1 A STUDY ON DAM CONSTRUCTION, HY DROLOGIC AL CHANGE S , AND THE SHIFT IN DIETARY PROTEIN I N THE LOWER MEKONG R IVER BASIN By Mateo Burbano A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Civil Engineering Master of Science 2019 ii ABSTRACT A STUDY ON DAM CONSTRUCTION, HY DROLOGIC AL CHANGE S , AND THE SHIFT IN DIETARY PROTEIN I N THE LOWER MEKONG R IVER BASIN By Mateo Burbano An important ramification that could be linked to the accelerated dam construction in the Lower Mekong Basin (LMB) is a dietary shift from fish as a source of pro tein to land - animal - based protein. T he proposed chain of events that lead to this conclusion starts with a disruption in - pulse, annual discharge, seasonality, water level), combined with physical barrier (i. e., a dam structure ) lead to lower fish catch rates and reproduction from migratory impedim ent. A shift to a westernized diet by th e developing country population of the LMB countries , can be observed as well. A relationship between fish catch and flooded area is developed to downscale fish catch to a 10 km grid in the study area, which enables the quantification of yearly distributed catch per capita. It i s found that wet years yield higher catch per capita than dry years. Further, a statistical analysis on fish and meat production and consumption show an overall increase in local production of meat. The 2% cropland expansion is found to be largely attributed to an increase in meat production as most of the crops are grown for animal feed. Finally , a meat virtual water trade (VWT) network of is constructed between 1988 and 2016. Virtual water outflow is four to eight orders of magnitude larger than inflow from 1998 to 2003, when outflow drops significantly, and the direction of flow co mpletely reverses. In fact, ed similar values to that of past outflow (1998 - 2003). The abrupt virtual water flow change opens the door for a water savings analysis, where local and international producti on of meat is discussed to reach greater regional water savings ( R WS) (i.e., minimizing water footprint) . iii ACKNOWLEDGEMENT S At the outset, I would like to thank my family for their unconditional continued support and encouragement through out my academic life. I would like to express the deepest appreciation to my committee chair and principal advisor, Dr. Yadu Pokhrel , who gave me the neces sary tools and guidance to be able to complete my studies. He continually and convincingly displayed a spirit of exploration in regard to research and scholarship, and an ex citement in regard to teaching. It is a great honor to work under his supervision. Without his leadership and persistent help this thesis would not have been possible. In addition, a thank you to Dr. Phani Mantha and Dr. Nathan Moore, for serving as part of my thesis committee and providing me with invaluable instruction and advice . Their encouragement and comprehensive recommendation s not only aided this thesis flourishment but also elevated its quality. I wish to express my extreme sincere gratitude towards Dr. Piercy Pierre, who m made m y advanced degree studies possible. He took me under his wing and generously helped me finance my studies through multiple so urces of financial aid. I will be forever be grateful for his generosity . My studies and thesis work would not have been possible without support. Las t, a special thank you to my lab mates Suyog Chaud h ari, Sanghoon Shin, and Farshid Felfelani for the endless hours of help with technical and research related matters. I learned most of my computer analytical skills from them and their advice played a key role in the formation of this thesis. iv TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ .......................... v LIST OF FIGURES ................................ ................................ ................................ ....................... vi 1. INTRODUCTION ................................ ................................ ................................ ................... 1 1.1 Background ................................ ................................ ................................ ..................... 3 1.1.1 Flood Variations ................................ ................................ ................................ ........ 3 1.1.2 Fish catch and flooded area relationship ................................ ................................ ... 4 1.1.3 Water footprint, VWT network, and Regional water savings ................................ .. 5 1.1.4 Land use/land cover change ................................ ................................ ...................... 6 1.2 Research Motivation ................................ ................................ ................................ ....... 8 1.3 Research Que stions and Objectives ................................ ................................ ................ 9 1.4 Structure of Thesis ................................ ................................ ................................ ........ 11 2. STUDY AREA, DATA, AND MODEL ................................ ................................ ............... 12 2.1 Stud y Area Details ................................ ................................ ................................ ........ 13 2.2 Data ................................ ................................ ................................ ............................... 17 2.3 Hydrological and Flood Dynamics Model Settings ................................ ...................... 18 3. METHODOLOGY ................................ ................................ ................................ ................ 20 3.1 Hydrological and Flood Dynamics Model ................................ ................................ .... 21 3.2 Creating a Gridded Map of Estimated Riverine Fish Catch ................................ ......... 23 3.3 Virtual Water Trade Network ................................ ................................ ....................... 26 4. RES ULTS AND DISCUSSION ................................ ................................ ............................ 28 4.1 CaMa - Flood Model Validation and Results ................................ ................................ . 28 4.2 Empirical Relationship between Flooded Areas and Fish Catch ................................ .. 29 4.3 Fish Catch Downscale ................................ ................................ ................................ ... 35 4.4 Fish Catch Loss ................................ ................................ ................................ ............. 42 4.5 LMB Land Use Change ................................ ................................ ................................ 46 4.6 Meat Production and Consumption ................................ ................................ .............. 48 4. 7 Fish Production and Consumption ................................ ................................ ................ 53 4.8 Virtual Water Flows ................................ ................................ ................................ ...... 58 4.9 Virtual Water Trade Network ................................ ................................ ....................... 64 4.10 Regional Water Savings ................................ ................................ ................................ 66 5. CONCLUD ING REMARKS ................................ ................................ ................................ 71 APPENDIX ................................ ................................ ................................ ................................ ... 73 BIBLIOGRAPHY ................................ ................................ ................................ ......................... 97 v LIST OF TABLES Table 1: Summary of the LMB spatial distribution (Frenken & FAO, 2011) .............................. 13 Table 2: Comparison between observation fish catch (Kian et al., 2005) and distributed fish catch in Cambodia using Equation 3 between the years 2000 and 2004. ................................ ............. 34 Table 3: Tonle Sap Lake summary of estimated fish catch loss for each outflow restriction scenario presented above (Figure 10). ................................ ................................ ................................ ......... 45 Table 4: Summary of water savings results in Mm 3 /year. ................................ ............................ 68 Table 5: Yearly fish catch data from FAO FishStat and flooded area from CaMa - Flood modeling results (Pokhrel, Shin, et al., 2018a). ................................ ................................ ............................ 74 Table 6: Population (millions) data from IMF, retrieved October 9 th , 2018. ............................... 79 Table 7: Virtual water trade network results summary at a regional level for years 1988, 1995, 2004, 2008, 2012, and 2016. ................................ ................................ ................................ ......... 81 Table 8: Calculated virtual water contents of all the meat commodities for every nation (m 3 /ton) (A K Chapagain & Hoekstra, 2003). ................................ ................................ ............................ 86 vi LIST OF FIGURES Figure 1. Dams in the Mekong River Basin. The red and black contour lines show the Upper and Lower Mekong Basin respectively breaking at the border between Laos, Myanmar and China. The background shows land use types and irrigated croplands obtained from (Salmon, Friedl, Frolking, Wisser, & Douglas, 2015). Dams are broken down into three categories (operational, proposed, and under construction) to highlight the number of dams that will become operational in the near future. The database for the dams was retrieved from the Research Program on Water, Land and Ecosystems (WLE), Greater Mekong. ................................ ................................ .......................... 14 Figure 2: Workflow diagram of the 10 km fish catch downscale applying a univariate power regression model. Flooded area data is retrieved from the CaMa - Flood model results while fish catch data is calculated from FAO FishStat (Pokhrel, Shin, et al., 2018a). ................................ .. 24 Figure 3: Observation and simulation data for nine major dams in the Lower Mekong Basin. Observation data retrieved from the MRC. ................................ ................................ ................... 28 Figure 4: Relationships between measured fish catch and modeled river flooded area for year s 1986 to 2010. Fish catch data are obtained from FAO FishStat and flooded area results are obtained from the CaMa - Flood model results (Pokhrel, Shin, et al., 2018b). ................................ ............. 30 Figure 5: Model fit validation through (A) comparison between regression fitted derived relationship between fish catch and discharge (Y - axis) from (McIntyre, Reidy Liermann, et al., 2016), and the derived relationship be tween fish catch and flooded area (X - axis) outlined above. The time series (B) and (C) show the limitations of using the univariate models by plotting average model form this study respectively. ................................ ................................ .............................. 32 Figure 6: Fish catch distribution map as eight - year averages from 1979 to 2010. Freshwater fish catch data from FAO FishStat was distributed using to the 10 km grids using Equation 3 based on the derived relationship of catch and flooded area. Grids with flood depths lower than 20 cm where masked as the fish of interest are located in deeper waters. Flooded area and depths data are taken directly from the CaMa - Flood model results (Pokhrel, Shin, et al., 2018b). Country borders are delineated in black. LMB border is delineated in red. The higher concentrations of total fish catch are shown in dark blu e, while lower concentrations are in light blue. ................................ .......... 36 Figure 7: Fish catch per capita distribution on four dry years: 1998, 2003, 2004, and 2005. Catch per capita distribution is used to normalize the effect of fish catch increases due to population demands. Population data is retrieved from IMF (October 9 th , 2018). ................................ ......... 38 Figure 8: Fish catch per capi ta distribution on four wet years: 2000, 2001, 2006, and 2007. Catch per capita distribution is used to normalize the effect of fish catch increases due to population demands. Population data is retrieved from IMF (October 9 th 2018). ................................ .......... 40 vii Figure 9: LMB CaMa - Flood simulation domain red box encloses the Lower Mekong sub - area, while green box encloses the Tonle Sap Lake area. ................................ ................................ ..... 43 Figure 10: Estimated fish catch loss for the Lower Mekong Basin (A) and for the Tonle Sap Lake (B). Monthly calculated potential fish catch loss in the total Lower Mekong study area presented by Figure 9 is shown here. Every month in the bar graph has a different bar with a light blue tone representing each scenario (baseline to 50 % outflow restriction) (A). Monthly calculated potential fish catch l oss in the Tonle Sap Lake shown in orange (B). The estimation is projected using Equation 3 on the CaMa - Flood flooded area results at each of the scenarios presented from (Pokhrel, Shin, et al., 2018a). ................................ ................................ ................................ ........ 44 Figure 11: Land use and land cover types for 1992, 2000, 2008, and 2015. The embedded pie charts show the fraction of the main land cover types with matching color coding except for the black pieces which are a lumped representation of the minor land use types. The urban areas are excluded from the pie charts since they represent a negligible amount of area. Data source: European Space Agency - Climate Impact Initiative ( ESA - CCI: https:// www.esa - landcover - cci.org/, accessed on 27 January 2018). ................................ ................................ ................................ ............................... 46 Figure 12: Time series comparing meat production and consumption between the years 1988 and 2013. The meat products (i.e., cattle, pig, sheep, goat, chicken, and horse). Additionally, a statistical mean changepoint is calculated and represented by the vertical lines on each graph (color coded accordingly). The consumption of th e commodity at hand is taken directly from the supply data at FAOSTAT. The production of meat above is taken from the production data at FAOSTAT. ................................ ................................ ................................ ................................ ....................... 49 Figure 13: Time series comparing fish production and consumption between the years 1988 and 2013. The fish catch products are exclusively freshwater fish in order to isolate fish from rivers and lakes. Additionally, a statistical mean change point for each variable is calculated and represented by the vertical line on each graph. The consumption of fish is taken directly from the supply data at FAOSTAT. The production of fish above is taken from the catch data at FAOSTAT. ................................ ................................ ................................ ................................ ....................... 54 Figure 14: Time series shows the virtual water imports, ROW to LMB (green line), and virtual water exports, LMB to ROW (blue line), through the years. Data shown is derived from VWT calculations (VWC data retrieved from A.K. Chapagain & Hoekstra (2003), and trade data from FAOSTAT detailed trade matrices. ................................ ................................ .............................. 59 Figure 15. VWT flows of animal - based protein products (cattle, pig, sheep, goat, chicken, and horse) between 1988 and 2016. The width of each band represents quantity of water in traded. The LMB basin c ountries are represented with green bands while the Rest of Asia and other continental regions have individual colors. The circular figure areas are scales to the total area traded. Data retrieved from A.K. Chapagain and Hoekstra (2003) and FAOSTAT detaile d trade matrices. This figure was created using the network visualization tool, Circos (Krzywinski et al., 2009). ................................ ................................ ................................ ................................ ............ 64 viii Figure 16: Water savings time series showing the difference between VWT outflow and VWT inflow using data from VWT network results. ................................ ................................ ............. 67 1 1. INTRODUCTION The Mekong River Basin (MRB hereafter) is the second most biodiverse river system in the world after the Amazon that host s the region for the largest inland fishery production (Ziv, Baran, Nam, Rodríguez - Iturbe, & Levin, 2012a) . The river originates in the Tibetan Plateau, and works its way down to and through Vietnam, passing Myanmar, Lao Republic (Lao PDR), Thailand, and Cambodia with a total length of ~4800 km (MRC & Mekong River Commission Lao PDR, 2005a, 2010b) . In the past two decades the river withstood signi ficant hydrological changes due to extensive manmade structural progress in the form of dams (Arias, Piman, Lauri, Cochrane, & Kummu, 2014; Baran, Guerin, & Nasielski, 2015; Baran & Myschowoda, 2009; Brownell, Reeves, Thomas, Smith, & R yan, 2017; Dugan, Barlow, Agostinho, Baran, Cada, Chen, Cowx, Ferguson, Jutagate, Mallen - Cooper, et al., 2010; Kondolf, Rubin, & Minear, 2014; Kummu & Sarkkula, 2008; Lu & Siew, 2006; Wild, Reed, Loucks, Mallen - Cooper, & Jensen, 2019; Xue, Liu, & Ge, 2011) ding water infrastructure (Baran & Myschowoda, 2009; Dugan, Barlow, Agostinho, Baran, Cada, Chen, Cowx , Ferguson, Jutagate, & Mallen - Cooper, 2010; Grumbine & Xu, 2011; Kummu & Sarkkula, 2008; Li et al., 2013; Piman, Cochrane, Arias, Green, & Dat, 2013a; Wild et al., 2019; Winemiller et al., 2016a) . e main stem and its additional 35,000 M W potential of its tributaries will result in the building of 16 dams at the main stem and over 100 dams which are at the tributaries by 2030 (ICEM & Management, 2010a; Pokhrel, Burbano, et al., 2018; Stone, 2011) . The accelerated development of dam construction is bound to disrupt logy as well as fish populations and specie biodiversity altogether. Projections predict a drop in migratory fish biomass by up to 51.3%, condemning 100 fish 2 species and moving them to the list o f critically endangered species (Ziv et al., 2012a) . second most aquatically biodiverse ecosystem (MRC & Mekong River Commission Lao PDR, 2010b; Winemiller et al., 2016b) . the environment as populations rise through food production is crucial because food production share s the highest portion of global freshwater consuming ; ~ 80% of the world water resource is used for this sector alone (Rost et al., 2008) . Today, the livelihood of 80% of the 60 million inhabitants of the Lower Mekong Basin ( LMB hereafter) is reliant on fishery and agriculture that depended heavily on the seasonal rainfall and flood pattern (Baran & Myschowoda, 2009; ICEM & Management, 2010b) . Mo re importantly, fish remains as the main source of protein for countries in the Lower Mekong River Basin (LMRB), namely flood pulse secures the reproduction of most migratory fish (ICEM & Management, 2010b) . With a compromised availability of their main protein source, a forced dietary shift towards alternative sources of protein such as land animal - based protein, or high protein crop derivatives should be considered (Orr, Pittock, Chapagain, & Dumaresq, 2012a; Pittock, Dumaresq, & Orr, 2017) . This in turn, triggers an array of measurable effects as the protein calories are replaced through locally grown or imported commodities. While hydrological variation is tracked using large - scale hydrological models, analyses regarding secondary effects such as a drop in fish catch per capita and its effects on alternative diets calls for an urgent need to better understand the dynamics between dam development , hydrologic variations, land use/land cover change s, and the potential effects of food production and fisheries . 3 1.1 Background The Mekong River faces many great challenge s like climate change, which is expected to rise basin - wide temperatures, change monsoon patterns, and ultimately result in unpredictable hydrological changes. For example, a - 3% to 15% change in total annual flows has been suggested to be heavily driven by irrigation expansions and climate change (Hoang et al., 2019) . The countries surrounding the basin are on the midst of fast - passed economic growth, creating a cycle of increased food and energy requirements, followed by deforestation, expansion of road networks, and dam construction (Pokhrel, Burbano, et al., 2018) . 1.1.1 Flood Variations Flood variation anomalies in the LMB have been observed through time (since before dam construction to present) showing an increasing trend towards new extreme we t and dry periods (Erban & Gorelick, 2016; Han, Long, Fang, Hou, & Hong, 2019; Hoang et al., 2019; Lauri et al., 2012a; Piman, Cochrane, Arias, Green, & Dat, 2013b; Piman, Lennaerts, & Southalack, 2013; Räsänen, Koponen, Lauri, & Kummu, 2012; Shrestha et al., 2013) . Granted, t hese anomalies could, to some extent, be attributed to changes in large - scale atmospheric mechanisms such as: Niño - Southern Oscillation (ENSO), Western North Pacific Monsson (WNPM) and the Indian Summer Monsson (ISM) (J. M. Delgado, Apel, & Merz, 2010; J. M. Delgado, Merz, & Apel, 2012; Räsänen & Kummu, 2013; Ward, Beets, Bouwer, Aerts, & Renssen, 2010) . Ho wever , research shows that the changes are also due to da m construction (J. M. Delgado et al., 2012; Lauri et al., 2012b; Sabo et al., 2017a) . In fac t, the direct impact of dam construction to the flood pulse dynamics of the LMB was modeled at varying flow regulation scenarios 10 - 50% peak flow reduction. The model results on the Tonle Sap Lake (TSL) show a reduction in peak flow of 7 - 37% and 7 - 34%, and a reduction of the reversed flow of 11 - 80% and 15 - 88% at LO (Lake Outlet) and PK (Prek Kdam) 4 station respectively (Pokhrel, Shin, Lin, Yamazaki, & Qi, 2018a) . The flood pulse reduction is a variable to consider when considering fish population drop, however, a much more heavily studied variable is the physical barrier created by dams is an other important issue because it prevents spawning of migratory species which are of biological and economical importance (Dugan, Barlow, Agostinho, Baran, Cada, Chen, Cowx, Ferguson, Jutagate, Mallen - Cooper, et al., 2010) . This study strives to advanc e the understanding of the former as it s effect on fish catch remains vaguely understood. 1.1.2 Fish catch and flooded area relationship contribute to a drop in fish populations, especially migratory fish, which account for 71% of the fisheries yield in the LMB (Barlow, Baran, Halls, & Kshatriya, 2008) . For instance, an strong association between fish catch and water level for a small temporal resolution (1998 - 20 01) in Tonle Sap Lake (Van Zalinge, n.d. - a ) migration patterns is captured (N van Zalinge et al., 2003) . Years later a large study by McIntyre and colleagues developed a relationship between discharge and fish catch in order to downscale fish catch to the major river basins of the world and study freshwater fishery (McIntyre, Reidy Liermann, & Revenga, 2016) . A more complex approach was late taken by Sabo and colleagues, when they developed a multivariate relationship between fish catch and flood anomalies of the Mekong River in order to design river fl ows to improve food security futures in the LMB (Sabo et al., 2017b) . f lood and fish catch, over discharge and fish catch. This study develops a univariate model between flooded area and fish catch, taking advantage of the simpler requirement for one variable which 5 allows for higher resolution downscaling, but also incorporat ing a variable with greater correlation than that of discharge. 1.1.3 Water footprint , VWT n etwork , and Regional water savings Locally grown land - based food production explains land use conversion. However, water consumption from food production (local and imported) can be quantified through water footprint, virtual water trade (VWT), and regional water savings ( RWS ) changes (Cai et al., 2019; A K Chapagain, Hoekstra, & Savenije, 2006; Ashok K. Chapagain & Hoekstra, 2008; da Silva et al., 2016; Dalin, Konar, Hanasaki, Rinaldo, & Rodriguez - Iturbe, 2012; Dalin, Qiu, Hanasaki, Mauzerall, & Rodriguez - Iturbe, 2015; Dang, Lin, & Konar, 2015; Duarte, Pinilla, & Serrano, 2019; Hanasaki, Inuzuka, Kanae, & Oki, 2010; Konar, Hussein, Hanasaki, Mauzerall, & Rodriguez - Iturbe, 2013; Oki, Entekhabi, & Harrold, 2004; Suweis et al., 2011; Tian et al., 2018; Zhang, Zhang, Tang, Chen, & Wang, 2016) . A significant loss of fish protein is projected and as population grows the demand for an alternative will leave a quantifiable water footprint (Dugan, Barlow, Agostinho, Baran, Cada, Chen, Cowx, Ferguson, Jutagate, Mallen - Cooper, et al., 2010; Li et al., 201 3) . In fact, a study by Orr et al., (2012) projects water footprint changes under two separate scenarios. The first scenario models the effects of 11 main stem dams, while the second scenario models the effects of 88 main stem and tributary dams combined. The water foot print outcomes are a 4 - 7% increase use under scenario 1 and a 6 - 17% increase use under scenario 2. The results on the study mentioned are conservative nor do th ey account for the increase in demand for crop production such as rice, which would require even more water (Orr et al., 2012a) . VWT is not taken into account in th e study by Orr and colleagues , which would potentially shed li ght on regional interactions with the rest of the world 6 (ROW). Additionally, regional water savings (RWS) (sum of virtual water trade) is not taken into consideration. Knowing how much water is flowing in and out of the region can explain water saving pote ntial to alleviate water scarcity globally and/or locally. embedded in food demand and consumption combining a hydrological model and an economical model. Two scen arios are put forward for decadal projection stemming from the baseline scenario (BL), which provides a feasible picture of developments based on expected trends: Inner Mongolia Mongolia plus baseline (IM+B) scenario reduces both IM and Beijing area by 50% in 2020 and 2030. The model results find that virtual water transfers will increase by a volume of 206 km 3 from 239 km 3 to 445 km 3 . As importations rise in the form of virtual w ater in China, water savings (WS) rise (Dalin et al., 2015) . So long as the countries that produce a comm odity produce it more efficiently than other nations, their production is preferred in terms of RWS . The last section of t his study takes a similar approach for the Mekong region where RWS and the best scenario is presented in discussion form . 1.1.4 Land use /land cover change Land use /land cover cropland expansion is are directly related to food production including crops and livestock (Abdullah, 2006; Aleksandrowicz, Green, Joy, Smith, & Haines, 2016; Bonfils & Lobell, 2007; Carpenter, Stanley, & Vander Zanden, 2011; Gephart et al., 2016; Jacobi et al., 2018; Monfreda, Ramankutty, & Foley, 2008; Smajgl et al., 2015; Stonestrom, Scanlon, & Zhang, 2009) . In fact, t he deforestation of the Amazon rainforest, is strongly attribu ted to an increased soybean production ( Dalin et al., 2012) . The previously mentioned study by Orr and colleagues also calculates land use change and footprint. The land footprint outcomes are a 13 - 27% increase 7 use under scenario 1 and a 19 - 63% increase use under scenario 2 (Orr et al., 2012a) . This study will present similar land use/land cover observable changes from the dietary shift on the LMB. 8 1.2 Research Motivation Freshwater resources are under increasing pressure as the demand for water - intensive products r ise s . Because loss of fish catch is highly likely, a dietary shift towards other sources of animal loss and retaining fish exportation would come at a high eco nomical cost to the LMB countries (Pittock et al., 2017; Ziv et al., 2012a) . This calls for a robust analysis in spatiotemporal fish catch variations. Additionally, m anaging rise water consuming process taking up to 80% of the world water r esources (Rost et al., 2008) . Hence, a suggested effort to reduce fresh - water use globally is optimizing VWT (Dalin et al., 2012) . VWT has been studied at different spatial and temporal scales (Liu et al., 2019) . This study pro poses a fractal basin - wide study focused on temporal variation (1988 - 2016) embedded in meat ( e.g., beef, chicken, and pork) production and trade. While the LMB is often studied by itself, a step back to look at this inte raction with t he rest of the world . Therefore, this study proposes a RWS analysis on meat stemming from the VWT network. Lastly, i as RWS analyses on rice showed regional water losses ( R WL) while wheat showed RWS (Konar et al., 2013) . The last section of this study analyses meat com modities individually RWS . The study recognizes that c ountries in the LMB, specifically the coastal ones, are highly dependent on income from fishery exports. (Allison et al., 2009) . However, economic impacts of food are outside of the scope of this study. 9 1.3 Research Questions and Objectives The overall goal of this thesis is to explore the causes and effects of protein source dietary shift of the LMB countries. In the efforts of achieving this goal the study asks the following scientific questions: Is there a relationship between fish catch and hydrological variations induced by da m construction in the LMB? What is the impact that a dietary shift from fish to livestock products has on virtual water transfers b etween the LMB and the ROW? How can we optimize RWS by balancing livestock production between the LMB and the ROW? These questions are answered through the reach of the following objectives: Devel opment of a univariate relationship between fish catch and flood to downscale fish catch data using results from a hydrological model and study fish catch temporal variability. Construct a VWT network using virtual water contents (VWC) of meat (e.g., catt le, chicken, pork) commodities and trade data to study the impacts of adopting a higher meat - based diet. Compare the estimated effects of domestic (LMB) versus internationally sourced livestock commodities in terms of RWS to discuss optimal production loca tions for each commodity . The outcomes of this research will provide insight to the accelerated additio n of dams to the LMB. This study is relevant as it will determine whether or not hydrological changes will take a 10 toll on fish reproducti on abilities, wh ich in turn could hinder fish catch . These effects could be another driving force to a dietary shift from fish protein to livestock. Additionally, t he o bserved virtual water flows could p resent a significant enlargement of virtual water inflow and outflow in terms of livestock commodities. The study will show t he outcome of RWS of each individual commodity ( e.g., chicken, beef, pork). Meanin g, policy makers will be encouraged to push for more efficient trade policies for the LMB . 11 1.4 Structure of Thesis The thesis has five sections including the introduction. An introduction of the main topics and background is provided in section one. The study area, model, and data used are presented in section two. The third section outlines the methods used for every analysis and model presented in the results. The results and discussion, section four, of the thesis is structured as follows: First, the relationship between fish catch and flooded area is presented. Second, the study presents the process of fish downscal e and distribution to 10 km grids. Third, the study presents the observed land use change of nearly two and a half decades. Fourth, a statistical analysis and discussion of fish and meat production and consumption is presented . Fifth, the VWT network is co nstructed. Last, the RWS results are presented with their respective discussion. Section five of the study, conclusions, presents the closing remarks and findings and the thesis ends. 12 2. STUDY AREA, DATA, AND MODEL The thesis focuses on what is known as the LMB , which begins at the Golden Triangle where the national borders between Thailand, Laos, and Myanmar meet and ends down at the Mekong Delta at the bottom of Vietnam. The area is delineated in red in Figure 1 . The data retrieved for the analyses mainly come from the FAO website, including fish catch, fish consumption, meat production, and meat consumption. Population that is retrieved from IMF, spec ifically, the October 9 th , 2018 iteration. The hydrological model employed is CaMa - Flood and its results from the study by Pokhrel and colleagues (2018) are also used in this study. The model results of use include discharge, water storage, flood depth, and flooded area. The following sub - sections expand on these topics to a greater detail. 13 2.1 Study Area De tails The LMB, located in Southeast Asia , has a total area of ~ 606,000 km 2 , which is 76% of the entire Mekong Basin which has a total area of ~ 795,000 km 2 (Frenken & FAO, 2011) . The countries sharing the basin include Laos, Thailand, Cambodia, and Vietnam. Their area contribution to the basin is summarized in Table 1 . Table 1 : Summary of the LMB spatial distribution (Frenken & FAO, 2011) Area Countries included Area of country in basin ( Percentage of total area of basin % 606 000 Laos 202 000 33 Thailand 184 000 30 Cambodia 155 000 26 Vietnam 65 000 11 discharge is approximately ~ 475 km 3 /year or ~15,000 m 3 /s making it the 10 th largest river in the world in terms of annual flow at its mouth (MRC & M ekong River Commission Lao PDR, 2005b) . It flows through ~2600 km of channels from the Golden Triangle to the China Sea at the Mekong Delta (Mekong River Commission (MRC), 2010; MRC & Mekong Riv er Commission Lao PDR, 2005a) . The LMB is characterized by flat - fertile lands that stretch over long distances and strong climatic gradients. Human development and a plethora of managed ecosystems co - evolve in the basin resulting in rapidly emerging gl obal issues such as land cover change, river regulation, and habitat loss (Pokhrel, Burbano, et al., 2018) . 14 Figure 1 . Dams in the Mekong River Basin. The red and black contour lines show the U pper and L ower Mekong Basin respectively breaking at the border between Laos, Myanmar and China. The background shows land use types and irrigated croplands obtained from (Salmon, Friedl, Frolking, Wisser, & D ouglas, 2015) . Dams are broken down into three categories (operational, proposed, and under construction) to highlight the number of dams that will become operational in the near future. The database for the dams was retrieved from the Research Program on Water, Land and Ecosystems (WLE), Greater Mekong . 15 Contrary to popular belief, the vast potential for hydropower that the Mekong River presents remains relatively unaltered (Grumbine & Xu, 2011; Nilsson, Reidy, Dynesius, & Revenga, 2 005) . Despite having many dams constructed over the past few decades, the effects are minimal because most are located in the tributaries ( Figure 1 ) and only capture a small portion of the river flow (Grumbine & Xu, 2011; Winemiller et al., 2016b) . By extent, the hydrology of the Mekong remains largely governed by natural flow variation. The river system at hand still manifest distinctive wet and dry seasons as direct outcomes of the precipitation seasonal variability, which supports highly product ive riverine ecological systems and agriculture. Food production in the LMB is heavily reliant on timely rainfall, seasonal flood pulse, fisheries and rivers. In fact, crops are grown on naturally fertilized soils from nutrient - rich sediments and flows tim ed with plentiful seasonal rainfall, while wetlands benefit from abundant freshwater and nutrients supplied by the seasonal flood (Fredén, 2011) . These characteristics allow the LMB house an important ecosystem responsible for the la rgest inland fishery that feeds the locals and a significant fraction of rest of the world (Ziv et al., 2012a) harvest of wild fish from the Mekong worth between $2.2 - 3.9 billion at first scale and $4.3 - 7.8 billion on retail markets (Hortle, 2009) expectedly went up and it continues to rise. Unfortunately, the ideal of the unchanged ecosystem that is the MRB is beginning to show measurable changes. For starters, widespread alterations in land use along from the construction of several large dams in the main stem of the river and hundreds of other in the tributaries. For instance, by 2030 there are 16 dams in the mainstream and ~110 dams in the tributaries are planned (Grumbine & Xu, 2011; Keskinen, Kummu, Käkönen, & Varis, 2012; Lauri et al., 2012a; Stone, 2011; Winemiller et al., 2016b; Ziv, Baran, Nam, Rodríguez - Iturbe, & Levin, 2012b) . Main land 16 use categories along with operational proposed, and under construction dam projects are shown in Figure 1 . Additional due to climate change. Crop stagnation and dramatically altered aquatic ecosystems deeply disrupt rural livelihoods and are a direct result of more frequent floods and droughts (Adamson, 2006; Fredén, 2011; MRC & Mekong River Commission Lao PDR, 2010a) . Another measurable change is the result of temperature rises in the headwaters of the Mekong River (Lauri et al., 2012a; Lutz, Immerzeel, Shrestha, & Bierkens, 2014) , which in turn alters the seasonality of stream flows, affecting agricultural productivity and aquatic ecosystems. Additional downstream pressures come in the form of groundwater overexploitation, and sea level rise, which leads to salt water intrusion and aquifer contamination. Overall, the falling quality of the ecosystem from the proliferation in dam construction, climate change, and sea level rise will likely presen t the LMB with unwanted changes in the hydrologic, agricultural, and aquatic systems (Johnston & Kummu, 2012; Lauri et al., 2012a; MRC & Mekong River Commission L ao PDR, 2005b; Nesbitt, Johnston, & Solieng, 2004) . As aforementioned, this thesis focuses on the hydrological causes and effects of a dietary change from fish to land - based meat protein in the LMB. Studies focused on food security of the region at han d show that due to dam construction and associated fishery loss (Stone, 2016) maintaining current levels of food supply would require 19 to 63% expansion of agricultural land (Orr, Pittock, Chapagain, & Dumaresq, 2012b) . Such an expansion would require substantial amounts of additional water, leading to unknown potential consequences in the LMB. Thus, this study explores the ram ifications of affected aquatic systems, and analyses the increased water usage from agriculture and new dietary demands of the LMB nations. 17 2.2 Data In its first analysis, fish downscale, the thesis uses: yearly fish catch historical data (1979 - 2010) from the calculated dataset from FAO FishStat ( http://www.fao.org/fishery/statistics/en ), model results data from CaMa - Flood (1979 - 2010) (Pokhrel, Shin, et al., 2018a) , and historical population data (1986 - 2010) from the International Monetary Fund (IMF), specifically the October 9 th , 2018 iteration ( https://www.imf.org/en/Data ). The CaMa - Flood results include discharge, storage, flood depth, and flooded area. The land use/land cover change historical variation inspection is mapped and calculated from European Space Agency - Climate Impact Initiative ( ESA - CCI: https://www.esa - landcover - cci.org/, accessed on 27 January 2018). For the seco nd analysis, meat and fish production and consumption, is carried out using: production FAOSTAT data ( http://www.fao.org/faostat/en/#data/QL ), supply FAOSTAT data ( http://www.fao.org/faostat/en/#data/CL ) for both meat and fish, and fish catch from the afor ementioned FAO FishStat in the place of fish production. The VWT network construction required detailed trade matrix meat (e.g., cattle, chicken, pork) data from FAOSTAT ( http://www.fao.org/faostat/en/#data/TM ) and VWC values retrieved from Chapagain & Hoe kstra (2003) results. 18 2.3 Hydrological and Flood Dynamics Model Settings The hydrodynamic model CaMa - Flood (version - 3.6) requires runoff and TWS components (e.g., soil moisture, snow, river storage, and groundwater) data as part of the input variables. These data are fed from an alternative model, HiGW - MAT, which ran for the sa me temporal scale (1979 - 2010) using the same parameters as in Pokhrel et. al. (2015) . HiGW - MAT is a global model; thus, its results were extracted for the MRB (90 - 110 E, 5 - 35 N) at a 1 grid resolution. CaMa - Flood is driven using runoff and the TWS analysis is c arried out using the storage components. CaMa - Flood was run using similar approaches to that of previous studies (Yamazaki, Sato, Kana e, Hirabayashi, & Bates, 2014a; Zhao et al., 2017) , if fact all of the simulation settings are identical to Yamazaki et. al . , (2014). River - floodplain hydrodynamics a t 1 over the MRB were simulated using runoff . Further simulation s are conducted for va rious year combinations including wet and dry year specific simulations. For the purpose of this thesis, the relevant resulting simulations from the mentioned study were conducted at varying degrees of dampened flood peak (i.e., by 10, 20, 30, 40, and 50%) . These scenarios were designed to capture peak flow reduction in magnitude that come with hydropower and flood - control dams. actually capture the flow regulation effects of future dams, they present plausible flow regulation ef fects from cumulative upstream dams . The hydrological data are results from the global hydrodynamic model CaMa - Flood are a crucial component of this study . This model has been extensively validated globally and over the MRB (Chaudhari, Felfelani, Shin, & Pokhrel, 2018; Felfelani, Wada, Longuevergne, & Pokhrel, 2017; Kim, Yeh, Oki, & Kanae, 2009; Pokhrel, Felfelani, Shin, Yamada, & Satoh, 2017; Pokhrel et al., 2015; Pokhrel, Hanasaki, Koirala, et al., 2012; Pokhrel, Hanasaki, Yeh, et al., 2012) . In short, the model computes river discharge, water level, and flooded areas by solving the shallow 19 water equation for open ch annel flow. For a complete model description and more detailed simulation settings refer to the 2018 Pokhrel et. al., scientific report (Pokhrel, Shin, et al., 2018a) . 20 3. METHODOLOGY This study takes the LMB (see Figure 1 ) to create a gridded map of estimated riverine fish catch. The distribution method follows McIntyre and colleagues (2016) methodology, deriving a relationship potential fish catch and hydrology. The distributed data per country is then merged and clipped for the LMB area in order to extract yearly fish catch data within the LMB itself . The results are mapped as total fish catch and fish catch per capita in timeline maps and dry and wet years respectively. The study goes further explores the measurable drop in fish catch on the study area by measuring the increase of alternative meat consumption by conducting a changepoint a nalysis on meat and fish production and consumption . A rise in water use to compensate for the elevated meat consumption is studied constructing a meat (e.g., cattle, chicken, pork) VWT network from 1988 to 2016. The VWC of each commodity considered was re results (2003) . Lastly, the study tak es a look at the RWS from meat by comparing years with higher virtual water inflow to the LMB with years of higher virtual water outflow from the LMB following (Dalin et al., 2012) water savings calculations. 21 3.1 Hydrological and Flood Dynamics Model This study employs results from t he global hydrodynamic model, CaMa - Flood (Yamazaki, Kanae, Kim, & Oki, 2011; Yamazaki, Sato, Kanae, Hirabayashi, & Bates, 2014b) . The model produces hydrological results (e.g., inundated area, water level, river discharge, flow velocity) by computing river hydrodynam ics solving the shallow water equation of open channel flow . Local inertial approximation in CaMa - Flood accounts for backwater effects explicitly (Yamazaki, de Almeida, & Bates, 2013) . The model was set up at a 10 km resolution with regional level settings for the MRB (Yamazaki et al., 2014b) . The Mekong Delta requires accounting of channel bifurcation to realistically sim u late river - floodplain dynamics , which is achieved by using CaMa - Flood version - 3.6 . The river network at 10 km resolution was obtained by upscaling the 3 arc - second ( 90 m ) flow direction map from HydroSHED S (Lehner, Verdin, & Jarvis, 2008) . The digital elevation model was obtained from SRTM3 DEM (Yamazaki et al., 2014b) . roughness coefficient for rivers w as set at 0.03 and for floodplains it was set at 0.10 following (Yamazaki et al., 2012, 2011, 2014b) . The rest of the parameters remain unchanged f rom Yamazaki et al., ( 2014) . Some of the critical output variables used in this study are flo od depth, flooded area, discharge, and water storage. Flood depth and flooded area are diagnosed from water storage in each unit catchment and discharge is calculated from the shallow water equation (Pokhrel, Shin, Lin, Yamazaki, & Qi, 2018b) . Water storage at each unit catchment , on the other hand, requires three components for its computation, discharge input from upstream, discharge output at the downstream , and local runoff. The discharge variables are obtained from CaMa - Flood, however, local runoff variable is an input from HiGW - MAT (global hydrological model ) (Pokhrel, Shin, et 22 al., 2018b) . T hese are the three components that update the mass conservation equation that updates water storage. 23 3.2 Creating a Gridded Map of Estimated Riverine Fish Catch The first step in downscaling fish catch data is gathering the necessary data. Mean annual catches of freshwater fishes from inland waters were calculated from FAO FishStat database (FAO, 2018) . The exclusive query term s for downloading catch data from fishing area respectively. Th ese specifications actively exclude fish from sea waters and aquaculture . The analysis laid out in Figure 2 focuses on four nations (i.e., Cambodia, Laos, Thailand, and Vietnam) with reliable data from 1979 - 2010. Data outside of this specified temporal period is available , however, the analysis is limited to the results of the CaMa - Flood simulat ion. 24 Figure 2 : Workflow diagram of the 10 km fish catch downscale applying a univariate power regression model. Flooded area data is retrieved from the CaMa - Flood model results while fish catch data is calculated from FAO FishStat (Pokhrel, Shin, et al., 2018a) . The second step entails establishing an empirical power - function scaling relationship b etween mean annual flooded area and fish catch. Fitting the univariate power regression model requires using reduced major axis regression to fit data from 4 nations . This process yields Equation 4 . Next, the study distributes fish catch into a 10 km grid system. Here, an exclusion of any grid below 10 cm is carried out using flood depth raster data from the simulation results . The exclusion follow s the assumption that large fisheries, which are the sole reporters of fish catch to the FAO FishStat database, fish in such shallow waters . 25 One of the limitations to the aforementioned downscaling method arises from the FAO statistics and the rooted assumptions . First, the analysis is limited because national governments submit catch statistics i ndependently to FAO. This can lead to an unwanted favoritism towards large rivers, population centers, and commercial fisheries (Ba rtley, De Graaf, Valbo - Jørgensen, & Marmulla, 2015; R. Welcomme & Winfield, 2012) . Second, f ish catch is famously underreported by 100 - 200% (Dickson, Hutton, & Adams, 2009; Nations & Center, 2008; R. Welcomme & Winfield, 2012) . Third, the statistics are confined by only one source of the comprehensive regional statistics on freshwater fisheries (McIntyre, Liermann, & Revenga, 2016) . Fourth, the main assumption for the downscaling algorithm lies in that the calibration of the model is equipped for large river basin scale distribution and so it assumes accurate capture at smaller scales. The last calculation carried out using fish catch data is the estimation of fish catch loss at each scen ario presented in Pokhrel et. al., ( 2018a) . This calculation feeds mean flooded area estimated at each scenario to Equation 4 resulting in a potential fish catch calculation of said area. Percent fish catch loss at each scenario is calculated from the baseline scenario and presented in Table 3 . 26 3.3 Virtual Water Trade Network The VWT network is built following a series of steps, hence, a set of regressive steps will follow. To obtain the VWT values we employed the following equation that multiplies trade volume of a specific commodity by the virtual water content of this commodity in the LMB country of ROW country of export: Equation 1 : Where is the local (Cambodia, Thailand, Laos, Vietnam) virtual water trade in volume of commodity exported from a LMB country to a ROW country through trade. is the virtual w ater content of commodity x produced in country . is the volume of commodity x produced in the LMB and exported from to . Similarly, we calculated VWT for the importing valu es with the following equation: Equation 2 : Where is the foreign (from ROW countries) virtual water trade in volume of commodity expo rted from a LMB country to a ROW country through trade. is the virtual water content of commodity x produced in foreign country . is the volume of commodity x produced in the ROW and exported from to . 27 As previously mentioned, VWC is the quotient of ET and yield (Y) of crops. However, the VWC of livestock products takes into consideration the production of their feed, and water consumption of the animal i tself. The feed consumed consists of two components including the virtual water embedder inside the various feed ingredients and the mixing water required to put the feed mix together. For greater details on how virtual water content from feed is calculate d refer to (Hoekstra, 2003) . This study takes the VWC values computed and reported in Chapagain and Hoekstra (2003) . By definition regional water savings ( RWS ) is the difference between the VWT of the importing region and the VWT of an exporting region (see Equation 3 ). Equation 3 : Where is the RWS in m 3 /year, is the and m 3 /year. 28 4. RESULTS AND DISCUSSION 4.1 Ca M a - Flood Model Validation and Results Figure 3 : Observation and simulation data for nine major dams in the Lower Mekong Basin. Observation data retrieved from the MRC . 29 4.2 Empirical Relationship between Flooded Area s and Fish Catch From the available fish catch datasets at FAO FishStat, data can be retrieved at a sub - basin or a nation maximum resolution . However, f or the purpose of this study, the available data for the Mekong River is at a nationwide resolution, allowing the acquisition of the total fish c atch per country per year. An objective for this study is to estimate the loss of fish catch in the LMB due to changes in hydrology . For this, the first step is to distribute the low - resolution fish catch data from FAO to a higher resolution grid system. Fish catch fraction that corresponds to the LMB from each LMB country can then be obtained from the high - resolut ion system derived from downscaling. This method allows the measurement to be more precise as opposed to obtaining fish catch values from neighboring basins that also cross th e LMB countries. A simple relationship between estimated riverine fish catch and a hydrological variable c an be valuable in the quest of downscaling observed fish catch data. As previously mentioned, a model with of these characteristics has been previously developed between fish catch and river discharge variables (C=0.3264Q 1.256 , R 2 =0.64) (McIntyre, Reidy Liermann, et al., 2016) . This mode l is intended for the purpose of its study , which is to cater for many river basins. D ischarge is selected as the input variable given that it is a widely available dat a and does not require model results for each specific basin. Here, a model simple principle is implemented specifically to the LMB. Additionally, the concept of using flooded area as a linkage to fres hwater fish catch distribution is based on the findings presented in the study by Sabo and colle a g u es where a multivariate model to characterize fish production in is constructed (Sabo et al., 2017b) . The flooded are a present in the individual grids within the basin is controlled by multiple variables (e.g., discharge , topography, precipitation), thus , allowing for an implicit cons ideration 30 of additional hydrological variables while only requiring one input variable for downscal ing purposes. An important and unique behavior that characterizes the Mekong river is its seasonal flood pulse, which creates large flooded areas near large water bodies at different periods throughout the year. Therefore, common catch locations were examined and it was found that most common freshwater catch species ( e.g. Notopterus notopterus , Channastriata , Cyclocheilischthys enoplus ) predominantly live in seasonally flooded areas (Kian, Yeap, Eong, Sensereivorth, & Racy, 2005) . These considerations led to an attempt in link ing fish catch to flooded area s by utilizing the CaMa - Flood Simulations . Figure 4 : Relationships between measured fish catch and modeled river flooded area for year s 1986 to 2010 . Fish c atch data are obtained from FAO FishStat and flooded area results are obtained from the CaMa - Flood model results (Pokhrel, Shin, et al., 2018b) . 31 The relationship between fish catch and flooded area (C = 0.6604A fl 1.071 , R 2= 0.718) is presented in Figure 4 . Based on the R - square d value and the data fit, this relationship is confidently more adequate for focusing on fish distribution in the flooded areas of the Mekong t han the general relationship dev eloped by McIntyre et al. (2016) . Additionally, it is important to note that a relatively longer span of data (1986 - 2010 ; See Table 5 in Appendix ) was used in deriving this relationship ( Equation 4 ) as opposed to the 1999 - 2008 period used by McIntyre et al. (2016) ( Equation 5 ) . Equation 4 : Equation 5 : Now that a relationship with a hydrological variable has been established ( Equation 4 : ), it is essential to test its performance. T he next objective is to downscale the fish catch data to 10 km grid s to compare the model developed here to that of McIntyre and colle a g u es . This was performed by downscaling the obtained fish catch data using both relationships shown above ( Equation 4 : & Equation 5 : ). The data used to downscale fish catch using the discharge relationship was obtained from Ca Ma - Flood results t o keep the input data consistent between the compari s on. Figure 5 shows a close match between the results obtained from the two relationships. This relationship is expected since discharge is considered a strong determinant of flooding and both variables are closely related in the described model . Furthermore , t he uncertainty in each variable, shown in plots B and C are virtually identical . 32 Figure 5 : Model fit validation through (A) comparison between regress ion fitted derived relationship between fish catch and discharge (Y - axis) from (McIntyre, Reidy Liermann, et al., 2016) , and the derived relationship between fish catch and flooded area (X - axis) outlined above. The time series (B) and (C) show the limitations of using the univariate models by plotting average fish catch of Cambodia, Thailand, Vietnam, and Laos with error b the derived model form this study respectively . Figure 5 shows that the catch distribution derived from fish catch and flooded area relationship is not significantly different from that of McIntyre and colle a g u es. However, the usefulness of this relationship stems from allowing catch distribution in areas with no discharge. For instance, large bodies of water that experience considerab le quantities of water storage, ergo, no inflow or outflow reported by the model output. To add further confidence to our newly derived fish catch - flooded area relationship, the study attempted to conduct an independent validation of our downscaled data wi th the observational data ( Mekong River Commission, 2002) of fish catch in the Tonle Sap Lake in Cambodia. The 33 aforementioned 2002 MRC report presents an average observed catch of 235,000 tonnes for the years 1995 - 1996, however , during these same years the FAO FishStat repo rts a total catch for Cambodia of 72,420 and 63,440 tonnes respectively , averaging 67,930 tonnes for these two years. A fter downscaling and fish catch data distribution takes place, the total Tonle Sap Lake fish catch is attained to be 44 , 864 and 51 , 988 tonnes for the years of 1995 and 1996 respectively. This results from the FAO data only using officially reported fish ca tch data from large fisheries, although most of the fish catch in the region is ob tained from single Cambodian fishermen who typically do not report catch statistics. The River Commission report, hence, estimates the total amount assuming most of the fish catch comes from individual fishermen. A third validation form was carried out usi ng observation data for four years (2000 - 2004) in Cambodia and comparing it to that of the resulting yearly data aggregated in Cambodia (Table 2) . The percent difference in these results presented in Table 2 , reach a high of 0.16 %, meaning that the estimated catch values are profoundly reliable in this case. This comparison serves as independent confirmation that the methods behind catch distribution are adequate for large scale distribution. The limitations explained in the methodology section, though, still need to be considered. In short, for smaller scale downscaling, the overarching assumptions are biased towards large scale distribution. Meaning that the finer the scale, the less reliable the method of fish catch distribution becomes. 34 Table 2 : Comparison between observation fish catch (Kian et al., 2005) and distributed fish catch in Cambodia using Equation 4 : between the years 2000 and 2004 . Year Total Production, Observed (Mt) Total Production, Estimated (Mt) Difference (Mt) Difference (%) 2000 245,600 245,300 300 0.12 2001 385,000 384,500 500 0.13 2002 360,300 359,800 500 0.14 2003 308,750 308,250 500 0.16 2004 249,600 250,000 - 400 - 0.16 35 4.3 Fish Catch Downscale There are three river basins in the area that Thailand, Laos, Cambodia and Vietnam conceal. These three basins and the total area of LMB countries were accounted for in the CaMa - Flood model. Here, a downscale approach was developed using the CaMa - Flood model, in which freshwater fi sh catch data is distributed among all the possible main river stems, tributaries, lakes, and floods inside of the LMB countries. The temporal period for this downscaling was set for 1979 - 2010 , with a yearly timestep . Th is temporal period is limited by the years that the CaMa - Flood model was run (1979 - 2010) even though the fish catch data was available for a longer period (i.e., 1959 to 2017 ). However, it is important to note that older fish catch data could be less reliable given that there is a higher prevalence of gaps and inaccurate estimated data points . On e of the restrictions of the downscale carried out in this study was to eliminate grids with average water levels below 20 cm. This was achieved by using additional flood depth output data from CaMa - Flood. Water level was limited to above 20 cm since catch data is looking at large fish caught by large - scale fisher ies and it would be inefficient for them to focus on shallow areas, thus distrib ution of fish in shallow waters would be in appropriate. T he results obtained at 10 km grid scale (using Equation 4 ) are presented in three figures ( Figure 6 , Figure 7 , and Figure 8 ) . 36 Figure 6 : Fish catch distribution map as eight - year averages from 1979 to 2010. Freshwater fish catch data from FAO FishStat was distributed using to the 10 km grid s using Equation 4 based on the derived relationship of catch and flooded area. Grids with flood depths lower than 20 cm where masked as the fish of interest are located in deeper waters. Flooded area and depths data are taken directly from the CaMa - Flood model results (Pokhrel, Shin, et al., 2018b) . Country borders are delineated in black. LMB border is delineated in red. The higher concentrations of total fish catch are shown in dark blue, while lower concentrations are in light blue. 37 The total downs caled fish catch is presented in Figure 6 . Firstly, the data shows a finite increase in total fish catch through out the years. This is apparent at every sp atial point presented in the maps, however, the areas of highest concentration (e.g. T he Tonle Sap Lake ) show the most obvious increase in fish catch. It is important to present these data as an average of multiple years since year to year flood variation is significant; that is, it is important to minimize the effect of possible bias during extremely wet and dry years. One of the hypotheses behind the fish catch downscale relationship ( Equation 4 : ) was that the higher the prevalence of flooded areas, the higher the fish catch concentrations would be detected in the LMB. To test this hypothesis, four dry ( Figure 7 ) and four wet ( Figure 8 ) years were mapped . 38 Figure 7 : Fish catch per capita distribution on four dry years: 1998, 2003, 2004, and 2005. Catch per capita distribution is used to normalize the effect of fish catch increases due to population demands. Population data is retrieved from IMF ( October 9 th , 2018) . 39 The main driv er of fish catch increase through out the years is population growth, so, in an effort to normalize yearly data this study divides the observ ed annual catch data by the population for the corresponding year . This allows for downscaling of fish catch per cap ita and a more appropriate catch comparison with hydrology by eliminating a strong driver of fish catch fluctuations ( i.e. population growth). The temporal period for these downscal ing is 1986 to 2010 , which is determined by the availability of population data reported by the International Monetary Fund ( IMF ). The downscaling of Figure 7 and Figure 8 was carried out with fish catch per capita data instead of total fish catch data which can be seen in Figure 6 . The dry years chosen for Figure 7 are 1998, 2003, 2004, and 2005. Earlier years are not included in these maps ( Figure 7 & Figure 8 ) because fish catch data had a considerable increase at the end of the 20 th century due to the technological advances , thus prior years do not demonstrate a fair comparison. This provides an alternative way of reducing the influence of external variables to hydrology. T he years chosen for Figure 8 are 2000, 2001, 2006, and 2007. All of these eight years fall in a span of nine years (1998 - 2007) where external variables do not play a major role in affecting the catch dynamics due to the As ian financial crisis ending in 1998 . After this crisis was subsided, no major technological fishing advances are reported, and population growth is no longer an issue. 40 Figure 8 : Fish catch per capita distribution on four wet year s: 2000, 2001, 2006, and 2007. Catch per capita distribution is used to normalize the effect of fish catch increases due to population demands. Population data is retrieved from IMF ( October 9 th 2018 ) . 41 Figure 8 shows a denser distribution of fish in the LMB and the outer basin streams compared to that of the dry years catch distribution map ( Figure 7 ). This is especially noticeable between latitude 12 N and 13 N in Thailand, and inside of the LMB with the exception of the Tonle Sap Lake. The unnoticeable differences in the Tonle Sap are exp ected as this body of water is a suitable place for fishing for a wide range of water levels because of its large size and its ability to become full up to a sufficient amount to hos t plenty of fish every year. An important distinction is that for the year s closer to present time (e.g. 2007 versus 2005 ) variations are harder to come by as compared to past years (e.g. 1998 versus 2000 ) . Overall, the fish catch per capita appears to be strongly affected by the presence or lack of flooded areas in the LMB. As previously established, dam construction in the Mekong River and its tributaries have generally reduce d the peak of the flood pulse and normalize d the seasonal variation of the hydrograph (Kummu & Sarkkula, 2008; Lauri et al., 2012a; Pokhrel, Shin, et al., 2018a) . An a dditional effect of dam construction is the reduction of discharge, which consequently affects the flooded areas in the basin. In fact, in a recent s tudy (Pokhrel, Shin, et al., 2018a) , the effects of reduced discharge in the LMB were computed using hypothetical scenarios w ere d ischarge was restricted to an upstream point in the LMB . These results are next used to calculate potential fish catch losses from the response variability of flooded areas. 42 4.4 Fish Catch Loss The 2018 study by Pokhrel and colle a g u es present s five scenarios in which water flow is restricted at the (marker) in the map below. The flow percent flow r estrictions range from 10 to 50% at a 10 % step count. The resulting hydrological responses at each scenario are modeled using the CaMa - Flood hydrolog ical model. The 2018 study by Pokhrel and colle a g u es present five scenarios in which flow is restricted at the (marker) in Figure 9 . The flow percent flow restrictions ran ge from 10 to 50 % at a 10 % step count. The resulting hydrological responses at each scenario are modeled using the CaMa - Flood hydrological model . 43 Figure 9 : L MB CaMa - Flood simulation domain red box encloses the L ower Mekong sub - are a , while green box encloses the Tonle Sap Lake area. 44 Figure 9 shows the study area used to calculate the hypothetical scenarios at hand . The mean monthly flooded area results of this specific area are used to calculate the potential fish catch loss at each scenario. This is computed employing Equation 4 to calculate the potential fish catch at each scenario. Then , the percent loss is calculated from the potential fish catch of the baseline scenario. Figure 10 : Estimated fish catch loss for the Lower Mekong Basin (A) and for the Tonle Sap Lake (B). Monthly calculated potential fish catch loss in the total Lower Mekong study area presented by Figure 9 is shown here . Every month in the bar graph has a different bar with a light blue tone representing each scenario (baseline to 50 % outflow restriction) (A) . Monthly calculated potential fish c atch loss in the Tonle Sap Lake shown in orange (B) . The estimation is projected using Equation 4 on the CaMa - Flood flooded area results at each of the sce narios presented from (Pokhrel, Shin, et al., 2018a) . 45 Reduction of seasonal discharge variability shows a clear ly reduced but constant effect of the discharge that dams provide . This effect can be observed in Figure 10 , as potential fish catch reduction is present dur ing the months of January to May, and then we encounter a rise in potential fish catch from months Jun e to December as we move down t hrough the scenarios presented in both graphs shown above. Therefore , focusing on the Tonle Sap Lake , a yearly result fish catch response to the reduced discharge scenarios is summarized in Table 3 . Table 3 : Tonle Sap Lake summary of estimated fish catc h loss for each outflow restriction scenario presented above ( Figure 10 ). Scenario Baseline 10% 20% 30% 40% 50% Potential Catch (Mt) 91865 90305 89135 87374 86123 84741 % Loss 0.0 1.7 3.0 4.9 6.3 7.8 Total Loss (Mt) 0.0 1559 2730 4491 5742 7124 The modest scenario (10 % o utflow restriction) shows a 1.7 % reduction of fish catch. Translating to fish catch in mega tonnes the result is a 1,559 Mt reduction of f ish catch. At the 50 % outf low restriction we obtain a 7.8 % fish catch loss or 7 , 124 Mt reduction. Und oubtedly, these hypothetical fish catch losses would be devastating to the livelihood of the surrounding LMB communities. But from literature, it is known that this is starting to occur , and a major dietary shift has already been taking place in the LMB co untries (Orr et al., 2012b; Pittock et al., 2017) . Therefore , the next question becomes, what are the consequences at a land use/land cover and water usage level? The following sections will cover results answering this question. 46 4.5 LMB Land Use Change Land use changes have become easy to track with satellite data as a response to the amount of fish catch lost per capita in the past few years and an increase in the standard of living across social strata at LMB developing countries are experiencing. Looking at Figure 11 a two percent increase in cropland is reported from 1992 to 2015. We also observe a two percent rise in mosaic tree and shrub/herbace ous cover. Most of the land used to compensate for the rise in cropland and other land use types comes from tree cover, with a drop of four percent in 23 years. To put that into perspective, a one percent change in land use is equivalent to 60,600 km 2 , meaning that a total of 242,400 km 2 of rainforest was lost in the past two decades. Figure 11 : Land use and land cover types for 1992, 2000, 2008, and 2015. The embedded pie charts show the fraction of the main land cover types w ith matching color coding except for the black pieces which are a lumped representation of the minor land use types. The urban areas are excluded from the pie charts since they represent a negligible amount of area. Data source: European Space Agency - Clima te Impact Initiative ( ESA - CCI: https://www.esa - landcover - cci.org/, accessed on 27 January 2018). 47 The majority of the cropland is used to grow crops such as maize, soy, and rice. Maize and soy are primarily consumed by livestock. In the United States , for example , 87% of the maize yields are evenly split towards animal feed and ethanol production, leaving only 13% for human consumption (Ranum, Peña - Rosas, & Garcia - Casal, 2014) . An FAO report calculated at least 50% of the grain we grow is fed to livestock ( He & Food and Agriculture Org anization of the United Nations, 2004) . Regarding land use, this translates to 75% of all agricultural land, including crop and pasture , being dedicated to animal production (Foley et al., 2011) . As fish consumption per capita d eclines in the LMB countries, the loss of protein will eventually be replaced by land animal protein. Local production of meat in the LMB countries continues to rise (see Figure 12 ). This requires more feed and grassing lands for cows, pigs, chickens, and other animal products. Land use change requirements have been calculated for various scenarios and show a range of 16. 5 % to 55% of agricultural land increase in order to compensate for the loss of fish protein (Pittock et al., 2017) . This accounts for a large amount by itself, however , this calculation fails to include the amount of feed that the livestock animals require to live , which would translate to even more land being used . Land use change not only compromises the ecosystems and biodiversity of the LMB but also rises the water footprint of the entire region as crops require irrigation efforts. The question now becomes, where will all this water come from to compensate for the fish protein loss? But before this can be answer ed, there is a need to look at the dynamics of meat and fish production and consumption to understand how dietary habits are changing over time for the countries of the LMB. 48 4.6 Meat Production and Consumption Production of meat commodities in the LMB countries grew at a fast pace, presumably at similar rates from that of population and GDP in the past two decades. Figure 12 shows the difference between production and consumption of meat for the populations of Thailand, Cambodia, Vietnam, and Laos. As a whole, these four countries show a growth of production and co nsumption at similar rates. 49 Figure 12 : Time series comparing meat production and consumption between the years 1988 and 2013. The meat products (i.e., cattle, pig, sheep, goat, chicken, and horse ) . Additionally, a statistical mea n changepoint is calculated and represented by the vertical lines on each graph (color coded accordingly). The consumption of the commodity at hand is taken directly from the supply data at FAOSTAT . The production of meat above is taken from the production data at FAOSTAT . 50 At the statistical changepoint s for production and consumption , Thailand shows a growing gap between the production and consumption lines. In other words, Thailand produces much more meat than what its population consumes from 1992 to 2003. the introduction of evaporative - cooling and poultry taking over the livestock market share in early years (1998 - 2001) rising from 30% to 53% (Costales, 2004) . This is reflected on the LMB plot of Figure 12 , where we see statistical changepoints calculated for production and consumption in year s 2003 and 2005 respectively. Meaning the production and consumption growth of the LMB changes its pace between these two years. As th e meat ind ustry kept growing in Thailand, it moved away from contract farming and it safety and animal warfare requirements. Year 2004, the highly pathogenic avian influenza (HPAI), outbreaks and results in a ban by most importers of frozen broiler meat from Thailand, arguably their cash - cow livestock product (NaRanong, n.d.) . T he Thailand plot of Figure 12 , shows sudden large drop in production from 2003 to 2004 narrow ing the gap between production and consumption , which represent s a reduced opportunity for meat exportation. This behavior is in c ongruence with the aforementioned HPAI outbreak. There is a slow recovery from this massive drop in production for the next decade and by the year 2013 the large gap between production and consumption form s again with production surpassing consumption sign ificantly. Cambodia in Figure 12 is characterized by their smaller meat production market at an order of magnitude less than that of Thailand and Vietnam. Meaning that changes in production and consumption will likely not influence the overall trend of the LMB a significant amount. However, that Cambodia has nearly no gaps between production and consumption except for years 2005, 2006, 2008, and 2009. This is relevant because it explains the lack of data 51 in terms of importation and exportation of meat commodities during the VWT analysis of section 4.8 . The statistically calculated changepoint of Cambodia falls on year 1998, which is likely linked to the Asian financial crisis of 1997 - 1998, whe n the economic growth in general took a hit that year in the region (Costales, 2004) . A large contributor to the overall LMB countrie production and consumption of meat is Vietnam. The Vietnam plot of Figure 12 shows a slow rise in both production and consum ption between the years of 1988 to 2003. We see equivalent levels of production and consumption during this period of time, suggesting little to none import or export of meat, more on this in the VWT analys i s o f section 4.8 . In y ear 2003, Vietnam shows a statistically calculated point where production growth takes a hit. This is also explained by the HPAI outbreak and by additional low standards of food safety put forwa rd by the Vietnamese meat industry . Such standards of food safety affect the population of Vietnam to this date with hundreds of cases of death by food born disease every year (Nguyen - Viet, Tuyet - Hanh, Unger, Dang - Xuan, & Grace, 2017) . On the flip side, in year 2005 , following large economic growth s , Vietn am displays a boom in meat consumption despite the local underproduction of the commodity (Hansen, 2018) . This is corroborated by the statistically calculated changepoint of consumption which falls in 2005 . Vietnam a n increasing gap between consumption and production between the years 2007 and 2013 , whic h must be explained by large meat importations. Lastly, Laos plot in Figure 12 , makes up a very small portion of the overall growth of meat of the LMB countries, with a n under production difference of two orders of magnitude compared to that of Thailand and Vietnam. The consumption and production of meat for this small nation grows simultaneously and at the same pace . This trend will reflect the subtle imports and exports of meat in section 4.8 . The calculated changepoint s for this n ation falls on 2002 for both production and 52 consumption. This point is relevant because it shows a recuperation of growth after the significant production and consumption of meat drop in 1999 - 2000 likely correlated to the Asian financial crisis of 1997 - 199 8 (Costales, 2004) . 53 4.7 Fish Production and Consumption A steady increase in t he production of the fish commodity is seen for the LMB countries over time . Fish p roduction and consumption variables moved closer together on years from 1988 to 1994. The consumption of the commodity at hand becomes greater by four - f old by the end of 2013. This phenomenon is quantified in the FAO databases due to fish catch being grossly underreported given that a big portion of catch comes from local family catch and single fishers inland. There is no accurate system at place for rep orting fish catch, and only large fisheries report to FAO (Sverdrup - Jensen, 2002) . This trend, however, is the main focus of interest in this section. Here, the discussion of fish production and c onsumption will be analyzed in more detail, not focusing solely on total values, hence, the following discussion based on Figure 13 . 54 Figure 13 : Time series comparing fish production and consumption between the years 1988 and 2013. The fish catch products are exclusively freshwater fish in order to isolate fish from rivers and lakes. Additionally, a statistical mean cha ngepoint for each variable is calculated and r epresented by the vertical line on each graph. The consumption of fish is taken directly from the supply data at FAOSTAT . The production of fish above is taken from the catch data at FAOSTAT . 55 In 2005 , Figure 13 shows a noticeable leveling of fish consumption. This can be attributed to three main catalyzers; F irst, , a concept coined to explain the increase in livestock consumption from developing countries (C. Delgado, Rosegrant, Steinfeld, Ehui, & Courbois, n. d.) . By the livestock revolution, as the economy of a country expands, its population gravitates towards an increased consumption of livestock products . This concept can be seen in Southeast Asia where a total meat consumption change from 4 to 16 millio n metric tons (1983 - 2020) was projected (C. Delgado et al., n.d.) . Second, with accelerated construction of dams, 40 - 70% of fish species in the Mekong are now faced with incapacitated migration (Barlow et al., 2008) , meaning that higher placement of dams leads to higher physical barriers presented for mi gratory species. Additionally, hydrological changes such as a decrease d seasonal discharge variation pose a threat to fish catch dist ribution, which ultimately results in tampering of their reproductive cycles. Third , the export ation of fish is highly lucr ative for large fisheries and preferred over selling for local consumption. In fact, 1.6 million tonnes of fish were exported in 2009 from Thailand alone (FAOSTAT, 2009) . Figure 13 demonstrates that initially Thailand has a higher consumption rate when compared to production rate of fish. However, in 1992 it shows decline in calculated statistical changepoint production , but from then on, fish production remains approximately the same until year 2012 . The shift in 1992 occurs around the time efficient fishing technologies were emergi ng in Southeast Asia . In the year 1995 the calculated changepoint for consumption highlights a sud den change in rate of increase, which then a sustained increase in consumption is se en until the year 2007 when consumption stalls and drops a small amount. At this time , meat consumption (see Figure 12 ) star ts to increase after three yea rs of decreased production and consumption. 56 Shifting the focus to Cambodia, Figure 13 shows a low production and consumption of fish up to year 1998, when the Asian financial crisis was ending, and better fishing technology was ise significantly and the consumption follows it very closely. From 2001 to 2013, fish consumption fluct uates alongside production starting with a changepoint detected near 2000 were a slow rise f ollowed by a steep production r ise is seen moving together is an indication that there is little to none impor t and export of freshwater fish catch. However, Tonle Sap Lake is also , which indicates that most of the fish catch originates from this body of water. Notably, a decrease in fish catch and consu mption is observed in dry years, for example in 2003 and 2004 (see Figure 13 ). Congruently, wet years such as 2006 and 2007 (see Figure 13 ) are very productive in terms of fish catch . Vietnam is similar to Thailand in regard to the consumption and production rates being close together for years 1988 to 1993. From this timepoint on, production drops even further while consumption of fish starts to quickly r ise. In 1995 , a changepoint for production is encountered , indicating a change in productivit y. From there fish production recovers to the same level of the starting years and main tains those numbers through 2013. On the oth er hand, consumption steadily r ises from 1995 to 2003 where the consumption changepoint occurs , which highlights a steeper climb in consumption. This is close to the year of the HPAI outbreak, where a r ise in oth er sources of calories replacing that of poultry is expected in the Southeastern Asian diet. Laos follows a similar trend to that of Thailand and Vietnam, but at a much lower scale. For the first ten years (1988 - 1998) the production and consumption of fi sh is an entire order of magnitude smaller than the rest of the LMB countries. Th is is expected, however, since it is the smallest country in the LMB . In 1998, when the Asian financial crisis ends, we get a calculated 57 statistical changepoint on fish catch. Moving forward, the fish catch production continues to r ise very slightly up to year 2013. In 1999, there is a consumption changepoint detected where a steep raise in fish consumption can be observed . Presumably this is due to an economic recuperation aft er the financial crisis. Now, with a stronger understanding of the economical dynamics of the two main sources of protein for the LMB countries , meat and fish , the first analysis on the impact that these commodities have on water can be explored . As prev iously mentioned, every commodity has a specific water footprint, which varies depending on the production location and time. Since the water footprint of fish products is not comparable to that of livestock (green vs blue water), v irtual water flows of me at products will be investigated exclusively. These are the main products that are progressively replacing fish protein calories in the LMB and will continue to do so in the foreseeable future. 58 4.8 Virtual Water Flows The concept of virtual water trade shows the amount of water that flows from one point to the other in the form of a commodity at a given time period. The water allocated towards a specific commodity under this calculation includes production, manufacturing, processing, and transportation water usage attached to it. Understanding the amount of virtual water that flows in and out of nations is crucial when optimizing water use at regional and global or even local levels. A complete analysis of this concept includes the calculation of virtual water exports as well as virtual water imports. For the purpose of this study VWF of meat products including, cattle, pig, sheep , goat, chicken, and horse, are explored and summarized in Figure 14 . 59 Figure 14 : Time series shows the virtual water imports, ROW to LMB (green line) , and virtual water exports, LMB to ROW ( blue line) , through the years. Data shown is derived from VWT calculations (VWC data retrieved from A.K. Chapagain & Hoekstra (2003) , and trade data from FAOSTAT detailed trade matrices . T he LMB countries VWF exportations are mostly driven by the flow fluctuations of Thailand from year 1988 to around 2005, 2006 when Cambodia, Vietnam, and Laos start to contribute at an 60 equal level on the import side of VWF ( Figure 14 ) . The above discussion on meat production and consumption explains why this is the case. In short, Thailand sits at higher production levels than that of consumption (see Figure 12 ). Ergo, the gap between production and consumption put forward by Thailand is equal to the amount of meat products available for exportation. The same process is true for the three smaller economies remaining, except that in their case, the consumption grows at a higher rate than production, thus, importations of meat are reported to those nations. Changepoint statistical analyses where only carried ou t for the LMB countries section of Figure 14 because there are missing data points at varying time steps of the four individual countries. The statistically calculated changepoint corresponding to VWF exports fall in year 2003. This is the same year that Thailand takes a hit from the HPAI outburst and their chicken production and trade plummets. This point serves as independent confirmation for the validity of the observational data as well as the methodology behind the VWF calculations. The second calculated changepoint corresponds to VWF inflow, and it falls in the year 2009. This is in harmony with the raising spikes in meat consumption of Vietn am, Cambodia. We do observe a r production and consumption of meat products, where an equivalent production and consumption fluctuations is observed . Thai expected near zero inflow of virtual water until year 2013. From this year on, data is blinded by the lack of meat production and consumption data (see vertical dotted line on Figure 14 ). Through assumptions, some of the meat products that are starting to be imported to Thailand are not part of their main production commodities. Meaning that the m eat category most likely falls to sheep, goat, or horse. Year 2014 shows a recuperating 61 r ise virtual water outflow combined with a drop of virtual water inflow. This tells that that local production is picking up again and the country is able to sustain th e local and global demand of meat products (more on this in section 4.9 ). The virtual water outflow of Thailand stands out the most, as it follows a strong increase in meat exportation up to year 2003, where it drops dramatically, only to be able to recover a decade later (2013). While this drop is in harmony with the decline in production ( Figure 12 ), the production graph shows a quick recovery of three years. The VWF however, demonstrate a much harsher consequence. This is due to chicken being the most affected product, which was their main exportation product, and one of the most water resource intensive. The other products such as cattle and pig pick up due the lack of chicken production and thus, a fast recovery on the production side is seen . As previously mentioned, literature on this phenomenon points to the HPAI o utburst as the main cause for the drop in VWF (outflow from chicken). However, literature fails to explain why the recovery process lasts for ten years. Speculations about this strugg le point to three possibilities; First, the loss business relationships w ith partnering natio ns; Second, extended contracts being formed between former partner nations and their replacing suppliers; And third, the inability of the meat industry to overcome the financial burden of losing such massive amounts of products and rece iving nothing in return. In the case of Cambodia an expected close to z ero inflow of virtual water is shown in Figure 14 from years 1988 to 2007. From 2 007 to 2016 an accelerated rate of meat importation is observed . This coincided with the inflow changepoint analysis calculated at year 2008 for the LMB countries confirming. In Figure 12 these are the same years (2007 - 2013) in which a gap is created putting consumption over production. Anothe r point to highlight is that outflow data is only available for years 2005, 2011, 2014, and 2015 . Years 2005 and 2011 are the only years where 62 production exceed consumption of meat products ( Figure 12 ) . This, again, acts as independent validation for the methodology behind VWT calculations and the trade data retrieved. Similarly, this is the case for Vietnam with a near - zero inflow and outflow reported until 2007, when their meat consumption increased a t a faster rate than production, opening a gap of importation for meat products. The observed r ise in meat imports also gently coincides with the and then drops , which coincides with the ye ars when we a gap between meat consu mption and production gets smaller a significant amount of outflow is then not reported between 1988 and 2013. This instills Finally, Laos presents sparse data for its outflow of virtual data. Years 1990, 1993, 2000, 2001, 2002 and 2015 are the only years with reported exportation of meet. None of these years stand out when looking at Figure 12 , in fact, all of the production markers fall inside of the consumption markers. Admittedly, th e years with reported outflow in Figure 14 still possible that the difference between production and consumption is just outside of the naked ounts of inflow until year 2009, when the inflow of virtual water seems to skyrocket by triplicating its inflow. It is important to keep in mind, however, that at the scale that Laos operates this is still a small change compared to that of any of the othe r three nations. Nonetheless, in comparison to itself in the production and consumption side of the graph this is significant and the reliability of the trade data for this nation becomes questionable. To further investigate the VWT of meat commodities in the LMB countries, a VWT network is constructed and presented in the next section. This will highlight the trade dynamics of the LMB 63 countries with the rest of the world. It will provide a deeper understanding of the effects of a dietary shift of the four nations at focus onto the rest of the world. Ultimately, the response of the LMB will be visualized . 64 4.9 Virtual Water Trade Network The following VWT circle plots on Figure 15 represent the magnitude and direction of either exported o r imported virtual water from meat products (e.g., cattle, chicken, pork). Figure 15 . VWT flows of animal - based protein products (cattle, pig, sheep, goat, chicken, and horse) between 1988 and 2016. The width of each band represents quantity of water in traded. The LMB basin countries are represented with green bands while the R est of A sia and other continental regions have individual colors. The circular figure areas are scales to the total area traded. Data retrieved from A.K. Chapagain and Hoekstra (2003) and FAOSTAT detailed trade matrices . This figure was created using the networ k visualization too l, Circos (Krzywi nski et al., 2009) . 65 C ircle plot A shows that the majority of the virtual water outflow was carried out by Thailand (green) and was driven towards the R est of Asia (red) (i.e., the remaining Asian countries excluding Thailand, Cambodia, Vietnam, and Laos ) in 1988 . From circle A it seems as though none of the other regions trade meat products . However, due to the nature of these plots, the thickness of the arrows represents the shear amount of flow carried out compared to that of the regions. In other word large compared to that of the other regions, that the trade among the rest of the world becomes negligible , thus, very narrow . Circles A, C, F show noticeably large interactions between Thailand and the Rest of Asia , this is bec ause throughout history, Japan has been the main recipient of chicken/ poultry from Thailand (Costales, 2004) . Moving forward to year 1995, circle plot B, shows a dramatic increase in intercontinental virtual water flows when Vietnam becomes a large importer from the Rest of Asia, the Americas, and Europe. merely alludes that the other regions with thicker export bands became larger virtual water exporters than Thailand. In 2004 (circle C), one year after the aforementioned HPAI chicken export reduced to such a significant extent that its virtual water export is on par to that of Vietnam . By the year 2008, we see Thailand beginning to recover by starting to export their meat products to neighboring nations that are part of the LM B. This finally explains why Thailand in Figure 14 ny signs of recovery while Figure 12 still shows high levels of production. Vietnams increase in meat consumption is highlighted in year Circle D as it increases its virtual water i nflow amount s becoming the highest LMB country im porter of meat. Year 2012, Circle E, we see the first noticeable interaction virtual water importer. Thailand also shows signs of economic recovery by branching out its trade to countries regions like Africa and Europ e. The last circle plot of Figure 15 , circle F, shows Thailand full 66 trading recovery. Though, not noticeable in the circle, Thailand in fact returned to exporting levels almost comparable to those of year 1988. More on this in section 4.10 . 4.10 Regional Water Savings The amount of virtual water exported on imported to a region is an important marker to explore. However, when t hinking about constructing a trade system that is characterized by water use efficiency, we need to explore water savings (WS) . As mentioned before, WS is the difference between the total VWT outflow minus the total VWT inflow. This calculation can be carried out at a global, regional, nation, or even sub - national level. For the purpose of this study we conduct a regional water savings (RWS) analysis and present the results in Figure 16 and Table 4 . 67 Figure 16 : Water savings time series showing the difference between VWT outflow and VWT inflow using data from VWT network results . The WS timeseries , Figure 16 , shows a drastic turn of events through the years . Between the years 1988 and 2003, the LMB countries were collectively virtual water outflow dominant. The peak outflow in th is period of time was 1044 Mm 3 /year. Most of the measurable virtual water outflow can be attributed to massive chicken exportation. But the virtual water flow balance presented in Figure 16 reaches those high levels because there are barely no virtual water 68 flow imports from the other LMB nations to counterbalance the scale. Both of these phenomena translate to massive amounts of negative regional water savings for the Mekong (i.e., much water is used in the Mekong instead of saved through importation). Upon the previously discussed HPAI outbreak, the tables turn in 2004. In fact, between the years of 2007 and 2015, the LMB nations experience positive RWS collectively. Their WS peak in the year 2011, probably due to the massive amounts of meat product imports to V ietnam. And while, Thailand still exports virtual water, in the form of meat, it does so to other LMB nations such as Vietnam and Laos during this period of time. In 2016, Thailand seems to recover its previous trade relationships and begins driving large exportations of virtual water to the ROW turning the tables on RWS once again. Table 4 : Summary of w ater savings results in Mm 3 /year. Year VWT, Local VWT, Inflow VWT, Outflow WS 1988 0 1 494 493 1989 0 1 522 521 1990 0 2 624 622 1991 0 4 791 787 1992 0 2 839 837 1993 0 1 715 713 1994 0 2 691 689 1995 0 1 664 664 1996 0 1 586 586 1997 0 2 552 549 1998 0 1 743 742 1999 0 2 760 758 69 Table 4 ( c 2000 1 2 747 745 2001 1 3 900 897 2002 2 12 1056 1044 2003 3 4 1030 1026 2004 18 22 71 48 2005 3 6 40 34 2006 28 38 32 - 6 2007 100 65 42 - 23 2008 128 91 27 - 64 2009 86 131 48 - 83 2010 115 411 44 - 368 2011 190 816 34 - 781 2012 226 520 88 - 433 2013 291 478 81 - 397 2014 330 554 290 - 264 2015 269 624 532 - 92 2016 338 343 644 301 A scenario like that of year 2011 in Table 4 is probably the ideal scenario for the LMB in terms of water usage. Meat production represents a large water footprint on a region relative to other means of production and food security. The p roblem with making a broad declaration like The end goal is to reduce water footprint in 70 the regions with water scarcity problems. So, we need to analyze global water savings (GWS) and water scarcity in dices to do this, which falls outside of the scope for this thesis. Nonetheless, RWS points in the right direction of action. The amount of water needed to produce enough meat to sustain the LMB can be easily achieved at a balanced of inflow to outflow rat io, like that of year 2014 which still presents 92 Mm 3 /year of positive RWS. From the limited information that RWS provides, striving for the maximum amount of positive RWS without affecting the economy (i.e., small virtual water outflows) is the most sens ible thing to do. 71 5. CONCLUDING REMARKS The thesis studies two overarching topics, the first is an exploration of fish catch dynamics in the LMB and the second is a construction of the VWT system relating to meat products from the LMB countries. The developed relationship between fish catch and flooded area suggests that there is a correlation between flood and fish catch upon i solating fish catch years variation for dry and we years individually. This relationship is studied by isolating both variables as best as possible by ridding the trend of population growth. Ultimately , the overall trend and literature point to a significa nt transition of protein source from fish to land - based meat products . This opens the door to an extensive analysis in meat production and consumption as well as fish production and consumption. Plenty of conclusions can be drawn from these results, howeve r, the most important being that the so - called livestock revolution, which comes wit h the growth of the LMB developing economies, already causes measurable effects on the local ecosystem. This leads to the secondary set of analyses. To test for the extent of the impact from the studied dietary shift, land use is mapped for four separate years from 1992 o 2016 with 8 years between each plot. From these maps, the results show major shifts from tree cover and other native species to cropland. This is a direct confirmation of population and economical growth. Literature points at livestock production as most of the crops grown are meant for animal feed. The water resources impact of the dietary shift is further studied by constructing various VWT networks from 1 988 to 2016. Six main VWT network plots are presented in along with a discussion on each one, which explain the LMB countries trading dynamics with the ROW. This leads to a RWS short calculation and discussion, which sheds some light on the topic of constr interest to import most of its meat products as this would be beneficial to the ecosystem by saving 72 the precious resource that is water. However, two problems arise from this conclus ion. First, this does not consider the economic ramifications of exclusively importing these products. Second, th e water savings echoed at a global scale. Future studies should focus on th ese issues in order to further the discussion and aid polic y makers in making the right decisions in the interest of the environment and the economy. 73 APPENDIX 74 APPENDIX A.1. Fish Catch Data and CaMa - Flood Results Table 5 : Yearly fish catch data from FAO FishStat and flooded area from CaMa - Flood modeling results (Pokhrel, Shin, et al., 2018a) . Year C ountry Catch (tonnes) Population (millions) Flooded Area (m 2 ) 1986 Cambodia 64181 7.99 7449842.937 1987 Cambodia 62154 8.228 5743618.036 1988 Cambodia 61155 8.467 5367193.871 1989 Cambodia 50477 8.724 5978190.038 1990 Cambodia 65081 9.009 7591378.873 1991 Cambodia 74672 9.324 7461522.237 1992 Cambodia 68881 9.659 5466623.661 1993 Cambodia 67880 10.007 5601417.416 1994 Cambodia 64960 10.43 8023884.851 1995 Cambodia 72420 10.769 7306399.47 1996 Cambodia 63440 11.091 7736345.108 1997 Cambodia 72900 11.396 8433418.349 1998 Cambodia 75600 11.685 4715221.418 1999 Cambodia 230700 11.96 8688194.163 2000 Cambodia 245300 12.223 10836736.17 2001 Cambodia 384500 12.473 9042568.737 2002 Cambodia 359800 12.709 7871865.977 75 Table 5 ( c 2003 Cambodia 308250 12.934 5899931.059 2004 Cambodia 249600 13.149 6713935.735 2005 Cambodia 323500 13.356 5642441.882 2006 Cambodia 421400 13.555 7734468.114 2007 Cambodia 394500 13.747 7314523.61 2008 Cambodia 364600 13.941 6750178.82 2009 Cambodia 389700 14.144 5502564.432 2010 Cambodia 404600 14.365 3394043.589 1986 Laos 21000 3.618 773557.3388 1987 Laos 22000 3.721 639985.6546 1988 Laos 21000 3.828 424875.4926 1989 Laos 20000 3.938 698878.435 1990 Laos 18000 4.087 971051.2318 1991 Laos 18500 4.208 824015.0301 1992 Laos 16740 4.331 533118.4677 1993 Laos 17000 4.454 634470.5043 1994 Laos 20600 4.575 879182.4432 1995 Laos 23370 4.691 856762.2944 1996 Laos 19500 4.801 868674.2496 1997 Laos 16057 4.907 835684.809 1998 Laos 16642 4.991 460912.5461 1999 Laos 25541 5.076 973953.1085 76 Table 5 ( c 2000 Laos 24850 5.162 1142034.91 2001 Laos 26350 5.25 1048796.202 2002 Laos 28440 5.339 951912.8518 2003 Laos 25300 5.43 543321.3001 2004 Laos 25300 5.522 692934.2924 2005 Laos 22560 5.621 774996.377 2006 Laos 22825 5.702 828989.5548 2007 Laos 24110 5.784 682768.2364 2008 Laos 24700 5.867 796932.0428 2009 Laos 25950 5.952 382228.6704 2010 Laos 26000 6.038 394575.0301 1986 Thailand 44199 52.97 2110210.497 1987 Thailand 31860 53.87 1483368.398 1988 Thailand 44304 54.96 2888019.589 1989 Thailand 47440 55.29 2258039.379 1990 Thailand 47316 56.303 2172420.281 1991 Thailand 37012 56.961 1841197.065 1992 Thailand 37054 57.789 1360666.753 1993 Thailand 52763 58.096 1508179.352 1994 Thailand 64587 58.797 2597340.804 1995 Thailand 60272 59.47 2850923.489 1996 Thailand 105726 60.077 2934031.101 77 Table 5 ( c 1997 Thailand 108551 60.677 2282071.275 1998 Thailand 70011 61.277 1342449.494 1999 Thailand 59376 61.849 2639742.633 2000 Thailand 64241 62.321 4087454.618 2001 Thailand 69000 62.908 3353677.05 2002 Thailand 70300 63.488 2911686.213 2003 Thailand 75171 64.05 3288702.276 2004 Thailand 72500 64.603 2777687.513 2005 Thailand 56310 65.152 1712026.813 2006 Thailand 129200 65.632 2891868.983 2007 Thailand 159800 66.094 2506614.072 2008 Thailand 59700 66.533 1900246.74 2009 Thailand 50418 66.953 1160043.846 2010 Thailand 71254 67.341 952472.5536 1986 Vietnam 119061 60.92 6460020.44 1987 Vietnam 127914 62.3 5510192.656 1988 Vietnam 124736 63.5 5454192.075 1989 Vietnam 141757 64.774 5660954.278 1990 Vietnam 124915 66.017 6350668.234 1991 Vietnam 135822 67.242 6133982.851 1992 Vietnam 137154 68.45 5410931.14 1993 Vietnam 145839 69.645 5501023.083 78 Table 5 ( c 1994 Vietnam 79087 70.825 6517556.834 1995 Vietnam 94189 71.996 6080357.49 1996 Vietnam 163936 73.157 6621183.257 1997 Vietnam 176589 74.307 6765012.73 1998 Vietnam 137800 75.456 5318329.523 1999 Vietnam 168107 76.597 6985115.16 2000 Vietnam 180000 77.635 7467048.856 2001 Vietnam 188542 78.686 7062966.645 2002 Vietnam 163615 79.727 6289337.431 2003 Vietnam 148959 80.899 5966124.477 2004 Vietnam 194621 82.032 5875215.785 2005 Vietnam 188400 82.392 5457452.756 2006 Vietnam 193300 83.311 6153429.565 2007 Vietnam 187800 84.219 6231355.705 2008 Vietnam 178700 85.119 6147961.547 2009 Vietnam 177400 86.025 5263087.379 2010 Vietnam 168855 86.933 4160992.987 79 A.2. Nationwide Population Data Table 6 : Population (millions) data from IMF, retrieved October 9 th , 2018. Year Cambodia Laos Thailand Vietnam 1986 25.707 3.618 52.97 60.92 1987 17.116 3.721 53.87 62.3 1988 32.608 3.828 54.96 63.5 1989 39.704 3.938 55.29 64.774 1990 99.833 4.087 56.303 66.017 1991 215.671 4.208 56.961 67.242 1992 252.491 4.331 57.789 68.45 1993 242.479 4.454 58.096 69.645 1994 265.056 4.575 58.797 70.825 1995 319.537 4.691 59.47 71.996 1996 316.188 4.801 60.077 73.157 1997 302.166 4.907 60.677 74.307 1998 267.864 4.991 61.277 75.456 1999 293.699 5.076 61.849 76.597 2000 299.982 5.162 62.321 77.635 2001 320.046 5.25 62.908 78.686 2002 337.501 5.339 63.488 79.727 2003 360.659 5.43 64.05 80.899 2004 405.629 5.522 64.603 82.032 2005 470.683 5.621 65.152 82.392 8 0 Table 6 ( c 2006 536.151 5.702 65.632 83.311 2007 627.78 5.784 66.094 84.219 2008 741.855 5.867 66.533 85.119 2009 734.655 5.952 66.953 86.025 2010 781.912 6.038 67.341 86.933 2011 877.635 6.124 67.638 87.84 2012 945.702 6.213 67.956 88.809 2013 1,009.34 6.302 68.297 89.76 2014 1,090.71 6.393 68.657 90.728 2015 1,163.41 6.492 68.838 91.713 2016 1,270.48 6.585 68.981 92.691 81 A.3. Virtual Water Trade Network Results Table 7 : Virtual water trade network results summary at a regional level for years 1988, 1995, 2004, 2008, 2012, and 2016. Year Origin Destination Flow (km3) 1988 Americas Thailand 503150 Europe Viet Nam 2084 Cambodia 41520 Thailand 46056 Oceania Thailand 321993 Rest of Asia Thailand 715716 Thailand Viet Nam 24430 Laos 11526 Rest of Asia 494038938 Europe 3699846 1995 Africa Cambodia 15752 Americas Viet Nam 92584 Thailand 654095 Europe Thailand 115698 Oceania Viet Nam 84735 Thailand 321993 Rest of Asia Viet Nam 405396 Cambodia 161564 82 Table 7 ( c Thailand 506745 Thailand Cambodia 132549 Oceania 69156 Americas 795294 Europe 55757025 Rest of Asia 664301010 Africa 12378924 Viet Nam Rest of Asia 3167932 2004 Africa Thailand 5035 Americas Viet Nam 20805320 Thailand 473509 Europe Viet Nam 39200 Cambodia 7362 Thailand 353430 Oceania Viet Nam 1143786 Cambodia 241414 Thailand 3929550 Rest of Asia Viet Nam 22228065 Cambodia 63510 Thailand 648432 Thailand Viet Nam 1970946 Laos 7039 83 Table 7 ( c Cambodia 28815 Thailand 16165215 Americas 1716156 Europe 47896293 Rest of Asia 70579461 Africa 432225 Viet Nam Rest of Asia 53612894 2008 Africa Viet Nam 2154093 Americas Viet Nam 90645066 Cambodia 845292 Thailand 560720 Europe Viet Nam 15308929 Cambodia 11198 Thailand 419265 Oceania Viet Nam 3343557 Laos 25412 Cambodia 406592 Thailand 3303560 Rest of Asia Viet Nam 85981467 Cambodia 57159 Thailand 10948088 Thailand Viet Nam 126855156 84 Table 7 ( c Cambodia 1025814 Thailand 121023 Rest of Asia 23511088 Europe 213231 Viet Nam Rest of Asia 27069366 2012 Africa Viet Nam 484335 Thailand 402375 Americas Viet Nam 113269128 Laos 532032 Cambodia 734599 Thailand 1226514 Cambodia Thailand 109700 Europe Viet Nam 34243440 Laos 34650 Thailand 3080385 Oceania Viet Nam 7903132 Laos 64071 Cambodia 146119 Thailand 4281922 Rest of Asia Viet Nam 520242165 Laos 7013448 Cambodia 3844764 85 Table 7 ( c Thailand 36974068 Thailand Laos 223664473 Cambodia 1077681 Thailand 973947 Rest of Asia 87563022 Oceania 144075 Americas 478329 Europe 57572370 Africa 47740692 Viet Nam Rest of Asia 29224334 2016 Africa Viet Nam 57797145 Americas Viet Nam 107770384 Cambodia 1056615 Thailand 3054588 Cambodia Viet Nam 837012 Europe Viet Nam 15406080 Laos 86625 Thailand 3769920 Oceania Viet Nam 24128610 Laos 25412 Cambodia 1255110 Thailand 5242272 86 Table 7 ( c Rest of Asia Viet Nam 342814365 Laos 5623644 Cambodia 4251520 Thailand 26981884 Thailand Viet Nam 9422505 Laos 311536254 Cambodia 14516997 Thailand 1256334 Americas 28815 Africa 11335821 Rest of Asia 643825071 Europe 25576194 Viet Nam Rest of Asia 40337904 Table 8 : Calculated virtual water contents of all the meat commodities for every nation (m 3 /ton) (A K Chapagain & Hoekstra, 2003) . Country Cattle Pig Sheep Goat Chicke n Horse Afghanistan 13685 2794 6515 5518 7510 5646 Albania 13426 6586 6454 5440 5584 5675 Algeria 15142 8671 7400 5056 9625 6130 Andorra 10586 2802 5650 3770 1849 5408 87 Table 8 ( c Angola 12891 2245 6081 5365 10243 5614 Anguilla 11781 2245 5645 4776 2392 5390 Antigua and Barbuda 11348 4231 5862 4207 4860 5638 Argentina 11549 3471 5943 4052 2327 5782 Armenia 12983 2699 6227 5251 3963 5544 Aruba 10586 2802 5650 3770 1849 5408 Australia 11730 6126 6353 3351 2373 6251 Austria 8025 2307 5369 2494 884 4741 Azerbaijan 13814 13772 6707 5761 4467 5813 Bahamas 13865 6554 7030 4985 4095 5997 Bahrain 10586 2802 5650 3770 1849 5408 Bangladesh 15546 4359 7515 6388 7828 5938 Barbados 11007 3739 5767 4011 3513 5535 Belarus 11821 4536 6251 4241 3601 5442 Belgium 7670 1833 5203 2431 893 4813 Belize 11254 4415 5902 4213 4448 5766 Benin 12729 5938 6122 5479 8482 5754 Bermuda 13548 3083 6536 5521 4618 5658 Bhutan 11781 2245 5645 4776 2392 5390 Bolivia 14843 3974 7145 5996 7087 5823 Bosnia and Herzegovina 11599 2238 5618 4780 2842 5285 BI Ocean Territory 11781 2245 5645 4776 2392 5390 88 Table 8 ( c British Virgin Islands 11781 2245 5645 4776 2392 5390 Brazil 13133 5005 6281 4401 3342 6552 Brunei Darussalam 9391 3359 5655 2763 1305 5427 Bulgaria 11078 4059 5920 4085 3409 5568 Burkina Faso 12899 9265 6181 5622 9366 5793 Burundi 13153 4471 6182 5466 12645 5717 Cambodia 12309 2289 5886 5069 5485 5510 Cameroon 12442 4027 5949 5177 7299 5626 Canada 9636 3276 5674 2775 1358 5567 Cabo Verde 14953 10998 6867 6275 19116 6724 Cayman Islands 11781 2245 5645 4776 2392 5390 Central African Republic 12299 3875 5866 5140 5867 5533 Chad 13973 9055 6563 6034 16651 5915 Chile 10080 2667 5495 3686 1979 5312 China, mainland 12596 2522 5948 5076 3111 5671 Christmas Island 11781 2245 5645 4776 2392 5390 Cocos (Keeling) Islands 11781 2245 5645 4776 2392 5390 Colombia 11782 4035 5972 4187 3878 5835 Comoros 12035 2245 5748 4949 4236 5452 Congo 11781 2245 5645 4776 2392 5390 DR of the Congo 15810 7254 7656 6700 12181 6178 Cook Islands 11781 2245 5645 4776 2392 5390 89 Table 8 ( c Costa Rica 10761 3558 5779 3911 3078 5360 Cote d'Ivoire 12866 9659 6162 5594 9042 5774 Croatia 10717 3361 5807 3870 2667 5250 Cuba 12281 4516 5908 5003 4236 5508 Cyprus 12373 4829 6338 4317 4716 5688 Czechia 9901 2609 5514 3668 1716 5088 Denmark 7827 2232 5334 2484 865 4693 Djibouti 12549 2245 5957 5297 7966 5578 Dominica 12819 6915 6492 4737 5617 5893 Dominican Republic 10586 2802 5650 3770 1849 5408 Ecuador 14763 8102 7322 5150 7232 6134 Egypt 15752 4680 6743 4892 2268 8067 El Slavador 12591 4989 6034 4283 4285 6461 Equatorial Guinea 11781 2245 5645 4776 2392 5390 Eritrea 14258 4254 6742 5855 11688 5783 Estonia 12367 4479 6398 4270 3042 5627 Ethiopia 14882 5260 7089 5987 10135 5828 Faroe Islands 11781 2245 5645 4776 2392 5390 Falkland Islands (Malvinas) 11781 2245 5645 4776 2392 5390 Fiji 10586 2802 5650 3770 1849 5408 Finland 9596 4273 5818 2864 1981 5248 France 7744 1936 5257 2424 795 4689 90 Table 8 ( c French Guiana 12109 5660 6074 4644 7871 5867 French Polynesia 9391 3359 5655 2763 1305 5427 FS&A Territories 11781 2245 5645 4776 2392 5390 Gabon 10979 3601 5770 4006 3403 5553 Gambia 12640 6141 6046 5315 7165 5624 Georgia 14123 3031 6357 5535 5476 6324 Germany 7768 2110 5275 2440 877 4672 Ghana 11781 2245 5645 4776 2392 5390 Gibraltar 11781 2245 5645 4776 2392 5390 Greece 12763 5037 6315 4306 3916 6167 Greenland 11781 2245 5645 4776 2392 5390 Grenada 10704 3176 5709 3865 2317 5510 Guadeloupe 11781 2245 5645 4776 2392 5390 Guatemala 14861 7282 7259 5127 3418 6489 Guinea 12273 5274 5865 5132 5432 5541 Guinea - Bissau 12600 5181 5998 5353 7822 5620 Guyana 12420 2245 5904 5209 7026 5546 Haiti 12889 4732 6126 5507 9578 5678 Honduras 12534 4099 5936 5052 7250 5516 China, Hong Kong SAR 9391 3359 5655 2763 1305 5427 Hungary 10298 2776 5623 3740 1771 5207 Iceland 9391 3359 5655 2763 1305 5427 91 Table 8 ( c India 14379 7562 6629 5734 8499 6251 Indonesia 13383 4589 6053 5225 8676 5969 Iran 15783 9133 7881 5518 5203 6321 Iraq 15776 16607 7653 5488 11585 6572 Ireland 7586 2017 5251 2455 908 4664 Israel 15500 7734 7755 5343 5434 6241 Italy 9595 3465 5716 2756 1637 5436 Jamaica 12324 4859 6037 4479 6020 6147 Japan 9535 4082 5658 2913 2044 5549 Jordan 14478 8682 6985 4400 12876 5844 Kazakhstan 18959 6302 7233 5416 6837 9661 Kenya 12789 5156 6086 5291 8122 5640 Kiribati 11781 2245 5645 4776 2392 5390 Kuwait 9391 3359 5655 2763 1305 5427 Kyrgyzstan 12834 15588 6190 5418 6061 5754 Laos 12133 2742 5827 5048 4883 5539 Latvia 11741 4257 6173 4208 2773 5634 Lebanon 11106 4448 5911 4189 3908 5854 Liberia 11106 4448 5911 4189 3908 5854 Libya 15167 6239 7276 6168 8601 5880 Lithuania 12398 2831 5995 5069 3249 5492 Macau 10586 2802 5650 3770 1849 5408 92 Table 8 ( c Yugoslav 10963 3620 5805 3946 2190 5498 Madagascar 12989 8238 6199 5599 10110 5775 Malawi 12539 4833 5955 5121 7785 5584 Malaysia 12934 5913 6521 4498 7468 5882 Maldives 10586 2802 5650 3770 1849 5408 Mali 12700 5889 6028 5218 8087 5592 Malta 10264 2761 5610 3719 1849 5204 Marshall Islands 10586 2802 5650 3770 1849 5408 Martinique 11781 2245 5645 4776 2392 5390 Mauritania 19223 15006 9063 8275 27332 6693 Mauritius 10639 2902 5665 3801 2060 5425 Mexico 12493 4062 6080 4265 2806 6375 Micronesia 10586 2802 5650 3770 1849 5408 Republic of Moldova 12698 2881 6103 5201 4545 5566 Mongolia 12977 2245 5941 4878 12923 5455 Montserrat 12325 2245 5866 5145 6341 5523 Morocco 15876 12428 7531 6051 14457 6693 Mozambique 12264 4069 5818 4984 5928 5484 Myanmar 13781 4888 6601 5658 8208 5750 New Caledonia 10586 2802 5650 3770 1849 5408 Northern Mariana Islands 11781 2245 5645 4776 2392 5390 Nauru 11781 2245 5645 4776 2392 5390 93 Table 8 ( c Nepal 14007 3724 6747 5747 8422 5821 Netherlands Antilles 11781 2245 5645 4776 2392 5390 Netherlands 7676 2084 5259 2451 914 4658 New Zealand 10551 3764 5649 4045 4179 5445 Nicaragua 13496 3743 6167 5483 7713 6072 Niger 15594 22346 7601 6796 12703 6356 Nigeria 13193 8250 6313 5710 12563 5917 Niue 11781 2245 5645 4776 2392 5390 Norfolk Islands 11781 2245 5645 4776 2392 5390 North Korea 11884 4649 5961 4153 2860 6125 Norway 9439 3382 5690 2764 1054 5434 Oman 17777 9229 8066 5838 5427 7764 Pakistan 14610 10618 7100 5978 7853 5942 Palau 10586 2802 5650 3770 1849 5408 Panama 11934 5773 6121 4413 6865 5915 Papua New Guinea 11834 2151 5666 4812 2796 5402 Paraguay 13302 6648 6570 4640 4443 6253 Peru 14179 6301 6921 4890 5275 6379 Philippines 12203 4119 5847 5074 5288 5534 Pitcairn Islands 11781 2245 5645 4776 2392 5390 Poland 10164 2560 5566 3725 2003 5211 Portugal 15177 5279 6800 4739 4029 7318 94 Table 8 ( c Qatar 10893 3096 5776 3845 2172 5446 Reunion 10980 3369 5808 3859 2465 5486 Romania 11198 3681 5854 3993 2527 5600 Russian Federation 13089 6010 6510 4600 4702 6205 Rwanda 11781 2245 5645 4776 2392 5390 Saint Vincent - Grenadines 11781 2245 5645 4776 2392 5390 Samoa 10586 2802 5650 3770 1849 5408 Sao Tome Principe 11781 2245 5645 4776 2392 5390 Saudi Arabia 11359 4985 5904 4179 4146 5648 Senegal 12535 5368 5982 5284 7033 5600 Seychelles 10586 2802 5650 3770 1849 5408 Sierra Leone 12306 4757 5875 5151 5764 5544 Singapore 11261 6162 6351 3317 3107 5879 Slovakia 10513 3427 5715 3878 2323 5333 Slovenia 10625 2875 5661 3792 2002 5420 Solomon Islands 11781 2245 5645 4776 2392 5390 Somalia 11781 2245 5645 4776 2392 5390 South Africa 16095 8799 7476 5440 5035 7248 South Korea 10586 2802 5650 3770 1849 5408 Spain 11963 3057 5935 4042 1912 6114 Spec Cats 11781 2245 5645 4776 2392 5390 Sri Lanka 12736 7283 6134 5390 7414 5707 95 Table 8 ( c Saint Helena 11781 2245 5645 4776 2392 5390 Saint Kitts and Nevis 10586 2802 5650 3770 1849 5408 Saint Lucia 10586 2802 5650 3770 1849 5408 Saint Pierre and Miquelon 11781 2245 5645 4776 2392 5390 Sudan 13265 6548 6248 5689 13072 5783 Suriname 11842 5653 6087 4581 6819 6003 Sweden 8339 2645 5463 2597 929 4886 Switzerland 8778 2383 5425 2574 858 5288 Syrian Arab Republic 20992 12197 10379 8612 9820 6789 China, Taiwan Province of 10586 2802 5650 3770 1849 5408 Tajikistan 13135 8002 6244 5329 7575 5641 United Republic of Tanzania 13260 6647 6398 5729 11305 5948 Thailand 14668 6452 7039 4836 5763 6505 Togo 12805 6027 6081 5092 7986 5567 Tokelau 11781 2245 5645 4776 2392 5390 Tonga 10586 2775 5650 3770 1849 5408 Trinidad and Tobago 11082 3967 5825 4107 3788 5655 Tunisia 40101 13205 10331 9054 3435 22722 Turkey 11070 3242 5624 3828 2301 5821 Turkmenistan 19703 5445 9705 7990 6471 6439 Turks Caicos Islands 11781 2245 5645 4776 2392 5390 Tuvalu 11781 2245 5645 4776 2392 5390 96 Table 8 ( c Uganda 12645 3687 6026 5260 8198 5625 Ukraine 13070 4682 6209 5380 6160 5749 United Arab Emirates 12476 8472 6541 3007 9026 5791 United Kingdom 7759 2106 5309 2476 784 4659 Uruguay 13684 4277 6294 4410 2814 7024 US miscellaneous pacific 10586 2802 5650 3770 1849 5408 United States of America 10063 3374 5718 2843 1304 5874 Uzbekistan 13471 4010 6527 5570 4851 5715 Vanuatu 11781 2245 5645 4776 2392 5390 Venezuela 14615 6616 6912 4699 9072 6490 Viet Nam 13948 3226 6754 5718 4947 5729 Wallis and Futuna Islands 11781 2245 5645 4776 2392 5390 Yemen 13873 4140 6616 5707 8435 5721 Yugoslavia 11443 3682 5949 4042 3016 5618 Zambia 14766 6104 7164 6188 6340 5888 Zimbabwe 13119 5733 6335 5486 4939 5671 97 BIBLIOGRAPHY 98 BIBLIOGRAPHY Abdullah, K. bin. (2006). Use of water and land for food security and environmental sustainability. Irrigation and Drainage , 55 (3), 219 222. https://doi.org/10.1002/ird.254 Adamson, P. (2006). Hydrological and water resources modelling in the Mekong Region : A brief overview. Exploring Water Futures Together: Mekong Region Waters Dialogue , 69 74. Aleksandrowicz, L., Green, R., Joy, E. J. M., Smith, P., & Haines, A. (2016). The Impacts of Dietary Change on Greenhouse Gas Emissions, Land Use, Water Use, and He alth: A Systematic Review. PLOS ONE , 11 (11), e0165797. https://doi.org/10.1371/journal.pone.0165797 Allison, E. H., Perry, A. L., Badjeck, M. - N. K. (2009). Vulnerability of national economies to the impac ts of climate change on fisheries. Fish and Fisheries , 10 (2), 173 196. https://doi.org/10.1111/j.1467 - 2979.2008.00310.x Arias, M. E., Piman, T., Lauri, H., Cochrane, T. A., & Kummu, M. (2014). Dams on Mekong tributaries as significant contributors of hydro logical alterations to the Tonle Sap Floodplain in Cambodia. Hydrology and Earth System Sciences , 18 (12), 5303. Baran, E., Guerin, E., & Nasielski, J. (2015). Fish, sediment and dams in the Mekong: How hydropower development affects water productivity and food supply . Baran, E., & Myschowoda, C. (2009). Dams and fisheries in the Mekong Basin. Aquatic Ecosystem Health & Management , 12 (3), 227 234. Barlow, C., Baran, E., Halls, A. S., & Kshatriya, M. (2008). How much of the Mekong fish catch is at risk from m ainstream dam development? Catch and Culture , 14 (3), 16 21. Bartley, D. M., De Graaf, G. J., Valbo - Jørgensen, J., & Marmulla, G. (2015). Inland capture fisheries: status and data issues. Fisheries Management and Ecology , 22 (1), 71 77. https://doi.org/10.1111/fme.12104 Bonfils, C., & Lobell, D. (2007). Empirical evidence for a recent slowdown in irrigation - induced cooling. Proceedings of the National Academy of Sciences , 104 (34), 13582 13587. https://doi.org/10.1073/pnas.0700144104 Brown ell, R. L., Reeves, R. R., Thomas, P. O., Smith, B. D., & Ryan, G. E. (2017). Dams threaten rare Mekong dolphins. Science , 355 (6327), 805. Cai, B., Zhang, W., Hubacek, K., Feng, K., Li, Z., Liu, Y., & Liu, Y. (2019). Drivers of virtual water flows on regional water scarcity in China. Journal of Cleaner Production , 207 , 1112 1122. https://doi.org/10.1016/j.jclepro.2018.10.077 99 Carpenter, S. R., Ecosystems: Physical, Chemical, and Biological Changes. Annual Review of Environment and Resources , 36 (1), 75 99. https://doi.org/10.1146/annurev - environ - 021810 - 094524 Chapagai n, A. K., & Hoekstra, A. Y. (2003). Virtual water flows between nations in relation to trade in livestock and livestock products Value of Water. Retrieved from http://www.ayhoekstra.nl/pubs/Report13.pdf Chapagain, A. K., & Hoekstra, A. Y. (2008). The globa l component of freshwater demand and supply: an assessment of virtual water flows between nations as a result of trade in agricultural and industrial products. Water International , 33 (1), 19 32. https://doi.org/10.1080/02508060801927812 Chapagain, A. K., H oekstra, A. Y., & Savenije, H. H. G. (2006). Water saving through international trade of agricultural products Hydrology and Earth System Sciences Water saving through international trade of agricultural products . European Geosciences Union (Vol. 10). Retr ieved from www.hydrol - earth - syst - sci.net/10/455/2006/ Chaudhari, S., Felfelani, F., Shin, S., & Pokhrel, Y. (2018). Climate and Anthropogenic Contributions to the Desiccation of the Second Largest Saline Lake in the Twentieth Century. Journal of Hydrology . Costales, A. (2004). A Review of the Thailand Poultry Sector . Retrieved from http://www.fao.org/ag/againfo/resources/en/publications/sector_reports/lsr_THA.pdf da, R. (2016). Water Footprint and Virtual Water Trade of Brazil. Water , 8 (11), 517. https://doi.org/10.3390/w8110517 Dalin, C., Konar, M., Hanasaki, N., Rinaldo, A., & Rodriguez - Iturbe, I. (2012). Evolution of the global virtual water trade network. Proce edings of the National Academy of Sciences , 109 (16), 5989 5994. https://doi.org/10.1073/pnas.1203176109 Dalin, C., Qiu, H., Hanasaki, N., Mauzerall, D. L., & Rodriguez - Iturbe, I. (2015). Balancing water resource conservation and food security in China. Pro ceedings of the National Academy of Sciences , 112 (15), 4588 4593. https://doi.org/10.1073/pnas.1504345112 Dang, Q., Lin, X., & Konar, M. (2015). Agricultural virtual water flows within the United States. Water Resources Research , 51 (2), 973 986. https://do i.org/10.1002/2014WR015919 Delgado, C., Rosegrant, M., Steinfeld, H., Ehui, S., & Courbois, C. (n.d.). Livestock to 2020: The Next Food Revolution . Washington, DC. Retrieved from http://lib.icimod.org/record/10601/files/4441.PDF Delgado, J. M., Apel, H., & Merz, B. (2010). Flood trends and variability in the Mekong river. Hydrology and Earth System Sciences , 14 (3), 407 418. https://doi.org/10.5194/hess - 14 - 407 - 100 2010 Delgado, J. M., Merz, B., & Apel, H. (2012). A climate - flood link for the lower Mekong River. Hydrology and Earth System Sciences , 16 (5), 1533 1541. https://doi.org/10.5194/hess - 16 - 1533 - 2012 Dickson, B., Hutton, J., & Adams, W. M. (Eds.). (2009). Recreational Hunting, Conservation and Rural Livelihoods . Oxford, UK: Wiley - Blackwell. https://doi.org/ 10.1002/9781444303179 Duarte, R., Pinilla, V., & Serrano, A. (2019). Long Term Drivers of Global Virtual Water Trade: A Trade Gravity Approach for 1965 2010. Ecological Economics , 156 , 318 326. https://doi.org/10.1016/j.ecolecon.2018.10.012 Dugan, P. J., B - Cooper, M. (2010). Fish migration, dams, and loss of ecosystem services in the Mekong basin. AMBIO: A Journal of the Human Environment , 39 (4), 344 348. O. (2010). Fish Migration, Dams, and Loss of Ecosystem Services in the Mekong Basin. AMBIO , 39 (4), 344 348. https://doi.org/10.1007/s13280 - 010 - 0036 - 1 Erban, L. E ., & Gorelick, S. M. (2016). Closing the irrigation deficit in Cambodia: Implications for transboundary impacts on groundwater and Mekong River flow. Journal of Hydrology , 535 , 85 92. https://doi.org/10.1016/j.jhydrol.2016.01.072 FAO. (2018). FishStatJ Sof tware for Fishery Statistical Time Series. Retrieved November 23, 2018, from www. fao.org/fishery/statistics/software/fishstatj/en. FAOSTAT. (2009). Food and Agricultural Organization of the United Nations . Rome. Felfelani, F., Wada, Y., Longuevergne, L., & Pokhrel, Y. (2017). Natural and human - induced terrestrial water storage change: A global analysis using hydrological models and GRACE. Journal of Hydrology , 553 , 105 118. Foley, J. A., Ramankutty, N., Brauman, K. A., Cassidy, E. S., Gerber, J. S., Johnst D. P. M. (2011). Solutions for a cultivated planet. Nature , 478 (7369), 337 342. https://doi.org/10.1038/nature10452 Fredén, F. (2011). Impacts of dams on lowland agriculture in the Mekong River catchment. Lunds Universitets Naturgeografiska Institution - Seminarieuppsatser . Frenken, K., & FAO. (2011). Irrigation in Southern and Eastern Asia in Figures: Aquastat Survey, 2011 (Vol. FAO Water). FAO: Food and Agricultural Organization. Gephart, J. A., Davis, K. F., Emery, K. A., Leach, A. M., Gall oway, J. N., & Pace, M. L. (2016). The environmental cost of subsistence: Optimizing diets to minimize footprints. Science of 101 The Total Environment , 553 , 120 127. https://doi.org/10.1016/j.scitotenv.2016.02.050 Grumbine, R. E., & Xu, J. (2011). Mekong Hydr opower Development. Science , 332 (6026), 178 179. https://doi.org/10.1126/science.1200990 Han, Z., Long, D., Fang, Y., Hou, A., & Hong, Y. (2019). Impacts of climate change and human activities on the flow regime of the dammed Lancang River in Southwest Chi na. Journal of Hydrology . https://doi.org/10.1016/j.jhydrol.2018.12.048 Hanasaki, N., Inuzuka, T., Kanae, S., & Oki, T. (2010). An estimation of global virtual water flow and sources of water withdrawal for major crops and livestock products using a global hydrological model. Journal of Hydrology , 384 (3 4), 232 244. https://doi.org/10.1016/j.jhydrol.2009.09.028 https://thediplomat.com/2018/06/vietnams - meat - boom/ He, C., & Food and Agriculture Organization of the United Nations. (2004). Protein sources for - 3 May 2002. Food and Agriculture Organization of the United Nations. Retrieved from http:/ /www.fao.org/3/y5019e/y5019e00.htm#Contents developments and irrigation e xpansions drive hydrological changes. Science of The Total Environment , 649 , 601 609. https://doi.org/10.1016/j.scitotenv.2018.08.160 Hoekstra, A. Y. (2003). Virtual Water Trade: Proceedings of the international meeting on virtual water trade \ r . Value of W ater Research Report Series No. 12 . Retrieved from https://waterfootprint.org/media/downloads/Report12.pdf Hortle, K. G. (2009). Fisheries of the Mekong River Basin. In Ian C. Campbell (Ed.), The Mekong (pp. 197 249). Academic Press. https://doi.org/10.101 6/B978 - 0 - 12 - 374026 - 7.00009 - 7 ICEM, & Management, I. C. for E. (2010a). MRC Strategic Environmental Assessment (SEA) of hydropower on the Mekong mainstream, Hanoi, Viet Nam . Hanoi, Viet Nam: MRC. Retrieved from http://www.mrcmekong.org/assets/Publications/Consultations/SEA - Hydropower/SEA - FR - summary - 13oct.pdf ICEM, & Management, I. C. for E. (2010b). MRC Strategic Environmental Assessment (SEA) of hydropower on the Mekong mainstream, Hanoi, Viet Nam . Hanoi, Viet N am: MRC. (2018). Operationalizing food system resilience: An indicator - based assessment in agroindustrial, smallholder farming, and agroecological contexts i n Bolivia and Kenya. Land Use Policy , 79 , 433 446. https://doi.org/10.1016/j.landusepol.2018.08.044 102 Johnston, R., & Kummu, M. (2012). Water Resource Models in the Mekong Basin: A Review. Water Resources Management , 26 (2), 429 455. https://doi.org/10.1007/s 11269 - 011 - 9925 - 8 Keskinen, M., Kummu, M., Käkönen, M., & Varis, O. (2012). Mekong at the crossroads: Next steps for impact assessment of large dams. Ambio , 41 (3), 319. Kian, G., Yeap, H., Eong, S., Sensereivorth, T., & Racy, B. (2005). Maximizing Utilizati on of Fish Catch - Fresh Water Species . Singapore. Retrieved from http://www.seafdec.org/documents/2014/10/maximizing - utilization - fish - catch - freshwater - species.pdf Kim, H., Yeh, P. J. F., Oki, T., & Kanae, S. (2009). Role of rivers in the seasonal variations of terrestrial water storage over global basins. Geophysical Research Letters , 36 (17), L17402. https://doi.org/10.1029/2009GL039006 Konar, M., Hussein, Z., Hanasaki, N., Mauzerall, D. L., & Rodriguez - Iturbe, I. (2013). Virtual water trade flows and saving s under climate change. Hydrology and Earth System Sciences , 17 (8), 3219 3234. https://doi.org/10.5194/hess - 17 - 3219 - 2013 Kondolf, G. M., Rubin, Z. K., & Minear, J. T. (2014). Dams on the Mekong: cumulative sediment starvation. Water Resources Research , 50 ( 6), 5158 5169. (2009). Circos: An information aesthetic for comparative genomics. Genome Research , 19 (9), 1639 1645. https://doi.org/10.1101/gr.092759.109 Kummu, M., & Sarkkula, J. (2008). Impact of the Mekong River Flow Alteration on the Tonle Sap Flood Pulse. AMBIO: A Journal of the Human Environment , 37 (3), 185 192. Retrieved from http://www.bioone.org/doi/full/10.1579/0044 - 7447%282008%2937%5B185%3AIOTMRF Lauri, H., de Moel, H., Ward, P. J., Räsänen, T. A., Keskinen, M., & Kummu, M. (2012a). Future changes in Mekong River hydrology: impact of climate change and reservoir operation on discharge. Hydrol. Earth Syst. Sci. , 16 (12), 4603 4619. https://doi.org/10.5194/ hess - 16 - 4603 - 2012 Lauri, H., de Moel, H., Ward, P. J., Räsänen, T. A., Keskinen, M., & Kummu, M. (2012b). Future changes in Mekong River hydrology: impact of climate change and reservoir operation on discharge. Hydrol. Earth Syst. Sci. , 16 (12), 4603 4619. https://doi.org/10.5194/hess - 16 - 4603 - 2012 Lehner, B., Verdin, K., & Jarvis, A. (2008). New global hydrography derived from spaceborne elevation data. Eos Trans , (AGU 89), 93 94. Li, J., Dong, S., Peng, M., Yang, Z., Liu, S., Li, X., & Zhao, C. (2013). Effe cts of damming on the biological integrity of fish assemblages in the middle Lancang - Mekong River basin. Ecological Indicators , 34 , 94 102. https://doi.org/10.1016/j.ecolind.2013.04.016 103 Wiley Interdisciplinary Reviews: Water , 6 (1), e1320. https://doi.org/10.1002/wat2.1320 Lu, X. X., & Siew, R. Y. (2006). Water discharge and sediment flux changes over the past decades in the Lower Mekong River: possible impacts of the Chinese dams. Hydrol. Earth Syst. Sci. , 10 (2), 181 195. https://doi.org/10.5194/hess - 10 - 181 - 2006 Lutz, A. F., Immerzeel, W. W., Shrestha, A. B., & Bierkens, M. F. P. (2014). Consistent increase Nature Climate Change , 4 (7), 587 592. McIntyre, P. B., Liermann, C. A. R., & Revenga, C . (2016). Linking freshwater fishery management to global food security and biodiversity conservation. Proceedings of the National Academy of Sciences , 113 (45), 12880 12885. McIntyre, P. B., Reidy Liermann, C. A., & Revenga, C. (2016). Linking freshwater f ishery management to global food security and biodiversity conservation. Proceedings of the National Academy of Sciences , 113 (45), 12880 12885. https://doi.org/10.1073/pnas.1521540113 Mekong River Commission (MRC). (2010). Mekong River Commission: State of the Basin Report 2010 . Vientiane, Laos. Monfreda, C., Ramankutty, N., & Foley, J. A. (2008). Farming the planet: 2. Geographic distribution of crop areas, yields, physiological types, and net primary production in the year 2000. Global Biogeochemical Cycl es , 22 (1), GB1022. https://doi.org/10.1029/2007GB002947 MRC, & Mekong River Commission Lao PDR, V. (2005a). Overview of the Hydrology of the Mekong Basin . Mekong River Commission, Vientaine, Lao PDR: Mekong River Commission, Vientaine, Lao PDR. MRC, & Mek ong River Commission Lao PDR, V. (2005b). Water Used for Agriculture in the Lower Mekong River Basin, Discussion Paper . Mekong River Commission, Vientaine, Lao PDR: Mekong River Commission. MRC, & Mekong River Commission Lao PDR, V. (2010a). Assessment o f basin - wide development scenarios main report, Mekong River Commission (MRC), Vientiane, Lao PDR . Vientaine, Lao DPR. MRC, & Mekong River Commission Lao PDR, V. (2010b). Mekong River Commission: State of the Basin Report 2010 . Vientaine, Lao DPR: Mekon g River Commission, Vientaine, Lao PDR. NaRanong, V. (n.d.). . 104 Retrieved from http://www.fao.org/ag/againfo/home/events/bangkok2007/docs/part1/1_4.pdf Nations, F. and A. O. of the U ., & Center, W. (2008). Small - scale capture fisheries: a global overview with emphasis on developing countries: a preliminary report of the Big Numbers Project. Working Papers . Retrieved from https://ideas.repec.org/b/wfi/wfbook/37878.html Nesbitt, H., Joh nston, R., & Solieng, M. (2004). Mekong River water: will river flows meet future agriculture needs in the Lower Mekong Basin? Water in Agriculture , 116 . Nguyen - Viet, H., Tuyet - Hanh, T. T., Unger, F., Dang - Xuan, S., & Grace, D. (2017). Food safety in Vietn am: where we are at and what we can learn from international experiences. Infectious Diseases of Poverty , 6 (1), 39. https://doi.org/10.1186/s40249 - 017 - 0249 - 7 Nilsson, C., Reidy, C. A., Dynesius, M., & Revenga, C. (2005). Fragmentation and Flow Regulation o Science , 308 (5720), 405 408. https://doi.org/10.1126/science.1107887 Oki, T., Entekhabi, D., & Harrold, T. I. (2004). The Global Water Cycle. In R. S. J. Sparks & C. J. Hawkesworth (Eds.), The State of the Planet: Frontie rs and Challenges in Geophysics (pp. 225 237). American Geophysical Union. Retrieved from http://onlinelibrary.wiley.com/doi/10.1029/150GM18/summary Orr, S., Pittock, J., Chapagain, A., & Dumaresq, D. (2012a). Dams on the Mekong River: Lost fish protein an d the implications for land and water resources. Global Environmental Change , 22 (4), 925 932. https://doi.org/10.1016/j.gloenvcha.2012.06.002 Orr, S., Pittock, J., Chapagain, A., & Dumaresq, D. (2012b). Dams on the Mekong River: Lost fish protein and the i mplications for land and water resources. Global Environmental Change , 22 (4), 925 932. https://doi.org/10.1016/J.GLOENVCHA.2012.06.002 Piman, T., Cochrane, T. A., Arias, M. E., Green, A., & Dat, N. D. (2013a). Assessment of Flow Changes from Hydropower Development and Operations in Sekong, Sesan, and Srepok Rivers of the Mekong Basin. Journal of Water Resources Planning and Management , 13 9 (6), 723 732. https://doi.org/10.1061/(ASCE)WR.1943 - 5452.0000286 Piman, T., Cochrane, T. A., Arias, M. E., Green, A., & Dat, N. D. (2013b). Assessment of Flow Changes from Hydropower Development and Operations in Sekong, Sesan, and Srepok Rivers of the Me kong Basin. Journal of Water Resources Planning and Management , 139 (6), 723 732. https://doi.org/10.1061/(ASCE)WR.1943 - 5452.0000286 Piman, T., Lennaerts, T., & Southalack, P. (2013). Assessment of hydrological changes in the lower Mekong Basin from Basin - W ide development scenarios. Hydrological Processes , 27 (15), 2115 2125. https://doi.org/10.1002/hyp.9764 Pittock, J., Dumaresq, D., & Orr, S. (2017). The Mekong River: trading off hydropower, fish, and 105 food. Regional Environmental Change , 17 (8), 2443 2453. h ttps://doi.org/10.1007/s10113 - 017 - 1175 - 8 Pokhrel, Y., Burbano, M., Roush, J., Kang, H., Sridhar, V., & Hyndman, D. (2018). A Review of the Integrated Effects of Changing Climate, Land Use, and Dams on Mekong River Hydrology. Water , 10 (3), 266. https://doi. org/10.3390/w10030266 Pokhrel, Y., Felfelani, F., Shin, S., Yamada, T. J., & Satoh, Y. (2017). Modeling large - scale human alteration of land surface hydrology and climate. Geoscience Letters , 4 (1), 10. Pokhrel, Y., Hanasaki, N., Koirala, S., Cho, J., Yeh, Incorporating Anthropogenic Water Regulation Modules into a Land Surface Model. Journal of Hydrometeorology , 13 (1), 255 269. https://doi.org/10.1175/JHM - D - 11 - 013.1 Pokhrel, Y., Hanasaki, N., Yeh, P. J. F., Yamada, T. J., Kanae, S., & Oki, T. (2012). Model estimates of sea - level change due to anthropogenic impacts on terrestrial water storage. Nature Geoscience , 5 (6), 389 392. https://doi.org/10.1038/ngeo1476 Pokhrel, Y., Koirala, S., Yeh, P. J. F., Hanasaki, N., Longuevergne, L., Kanae, S., & Oki, T. (2015). Incorporation of groundwater pumping in a global Land Surface Model with the representation of human impacts. Water Resources Research , 51 (1), 78 96. https://doi. org/10.1002/2014WR015602 Pokhrel, Y., Shin, S., Lin, Z., Yamazaki, D., & Qi, J. (2018a). Potential Disruption of Flood Dynamics in the Lower Mekong River Basin Due to Upstream Flow Regulation. Scientific Reports , 8 (1), 17767. https://doi.org/10.1038/s41598 - 018 - 35823 - 4 Pokhrel, Y., Shin, S., Lin, Z., Yamazaki, D., & Qi, J. (2018b). Potential Disruption of Flood Dynamics in the Lower Mekong River Basin Due to Upstream Flow Regulation. Scientific Reports , In Revisio . Ranum, P., Peña - Rosas, J. P., & Garcia - Casa l, M. N. (2014). Global maize production, utilization, and consumption. Annals of the New York Academy of Sciences , 1312 (1), 105 112. https://doi.org/10.1111/nyas.12396 Räsänen, T. A., Koponen, J., Lauri, H., & Kummu, M. (2012). Downstream Hydrological Imp acts of Hydropower Development in the Upper Mekong Basin. Water Resources Management , 26 (12), 3495 3513. https://doi.org/10.1007/s11269 - 012 - 0087 - 0 Räsänen, T. A., & Kummu, M. (2013). Spatiotemporal influences of ENSO on precipitation and flood pulse in the Mekong River Basin. Journal of Hydrology , 476 , 154 168. https://doi.org/10.1016/j.jhydrol.2012.10.028 River Commission, M. (2002). Mekong River Commission Fisheries in the Lower Mekong Basin: Status and Perspectives . Retrieved from www.mrcmekong.org 106 Rost, S., Gerten, D., Bondeau, A., Lucht, W., Rohwer, J., & Schaphoff, S. (2008). Agricultural green and blue water consumption and its influence on the global water system. Water Resources Research , 44 (9). https://doi.org/10.1029/2007WR006331 Sabo, J. L., Designing river flows to improve food security futures in the Lower Mekong Basin. Science , 358 (6368), eaao1053. https://doi.org/10.1126/science.aao1053 Sabo, J. L., Ru Designing river flows to improve food security futures in the Lower Mekong Basin. Science , 358 (6368), eaao1053. Salmon, J. M., Friedl, M. A., Frolking, S., Wisser, D., & Douglas, E. M. (2015). Global rain - fed, irrigated, and paddy croplands: A new high resolution map derived from remote sensing, crop inventories and climate data. International Journal of Applied Earth Observation and Geoinformation , 38 , 321 334. https://do i.org/10.1016/j.jag.2015.01.014 Shrestha, B., Babel, M. S., Maskey, S., van Griensven, A., Uhlenbrook, S., Green, A., & Akkharath, I. (2013). Impact of climate change on sediment yield in the Mekong River basin: a case study of the Nam Ou basin, Lao PDR. H ydrol. Earth Syst. Sci. , 17 (1), 1 20. https://doi.org/10.5194/hess - 17 - 1 - 2013 Responding to rising sea levels in the Mekong Delta. Nature Climate Change , 5 (2), 167 174. https://doi.org/10.1038/nclimate2469 Stone, R. (2011). Mayhem on the Mekong. Science , 333 (6044), 814 818. Stone, R. (2016). Dam - building threatens Mekong fisheries. Science , 354 (6316), 1084 1085. Stonestrom, D. A., Scanlon, B. R., & Zhang, L. (2009). Introduction to special section on Impacts of Land Use Change on Water Resources. Water Resources Research , 45 . https://doi.org/W00a0010.1029/2009wr007937 Suweis, S., Konar, M., Dalin, C., Hanasaki, N., Rinaldo, A., & Rodriguez - Iturbe, I. (2011). Structur e and controls of the global virtual water trade network. Geophysical Research Letters , 38 (10), n/a - n/a. https://doi.org/10.1029/2011GL046837 Sverdrup - Jensen, S. (2002). Fisheries in the Lower Mekong Basin: Status and Perspectives . Mekong River Commission . Phnom Penh. Retrieved from www.mrcmekong.org Tian, X., Sarkis, J., Geng, Y., Qian, Y., Gao, C., Bleischwitz, R., & Xu, Y. (2018). Evolution of Environment International , 121 , 178 188. https://doi.org/10.1016/j.envint.2018.09.011 Van Zalinge, N. (n.d.). Data Requirements for Fisheries Management in the Tonle Sap. Retrieved 107 March 28, 2019, from http://www.fao.org/3/ad070e/ad070e0a.htm van Zalinge, N., Degen, P., Pongsri, C., Nuov, S. , Jensen, J. G., Nguyen, V. H., & Choulamany, X. (2003). The Mekong River System. In R. L. Welcomme & T. Petr (Eds.), Proceedings of the second international symposium on the management of large rivers for fisheries: Volume I . Phnom Penh. Retrieved from ht tp://www.fao.org/3/ad525e/ad525e0l.htm Ward, P. J., Beets, W., Bouwer, L. M., Aerts, J. C. J. H., & Renssen, H. (2010). Sensitivity of river discharge to ENSO. Geophysical Research Letters , 37 (12), n/a - n/a. https://doi.org/10.1029/2010GL043215 Welcomme, R. , & Winfield, I. J. (2012). Review Of The State Of The World Fishery Resources: Inland Fisheries . Journal of Fish Biology (Vol. 81). https://doi.org/10.1111/j.1095 - 8649.2012.03467.x Wild, T. B., Reed, P. M., Loucks, D. P., Mallen - Cooper, M., & Jensen, E. D. (2019). Balancing Hydropower Development and Ecological Impacts in the Mekong: Tradeoffs for Sambor Mega Dam. Journal of Water Resources Planning and Management , 145 (2), 05018019. https://doi.org/10.1061/(ASCE)WR.1943 - 5452.0001036 Winemiller, K. O., McIntyre, P. B., Castello, L., Fluet - Sáenz, L. (2016a). Balancing hydropower and biodiversity in the Amazon, Congo, and Mekong. Science , 351 (6269 ), 128 129. https://doi.org/10.1126/science.aac7082 Winemiller, K. O., McIntyre, P. B., Castello, L., Fluet - Sáenz, L. (2016b). Balancing hydropower and biodiversity in the Amazon, Congo, and Mekong. Science , 351 (6269), 128 129. https://doi.org/10.1126/science.aac7082 Xue, Z., Liu, J. P., & Ge, Q. (2011). Changes in hydrology and sediment delivery of the Mekong River in the last 50 years: connection to damming, monsoon, and ENSO. Earth Surface Processes and Landfo rms , 36 (3), 296 308. https://doi.org/10.1002/esp.2036 Yamazaki, D., Baugh, C. A., Bates, P. D., Kanae, S., Alsdorf, D. E., & Oki, T. (2012). Adjustment of a spaceborne DEM for use in floodplain hydrodynamic modeling. Journal of Hydrology , 436 437 , 81 91. h ttps://doi.org/10.1016/j.jhydrol.2012.02.045 Yamazaki, D., de Almeida, G. A. M., & Bates, P. D. (2013). Improving computational efficiency in global river models by implementing the local inertial flow equation and a vector - based river network map. Water R esources Research , 49 (11), 7221 7235. https://doi.org/10.1002/wrcr.20552 Yamazaki, D., Kanae, S., Kim, H., & Oki, T. (2011). A physically based description of floodplain inundation dynamics in a global river routing model. Water Resources Research , 47 (4). https://doi.org/10.1029/2010WR009726 Yamazaki, D., Sato, T., Kanae, S., Hirabayashi, Y., & Bates, P. D. (2014a). Regional flood 108 dynamics in a bifurcating mega delta simulated in a global river model. Geophysical Research Letters , 41 (9), 3127 3135. Yamazaki , D., Sato, T., Kanae, S., Hirabayashi, Y., & Bates, P. D. (2014b). Regional flood dynamics in a bifurcating mega delta simulated in a global river model. Geophysical Research Letters , 41 (9), 3127 3135. https://doi.org/10.1002/2014GL059744 Zhang, Y., Zhang , J., Tang, G., Chen, M., & Wang, L. (2016). Virtual water flows in the international trade of agricultural products of China. Science of The Total Environment , 557 558 , 1 11. https://doi.org/10.1016/j.scitotenv.2016.02.166 Zhao, F., Veldkamp, T. I. E., Fr (2017). The critical role of the routing scheme in simulating peak river discharge in global hydrological models. Environmental Research Letters , 12 (7), 75003. Ziv, G., Baran, E., Nam, S., Rod ríguez - Iturbe, I., & Levin, S. A. (2012a). Trading - off fish biodiversity, food security, and hydropower in the Mekong River Basin. Proceedings of the National Academy of Sciences of the United States of America , 109 (15), 5609 5614. https://doi.org/10.1073/ pnas.1201423109 Ziv, G., Baran, E., Nam, S., Rodríguez - Iturbe, I., & Levin, S. A. (2012b). Trading - off fish biodiversity, food security, and hydropower in the Mekong River Basin. Proceedings of the National Academy of Sciences , 109 (15), 5609 5614. https:// doi.org/10.1073/pnas.1201423109