6.2.1 Toward a Common Ontology of Scaling Up in Development Chapter 2 explores the lack of ontolog Development (R&D) and Natural Resource Management (NRM) literature. This dissertation is conducted against a backdrop of policies and development initiatives on the part of government and internati onal development partners aimed at scaling irrigated agriculture to mitigate food insecurity and boost economic growth. Ontological ambiguity devalues scaling up by contributing not only to the failure of these programs to scale as both a product and process . Ultimately , how scaling up is conceptualized impacts how it is operationalized, influences monitoring and evaluation, and affects program success. To explore the varying definitions of scaling up in development literature, definitions from available Consultative Group on International Agricultural Research (CGIAR) Centers, the United States Agency for International Development (USAID) and the International Fund for Agricultural development (IFAD) w ere analyzed. Findings suggested that across institutions there w as a distinct categorization s combination s of Interventions, Mechanism, or Outcomes. Or, more directly, what is being scal ed, how is scaling verb and a noun. The verb form is demonstrated directly in the mechanisms of scaling up. As a noun, to take an innovation to scale im plies project success; there is greater emphasis on project outcomes. Rather than call for improved definitions, a conceptual framework was constructed to assist in how to define scaling up through separation of actions related to process (verb) and outcom es (noun). Importantly, the model also emphasizes the necessary role of monitoring and evaluation on both the innovation being brought to scale and scaling up efforts. The implementation of the proposed framework not only works to define a clear pathway fo r s uccess but allows for critical examination of value judgments regarding development to be examined. 144 6.2.2 LULCC for Irrigated Agriculture and it s Impact on Mosquito Distribution Aim #1: Address land cover and land use decisions and their impact on the spatio - temporal structure of agricultural growth and mosquito d istribution by: 1. Describing the LULC of BVIS and the Bwanje Valley its impact on b reeding pool formation through development of a land classification s system for irrigated agriculture in r ainy and dry seasons Aim # 2 Demonstrate the influence of irrigation schemes for agriculture on mosquito breeding pool formation and persistence by: 1. Modeling breeding pool scenarios based on spatio - temporal, environmental, and anthropogenic charact eristics at BVIS under three scenarios 2. Describing the association between LULC and breeding potential at BVIS contrasted with the 8 - km area surrounding the scheme Scaling irrigated agriculture to enhance crop productivity is a common tool for ensuring food security. Yet, LULCC for irrigated agriculture impact s biotic interactions within ecosystems that have been shown to encourage vector and pathogen transmission, including malaria (Boelee & Madsen, 2006; Ijumba & Lindsay, 2001; J.M.Hunter, L.Rey, K.Y. Chu, E.O.Adekolu - John, 1993; Patz, Graczyk, Geller, & Vittor, 2000 ) . The associat ion between irrigated schemes and the proliferation of malaria mosquitoes is well documented ( see e.g. Ijumba & Lindsay, 2001; Kibret et al., 2010) , yet irrigated spaces are treated as homogenous spatial units when in fact important distinctions can be made. Land cove r, specifically crop type, engineering, water management, and scheme management play critical roles in defining risk potential for breeding pool formation and persistence within irrigated spaces. The aim of Chapter 3 was two - fold: first, describe the resul ts of a characterization study conducted at the BVIS second, to generate spatio - temporal breeding scenarios at BVIS to illustrate how perturbations to 145 irrigated syst ems influence the distribution of breeding site potential and overall disease ecology for the local area. The distribution of aggregated water bodies in an irrigated space is influence d by a number of factors. To that end, BVIS was characterized using mu ltiple sources of information including: (1) satellite imagery from the SPOT - 6 sensor at two time periods to assist in LULC classification; (2) dry season s; (3) Soils data from JICA ( 1994) ; and (4 ) onsite interviews conducted with BVIS personnel at three time periods. Field based surveys of BVIS demonstrated that the spatial d istribution of water bodies was influenced by seasonality, soil properties, timing and intensity of irrigation, drainage, la nd cover, crop water requirements, and management. Projected distributions of breeding were constructed under three scenarios: rainy season, dry season with limited water resources, and dry season with abundant water resources. Importantly, while each mode l depicts similar areas of maximally suitable area for breeding, the abundance of adult stage vectors produced within these areas is seasonally variable. It is expected that during prototypical dry seasons with limited water resources, the number of vector s produced will be lower than the number produced from the same area during the rainy season and dry season with abundant water resources. This is a direct result of water availability and crop water requirements. The area of highest breeding potential dur ing the dry season is occupied by maize, a crop that unlike rice does not require continuous flooding , limiting breeding potential . R ice is cultivated throughout the same area during the rainy season and is the proposed crop type to be cultivated once wate r resources from the Bwanje D am are available during the dry season. Rice dependency on water provides a consistent, favorable environment for mosquito oviposition and subsequent rearing of adult stage vectors (IRRI, 1988) . Results elucidate how the heterogeneity of the landscape bo th in land cover and in land and water management influence the spatio - temporal d istribution of risk for mosquito breeding. 146 Chapter 4 addressed how LULCC for irrigated agriculture changes the distribution of breeding risk for malaria mosquitoes across la ndscapes and in turn, malaria disease ecology for local areas during the dry season. Irrigated agriculture provides a stable source of surface water availability throughout the dry season; a period when mosquito breeding is often limited . To estimate breed ing potential, the study had two primary objectives: (1) Develop a dry season LULC clas sification system through field - based surveys of the Bwanje Valley; and (2) Construct a breeding pool suitability model to differentiate breeding risk potential between defined at the 8km area surrounding BVIS. An 8km distance gave consideration to the maximum recorded flight distance of the Anopheles gambiae s.s. mosquito, while also providing adequate es timation of LULC attributable to BVIS and the Bwanje Valley that may promote mosquito development. Findings demonstrated that suitable breeding areas throughout the Bwanje Valley had a distinct spatial structure. Categorization of suitability showed suprao ptimal areas for mosquito breeding were concentrated within BVIS. In comparison, the surrounding Bwanje Valley ranged from merely satisfactory to unsuitable. Models demonstrated the asymmetrical breeding potential for Anopheles mosquitoes as a result of ir rigated agriculture. In turn, not only does irrigation change the geography of mosquito breeding, but seasonal malaria disease dynamics through lengthening of transmission windows and risk potential for those residing in close proximity to irrigated spaces . 6.2. 3 Malaria Vulnerability and Dynamic Changes to Irrigated Agriculture Aim #3: Assess the impact of dynamic changes in irrigated agriculture on malaria vulnerability in Malawi by: 1. Examining the spatio - temporal change in irrigated agriculture at the national scale 2 . Describing habitat suitability for Anopheles gambiae s.s. mosquitoes in Malawi through construction of a habitat suitability model in Google Earth Engine 147 3. Address ing the historical and plausible future impact on malaria vulnerabi lity driven by scaling irrigated agriculture spurred by national policy frameworks Irrigation Master Plan, irrigable area is intended to expand to 220,00 0 hectares (ha) by 2035; 104,299 ha had been developed by 2015 (GoM, 2015b) . Further, through the Green Belt Initiative (GBI), the GoM has committed to offering nearly 1 million hectares for irrigation development. Chapters 3 and 4 highlight the impact of irrigat ed agriculture on mosquito breeding pool formation and persistence. It is expected that the production of adult stage Anopheles mosquitoes from irrigated areas will result in greater risk exposure to infectious bites for those living in close proximity to irrigated schemes compared with those living further away. Further, where irrigation is conducted during the dry season, malaria prevalence will be higher for those living nearer to irrigated sites. The impact of expansion of irrigated agriculture on the s patial distribution of malaria vulnerability in Malawi was examined by: (1) In vestigating the historical and current (2015) distribution of irrigated agriculture in Malawi; (2) Developing a habitat suitability model in Google Earth Engine (GEE) for the mal aria vector, Anopheles gambiae s.s . to determine suitable landscape for t he species across Malawi; (3) A ssessing malaria prevalence data from the Demographic and Health Survey s (DHS) Malaria Indicator Survey (MIS) at the cluster level ; (4) Examining relat ionships between habitat suitability drivers and malaria prevalence from the 2012, 2014, and 2017 MIS s urvey s ; and ( 5) M odel ing maximum estimations of habitat suitability using the produced habitat suitability model for 2017 and potential GBI area data. S tudy findings sho w habitat suitability drivers among MIS clusters for 2012, 2014, and 2017 are most often associated with suitable land characteristics (NDVI and Land Cover) , followed by the combination of suitable land and water (precipitation, water bodi es, and flow accumulation) 148 characteristics. Areas described as supraoptimal for breeding occurred most often in rural, as opposed to urban clusters for each survey year. These areas exhibited suitable conditions based on climate, land, and water characteristics. A maximum estimation of habitat suitability using the results of the 2017 habitat suitability model and potential GBI area data elucidated the expansion of An. gambiae s.s . niche driven by LULCC for irrigated agriculture. Notable changes included the expansion of supraoptimal area from 6.86% (~6.7 million km²) in the 2017 product to 7.06% (~6.9 million km²) under the GBI scenario. Further, the expansion of areas suitable according to land and water characteristics from 24.90% in the 2017 product to 40.66% in the GBI scenario. These data along with information on proposed crop type for each GBI site were used to provide a ranked categorization of GBI sites based on their likelihood to increase malaria risk through the production of suitable breeding sites for malaria mosquitoes. Twelve sites were categorized as least likely to increase malaria risk, eleven as most likely to increase malaria risk, and two where additional consideration is needed. As irrigation expands under the Irrigation Master Plan in Malawi, geographically informed analysis of irrigation sites in relationship to habitat suitability and historical malaria prevalence can assist in estimating to what extent the epidemiology of malaria will change with LULCC for irrigated agriculture. Further, what re-consideration or provisioning may be necessary to mitigate malaria transmission risk through the production of Anopheles mosquitoes. 6.2.4 Measuring agreement of predictive mosquito habitat This work produced models of suitable breeding habitat for malaria mosquitoes at different spatial units and scales. Results of the Malawi habitat suitability model in GEE presented in Chapter 5 are at a 250-m resolution. Habitat suitability models at BVIS (Chapter 3) and the Bwanje Valley (Chapter 4) are at a 6-m resolution. To estimate agreement of model outputs for the GEE and Bwanje Valley (BV) model, the Kappa ( ) coefficient was used after clipping the GEE model to the Bwanje Valley perimeter . Confusion matrices were produced in . and reference (test pixel) data for 149 each class were selected from the BV map; 60 test pixels were selected per class (see e.g. Jensen, 2005). The coefficient value indicates poor agreement between the two models ( 0.07 ) , this is in part a result of differences in spatial resolutions but also data product inputs for each model. Visual inspection scale . The GEE model shows four suitability classes: unsuitable, suboptimal, marginal, and satisfactory . The majority (80.1%) is classified as suboptimal for breeding, with 17.7% classified as Marginal. Marginal areas in the GEE model are strongly influenced by the Flow Accumulation variable, following river and ephemeral stream channels in the area. Less than 1% is satisfactory for breeding, located in the far western portion of the Bwanje Valley. A notable difference between the GEE and BV model is the absence of optimal and supraoptimal classified area in the GEE model. This is due to the absence of land use information on irrigated area encoded within the LULC data used to construct the GEE model. It is important to note that while global scale data are useful, they do not provide specific local scale information. Irrigation at BVIS criti cal to de fining suitable breeding habitat in the Bwanje Valley , yet it is not reflected in the GEE model. Contrastingly, t he LULC data product used in the construction of the BV model were collected from in - situ field measurements of the Bwanje Valley whic h provide a more accurate estimation of LULC across the Bwanje Valley, and thereby representation of mosquito breeding opportunity. Finally, there are distinct differences in temporal perio ds for each model . The GEE habit suitability model is constructed o n an annual, not seasonal basis. The BV model is a representation of habitat suitability during the dry season at the Bwanje Valley. 6.3 Future Research 6.3.1 Interpretations of Scaling and Their Influence on D evelopment How institutions define scaling up influences how scaling is operationalized (i.e., the intervention and/or pathway to scaling), influences monitoring and evaluation, and affects program success. As was discussed in Chapter 2, scalin g up lacks ontological commitment across R&D instituti ons. With 150 this being the case, it would not be unreasonable to assume that how scalin g programs are concep tualized by different governments vary too. The research and conceptual model presented in Chapter 2 provides a framework for engendering further cons ideration on the precision of scaling up terminology. Multiple development partners and foreign governments have partnered with Malawi to scale irrigated agriculture (e.g., USAID, JICA, European Union). Future work could present commonalities and discrepan cies in the meaning of scale, not only over time, but among institutions and partners to describe how conceptions of scale have and could continue to influence development interventions. 6.3.2 Heterogeneity in Irrigated L andscapes This dissertation found that not only do irrigated agricultural spaces create asymmetrical breeding potential for malaria mosquitoes across landscapes, but even within irrigated spaces, the distribution of breeding risk is heterogeneous. The aggregation of water bodies suitable f or oviposition in irrigated schemes is the result of a number of factors . As a result, there is no single solution to mitigating and controlling vector populations. Engineering, water management, scheme management, and crop selection and geography individu ally and collectively influence where and to what extent mosquitoes will reach adult stage. To effectively irrigate for mosquito control , the irrigation system as a whole from movement of water from irrigation source, to field application, timing and inten sity of watering c oincident with soil properties, ponding potential, and root water uptake of plants must be given adequate consideration. Mosquito larval surveying was conducted at the Bwanje Valley Irrigation Scheme in May 2017 and March 2019 to coincid e with r ice growing stages; mature and newly transplanted , respectively . Sampling was designed with specific consideration for irrigation structure at BVIS . Sample transects followed alternating tertiary canals across the scheme (N=41). At each transect, r esearch assistants sampled at 30m intervals to coincide with average plot size. Collection sites were recorded using a 151 handheld Global Positioning System (GPS), then inspected for water status: Present/Absent; Moving/Still, and stage of rice growth within were s ampled with a standard 350 ml mosquito dipper (Kenea et al., 2011; Kweka et al., 2012) with maximum of ten dips. Larvae were transferred to a container and identified on the basis of larval and pupal characteristics, and counted. While present, Culicines w ere excluded from analysis. For the survey conducted in March 2019, all sites were sampled 2 3 days after irrigation had been applied to plots. Future work will cross reference the results of these surveys with the modeled predictions of breeding pool risk produced in Chapter 3. The research in Chapter 3 provides an initial step in estimating the spatial distribution of breeding pools within irrigated agricultural schemes. However, at present these estimates are static and do not reflect the influence of changes in crop phenology, or timing and intensity of irrigation. As is discussed throughout this dissertation, Anopheles mosquitoes have characteristic breeding preferences and vary in their efficiency in transmitting malaria . In Chapter 3, the difference s related to vegetative growth stage and An. funestus , An. gambiae , and An. arabiensis are described. Developing pixel - level time series of phenology would not only assist in measuring spatio - temporal changes in vegetation dynamics and breeding potential b ut would be useful in determining the specific mosquito species breeding within the irrigated space as well. Beyond phenology alone, future stud ies should consider the use of M on te C arlo simulations to model the probability of breeding pool potential based on perturbations to the irrig ated system including crop type(s) and their spatial arrangement within the irrigated space, timing and intensity of irrigation, precipitation events, and drainage. This technique could further elucidate the geography of malar ia transmission risk through the production of adult stage vectors for those living in close proximity to irrigated schemes. 152 6 .3.3 Spatial Structure of Active Agriculture Chapter 5 examined the association between breeding pool suitability at BVIS contras ted with the Bwanje Valley. Primary findings showed surpraoptimal breeding was most prominent within the irrigated portion of BVIS demonstrating the impact of LULCC for irrigated agriculture on the Bwanje Valley and its inhabitants. In the absence of irrig ation, active agricultural areas would be considered the most favorable areas for dry season mosquito breeding. The suitability model presented does not adequately address the spatial distribution of active agricultural spaces throughout the study area. Du ring field surveying for LULC in the Bwanje Valley , instances were rare resulting in the absence of an active agricultural LULC class beyond BVIS. Future studies should work to specifically identify these areas either th rough field based observation or rem otely sensed methods to more accurately assess risk poten tial for breeding. 6.3.4 Surface Wetness Water availability is a requisite for mosquito development. During oviposition, females will lay their eggs either on or in the water, or on solid substrates that are likely to become inundated (Foster & Walker, 2009). As such, malaria transmission is limited to regions that allow for the formation and persistence of water bodies suitable for oviposition. These water bodies are not a linear function of rainfal l, but rather are a product of precipitation, antecedent soil moisture, soil type, and rates of evapotranspiration (Shaman & Day, 2005). The relationship between vector populations and water availability make hydrology models and hydrologic monitoring usef ul as a predictive tool for mosquito abundance. To further estimate the spatial distribution of mosquito breeding pool formation in irrigated agricultural schemes and/or the surrounding landscapes , time series simulations of surface wetness could be produc ed through combination of Synthetic Aperture Radar (SAR) data and high spatial resolution Digital Elevation Models (DEMs). Comparisons of estimated soil moisture between irrigated and non - irrigated spaces in combination with data on LULC may provide import ant insights into understanding the distribution of breeding risk. 153 6 .3.5 Spatio - T emporal Expansion of Irrigated A griculture Irrigated agriculture in Malawi began in the late 1940s and has since expanded to 104,299 ha as of 2015 (GoM, 2015b) . W hile statistics on the estimates of irrigated area exist in the literat ure, the geography of irrigated area in Malawi was unrecorded until the completion of the Irrigation Master Plan and Investment Framework in 2015. Were these data available, it would have been enlightening to investigate correlations between the geography of irrigation expansion along with changes to malaria prevalence data according to the MIS surveys. One method to estimate the geographical extent o f irrigated agricultural area over time would be through examination of remotely sensed, time - series NDVI an d Enhanced Vegetation Index (EVI) data for Malawi during the dry season. NDVI is a measure of both vegetation presence and health (Jensen, 2005) ; EVI is similar to NDVI but has the added benefit of canopy reflectance properties and soil reflectance correction. There are considerable difference s in NDVI for irrigated and non - irrigated crops; irrigated areas demonstrate higher NDVI values than non - irrigated (K rishnankutty Ambika, Wardlow, & Mishra, 2016) . These differences in values during the dry season may assist in differentiating between irrigated and non - irrigated areas to create high spatial - resolution irrigated maps for Malawi. This method has successfu lly been applied to mapping irrigated area in India (Krishnankutty Ambika et al., 2016; Sharma et al., 2018) and Arizona (Zheng, Myint, Thenkabail, & Aggarwal, 2015) . 6.3.6 Habitat Suitability and Malaria Prevalence Malaria is an ongoing, significant public health issue in Malawi (DHS, 2017) . While this research demonstrates habitat suitabilit y drivers for MIS clusters, there is a need to explore the relationship between habitat suitability defined by the HSM product produced as a part of this work and malaria prevalence . Distributions of disease vectors are directly related to environmental co nditions that support the biological requirements for species survival. Thus, malaria cases are assumed to be positively linked to potential habitat suitability. One method for examining the relationship between habitat suitability variables and malaria pr evalence is Geographically Weighted Regression (GWR). 154 Rather than examining global estimates, GWR allows for local variation and rates of change in model coefficients in order to better understand the spatial structure of observed relationships (Brunsdon, Fot he ullivan & Unwin, 2014) . GWR could be implemented to explore local effects of habitat suitability variables on malaria prevalence data for 2012, 2014, and 2017. There is significant interest in the relationship between irri gated agriculture and malaria as irrigation is scaled to meet global food demands . By ex amining LULCC for irrigated agriculture and malaria risk, this dissertation contributes important information on the impact of scaling irrigated agriculture in Malawi o n spatio - temporal malaria disease dynamics . LULCC for irrigated agriculture alters landscapes and encourages the production of malaria vectors. The purpose of this body of work is not to undermine irrigation, nor to contend that efforts to increase irrigat ed agriculture should cease. Rather, in light of continued LULCC for irrigated agriculture across scales, there should be specific consideration given to how human - environment interactions, even those whose intentions are rooted in promoting positive heal th outcom es, can simultaneously lead to declining o utcomes in other health arenas. Scaling irrigated agriculture is a n effective , existing intensification strategy for meeting food production demands to ensure global food security. It is necessary then to consider how irrigation can be effectively scaled such that it does not encourage the proliferation of Anopheles mosquitoes, particula rly in malaria - endemic areas . 155 APPENDICES 156 APPENDIX A : Rainy Season LULC Data: BVIS 157 Rainy Season BVIS LULC Classification Table BVIS TOTAL Cl ass Descriptions Sample AA II. IRRIGATION INFRASTRUCTURE Concrete Canal w/ Flowing Water, Bare Earth & Dried Grasses 1 0 1 Concrete Canal w/ Flowing Water & Emergent Grasses 2 1 3 Concrete Canal w/ Flowing Water, Emergent Grasses & Maize 0 1 1 Concrete Canal w/ Flowing Water, Emergent Grasses & Dried Grasses 0 3 3 Concrete Bridge Over Dirt Canal w/ Emergent Grasses 0 1 1 Dirt Canal w/ Standing Water & Dried Grasses 0 1 1 Dirt Canal w/ Standing Water & Panicle Rice 0 1 1 III. TREES Ba nana Trees 1 1 2 Green foliage on flooded rice plot 1 0 1 Trees: Roadway Adjacent 1 0 1 BVIS TOTAL Class Descriptions Sample AA I. AGRICULTURAL LAND I. ACTIVE Rice Flooded Field 2 0 2 Flooded Field w/ Dried Grasses 2 0 2 Early Growth w/ Visible Water & Grasses 2 1 3 Early Growth w/ Visible Light Standing Water 1 0 1 Early Growth w/ Visible Dark Standing Water 0 2 2 Mid - Growth w/ Visible Light Standing Water 1 0 1 Mid - Growth w/ Visibl e Dark Standing Water 0 2 2 Panicle 14 8 22 Panicle w/ Dried Grasses 3 1 4 Panicle & Grasses: Tall 8 1 9 Varied Growth Stages 2 1 3 Maize Mature 7 0 7 Past Maturity 5 2 7 Sweet Potato Mature 1 1 2 Young 2 0 2 Sweet Potato & Grasse s 0 1 1 Sweet Potato & Emergent Shrubbery 1 0 1 Pumpkin Intermediate Growth Stage & Grasses 1 0 1 Mixed Cropping Sweet Potato: Young & Panicle Rice 1 1 2 Sweet Potato: Mature, Maize: Mature, & Rice: Panicle 0 1 1 Rice: Panicle & Maize: Mature 3 2 5 II. FALLOW Grasses Tall 3 1 4 Tall and Standing Water 1 1 2 Human Influence Road 1 0 1 Straw Structure on Bare Soil 1 1 2 158 Rainy Season BVIS LULC Classification Photos I. AGRICULTURAL LAND II. ACTIVE Rice Floo ded Field Sample Photos: LULC_73 LULC_86 Accuracy Assessment Photos: None I. AGRICULTURAL LAND II. ACTIVE Rice Flooded Field w/ Dried Grass Sample Photos: LULC_4 LULC_9 Accuracy Assessment Photos: None 159 I. AGRIC ULTURAL LAND II. ACTIVE Rice Early Growth w/ Visible Water & Grasses Sample Photos: LULC_11 LULC_17 Accuracy Assessment Photos: AA_91 16 0 I. AGRICULTURAL LAND II. ACTIVE Rice Early Growth w/ Visible Light Standing Water Sample Photos: LULC_81 Accuracy Assessment Photos: None I. AGRICULTURAL LAND II. A CTIVE Rice Early Growth w/ Visible Dark Standing Water Sample Photos: None Accuracy Assessment Photo: AA_167 AA_144 161 I. AGRICULTURAL LAND II. ACTIVE Rice Mid - Growth w/ Visible Light Standing Water Sample Photos: LULC_80 Accuracy Assessment Photos: None I. AGRICULTURAL LAND II. ACTIVE Rice Mid - Growth w/ Visible Dark Standing Water Sample Photos: None Accuracy Assessment Photos: AA_150 AA_97 162 I. AGRICULTURAL LAND II. ACTIVE Rice Mature Sample Photos: LULC_70 LULC_74 LULC_75 LULC_76 LULC_6 LULC_85 LULC_157 LULC_158 163 LULC_159 LULC_160 LULC_161 LULC_162 Accuracy Assessment Photos: AA_127 AA_154 AA_171 AA_43 AA_ 59 AA_7 164 AA_75 AA_148 165 I. AGRICULTURAL LAND II. ACTIVE Rice Mature w/ Dried Grasses Sample Photos: LULC_20 LULC_18 LULC_72 Accuracy Assessment Photos: AA_62 166 I. AGRICULTURAL LAND II. ACTIVE Rice Mature & Grasses: Tall Sample Photos: LULC_23 LULC_21 LULC_144 LULC_5 LULC_15 LULC_10 LULC_7 LULC_32 167 Accuracy Assessment Photos: AA_31 168 I. AGRICULTURAL LAND II. ACTIVE Rice Varied Growth Stage Sample Ph otos: LULC_13 LULC_84 Accuracy Assessment Photos AA_13 169 I. AGRICULTURAL LAND II. ACTIVE Maize Mature Sample Photos: LULC_91 LULC_28 LULC_50 LULC_58 LULC_94 LULC_97 170 LULC _83 Accuracy Assessment Photos 171 I. AGRICULTURAL LAND II. ACTIVE Maize Past Maturity Sample Photos LULC_37 LULC_64 LULC_93 LULC_95 LULC_96 Accuracy Assessment Photos: AA_25 AA_110 172 I. AGRICULTURAL LAND II. ACTIVE Sweet Potato Mature Sample Photos: LULC_69 Accuracy Assessment Photos: AA_137 173 I. AGRICULTURAL LAND I. ACTIVE Sweet Potato Sweet Potato & Grasses Sample Photos: None Accuracy Assessment Photos: AA_33 I. AGRICULTURAL LAND I. ACTIVE Sweet Potato Young Sample Photos: LULC_3 LULC_78 Accuracy Assessment Photos: None 174 I. AGRICULTURAL LAND I. ACTIVE Sweet Potato Sweet Potato & Emergent Shrubbery Sample Phot os: LULC_68 Accuracy Assessment Photos: None I. AGRICULTURAL LAND I. ACTIVE Pumpkin Intermediate Growth Stage & Grasses Sample Photos: LULC_92 Accuracy Assessment Photos: None 175 I. AGRICULTURAL LAND I. ACTIVE Mix ed Cropping Sweet Potato: Young & Rice: Mature Sample Photos: LULC_60 Accuracy Assessment Photos: None I. AGRICULTURAL LAND I. ACTIVE Mixed Cropping Sweet Potato: Mature & Rice: Mature Sample Photos: None Accuracy Asses sment Photos: AA_142 176 I. AGRICULTURAL LAND I. ACTIVE Mixed Cropping Sweet Potato: Mature Maize: Mature & Rice: Mature Sample Photos: None Accuracy Assessment Photos: AA_141 177 I. AGRICULTURAL LAND I. ACTIVE Mixed Cropping Rice : Mature & Maize: Mature Sample Photos: LULC_53 LULC_29 LULC_41 Accuracy Assessment Photos: AA_162 AA_165 178 I. AGRICULTURAL LAND II. FALLOW Grasses Tall Sample Photos: LULC_1 LULC_2 LULC_22 Accuracy A ssessment Photos: AA_29 179 I. AGRICULTURAL LAND II. FALLOW Grasses Tall & Standing Water Sample Photos: None Accuracy Assessment Photos: AA_104 I. AGRICULTURAL LAND II. FALLOW Human Influence Dried Grass Sample Photos: LULC_67 Accuracy Assessment Photos: None 180 I. AGRICULTURAL LAND II. FALLOW Human Influence Straw Structure on Bare Soil Sample Photos: LULC_79 Accuracy Assessment Photos: AA_70 181 I. AGRICULTURAL LAND III. IRRIGATION INFRAS TRUCTURE Concrete Canal w/ Flowing Water & Bare Earth + Dried Grass Sample Photos: LULC_77 Accuracy Assessment Photos: None 182 I. AGRICULTURAL LAND III. IRRIGATION INFRASTRUCTURE Concrete Canal w/ Flowing Water & Emergent Grasses Sample Photos: LULC_63 LULC_88 Accuracy Assessment Photos: AA_6 183 I. AGRICULTURAL LAND III. IRRIGATION INFRASTRUCTURE Concrete Canal w/ Flowing Water, Emergent Grasses & Maize Sample Photos: None Accuracy Assessment Photos: AA_47 184 I. AGRICULTURAL LAND III. IRRIGATION INFRASTRUCTURE Concrete Canal w/ Emergent Grasses & Dried Grasses Sample Photos: None Accuracy Assessment Photos: AA_2 AA_77 AA_19 185 I. AGRICULTURAL LAND III. IRRIGATION INFRASTRUCTURE Concrete Bridge Over Dirt Canal w/ Emergent Grasses Sample Photos: None Accuracy Assessment Photos: AA_46 I. AGRICULTURAL LAND III. IRRIGATION INFRASTRUCTURE Dirt Canal w/ Stan ding Water & Emergent Grasses Sample Photos: LULC_39 Accuracy Assessment Photos: None 186 I. AGRICULTURAL LAND III. IRRIGATION INFRASTRUCTURE Dirt Canal w/ Standing Water & Dried Grasses Sample Photos: None Accuracy Assessment Phot os: AA_73 I. AGRICULTURAL LAND III. IRRIGATION INFRASTRUCTURE Dirt Canal w/ Standing Water & Mature Rice Sample Photos: None Accuracy Assessment Photos: AA_123 187 I. AGRICULTURAL LAND IV. TREES Banana Trees Sample Ph otos: LULC_90 Accuracy Assessment Photos: AA_26 I. AGRICULTURAL LAND IV: TREES Green Foliage on Flooded Rice Plot Sample Photos: LULC_65 Accuracy Assessment Photos: None 188 I. AGRICULTURAL LAND IV: TREES Banana Trees & Dense Sh rubbery Sample Photos: None Accuracy Assessment Photos: BVIS_AA_159 I. AGRICULTURAL LAND IV: TREES Green Foliage on Roadway Sample Photos: LULC_66 Accuracy Assessment Photos: None 189 Rainy Season BVIS Data ID Name X Y Elev P nt_Type PntAcq Offset_Dis Class SubClass Land_Cover Land_Use Notes 1 LULC_1 34.63208 - 14.234 452.6 Sample On 0 Fallow Grasses Grasses: Tall Agricultural Land Tall grasses lining tertiary canal 2 LULC_10 34.62554 - 14.24431 472.4 Sample On 0 Active Rice Ri ce: Mature & Grasses: Tall Agricultural Land Boundary - tertiary canal ridge . T he canal is not functional (tall grass growing) 3 LULC_11 34.62472 - 14.24533 471.8 Sample On 0 Active Rice Rice: Early Growth w/ Visible Water & Grasses Agricultural Land Bound ary - just next (close) to the main drain. 4 LULC_13 34.62251 - 14.24478 471.5 Sample On 0 Active Rice Rice: Varied Growth Stages Agricultural Land Point within the scheme on the ridge between rice paddy plots ~90m to the boundary on the southern part. 5 LULC_144 34.62185 - 14.24439 469.3 Sample On 0 Active Rice Rice: Mature & Grasses: Tall Agricultural Land 6 LULC_15 34.61904 - 14.24531 471.5 Sample On 0 Active Rice Rice: Mature & Grasses: Tall Agricultural Land Boundary - tall grass on the southern side . 7 LULC_157 34.6058 - 14.2559 493.5 Sample On 0 Active Rice Rice: Mature Agricultural Land Visible water present 190 8 LULC_158 34.585 - 14.2549 500.8 Sample On 0 Active Rice Rice: Mature Agricultural Land Point taken from a bund 9 LULC_159 34.595 - 14.2521 496.6 Sample On 0 Active Rice Rice: Mature Agricultural Land Visible water present. Rice not as tightly planted 10 LULC_160 34.5864 - 14.2594 501.7 Sample On 0 Active Rice Rice: Mature Agricultural Land Adjacent field has visible standing water; noting pl anted yet 11 LULC_161 34.6268 - 14.2393 481.0 Sample On 0 Active Rice Rice: Mature Agricultural Land Adjacent field has visible dark standing water 12 LULC_162 34.6213 - 14.2414 483.1 Sample On 0 Active Rice Rice: Mature Agricultural Land Transition betwee n two fields. One is at a slightly earlier growth stage 13 LULC_163 34.6086 - 14.2524 491.9 Sample On 0 Active Rice Rice: Mature Agricultural Land 14 LULC_164 34.5798 - 14.2666 504.5 Sample On 0 Active Rice Rice: Mature Agricultural Land Some visible sta nding water between plantings. 15 LULC_17 34.61735 - 14.24621 471.2 Sample On 0 Active Rice Rice: Early Growth w/ Visible Water & Grasses Agricultural Land Boundary: Tall grass grown on the southern part and some trees. 191 16 LULC_18 34.61707 - 14.24602 474. 5 Sample On 0 Active Rice Rice: Mature w/ Dried Grasses Agricultural Land Grasses lining non functioning tertiary canal 17 LULC_2 34.63451 - 14.23507 468.7 Sample On 0 Fallow Grasses Grasses: Tall Agricultural Land Tall grasses lining non functioning terti ary canal 18 LULC_20 34.61568 - 14.24639 475.4 Sample On 0 Active Rice Rice: Mature w/ Dried Grasses Agricultural Land Ridge on tertiary canal between rice fields running from NW to SW. 19 LULC_21 34.61462 - 14.24774 472.1 Sample On 0 Active Rice Rice: Mat ure & Grasses: Tall Agricultural Land Boundary. Tall grasses on other side. 20 LULC_22 34.61443 - 14.24745 475.1 Sample On 0 Fallow Grasses Grasses: Tall Agricultural Land Tall grasses lining non functioning tertiary canal 21 LULC_23 34.61422 - 14.24756 47 7.0 Sample On 0 Active Rice Rice: Mature & Grasses: Tall Agricultural Land Grass walkway between plots (one plot away from boundary (on a strip of 4 plots cultivate). 22 LULC_28 34.61056 - 14.2495 477.0 Sample On 0 Active Maize Maize: Mature Agricultural L and Boundary of scheme. Maize is growing on the upland here. 192 23 LULC_29 34.61005 - 14.25092 478.5 Sample On 0 Active Mixed Cropping Rice: Mature & Maize: Mature Agricultural Land Scheme edge. Road to west rice planted in fields. Maize on upland with a smal l border ~1ft of tall grasses 24 LULC_3 34.63435 - 14.23691 468.7 Sample On 0 Active Sweet Potato Sweet Potato: Young Agricultural Land W ithin the scheme 1 field of sweet potatoes on the upland. 25 LULC_32 34.57975 - 14.25081 488.5 Sample On 0 Active Rice Rice: Mature & Grasses: Tall Agricultural Land Boundary - rice paddy field on SE and tall grass grown on NE. Several trees next to grass on NE. 26 LULC_37 34.57704 - 14.25372 489.2 Sample On 0 Active Maize Maize: Past Maturity Agricultural Land Almost 1 ac re of maize field within the scheme near the boundary. 27 LULC_39 34.57596 - 14.25475 500.1 Sample On 0 Fallow Grasses Grasses: Tall & Standing Water Agricultural Land Diversion point from the drain to the scheme fields. 28 LULC_4 34.63375 - 14.23818 466. 3 Sample On 0 Active Rice Rice: Flooded Field w/ Agricultural Land Boundary 193 Dried Grasses 29 LULC_41 34.57477 - 14.25682 496.8 Sample On 0 Active Mixed Cropping Rice: Mature & Maize: Mature Agricultural Land Within the scheme. On the ridge between the mai ze (grown next to the drain - strip running from east to west (see diagram in notebook) field and rice 30 LULC_5 34.63121 - 14.23965 469.0 Sample On 0 Active Rice Rice: Mature & Grasses: Tall Agricultural Land Tertiary canal downstrea m not functional. 31 LULC_50 34.57161 - 14.26163 498.0 Sample On 0 Active Maize Maize: Mature Agricultural Land On the footpath in the maize field. The maize field is located between the drain and rice paddy field supplied with drainage water. 32 LULC_53 34.57055 - 14.26251 494 .9 Sample On 0 Active Mixed Cropping Rice: Mature & Maize: Mature Agricultural Land Rice grown in the main drain and there is maize grown on the upland between 194 the drain and the rice paddy fields. 33 LULC_58 34.56821 - 14.26424 506.8 Sample On 0 Active Mai ze Maize: Mature Agricultural Land Maize field 34 LULC_6 34.62929 - 14.24205 470.3 Sample On 0 Active Rice Rice: Mature Agricultural Land Boundary 35 LULC_60 34.5672 - 14.26503 502.9 Sample On 0 Active Mixed Cropping Sweet Potato: Young w/ Rice: Mature Agr icultural Land Drain running from west to east there is rice cultivated within the drain and sweet potatoes on the side. 36 LULC_63 34.59582 - 14.26065 485.8 Sample On 0 Ag Infrastructure Irrigation Structure Concrete Canal w Flowing Water & Emergent Grass es Agricultural Land drainage meets canal 37 LULC_64 34.59565 - 14.26084 488.2 Sample On 0 Active Maize Maize: Past Maturity Agricultural Land maize middle od drain and old canal used as drain 38 LULC_65 34.59061 - 14.25986 505.0 Sample Offset 25 Trees Tre es Trees: Green foliage on flooded rice plot Agricultural Land Original point is a large tree shading most of a farmers field. Point taken on road. 39 LULC_66 34.59708 - 14.25299 492.2 Sample On 0 Trees Trees Trees: Green Foliage on Roadway Agricultural La nd Tree between canal and village 195 along roadway 40 LULC_67 34.60095 - 14.25116 488.2 Sample On 0 Ag Infrastructure Human Influence Road Agricultural Land Point taken on roadway. Village adjacent 41 LULC_68 34.60478 - 14.24897 474.2 Sample On 0 Active Sweet Potato Sweet Potato & Emergent Shrubbery Agricultural Land Not rice; taller unmanaged weeds one rice field to the east. Two plots past the end of irrigation canal. 42 LULC_69 34.55937 - 14.27472 499.2 Sample On 0 Active Sweet Potato Sweet Potato: Mature Agricultural Land sweet potato rice on all sides 43 LULC_7 34.62824 - 14.2424 472.4 Sample On 0 Active Rice Rice: Mature & Grasses: Tall Agricultural Land Boundary next to the main drain. 44 LULC_70a 34.568199 - 14.2657 498.0 Sample On 0 Active Rice Rice: Mature Agricultural Land 45 LULC_70 34.56821 - 14.26569 498.0 Sample On 0 Active Rice Rice: Mature Agricultural Land Point is within rice paddy filed; great deal of standing H20. Almost to the top of boots ~7ft from bund (to east). 46 LULC_72 34.58018 - 1 4.2568 487.3 Sample On 0 Active Rice Rice: Mature w/ Dried Grasses Agricultural Land Point just off the road immediatel y adjacent to rice paddies. This point 196 has mounded up grasses. 47 LULC_73 34.57882 - 14.25982 488.2 Sample On 0 Active Rice Rice: Flooded Field Agricultural Land field ponded with excess water 48 LULC_74 34.57863 - 14.26104 491.6 Sample On 0 Active Rice Rice: Mature Agricultural Land Taken from bund. Original point is inside rice paddy. Bund has tall grasses. Taller than Willy. 49 LULC_75a 34.5774 - 14.2617 494.3 Sample On 0 Active Rice Rice: Mature Agricultural Land 50 LULC_75 34.57745 - 14.26167 494.3 Sample On 0 Active Rice Rice: Mature Agricultural Land Taken on a bund point is inside of a rice field. Bund = mounded grasses. 51 LULC_76 34.57612 - 14.26408 493.4 Sample On 0 Active Rice Rice: Mature Agricultural Land Rice paddies potentially the bund separating two fields. 52 LULC_77 34.57581 - 14.26449 496.8 Sample Offset 1 Ag Infrastructure Irrigation Structure Concrete Canal w/ Flowing Water, Bare Earth & Dried Grasses Agricultural Land 53 LULC_78 34.57601 - 14.2642 500.4 Sample On 0 Active Sweet Potato Sweet Potato: Young Agricultural Land Exposed bare earth (for sweet potatoes) adjacent to flooded 197 rice paddies. More water and this a rea could be flooded too 54 LULC_79 34.57528 - 14.26512 498.6 Sample Offset 10 Ag Infrastructure Human Influence Straw structure on Bare Soil Agricultural Land Straw structure 55 LULC_80 34.568 - 14.26469 496.8 Sample On 0 Active Rice Rice: Mid - Growth w/ V isible Light Standing Water Agricultural Land Rice adjacent to unintended maize to the north seeming the northern boundary of the scheme. 56 LULC_81 34.56274 - 14.27414 501.7 Sample On 0 Active Rice Rice: Early Growth w/ Visible Light Standing Water Agricu ltural Land Used point: 3 - 6 - 2017 @ 3:18 pm. 14.27349, 34.55320, maize 57 LULC_83 34.6022 - 14.2573 485.8 Sample On 0 Active Maize Maize: Mature Agricultural Land Three rows of maize planted immediatel y adjacent to the road in upland area. 58 LULC_84 34.6 073 - 14.2508 482.4 Sample On 0 Active Rice Rice: Varied Growth Stages Agricultural Land At almost the corner of the last section of fields and a tertiary canal. Located in the rice 198 fields immediatel y adjacent to a bund. 59 LULC_85 34.60962 - 14.25091 473.0 Sample On 0 Active Rice Rice: Mature Agricultural Land Cluster of unmanaged taller grasses immediatel y adjacent to these rice fields. Maize planted two fields to the west. 60 LULC_86 34.61084 - 14.25165 477.9 Sample Offset 5 Active Rice Rice: Flooded Fiel d Agricultural Land Water is too deep to get to the point. Up to thighs. Rice is growing surrounded by the road to the east and maize to the west. See 360 degree on Nikon. 61 LULC_88 34.62309 - 14.2388 471.5 Sample Offset 2 Ag Infrastructure Irrigation Str ucture Concrete Canal w Flowing Water & Emergent Grasses Agricultural Land Point is between the roa and the canal. Immediatel y adjacent to the canal. Small patch of weeds growing. 62 LULC_9 34.6265 - 14.24317 472.1 Sample Offset 15 Active Rice Rice: Floode d Field w/ Agricultural Land Tertiary canal downstrea 199 Dried Grasses m on the scheme boundary. 63 LULC_97 34.57532 - 14.26512 499.5 Sample On 0 Active Maize Maize: Mature Agricultural Land Small outcrop of maize on the upland tree adjacent and straw 64 AA_ 104.jpg 34.5888 - 14.2622 - 9999 AA On 0 Fallow Grasses Grasses: Tall & Standing Water Agricultural Land 65 AA_110.jpg 34.5957 - 14.2608 - 9999 AA On 0 Active Maize Maize: Past Maturity Agricultural Land 66 AA_123.jpg 34.5583 - 14.274516 - 9999 AA On 0 Ag Infrastructure Irrigation Structure Dirt Canal w/ Standing Water & Mature Rice Agricultural Land 67 AA_127.jpg 34.5584 - 14.2746 - 9999 AA On 0 Active Rice Rice: Mature Agricultural Land 68 AA_13.jpg 34.6267 - 14.237001 - 9999 AA On 0 Active Rice Ri ce: Varied Growth Stages Agricultural Land 69 AA_137.jpg 34.5592 - 14.2747 - 9999 AA On 0 Active Sweet Potato Mature Agricultural Land 70 AA_141.jpg 34.5713 - 14.2688 - 9999 AA On 0 Active Mixed Cropping Sweet Potato: Mature, Maize: Mature, & Rice: Matur e Agricultural Land 71 AA_142.jpg 34.5736 - 14.268601 - 9999 AA On 0 Active Mixed Cropping Sweet Potato: Young & Mature Rice Agricultural Land 72 AA_144.jpg 34.5738 - 14.2668 - 9999 AA On 0 Active Rice Rice: Early Growth w/ Visible Dark Standing Water A gricultural Land 73 AA_148.jpg 34.6344 - 14.2326 - 9999 AA On 0 Active Rice Rice: Mature Agricultural Land 74 AA_150.jpg 34.6103 - 14.2436 - 9999 AA On 0 Active Rice Rice: Mid - Growth w/ visible Dark Agricultural Land 200 Standing Water 75 AA_154.jpg 34.6 162 - 14.2416 - 9999 AA On 0 Active Rice Rice: Mature Agricultural Land 76 AA_162.jpg 34.61 - 14.2506 - 9999 AA On 0 Active Mixed Cropping Rice: Mature & Maize: Mature Agricultural Land 77 AA_165.jpg 34.6107 - 14.2493 - 9999 AA On 0 Active Mixed Cropping R ice: Mature & Maize: Mature Agricultural Land 78 AA_167.jpg 34.612 - 14.2491 - 9999 AA On 0 Active Rice Rice: Early Growth w/ Visible Dark Standing Water Agricultural Land 79 AA_171.jpg 34.6231 - 14.239033 - 9999 AA On 0 Active Rice Rice: Mature Agricult ural Land 80 AA_19.jpg 34.6333 - 14.2331 - 9999 AA On 0 Ag Infrastructure Irrigation Structure Concrete Canal w/ Emergent Grasses & Dried Grasses Agricultural Land 81 AA_2.jpg 34.6243 - 14.2382 - 9999 AA On 0 Ag Infrastructure Irrigation Structure Co ncrete Canal w/ Emergent Grasses & Dried Grasses Agricultural Land 82 AA_29.jpg 34.6231 - 14.2388 - 9999 AA On 0 Fallow Grasses Grasses: Tall Agricultural Land 83 AA_31.jpg 34.6093 - 14.2506 - 9999 AA On 0 Active Rice Rice: Mature & Grasses: Tall Agricul tural Land 84 AA_33.jpg 34.6099 - 14.251 - 9999 AA On 0 Active Sweet Potato Sweet Potato & Grasses Agricultural Land 85 AA_43.jpg 34.6072 - 14.2508 - 9999 AA On 0 Active Rice Rice: Mature Agricultural Land 86 AA_59.jpg 34.5687 - 14.2647 - 9999 AA On 0 A ctive Rice Rice: Mature Agricultural Land 87 AA_6.jpg 34.625 - 14.2378 - 9999 AA On 0 Ag Infrastructure Irrigation Structure Concrete Canal w/ Flowing Water & Agricultural Land 201 Emergent Grasses 88 AA_62.jpg 34.5688 - 14.2649 - 9999 AA On 0 Active Rice Rice: Mature w/ Dried Grass Agricultural Land This is actually a tertiary canal 89 AA_7.jpg 34.6256 - 14.2376 - 9999 AA On 0 Active Rice Rice: Mature Agricultural Land 90 AA_70.jpg 34.5753 - 14.2651 - 9999 AA On 0 Ag Infrastructure Human Influence Straw S tructure on Bare Soil Agricultural Land 91 AA_73.jpg 34.5753 - 14.265 - 9999 AA On 0 Ag Infrastructure Irrigation Structure Dirt Canal w/ Standing Water & Dried Grasses Agricultural Land 92 AA_75.jpg 34.5784 - 14.2609 - 9999 AA On 0 Active Rice Rice: Mature Agricultural Land 93 AA_77.jpg 34.5787 - 14.2601 - 9999 AA On 0 Ag Infrastructure Irrigation Structure Concrete Canal w/ Emergent Grasses & Dried Grasses Agricultural Land 94 AA_91.jpg 34.5799 - 14.264 - 9999 AA On 0 Active Rice Rice: Early Grow th w/ Visible Water & Grasses Agricultural Land 95 AA_97.jpg 34.5837 - 14.2632 - 9999 AA On 0 Active Rice Rice: Mid - Growth w/ Visible Dark Standing Water Agricultural Land 202 Rainy Season BVIS Analysis Classes Class Name Samples in Class Land Cover 2 4 Rice 3 6 Rice 4 8 Rice 5 3 Rice 6 4 Rice 7 12 Rice 8 8 Non - Vegetated 9 26 Rice 10 25 Non - Vegetated Class 2: Rice (Panicle) NAME CLASS SUBCLASS LAND COVER LULC_161 Active Rice Rice: Panicle LULC_144 Active Rice Rice: Panicle & Tall Grasses LULC_163 Active Rice Rice: Panicle LULC_158 Active Rice Rice: Panicle Class 3: Rice (Panicle) NAME CLASS SUBCLASS LAND COVER LULC_5 Active Rice Rice: Panicle & Tall Grasses AA_162 Active Mixed Cropping Mixed Cropping: Rice: Panicle & Maize: Panicl e LULC_159 Active Rice Rice: Panicle LULC_83 Active Maize Maize: Panicle LULC_42 Active Mixed Cropping Mixed Cropping: Rice: Panicle & Maize: Panicle LULC_75 Active Rice Rice: Panicle Class 4: Rice (Panicle or Past Maturity) NAME CLASS SUBCLASS LAND COVER AA_148 Active Rice Rice: Panicle LULC_4 Active Rice Flooded Field w/ Dried Grass LULC_23 Active Rice Rice: Panicle & Tall Grasses LULC_86 Active Rice Flooded Field LULC_75 Active Rice Rice: Panicle LULC_80 Active Rice Rice: Early Growth w/ Vi sible Light Standing Water AA_142 Active Mixed Cropping Mixed Cropping: Sweet Potato: Young & Rice: Panicle LULC_81 Active Rice Rice: Early Growth w/ Visible Light Standing Water Class 5: Rice (Panicle) NAME CLASS SUBCLASS LAND COVER LULC_22 Fallow Grasses Grasses: Tall LULC_85 Active Rice Rice: Panicle LULC_157 Active Rice Rice: Panicle 203 Class 6: Rice (Panicle & Dried Grasses) NAME CLASS SUBCLASS LAND COVER LULC_2 Fallow Grasses Grasses: Tall AA_165 Active Mixed Cropping Mixed Cropping: Rice: Panicle & Maize: Panicle AA_62 Active Rice Rice: Panicle & Dried Grasses Class 7: Rice NAME CLASS SUBCLASS LAND COVER LULC_9 Active Rice Rice: Flooded Field w/ Dried Grasses LULC_18 Active Rice Rice: Panicle w/ Dried Grasses LULC_67 Ag. Inf. Human Influence Road AA_33 Active Sweet Potato Sweet Potato & Grasses LULC_36 Active Maize Maize: Panicle LULC_37 Active Maize Maize: Past Maturity LULC_39 Fallow Grasses Grasses: Tall LULC_74 Active Rice Rice: Panicle AA_97 Active Rice Rice: Mid - Growth w/ Visible Dark Standing Water AA_91 Active Rice Rice: Early Growth w/ Visible Water & Grasses AA_59 Active Rice Rice: Panicle AA_144 Active Rice Rice: Early Growth w/ Visible Dark Standing Water Class 8: Non - Vegetated NAME CLASS SUBCLASS LAND COVER AA_7 Active Rice Rice: Panicle AA_6 Ag. Inf. Ag. Inf. Concrete Canal w/ Flowing Water & Emergent Grasses LULC_73 Active Rice Rice: Flooded Field LULC_160 Active Rice Rice: Panicle AA_77 Ag. Inf. Ag. INf. Concrete Canal w/ Flowing Water, Emergent Gr asses & Dried Grasses LULC_70 Active Rice Rice: Panicle AA_127 Active Rice Rice: Panicle AA_123 Ag. Inf. Ag. Inf. Dirt Canal w/ Standing Water & Panicle Rice 204 Class 9: Rice NAME CLASS SUBCLASS LAND COVER LULC_3 Active Sweet Potato Sweet P otato: Young AA_13 Active Rice Rice: Varied Growth Stages AA_171 Active Rice Rice: Panicle LULC_162 Active Rice Rice: Panicle LULC_6 Active Rice Rice: Panicle LULC_7 Active Rice Rice: Panicle & Grasses: Tall LULC_10 Active Rice Rice: Panicle & Gras ses: Tall AA_150 Active Rice Rice: Mid - Growth w/ Visible Standing Water LULC_11 Active Rice Rice: Early Growth w/ Visible Standing Water and Grasses LULC_20 Active Rice Rice: Panicle w/ Dried Grasses LULC_21 Active Rice Rice: Panicle w/ Tall Grasses L ULC_28 Active Maize Maize: Panicle AA_167 Active Rice Rice: Early Growth w/ Visible Dark Standing Water LULC_32 Active Rice Rice: Panicle & Tall Grasses LULC_63 Ag. Inf. Ag. Inf. Concrete Canal w/ Flowing Water & Emergent Grasses AA_75 Active Rice Ri ce: Panicle LULC_53 Active Mixed Cropping Mixed Cropping: Rice: Panicle & Maize Panicle LULC_76 Active Rice Rice: Panicle LULC_58 Active Maize Maize: Panicle LULC_77 Ag. Inf. Ag. Inf. Concrete Canal w/ Flowing Wtaer, Bare Earth, & Dried Grasses LU LC_70 Active Rice Rice: Panicle LULC_60 Active Mixed Cropping Mixed Cropping: Sweet Potato: Young & Rice: Panicle LULC_164 Active Rice Rice: Panicle AA_137 Active Sweet Potato Sweet Potato: Panicle 205 Class 10: Non - Vegetated NAME CLA SS SUBCLASS LAND COVER AA_19 Ag. Inf. Ag. Inf. Concrete Canal w/ Flowing Water, Emergent Grasses, & Dried Grasses LULC_1 Fallow Grasses Grasses: Tall AA_2 Ag. Inf. Ag. Inf. Concrete Canal w/ Flowing Water, Emergent Grasses, & Dried Grasses LULC_88 Ag. Inf. Ag. Inf. Concrete Canal w/ Flowing Water & Emergent Grasses AA_29 Fallow Grasses Grasses: Tall AA_154 Active Rice Rice: Panicle LULC_17 Active Rice Rice: Early Growth w/ Visible Water & Grasses LULC_68 Active Pumpkin Pumpkin: Intermediate Growth Stage & Grasses AA_31 Active Rice Rice: Panicle & Grasses: Tall LULC_84 Active Active Rice: Varied Growth Stages LULC_29 Active Mixed Cropping Mixed Cropping: Rice: Panicle & Maize: Panicle AA_43 Active Rice Rice: Panicle LULC_66 Trees Tree s Trees: Green Foliage over Roadway LULC_72 Active Rice Rice: Panicle w/ Dried Grasses LULC_65 Trees Trees Trees: Green Foliage on Flooded Rice Plot LULC_64 Active Maize Maize: Past Maturity AA_110 Active Maize Maize: Past Maturity LULC_50 Active Maize Maize: Panicle AA_104 Fallow Grasses Grasses: Tall & Standing Water LULC_78 Active Sweet Potato Sweet Potato: Young AA_73 Ag. Inf. Ag. Inf. Dirt Canal w/ Standing Water & Tall Grasses LULC_97 Active Maize Maize: Panicle LULC_79 Ag. Inf. Hum an Influence Straw Structure on Bare Soil AA_70 Ag Inf. Human Influence Straw Structure on Bare Soil LULC_69 Active Pumpkin Sweet Potato: Panicle 206 APPENDIX B : Dry Season LULC Data: BVIS 207 Dry Season BVIS LULC Classification Table BVIS TO TAL Class Descriptions Sample AA I. AGRICULTURAL LAND I. ACTIVE Maize Mature 9 7 16 Young 4 0 4 Young & Dense Shrubbery 2 0 2 Beans Mature 2 1 3 Mature & Emergent Weeds 2 0 2 Young 5 1 6 Cowpea Mature 2 1 3 Cowpe a & Shrubbery 2 1 3 Young 3 0 3 General Agriculture Intercrop: Young Maize & Beans 1 1 2 Mature Maize & Beans 1 0 1 Mustard Greens 0 1 1 II. FALLOW Bare Earth Dark Soil 7 0 7 Dark Soil w/ Straw 14 2 16 Dark Soil w/ Straw & Ridg ing 0 2 2 Light Soil 2 0 2 Light Soil w/ Straw & Ridging 3 0 3 Medium Soil w/ Straw & Ridging 3 0 3 Charred Ground Charred Ground w/ Beans 1 0 1 Charred Ground w/ Cowpea 2 0 2 Charred Ground w/ Emergent Weeds 3 2 5 Charred Ground w/ Shru bbery 1 0 1 Charred Ground on Bare Earth 1 0 1 Dried Fields Dried Grass 10 3 13 Dried Grass & Emergent c3 Vegetation 6 0 6 Vertical Rice Straw w/ Little Vegetation 4 0 4 Vertical Rice Straw w/ Emergent Weeds 7 10 17 Vertical Rice Straw w/ Green Tops 1 0 1 Horizontal Rice Straw w/ Emergent Weeds 31 0 31 BVIS TOTAL Class Descriptions Sample AA II. FALLOW (CONT.) Shrubbery Dense 30 5 35 Sparse 15 2 17 Dense & Emergent Red Weeds 2 2 4 Blackjack w/ Emergent Weeds 1 1 2 Dense Shrubbery w/ Medium Soil & Straw 1 0 1 III. AGRICULTURAL INFRASTRUCTURE Irrigation Canal Concrete 1 2 3 Human Influence Road 0 3 3 IV. TREES Banana Trees & Dried Grass 2 0 2 Banana Trees & Dense Shrubbery 0 1 1 Green F oliage on Active Agricultural Land 1 0 1 Green Foliage on Fallow Field 3 1 4 Green Foliage on Fallow Land w/ Green Vegetation 0 1 1 208 Dry Season BVIS LULC Classification Photos I. AGRICULTURAL LAND II. ACTIVE Maize Mature Sample Photos : 2_57 5_52 4_53 5_53 1_39 3_39 5_39 4_45 209 1_3 Accuracy Assessment Photos: BVIS_AA_108 BVIS_AA_109 BVIS_AA_116 BVIS_AA_15 BVIS_AA_16 BVIS_AA_80 210 BVIS_AA_86 211 I. AGRICULTURAL LAND II. ACTIVE Maize Young Sample Photos: 9_38 2_12 1_9 2_3 Accuracy Assessment Photos: None 212 I. AGRICULTURAL LAND II. ACTIVE Maize Young w Dense Shrubbery Sample Photos: 5_33 5_38 Accuracy Assessment Photos None 213 I. AGRICULTURA L LAND II. ACTIVE Beans Mature Sample Photos: 1_18 6_36 Accuracy Assessment Photos: BVIS_AA_131 214 I. AGRICULTURAL LAND II. ACTIVE Beans Mature & Emergent Weeds Sample Photos: 3_31 1_24 Accuracy Assessmen t Photos None 215 I. AGRICULTURAL LAND II. ACTIVE Beans Young Sample Photos: 3_38 4_38 2_19 2_18 1_6 Accuracy Assessment Photos: BVIS_AA_132 216 I. AGRICULTURAL LAND II. ACTIVE Cowpea Mature Sample Photos: 2_ 10 1_15 Accuracy Assessment Photos: BVIS_AA_123 217 I. AGRICULTURAL LAND II. ACTIVE Cowpea Cowpea & Shrubbery Sample Photos: 2_15 3_15 Accuracy Assessment Photo: BVIS_AA_128 218 I. AGRICULTURAL LAND II. ACTIVE Cow pea Young Sample Photos: 2_23 1_10 7_21 Accuracy Assessment Photos: None 219 I. AGRICULTURAL LAND II. ACTIVE General Agriculture Intercrop: Young Maize & Beans Sample Photos: 2_38 Accuracy Assessment Photos BVIS_AA _144 I. AGRICULTURAL LAND II. ACTIVE General Agriculture Mature Maize & Beans Sample Photos: 4_39 Accuracy Assessment Photos None 220 I. AGRICULTURAL LAND II. ACTIVE General Agriculture Mustard Greens Sample Photos: No ne Accuracy Assessment Photos: BVIS_AA_151 221 I. AGRICULTURAL LAND II. FALLOW Bare Earth Dark Soil Sample Photos: 1_58 1_50 3_50 7_53 3_49 3_47 2_49 Accuracy Assessment Photos: None 222 I. AGRICULTURAL LAND I I. FALLOW Bare Earth Dark Soil w Straw Sample Photos: 6_7 1_47 2_47 4_47 2_46 5_46 6_46 1_45 223 2_45a 2_40 1_51 1_52 1_55 4_57 Accuracy Assessment Photos: BVIS_AA_5 BVIS_AA_47 224 I. AGRICULTURAL LA ND II. FALLOW Bare Earth Bare Earth: Dark Soil w Straw and Ridging Sample Photos: None Accuracy Assessment Photos: BVIS_AA_158 BVIS_AA_163 I. AGRICULTURAL LAND II. FALLOW Bare Earth Bare Earth: Light Soil Sample Photos: 1_1 4_2 Accuracy Assessment Photos: None 225 I. AGRICULTURAL LAND II. FALLOW Bare Earth Light Soil w Straw and Ridging Sample Photos: 2_1 3_2 3_6 Accuracy Assessment Photos: None 226 I. AGRICULTURAL LAND II. FA LLOW Bare Earth Medium Soil w Straw and Ridging Sample Photos: 1_2 2_2 7_7 Accuracy Assessment Photos: None I. AGRICULTURAL LAND II. FALLOW Charred Ground Charred Ground w Beans Sample Photos: 5_19 Accuracy As sessment Photos: None 227 I. AGRICULTURAL LAND II. FALLOW Charred Ground Charred Ground w Cowpea Sample Photos: 2_9 3_10 Accuracy Assessment Photos: None 228 I. AGRICULTURAL LAND II. FALLOW Charred Ground Charred Ground w Emer gent Weeds Sample Photos: 4_52 2_31 4_25 Accuracy Assessment Photos: BVIS_AA_138 BVIS_AA_75 229 I. AGRICULTURAL LAND II. FALLOW Charred Ground Charred Ground w Shrubbery Sample Photos: 2_22 Accuracy Assessment Pho tos: None I. AGRICULTURAL LAND II. FALLOW Charred Ground Charred Ground on Bare Earth Sample Photos: 1_21 Accuracy Assessment Photos: None 230 I. AGRICULTURAL LAND II. FALLOW Dried Fields Dried Grass Sample Photos: 3_57 b 3_57 1_54 1_46 6_16 1_23 2_51 4_27 231 3_46 1_29 Accuracy Assessment Photos: BVIS_AA_28 BVIS_AA_68 BVIS_AA_67 232 I. AGRICULTURAL LAND II. FALLOW Dried Fields Dried Grass & Emergent c3 Vegetation Sample Pho tos: 2_56 1_42 1_38 1_19 1_12 7_16 Accuracy Assessment Photos: None 233 I. AGRICULTURAL LAND II. FALLOW Dried Fields Vertical Rice Straw w/ Little Vegetation Sample Photos: 1_25 1_40 1_17 4_51 Accuracy Asse ssment Photos: None 234 I. AGRICULTURAL LAND II. FALLOW Dried Fields Vertical Rice Straw w Emergent Weeds Sample Photos: 1_56 1_53 5_51 6_53 1_49 3_22 3_20 235 Accuracy Assessment Photos: BVIS_AA_103 BVIS_AA_104 BVIS _AA_34 BVIS_AA_72 BVIS_AA_71 BVIS_AA_76 BVIS_AA_77 BVIS_AA_148 BVIS_AA_156 BVIS_AA_35 236 I. AGRICULTURAL LAND II. FALLOW Dried Fields Vertical Rice Straw w Green Tops Sample Photos: 1_8 Accuracy Assessment Photos: None 237 I. AGRICULTURAL LAND II. FALLOW Dried Fields Horizontal Rice Straw w/ Emergent Weeds Sample Photos: 2_55 3_51 10_53 1_48 3_44 3_48 2_37 6_38 4_46 10_38 4_48 5_48 238 6_48 2_36 4_36 3_40 5_36 1_35 1_28 2_28 1_22 7_22 3_28 4_28 239 5_28 2_21 2_7 3_21 4_21 6_21 2_52 Accuracy Assessment Photos: None 240 I. AGRICULTURAL LAND II. FALLOW Shrubbery Shrubbery: Dense Sample Photos: 7_25 3_8 2_27 3_ 27 3_7 4_7 2_20 4_20 2_6 1_14 7_44 1_43 241 2_33 3_33 1_26 4_33 1_31 2_25 1_36 5_31 1_30 3_36 3_52 1_44 242 4_31 3_16 1_11 2_44 8_38 4_19 Accuracy Assessment Photos: BVIS_AA_133 BVIS_AA_4 BVIS_A A_166 BVIS_AA_100 BVIS_AA_134 243 I. AGRICULTURAL LAND II. FALLOW Shrubbery Sparse Sample Photos: 2_48 1_33 2_16 4_22 5_22 3_23 6_22 1_27 5_7 5_21 5_2 4_44 244 5_44 6_44 2_8 Accuracy Assessment Pho tos: BVIS_AA_58 BVIS_AA_99 245 I. AGRICULTURAL LAND II. FALLOW Shrubbery Dense & Emergent Red Weeds Sample Photos: 5_25 6_25 Accuracy Assessment Photos: BVIS_AA_125 BVIS_AA_126 246 I. AGRICULTURAL LAND II. FALLOW Shrubbery Blackjack w Emergent Weeds Sample Photos: 4_9 Accuracy Assessment Photos: BVIS_AA_145 247 I. AGRICULTURAL LAND II. FALLOW Shrubbery Dense Shrubbery w Medium Soil & Straw Sample Photos: 3_19 Accuracy Assess ment Photos: None 248 I. AGRICULTURAL LAND III. AGRICULTURAL INFRASTRUCTURE Irrigation Canal Concrete Sample Photos: 3_25 Accuracy Assessment Photos: BVIS_AA_118 BVIS_AA_147 249 I. AGRICULTURAL LAND III. AGRICULTURAL INFRAS TRUCTURE Human Influence Road Sample Photos: None Accuracy Assessment Photos: BVIS_AA_124 BVIS_AA_149 BVIS_AA_97 I. AGRICULTURAL LAND IV: TREES Banana Trees & Dried Grass Sample Photos: 1_7 1_5 Accuracy Assessment Photos: None 250 I. AGRICULTURAL LAND IV: TREES Banana Trees & Dense Shrubbery Sample Photos: None Accuracy Assessment Photos: BVIS_AA_159 I. AGRICULTURAL LAND IV: TREES Green Foliage on Active Agricultural Land Sample Photo s: 2_4 Accuracy Assessment Photos: None 251 I. AGRICULTURAL LAND IV: TREES Green Foliage on Fallow Land Sample Photos: 7_46 11_38 2_5 Accuracy Assessment Photos: BVIS_AA_164 252 I. AGRICULTURAL LAND IV: TREES Green Foliage on Fallow Land w/ Green Vegetation Sample Photos: None Accuracy Assessment Photos: BVIS_AA_122 253 Dry Season BVIS Data ID Name X Y Elev Date_ Time Pnt_type Pnt_Acq Offset_Dis Class SubClass Land_Co ver Land_Use Notes 0 1_58 34.6331 - 14. 229 456.8 8/16/ 2016 13:44 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil Agricultural Land Bare earth. Straw to the east north & south. Bare earth to the west. No evidence of standing water. 6 4_57 34.636 - 14.233 455.3 8/16/ 2016 14:46 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land Bare. To the south and west bare with dried grasses. To the east is bare. 10 1_55 34.628 - 14.233 456.5 8/16/ 2016 15:30 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w S traw Agricultural Land Dry harvested rice field. To the North are Baobab trees. Active animal grazing occurring here. 25 7_53 34.634 - 14.238 461.1 8/16/ 2016 17:21 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil Agricultural Land Bare. To the North are Trees. To the South are trees and dried grass. To the West is maize and to the East is dried grass. 41 1_50 34.632 - 14.24 466.3 8/17/ 2016 11:41 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil Agricultural Land Dark Soil. Bare earth. T o the north, grass, shrubs, maize and a drain. To the South, grass and trees. To the west, grass shrubs, drain and maize field To the East, grass and trees. 46 3_50 34.632 - 14.241 463.6 8/17/ 2016 12:23 Sample On 0 Fallow Bare Earth Bare Earth: Dark S oil Agricultural Land Bare earth. Trees and grass located to the East and South. Grass to the West. Shrubs and Grass to the North. 254 48 2_49 34.63 - 14.242 465.4 8/17/ 2016 12:46 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil Agricultural Land Ba re earth. To the north are shrubs and trees. To the south is bare with burned grasses. To the West and East there are shrubs. 50 3_49 34.628 - 14.243 462.0 8/17/ 2016 13:00 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil Agricultural Land Dry harv ested rice field. Maize field & drain to the north. Drain is covered in grass and shrubs. Shrubs in a drain to the West. Harvested rice fields to the south and east. 59 3_47 34.627 - 14.246 466.3 8/17/ 2016 13:48 Sample On 0 Fallow Bare Earth Bare Eart h: Dark Soil Agricultural Land Bare Earth. Surrounding fields have trees and grass. 181 6_7 34.562 - 14.268 500.1 8/20/ 2016 10:51 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land Out. 54 1_47 34.629 - 14.244 466.9 8/17 / 2016 13:24 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land Dry harvested rice field. Trees to the south and east. Dry rice fields to the north and west. 57 2_47 34.627 - 14.245 464.5 8/17/ 2016 13:38 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land Dry rice field. All surrounding fields are dry rice fields. To the South and East are trees. 60 4_47 34.626 - 14.246 465.7 8/17/ 2016 13:57 Sample On 0 Fallow Bare Earth Bare Earth: Dark So il w Straw Agricultural Land Bare earth. Trees and grass located to the West and East. Bare to the North. To the South is grass. 255 64 2_46 34.624 - 14.247 468.1 8/17/ 2016 14:17 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural L and Bare earth. Bare and trees to the north. Trees and grass in all other directions. 68 5_46 34.625 - 14.244 467.5 8/18/ 2016 11:06 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land Bare / Harvested maize field. Harvested maize field to the North and East. Grass and harvested maize to the south. Grass to the west. 69 6_46 34.624 - 14.244 471.5 8/18/ 2016 11:13 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land Harvested maize fields to the No rth and East. Grass to the West and Grass and Harvested Maize to the South. 76 1_45 34.621 - 14.245 471.2 8/18/ 2016 11:46 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land Harvested rice field. To the north, west and east are rice fields. To the south there is a maize between rice fields and grasses. 79 2_45a 34.618 - 14.244 471.2 8/18/ 2016 12:11 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land 107 2_40 34.598 - 14.248 481.8 8/18/ 2016 15 :44 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land Trees and shrubs to the north and west. To the east there are dry harvested rice fields. To the south are rice fields. 13 1_51 34.625 - 14.236 464.5 8/16/ 2016 15:54 Sam ple On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land Dry harvest fields but the soil here is no longer cracked 256 12 1_52 34.626 - 14.235 462.9 8/16/ 2016 15:45 Sample On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricu ltural Land Dry harvested rice field. To the south is dry irrigation canal. 190 1_1 34.55 - 14.273 507.1 8/20/ 2016 11:31 Sample On 0 Fallow Bare Earth Bare Earth: Light Soil Agricultural Land Bare earth. Harvested maize field where only straw remains . 187 4_2 34.552 - 14.272 504.4 8/20/ 2016 11:20 Sample On 0 Fallow Bare Earth Bare Earth: Light Soil Agricultural Land Bare earth. Harvested Cotton Field. 191 2_1 34.549 - 14.274 507.1 8/20/ 2016 11:34 Sample On 0 Fallow Bare Earth Bare Earth: Li ght Soil w Straw & Ridging Agricultural Land Dry, harvested maize field. 174 3_2 34.556 - 14.272 500.7 8/20/ 2016 10:18 Sample On 0 Fallow Bare Earth Bare Earth: Light Soil w Straw & Ridging Agricultural Land Bare field covered only by straw and spiney plants 179 3_6 34.559 - 14.268 500.7 8/20/ 2016 10:42 Sample On 0 Fallow Bare Earth Bare Earth: Light Soil w Straw & Ridging Agricultural Land 159 1_2 34.553 - 14.273 505.3 8/19/ 2016 16:15 Sample On 0 Fallow Bare Earth Bare Earth: Medium Soil w Straw & Ridging Agricultural Land Dry harvested maize field. Bare earth. 160 2_2 34.552 - 14.273 503.2 8/19/ 2016 16:18 Sample On 0 Fallow Bare Earth Bare Earth: Medium Soil w Straw & Ridging Agricultural Land Bare earth with sporadic spiked weeds. 183 7_7 34.563 - 14.266 501.7 8/20/ 2016 10:57 Sample On 0 Fallow Bare Earth Bare Earth: Medium Soil w Straw & Ridging Agricultural Land 257 110 1_18 34.585 - 14.265 487.9 8/18/ 2016 16:02 Sample On 0 Active Beans Beans: Mature Agricultural Land Beans are nearly mature. No evidence of beans on plants yet though. Other fields are covered in shrubs with nothing planted. 113 6_36 34.594 - 14.25 483.7 8/18/ 2016 16:11 Sample On 0 Active Beans Beans: Mature Agricultural Land Peas. Trees to the north and east. To the west is grass. To the south is trees and peas. 108 1_24 34.592 - 14.261 485.2 8/18/ 2016 15:46 Sample On 0 Active Beans Beans: Mature & Emergent Weeds Agricultural Land Newly planted bean field. Bean field to the immediate south, otherwise a ll other fields are shrubs 83 3_31 34.597 - 14.257 480.6 8/18/ 2016 12:36 Sample On 0 Active Beans Beans: Mature & Emergent Weeds Agricultural Land Edge of a bean field. Mature beans on the vines. Mixture of weeds present. 61 3_38 34.605 - 14.253 475.1 8 /17/ 2016 13:57 Sample On 0 Active Beans Beans: Young Agricultural Land Newly turned bean field surrounded by shrubbery and fallow land. 62 4_38 34.6034 - 14.252 476.0 8/17/ 2016 14:08 Sample On 0 Active Beans Beans: Young Agricultural Land Newly turned bean field. Maize planted adjacent. 93 2_19 34.591 - 14.265 482.8 8/18/ 2016 13:41 Sample On 0 Active Beans Beans: Young Agricultural Land Young bean field. To the north is bare. Trees to the west. 112 2_18 34.587 - 14.266 487.0 8/18/ 2016 16:09 Sample On 0 Active Beans Beans: Young Agricultural Land Maize planted with beans immediately east. To the West, a field intercropped with maize and sweet potato and another field with maize and beans. 258 170 1_6 34.56 - 14.271 500.4 8/20/ 2016 10:01 Sample On 0 Active Beans Beans: Young Agricultural Land 137 4_9 34.571 - 14.271 490.7 8/19/ 2016 13:45 Sample On 0 Fallow Shrubbery Blackjack w Emergent Weeds Agricultural Land Full field of blackjack. Maize located along the periphery. 161 1_21 34.579 - 14.259 484.6 8/20/ 2016 9:29 Sample On 0 Fallow Charred Ground Charred Ground on Bare Earth Agricultural Land Ash in the center of a harvested rice field. Rice fields in all directions. 96 5_19 34.59 - 14.263 484.0 8/18/ 2016 13:55 Sample On 0 Fallow Charre d Ground Charred Ground w Beans Agricultural Land Young bean fields. Adjacent plots are all beans. 133 2_9 34.574 - 14.269 490.1 8/19/ 2016 13:25 Sample On 0 Fallow Charred Ground Charred Ground w Cowpea Agricultural Land Cow pea field. In the center is a burned area. Surroundings are all cowpea with one plot of maize. 149 3_10 34.578 - 14.269 486.4 8/19/ 2016 14:54 Sample On 0 Fallow Charred Ground Charred Ground w Cowpea Agricultural Land In the center of a cowpea field directly in the charred ar ea. Surrounded by cowpea. Fields to the south are barren. 20 4_52 34.629 - 14.238 463.2 8/16/ 2016 16:41 Sample Offset 4 Fallow Charred Ground Charred Ground w Emergent Weeds Agricultural Land Ash circle located in the center of a harvest field with gr een grasses growing. To the North and West are dry rice fields. To the South and East are maize fields. 82 2_31 34.596 - 14.256 481.8 8/18/ 2016 12:28 Sample On 0 Fallow Charred Ground Charred Ground w Emergent Weeds Agricultural Land Inside of a barr en field covered with shrubbery. Specifically in a burned central circle. Field is otherwise covered 259 in low vegetative shrubbery. 90 4_25 34.595 - 14.261 482.4 8/18/ 2016 13:13 Sample On 0 Fallow Charred Ground Charred Ground w Emergent Weeds Agricult ural Land Dry, harvested rice field. Low shrubbery including red burned grass and charring in the center. 125 2_22 34.582 - 14.259 488.5 8/19/ 2016 12:41 Sample Offset 3 Fallow Charred Ground Charred Ground w Shrubbery Agricultural Land Dry harvested r ice fields. There is evidence of ash here. 186 2_15 34.575 - 14.263 496.2 8/20/ 2016 11:18 Sample On 0 Active Cowpea Cowpea & Shrubbery Agricultural Land Cowpea fields completely surrounding this one. 188 3_15 34.574 - 14.264 494.6 8/20/ 2016 11:23 Sampl e On 0 Active Cowpea Cowpea & Shrubbery Agricultural Land Cow pea field surrounded by other cowpea fields. 146 2_10 34.576 - 14.271 491.9 8/19/ 2016 14:38 Sample On 0 Active Cowpea Cowpea: Mature Agricultural Land Cowpea field surrounded by other cowpe a fields. 184 1_15 34.573 - 14.263 498.0 8/20/ 2016 11:11 Sample On 0 Active Cowpea Cowpea: Mature Agricultural Land Cowpea fields completely surrounding this plot. 135 2_23 34.586 - 14.262 485.8 8/19/ 2016 13:37 Sample On 0 Active Cowpea Cowpea: Young Agricultural Land Cow pea field surrounded by other cow pea fields. 145 1_10 34.575 - 14.271 489.8 8/19/ 2016 14:32 Sample On 0 Active Cowpea Cowpea: Young Agricultural Land Primarily bare dirt. Cowpea field but young and barely sprouted. Surrounding fi elds are all cowpea. 182 7_21 34.577 - 14.262 492.5 8/20/ 2016 10:53 Sample On 0 Active Cowpea Cowpea: Young Agricultural Land Cowpea fields to the south and east. Dry harvested rice fields to the north and west. 260 94 3_19 34.59 - 14.266 478.5 8/18/ 2016 1 3:46 Sample On 0 Fallow Shrubbery Dense Shrubbery w/ Medium Soil & Straw Agricultural Land Division mound between fields. Field to the south is turned for new planting. Surrounding fields are shrubbery. 2 3_57b 34.638 - 14.232 458.7 8/16/ 2016 14:15 Sa mple On 0 Fallow Dried Fields Dried Grass Agricultural Land To the north and east are maize. Dried grasses and brush are located to the south and west 3 3_57 34.638 - 14.232 458.4 8/16/ 2016 14:17 Sample On 0 Fallow Dried Fields Dried Grass Agricul tural Land To the north and east are maize. Dried grasses and brush are located to the south and west 5 1_54 34.637 - 14.235 457.1 8/16/ 2016 14:37 Sample On 0 Fallow Dried Fields Dried Grass Agricultural Land To the north, east, and west are dried gr asses. To the south are grasses but it is also bare here. 63 1_46 34.625 - 14.247 466.9 8/17/ 2016 14:08 Sample Offset 6 Fallow Dried Fields Dried Grass Agricultural Land Bare field. Termite mound about 2m from point. Grass to the North. Trees and grass in all other directions. 8 2_56 34.632 - 14.234 457.1 8/16/ 2016 15:08 Sample On 0 Fallow Dried Fields Dried Grass & Emergent c3 Vegetation Agricultural Land To the north is the road. Dry harvested fields surround this entire area 39 1_42 34.607 - 1 4.248 475.4 8/17/ 2016 11:30 Sample On 0 Fallow Dried Fields Dried Grass & Emergent c3 Vegetation Agricultural Land Out. Shrubs amid grassland that backs up to a village. There is a pump nearby with standing water. 261 45 1_38 34.607 - 14.250 475.4 8/17/ 2016 12:14 Sample On 0 Fallow Dried Fields Dried Grass & Emergent c3 Vegetation Agricultural Land Dried grasses along the roadway. Fields to the north were planted with rice. Now beans and maize and growing in the field. Maize is planted closest to th e village. 92 1_19 34.593 - 14.264 487.0 8/18/ 2016 13:34 Sample On 0 Fallow Dried Fields Dried Grass & Emergent c3 Vegetation Agricultural Land Tall, dried grasses along the scheme periphery 116 1_12 34.586 - 14.267 488.2 8/18/ 2016 16:46 Sample On 0 Fallow Dried Fields Dried Grass & Emergent c3 Vegetation Agricultural Land Division between crops. Beans to the east, sweet potatoes to the north. Maize to the west and east and south, enveloping this field. 129 7_16 34.58 - 14.266 488.8 8/19/ 2016 12 :57 Sample On 0 Fallow Dried Fields Dried Grass & Emergent c3 Vegetation Agricultural Land Dry, mounded grasses separating dry, harvested rice fields. 18 2_52 34.627 - 14.239 466.0 8/16/ 2016 16:29 Sample On 0 Fallow Dried Fields Horizontal Rice St raw w Emergent Weeds Agricultural Land Division mound between plots. Green maize growing to the west. All other surrounding plots are dry harvested maize. 11 2_55 34.626 - 14.234 461.4 8/16/ 2016 15:38 Sample On 0 Fallow Dried Fields Horizontal Rice S traw w Emergent Weeds Agricultural Land To the north are Baobab Trees along the periphery of the scheme. Otherwise all other fields are 262 dry, harvested fields. 15 3_51 34.623 - 14.239 468.4 8/16/ 2016 16:10 Sample On 0 Fallow Dried Fields Horizontal Ri ce Straw w Emergent Weeds Agricultural Land Mounded division between two fields nearby. Dry canal with creeping grass. To the north is an irrigation canal and road. Other fields nearby are dry harvested rice. 26 10_53 34.635 - 14.237 459.3 8/16/ 2016 17:3 3 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Bare earth. Maize fields to the north and west. Dried grasses to the south and east. 27 1_48 34.623 - 14.24 466.0 8/17/ 2016 10:30 Sample On 0 Fallow Dried Fi elds Horizontal Rice Straw w Emergent Weeds Agricultural Land Harvested rice field. Canal to the north. All surrounding fields are harvested rice fields. 30 3_44 34.616 - 14.245 475.4 8/17/ 2016 10:44 Sample On 0 Fallow Dried Fields Horizontal Rice S traw w Emergent Weeds Agricultural Land Dry harvested rice field. Canal to the right. Maize to the southeast. 33 3_48 34.625 - 14.24 470.9 8/17/ 2016 10:55 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Dark S oil. Dry harvested rice field surrounded by the same. 49 2_37 34.6 - 14.253 480.9 8/17/ 2016 12:53 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Dark Soil. Harvested, dry rice field with emergent weeds. 66 6_38 34.606 - 14.252 477.0 8/17/ 2016 14:26 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Harvested rice fields. Can see planted maize closer to the roadway 263 73 4_46 34.623 - 14.245 470.9 8/18/ 2016 11:32 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Dark Soil. Harvested rice field. Surrounding fields are a combination of harvested rice and grasses. 74 10_38 34.61 - 14.252 478.5 8/18/ 2016 11:36 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Bare field surrounded by planted maize. Trees located to the southwest 85 4_48 34.622 - 14.243 467.5 8/18/ 2016 12:44 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Dark Soil. Dry harvested rice fields completely surrounding this plot. 88 5_48 34.625 - 14.243 465.7 8/18/ 2016 13:08 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land To the North and We st are grass. Maize to the south and east. 91 6_48 34.624 - 14.242 467.2 8/18/ 2016 13:18 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Dry harvested rice field. 99 2_36 34.597 - 14.251 479.1 8/18/ 2016 14:59 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Harvested rice fields. Rice fields and brush. 102 4_36 34.598 - 14.25 480.3 8/18/ 2016 15:21 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Dark Soil. Rice fields and brush. All surrounding areas and rice fields 109 3_40 34.597 - 14.248 482.4 8/18/ 2016 15:52 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Creeping grass. To the north is a termite mound and shrubs. To the west, east, and south are trees. 264 111 5_36 34.595 - 14.25 483.7 8/18/ 2016 16:04 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Out. Dark Soil. TO the north are t rees. To the south are trees and grass. To the west is a tilled field. To the east is a rice field. 114 1_35 34.589 - 14.252 486.7 8/18/ 2016 16:30 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Green peas grow ing here. Elephant grass to the west. East and South to the peas 119 1_28 34.581 - 14.255 488.2 8/19/ 2016 12:08 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land All surrounding fields are dry harvested rice 121 2_28 34.581 - 14.256 492.2 8/19/ 2016 12:21 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land A canal to the North, east, and West. Dry irrigated rice fields surround. 123 1_22 34.582 - 14.258 489.2 8/19/ 2016 12: 33 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Dry irrigated rice fields. 143 7_22 34.583 - 14.258 486.1 8/19/ 2016 14:15 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultur al Land Trees to the north. Dry harvested rice in all other directions 144 3_28 34.584 - 14.256 483.7 8/19/ 2016 14:31 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Dark Soil. Dry harvested rice fields. 148 4 _28 34.582 - 14.254 476.7 8/19/ 2016 14:49 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Dry harvested rice field 265 150 5_28 34.581 - 14.253 479.1 8/19/ 2016 14:58 Sample On 0 Fallow Dried Fields Horizontal Ric e Straw w Emergent Weeds Agricultural Land Dry harvested rice field. Medium soil. 163 2_21 34.577 - 14.259 484. 8/20/ 2016 9:37 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Harvested rice field with brush. Th e plots immediately surrounding this one are all the same. 164 2_7 34.566 - 14.268 494.6 8/20/ 2016 9:42 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land 166 3_21 34.575 - 14.258 488.5 8/20/ 2016 9:47 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Cowpeas to the north. Dry harvested rice in all other directions. 169 4_21 34.576 - 14.259 487.9 8/20/ 2016 9:57 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agricultural Land Brush beginning to grow along dry harvested rice field. All surrounding fields appear similar. 180 6_21 34.576 - 14.261 493.7 8/20/ 2016 10:48 Sample On 0 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds Agric ultural Land Dry harvested rice field with horizontal straw. Lots of straw. 43 4_39 34.606 - 14.251 475.1 8/17/ 2016 12:07 Sample On 0 Active General Agriculture Intercrop: Mature Maize & Beans Agricultural Land Bund separating a maize field to the no rth and a field planted with beans to the south. 47 2_38 34.603 - 14.252 476.7 8/17/ 2016 12:36 Sample On 0 Active General Agriculture Intercrop: Young Maize & Beans Agricultural Land Division between two plots. To the south, young maize is planted. T o the north, beans are planted through to the 266 northern edge of the scheme. 89 3_25 34.594 - 14.261 485.8 8/18/ 2016 13:10 Sample Offset 2 Ag Infrast ructure Irrigation Canal Irrigation Canal: Concrete Infrastructu re Irrigation Canal. No water present. 1 2_57 34.635 - 14.230 458.4 8/16/ 2016 14:02 Sample On 0 Active Maize Maize: Mature Agricultural Land Plots nearby are a mixture between bare and dried grasses. 21 5_52 34.629 - 14.239 465.1 8/16/ 2016 16:48 Sample Offset 8 Active Maize Maize: Mature Agricultural Land From this vantage point (mounded division) maize fields in all directions. 22 4_53 34.631 - 14.239 464.2 8/16/ 2016 17:01 Sample On 0 Active Maize Maize: Mature Agricultural Land Inside of a maize field. From this vantage point, maiz e fields in all directions. 23 5_53 34.631 - 14.238 464.8 8/16/ 2016 17:05 Sample On 0 Active Maize Maize: Mature Agricultural Land Inside of a maize field. From this vantage point, maize fields in all directions. 40 1_39 34.607 - 14.25 476.4 8/17/ 201 6 11:40 Sample On 0 Active Maize Maize: Mature Agricultural Land Maize field. Maize to the North. Small outcrop of maize to the south but then fields turn to dry harvested rice. 42 3_39 34.605 - 14.251 477.3 8/17/ 2016 12:03 Sample On 0 Active Maize Maize: Mature Agricultural Land Maize field. No intercropping. Field to the immediate south is newly planted maize field. Adjacent field to the west is beans. 75 5_39 34.609 - 14.252 479.4 8/18/ 2016 11:43 Sample On 0 Active Maize Maize: Mature Agricul tural Land Located in maize field. Tallest maize in the area; adjacent to canal. 267 80 4_45 34.617 - 14.245 468.1 8/18/ 2016 12:22 Sample On 0 Active Maize Maize: Mature Agricultural Land Field is divided in half. From this vantage point, all areas around are maize. 158 1_3 34.5597 - 14.274 501.0 8/19/ 2016 16:03 Sample Offset 10 Active Maize Maize: Mature Agricultural Land Irrigated maize fields. First signs of flowing water. 72 9_38 34.608 - 14.252 479.7 8/18/ 2016 11:28 Sample On 0 Active Maize Maize : Young Agricultural Land Young maize growth. One plot to the east is beans. The rest are maize. 117 2_12 34.585 - 14.268 492.8 8/18/ 2016 16:51 Sample On 0 Active Maize Maize: Young Agricultural Land Edge of maize field. Bounded up dirt covered in veget ation that is low to the ground. Termite mound is present here. There are signs of water here, but no water in the irrigation canal. 132 1_9 34.572 - 14.269 494.3 8/19/ 2016 13:19 Sample Offset 10 Active Maize Maize: Young Agricultural Land Young, planted maize field immediately adjacent to a tertiary canal/ 176 2_3 34.559 - 14.272 500.7 8/20/ 2016 10:28 Sample On 0 Active Maize Maize: Young Agricultural Land irrigated maize fields. 58 5_33 34.605 - 14.255 472.4 8/17/ 2016 13:48 Sample On 0 Active Maiz e Maize: Young w Dense Shrubbery Agricultural Land Planted, young maize field. Young maize all around and shrubbery / grasses growing among the maize 65 5_38 34.607 - 14.253 474.2 8/17/ 2016 14:19 Sample On 0 Active Maize Maize: Young w Dense Shrubbery Ag ricultural Land Newly turned maize field mixed with shrubbery 268 70 8_38 34.609 - 14.251 477.9 8/18/ 2016 11:23 Sample On 0 Active Maize Shrubbery: Dense Agricultural Land Newly planted maize fields covered in shrubbery and low vegetation. Surrounding field s are the same with some maize in various stages of growth. 151 1_11 34.581 - 14.268 485.8 8/19/ 2016 15:06 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Division between two dry, harvested rice fields that are covered in red shrubber y. 106 7_25 34.592 - 14.258 485.2 8/18/ 2016 15:34 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Dry, harvested maize area covered in low shrub growth 142 3_8 34.569 - 14.271 493.7 8/19/ 2016 14:09 Sample On 0 Fallow Shrubbery Shru bbery: Dense Agricultural Land Out. Overgrowth of vegetation with an irrigation canal 154 2_27 34.578 - 14.256 486.4 8/19/ 2016 15:34 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Dry harvested rice field. Dry harvested rice fields com pletely surrounding this one. 155 3_27 34.577 - 14.257 488.2 8/19/ 2016 15:42 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Dry harvested rice field. Cowpea fields to the north. Dry rice plots to the south and west. Dry rice field wit h cowpea growing to the east. 165 3_7 34.565 - 14.268 493.4 8/20/ 2016 9:44 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land 167 4_7 34.564 - 14.268 495.6 8/20/ 2016 9:48 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural La nd 269 171 2_20 34.573 - 14.26 487.6 8/20/ 2016 10:07 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Dry harvested rice fields in all directions. Canal located to the east. 175 4_20 34.575 - 14.261 491.3 8/20/ 2016 10:26 Sample On 0 Fal low Shrubbery Shrubbery: Dense Agricultural Land Dry harvested rice fields with shrubbery in all directions. 177 2_6 34.56 - 14.27 500.7 8/20/ 2016 10:34 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land 192 1_14 34.57 - 14.266 499.5 8/20/ 2016 11:49 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Dry harvested rice field. Dry grass to the north and south. To the west has been cultivated. To the East are sweet potato and cowpeas intercropped. 28 1_44 34.615 - 14.244 472.7 8/17/ 2016 10:34 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Division mound between plots. To the North are dry harvested rice fields, brush and weeds on all sides. 84 4_31 34.596 - 14.257 481.2 8/18/ 2016 12:41 Sample Offs et 4 Fallow Shrubbery Shrubbery: Dense Agricultural Land Edge of bean field. All barren to the north. 122 3_16 34.576 - 14.266 491.0 8/19/ 2016 12:25 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Division between plots marked wi th lots of rice straw. Surrounding fields are fallow. 29 2_44 34.616 - 14.244 475.4 8/17/ 2016 10:40 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Footpath along a canal. Located in a dry harvested field. 37 7_44 34.612 - 14.247 474 .2 8/17/ 2016 11:09 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Medium Soil. Dry, harvested rice field with a plot division located directly behind. 270 38 1_43 34.609 - 14.248 473.6 8/17/ 2016 11:17 Sample On 0 Fallow Shrubbery Sh rubbery: Dense Agricultural Land Medium Soil. Bare earth along the roadway and canal. Dry, harvested fields to the south. Evidence of more brush cover now in addition to the presence of children in the field. Village to North. 52 2_33 34.6 - 14.255 482. 8 8/17/ 2016 13:08 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Harvested rice fields covered in shrubbery. Burned areas in the center of the field. Low growing red vegetation growing sporadically. 53 3_33 34.603 - 14.257 477.0 8/17 / 2016 13:20 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Harvested rice field with shrubbery and cracked, dry soil. Presence of red vegetation to the east. 55 1_26 34.604 - 14.258 476.4 8/17/ 2016 13:25 Sample On 0 Fallow Shrubbe ry Shrubbery: Dense Agricultural Land Harvested, dry field covered in red and green shrubbery. Maize growing along the south of the field. 56 4_33 34.603 - 14.256 479.4 8/17/ 2016 13:33 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Lan d Harvest field. Red and green shrubbery growing. No other crops growing in view. 81 1_31 34.597 - 14.255 481.2 8/18/ 2016 12:23 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Barren field covered in low, red hued vegetation. 271 87 2_2 5 34.597 - 14.258 481.2 8/18/ 2016 12:58 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Cleared field. Beans to the immediate south. Otherwise fields in various states from cleared to turned for beans 97 1_36 34.595 - 14.252 473.9 8/18/ 2016 14:49 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Dry harvested rice field. Grass and trees to the north. Dry harvested rice fields to the south, west, and east. 98 5_31 34.594 - 14.254 477.6 8/18/ 2016 14:49 Sample On 0 Fal low Shrubbery Shrubbery: Dense Agricultural Land Dry, harvest field. All surrounding fields are planted. 100 1_30 34.591 - 14.256 480.6 8/18/ 2016 15:04 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Dry, harvested rice fields. Con siderable overgrowth here. Large trees nearby 101 3_36 34.597 - 14.25 475.7 8/18/ 2016 15:12 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Dry harvested rice fields and brush. Dry harvested rice fields all around. 19 3_52 34.628 - 14 .237 463.6 8/16/ 2016 16:36 Sample On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land Harvested field with green grasses growing. To the north are shrubs on a harvested field. In all other directions there are dry harvested fields. 103 5_25 34 .594 - 14.258 485.8 8/18/ 2016 15:24 Sample On 0 Fallow Shrubbery Shrubbery: Dense & Emergent Red Weeds Agricultural Land Harvested rice field with red weeds. Maize planted to the west. 104 6_25 34.592 - 14.259 485.2 8/18/ 2016 15:30 Sample On 0 Fallow Shrubbery Shrubbery: Dense & Emergent Red Weeds Agricultural Land Dry, harvested rice field covered in red weeds. Everything to the south of this point appears similar. 272 32 4_44 34.614 - 14.246 473.3 8/17/ 2016 10:53 Sample On 0 Fallow Shrubbery Shru bbery: Sparse Agricultural Land Medium Soil. Dry harvest rice field surrounded by the same. 34 5_44 34.614 - 14.247 474.8 8/17/ 2016 10:56 Sample Offset 1 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Dark Soil. Grasses to the North and West. Maize to the South and East. 35 6_44 34.612 - 14.246 476.0 8/17/ 2016 11:05 Sample Offset 3 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Medium Soil. Grasses. Dry harvested rice fields. 140 2_8 34.568 - 14.271 494.9 8/19/ 2016 14:05 Sample O n 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Mounded division between fields. The straw is hard packed. 31 2_48 34.624 - 14.24 468.4 8/17/ 2016 10:45 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Harvested rice field covered in brush. Dry harvested rice fields surrounding 51 1_33 34.601 - 14.254 480.9 8/17/ 2016 13:03 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Harvested rice field with small, sparse shrubs growing across the plot. Barren fields all around. 120 2_16 34.576 - 14.265 491.9 8/19/ 2016 12:19 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Small triangular plot with dirt turned and ready for planting. Low vegetative growth. 130 4_22 34.581 - 14.261 486.4 8/19/ 201 6 12:58 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Dry harvested rice fields surrounding. 131 5_22 34.58 - 14.262 496.2 8/19/ 2016 13:05 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Dark Soil. Dry harveste d rice fields. 138 3_23 34.585 - 14.258 485.5 8/19/ 2016 13:54 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land To the north are grasses growing on dry harvest rice fields. To the south is rice. 273 141 6_22 34.584 - 14.258 483.7 8/19/ 2016 14:09 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Dry irrigated rice fields. 152 1_27 34.577 - 14.254 485.5 8/19/ 2016 15:24 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Dry harvested rice field. Trees and grass to the north. Bean to the south. More harvested rice to the west. Trees and grass to the east. 168 5_7 34.563 - 14.269 496.8 8/20/ 2016 9:52 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land 178 5_21 34.576 - 14.262 492.2 8/20/ 20 16 10:34 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Field is divided in half. Half is cow peas and half is dry harvested rice. Plots to the immediate south and west are the same. In other directions, plots are dry harvested rice. 189 5_2 34.553 - 14.272 506.2 8/20/ 2016 11:23 Sample On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land Bare earth covered in spiked plants. 14 2_51 34.625 - 14.237 463.9 8/16/ 2016 16:00 Sample On 0 Fallow Dried Fields Dried Grass Agricultu ral Land Some minor cracks in the soil. There are dry harvest rice fields here. 128 6_16 34.579 - 14.265 488.5 8/19/ 2016 12:52 Sample On 0 Fallow Dried Fields Dried Grass Agricultural Land Mounded dried grasses and rice straw 134 1_23 34.585 - 14.261 486.4 8/19/ 2016 13:28 Sample On 0 Fallow Dried Fields Dried Grass Agricultural Land Dry harvested rice fields 147 1_29 34.585 - 14.255 482.4 8/19/ 2016 14:38 Sample On 0 Fallow Dried Fields Dried Grass Agricultural Land Dry harvested rice fields wi th straw. 157 4_27 34.576 - 14.257 487.3 8/19/ 2016 15:50 Sample On 0 Fallow Dried Fields Dried Grass Agricultural Land Mounded rice straw in a dry, harvested rice 274 field. Cowpea planted to the north and west. Dry rice to the south and east. 67 3_46 3 4.622 - 14.246 466.0 8/17/ 2016 14:33 Sample On 0 Fallow Dried Fields Dried Grass Agricultural Land Surrounding fields are dried rice fields with grasses 95 4_19 34.59 - 14.265 484.6 8/18/ 2016 13:49 Sample On 0 Ag Infrast ructure Irrigation Canal Shru bbery: Dense Infrastructu re Cement gate to field along tertiary canal/ Large termite mound is also present here. Surround maize along the canal line. 162 1_7 34.565 - 14.266 491.9 8/20/ 2016 9:36 Sample Offset 15 Trees Trees Trees: Banana Trees & Dried G rass Agricultural Land 172 1_5 34.555 - 14.271 498.6 8/20/ 2016 10:12 Sample Offset 10 Trees Trees Trees: Banana Trees & Dried Grass Agricultural Land 156 2_4 34.564 - 14.274 498.6 8/19/ 2016 15:48 Sample Offset 15 Trees Trees Trees: Green Foliage on A ctive Agricultur al Land Agricultural Land Tree located along the scheme periphery (road) 71 7_46 34.622 - 14.244 471.2 8/18/ 2016 11:25 Sample Offset 8 Trees Trees Trees: Green Foliage on Fallow Field Agricultural Land Bare with a tree present. Patches of grass in the North, South, and West. Grass in the East. 77 11_38 34.6102 - 14.2509 478.2 8/18/ 2016 11:54 Sample On 0 Trees Trees Trees: Green Foliage on Fallow Field Agricultural Land Trees located along the periphery of the scheme 275 185 2_5 34.553 - 14. 269 505.3 8/20/ 2016 11:13 Sample Offset 40 Trees Trees Trees: Green Foliage on Fallow Field Agricultural Land 7 1_56 34.634 - 14.233 457.1 8/16/ 2016 14:54 Sample On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land To th e south, east, and west are dry, harvested rice fields. Termite mound north 9 1_53 34.632 - 14.235 456.2 8/16/ 2016 15:16 Sample On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land Plot surrounded by dry harvested rice fie lds. 17 5_51 34.625 - 14.239 464.8 8/16/ 2016 16:20 Sample On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land Dry harvested rice fields. All surrounding plots are the same. 24 6_53 34.631 - 14.237 462.3 8/16/ 2016 17:11 S ample On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land Harvested rice field. To the North and West are harvested rice fields. To the South and East are Maize fields. 36 1_49 34.627 - 14.24 466.0 8/17/ 2016 11:06 Sample On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land Harvested rice field. To the North is evidence of rice straw. All surrounding areas are dry harvested rice fields. 127 3_22 34.58 - 14.259 487.0 8/19/ 2016 12:49 Sample On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land To the north is a termite mound. Surrounding fields are all dry harvested rice. 173 3_20 34.573 - 14.261 488.8 8/20/ 2016 10:16 Sample On 0 Fallow Dried Fields Vertical R ice Straw w Emergent Weeds Agricultural Land All surrounding fields are dry harvested rice 276 139 1_8 34.568 - 14.27 491.9 8/19/ 2016 14:00 Sample On 0 Fallow Dried Fields Vertical Rice Straw w Green Tops Agricultural Land Former rice fields. Green rice tops are still exposed but bare earth turned for new plantings (holes dug). 86 1_25 34.595 - 14.259 481.8 8/18/ 2016 12:50 Sample On 0 Fallow Dried Fields Vertical Rice Straw w Little Vegetation Agricultural Land Dried rice field with straw collected. Beans to the immediate east. Remaining fields are dried rice. 105 1_40 34.6 - 14.247 481.5 8/18/ 2016 15:33 Sample On 0 Fallow Dried Fields Vertical Rice Straw w Little Vegetation Agricultural Land Dry harvested rice field. Surrounding areas are dry and harvested. 115 1_17 34.582 - 14.266 491.0 8/18/ 2016 16:32 Sample On 0 Fallow Dried Fields Vertical Rice Straw w Little Vegetation Agricultural Land Dry harvested rice field with little other ground cover. Mostly dark soil. 16 4_51 34.624 - 14.23 9 466.9 8/16/ 2016 16:17 Sample On 0 Fallow Dried Fields Vertical Rice Straw w Little Vegetation Agricultural Land Plot is surrounded by dry harvested rice fields. 34 BVIS_ AA_15 9 34.562 - 14.268 449 8/20/ 2016 Accuracy Offset 40 Trees Trees Banana Tr ees & Dense Shrubbery Agricultural Land 46 BVIS_ AA_5 34.6318 - 14.2303 395 8/16/ 2016 Accuracy On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw Agricultural Land 33 BVIS_ AA_15 8 34.5585 - 14.2678 8/16/ 2016 Accuracy On 0 Fallow Bare Earth Bare Earth: Dark Soil w Straw & Ridging Agricultural Land 37 BVIS_ AA_16 3 34.5622 - 14.267 448 8/20/ 2016 Accuracy On 0 Fallow Bare Earth Bare Earth: Dark Soil Agricultural Land 277 w Straw & Ridging 16 BVIS_ AA_13 1 34.5953 - 14.2585 424 8/18/ 2016 Accuracy On 0 Active Beans Beans: Mature Agricultural Land 17 BVIS_ AA_13 2 34.5953 - 14.2585 437 8/18/ 2016 Accuracy On 0 Active Beans Beans: Young Agricultural Land 22 BVIS_ AA_14 5 34.5667 - 14.2655 455 8/19/ 2016 Accuracy On 0 Fallow Shrubbery Blackjack w Emerge nt Weeds Agricultural Land 20 BVIS_ AA_13 8 34.5927 - 14.257 442 8/18/ 2016 Accuracy Offset 25 Fallow Charred Ground Charred Ground w Emergent Weeds Agricultural Land 55 BVIS_ AA_75 34.6274 - 14.2382 410 8/16/ 2016 Accuracy On 0 Fallow Charred Ground Charred Ground w Emergent Weeds Agricultural Land Lots of straw is heaped up around this circle 15 BVIS_ AA_12 8 34.597 - 14.2571 421 8/18/ 2016 Accuracy On 0 Active Cowpea Cowpea & Shrubbery Agricultural Land 11 BVIS_ AA_12 3 34.6101 - 14.2519 407 8/18/ 20 16 Accuracy On 0 Active Cowpea Cowpea: Mature Agricultural Land 40 BVIS_ AA_28 34.637 - 14.235 396 8/16/ 2016 Accuracy On 0 Fallow Dried Fields Dried Grass Agricultural Land 21 BVIS_ AA_14 4 34.5865 - 14.2658 450 8/18/ 2016 Accuracy On 0 Active Genera l Agriculture Intercrop: Young Maize & Beans Agricultural Land 8 BVIS_ AA_11 8 34.5966 - 14.253 398 8/18/ 2016 Accuracy On 0 Ag Infrast ructure Irrigation Canals Irrigation Canal: Concrete Infrastructu re Concrete tertiary canal. Charred ground & straw 2 4 BVIS_ AA_14 7 34.5741 - 14.2718 459 8/19/ 2016 Accuracy On 0 Ag Infrast ructure Irrigation Canals Irrigation Canal: Concrete Infrastructu re Surrounded by green grasses and shrubbery 45 BVIS_ AA_47 34.6281 - 14.233 401 8/16/ 2016 Accuracy On 0 Ag Infrast ructur e Irrigation Canals Bare Earth: Dark Soil w Straw Infrastructu re 278 5 BVIS_ AA_10 8 34.6069 - 14.2482 404 8/17/ 2016 Accuracy On 0 Active Maize Maize: Mature Agricultural Land Emergent vegetation growing between rows; not intercropped 6 BVIS_ AA_10 9 34.607 - 14.25 404 8/17/ 2016 Accuracy On 0 Active Maize Maize: Mature Agricultural Land Field is adjacent to a fallow field covered in emergernt vegetation 7 BVIS_ AA_11 6 34.6061 - 14.25 403 8/17/ 2016 Accuracy On 0 Active Maize Maize: Mature Agricultural L and Potentially an intercropped field 27 BVIS_ AA_15 34.6351 - 14.2308 394 8/16/ 2016 Accuracy On 0 Active Maize Maize: Mature Agricultural Land On dark soil. No intercropping 35 BVIS_ AA_16 34.635 - 14.2304 395 8/16/ 2016 Accuracy On 0 Active Maize Mai ze: Mature Agricultural Land No intercropping. Emergent weeds present. 58 BVIS_ AA_80 34.629 - 14.2389 405 8/16/ 2016 Accuracy On 0 Active Maize Maize: Mature Agricultural Land No intercropping or emergent weeds 60 BVIS_ AA_86 34.6311 - 14.2381 404 8/16/ 2016 Accuracy On 0 Active Maize Maize: Mature Agricultural Land No intercropping w little emergent weeds 29 BVIS_ AA_15 1 34.552 - 14.2734 470 8/19/ 2016 Accuracy On 0 Active General Agriculture Mustard Greens Agricultural Land What crop is this? 12 BVIS_ AA_12 4 34.5981 - 14.2578 416 8/18/ 2016 Accuracy Offset 10 Ag Infrast ructure Human Influence Road Infrastructu re 26 BVIS_ AA_14 9 34.564 - 14.2742 469 8/19/ 2016 Accuracy On 0 Ag Infrast ructure Human Influence Road Infrastructu re 64 BVIS_ AA_97 34 .613 - 14.253 387 8/17/ 2016 Accuracy On 0 Ag Infrast ructure Human Influence Road Agricultural Land Looking out toward drain at southwest corner of irrigation scheme 18 BVIS_ AA_13 3 34.591 - 14.2648 436 8/18/ 2016 Accuracy On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land 279 39 BVIS_ AA_16 6 34.5529 - 14.272 455 8/20/ 2016 Accuracy On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Land 0 BVIS_ AA_10 0 34.6161 - 14.2439 385 8/17/ 2016 Accuracy On 0 Fallow Shrubbery Shrubbery: Dense Agricultural Lan d 44 BVIS_ AA_4 34.6332 - 14.2331 400 8/16/ 2016 Accuracy Offset 3 Fallow Shrubbery Shrubbery: Dense Agricultural Land 19 BVIS_ AA_13 4 34.59 - 14.265 438 8/18/ 2016 Accuracy Offset 10 Fallow Shrubbery Shrubbery: Dense Infrastructu re 13 BVIS_ AA_12 5 34.5981 - 14.2578 414 8/18/ 2016 Accuracy On 0 Fallow Shrubbery Shrubbery: Dense & Emergent red weeds Agricultural Land 14 BVIS_ AA_12 6 34.597 - 14.255 411 8/18/ 2016 Accuracy On 0 Fallow Shrubbery Shrubbery: Dense & Emergent red weeds Agricultural La nd 47 BVIS_ AA_58 34.625 - 14.2359 405 8/16/ 2016 Accuracy On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land 65 BVIS_ AA_99 34.6151 - 14.2437 379 8/17/ 2016 Accuracy On 0 Fallow Shrubbery Shrubbery: Sparse Agricultural Land 50 BVIS_ AA_67 34 .6248 - 14.2366 404 8/16/ 2016 Accuracy On 0 Fallow Dried Fields Dried Grass Agricultural Land 51 BVIS_ AA_68 34.625 - 14.237 405 8/16/ 2016 Accuracy On 0 Fallow Dried Fields Dried Grass Agricultural Land 10 BVIS_ AA_12 2 34.609 - 14.2517 405 8/18/ 2016 A ccuracy Offset 45 Trees Trees Trees: Green Foliage on Fallow Land w Green Vegetation Agricultural Land 38 BVIS_ AA_16 4 34.553 - 14.269 452 8/20/ 2016 Accuracy Offset 40 Trees Trees Trees: Green foliage on Fallow Field Agricultural Land 2 BVIS_ AA_10 3 34.612 - 14.247 393 8/17/ 2016 Accuracy On 0 Fallow Dried Fields Vertical Rice Straw w Agricultural Land 280 Emergent Weeds 3 BVIS_ AA_10 4 34.6112 - 14.247 393 8/17/ 2016 Accuracy On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural L and ~15ft away is a charred circle 25 BVIS_ AA_14 8 34.581 - 14.2679 458 8/19/ 2016 Accuracy On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land Charred circle roughly 20ft away 32 BVIS_ AA_15 6 34.5599 - 14.27 449 8/20/ 2016 Ac curacy On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land 42 BVIS_ AA_34 34.6352 - 14.2329 396 8/16/ 2016 Accuracy On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land 43 BVIS_ AA_35 34.6339 - 1 4.233 398 8/16/ 2016 Accuracy On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land 53 BVIS_ AA_71 34.623 - 14.2392 407 8/16/ 2016 Accuracy On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land Charre d cirlce roughly 15ft away 54 BVIS_ AA_72 34.6248 - 14.2392 408 8/16/ 2016 Accuracy On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land 56 BVIS_ AA_76 34.6273 - 14.2378 409 8/16/ 2016 Accuracy On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land 281 57 BVIS_ AA_77 34.6278 - 14.2368 406 8/16/ 2016 Accuracy On 0 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Agricultural Land 282 Dry Season BVIS Analysis Classes Class Name Samples in Clas s # Crop Samples Land Cover 1 18 1 Dry Season Fallow 2 20 5 Dry Season Fallow 3 43 4 Dry Season Fallow 4 13 2 Dry Season Fallow 5 19 6 Dry Season Fallow 6 35 10 Active Agriculture 7 3 0 Not Vegetated 8 10 4 Active Agriculture 9 12 3 Not Vegetated 10 7 3 Not Vegetated 11 33 8 Not Vegetated Class 1: Dry Season Fallow NAME CLASS SUBCLASS LAND COVER 1_10 Active Cowpea Cowpea: Young 1_22 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 1_30 Fallow Shrubbery Shrubbery: Dense 1_44 F allow Shrubbery Shrubbery: Dense 1_54 Fallow Dried Fields Dried Grass 2_46 Fallow Bare Earth Bare Earth: Dark Soil w Straw 2_48 Fallow Shrubbery Shrubbery: Sparse 3_10 Fallow Charred Ground Charred Ground w Cowpea 3_22 Fallow Dried Fields Vertical Ric e Straw w Emergent Weeds 4_36 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 4_52 Fallow Charred Ground Charred Ground w Emergent Weeds 6_53 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds 7_16 Fallow Dried Fields Dried Grass & Eme rgent c3 Vegetation 7_44 Fallow Shrubbery Shrubbery: Dense BVIS_AA_103 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds BVIS_AA_104 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds BVIS_AA_28 Fallow Dried Fields Dried Grass BVIS_AA_3 4 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds 283 Class 2: Dry Season Fallow NAME CLASS SUBCLASS LAND COVER 2_57 Active Maize Maize: Mature 3_57b Fallow Dried Fields Dried Grass 3_57 Fallow Dried Fields Dried Grass 1_56 Fallow Dried Fie lds Vertical Rice Straw w Emergent Weeds 1_53 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds 10_53 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 5_53 Active Maize Maize: Mature 2_52 Fallow Dried Fields Horizontal Rice Straw w Em ergent Weeds 5_28 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 2_28 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 4_27 Fallow Dried Fields Dried Grass 6_22 Fallow Shrubbery Shrubbery: Sparse 7_22 Fallow Dried Fields Horizont al Rice Straw w Emergent Weeds 1_21 Fallow Charred Ground Charred Ground on Bare Earth 2_20 Fallow Shrubbery Shrubbery: Dense 4_22 Fallow Shrubbery Shrubbery: Sparse 7_21 Active Cowpea Cowpea: Young 5_21 Fallow Shrubbery Shrubbery: Sparse BVIS_AA_16 Active Maize Maize: Mature BVIS_AA_15 Active Maize Maize: Mature 284 Class 3: Dry Season Fallow NAME CLASS SUBCLASS LAND COVER 1_58 Fallow Bare Earth Bare Earth: Dark Soil 4_57 Fallow Bare Earth Bare Earth: Dark Soil w Straw 3_52 Fallow Shrubbery Shrubbery: Dense 2_51 Fallow Dried Fields Dried Grass 7_53 Fallow Bare Earth Bare Earth: Dark Soil 4_53 Active Maize Maize: Mature 4_51 Fallow Dried Fields Vertical Rice Straw w Little Vegetation 3_48 Fallow Dried Fields Horizontal Rice Straw w Emer gent Weeds 3_44 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 2_47 Fallow Bare Earth Bare Earth: Dark Soil w Straw 3_46 Fallow Dried Fields Dried Grass 1_40 Fallow Dried Fields Vertical Rice Straw w Little Vegetation 2_36 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 5_31 Fallow Shrubbery Shrubbery: Dense 3_28 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 2_27 Fallow Shrubbery Shrubbery: Dense 3_33 Fallow Shrubbery Shrubbery: Dense 7_25 Fallow Shrubbery Shrubbery: Dense 2_22 Fallow Charred Ground Charred Ground w Shrubbery 6_21 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 1_18 Active Beans Beans: Mature 4_19 Ag Inf Irrigation Canal Shrubbery: Dense 6_16 Fallow Dried Fields Dried Gr ass 1_11 Fallow Shrubbery Shrubbery: Dense 4_7 Fallow Shrubbery Shrubbery: Dense 2_10 Active Cowpea Cowpea: Mature BVIS_AA_47 Ag Inf Irrigation Canals Bare Earth: Dark Soil w Straw BVIS_AA_4 Fallow Shrubbery Shrubbery: Dense BVIS_AA_35 Fallow Drie d Fields Vertical Rice Straw w Emergent Weeds BVIS_AA_67 Fallow Dried Fields Dried Grass BVIS_AA_77 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds BVIS_AA_68 Fallow Dried Fields Dried Grass BVIS_AA_75 Fallow Charred Ground Charred Ground w/ Emergent Weeds BVIS_AA_86 Active Maize Maize: Mature BVIS_AA_71 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds BVIS_AA_72 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds BVIS_AA_99 Fallow Shrubbery Shrubbery: Sparse BVIS_AA_134 Fall ow Shrubbery Shrubbery: Dense BVIS_AA_156 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds 285 Class 4: Dry Season Fallow NAME CLASS SUBCLASS LAND COVER 1_55 Fallow Bare Earth Bare Earth: Dark Soil w Straw 1_52 Fallow Bare Earth Bare Earth: D ark Soil w Straw 5_51 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds 1_48 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 1_49 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds 3_50 Fallow Bare Earth Bare Earth: Dark Soil 3_36 Fallow Shrubbery Shrubbery: Dense 3_27 Fallow Shrubbery Shrubbery: Dense 3_31 Active Beans Beans: Mature & Emergent Weeds 3_23 Fallow Shrubbery Shrubbery: Sparse 4_20 Fallow Shrubbery Shrubbery: Dense BVIS_AA_5 Fallow Bare Earth Bare Earth: Dark Soil w Straw BVIS_AA_128 Active Cowpea Cowpea & Shrubbery Class 5: Dry Season Fallow NAME CLASS SUBCLASS LAND COVER 2_55 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 5_52 Active Maize Maize: Mature 3_51 Fallow Dried Fields Hori zontal Rice Straw w Emergent Weeds 1_50 Fallow Bare Earth Bare Earth: Dark Soil 6_48 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 3_47 Fallow Bare Earth Bare Earth: Dark Soil 1_29 Fallow Dried Fields Dried Grass 3_21 Fallow Dried Field s Horizontal Rice Straw w Emergent Weeds 4_25 Fallow Charred Ground Charred Ground w Emergent Weeds 2_23 Active Cowpea Cowpea: Young 1_14 Fallow Shrubbery Shrubbery: Dense 7_7 Fallow Bare Earth Bare Earth: Medium Soil w Straw & Ridging 2_12 Active Ma ize Maize: Young 1_6 Active Beans Beans: Young 2_1 Fallow Bare Earth Bare: Earth: Light Soil w Straw & Ridging BVIS_AA_76 Fallow Dried Fieds Vertical Rice Straw w Emergent Weeds BVIS_AA_80 Active Maize Maize: Mature BVIS_AA_100 Fallow Shrubbery Shru bbery: Dense BVIS_AA_144 Active Gen Ag Intercrop: Young Maize & Beans 286 Class 6: Active Agriculture NAME CLASS SUBCLASS LAND COVER 3_49 Fallow Bare Earth Bare Earth: Dark Soil 7_46 Trees Trees Trees: Green Foliage on Fallow Field 1_47 Fallow Bar e Earth Bare Earth: Dark Soil w Straw 6_46 Fallow Bare Earth Bare Earth: Dark Soil w Straw 1_45 Fallow Bare Earth Bare Earth: Dark Soil w Straw 4_47 Fallow Bare Earth Bare Earth: Dark Soil w Straw 6_44 Fallow Shrubbery Shrubbery: Sparse 3_39 Active M aize Maize: Mature 6_38 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 9_38 Active Maize Maize: Young 3_38 Active Beans Beans: Young 5_38 Active Maize Maize: Young w Dense Shrubbery 4_28 Fallow Dried Fields Horizontal Rice Straw w Emergen t Weeds 1_33 Fallow Shrubbery Shrubbery: Sparse 4_33 Fallow Shrubbery Shrubbery: Dense 2_31 Fallow Charred Ground Charred Ground w Emergent Weeds 4_31 Fallow Shrubbery Shrubbery: Dense 4_21 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 6_25 Fallow Shrubbery Shrubbery: Dense & Emergent Red Weeds 3_20 Fallow Dried Fields Vertical Rice Straw w Emergent Weeds 1_15 Active Cowpea Cowpea: Mature 5_19 Fallow Charred Ground Charred Ground w Beans 1_19 Fallow Dried Fields Dried Grass & Emerge nt c3 Vegetation 3_19 Fallow Shrubbery Dense Shrubbery w/ Medium Soil & Straw 1_17 Fallow Dried Fields Vertical Rice Straw w Little Vegetation 1_12 Fallow Dried Fields Dried Grass & Emergent c3 Vegetation 2_7 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 1_9 Active Maize Maize: Young 2_8 Fallow Shrubbery Shrubbery: Sparse 2_3 Active Maize Maize: Young BVIS_AA_116 Active Maize Maize: Mature BVIS_AA_122 Trees Trees Trees: Green foliage on fallow land w green vegetation BVIS_AA_138 Fall ow Charred Ground Charred Ground w Emergent Weeds BVIS_AA_131 Active Beans Beans: Mature BVIS_AA_132 Active Beans Beans: Young Class 7: Not Vegetated NAME CLASS SUBCLASS LANDCOVER 2_56 Fallow Dried Fields Dried Grass & Emergent c3 Vegetation 2_40 Fallow Bare Earth Bare Earth: Dark Soil w Soil 1_35 Fallow Dried fields Horizontal Rice Straw w Emergent Weeds 287 Class 8: Active Agriculture NAME CLASS SUBCLASS LAND COVER 2_38 Active General Ag Intercrop: Young Maize & Beans 5_39 Active Maize Ma ize: Mature 2_37 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 2_33 Fallow Shrubbery Shrubbery: Dense 2_21 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 1_23 Fallow Dried Fields Dried Grass 1_24 Active Beans Beans: Mature & Emergent Weeds 3_7 Fallow Shrubbery Shrubbery: Dense 2_9 Fallow Charred Ground Charred Ground w Cowpea BVIS_AA_133 Fallow Shrubbery Shrubbery: Dense Class 9: Not Vegetated NAME CLASS SUBCLASS LAND COVER 2_49 Fallow Bare Earth Bare Earth: Dark So il 1_28 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 3_25 Ag Inf Irrigation Canal Irrigation Canal: Concrete 3_15 Active Cowpea Cowpea & Shrubbery 2_16 Fallow Shrubbery Shrubbery: Sparse 1_8 Fallow Dried Fields Vertical Rice Straw w G reen Tops 3_8 Fallow Shrubbery Shrubbery: Dense 1_3 Active Maize Maize: Mature BVIS_AA_58 Fallow Shrubbery Shrubbery: Sparse BVIS_AA_123 Active Cowpea Cowpea: Mature BVIS_AA_124 Ag Inf Human Influence Road BVIS_AA_125 Fallow Shrubbery Shrubbery: Den se & Emergent Red Weeds 288 Class 10: Not Vegetated NAME CLASS SUBCLASS LAND COVER 1_51 Fallow Bare Earth Bare Earth: Dark Soil w Straw 1_39 Active Maize Maize: Mature 1_38 Fallow Dried Fields Dried Grass & Emergent c3 Vegetation 10_38 Fallow Dried Fi elds Horizontal Rice Straw w Emergent Weeds BVIS_AA_108 Active Maize Maize: Mature BVIS_AA_109 Active Maize Maize: Mature BVIS_AA_147 Ag Inf Irrigation Canals Irrigation Canal: Concrete Class 11: Not Vegetated NAME CLASS SUBCLASS LAND COVER 4_48 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 5_48 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 5_46 Fallow Bare Earth Bare Earth: Dark Soil w Straw 2_44 Fallow Shrubbery Shrubbery: Dense 2_45A Fallow Bare Earth Bare Eart h: Dark Soil w Straw 4_46 Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds 4_45 Active Maize Maize: Mature 4_44 Fallow Shrubbery Shrubbery: Sparse 1_46 Fallow Dried Fields Dried Grass 5_44 Fallow Shrubbery Shrubbery: Sparse 1_43 Fallow S hrubbery Shrubbery: Dense 4_39 Active Gen Ag Intercrop: Mature Maize & Beans 8_38 Active Maize Shrubbery: Dense 1_36 Fallow Shrubbery Shrubbery: Dense 4_38 Active Beans Beans: Young 1_27 Fallow Shrubbery Shrubbery: Sparse 1_31 Fallow Shrubbery S hrubbery: Dense 5_33 Active Maize Maize: Young w Dense Shrubbery 2_25 Fallow Shrubbery Shrubbery: Dense 5_25 Fallow Shrubbery Shrubbery: Dense & Emergent Red Weeds 1_26 Fallow Shrubbery Shrubbery: Dense 1_25 Fallow Dried Fields Vertical Rice Straw w Little Vegetation 5_22 Fallow Shrubbery Shrubbery: Sparse 2_15 Active Cowpea Cowpea & Shrubbery 2_19 Active Beans Beans: Young 3_16 Fallow Shrubbery Shrubbery: Dense 2_18 Active Beans Beans: Young 5_7 Fallow Shrubbery Shrubbery: Sparse 2_6 Fal low Shrubbery Shrubbery: Dense 4_9 Fallow Shrubbery Blackjack w Emergent Weeds BVIS_AA_126 Fallow Shrubbery Shrubbery: Dense & Emergent Red Weeds BVIS_AA_145 Fallow Shrubbery Blackjack w Emergent Weeds BVIS_AA_148 Fallow Dried Fields Vertical Rice Str aw w Emergent Weeds 289 APPENDIX C : Dry Season LULC Data: Bwanje Valley 290 Bwanje Valley Dry Season Cl assification Table Study Area TOTAL Class Descriptions Sample AA IV. TREES Green Foliage on Bare Soil 1 0 1 Green Foliage on Dry Harvested Agricultural Land w Ridging 3 2 5 I. NON - AGRICULTU RAL LAND Bare Earth Dark Soil w/ Straw 2 1 3 Light Soil 5 1 6 Light Soil w/ Green Vegetation 3 4 7 Charred Ground Charred Ground w/ Dry Vegetation 3 1 4 Dried Grass Dried Grass 0 5 5 Trees Green Foliage & Dry Undergrowth 2 1 3 Gr een Foliage on Bare Soil 3 1 4 Little Green Foliage w/ Dry Undergrowth 4 7 11 Mostly Dried Leaves 0 3 3 Dried Trees on Bare Rock 1 0 1 Human Influence Road 1 0 1 Human Dwellings 1 1 2 Riverbed Riverbed 0 3 3 Study Area TOTAL Class Descriptions Sample AA I. AGRICULTURAL LAND II. ACTIVE Maize Young 1 0 1 Cowpea Cowpea & Shrubbery 2 0 2 Cotton Cotton 0 1 1 II. FALLOW Bare Earth Dark Soil 2 0 2 Dark Soil w/ Straw 0 1 1 Dark Soil w/ Straw & Ridging 0 1 1 Light Soil 1 0 1 Light Soil w/ Straw 1 1 2 Light Soil w/ Straw & Ridging 8 4 12 Light Soil w/ Ridging 2 5 7 Medium Soil w/ Ridging & Emergent Weeds 3 0 3 Charred Ground Charred Ground on Bare Earth 1 1 2 Dried Fields Vertical Rice Straw w/ Emergent Weeds 1 4 5 Horizontal Rice Straw w/ Emergent Weeds 0 1 1 Dried Grass 1 0 1 Long Term Fallow Medium Soil w/ Ridging & Dried Vegetation 1 1 2 Shrubbery Dense 2 3 5 291 Bwanje Valley Dry S eason Classification Photos I. AGRICULTURAL LAND II. ACTIVE Maize Young Sample Photos: 3_31 Accuracy Assessment Photos: None I. AGRICULTURAL LAND II. ACTIVE Cowpea Cowpea & Shrubbery Sample Photos: 3_43 Other_3 Accuracy Assessment Photos: None 292 I. AGRICULTURAL LAND II. ACTIVE Cotton Cotton Sample Photos: None Accuracy Assessment Photos: SA_AA_35 I. AGRICULTURAL LAND II. FALLOW Bare Earth Dark Soil Sample Photos: 2_87 4_34 Accuracy Assessment Photos: None 293 I. AGRICULTURAL LAND II. FALLOW Bare Earth Dark Soil w Straw Sample Photos: None Accuracy Assessment Photos: SA_AA_137 I. AGRICULTURAL LAND II. FALLOW Bare Earth Dar k Soil w Straw & Ridging Sample Photos: None Accuracy Assessment Photos: SA_AA_114 294 I. AGRICULTURAL LAND II. FALLOW Bare Earth Light Soil Sample Photos: 1_18 Accuracy Assessment Photos: None I. AGRICULTURAL LAND II . FALLOW Bare Earth Light Soil w/ Straw Sample Photos: 1_17 Accuracy Assessment Photos: SA_AA_23 295 I. AGRICULTURAL LAND II. FALLOW Bare Earth Light Soil w Straw & Ridging Sample Photos: 1_34 1_18a 1_56 1_82 3_64 1_64 2_22 1_22 Accuracy Assessment Photos: SA_AA_41 SA_AA_48 SA_AA_49 296 SA_AA_111 I. AGRICULTURAL LAND II. FALLOW Bare Earth Light Soil w Ridging Sample Photos: 2_82 1_32 Accuracy Assessment Photos: SA_AA_88 SA_AA_89 SA_AA_108 SA_AA_112 SA_AA_6 297 I. AGRICULTURAL LAND II. FALLOW Bare Earth Medium Soil w Ridging & Emergent Weeds Sample Photos: 1_58 1_81 2_32 Accuracy Assessment Photos: None I. AGRICULT URAL LAND II. FALLOW Charred Ground Charred Ground on Bare Earth Sample Photos: 1_100 Accuracy Assessment Photos: SA_AA_69 298 I. AGRICULTURAL LAND II. FALLOW Dried Fields Vertical Rice Straw w Emergent Weeds Sample Phot os: 1_37 Accuracy Assessment Photos: SA_AA_17 SA_AA_158 SA_AA_18 SA_AA_135 299 I. AGRICULTURAL LAND II. FALLOW Dried Fields Horizontal Rice Straw w Emergent Weeds Sample Photos: None Accuracy Assessment Photos: SA_AA_157 I. AGRICULTURAL LAND II. FALLOW Dried Fields Dried Grass Sample Photos: 1_20 Accuracy Assessment Photos: None 300 I. AGRICULTURAL LAND II. FALLOW Long Term Fallow Medium Soil w Ridging & Dried Vegetation Sample Photos: 1_3 Accuracy Assessment Photos: SA_AA_110 301 I. AGRICULTURAL LAND II. FALLOW Shrubbery Shrubbery: Dense Sample Photos: 2_56 3_56 Accuracy Assessment Photos: SA_AA_162 SA_AA_163 SA_AA_164 I. AGRICUL TURAL LAND IV. TREES Green Foliage on Bare Soil Sample Photos: 1_33 Accuracy Assessment Photos: None 302 I. AGRICULTURAL LAND IV. TREES Green Foliage on Dry Harvested Agricultural Land w Ridging Sample Photos: 1_27 1_87 2_64 Accuracy Assessment Photos: SA_AA_50 SA_AA_118 303 II. NON - AGRICUTURAL LAND Bare Earth Dark Soil w Straw Sample Photos: 1_61 2_14 Accuracy Assessment Photos: SA_AA_61 304 II. NON - AGRICUTURAL LAND Bare Earth Light Soil Sample Photos: 1_93 2_95 2_100 1_49 1_74 Accuracy Assessment Photos: SA_AA_103 305 II. NON - AGRICUTURAL LAND Bare Earth Light Soil w Green Vegetation Sample Photos: 1_36 3_79 2_61 Accuracy Assessment Photos : SA_AA_51 SA_AA_52 SA_AA_119 SA_AA_120 306 II. NON - AGRICUTURAL LAND Charred Ground Charred Ground & Dry Vegetation Sample Photos: 3_34 3_77 1_31 Accuracy Assessment Photos: SA_AA_30 307 II. NON - AGRICUTURAL LAND Dried Grass Dried Grass Sample Photos: None Accuracy Assessment Photos: SA_AA_60 SA_AA_78 SA_AA_73 SA_AA_74 SA_AA_75 308 II. NON - AGRICUTURAL LAND Trees Green Foliage & Dry Undergrowth Sample Photos: 2_20 1_2 4 Accuracy Assessment Photos: SA_AA_2 309 II. NON - AGRICUTURAL LAND Trees Green Foliage on Bare Soil Sample Photos: 1_14 2_36 4_91 Accuracy Assessment Photos: SA_AA_5 310 II. NON - AGRICUTURAL LAND Trees Little Green Foli age w Dry Undergrowth Sample Photos: 2_24 1_8 2_77 1_35 Accuracy Assessment Photos: SA_AA_9 SA_AA_10 SA_AA_11 SA_AA_28 SA_AA_29 SA_AA_38 311 SA_AA_39 II. NON - AGRICUTURAL LAND Trees Mostly Dried Leaves Sample Photos: None Accuracy Assessment Photos: SA_AA_37 SA_AA_40 SA_AA_64 II. NON - AGRICUTURAL LAND Trees Dried Trees on Bare Rock Sample Photos: 2_31 Accuracy Assessment Photos: None 312 II. NON - AGRICUTURAL LAND Human I nfluence Road Sample Photos: 2_34 Accuracy Assessment Photos: II. NON - AGRICUTURAL LAND Human Influence Human Dwellings Sample Photos: 2_3 Accuracy Assessment Photos: SA_AA_36 313 II. NON - AGRICULTURAL LAND IV. TREES Riverbed Sample Photos: None Accuracy Assessment Photos: SA_AA_92 SA_AA_155 SA_AA_156 314 Bwanje Valley Dry Season Data ID Name X Y Elev Date_ Time Pnt_type Pnt_Acq Offset_ Dist Land_Use_ ANA Land_ Cover Land_Use Notes 56 3_4 3 34.56 - 14.27 498.6 8/25/201 6 13:50 Sample On 0 Agriculture Cowpea & Shrubbery Agricultural Land Inside the scheme within a cowpea field with weeds. Harvested rice field in between cow pea field and a maize field to the south. Cowpea fields to the E, N & W. 26 2_87 34.615 - 14.222 462.6 8/23/201 6 13:08 Sample On 0 Agriculture Bare Earth: Dark Soil Agricultural Land Charred former cropland under the canopy of trees Same to the ENW & South. Open dry land. 6 4_34 34.561 - 14.292 503.5 8/22/201 6 11:57 Sa mple On 0 Agriculture Bare Earth: Dark Soil Agricultural Land Open dirt field. Appears to at one time been used for agriculture (maize?) 52 1_61 34.657 - 14.26 468.7 8/25/201 6 10:15 Sample On 0 Non - Agriculture Bare Earth: Dark Soil w Straw Unsure Tree s in the middle of a mostly open landscape dotted with sporadic trees and dried grasses. 17 2_14 34.675 - 14.286 461.7 8/22/201 6 16:58 Sample On 0 Non - Agriculture Bare Earth: Dark soil w Straw Unsure 35 2_100 34.598 - 14.195 465.7 8/23/201 6 16:09 Samp le On 0 Non - Agriculture Bare Earth: Light soil Unsure Open, barren land covered in dried grass. A small village to the west and southeast. 28 2_95 34.611 - 14.208 465.1 8/23/201 6 13:59 Sample On 0 Non - Agriculture Bare Earth: Light soil Unsure Dry gras sland with a few sporadic trees. The soil is dark grey. 51 1_49 34.661 - 14.273 471.5 8/25/201 6 10:02 Sample On 0 Non - Agriculture Bare Earth: Light soil Unsure Bare land with sporadic palm trees and baobab trees. Houses located to the immediate south 33 1_93 34.582 - 14.204 471.5 8/23/201 6 15:38 Sample On 0 Non - Agriculture Bare Earth: Light soil Unsure Dry, barren area. A few homes to the 315 south and a few sporadic trees. Otherwise a traditional looking savannah landscape. 15 1_18 34.662 - 14.309 486 .4 8/22/201 6 16:03 Sample On 0 Agriculture Bare Earth: Light Soil Agricultural Land Open field. Soils are more red and sandy. Adjacent to a village. Covered in tall, dried reeds. Evidence of some type of cropping here. Maize (?) 46 1_74 34.515 - 14.238 537.0 8/24/201 6 14:23 Sample On 0 Non - Agriculture Bare Earth: Light Soil Built Up Torn down house surrounded by open land with sporadic tree cover. Soccer field to the immediate south. 1 1_32 34.537 - 14.297 519.0 8/24/201 6 13:21 Sample On 0 Agricul ture Bare Earth: Light soil w ridging Agricultural Land Dry, harvested maize fields. Homes to the immediate north. Fields in the direction of the road and homes to the west. 10 1_36 34.586 - 14.29 486.7 8/22/201 6 13:36 Sample On 0 Non - Agriculture Bar e Earth: Light Soil w Green Vegetation Unsure 23 3_79 34.619 - 14.228 465.1 8/23/201 6 12:09 Sample On 0 Non - Agriculture Bare Earth: Light Soil w Green Vegetation Unsure 53 2_61 34.655 - 14.253 465.7 8/25/201 6 10:30 Sample On 0 Non - Agriculture Bar e Earth: Light Soil w Green Vegetation Infrastructure Inside a dried river bed surrounded by open land with sporadic tree and grasses. This is a diversion weir for another rice scheme further down. Namboona River. 316 48 2_82 34.523 - 14.222 506.2 8/24/201 6 14:59 Sample On 0 Agriculture Bare Earth: Light soil w ridging Agricultural Land Dry, harvested field left fallow. Sporadic trees but mainly dry fields and a few houses. 14 1_17 34.636 - 14.313 484.9 8/22/201 6 15:30 Sample On 0 Agriculture Bare Eart h: Light soil w Straw Agricultural Land 24 1_18a 34.625 - 14.23 462.3 8/23/201 6 12:23 Sample On 0 Agriculture Bare Earth: Light soil w straw & ridging Agricultural Land 2 1_34 34.562 - 14.289 503.5 8/22/201 6 11:33 Sample On 0 Agriculture Bare Ea rth: Light soil w straw & ridging Agricultural Land Dry, harvested maize field. To the north is a village. To the south is open land. Grey - Brown soil. 57 1_56 34.561 - 14.267 497.4 8/25/201 6 14:06 Sample On 0 Agriculture Bare Earth: Light soil w straw & ridging Agricultural Land Harvested field of maize but outside o the scheme. Banana and sugarcane on the SE. Nsangu trees on western side and a main farm road on the north. 43 1_64 34.512 - 14.256 551.3 8/24/201 6 13:51 Sample On 0 Agriculture Bare Ea rth: Light soil w straw & ridging Agricultural Land Drainage canal surrounded by dry maize fields that have been harvested and left fallow. 47 1_82 34.52 - 14.217 497.7 8/24/201 6 14:41 Sample On 0 Agriculture Bare Earth: Light soil w straw & ridging Ag ricultural Land Dry field. Irrigated maize field closer to the road. Village to the east. 45 3_64 34.517 - 14.25 535.2 8/24/201 6 14:11 Sample On 0 Agriculture Bare Earth: Light soil w straw & ridging Agricultural Land Dry, harvested maize field immedi ately adjacent to homes. Sporadic trees to the east and some to the west. 317 50 2_22 34.537 - 14.306 531.2 8/24/201 6 15:43 Sample On 0 Agriculture Bare Earth: Light Soil w Straw & Ridging Agricultural Land Dry, harvested field. Sporadic trees in all dir ections. 49 1_22 34.535 - 14.304 528.8 8/24/201 6 15:34 Sample On 0 Agriculture Bare Earth: Light Soil w Straw & Ridging Agriculture Dry, harvested land covered in grasses with sporadic trees dotting the landscape 60 1_81 34.639 - 14.235 464.2 8/26/201 6 9:31 Sample On 0 Agriculture Bare Earth: Medium Soil w Ridging & Emergent Weeds Agriculture Appearance of previous harvested rice field. Now this area is left fallow and covered in dried grasses. 61 2_32 34.536 - 14.288 519.6 8/26/201 6 10:26 Sample O n 0 Agriculture Bare Earth: Medium Soil w Ridging & Emergent Weeds Agriculture Dry, harvested maize field. Tree immediately to the E. Sporadic trees dot the landscape where most maize fields have been left fallow and are covered in weeds and tall grasses 5 3_34 34.564 - 14.293 501.7 8/22/201 6 11:51 Sample On 0 Non - Agriculture Charred Ground & Dry Vegetation Unsure Charred ground. Dark grey soil. Surrounded by random grasses and trees 31 3_77 34.573 - 14.232 483.1 8/23/201 6 14:54 Sample On 0 Non - Agricul ture Charred Ground & Dry Vegetation Unsure Charred ground surrounded by grasses that are dry. Outside of a small tree grove. To the north are more grasses and larger trees. 37 1_31 34.514 - 14.286 557.1 8/24/201 6 10:02 Sample On 0 Non - Agriculture Cha rred Ground & Dry Vegetation Unsure 34 1_100 34.602 - 14.191 459.9 8/23/201 6 15:54 Sample On 0 Agriculture Charred Ground on Bare Earth Agricultural Land Charred ground amidst open fields. A few palms in the 318 area. A few homes nearby just off the lak e. 55 Other_3 34.562 - 14.27 495.6 8/25/201 6 13:38 Sample On 0 Agriculture Cowpea & Shrubbery Agricultural Land 39 1_20 34.496 - 14.295 615.0 8/24/201 6 10:33 Sample On 0 Agriculture Dried Grass Agriculture Small tree adjacent to a channel for wate r. Maize plots with nothing planted located here. Tall, dry grasses dominate. 38 2_31 34.5131 - 14.286 7 555.6 8/24/201 6 10:09 Sample On 0 Non - Agriculture Dried Trees on Bare Rock Unsure Bare rock along a cliff face. Former perineal river. 22 1_58 34. 601 - 14.259 482.8 8/23/201 6 11:00 Sample On 0 Agriculture Bare Earth: Medium Soil w Ridging & Emergent Weeds Agriculture Harvested maize field. Spiny grasses. Weeds to the south and west. To the E, low grasses. N - Line of trees, then the scheme. Right a t this point are two small banana trees. 20 2_3 34.561 - 14.337 525.1 8/23/201 6 9:30 Sample On 0 Non - Agriculture Human Dwellings Built Up Housing compound. Point is at the outhouse structure. Made of dried grasses adjacent to cotton fields. In a village with grasses to the north in the direction of the scheme. 42 3_31 34.5251 - 14.292 516.3 8/24/201 6 13:32 Sample On 0 Agriculture Maize: Young Agricultural Land Irrigated maize field. Point taken from a small road between fields. To the south are house s. A few large trees dot the scheme. All maize is green. 18 1_3 34.558 - 14.339 532.7 8/23/201 6 9:17 Sample On 0 Agriculture Medium Soil w Ridging & Agriculture Bare maize field overgrown with 319 Dried Vegetation weeds adjacent to the railway lines. 3 2 _34 34.566 - 14.289 493.1 8/22/201 6 11:40 Sample On 0 Non - Agriculture Road Infrastructure Dirt road through to a village surrounded by dead, drying vegetation. 58 2_56 34.571 - 14.263 489.5 8/25/201 6 14:29 Sample On 0 Agriculture Shrubbery: Dense Agric ultural Land Inside the scheme. Cleared field covered with red and green shrubs. Barren field as well on the N, E, & W but green maize fields to the south. 59 3_56 34.587 - 14.255 484.3 8/25/201 6 15:03 Sample On 0 Agriculture Shrubbery: Dense Agricult ural Land Inside the scheme. Harvested rice field surrounded by harvested rice fields with some grass and dry vegetation. There is a green maize plot almost 50m on the eastern side. 8 1_24 34.579 - 14.296 491.6 8/22/201 6 12:33 Sample On 0 Non - Agricultur e Trees: Green Foliage & Dry Undergrowt h Unsure Small cluster of trees surrounded by open, bare land 40 2_20 34.505 - 14.31 629.1 8/24/201 6 10:58 Sample On 0 Non - Agriculture Trees: Green Foliage & Dry Undergrowt h Unsure Thick tree and vegetative growt h. Tall grass undergrowth. Adjacent to a road. House located to the south. 11 2_36 34.592 - 14.288 481.2 8/22/201 6 13:50 Sample On 0 Non - Agriculture Trees: Green Foliage on Bare Soil Unsure 36 4_91 34.553 - 14.20 491.0 8/23/201 6 16:26 Sample On 0 No n - Agriculture Trees: Green Foliage on Bare Soil Built Up Taller clustering of grasses and low trees. To the E is a grouping of homes. This point is located within a village. 320 Some of the homes have tin roofs. 0 1_33 34.55 - 14.292 515.7 8/22/201 6 11:18 S ample On 0 Agriculture Trees: Green Foliage on Bare Soil Agriculture 16 1_14 34.664 - 14.296 478.5 8/22/201 6 16:28 Sample On 0 Non - Agriculture Trees: Green Foliage on Bare Soil Unsure 13 1_27 34.624 - 14.308 479.4 8/22/201 6 15:03 Sample On 0 Agr iculture Trees: Green Foliage on Dry Harvested Agricultural Land w Ridging Agriculture Tall tree cover amidst a field of changing brush to barren lots surrounded by tall trees. We have not seen these types of trees elsewhere. Brush is almost to our waist s. It looks like a prairie. 27 1_87 34.61 - 14.223 460.2 8/23/201 6 13:25 Sample On 0 Agriculture Trees: Green Foliage on Dry Harvested Agricultural Land w Ridging Agriculture Grey, dry ground surrounded by green trees and small shrubs. Dead leaves and d ry grasses cover the floor. 44 2_64 34.514 - 14.251 542.5 8/24/201 6 14:02 Sample On 0 Agriculture Trees: Green Foliage on Dry Harvested Agricultural Land w Ridging Agriculture Mango tree sitting within a dry harvested maize field. 7 1_35 34.581 - 14.2 92 487.9 8/22/201 6 12:24 Sample On 0 Non - Agriculture Trees: Little foliage w Dry Undergrowt h Unsure Open, dry area covered in sticks. Houses to the E 21 1_8 34.549 - 14.327 534.9 8/23/201 6 9:42 Sample On 0 Non - Agriculture Trees: Little foliage w Dry Un dergrowt h Unsure Small grouping of trees with dried orange leaves. Adjacent to powerlines with dried grasses below. 321 9 2_24 34.572 - 14.298 495.3 8/22/201 6 12:48 Sample On 0 Non - Agriculture Trees: Little foliage w Dry Undergrowt h Unsure Clustering of sma ll trees surrounded by dry grasses. 30 2_77 34.575 - 14.233 476.0 8/23/201 6 14:47 Sample On 0 Non - Agriculture Trees: Little foliage w Dry Undergrowt h Unsure Charred ground surrounded by grasses that are dried out. We're outside of a small tree grove now . To the north are more grasses and larger trees. 12 1_37 34.612 - 14.284 474.5 8/22/201 6 14:14 Sample On 0 Agriculture Vertical Rice Straw w Emergent Weeds Agricultural Land Harvested dry rice field adjacent to maize planted to the east. 4 SA_AA_10 34.5790 - 14.296 453 2016:08:22 Accuracy On 0 Non - Agriculture Trees: Little Green Foliage with Dry Undergrowt h Unsure 37 SA_AA_103 34.5169 - 14.250 466 2016:08:24 Accuracy Offset 5 Non - Agriculture Bare Earth: Light Soil Built up 38 SA_AA_108 34.522 7 - 14.222 441 2016:08:24 Accuracy On 0 Agriculture Bare Earth: Light soil w Ridging Agriculture 5 SA_AA_11 34.576 - 14.296 459 2016:08:22 Accuracy On 0 Non - Agriculture Trees: Little Green Foliage with Dry Undergrowt h Unsure 39 SA_AA_110 34.5350 - 14 .304 468 2016:08:24 Accuracy On 0 Agriculture Medium Soil w Ridging & Dried Vegetation 40 SA_AA_111 34.5350 - 14.304 468 2016:08:24 Accuracy On 0 Agriculture Bare Earth: Light Soil w Straw & Ridging Agriculture Trees with green foliage present acros s the area. Spotty. 322 41 SA_AA_112 34.5350 - 14.304 466 2016:08:24 Accuracy On 0 Agriculture Bare Earth: Light soil w Ridging Agriculture One tree located here. No foliage. 42 SA_AA_114 34.5350 - 14.304 466 2016:08:24 Accuracy On 0 Agriculture Bare Earth: Dark soil w straw & ridging Agriculture 46 SA_AA_118 34.5343 - 14.306 473 2016:08:24 Accuracy On 0 Agriculture Trees: Green Foliage on Dry Harvested Agricultural Land w Ridging Agriculture There is a pathway here flanked by trees. Otherwise trees are o nly sporadic. 47 SA_AA_119 34.5343 - 14.306 396 2016:08:24 Accuracy On 0 Non - Agriculture Light Soil w Green Vegetation Unsure 48 SA_AA_120 34.5343 - 14.306 396 2016:08:24 Accuracy On 0 Non - Agriculture Light Soil w Green Vegetation Unsure 52 SA_AA _135 34.6616 - 14.249 380 2016:08:25 Accuracy On 0 Agriculture Vertical Rice Straw w Emergent Weeds Agriculture Nambuona Irrigation Scheme: 200ha 54 SA_AA_137 34.6616 - 14.249 378 2016:08:25 Accuracy On 0 Agriculture Bare Earth: Dark Soil w Straw Agricu lture Active cultivation going on nearby. 55 SA_AA_155 34.5452 - 14.279 394 2016:08:25 Accuracy On 0 Non - Agriculture Riverbed Unsure This is the end of available water from Namikokwe 56 SA_AA_156 34.5452 - 14.279 394 2016:08:25 Accuracy On 0 Non - Agricult ure Riverbed Unsure 57 SA_AA_157 34.5452 - 14.279 395 2016:08:25 Accuracy On 0 Agriculture Horizontal Rice Straw w Emergent Weeds Agriculture 58 SA_AA_158 34.5452 - 14.279 395 2016:08:25 Accuracy On 0 Agriculture Vertical Rice Straw w Emergent Weeds Agricultures 60 SA_AA_162 34.536 - 14.288 426 2016:08:26 Accuracy On 0 Agriculture Shrubbery: Dense Agriculture With ridging 61 SA_AA_163 34.5360 - 14.288 426 2016:08:26 Accuracy On 0 Agriculture Shrubbery: Dense Agriculture Without ridging 323 62 SA_AA_ 164 34.5360 - 14.288 427 2016:08:26 Accuracy On 0 Agriculture Shrubbery: Dense Agriculture Ridging. Fallow field. Trees with green foliage present. 6 SA_AA_17 34.6119 - 14.284 450 2016:08:22 Accuracy On 0 Agriculture Vertical Rice Straw w Emergent Weeds Agriculture Charred circle in the center of the plot. Maize planted nearby 7 SA_AA_18 34.612 - 14.284 451 2016:08:22 Accuracy On 0 Agriculture Vertical Rice Straw w Emergent Weeds Agriculture 1 SA_AA_2 34.5162 - 14.275 487 2016:08:20 Accuracy On 0 Non - Agriculture Trees: Green Foliage & Dry Undergrowt h Unsure 9 SA_AA_23 34.6242 - 14.307 462 2016:08:22 Accuracy On 0 Agriculture Bare Earth: Light Soil w Straw & Ridging Agriculture Tree with green foliage about 75ft away 10 SA_AA_28 34.6639 - 14.296 46 1 2016:08:22 Accuracy On 0 Non - Agriculture Trees: Little Green Foliage with Dry Undergrowt h Unsure 11 SA_AA_29 34.6639 - 14.296 460 2016:08:22 Accuracy On 0 Non - Agriculture Trees: Little Green Foliage with Dry Undergrowt h Unsure 12 SA_AA_30 34.6701 - 14.284 451 2016:08:22 Accuracy On 0 Non - Agriculture Charred Ground & Dry Vegetation Unsure 13 SA_AA_35 34.5612 - 14.337 466 2016:08:23 Accuracy On 0 Agriculture Cotton Agriculture 14 SA_AA_36 34.5612 - 14.337 466 2016:08:23 Accuracy On 0 Non - Agric ulture Human Dwellings Built up 15 SA_AA_37 34.5611 - 14.337 475 2016:08:23 Accuracy On 0 Non - Agriculture Trees: Mostly Dried Leave Unsure 16 SA_AA_38 34.5489 - 14.327 474 2016:08:23 Accuracy On 0 Non - Agriculture Trees: Little Green Unsure 324 Foliage with Dry Undergrowt h 17 SA_AA_39 34.5489 - 14.327 473 2016:08:23 Accuracy On 0 Non - Agriculture Trees: Little Green Foliage with Dry Undergrowt h Unsure 18 SA_AA_40 34.5489 - 14.327 475 2016:08:23 Accuracy On 0 Non - Agriculture Trees: Mostly Dried Leave Unsure 19 SA_AA_41 34.5489 - 14.327 423 2016:08:23 Accuracy On 0 Agriculture Bare Earth: Light Soil w Straw & Ridging Agriculture 20 SA_AA_48 34.625 - 14.23 414 2016:08:23 Accuracy On 0 Agriculture Bare Earth: Light soil w Straw & Ridging Agriculture 21 SA_AA_49 34.6250 - 14.23 414 2016:08:23 Accuracy On 0 Agriculture Bare Earth: Light soil w Straw & Ridging Agriculture Fallow field. Green wild vegetation growing in plots nearby. 2 SA_AA_5 34.5639 - 14.293 459 2016:08:22 Accuracy On 0 Non - Agricult ure Trees: Green Foliage on Bare Soil Unsure 22 SA_AA_50 34.6250 - 14.23 414 2016:08:23 Accuracy On 0 Agriculture Trees: Green Foliage on Dry Harvested Agricutural Land w Ridging Agriculture 23 SA_AA_51 34.6151 - 14.222 406 2016:08:23 Accuracy On 0 Non - Agriculture Light Soil w Green Vegetation Unsure 24 SA_AA_52 34.6150 - 14.222 407 2016:08:23 Accuracy On 0 Non - Agriculture Light Soil w Green Vegetation Unsure Trees with green foliage 0 SA_AA_6 34.5617 - 14.292 464 2016:08:22 Accuracy On 0 Agri culture Bare Earth: Light soil w Ridging Agriculture Large mound of dirt in the distance. The ridging is not very 325 high here, but is evident. 25 SA_AA_60 34.6238 - 14.217 412 2016:08:23 Accuracy On 0 Non - Agriculture Dried Grass Unsure 26 SA_AA_61 34.6 236 - 14.216 412 2016:08:23 Accuracy On 0 Non - Agriculture Bare Earth: Dark Soil w Straw Unsure 27 SA_AA_64 34.5750 - 14.233 435 2016:08:23 Accuracy On 0 Non - Agriculture Trees: Mostly Dried Leave Unsure 28 SA_AA_69 34.602 - 14.191 417 2016:08:23 Accu racy On 0 Agriculture Charred Ground on Bare Earth Agriculture 29 SA_AA_73 34.6037 - 14.191 417 2016:08:23 Accuracy On 0 Non - Agriculture Dried Grass Agriculture 30 SA_AA_74 34.6038 - 14.191 418 2016:08:23 Accuracy On 0 Non - Agriculture Dried Grass Agriculture 31 SA_AA_75 34.6038 - 14.191 417 2016:08:23 Accuracy On 0 Non - Agriculture Dried Grass Unsure 32 SA_AA_78 34.5980 - 14.194 421 2016:08:23 Accuracy On 0 Non - Agriculture Dried Grass Unsure 33 SA_AA_88 34.5371 - 14.296 443 2016:08:24 Acc uracy On 0 Agriculture Bare Earth: Light soil w Ridging Agriculture 34 SA_AA_89 34.5371 - 14.296 445 2016:08:24 Accuracy On 0 Agriculture Bare Earth: Light soil w Ridging Agriculture Large dirt mound located here in the center of the field 3 SA_AA_9 34.5760 - 14.29 450 2016:08:22 Accuracy On 0 Non - Agriculture Trees: Little Green Foliage with Dry Undergrowt h Unsure Lots of sticks mounded up here for later gathering (?) 36 SA_AA_92 34.52273 - 14.291 449 2016:08:24 Accuracy On 0 Non - Agriculture Riverbed Unsure Point taken from bridge 326 Bwanje Valley Dry Season Analysis Classes Class Name Samples in Class Land Cover 29 4 Fallow Agriculture 32 2 Fallow Agriculture 33 2 Fallow Agriculture 40 3 Fallow Agriculture 41 6 Fallow Agriculture 50 10 Fallo w Agriculture 56 6 Mixed Foliage 59 7 Fallow Agriculture 63 1 Mixed Forest 69 3 Mixed Foliage 70 5 Mixed Forest 75 6 Mixed Foliage 77 6 Bare Earth 80 4 Mixed Foliage 85 2 Bare Earth 90 29 Mixed Foliage Class 29: Fallow Agriculture NAME LAND USE CLASS SUBCLASS LAND COVER 2_95 Fallow Bare Earth Light Soil 1_61 Fallow Bare Earth Dark Soil w Straw 2_32 Fallow Bare Earth Medium Soil w Ridging & Emergent Reeds 3_31 Active Maize Young Class 32: Fallow Agriculture NAME LAND USE CLASS SUBC LASS LAND COVER 1_100 Fallow Charred Ground Charred Ground on Bare Earth 1_87 Trees Trees Trees: Green Foliage on Dry Harvested Agricultural Land w Ridging 1_27 Trees Trees Trees: Green Foliage on Dry Harvested Agricultural Land w Ridging SA_AA_6 9 AG Fallow Charred Ground Charred Ground on Bare Earth Class 33: Fallow Agriculture NAME LAND USE CLASS SUBCLASS LAND COVER SA_AA_164 Ag Fallow Shrubbery Dense SA_AA_162 Ag Fallow Shrubbery Dense 327 Class 40: Fallow Agriculture NAME LAND USE CLA SS SUBCLASS LAND COVER SA_AA_35 Ag Active Cotton Cotton SA_AA_36 Non Ag Human Inf Human Inf Human Dwellings SA_AA_37 Non Ag Trees Trees Mostly Dried Leaves Class 41: Fallow Agriculture NAME LAND USE CLASS SUBCLASS LAND COVER 1_81 Fallow Bare Eart h Medium Soil w Riding & Emergent Weeds 3_64 Fallow Bare Earth Light Soil w Straw & Ridging 1_37 Fallow Long Term Fallow Medium Soil w Ridging & Dried Vegetation SA_AA_74 Ag Fallow Dried Fields Dried Grass SA_AA_17 Ag Fallow Dried Fields Vertical Rice Straw w Emergent Weeds SA_AA_18 Ag Fallow Dried Fields Vertical Rice Straw w Emergent Weeds Class 50: Fallow Agriculture NAME LAND USE CLASS SUBCLASS LAND COVER 4_34 Fallow Bare Earth Dark Soil 2_3 Non - Ag Human Inf Human Dwellings 1_3 Fal low Long Term Fallow Medium Soil w Ridging & Dried Vegetation SA_AA_108 Ag Fallow Bare Earth Light Soil w Ridging SA_AA_137 Ag Fallow Bare Earth Dark Soil w Straw SA_AA_135 Ag Fallow Dred Fields Vertical Rice Straw w Emergent Weeds SA_AA_103 Non Ag Bar e Earth Bare Earth Light Soil SA_AA_9 Non Ag Trees Trees Little Green Foliage w Dry Undergrowth Sa_AA_89 Ag Fallow Bare Earth Light Soil w Ridging Sa_AA_11 Non Ag Trees Trees Little Green Foliage w Dry Undergrowth Class 56: Mixed Foliage NAME LAND U SE CLASS SUBCLASS LAND COVER 3_77 Non - Ag Charred Ground Charred Ground Charred Ground w/ Dry Vegetation 2_61 Non - Ag Bare Earth Bare Earth Light Soil w/ Green Vegetation 2_24 Non - Ag Trees Trees Little Foliage w Dry Undergrowth 1_14 Non - Ag Trees Trees Green Foliage on Bare Soil SA_AA_64 Non - Ag Trees Trees Mostly Dried Leaves SA_AA_28 Non - Ag Trees Trees Little Green Foliage w Dry Undergrowth 328 Class 59: Fallow Agriculture NAME LAND USE CLASS SUBCLASS LAND COVER SA_AA_112 Ag Fallow Bare Earth Light Soil w Ridging SA_AA_111 Ag Fallow Bare Earth Light Soil w Straw & Ridging SA_AA_110 Ag Fallow Bare Earth Medium Soil w Ridging & Dried Vegetation SA_AA_114 Ag Fallow Bare Earth Dark soil w Straw & Ridging 2_82 Ag Fallow Bare Earth Light Soil w Ridgi ng 1_58 Ag Fallow Bare Earth Medium Soil w Ridging & Emergent Weeds 2_22 Ag Fallow Bare Earth Light Soil w Straw & Ridging Class 63: Mixed Forest NAME LAND USE CLASS SUBCLASS LAND COVER 1_31 Non - Ag Charred Ground Charred Ground Charred Ground & Dry Vegetation Class 69: Mixed Foliage NAME LAND USE CLASS SUBCLASS LAND COVER 2_64 Ag Trees Trees Green Foliage on Dry Harvested Agricultural Land w/ Ridging 1_20 Ag Fallow Dried Fields Dried Grass 1_22 Ag Fallow Bare Earth Light Soil w Straw & Ridg ing Class 70: Mixed Forest NAME LAND USE CLASS SUBCLASS LAND COVER 3_34 Non - Ag Charred Ground Charred Ground Charred Ground & Dry Vegetation SA_AA_5 Non - Ag Trees Trees Green Foliage on Bare Soil SA_AA_39 Non - Ag Trees Trees Little Green Foliage w D ry Undergrowth SA_AA_40 Non - Ag Trees Trees Mostly Dried Leaves SA_AA_41 Ag Fallow Bare Earth Light Soil w Straw & Ridging Class 75: Mixed Foliage NAME LAND USE CLASS SUBCLASS LANDCOVER 2_31 Non - Ag Trees Trees Dried trees on bare rock 1_35 Non - Ag Trees Trees Little Foliage w Dry Undergrowth 2_36 Non - Ag Trees Trees Green Foliage on Bare Soil 1_24 Non - Ag Trees Trees Green Foliage & Dry Undergrowth SA_AA_10 Non - Ag Trees Trees Little Green Foliage w Dry Undergrowth SA_AA_29 Non - Ag Trees Trees L ittle Green Foliage w Dry Undergrowth 329 Class 77: Bare Earth NAME LAND USE CLASS SUBCLASS LANDCOVER 2_100 Non - Ag Bare Earth Bare Earth Light Soil 2_87 Ag Fallow Bare Earth Dark Soil SA_AA_78 Non - Ag Dried Grass Dried Grass Dried Grass SA_AA_60 N on - Ag Dried Grass Dried Grass Dried Grass SA_AA_61 Non - Ag Bare Earth Bare Earth Dark Soil w Straw SA_AA_30 Non - Ag Charred Ground Charred Ground Charred Ground & Dry Vegetation Class 80: Mixed Foliage NAME LAND USE CLASS SUBCLASS LAND COVER 2_34 Non - Ag Human Inf Human Inf Road 1_33 Ag Trees Trees Green Foliage on Bare Soil 1_8 Non - Ag Trees Trees Little Foliage w Dry Undergrowth SA_AA_38 Non - Ag Trees Trees Little Foliage w Dry Undergrowth Class 85: Bare Earth NAME LAND USE CLASS SUBCLASS LANDCO VER 3_79 Non - Ag Human Inf Human Inf Road SA_AA_52 Non - Ag Bare Earth Bare Earth Little Soil w Green Vegetation 330 Class 90: Mixed Foliage NAME LAND USE CLASS SUBCLASS LANDCOVER 1_93 Non - Ag Bare Earth Bare Earth Light Soil 4_91 Non - Ag Trees Trees Green Foliage on Bare Soil 1_82 Ag Fallow Bare Earth Light Soil w Straw & Ridging 1_18a Ag Fallow Bare Earth Light Soil w Straw & Ridging 2_77 Non - Ag Trees Trees Little Foliage w Dry Undergrowth 1_74 Non - Ag Bare Earth Bare Earth Light Soil 1_64 Ag Fallow Bare Earth Light Soil w Straw & Ridging 1_49 Non - Ag Bare Earth Bare Earth Light Soil 1_34 Ag Fallow Bare Earth Light Soil w Straw & Ridging 2_14 Non - Ag Bare Earth Bare Earth Dark Soil w Straw 1_36 Non - Ag Bare Earth Bare Earth Light S oil w Green Vegetation 1_32 Ag Fallow Bare Earth Bare Earth w ridging 1_18 Ag Fallow Bare Earth Light Soil 2_20 Non - Ag Trees Trees Green Foliage & Dry undergrowth 1_17 Ag Fallow Bare Earth Light Soil w Straw SA_AA_51 Non - Ag Bare Earth Bare Earth Lit tle Soil w Green Vegetation SA_AA_49 Ag Fallow Bare Earth Light Soil w Straw & Ridging SA_AA_48 Ag Fallow Bare Earth Light Soil w Straw & Ridging SA_AA_50 Ag Trees Trees Green Foliage on Dry Harvested Agricultural Land w Ridging SA_AA_2 Ag Trees Tre es Green Foliage & Dry Undergrowth SA_AA_155 Non Ag Riverbed Riverbed Riverbed SA_AA_157 Ag Fallow Dried Fields Horizontal Rice Straw w Emergent Weeds SA_AA_158 Ag Fallow Dried Fields Vertical Rice Straw w Emergent Weeds SA_AA_6 Ag Fallow Bare Earth L ight Soil w Ridging SA_AA_92 Non - Ag Riverbed Riverbed Riverbed SA_AA_88 Ag Fallow Bare Earth Light Soil w Ridging SA_AA_118 Ag Trees Trees Green Foliage on Dry Harvested Ag Land w Ridging SA_AA_23 Ag Fallow Bare Earth Light Soil w Straw & Ridging SA _AA_119 Non - Ag Bare Earth Bare Earth Light Soil w Green Vegetation 3 31 BIBLIOGRAPHY 332 BIBLIOGRAPHY ADB. (2013). India Agriculture : Better Irrigation Boosts Rice Yields . Retrieved from http s://www.adb.org/results/india - agriculture - better - irrigation - boosts - rice - yields Adhikari, U., & Nejadhashemi, A. P. (2016). Impacts of Climate Change on Water Resources in Malawi. http://doi.org/10.1061/(ASCE)HE.1943 - 5584.0001436 Afrane, Y. A., Klinkenber g, E., Drechsel, P., Owusu - Daaku, K., Garms, R., & Kruppa, T. (2004). Does irrigated urban agriculture influence the transmission of malaria in the city of Kumasi, Ghana? Acta Tropica , 89 , 125 134. http://doi.org/10.1016/j.actatropica.2003.06.001 Afrane, Y. A., Lawson, B. W., Githeko, A. K., & Yan, G. (2005). Effects of Microclimatic Changes Caused by Land Use and Land Cover on Duration of Gonotrophic Cycles of Anopheles gambiae (Diptera: Culicidae) in Western Kenya Highlands . P opulation and Community Ecol ogy J. Med. Entomol (Vol. 42). Retrieved from https://academic.oup.com/jme/article - abstract/42/6/974/886946 AGRA. (2017). Africa Agriculture Status Report: The Business of Smalholder Agriculture in Sub - Saharan Africa (Issue 5) . Nairobi. Retrieved from htt ps://agra.org/wp - content/uploads/2017/09/Final - AASR - 2017 - Aug - 28.pdf AICC. (2016). Malawi Rice Outlook . AICC. (2017). Malawi Rice Outlook . Alimi, T. O., Fuller, D. O., Qualls, W. A., Herrera, S. V, Arevalo - Beier, J. C. (20 15). Predicting potential ranges of primary malaria vectors and malaria in northern South America based on projected changes in climate, land cover and human population. Parasites & Vectors , 8 , 431. http://doi.org/10.1186/s13071 - 015 - 1033 - 9 Almeida, M., Al exandre, A., Gomes Benzecry, S., Siqueira, A. M., Vitor - Alecrim, C. (2015). The Association between Nutritional Status and Malaria in Children from a Rural Community in the Amazonian Region: A Longitudinal Study. http://doi.org/10 .1371/journal.pntd.0003743 AQUASTAT. (2014). Irrigated Crops. Retrieved June 29, 2018, from http://www.fao.org/nr/water/aquastat/infographics/Irrigated_eng.pdf Assi, S. - B., Henry, M. - C., Rogier, C., Dossou - le, P. (2013). Inland valley rice production systems and malaria infection and disease in the forest Malaria Journal , 12 , 1. http://doi.org/10.1186/1475 - 2875 - 12 - 233 Ayala, D., Costantini, C., Ose, K., Kamdem, G. C., Antoni o - Nkondjio, C., Agbor, J. - F. (2009). Malaria Journal Habitat suitability and ecological niche profile of major malaria vectors in Cameroon. http://doi.org/10.1186/1475 - 2875 - 8 - 307 333 Bass, C., Williamson, M. S., Wilding, C. S., Donnelly, M. J., & Field, L. M. (2007). Identification of the main malaria vectors in the Anopheles gambiae species complex using a TaqMan real - time PCR assay. Malaria Journal , 6 , 155. http://doi.org/10.1186/1475 - 2875 - 6 - 155 Bayoh, M. N., & Lindsay, S. W. (2004). Temperatur e - related duration of aquatic stages of the Afrotropical malaria vector mosquito Anopheles gambiae in the laboratory. Medical and Veterinary Entomology , 18 (2), 174 179. http://doi.org/10.1111/j.0269 - 283X.2004.00495.x Beck - Johnson, L. M., Nelson, W. A., Pa aijmans, K. P., Read, A. F., Thomas, M. B., & Bjørnstad, O. N. (2013). The Effect of Temperature on Anopheles Mosquito Population Dynamics and the Potential for Malaria Transmission. PLoS ONE , 8 (11), e79276. http://doi.org/10.1371/journal.pone.0079276 Bec Mosquitoes and their control (Vol. 2). Springer. Retrieved from https://link.springer.com/content/pdf/10.1007/978 - 3 - 540 - 92874 - 4.pdf Beer, S. (1968). Management science. Aldus, London . Binswanger, H. P., & Aiyar, S. S. (2003). Scaling up community - driven development: theoretical underpinnings and program design implications. Blanford, J. I., Blanford, S., Crane, R. G., Mann, M. E., Paaijmans, K. P., Schreiber, K. V, & Th omas, M. B. (2013). Implications of temperature variation for malaria parasite development across Africa. Scientific Reports , 3 (1300). http://doi.org/10.1038/srep01300 Boelee, E., & Madsen, H. (2006). Irrigation and schistosomiasis in Africa: ecological a spects (Vol. 99). IWMI. Boucher, C. (2012). When animals sing and spirits dance: gule wamkulu, the great dance of the Chewa people of Malawi. Briegel, H. (1990). Fecundity, metabolism, and body size in Anopheles (Diptera: Culicidae), vectors of malaria. Journal of Medical Entomology , 27 (5), 839 850. Brouwer, C., Goffeau, A., & Heibloem, M. (1985). Irrigation Water Management: Training Manual No. 1 - Introduction to Irrigation . Retrieved from http://www.fao.org/docrep/R4082E/r4082e00.htm#Contents Brouwer, C., Prins, K., Kay, M., & Heibloem, M. (1988). Irrigation Water Management: Irrigation Methods . Retrieved from http://www.fao.org/docrep/s8684e/s8684e00.htm#Contents Brown, L. A. (1981). Innovation diffusion; a new perspective . Methuen. Brunsdon, C., Fo theringham, A. S., & Charlton, M. E. (1996). Geographically Weighted Regression: A Method for Exploring Spatial Nonstationarity. Geographical Analysis , 28 (4), 281 298. http://doi.org/10.1111/j.1538 - 4632.1996.tb00936.x Burgert, C. R., Colston, J., Roy, T., & Zachary, B. (2013). DHS S patial Analysis Reports 7 Geographic 334 Displacement Procedure and Georeferenced Data Release Policy for the Demographic and Health Surveys . Retrieved from www.measuredhs.com Burney, J. A., & Naylor, R. L. (2012). Smallholder Irri gation as a Poverty Alleviation Tool in Sub - Saharan Africa. http://doi.org/10.1016/j.worlddev.2011.05.007 Carter, R., Mendis, K. N., & Roberts, D. (2000). Spatial targeting of interventions against malaria . Retrieved from https://pdfs.semanticscholar.org/ 78c3/b4654f06eabaadb1e57b8e0d242421bd0201.pdf Caulfield, L. E., Richard, S. A., & Black, R. E. (2004). Undernutrition as an underlying cause of malaria morbidity and mortality in children less than five years old. The American Journal of Tropical Medicine and Hygiene , 71 (2_suppl), 55 63. CDC. (2015). Malaria. Retrieved from https://www.cdc.gov/malaria/about/biology/mosquitoes/index.html CDC. (2017). Anopheles Mosquitoes. Retrieved July 2, 2018, from https://www.cdc.gov/malaria/about/biology/mosquitoes/in dex.html https://www.cdc.gov/malaria/malaria_worldwide/impact.html CDC - Centers for Disease Control and. (2016). Malaria - Biology. Retrieved January 4, 2018, from https://www.cdc.go v/malaria/about/biology/index.html CGIAR. (2015). CGIAR Strategy and Results 2016 - 2030 . CGIAR (Vol. 1). Chester, R. (2005). Achieving scale in agriculture past efforts, present promises . Retrieved from http://pdf.usaid.gov/pdf_docs/Pnadd294.pdf Chidanti - Malunga, J. (2009). When government acts against the will of the people: The case of Bwanje Valley Irrigation Scheme in Malawi. Journal of Applied Sciences Research , 5 (3), 323 328. Retrieved from http://www.scopus.com/inward/record.url?eid=2 - s2.0 - 65449168 367&partnerID=tZOtx3y1 Chinsinga, B. (2017). The Green Belt Initiative, Politics and Sugar Production in Malawi Blessings Chinsinga. Journal of South African Studies , 43 (3), 501 515. http://doi.org/10.1080/03057070.2016.1211401org/10.1080/03057070.2016.12 11401 Chinsinga, B., & Chasukwa, M. (2012). The Green Belt Initiative and Land Grabs in Malawi . Retrieved from http://www.future - agricultures.org/policy - engagement/policy - briefs/1593 - the - green - belt - initiative - and - land - grabs - in - malawi/file Chinsinga, B., & Chasukwa, M. (2012). Youth, Agriculture and Land Grabs in Malawi. IDS Bulletin , 43 (6), 67 77. http://doi.org/10.1111/j.1759 - 5436.2012.00380.x Chirwa, W. C. (2002). Land use and extension services at wovwe rice scheme, Malawi. Development 335 Southern Africa , 19 (2), 307 328. http://doi.org/10.1080/03768350020013543 Christiansen - Jucht, C., Parham, P. E., Saddler, A., Koella, J. C., & Basáñez, M. - G. (2014). Temperature during larval development and adult maintenance influences the survival of Anopheles gambiae s.s. Parasites & Vectors , 7 (1), 489. http://doi.org/10.1186/s13071 - 014 - 0489 - 3 Cohuet, A., Harris, C., Robert, V., & Fontenille, D. (2010). Evolutionary forces on Anopheles: what makes a malaria vector? Trends in Parasitology , 26 (3), 130 136. http://doi.o rg/10.1016/j.pt.2009.12.001 Collins, F. H., & Paskewitz, S. M. (1995). MALARIA: Current and Future Prospects fo r Control Further Annual Reviews . AflflU. Rev. ElIIomol (Vol. 40). Retrieved from www.annualreviews.org Cooley, L., & Linn, J. F. (2014). Takin g Innovations to Scale: Methods, Applications and Lessons . Washington, D.C. Retrieved from https://www.usaid.gov/sites/default/files/documents/1865/v5web_R4D_MSI - BrookingsSynthPaper0914 - 3.pdf Costantini, C., Li, S. G., Della Torre, A., Sagnon, N., Coluzzi , M., & Taylor, C. E. (1996). Density, survival and dispersal of Anopheles gambiae complex mosquitoes in a West African Sudan savanna village. Medical and Veterinary Entomology , 10 (3), 203 219. http://doi.org/10.1111/j.1365 - 2915.1996.tb00733.x Crawford, P ., & Bryce, P. (2003). Project monitoring and evaluation: a method for enhancing the efficiency and effectiveness of aid project implementation. http://doi.org/10.1016/S0263 - 7863(02)00060 - 1 Crutcher, J. M., & Hoffman, S. L. (1996). Malaria . Medical Microb iology . University of Texas Medical Branch at Galveston. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/21413352 Dambach, P., Machault, V., Lacaux, J. - P., Vignolles, C., Sié, A., & Sauerborn, R. (2012). Utilization of combined remote sensing techniques to detect environmental variables influencing malaria vector densities in rural West Africa . Retrieved from http://www.ij - healthgeographics.com/content/11/1/8 Debebe, Y., Hill, S. R., Tekie, H., Ignell, R., & Hopkins, R. J. (2018). Shady business: unders tanding the spatial ecology of exophilic Anopheles mosquitoes. Malaria Journal , 17 , 351. http://doi.org/10.1186/s12936 - 018 - 2499 - 7 DeCapua, S. (2009). Malawi in Pictures . Minneapolis: Lerner Publishing Group, Inc. Denning, G., Kabambe, P., Sanchez, P., Ma Essay Input Subsidies to Improve Smallholder Maize Productivity in Malawi: Toward an African Green Revolution . PLoS Biology | www.plosbiology.org (Vol. 0002). Retrieved from www.plosbiology.org Detinova, T. S., Bertram, D. S., & Organization, W. H. (1962). Age - grouping methods in Diptera of medical importance, with special reference to some vectors of malaria. 336 DHS. (2015). Malawi Malaria Indicator Survey 2014 Ministry of Health National Malaria Control Pr ogramme Lilongwe, Malawi . Retrieved from http://www.dhsprogram.com DHS. (2017). Malawi Malaria Indicator Survey . Retrieved from https://dhsprogram.com/pubs/pdf/MIS28/MIS28.pdf DHS. (2018). Malaria Indicator Surveys - Access to Malaria Indicator Surveys, Datasets. Retrieved May 21, 2019, from http://www.malariasurveys.org/ DHS. (2019a). Malaria Corner Main Page. Retrieved May 21, 2019, from https://dhsprogram.com/topics/malaria - Corner/index.cfm DHS. (2019b). Malaria Indicators Survey (MIS) Overview. Retr ieved May 21, 2019, from https://dhsprogram.com/What - We - Do/Survey - Types/MIS.cfm Diakité, N. R., Guindo coulibaly, N., Adja, A. M., Ouattara, M., Coulibaly, J. T., Utzinger, J., & N onset of a hydro - Malaria Journ al , 14 . http://doi.org/10.1186/s12936 - 015 - 0871 - 4 Doolan, D. L., Dobaño, C., & Baird, J. K. (2009). Acquired immunity to malaria. Clinical Microbiology Reviews , 22 (1), 13 36, http://doi.org/10.1128/CMR.00025 - 08 Dougherty, T. C., & Hall, A. W. (1995). Envi ronmental impact assessment of irrigation and drainage projects . Food & Agriculture Org. Dueri, S., Bopp, L., & Maury, O. (2014). Projecting the impacts of climate change on skipjack tuna abundance and spatial distribution. Global Change Biology , 20 (3), 7 42 753. http://doi.org/10.1111/gcb.12460 Duflo, E. (2004). Scaling up and evaluation. In Annual World Bank Conference on Development Economics 2004 (pp. 341 369). ESRI. (2018). Flow Accumulation. Retrieved May 31, 2019, from https://pro.arcgis.com/en/pro - app/tool - reference/spatial - analyst/flow - accumulation.htm FAO, IFAD, UNICEF, WFP, W. (2017). The State of Food Insecurity and Nutrition in the World 2017. Building resilience for peace and food security. Rome. Retrieved from http://www.fao.org/3/a - I7695e. pdf FAO, IFAD, UNICEF, WFP, W. (2018). The State of Food Security and Nutrition in the World 2018. Building climate resilience for food security and nutrition. Rome. Retrieved from www.fao.org/publications FAO. (1991). Water Harvesting A Manual for the Design and Construction of Water Harvesting Schemes for Plant Production. Retrieved March 6, 2019, from http://www.fao.org/3/u3160e/u3160e00.htm#Contents 337 FAO. (2003). Trade Reforms and Food Security: Conceptualizing the linkages . Rome. Retrieved from http: //www.fao.org/tempref/docrep/fao/005/y4671e/y4671e00.pdf FAO. (2004). The ethics of sustainable agricultural intensification . Rome. Retrieved from http://www.fao.org/tempref/docrep/fao/007/j0902e/j0902e00.pdf FAO. (2005). Irrigation in Africa in figures - AQUASTAT Survey 2005 . Retrieved from http://www.fao.org/nr/water/aquastat/countries_regions/MWI/MWI - CP_eng.pdf FAO. (2011). Save and grow . Food and Agriculture Organization of the United Nations. FAO. (2014). Facts and Figures about Irrigation areas, irrigated crops, en. Retrieved June 29, 2018, from http ://www.fao.org/nr/water/aquastat/didyouknow/index3.stm FAO. (2015). The food systems of the future need to be smarter, more efficient. Retrieved December 22, 2017, from http://www.fao.org/news/story/en/item/275009/icode/ FAO. (2016). Global Map of Irriga tion Areas (GMIA). Retrieved May 1, 2019, from http://www.fao.org/nr/water/aquastat/irrigationmap/index10.stm FAO. (2017). The future of food and agriculture: Trends and challenges . Rome. Retrieved from http://www.fao.org/3/a - i6583e.pdf FAO. (2018). GIEW S - Global Information and Early Warning Systems. Countries requiring external assistance for food. Retrieved January 25, 2019, from http://www.fao.org/giews/country - analysis/external - assistance/en/ FAO, IFAD, & WFP. (2015). The State of Food Insecurity in the World 2015. Meeting the 2015 international hunger targets: taking stock of uneven progress . Rome. Retrieved from http://www.fao.org/3/a - i4646e.pdf rice growing in the Senegal River delta (Senegal). Annales de la Societe belge de medecine tropicale , 75 (3), 179 189. Malaria Journal Impact of child malnut rition on the specific anti - Plasmodium falciparum antibody response. http://doi.org/10.1186/1475 - 2875 - 8 - 116 Foley, D. H., Torres, E. P., & Mueller, I. (2002). Stream - bank shade and larval distribution of the Philippine malaria vector Anopheles flavirostri s. Medical and Veterinary Entomology , 16 (4), 347 355. http://doi.org/10.1046/j.1365 - 2915.2002.00382.x Foody, G. M. (2002). Status of land cover classification accuracy assessment. Remote Sensing of Environment , 80 (1), 185 201. http://doi.org/10.1016/S0034 - 4257(01)00295 - 4 338 Ciufolini, M. (2015). Experimental studies on comparison of the vector competence of four Italian Culex pipiens populations for West Nile virus. h ttp://doi.org/10.1186/s13071 - 015 - 1067 - z Foster, W., & Walker, E. (2009 ). Mosquitoes (culicidae). In Medical and veterinary entomology (pp. 207 260). Elsevier. Frake, A. (2019a). Replication Data for: BVIS Management Survey Questionnaire - Dry Season. htt p://doi.org/https://doi.org/10.7910/DVN/PZEJWA Frake, A. (2019b). Replication Data for: BVIS Management Survey Questionnaire - Rainy Season. http://doi.org/https://doi.org/10.7910/DVN/9BZU2C Frake, A. (2019c). Replication Data for: BVIS Management Survey Questionnaire - Rainy Season V.2. http://doi.org/https://doi.org/10.7910/DVN/QKFEVM Friedman, J. F., Kwena, A. M., Mirel, L. B., Kariuki, S. K., Terlouw, D. J., Phillips - Shi, Y. P. (2005). Malaria and Nutritional Status Among Pre - School Children: Results from Cross - Sectional Surveys in Western Kenya. Retrieved from http://www.ajtmh.org/docserver/fulltext/14761645/73/4/0730698.pdf?expires=1563369131 &id=id&accname=guest&checksum=E914D96F0EB698B1B8E75768BF251920 Gachelin, G., Garner, P., Fe rroni, E., Verhave, J. P., & Opinel, A. (2018). Evidence and strategies for malaria prevention and control: a historical analysis. Malaria Journal , 17 (1), 96. http://doi.org/10.1186/s12936 - 018 - 2244 - 2 Garrett - Jones, C. (1964). Prognosis for interruption of malaria transmission through assessment of Nature , 204 , 1173 1175. Retrieved from https://www.nature.com/articles/2041173a0.pdf Ghebreyesus, T. A., Haile, M., Witten, K. H., Getachew, A., Yohannes, A. M., Yohannes, M., Byass, P. (1999). Incidence of malaria among children living near dams in northern Ethiopia: community based incidence survey. BMJ , 319 (7211), 663 666. http://doi.org/10.1136/bmj.319.7211.663 Giertz, Å., Caballero, J., Dileva, M., Galperin, D., & Backgr ound, T. J. (2015). Managing Agricultural Risk for Growth and Food Security in Malawi . Retrieved from http://documents.worldbank.org/curated/en/390161467998464934/pdf/100124 - BRI - P148140 - PUBLIC - Box393225B - Malawi - Policy - Note - web.pdf Gimnig, J. E., Ombok, M. , Kamau, L., & Hawley, W. A. (2001). Characteristics of Larval Anopheline (Diptera: Culicidae) Habitats in Western Kenya. J. Med. Entomol , 38 (2), 282 288. n, C. (2010). Food security: the challenge of feeding 9 billion people. Science (New York, N.Y.) , 327 (5967), 812 8. http://doi.org/10.1126/science.1185383 339 GoM. (2000a). National irrigation policy and development strategy . The Ministry. Retrieved from http s://books.google.com/books?id=U9CGAAAAIAAJ GoM. (2000b). National Irrigation Policy and Development Strategy . Lilongwe. Go M. (2002). Malawi Poverty Reduction Strategy Paper . Lilongwe. Retrieved from https://www.imf.org/External/NP/prsp/2002/mwi/01/043002 .pdf GoM. (2006). Malawi Growth and Development Strategy From Poverty to Prosperity 2006 - 2011 . Retrieved from http://www.malawi.gov.mw/Publications/MGDS November 2006 - MEPD.pdf GoM. (2010). itised and harmonised Agricultural Development Agenda . Retrieved from http://www.caadp.net/pdf/Investment plan - Malawi.pdf GoM. (2011). Malawi Growth and Development Strategy II 2011 - 2016 . Lilongwe. Retrieved from http://www.mw.one.un.org/wp - content/uplo ads/2014/04/Malawi - Growth - and - Dedvelopment - Strategy - MGDS - II.pdf GoM. (2012a). Malawi Sector Performance Report 2011 Irrigation, Water & Sanitation . Retrieved from www.deltapartnership.com GoM. (2012b). Malawi Water Sector Investment Plan Volume II . L ilongwe. Retrieved from http://water.worldbank.org/water/wpp. GoM. (2015a). National Irrigation Master Plan and Investment Framework Appendix 5 Inventory of Existing Schemes . GoM. (2015b). National Irrigation Master PLan and Investment Framework Main Rep ort . Lilongwe. GoM. (2015c). National Irrigation Master Plan and Invetment Framework Appendix 2: Agriculture . Lilongwe. GoM. (2016). National Irrigation Policy . Lilongwe. Retrieved from https://reliefweb.int/sites/reliefweb.int/files/resources/NIP_Final_S igned.pdf GoM. (2017). The Malawi Growth and Development Strategy (MGDS) iii. Government of Malawi , (August). Gonsalves, J. (2000). Going to scale: can we bring more benefits to more people more quickly. In Workshop highlights presented by the CGIAR - NGO Committee and The Global Forum for Agricultural Research with BMZ, MISEREOR, Rockefeller Foundation, IRRI and IIRR (pp. 10 14). Goodchild, M. F. (2001). Metrics of scale in remote sensing and GIS. International Journal of Applied Earth Observation and Geo information , 3 (2), 114 120. http://doi.org/10.1016/S0303 - 2434(01)85002 - 9 Gorelick, N., Hancher, M., Dixon, M., Ilyushchenko, S., Thau, D., & Moore, R. (2017). Google Earth Engine: Planetary - scale geospatial analysis for everyone. Remote Sensing of Environ ment , 202 , 18 27. http://doi.org/10.1016/j.rse.2017.06.031 340 Gu, W., Nter, G., Ller, M., Schlein, Y., Novak, R. J., & Beier, J. C. (2011). Natural Plant Sugar Sources of Anopheles Mosquitoes Strongly Impact Malaria Transmission Potential. PLoS ONE , 6 (1). htt p://doi.org/10.1371/journal.pone.0015996 Guthmann, J. P., Llanos - Cuentas, A., Palacios, A., & Hall, A. J. (2002). Environmental factors as determinants of malaria risk. A descriptive study on the northern coast of Peru. Tropical Medicine and International Health , 7 (6), 518 525. http://doi.org/10.1046/j.1365 - 3156.2002.00883.x Gwiyani - Nkhoma, B. (2005). The State and Irrigation Farming in Malawi, 1946 - 2002 . Gwiyani - Nkhoma, B. (2011). Irrigation development and its socioeconomic impact on rural communities in Malawi, 28 (2). http://doi.org/10.1080/0376835X.2011.570068 Haque, U., Hashizume, M., Glass, G. E., Dewan, A. M., Overgaard, H. J., & Yamamoto, T. (2010). The Role of Climate Variability in the Spread of Malaria in Bangladeshi Highlands. PLoS ONE , 5 (12) . http://doi.org/10.1371/journal.pone.0014341 Hartmann, A., Kharas, H., Kohl, R., Linn, J., Massler, B., & Sourang, C. Global Economy and Development Scaling Up Programs for the Rural Poor: IFAD'S Experience, Lessons, and Prospects (Phase 2) (2013). DC. R etrieved from https://www.brookings.edu/wp - content/uploads/2016/07/ifad - rural - poor - kharas - linn.pdf Hartmann, A., & Linn, J. F. (2007). Scaling up A Path to Effective Development. 2020 Focus Brief on the Washington, DC. Retr ieved from https://www.brookings.edu/wp - content/uploads/2016/06/200710_scaling_up_linn.pdf Hartmann, A., & Linn, J. F. (2008). Scaling Up: A framework and lessons for development effectiveness from literature and practice. Retrieved from http://ssrn.com/a bstract=1301625 Hay, S.I., Snow, R.W., ROgers, D. J. (1998). Predicting malaria season in Kenya using multitemporal meteorological satellit sensor data. Transactions of the Royal Society of Tropical Medicine and Hygiene , 92 , 12 20. Hay, S. ., Snow, R. ., & Rogers, D. . (1998). From Predicting Mosquito Habitat to Malaria Seasons Using Remotely Sensed Data: Practice, Problems and Perspectives. Parasitology Today , 14 (8), 306 313. http://doi.org/10.1016/S0169 - 4758(98)01285 - X Hedrick, P. W. (2011). Populati on genetics of malaria resistance in humans. Heredity , 107 (4), 283 304. http://doi.org/10.1038/hdy.2011.16 Hendrickx, S., Ballantyne, P., Duncan, A., Teufel, N., & Ravichandran, T. (2015). Scaling and Innovation Platforms . Retrieved from https://cgspace.c giar.org/bitstream/handle/10568/67884/Brief13.pdf?sequence=3 Herzog, M., Striker, G. G., Colmer, T. D., & Pedersen, O. (2016). Mechanisms of waterlogging tolerance in wheat - a review of root and shoot physiology. http://doi.org/10.1111/pce.12676 Hirzel, A . H., & Le Lay, G. (2008). Habitat suitability modelling and niche theory. Journal of Applied 341 Ecology , 45 , 1372 1381. http://doi.org/10.1111/j.1365 - 2664.2008.01524.x IEG. (2015). Malawi Irrigation, Rural Livelihoods and Agricultural Development Project Im plementation Completion Reprt (ICR) . Retrieved from http://documents.worldbank.org/curated/en/878781475068742168/pdf/ICRR - Disclosable - P084148 - 09 - 28 - 2016 - 1475068718165.pdf IFAD. (2011). Updated Guidelines and Source Book for Preparation and Implementation of a Results - Based Country Strategic Opportunities Programme (RB - COSOP) Volume 1: Guidelines . Retrieved from https://www.ifad.org/documents/10180/546394d5 - d8e7 - 475c - 84db - 34b9341602cf Ijumba, J. N., & Lindsay, S. W. (2001). Impact of irrigation on malaria in Africa: Paddies paradox. Medical and Veterinary Entomology , 15 (1), 1 11. http://doi.org/10.1046/j.1365 - 2915.2001.00279.x Ijumba, J. N., Mosha, F. W., & Lindsay, S. W. (2002). Malaria transmission risk variations derived from different agricultural prac tices in an irrigated area of northern Tanzania. Med Vet Entomol , 16 (1), 28 38. http://doi.org/DOI 10.1046/j.0269 - 283x.2002.00337.x IMF. (2017). Malawi Economic Development Document. IMF Country Report NO. 17/184 , (17). Retrieved from http://www.imf.org IRRI. (1988). Vector - borne Disease Control in Humans Through Rice Agroecosystem Management . Int. Rice Res. Inst. J.M.Hunter, L.Rey, K.Y.Chu, E.O.Adekolu - John, K. E. M. (1993). 3371.Parasitic diseases in water ressources developmen. the need for intersecto ral negociation.pdf. Jaleta, K. T., Hill, S. R., Seyoum, E., Balkew, M., Gebre - Michael, T., Ignell, R., & Tekie, H. (2013). Agro - ecosystems impact malaria prevalence: large - scale irrigation drives vector population in western Ethiopia. Malaria Journal , 12 (350). http://doi.org/10.1186/1475 - 2875 - 12 - 350 Jensen, J. R. (2005). Introductory digital image processing 3rd edition. In Upper saddle river: Prentice hall . JICA. (n.d.). FIGURES . Retrieved from http://open_jicareport.jica.go.jp/pdf/11809936_02.pdf JIC A. (1994 ). Feasibility study on bwanje valley smallholder irrigation development project . Retrieved from http://open_jicareport.jica.go.jp/pdf/11187432_01.pdf JICA. (2005). Basic Design Study Report on the Project for the Rehabilitation of the Bwanje Vall ey Irrigation System in the Republic of Malawi . Retrieved from http://open_jicareport.jica.go.jp/pdf/11809936_01.pdf Johnson, T. L., Bjork, J. K. H., Neitzel, D. F., Dorr, F. M., Schiffman, E. K., & Eisen, R. J. (2016). Habitat Suitability Model for the D istribution of Ixodes scapularis (Acari: Ixodidae) in Minnesota. Journal of Medical Entomology , 53 (3), 598 606. http://doi.org/10.1093/jme/tjw008 Johnstone, J. (2011). Vulnerability of a Run - Of River Irrigation Scheme to Extreme Hydrological 342 Conditions - A Case Study of the Bwanje Valley Irrigation Scheme. Retrieved from https://atrium2.lib.uoguelph.ca/xmlui/bitstream/handle/10214/3008/Thesis v13.pdf?sequence=3&isAllowed=y nberg, M. (2014). Satellite - derived NDVI , LST , and climatic factors driving the distribution and abundance of Anopheles mosquitoes in a former malarious area in northwest Argentina. Journal of Vector Ecology , 40 (1), 36 45. Kamwamba - Mtethiwa, J., Weather head, K., & Knox, J. (2015). Assessing performance of small - scale pemped irrigation systems in Sub Saharan Africa: Evidence from a systematic review. Irrigation and Drainage , In Press (March), 308 318. Keiser, J., Caldas, M., Castro, D., Maltese, M. F., Bo Effect of Irrigation and Large Dams on the Burden of Malaria on a Global and Regional Scale . American Journal of Tropical Medicine and Hygiene . Retrieved from http://www.ajtmh.org/docserver/fulltext/14761645/72/4/0 720392.pdf?expires=1528821946 &id=id&accname=guest&checksum=3650B440F690FAE9F9847ED73375DE96 Keiser, J., de Castro, M. C., Maltese, M. F., Bos, R., Tanner, M., Singer, B. H., & Utzinger, J. (2005). Effect of irrigation and large dams on the burden of malar ia on a global and regional scale. The American Journal of Tropical Medicine and Hygiene , 72 (4), 392 406. Kelly - Hope, L. A., Hemingway, J., & McKenzie, F. E. (2009). Environmental factors associated with the malaria vectors Anopheles gambiae and Anopheles funestus in Kenya. Malaria Journal , 8 (1), 268. http://doi.org/10.1186/1475 - 2875 - 8 - 268 Kibret, S., Alemu, Y., Boelee, E., Tekie, H., Alemu, D., & Petros, B. (2010). The impact of a small - scale irrigation scheme on malaria transmission in Ziway area, Centr al Ethiopia. Tropical Medicine and International Health , 15 (1), 41 50. http://doi.org/10.1111/j.1365 - 3156.2009.02423.x Klein, E. Y. (2013). Antimalaria drug resistance: a review of the biology and strategies to delay emergence and spread. International Jo urnal of Antimicrobial Agents , 41 , 311 317. Retrieved from https://ac.els - cdn.com/S0924857913000034/1 - s2.0 - S0924857913000034 - main.pdf?_tid=8a99e3b6 - f16b - 11e7 - b5fd - 00000aab0f6b&acdnat=1515083165_aa709a8e817387235050175b6c9ab0e7 Klinkenberg, E., Van Der Ho ek, W., & Amerasinghe, F. P. (2004). A malaria risk analysis in an irrigated area in Sri Lanka. Acta Tropica , 89 , 215 225. http://doi.org/10.1016/j.actatropica.2003.08.007 Kobayashi, J., Somboon, P., Keomanila, H., Inthavongsa, S., Nambanya, S., Inthakone Miyagi, I. (2000). Malaria prevalence and a brief entomological survey in a village surrounded by rice fields in Khammouan province, Lao PDR. Tropical Medicine and International Health , 5 (1), 17 21. http://doi.org/10.1046/j.1365 - 3156.2000.00516.x Krishnankutty Ambika, A., Wardlow, B., & Mishra, V. (2016). Remotely sensed high resolution irrigated area mapping in India for 2000 to 2015. Scientific Data , 3 . 343 http://doi.org/10.1038/sdata.2016.118 Landis, J., & Koch, G. (1977). The Measurement of Obser ver Agreement for Categorical Data. Biometrics , 33 , 159 174. Lehner, B., Verdin, K., & Jarvis, A. (2008). New global hydrography derived from spaceborne elevation data. Eos Transactions AGY , 89 (10), 93 94. Levin, S. A. (1992). The Problem of Pattern and Scale in Ecology. Ecology , 73 (6), 1943 1967. Retrieved from http://www.esa.org/history/Awards/papers/Levin_SA_MA.pdf Lindblade, K. A., Walker, E. D., Onapa, A. W., Katungu, J., & Wilson, M. L. (2000). Land use change alters malaria transmission parameter s by modifying temperature in a highland area of Uganda. Tropical Medicine and International Health , 5 (4), 263 274. http://doi.org/10.1046/j.1365 - 3156.2000.00551.x Lindsay, S. W., Parson, L., & Thomas, C. J. (1998). Mapping the ranges and relative abundan ce of the two principal African malaria vectors, Anopheles gambiae sensu stricto and An. arabiensis, using climate data . Retrieved from https://royalsocietypublishing.org/doi/pdf/10.1098/rspb.1998.0369 Linn, J. (2014). How to Reach Millions of Poor Farmer s by Scaling Up Agricultural Technology | Feed the Future. Retrieved October 24, 2017, from https://www.feedthefuture.gov/article/how - reach - millions - poor - farmers - scaling - agricultural - technology Linn, J. F. (2012). Scaling up in agriculture, rural developm ent, and nutrition (Vol. 19). Intl Food Policy Res Inst. Lourenço, P. M., Sousa, C. A., Seixas, J., Lopes, P., Novo, M. T., & Almeida, A. P. G. (2011). Anopheles atroparvus density modeling using MODIS NDVI in a former malarious area in Portugal. Journal of Vector Ecology , 36 (2), 279 291. http://doi.org/10.1111/j.1948 - 7134.2011.00168.x Lyons, C. L., Coetzee, M., & Chown, S. L. (2013). Stable and fluctuating temperature effects on the development rate and survival of two malaria vectors, Anopheles arabiens is and Anopheles funestus . Retrieved from http://www.parasitesandvectors.com/content/6/1/104 Mango, N., Makate, C., Tamene, L., Mponela, P., & Ndengu, G. (2018). Adoption of Small - Scale Irrigation Farming as a Climate - Smart Agriculture Practice and Its In fluence on Household Income in the Chinyanja Triangle, Southern Africa. Land , 7 (2), 49. http://doi.org/10.3390/land7020049 Manson, S. M. (2008). Does scale exist? An epistemological scale continuum for complex human environment systems. Geoforum , 39 (2), 7 76 788. http://doi.org/10.1016/j.geoforum.2006.09.010 Mansuri, G., & Rao, V. (2004). Community - Based and - Driven Development: A Critical Review. The World Bank Research Observer , 19 (1), 1 39. Retrieved from 344 http://dx.doi.org/10.1093/wbro/lkh012 Martens, P. ., & Hall, L. (2000). Malaria on the Move: Human Population Movement and Malaria Transmission. Emerging Infectious Diseases , 6 (2), 103 109. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2640853/pdf/10756143.pdf Mathanga, D. P., Walker, E . D., Wilson, M. L., Ali, D., Taylor, T. E., & Laufer, M. K. (2012). Malaria control in Malawi: current status and directions for the future. Acta Tropica , 121 (3), 212 7. http://doi.org/10.1016/j.actatropica.2011.06.017 May, J. M. (1961). Studies in Disea se Ecology. Studies in Disease Ecology. Mayer, K., & Pizer, H. (2011). The Social Ecology of Infectious Diseases . Elsevier. McCann, R. S., Messina, J. P., MacFarlane, D. W., Bayoh, M. N., Vulule, J. M., Gimnig, J. E., & Walker, E. D. (2014). Modeling lar val malaria vector habitat locations using landscape features and cumulative precipitation measures. International Journal of Health Geographics , 13 (1), 17. http://doi.org/10.1186/1476 - 072X - 13 - 17 Mccann, R. S., Messina, J. P., Macfarlane, D. W., Bayoh, N. , Vulule, J. M., Gimnig, J. E., & Walker, E. D. (2014). Modeling larval malaria vector habitat locations using landscape features and cumulative precipitation measures (Vol. 13). Retrieved from http://www.ij - healthgeographics.com/content/13/1/17 Mcdonald, G. (1957). The epidemiology and control of malaria Oxford University Press. London. Melaine, G., & Nonvide, A. (2017). Effect of Adoption of Irrigation on Rice Yield in the Municipality of Malanville, Benin . Retrieved from https://onlinelibrary.wiley.com /doi/pdf/10.1111/1467 - 8268.12266 Menter, H., Kaaria, S., Johnson, N., & Ashby, J. (2004). Scaling Up. In D. Pachico & S. Fujisaka (Eds.), Scaling Up and Out: Achieving Widespread Impact through Agricultural Research (pp. 9 22). Retrieved from http://ciat - library.ciat.cgiar.org:8080/xmlui/bitstream/handle/123456789/1096/Scaling_up_and_out.pdf? sequence=1 Metzger, M. E. (2204). Managing Mosquitoes in Stormwater Treatment Devices . Oakland. Retrieved from http://anrcatalog.ucdavis.edu Mkwanda, A. (2017). Gree n Belt Initiative: What has gone wrong? - The Nation Online. Retrieved August 10, 2018, from https://mwnation.com/green - belt - initiative - gone - wrong/ Moshkovsky, S. D. (1946). On the dependency of the speed of development of plasmodia malaria on temperature . Med Para Para Bol , 15 , 19 32. Mosquito Taxonomic Inventory. (2018). Genus Anopheles Meigen, 1818 | Mosquito Taxonomic Inventory. Retrieved May 14, 2018, from http://mosquito - taxonomic - inventory.info/genus - anopheles - meigen - 1818# 345 Moyo - Mana, L. (2018). Min ister Mwanamvekha impressed with Plem Construction on progress of Bwanje dam construction. Retrieved January 16, 2019, from https://www.nyasatimes.com/minister - mwanamvekha - impressed - plem - construction - progress - bwanje - dam - construction/ Mundial, B. (2004). R educing Poverty, Sustaining Growth. Scaling Up Poverty Reduction. Washington, D.C.: The World Bank. Land use and land cover changes and spatiotemporal dynamics of anopheline larval habitats during a four - year period in a highland community of Africa. The American Journal of Tropical Medicine and Hygiene , 81 (6), 1079 84. http://doi.org/10.4269/ajtmh.2009.09 - 0156 Mutero, C. M., Kabutha, C., Kimani, V., Kabuage, L., G (2004). A transdisciplinary perspective on the links between malaria and agroecosystems in Kenya. Acta Tropica , 89 (89), 171 186. http://doi.org/10.1016/j.actatropica.2003.07.003 Mutharika launches K6 billion dam proj ect. (2017). Retrieved January 16, 2019, from https://www.nyasatimes.com/mutharika - launches - k6 - billion - dam - project/ (2011). Impact of insecticide - treated bed ne ts on malaria transmission indices on the south coast of Kenya . Malaria Journal (Vol. 10). Retrieved from http://www.malariajournal.com/content/10/1/356 NASA. (2000). Measuring Vegetation (NDVI & EVI). Retrieved from https://earthobservatory.nasa.gov/ features/MeasuringVegetation/measuring_vegetation_2.ph p National Economic Council. (2000). Vision 2020 - The National Long Term Development Perspective for Malawi.pdf. News, W. H. (2018, December 13). APM walking the talk as construction of Bwanje Dam fin ished. Retrieved from https://warmheartnews.com/apm - walking - the - talk - as - construction - of - bwanje - dam - finishes/ NIH. (2016). Malaria Parasite, Mosquito, and Human Host. Retrieved May 14, 2018, from https://www.niaid.nih.gov/diseases - conditions/malaria - parasi te Noordin, Q., Niang, A., Jama, B., & Nyasimi, M. (2001). Development in Practice Scaling up adoption and impact of agroforestry technologies: Experiences from western Kenya Scaling up adoption and impact of agroforestry technologies: experiences from we stern Kenya. Development in Practice , 11 (4). http://doi.org/10.1080/09614520120066783 Nyondo, E. (2016). 15 years on, Irrigation Act still gathering dust. Retrieved August 29, 2018, from https://mwnation.com/15 - years - on - irrigation - act - still - gathering - dust / ullivan, D., & Unwin, D. (2014). Geographic information analysis . John Wiley & Sons. 346 Olson, S. H., Gangnon, R., Silveira, G. A., & Patz, J. A. (2010). Deforestation and malaria in Mâncio Lima County, Brazil. Emerging Infectious Diseases , 16 (7), 1108 15. http://doi.org/10.3201/eid1607.091785 Paaijmans, K. P., Read, A. F., & Thomas, M. B. (2009). Understanding the link between malaria risk and climate . Proceedings of the National Academy of Sciences (Vol. 106). Retrieved from http://www.pnas.org/cgi/d oi/10.1073/pnas.0903423106 Paaijmans, K. P., Wandago, M. O., Githeko, A. K., & Takken, W. (2007). Unexpected High Losses of Anopheles gambiae Larvae Due to Rainfall. PLoS One , (11). http://doi.org/10.1371/journal.pone.0001146 Packard, R. M. (1986). Agric ultural development, migrant labor and the resurgence of malaria in Swaziland. Social Science & Medicine , 22 (8), 861 867. Packard, R. M. (2007). The making of a tropical disease: a short history of malaria . JHU Press. Pates, H., & Curtis, C. (2005). Mosq uito Behavior and Vector Control Resting Behavior in Relation to House Spraying . Annu. Rev. Entomol , 50 , 53 70. http://doi.org/10.1146/annurev.ento.50.071803.130439 Patz, J. A., Graczyk, T. K., Geller, N., & Vittor, A. Y. (2000). Effects of environmental change on emerging parasitic diseases. International Journal of Parasitology , 30 (12 13), 1395 1405. Retrieved from https://ac.els - cdn.com/S0020751900001417/1 - s2.0 - S0020751900001417 - main.pdf?_tid=2c64b562 - 28b1 - 4397 - a5aa - b6fc7427ae5f&acdnat=1530388622_d35bbb 2baf6db8b0f4e9e166321634b7 Patz, J. A., & Olson, S. H. (2006). Malaria risk and temperature: Influences from global climate change and local land use practices. PNAS , 103 (15), 5635 5636. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1458623/ pdf/zpq5635.pdf Paul, P., Kangalawe, R. Y. M., & Mboera, L. E. G. (2018). Land - use patterns and their implication on malaria transmission in Kilosa District, Tanzania. Tropical Diseases, Travel Medicine and Vaccines , 4 (6). http://doi.org/10.1186/s40794 - 01 8 - 0066 - 4 Pekel, Jean - François, Cottam, A., Gorelick, N., & Belward, A. S. (2016). High - resolution mapping of global surface water and its long - term changes. Nature , 540 , 418 422. http://doi.org/10.1038/nature20584 Pérez - Escamilla, R., Dessalines, M., Fin nigan, M., Pachó, H., Hromi - Fiedler, A., & Gupta, N. (2009). The Journal of Nutrition Community and International Nutrition Household Food Insecurity Is Associated with Childhood Malaria in Rural Haiti 1,2. J. Nutr , 139 , 2132 2138. http://doi.org/10.3945/j n.109.108852 Perez - Heydrich, C., & Emch, M. E. (2013). Guidelines on the Use of DHS GPS Data [SAR8]. Retrieved from https://www.dhsprogram.com/pubs/pdf/SAR8/SAR8.pdf Pitt, H., & Jones, M. (2016). Scaling up and out as a pathway for food system transition s. 347 Sustainability (Switzerland) , 8 (10), 1 16. http://doi.org/10.3390/su8101025 Quattrochi, D. A., & Goodchild, M. F. (1997). Scale in remote sensing and GIS . CRC press. Record, R. (2007). From Policy to Practice: Changing Goverment Attitudes Towards the Private Sector in Malawi. Journal of International Development , 19 , 805 816. http://doi.org/10.1002/jid Republic of South Africa. (2011). Production Guidelines for Cowpeas . Pretoria. Retrieved from http://www.arc.agric.za/arc - gci/Fact Sheets Library/Cowpe a - Production guidelines for cowpea.pdf Sustainable agricultural intensification: The promise of innovative farming practices. 2013 Global Food Policy Report , 42 51. Roca - Heyderman, R. S. (2012). Lack of Decline in Childhood Malaria, Malawi, 2001 2010. Emerging Infectious Diseases , 18 (2), 272 278. http://doi.org/10. 3201/eid1802.111008 Sustainable intensification of agriculture for human prosperity and global sustainability. Ambio , 46 (1), 4 17. http://doi.org/10.1007/s132 80 - 016 - 0793 - 6 http://www.cimmyt.org/scaling - up - research - for - impact/ in mosquitoes and their consequences for malaria transmission. Insects , 7 (2). http://doi.org/10.3390/insects7020014 Sajidu, S., Monjerezi, M., Ngongoro, C., & Namangale, J. (2013). Country Water Resource Profile Malawi . Retrieved from http://nepadwatercoe .org/wp - content/uploads/CountryWaterResourceReport - MalawiFinal1.pdf Sarkar, S., Strutz, S. E., Frank, D. M., Rivaldi, C. - L., Sissel, B., & Sá Nchez - Cordero, V. (2010). Chagas Disease Risk in Texas. PLoS Neglected Tropical Diseases , 4 (10). http://doi.org/1 0.1371/journal.pntd.0000836 Secretariat, P., & WHO, G. (1996). Agricultural Development and Vector - Borne Diseases. Service, M., & Townson, H. (2002). Essential Malariology. In W. DA Gilles HM (Ed.), (4th ed., pp. 59 84). London: Arnold. Sewe, M. O., Ahl m, C., & Rocklöv, J. (2016). Remotely Sensed Environmental Conditions and Malaria Mortality in Three Malaria Endemic Regions in Western Kenya. PloS One , 11 (4), e0154204. http://doi.org/10.1371/journal.pone.0154204 Shankar, A. H. (2000). Nutritional Modula tion of Malaria Morbidity and Mortality . Retrieved from 348 https://academic.oup.com/jid/article - abstract/182/Supplement_1/S37/2191323 Shapiro, L. L. M., Whitehead, S. A., & Thomas, M. B. (2017). Quantifying the effects of temperature on mosquito and parasite traits that determine the transmission potential of human malaria. PLOS Biology , 15 (10), e2003489. http://doi.org/10.1371/journal.pbio.2003489 Sharma, A. K., Hubert - Moy, L., Buvaneshwari, S., Sekhar, M., Ruiz, L., Bandyopadhyay, S., & Corgne, S. (2018). Irrigation history estimation using multitemporal landsat satellite images: Application to an intensive groundwater irrigated agricultural watershed in India. Remote Sensing , 10 (6). http://doi.org/10.3390/rs10060893 Sheppard, E., & McMaster, R. B. (2008). Scale and geographic inquiry: Nature, society, and method . John Wiley & Sons. Ecological niche partitioning between Anopheles gambiae molecular forms in Camero on: the ecological side of speciation. BMC Ecology , 9 , 17. http://doi.org/10.1186/1472 - 6785 - 9 - 17 S. (2011). The Science of Scaling Up: An Agroforestry Perspec tive. Retrieved February 13, 2018, from https://www.slideshare.net/ifad/scaling - up - icraf Sinka, M. E., Bangs, M. J., Manguin, S., Coetzee, M., Mbogo, C. M., He (2010 ). The dominant Anopheles vectors of human malaria in Africa, Eur ope and the Middle East: occurrence data, distribution maps and bionomic précis. http://doi.org/10.1186/1756 - 3305 - 3 - 117 S inka, M. E., Bangs, M. J., Manguin, S., Rubio - S. I. (2012). A global map of domi nant malaria vectors. Parasites and Vectors , 5 (1), 1 11. http://doi.org/10.1186/1756 - 3305 - 5 - 69 Malaria incidence in relation to rice cultivation in the irrig ated Sahel of Mali. Acta Tropica , 89 , 161 170. http://doi.org/10.1016/j.actatropica.2003.10.015 Smelser, N. J., & Baltes, P. B. (2001). International encyclopedia of the social & behavioral sciences (Vol. 11). Elsevier Amsterdam. Smith, D. L., Battle, K. E., Hay, S. I., Barker, C. M., Scott, T. W., & Mckenzie, F. E. (2012). Ross, Macdonald, and a Theory for the Dynamics and Control of Mosquito - Transmitted Pathogens. PLoS Pathogens , 8 (4). http://doi.org/10.1371/journal.ppat.1002588 Smith, M., Muñoz, G., S anz, J., Series, A., Sanz Alvarez, J., & O´brien, E. (2014). Irrigation Techniques for Small - scale Farmers: Key Practices for DRR Implementers . Retrieved from www.fao.org/publications Sutherst, R. W. (2004). Global Change and Human Vulnerability to Vector - Borne Diseases. CLINICAL MICROBIOLOGY REVIEWS , 17 (1), 136 173. 349 http://doi.org/10.1128/CMR.17.1.136 173.2004 (2009). Global irrigated area map (GIAM), derived from remote sensing, for the end of the last millennium. International Journal of Remote Sensing , 30 (14), 3679 3733. Thomas, C. J., Nall, D., Cross, E., & Bøgh, C. (2013). Landscape Movements of Anopheles gambiae Malaria Vector Mosquitoes in Rural Gambia . PLoS ONE , 8 (7). http://doi.org/10.1371/journal.pone.0068679 Tonnang, H. E., Kangalawe, R. Y., & Yanda, P. Z. (2010). Predicting and mapping malaria under climate change scenarios: the potential redistribution of malaria vectors in Africa. Malaria Journal , 9 . http://doi.org/10.1186/1475 - 2875 - 9 - 111 Turner, B. (1986). The importance of dambos in African agriculture. Land Use Policy , 343 347. Retrieved from https://ac.els - cdn.com/026483778690030X/1 - s2.0 - 026483778690030X - main.pdf?_tid=e00e9413 - 5b7c - 4b8d - 98db - c834711cc8f4&acdnat=1535465249_034c2a9af103fbb71e919ac8a38910cd United Nations. (1948). Universal declaration of human rights . UN General Assembly . United Nations. (2016). Zero Hunger: Why it Matters . Retrieved from http://www.un.org/sustainabledevelopm ent/wp - content/uploads/2016/08/2_Why - it - Matters_ZeroHunger_2p.pdf United Nations. (2017). World Population Prospects The 2017 Revision Key Findings and Advance Tables . new York. Retrieved from https://esa.un.org/unpd/wpp/Publications/Files/WPP2017_KeyFind ings.pdf Urama, K. C. (2005). Land - use intensification and environmental degradation: empirical evidence from irrigated and rain - fed farms in south eastern Nigeria. Journal of Environmental Management , 75 (3), 199 217. http://doi.org/10.1016/j.jenvman.2004 .11.018 USAID. (2017). Sustainable Agriculture. Retrieved December 22, 2017, from https://2012 - 2017.usaid.gov/what - we - do/agriculture - and - food - security/investing - sustainable - agriculture USAID. (2019a). About DHS. Retrieved May 21, 2019, from https://dhspr ogram.com/faq.cfm USAID. (2019b). The DHS Program. Retrieved May 21, 2019, from https://dhsprogram.com/Who - We - Are/About - Us.cfm Utzinger, J., Tozan, Y., & Singer, B. H. (2001). Efficacy and cost effectiveness of environmental management for malaria control. Tropical Medicine & International Health , 6 (9), 677 687. Uvin, P. (1995). Fighting Hunger at the Grassroots: Paths to Scaling Up. World Develuptnenl , 23 (6), 927 939. Retrieved from https://ac .els - cdn.com/0305750X9500028B/1 - s2.0 - 0305750X9500028B - main.pdf?_tid=fcd5a04a - b98b - 11e7 - 944a - 00000aacb35f&acdnat=1508939835_b8cd3b477c2a844c3de0317b436079e1 350 Veldkamp, T., Polman, N., Reinhard, S., & Slingerland, M. (2011). From scaling to governance of the land system: Bridging ecological and economic perspectives. Ecology and Society , 16 (1). http://doi.org/1 Veldwisch, G. J., Bolding, A., & Wester, P. (2009). Sand in the engine: The travails of an irrigated rice scheme in Bwanje valley, Malawi. Journal of Development Studies , 45 (2), 197 226. http://doi.org/10.1080/00220380802265587 deforestation to malaria in the Amazon: characterization of the breedin g habitat of the principal malaria vector, Anopheles darlingi. The American Journal of Tropical Medicine and Hygiene , 81 (1), 5 12. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/19556558 Wesolowski, A., Eagle, N., Tatem, A. J., Smith, D. L., Noor, A. M ., Snow, R. W., & Buckee, C. O. (2012). Quantifying the Impact of Human Mobility on Malaria. Science , 338 (6104), 267 270. http://doi.org/10.1126/SCIENCE.1223467 White, M. T., Griffin, J. T., Churcher, T. S., Ferguson, N. M., Basáñez, M. - G., & Ghani, A. C. (2011). Modelling the impact of vector control interventions on Anopheles gambiae population dynamics. Parasites & Vectors , 4 , 153. http://doi.org/10.1186/1756 - 3305 - 4 - 153 White, N. J. (2008). Malaria Journal The role of anti - malarial drugs in eliminating malaria. http://doi.org/10.1186/1475 - 2875 - 7 - S1 - S8 WHO. (2016). Malawi . Geneva. Retrieved from http://www.who.int/malaria/publications/country - profiles/profile_mwi_en.pdf?ua=1 WHO. (2017). World Malaria Report 2017 . Geneva. Retrieved from http://apps.who .int/iris/bitstream/10665/259492/1/9789241565523 - eng.pdf?ua=1 WHO. (2018 ). Malawi . Retrieved from https://www.who.int/malaria/publications/country - profiles/profile_mwi_en.pdf?ua=1 Wiebe, A., Longbottom, J., Gleave, K., Shearer, F. M., Sinka, M. E., Masse (2017). Geographical distributions of African malaria vector sibling species and evidence for insecticide resistance. Malaria Journal , 16 (1), 85. http://doi.org/10.1186/s12936 - 017 - 1734 - y Wigboldus, S., & Brouwers, J. (2016). Using a Theory of Scaling to guide decision making. Towards a structured approach to support responsible scaling of innovations in the context of agrifood systems. Wageningen. Retrieved from http://www.theoryofchange.nl/sites/default/files/resource/seerp_wigbol dus_ - _theory_of_scaling_lr.pdf Wigboldus, S., Klerkx, L., Leeuwis, C., Schut, M., Muilerman, S., Jochemsen, H., & Nl, S. W. (2016). Systemic perspectives on scaling agricultural innovations. A review. Agronomy for Sustainable Development , 36 . http://doi.o rg/10.1007/s13593 - 016 - 0380 - z Wigboldus, S., & Leeuwis, C. (2013). Towards responsible scaling up and out in agricultural development An 351 exploration of concepts and principles . Wageningen UR. Retrieved from http://edepot.wur.nl/306491 Wilson, M. L., Walke r, E. D., Mzilahowa, T., Mathanga, D. P., & Taylor, T. E. (2012). Malaria elimination in Malawi: Research needs in highly endemic, poverty - stricken contexts. Acta Tropica , 121 (3), 218 226. http://doi.org/10.1016/j.actatropica.2011.11.002 Woodhouse, P., Ja n Veldwisch, G., Venot, J. - P., Brockington, D., Komakech, H., & Manjichi, Â. (2017). African farmer - led irrigation development: re - framing agricultural policy and investment? The Journal of Peasant Studies , 44 (1), 213 233. http://doi.org/10.1080/03066150.2 016.1219719org/10.1080/03066150.2016.1219719 World Bank. (2010). The Zambezi River Basin A Multi - Sector Investment Opportunities Analysis Modeling, Analysis and Input Data . Washington D.C. . Retrieved from http://documents.worldbank.org/curated/en/5991914 68203672747/pdf/584040V40WP0Wh 1lysis0and0Input0Data.pdf Wu, J., & Li, H. (2006). Concepts of Scale and Scaling. In Scaling and Uncertainty Analysis in Ecology Methods and Applications (pp. 3 15). Springer. Retrieved from https://link.springer.com/content/ pdf/10.1007/1 - 4020 - 4663 - 4.pdf Yamana, T. K., & Eltahir, E. A. B. (2013). Incorporating the effects of humidity in a mechanistic model of Anopheles gambiae mosquito population dynamics in the Sahel region of Africa . Retrieved from http://www.parasitesandve ctors.com/content/6/1/235 Yasuoka, J., & Levins, R. (2007). Impact of Deforestation and Agricultural Development on Anopheline Ecology and Malaria Epidemiology. Retrieved from http://www.ajtmh.org/docserver/fulltext/14761645/76/3/0760450.pdf?expires=15489 67924 &id=id&accname=guest&checksum=7A656DB7D53C264E534BF598C5ED4F0F The Effect of Deforestation on the Human - Biting Rate of Anopheles Darlingi, the Pr imary Vector of Falciparum Malaria in the Peruvian Amazon. Retrieved from www.ajtmh.org. You, L., Ringler, C., Nelson, G., Wood - What Is the Irrigation Potential for Africa? A Combined Biophysical and Socioeconomic Approach . Retrieved from http://www.ifpri.org/publications/results/taxonomy%3A468. productivity from patch to landscape: a case study of Chang bai Mountain Nature Reserve, China. Landscape Ecol , 22 , 303 315. http://doi.org/10.1007/s10980 - 006 - 9027 - 9 Zheng, B., Myint, S. W., Thenkabail, P. S., & Aggarwal, R. M. (2015). A support vector machine to identify irrigated crop types using time - series Lan dsat NDVI data. International Journal of Applied Earth Observation and Geoinformation , 34 , 103 112. http://doi.org/10.1016/J.JAG.2014.07.002 Zuehlke, E. (2009). The Demographic and Health Surveys at 25 Years and Beyond Population Reference Bureau. Ret rieved May 21, 2019, from https://www.prb.org/dhspolicyseminar/