ANALYSIS OF TREE COMMUNITY COMPOSITION, DIVERSITY AND NATURAL REGENERATION IN MIOMBO WOODLANDS OF MALAWI By Francis Kamangadazi A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Forestry - Master of Science 2019 ABSTRACT ANALYSIS OF TREE COMMUNITY COMPOSITION , DIVERSITY AND NATURAL REGENERATION IN MIOMBO WOODLANDS OF MALAWI By Francis Kamangadazi The main aim of the study was to ascertain whether Miombo woodlands have the potential of replacing themselves in terms of demography of seedling regeneration in three different Miombo forests of Liwonde, Ntchisi and Perekezi in Malawi. The study analyzed different descriptors that characterize the woodlands including: species composition, div ersity, dominance, evenness and carbon in order to achieve the main objective. The results reveal a normal reverse J - shaped, which is an indicator of increased seedling populations , active regeneration and recruitment across the reserves hence indicating a steady and expanding communities in the woodlands. However, the forests will physiognomically change. This is because analysis of the mature tree cohorts against the seedling and young tree cohorts indicate that the species profiles are in transition for a greater change in terms of structure. The study suggests that the current scenarios will compromise sustainability of Miombo woodlands species. Therefore, it is vital that urgent attention and interventions are done to address the current situation. iii Dedicated to my beautiful wife Rebecca and son Theophilus. iv ACKNOWLEDGEMENTS First and foremost, I am thankful to God Almighty for His blessings and mercies during the study period. I am very grateful to my Guidance Committee, in particula r Dr. David Skole (Chair of Guidance C ommittee), and Dr. Pascal Nzokou and Dr. Leo Zulu (Co - supervisors), Dr. Alfred Chioza (resident supervisor) . I would like to acknowledge the generous technical training and guidance from Mr. Jay S a m e ck . I provide s pec ial recognition for the financial support I received under a scholarship from USAID through the project, Protecting Ecosystems and Restoring Forests in Malawi (PERFORM) . Additional critical financial support was provided by the laboratory of Dr. Skole and Michigan State University . Appreciation extends to inventory teams that participated in initial 2016 forest inventory conducted in Liwonde, Ntchisi and Perekezi Forest reserves. Dan Ndalowa, I salute you for being the drive throughout this journey. A spe cial t hank you to my parents, brothers, relations, workmates and friends for your prayers, encouragement and moral support. v TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ................................ vii LIST OF FIGURES ................................ ................................ ................................ .............................. viii KEY TO ABBREVIATIONS ................................ ................................ ................................ .................... ix INTRODUCTION ................................ ................................ ................................ ................................ . 1 CHAPTER 1: BACKGROUND, OBJECTIVES AND RESEARCH QUESTIONS ................................ .................. 4 1.1 Background Information ................................ ................................ ................................ ........... 4 1.2 Problem statement ................................ ................................ ................................ ................... 5 1.3 Significance of the study ................................ ................................ ................................ ........... 6 1.4 Objectives ................................ ................................ ................................ .............................. 10 1.4.1 Main objective ................................ ................................ ................................ ........................... 10 1.4.2 Specific objectives ................................ ................................ ................................ ...................... 10 1.5 Research Questions ................................ ................................ ................................ ................ 11 CHAPTER 2: LITERATURE REVI EW ................................ ................................ ................................ ..... 12 2.1 Composition ................................ ................................ ................................ ........................... 12 2.2 Carbon storage ................................ ................................ ................................ ....................... 14 2.3 Structure ................................ ................................ ................................ ................................ 15 2.4 Regeneration ................................ ................................ ................................ .......................... 18 2.4.1 Coppicing ................................ ................................ ................................ ................................ ... 18 2.4.2 Seedling recruitment ................................ ................................ ................................ ................. 19 CHAPTER 3: METHODS ................................ ................................ ................................ ..................... 22 3.1 Study procedure ................................ ................................ ................................ ..................... 23 3.2 Study sites ................................ ................................ ................................ .............................. 24 3.2 Forest Inventory ................................ ................................ ................................ ..................... 26 3.2.1 Determination of sample frame, number of plots and their location ................................ ....... 26 3.2.2 Establishing cluster plots ................................ ................................ ................................ ........... 27 3.2.3 Establishing sample plot areas ................................ ................................ ................................ ... 27 3.2.3 Data Collection ................................ ................................ ................................ ........................... 28 3.3 Analysis of Species Composition and Diversity Data from the Forest Inventory ......................... 28 3.3.1 Composition ................................ ................................ ................................ ............................... 28 3.3.2 Shannon - ................................ ................................ .................... 29 3.3.3 Simpsons Index of Dominance (D) ................................ ................................ ............................. 29 3.3.4 Evenness (E) ................................ ................................ ................................ ............................... 30 3.3.5 Biomass and Carbon estimations ................................ ................................ ............................... 30 3.4 Analysis of Size and Age Class Data, including regeneration ................................ ..................... 31 CHAPTER 4: RESULTS ................................ ................................ ................................ ........................ 36 4.1 Analysis of Species Composition and Diversity ................................ ................................ ......... 36 4.1.1 Species composition and richness ................................ ................................ ............................. 36 vi 4.1.2 Species Diversity ................................ ................................ ................................ ........................ 38 4.1.3 Carbon estimation ................................ ................................ ................................ ..................... 39 4.2 Analysis of Size C lass and Species regeneration ................................ ................................ ....... 34 4.2.1 Analysis of size class ................................ ................................ ................................ ................... 34 4.2.2 Species regeneration ................................ ................................ ................................ ................. 35 CHAPTER 5: DISCUSSION ................................ ................................ ................................ .................. 39 5.1 Analysis of Species Composition and Diversity Data from the Forest Inventory ......................... 39 5.1 .1 Species composition and richness ................................ ................................ ............................. 39 5.1.2 Species Diversity ................................ ................................ ................................ ........................ 39 5.1.3 Carbon estimation ................................ ................................ ................................ ..................... 40 5.2 Analysis of Size class and regeneration ................................ ................................ .................... 41 5.2.1 Analysis of size class ................................ ................................ ................................ ................... 41 5.2.2 Species regeneration ................................ ................................ ................................ ................. 41 CHAPTER 6: CONCLUSION ................................ ................................ ................................ ................. 43 APPENDICES ................................ ................................ ................................ ................................ .... 44 APPENDIX A: FIELD INVENTORY TOOLS ................................ ................................ ......................... 45 APPENDIX B: TREE SPECIES COMPOSITION IN THREE FRS ................................ ............................... 48 REFERENCES ................................ ................................ ................................ ................................ .... 51 vii LIST OF TABLES Table 1: Geographic location, silvicultural classification and climatic conditions ................................ ...... 26 Table 2: Biodiversity indices and growth parameters for tree species in three study sites ....................... 38 Table 3: Kachamba allometric equation 1 summary of calculations for carbon ................................ ........ 39 Table 4: Field inventory equipme nt ................................ ................................ ................................ ............ 46 Table 5: Species composition in three FRs ................................ ................................ ................................ .. 48 viii LIST OF FIGURES Figure 1: Map of Malawi showing l ocation of study sites ................................ ................................ .......... 25 Figure 2: Seedling gain recruitment model ................................ ................................ ................................ . 32 Figure 3: Diameter class distribution in Bereku forest Tanzania ................................ ................................ 3 3 Figure 4: Model for mature trees size class distribution ................................ ................................ ............ 34 Figure 5: Model for young trees size class distribution ................................ ................................ .............. 34 Figure 6: Tree species composition in Liwonde FR ................................ ................................ ..................... 36 Figure 7: Tree species composition in Ntchisi FR ................................ ................................ ........................ 37 Figure 8: Tree species composition in Perekezi FR ................................ ................................ ..................... 38 Figure 9: Individual size class distribution in Liwonde FR ................................ ................................ ........... 34 Figure 10: Individual size class distribution in Ntchisi FR ................................ ................................ ............ 34 Figure 11: Individual size class distribution in Perekezi FR ................................ ................................ ......... 35 Figure 12: Mature trees cohort abundance in Liwonde FR ................................ ................................ ........ 36 Figure 13: Young trees cohort abundance in Liwonde FR ................................ ................................ .......... 36 Figure 14: Mature trees c ohort abundance in Ntchisi FR ................................ ................................ ........... 37 Figure 15: Young trees cohort abundance in Ntchisi FR ................................ ................................ ............. 37 Figure 16: Mature trees cohort abundance in Perekezi FR ................................ ................................ ........ 38 Figure 17: Young trees cohort abundance in Perekezi FR ................................ ................................ .......... 38 Figure 18: Recording Standard Form for Sample Cluster ................................ ................................ ........... 45 Figure 19: Cluster plots ................................ ................................ ................................ ............................... 47 Figure 20: Concentric plot ................................ ................................ ................................ ........................... 47 ix KEY TO ABBREVIATIONS DoF Department of Forestr y DBH Diameter at Breast Height D Dominance FR Forest Reserve Ha hectare Shannon - Wiener Index of Diversity E Evenness MRV Measurement Reporting and V erification NFMS National Forest Management Strategy PERFORM Protecting Ecosystems and Rest oring Forests in Malawi REDD+ Reducing Emissions from Deforestation and Degradation SOPs standard of procedures Spp species 1 INTRODUCTION The Miombo woodland ecosystem is the most extensive vegetation type in Africa, covering an estimated 2.7 million km 2 in regions receiving greater than 700 mm mean annual rainfall on nutrient - poor soils (Campbell et al. 1996). It is the most ecologically widespread woodland in Africa after the tropical rainforests. Miombo is a type of tropical woodland which is domin ated by the genera Brachystegia, Julbernadia and Isoberlinia . The three closely related genera are from the legume family Leguminaceae (White 1983). These woodlands cover vast areas of Africa stretching from Angola through Zimbabwe, Zaire to Mozambique, th e entire Zambia, Tanzania and most of Malawi. It is estimated that 1.4 million ha of woodlands are cleared annually, leading to a loss of carbon stocks, biodiversity, and, through soil degradation , loss of plant nutrients (Campbell, 2007). Although there are no reliable data on the area of Miombo that is degraded, it is believed to be more extensive than the area cleared outright. Anthropogenic activities play an important role in the land use dynamics and ecological impacts in Miombo woodlands (Chidumayo 1989a). Much of the threat to Miombo woodland comes from land use activities that either degrade or convert these forests. Charcoal production, firewood collection for subsistence use and for tobacco curing, conversion of woodlands to farmland, and seasona l forest fires are among the major drivers of deforestation and forest degradation in the Miombo region (Campbell, et al., 2006). Habitat loss due to deforestation reduces not only the number of species in the ecosystem but also the number and extent of pl aces where species coexist. Syampungani et al. (2009) observed current intensive exploitation of Miombo 2 It is this nexus of con cerns that knowledge of regeneration and succession in Miombo woodlands are key to their sustainability. Forest regeneration in Miombo woodlands is critical in maintaining biomass and carbon sequestration (May, 2013). Regeneration consists of stump/root s ucker shoots and recruitment from old stunted seedlings already present in the grass layer at the time of cutting (Chidumayo and Frost, 1996; Ernst, 1988). Seedling abundance is crucial in the maintenance of tree populations in these woodlands (Chidumayo 1 991). Seedlings of Miombo trees generally grow slowly even in the absence of shoot die - back (Chidumayo 1992a). Several studies have reported that Miombo species regenerate largely through coppice regrowth and root suckers rather than through seeds. This is because of the prolonged seedling phase, where growth is concentrated in the roots rather than shoot growth. (Chidumayo 1993a). This leads to analysis of two critical questions: (1) does the Miombo regenerate in the sense that their composition, diversit y and biomass is holding steady, replacement is occurring, and the forest contains younger cohorts/classes that will readily replace the older classes? (2), if 1 is true, is the structure species composition in the younger classes the same as the older cla sses they will replace? will the structure of the forest be maintained in the future? It is possible to ascertain indicators of future forest stature and composition by an analysis of the seedling and young classes of trees measured in a forest inventory. In 2016 the GoM conducted a systematic forest biomass and carbon inventory in three forest reserves. Although the primary aim of the inventory was to measure carbon stocks, detailed data were collected on the species composition of individual trees that we re measured. These measurements were tallied by size class and included an explicit measurement count of individuals in the seedling class as well as large diameter classes. It is there fore possible to use the carbon inventory dataset to derive both 3 biomas s and tree species diversity (composition) by size and age classes. The objectives on this research are thus twofold: (1) assess the biomass and carbon stocks in the woodland, including the contribution from different age classes, and (2) assess the specie s composition in the woodland, including contribution from different age classes to evaluate the regeneration potential of these systems. This work is important because considerable policy and management emphasis is being placed on protecting the Miombo wo odlands of Malawi from deforestation and degradation by human land use i.e. culling of trees for fuelwood and charcoal, and outright clearing of trees for cultivated land. However, there has been less emphasis on measures to protect the forests from unders tory grazing or fire, which tends to remove seedling and younger classes of trees. It is also suspected that climate change could affect the composition of the seedling and younger classes, either through increased fire or by changes in site conditions. Si milarly culling of trees for charcoal is often selective, thus emphasizing a direct loss of seed banks for particular species. Thus, even with effective or significantly reduced deforestation and degradation , effects of grazing, climate change, fire and se ed bank disruptions could have a long - term, and perhaps undetectable strong influence on system demography, recruitment and regeneration in terms of both stature (biomass) and ecological composition. Even protected forests could be on a downward path of ch ange that would not be readily detected for decades in these fragile woodland ecosystems. With a detailed view of age classes and composition, evaluate the future health , suggest management recommendations for these forests long term survivability. 4 CHAPT ER 1: BACKGROUND, OBJECTIVES AND RESEARCH QUESTIONS 1.1 Background Information Deforestation and forest degradation in developing countries results in 20 percent of global anthropogenic CO 2 emissions, and are the major source of CO 2 emissions after the fos sil fuels usage (REDD Research and Development Center, 2012). Dry and semi - dry tropical forests experience some of the highest rates of change among the world's forests (Vicharnakorn et al., 2014). Tropical forest lands comprise the largest proportion of t of global forest area. Forest resources in most African countries are threatened by various natural and anthropogenic disturbances, especially deforestation and degradation due to overexploitation and land conversion (Gonça lves et al., 2017). The problem of unsustainable use of forest resources is of increasing concern in many African countries, due to forest land use and management practices that are causing deforestation and forest degradation . There has been significant attention given to dense closed canopy tropical forests, but less focus on semi - dry woodlands such as the Miombo ecosystem. Parrotta et al., (2012) noted that forest degradation in the tropics and sub - tropics has a large negative impact on terrestrial biod iversity. The forests in Malawi fall in this category; they too are threatened by loss of tree diversity due to high forest degradation rates ( Smith, 2016) . Tree diversity conservation helps ecosystems and people cope with climate change (Perrings, 2010). Forests are degraded in two ways. First, the loss of individual mature trees for harvesting fuelwood and charcoal making, as well as intense fires, results in the transformation of the natural woodlands into degraded scrubland or grasslands with much lowe r levels of biodiversity and biomass; the loss of biomass is a source of carbon emissions (Ribeiro et al., 2013). Second, 5 the much lower intensity effects of grazing, fire, and loss of seed sources depletes the potential for these forests to provide a recr uitment class to regenerate the forest when mature trees are lost to mortality. Both human pressure and climate change can deplete the recruitment class. Even with a sufficiently large number of recruitment seedlings or juvenile trees, human degradation pr essures can alter the tree composition and diversity in latter generations. 1.2 Problem statement Studies that have been conducted on semi - dry tropical Miombo woodlands have focused on land use change (Chidumayo, 2002) and selective logging (Schwartz and Caro, 2003). In Malawi, Miombo woodlands are in a state of rapid change; these woodlands are under increasing pressure as more areas are converted or transformed. Direct causes include fuel wood harvesting, charcoal production, extraction of construction m aterial, burning and conversion to cultivated fields (Desanker et al., 1997). However, there has been less attention given to human caused influences on the demography of intact forests and forests that have only modest degradation . Efforts to reduced defo restation and degradation have often overshadowed consideration of subtle effects on ecosystem demography, particularly recruitment dynamics that come from both endogenous factors (e.g. light understory grazing) and exogenous factors (e.g. climate change e ffects on site conditions). Backéus et al., (2006) note against this backdrop of factors in intact forest, demography and successional dynamics in woodland systems are poorly understood, and it is difficult to predict the effects on future forest stature and composition from these influencing factors the ecosystem response to different forms of low gr ade understory disturbance or long - term response to climate change, particularly as they impact recruitment and regeneration. 6 This study aims to understand d ifferent forest structure dynamics in Miombo woodlands, considering recruitment potential of seedling population to replenish older cohorts of mature trees in terms of a) tree abundance and b) tree species composition. Thus, this study first aims to examin e whether there are enough individuals in the recruitment classes to sustain a healthy forest of stature comparable to today. This study also aims to understand the course of succession based on the composition of younger trees being consistent with a futu re composition that is the same as today. If not does this create a possibility of redefining the Miombo woodlands differently than that based on species of the genera of Brachystegia, Julbernardia and Jsoberlinia. Indeed, natural regeneration is important to maintaining forest condition, but knowledge on regeneration dynamics of Miombo forests is insufficient (Piiroinen, et al., 2008). Therefore, it was important to understand whether understory species recruitment is sufficient to maintain and replace a m atured forest. For instance, is mortality balanced with recruitment for forest stability and maintenance of composition and diversity. It is against this background t hat this research study analyzes tree species composition and natural regeneration in Miom bo woodlands in Malawi using detailed data from and extensive forest inventory that was conducted in three different landscapes: Liwonde, Ntchisi and Perekezi forest reserves in Malawi which are also being examined in the USAID project, Protecting Ecosyste ms and Restoring Forests in Malawi (PERFORM). 1.3 Significance of the study This project looks at composition in a particular way with specific reference to the composition characteristics of abundance and diversity across age classes to evaluate long term assessment of the forest structural sustainability, and hence the long term viability of the Miombo . There is 7 considerable, and important, emphasis in climate mitigation and sustainable forest management discourse on reducing forest loss and degradation f rom land use and cover change. It is obvious why this is important in terms of taking positive aim at reducing emissions. But even with highly effective management interventions aimed at reducing deforestation and degradation , subtle effects from understor y grazing, fire or climate change may have insidious long term impacts for biomass collapse. For instance, in a classical study of long term plots in tropical wet forests in the Amazon Philips and Gentry (1994) found that changes in the dynamics of mortali ty and recruitment were leading to important loss of forest biomass as new recruitment was not able to keep up with losses from mortality over a period of decades. This has lead researchers to ask important questions about less intensive influences on fore sts than those from land use alone. In the Philips and Gentry (1994) case it was believed to be climate change was reducing seedling survivorship and changing the demographics of the forest. This study recognizes the importance of land use and cover change in Miombo woodlands of southern Africa. Yet at the same time, more - subtle influences, such as low grade selective harvest of charcoal - specific species, or livestock grazing, or climate change influence on edaphic conditions are also present even in high p rotected forests. These impacts have more of an influence on species composition and biomass indirectly by altering patterns or recruitment through destruction of the seedlings and young trees , leading to a nearly invisible form of forest degradation over demographic cycles . Thus, by looking at both the species - specific data from a forest carbon invento ry across general age classes, the study ha s attempted to explore this problem of demographic - scale changes in both stature (biomass) and composition (diver sity) . This will be vital to make a broad preliminary, but evidence - based, evaluation of the future of these woodlands in a way that would provide 8 information for restoration or other management interventions that would otherwise be overlooked. The focus i s on age - dependent factors, considering recruitment potential from data on seedlings and young trees. The study examines two scenarios : firstly, is to attempt obtaining an expected result that of a normal J - curve, applicable across all age and size classe s . This would be realized by answering the question, does distribution follow an expected J - curve where regenerating small - tree classes are high relative to the large - tree mature classes? As shown in Figure 2, which is conceptual framework of an inverted J - curve model depicting presence of larger number of seedlings and regenerating small - tree classes at 5 cm DBH as compared to fewer mature tree classes at 40 cm DBH . This will further be achieved by comparing the study results with similar studies from othe r countries such as Tanzania and Zambia. On the other hand, figures 4 and 5 help the study answer the second question, does species abundance and composition change between small - tree regenerating classes and large - tree mature classes ? The study consider s frequency of trees in each species, along a rank ordering from the species with the most on the left and the species with the least on the right. This shows the distribution o f species and the overall composition of the forest. The differences in the profiling will help conclude that the two age cohorts have different species composition distribution. This study is a preliminary foray into these complex questions. It focuses on an examination of readily available and basic data from an extensive forest carbon inventory in three forest reserves. It is an exploratory and prefatory examination of evidence that exists recorded in the inventory data. The study propose s some simple an d straightforward precepts about the relationship between abundance and composition of the recruitment class compared to the 9 mature classes of trees. The recruitment class of trees is represented in the inventory data as seedlings and very small trees. Age is assumed to be directly related to size, and although there is no precise data on specific ages to the year, the dataset separate s young from old tree classes. The mature class is represented in the inventory data as the larger class of trees. The aim is a first order comparison of the recruits and the older trees . However, the inventory dataset falls short of data nor the time to invoke a detailed species - specific demographic survivorship model. On the other hand, a close examination of the inventory d ata can shed light on potential outcomes of forest demography over the long term (10 years or more) in the most general way. I t is recognize d that the problem of disturbance effects on recruitment - mortality dynamics is likely to differ across different scales of analysis. The researcher is aware that both mortality in the older classes of trees and selected culling of trees by harvesters may be spatially localized. Likewise, it is understood that recruitment disturbances from grazing and other factors m ay be localized. This study is aimed at a n indicative examination at the scale of whole reserves, which would be a necessary first pass for any subsequent detailed spatial analysis, using more data and more time. The study recognize s that forest growth an d replacement is a multi - factor process, and involves aspects of both demography and physiology. As such this study is not able to consider the dynamics of coppicing or vegetative propagation as a form of regeneration, which would also be important as a fo llow up. Nonetheless this focus on demography is important by itself and provides a first step in narrowing the examination of a question that simply has not been addressed before in Malawi Miombo . T he limitation s to the current study are thus the basis fo r future work, specifically looking at coppice regrowth given that this is a major form of recruitment 10 in such disturbed forests. T his assessment is the starting point for an important new approach to forest management and conservation in the Miombo woodla nds of Malawi. A particular composition of genera defines the Miombo woodlands ( Brachystegia, Julbernardia and Isoberlinia ). Therefore, understanding species composition and diversity will be of great necessity towards identification of tree species that have superior economic value as well as endangered. In addition, evaluation of understory species regeneration status will be needed to manage and restore degraded intact forests, across the selected forest landscapes of Liwonde, Ntchisi and Perekezi. I n this case, the results will offer Department of Forestry (DoF) the options on efforts to develop its own national forest management strategy (NFMS) for Miombo forest landscapes in the country as whole. 1.4 Objectives 1.4.1 Main objective The main aim of this study is to analyze tree community composition, diversity and natural regeneration in Miombo woodlands in Malawi. 1.4.2 Specific objectives The study has the following specific objectives, which were to: Examine the relative abundance of recruitment class trees compared to mature class trees, by using measures of tree community composition, diversity, dominance, evenness and carbon in three different Miombo forests in Malawi Examine the composition similarities and differences between recruitment clas ses and mature classes by using measures of abundance as well as stocks of biomass and other parameters from a forest inventory in three different Miombo forests in Malawi. 11 1.5 Research Questions The following are the questions this study was try to answer ing: What is tree community composition, diversity, dominance, evenness, and carbon in three different Miombo woodlands in Malawi? Are the forests replacing themselves in terms of demography of seedling regeneration in three different Miombo forests in Mal awi? 12 CHAPTER 2: LITERATURE REVIEW Miombo woodland s are a dry forest that is characterized by the dominance of trees in the genera Brachystegia, Julbernardia and Jsoberlinia . This woodland type covers an estimated 2.7 million km2 in southern, central and eastern Africa (Frost, 1996). Miombo - woodland area has decreased substantially because of unsustainable management practices (Piiroinen, et al., 2008). Tree diversity is an important driver of forest ecosystem functioning (Haase et al., 2015). According to Coder, (2013) tree diversity is the number of tree species per acre plus their relative abundance per acre and tree diversity is limited by loss of total habitat area. Chidumayo (1997) indicated that natural regeneration in Miombo is influenced by the typ e of harvesting. Therefore, observing how Miombo woodlands are exploited will help understand its regeneration potential other than through harvesting alone. 2.1 Composition Vegetation composition refers to the floristic characteristic of the vegetation a nd can be measured by species richness, diversity and evenness (Begon et al., 1990 in Lumbwe, 2010). Dominance of the genera Brachystegia, Julbernardia and Isoberlinia makes Miombo woodland floristically distinct from other African woodlands. However, fact ors that favour this dominance is an interesting but yet largely remain unanswered question. Although data on the Miombo woodlands is limited, several studies report clear trends in other biomes which show that change in climate has already altered the com position of many ecosystems. Some observational studies have already documented species turnover and attendant changes in species richness in the tropics (Phillips et al. 2008 in Shilima et al., 2011). 13 Miombo woodland is divided into dry and wet Miombo wo odland. Dry Miombo woodland occurs in southern Malawi, Mozambique and Zimbabwe, in areas receiving less than 1000 mm rainfall annually. Canopy height is less than 15 m and the vegetation is floristically impoverished. Wet Miombo woodland occurs over much o f eastern Angola, northern Za mbia, south western Tanzania, central and northern Malawi in areas receiving more than 1000 mm rainfall per year. Canopy height is usually greater than 15 m, reflecting the generally deeper and moister soils which create favour able conditions for growth. The vegetation is floristically rich and includes nearly all of the characteristic Miombo species (White, 1983). et In other studies, species composition was examined in greater detail using the Importance Value Index (IVI). The IVI describes the floristic structure and composition of the woodland and has been used frequently in Miombo systems (e.g. Kalaba et al., 2013; Giliba et al., 2011; Munishi et al., 2011; Mwakalukwa et al., 2014). It demonstrates how often a species occurs at a site, the size of the trees and how abundant they are. Kalaba et al., (2013) studied floristic composition in charcoal and agriculture fallows in Miombo woodland systems of Zambia. Importance values (IVI) of tree species show low presence of less fir e - resistant tree species such as Uapaca kirkiana in the initial regrowth of post - agriculture fallows. Shannon diversity indices showed high diversity in both woodlands and fallows, suggesting that though Miombo systems recover relatively fast in terms of s pecies diversity and species composition takes longer to recuperate. In Zimudzi et al., (2013) the species composition of the plots was described using Importance Value Index, which is commonly used to describe vegetation structure and species composition of forests. 14 2.2 Carbon storage The Miombo woodlands have a great potential to either add to the growing carbon dioxide content of the atmosphere, or help reduce it (Frost, 1996). Given the importance of the Miombo woodlands as a reservoir of above - and be low - ground carbon, it presents potential for implementation of REDD+ policies towards environmental sustainability and socio - economic development. However, research on carbon dynamics in Malawi is still incipient. Within this, an investigation of the chang e in vegetation dynamics in Miombo woodlands was conducted. The results indicate that Miombo have good potential for carbon sequestration woody biomass, represe nting an important potential carbon sink (Ribeiro et al., 2013). Carbon storage estimation in this study is anticipated to be similar to that in other studies, which is to demonstrate that Miombo generally holds similar carbon stocks. In previous studies, the carbon values have been derived from applying allometric equations that relate woody biomass to an index of plant size, usually diameter at breast height (DBH, 1.3 m) or the product of DBH and tree height, to the enumerated and measured stands (Malimbw i et al. , 1994). Missanjo et al., (2015) nderstanding the capacity of forest ecosystems to store carbon is fundamental in quantifying the contribution of trees to climate mitigation because they indicate the amount of carbon that can be offs et . In their study they estimate d living biomass and carbon stock for selected part of Miombo woodland in Chongoni forest reserve with the purpose of providing data for sustainable forest management and baseline d ata for carbon monitoring. R esults show si gnificant amount of biomass and carbon stock in the reserve. 15 B iomass estimation models can be used to determine total dry woody biomass from stem measurements hence will have useful applications for management of Miombo woodland (Grundy et al., 1994). Esti mation of forest biomass is the first step towards calculation of carbon stocks. Due to the natural capacity of trees to sequester carbon dioxide, Miombo woodlands are considered an important element in global climate change mitigation programs such as the REDD+ Kachamba et al., (2016). In essence, models developed by Kachamba et al., (2016) will be used since their applicability other than the initial study remains to be tested. In their study they present general (multiple tree species from several sites) above - and belowground biomass models for trees in the Miombo woodlands of Malawi. The modelling was based on 74 trees comprising 33 different species with DBH and total tree height (ht). Trees were collected from four silvicultural zones covering a wide range of conditions. Model performance indicated that the models can be used over a wide range of geographical and ecological conditions in Malawi. 2.3 Structure The biophysical characteristics of Miombo woodland, provide a framework for understanding whet her mature undisturbed Miombo is physiognomically a closed deciduous woodland (Walker 1981, Huntley 1982, in Frost, 1996). Vegetation structure is the horizontal and vertical distribution of plant biomass. Horizontal distribution of plant biomass refers t o the pattern of spacing of plant stems on the ground. It can be described by plant density and basal area cover. Plant density is the number of plant stems in a unit area while basal area cover is the area or proportion of the soil surface occupied by bas es of plants. The vertical distribution of plant biomass refers to the height distribution of plants. 16 It can be described by plant height and canopy cover. Canopy cover is the proportion of ground covered by the vertical projection of the tree canopy (Hea dy & Heady, 1982 in Lumbwe, 2010). The structure of Miombo woodland appears superficially to be relatively uniform over large regions, suggesting a broad similarity in key environmental conditions. Woody plants comprise 95 - 98% of the aboveground biomass of undisturbed stands; grasses and herbs make up the remainder. Mbwambo (2008) noted structure has three main components: species composition, tree size distribution, and spatial distribution. The number, proportion, and spatial status of individual trees of different species, forms, sizes and canopy layers are the primary components. Changes in species composition related to climate change have led to changes in the physical and trophic structure of ecosystems, with resulting further effects on ecosystem fu nction, and composition. Although data for the Miombo woodlands is limited, examples of structural changes related to climate change have been observed in other systems, including: accelerated forest turnover and associated gap formation in the tropics (Ph illips et al. 2008 in Shilima et al., 2011). The woodlands typically comprise an upper canopy of pagoda or umbrella shaped trees; a scattered layer, often absent, of sub - canopy trees; a discontinuous understorey of shrubs and saplings; and a patchy layer o f grasses, forbs and suffrutices. The uniformity in appearance is due in part to the remarkably similar physiognomy of the dominant canopy trees, no doubt a reflection of their origins in the Caesalpinioideae family . Mwakalukwa et al., (2014) argue that mo st of forest reserves in Tanzania, biodiversity is poorly documented. Therefore, their study was conducted to assess species richness, diversity, and forest structure and to examine relationships between species occurrence and topographic and edaphic facto rs in the Gangalamtumba Village Land Forest Reserve, a dry Miombo woodland area 17 in Tanzania. In general, forest structure parameters and diversity indices indicated the forest to be in a good condition and have high species richness and diversity. Accordi ng to Shirima et al., (2015) dominant woody species can determine the structure and composition of a plant community by affecting environmental conditions experienced by other species. In this study they explored how dominant tree species affect the tree s pecies richness, diversity, evenness and vertical structural heterogeneity of non - dominant species in wet and dry Miombo woodlands of Tanzania. Shannon diversity and evenness had strong non - linear negative relationships with relative abundance of dominant tree species. There were many small individual stems from dominant and non - dominant tree species suggesting good regeneration conditions, and intensive competition affecting survival. Missanjo et al., (2014) conducted a study to determine the best silvicul tural practice to maximise natural regeneration and tree species diversity in Miombo woodland in Chongoni Forest Reserve in Malawi. The results indicate that selective thinning had higher natural regeneration, mainly because of one species, although the di fferences among silvicultural practices were not significant. Coppice with standard was observed to have the highest tree species diversity hence recommended as one of the future management options in management of Miombo woodlands in Malawi in order to ma ximize tree species diversity whilst maintaining good site cover. Miombo woodlands show potential pathways of forest recovery in terms of faster regeneration after agricultural abandonment (Goncalves et al., 2017). They aimed to assess the floristic divers ity, the species composition, and stand structure of Miombo woodlands during regeneration after shifting cultivation out in the Cusseque area of the Municipality of Chitembo 18 in south - central Angola. Shannon Diversity and Evenness were highest in mature for ests and young fallows, while the mature forest stands showed the highest species richness. 2.4 Regeneration Successional relationships in Miombo woodlands, is poorly understood, hence are important tasks for research. This includes their reactions to dif ferent kinds of disturbance but also long - term dynamics and reactions to climate change ( Backéus, et al., 2006). Most regeneration is in the form of coppice shoots, root suckers and seedlings (Grundy et al., 1994). 2.4.1 Coppicing Trees of the distinctive Miombo woodlands of south - central and eastern Africa resprout from roots and stumps once the above ground parts have been removed or killed by harvesting or fire damage (Frost, 1996). The ability of Miombo woodland species to regenerate through coppicing m itigates against the degrading effects of local harvesting practices ( Syampungani, 2009). According to Chidumayo (2013) many studies have documented high levels of regeneration through sprouting after tree cutting in African dry forests and woodlands. In a Marquesia woodland in Zambia stump sprouting ranged from 70 76% but regeneration was dominated by saplings of seedling origin (Chidumayo, 1989). In a Tanzanian Miombo woodland 83 90% of stumps sprouted (Luoga et al., This study compared and cont rasted natural regeneration by coppicing of tree species in a forest reserve and more disturbed adjacent public lands in eastern Tanzanian Miombo woodlands. Coppicing effectiveness (mean number of shoots per stump) varied among species and depended on plan t size at the time of cutting, stump height and percentage of the stand removed 19 2.4.2 Seedling recruitment Little is known concerning regeneration by saplings of seedling origin. Disturbances such as harvesting of trees induces natural regeneration. Thi s is because the germination and recruitment of young forest is enhanced through increase of gaps, light, raised soil temperature and reduced nutrient competition (Augspurger 1984 in Chamshama et al., 2004). Seed is dispersed by mechanical expulsion from t he pod, but in most cases are not propelled more than about 6 m from the parent tree. In addition, seedlings are more likely to survive in under - canopy environments where moisture levels are higher (Grundy et al., 1994). Increased number of seedlings and y oung trees as compared to mature trees is expected in the current study. This condition results into a distribution of size classes that determine relative age of individual trees hence indicating survivorship of Miombo species. The distribution of stem nu mbers per hectare in all strata follows the usual expected reversed J - shaped trend. This is an indication of good forest regeneration and recruitment trend (Phi l lip 1983). The reverse J - shaped curve of woodland structure is an indicator of a steady and exp anding population, which has more trees in the smaller classes (Peters, 1994), indicating continuous recruitment in a sustainable system (Hörnberg et al., 1995 in Jew et al., 2016). Other studies in Miombo woodland in protected areas demonstrate this struc ture (Giliba et al., 2011; Mwakalukwa et al., 2014; Shirima et al., 2011), Giliba et al., (2011) wrote about increased interest in the academic world with regards to tree and shrub species richness and diversity within the Miombo woodlands. The problem of trees dying - off is becoming an important topic in global change science, since climate change and human pressures are potentially having detrimental impacts on the balance between mortality and recruitment. Tropical forest studies in the neotropics have su ggested that there is 20 widespread decline in recruitment relative to mortality (Philips and Gentry 1994). Most models of forest management in the context of REDD+ interventions have emphasized how forest cover change affects carbon stocks with less attentio n given to the co - benefits of biodiversity and thus are inadequate for REDD+ compliance (Mbow et al, 2012). Giliba et al., (2011) argue that despite the Miombo woodland providing products and services to surrounding communities the woodland is still fairly stocked with high tree and shrub species diversity. This study assessed species richness, diversity, dominance and exploitation of tree and shrub in Bereku Forest Reserve. Results showed high Shannon - Wiener Index of 4.27 and low Simpson Diversity Index of 0.043 for the Miombo woodland of Bereku. May (2013) considered the effects of the institutional structure of forest governance and associated land uses on forest regeneration in REDD+ villages operating under community - based forest management in compariso n to centrally - managed forests. Seedling density was used as a proxy for recruitment and modeled using a generalized linear model with several environmental parameters to test for the effects of forest governance and land use on tree regeneration. Predicte d recruitment was significantly higher in community - managed forests than centrally - managed forests. This suggests that community managed forests are healthier than centrally - managed lands. According to Chidumayo (1993) Miombo woodland does not appear to h ave a soil seed bank. This implies that if seed production and germination fail to replenish the seedling pool, the population of seedlings will decline with each successive clearing of regrowth Miombo . Shoots of seedlings in Miombo woodland grow very slow ly as they initially allocate more biomass to root growth. Apparently this growth strategy may be both genetically and environmentally (fire and 21 drought) based. A comparison of ring counts of root stocks and their established shoots indicate that it takes eight to ten years for Miombo woodland seedlings to reach the sapling phase. Many plants in the seedling phase are therefore old, albeit stunted. The high light intensities that arise after woodland clearing seem to enhance the shoot growth of such stunted old seedlings. Since competition for resources ultimately determines the maximum tree density that can be supported in regrowth Miombo , a higher contribution by coppiced stumps to regeneration in successive regrowth may prolong the life of the seedling po ol even when seed germination fails to replenish it. Destruction of woody plant root stocks at charcoal spots during carbonization implies that regeneration on such spots originates solely from seed. Given the very slow growth rate of Miombo woodland seedl ings, such a regeneration process probably takes many years. It is therefore against this knowledge that species regeneration through seedling recruitment is vital to this study. The ultimate goal to the current study is to understand Miombo woodland dynam ics and successional changes with respect to potential of seedling and young recruits to replace undisturbed mature trees. An understanding of seedling recruitment potential and structural composition of trees is important in forest succession stages as we ll as sustainability in forest management. This is because coppicing system which makes no allowance for seed production may therefore result in an erosion of the genetic base (Chidumayo, 1993). 22 CHAPTER 3: METHODS This study uses data from a large forest ry inventory to provide data for analysis. The protocols followed in this s tudy, are the standard of procedures (SOPs) for field measurements, data collection and recording for carbon stock and sequestration in woodland forest systems in developing countri es. These protocols have been developed to support an inventory activity in Miombo woodlands in three pilot landscapes (Liwonde, Ntchisi and Perekezi FRs) in the Protecting Ecosystems and Restoring Forests in Malawi (PERFORM) project (USAID, 2016). Data us ed in this study came from forest inventory plots in which tree species and their allometric parameters were collected in a field campaign in 2016. This inventory was primarily designed to estimate carbon stocks for Reducing Emissions from Deforestation an d Degradation (REDD+) measurement, reporting and verification (MRV). The forest inventory included 250 sample clusters of nested fixed radius plots in the three forest reserves of the study sites mentioned above. An IPCC approved method was used to determi ne the sample allocation, where the number and geographic location of sample plot cluster for a specific level of accuracy of 10% was determined. The forest carbon inventory dataset then formed the foundation for subsequent analysis of species composition and age class distribution following a prescribed field inventory technique. It has to be mentioned that dataset is limited, it does provide data for species survivorship through natural coppicing, which according to several studies is the most profound m eans of and so the approach in this study is a preliminary indication. 23 3.1 Study procedure The procedure for this study focuses on the basic structure of the Miombo woodland including it biophysical characteristics and species composition, with referenc e to age class distribution. The aim is to understand the influence of understory disturbance and other factors that may have influenced the demography and age - dependent recruitment and regeneration of trees in the forest. The study considers the need to u se locations that are protected from large scale and intensive deforestation but may have been disturbed by more - subtle disturbances within the intact forests. To a larger extent, locations selected for study that are potentially influenced by fire, climat e, seedling grazing, because these factors are widely prevalent, and which represent the current state of protected Miombo . Even though the sites have differences in structure, the datasets provide common descriptors such as species composition, diversity measures, biomass and carbon among others to describe Miombo woodlands. The study took advantage of a recent inventory that contains data on the aforesaid biophysical characteristics. Furthermore, the study sought to compare the results with similar studi es from other countries like Tanzania on size class distribution (J - curve reference model). The essence is to observe the expected normal reverse J - curve depicting a normal and vibrant re - growing natural Miombo woodlands (Giliba et al., 2011). The most imp ortant part to this study was to get data that relates to the age classes of which the inventory provides a size class distribution by species for established trees and also for seedlings. In this case, the inventory was used to make an inference on size c lasses with respect to demography. This was generally achieved by dividing the sample population of trees into three groups: i) seedlings with very young recruits, ii) medium size trees which are middle age (juvenile) trees and iii) large trees which are m ature. 24 3.2 Study sites The study sites were selected based on their unique challenges and opportunities for PERFORM household farm size. Its dependence on the forest reserve for firewood is highest and, consequently, so is its food insecurity. Opportunities for this site are the relatively short distance to market and familiarity with crop diversification. Ntchisi is an interesting case of intermediate farm size and po pulation density; the site of much donor investment, households report higher use of woodlots and a better social and governance environment. Opportunities there are familiarity with woodlot management and a cohesive governance environment with which to im prove farm productivity and woodlot prevalence. Perekezi/Mzimba has the lowest population density, the most forest remaining, and the highest levels of land clearing and wood use as a result. Household confidence in its leadership is also lower than other sites. The relative abundance of resources to improve local governance of natural resources and the productivity of farms and woodlots, as well as to decrease clearing of land for agriculture are the main The general character istics for the sites (Figure 1 and Table 1), are based on geographical location, management regime, silvicultural classification, and climatic conditions to capture a wide range of factors influencing tree growth (Kachamba, et al., 2016). 25 Figure 1 : Map of Malawi showing location of study sites Perekezi Ntchisi Liwonde Legend Lake _Malawi 26 Table 1 : Geographic location, silvicultural classification and climatic conditions Forest Reserve Liwonde Ntchisi Perekezi Region Southern Central Northern Dis trict Machinga Ntchisi Mzimba Silvicultural zone C E E Location 15 0 13 0 12 0 35 0 34 0 33 0 Area (ha) 26,889.18 9,410 15,370 Altitude 800 m 2080 m 1500 m 1700 m 1566 m Type of soils Ferrallit ic latosols Ferruginous Sandy ferrallitic Annual Temperature 21 23 22 36 19 21 Annual Rainfall 840 960 900 1500 960 1050 Management regime co - management co - management co - management Source: Department of Forestry in Malawi 3.2 Forest Inven tory This section describes Sta ndard of Procedures (SOPs) developed by MSU which were deployed during forest field inventory exercises. Different tools were used during the forest field inventory as well as computation of the data into MRV Toolbox (Skole a nd Samek, 2016). The data tabulation sheets, basic equipment and sample plot designs needed for field measurements are shown in Appendix 1. 3.2.1 Determination of sample frame, number of plots and their location The minimum number of sample plots and sam ple allocation for a given level of accuracy, was determined using standard IPCC methods as provided by the MRV Toolbox which is derived from Winrock sample estimator tool ( http ://www.winrock.org/resources/winrock - sample - plot - calculator ). The Toolbox is used to ascertain the number of samples and their location. A random unstratified sample allocation was prescribed with e very management block having at least one (1) sample plot . 27 Sample plots adhered to the following conditions: Minimum 50 m from a road or forest edge Not be located too near or include certain geographic features such as cliffs, ravines, an d large rivers Not located across two strata 3.2.2 Establishing cluster plots Nested cluster plots were laid out in the field. The starting point for establishing the cluster is taken from the random location generated by the Toolbox sample frame. Cluster ing of plots within one area can often allow field crews to sample a large area while reducing travel time between plots. Clustering of plots at each sampling unit is often recommended for natural forest areas and areas that have been selectively logged. A cluster sample using 3 center points for 3 nested circular plots is recommended as the most suitable design to cope with access in terms of access and ease of implementation. Once at the initial latitude/longitude coordinates determined prior to field dat a collection, an additional 10 steps in the direction of travel are taken to establish the cluster center. The additional steps reduce bias in choosing the cluster center. Layout of the nested plot cluster is shown in Appendix 1.3. Each center point was mo numented by taking coordinates with aid of a geographical positioning system (GPS), inserting a piece of rebar iron rod into the ground and marking the location with a flag. 3.2.3 Establishing sample plot areas The selection of sample trees was done in fi xed radius nested circular plots, which are easy to establish in the field and can be demarcated with simple tools and procedures. The sampling 28 point design applied in the field inventory follows this scheme for a nest plots design to capture both large, s mall trees as well as seedlings. The fixed radius plot is circular with concentric plots at the radii below. The design of each nested fixed radius plot is shown in appendix 1.4. 6 m Radius Circular Plot small trees 5 15 cm DBH 12 m Radius Circular Plo t medium trees 15 30 cm DBH 20 m Radius Circular Plot big trees > 30 cm DBH In addition, there was one 2m radius clip plot used to collect seedlings for estimation regeneration estimates. This plot will be placed due East with a center at 6 m radius. 3.2.3 Data Collection For the purposes of this study, all sampled trees were identified and their diameters measured at a standard height of 1.3 m long from the foot of the tree using calibrated diameter at breast height (DBH) tapes. The data is recorded on the MRV Toolbox field data tabulation sheet (see Appendix 1.1). Thereafter, the data is computed into the MRV Toolbox designed by MSU. The data collected helps to estimate tree species composition, diversity, dominance, evenness, carbon stocks, and to e valuate species regeneration. 3.3 Analysis of Species Composition and Diversity Data from the Forest Inventory The first analytical examination of the forest inventory data considers the detailed information on species composition and tree diversity. 3. 3.1 Composition Species composition is defined as the relative contribution of individual species to the total composition of trees in each plot and for each entire forest reserve. Composition is computed 29 based on density, which is the total number of tre es sampled in hectares. The metric of richness is computed as the number of species in each plot and for each reserve. 3.3.2 Shannon - This Index is most widely used because it combines species richness the relative abundances of species (Begon et al., in the organizing of plots into high, low and medium level of tree diversity across the three forest nfluence of biotic disturbance, and the state of succession and stability in the environment (Misra 1989 in Giliba et al., 2011). Shannon - Wiener Index is expressed as: Where: = Shannon - in the area, ln = natural logarithm of the number, pi = proportion of individuals in the i th of species, n i = number of individuals in the i th sp ecies and N = total number of individual species for the site. Values range from 0 to 5, with values usually between 1.5 and 3.5. The larger the 3.3.3 Simpsons Index of Dominance (D) This index me asures the distribution of individuals among the species in a community. Dominance is defined as relative importance of a species related to degree of influence it has on ecosystem components: Soils, other plants, animals. Dominance is useful to characte rize plant communities, habitat types and ecological sites. 30 Simpson Index is given by: 2 ), n i = number of individuals in the i th species and N = total number of individual species for the site. The lower the index value, the lower the dominance and greater the value of index of dominance the lower the species diversity and vice versa in the scal e of 0 to 1 (Edward 1996; Misra, 1989 in Giliba et al., 2011) 3.3.4 Evenness (E) To have a metric more closely descriptive of diversity, evenness computed for each plot and forest reserve. Species evenness is considered as the measure of equality of abund ances in a community. It ranges from zero to one, with zero signifying no evenness and one, a complete evenness. Mathematically, evenness is the ratio between Shannon's entropy and the maximum H', which may arise with a given S (Pielou 1966, in Alatalo, 19 81). Where: E = Evenness, H' = Shannon - Wiener Index, S = number of species. 3.3.5 Biomass and Carbon estimations Above ground biomass (AGB) and below ground biomass (BGB) of a tree were estimated usi ng the following site - specific equations developed by (Kachamba et al., 2016) with the following standard inputs: diameter at breast height (DBH), total height, and tree density. Biomass is converted to carbon stocks using IPCC methods and a conversion fac tor of 0.5. 31 Below are the equations: AGB = 0.21691 x dbh BGB = 0.21691 x dbh Where: AGB and BGB are above ground biomass and below ground biomass (kg dry matte r per tree), respectively; DBH is a diameter at breast height (1.3 m above the ground level). 3.4 Analysis of Size and Age Class Data, including regeneration Two questions were posed to examine the distribution of size and age classes in each FR: Firstly, does distribution follow an expected J - curve where regenerating small - tree classes are high relative to the large - tree mature classes? To consider the second question, does species abundance and composition change between small - tree regenerating classes a nd large - tree mature classes? Figure 2 below shows conceptual framework of an inverted J - curve. The model depicts a scenario where there are larger number of seedlings and regenerating small - tree classes, the expected result is that of the normal J - curve, applicable across all age and size classes, hence seedlings with active regeneration and recruitment (Phillip , 1983). On the other hand, if the counts and dis tribution of the seedlings are fewer than in other age class counts, presumably young trees at 5 cm DBH to mature trees at 40 cm DBH. This suggests that seedling cohort alone will not replace the number of trees in mature class when they senesce and die, e specially if there is mortality as this seedling cohort continues through time. This distribution can be created by two conditions: either the loss of small regenerates as seedlings is 32 a recent phenomenon and this age class has been impacted more than the next older classes, or there are significant small trees (5cm and 10cm) that are being created not from seedlings but from sprouting of cut trees. Figure 2 : Seedling gain recruitment model The study seeks to compare the expected result on recruitment with similar studies from other countries. Active regeneration and recruitment in Miombo woodland of Bereku Forest Reserve as shown in figure 3 is a good sign of sustainability of the woodland stock which has chances of insuring susta inable supply of products and services (Giliba, et al., 2011). However, the Tanzanian study would help ask if this J - curve is maintained with the counts of seedlings are included. 0 20 40 60 80 100 120 140 160 A B C D E F G H I J K NUMBER OF INDIVIDUALS TREES SIZE AND RELATIVE AGE CLASSES 33 Figure 3 : D iameter class distribution in Bereku forest Tanzania The second question above consideration is made of the frequency of trees in each species, along a rank ordering from the species with the most on the left and the species with the least on the right. This shows the distribution of species and the overall composition of the forest. It can be shown as raw counts (frequency) or abundance or dominance. To do this we separate the young 34 Figure 4 : Model for mature trees size class distribution Figure 4 shows a distribution of tree species in the Mature class labeled as A, B, C, etc. The y axis is frequency, and it is a profile of the forest today based on the older and larger trees. Figure 5 : Model for young trees size class distribution 0 5 10 15 20 25 30 35 40 A B C D E F G Number of trees per ha Species 0 10 20 30 40 50 60 70 G F C D E A B H I J L K Number of seedlings Species 35 Figure 5 shows a distribution of tree species in the Young class labeled as A, B, C, etc. The y axis is frequency, and it is a profile of the forest today based on the younger and sm aller trees. By looking at the two profiles ( figures 4 and 5), the younger cohorts are dominated by different tree species (G and F) than the mature cohorts (A and B). The profile has changed considering the type and numbers of species encountered. Thus, by taking into account the J - curve figures of size class distributions above, we can propose that while the forest is sustaining and trees in the younger understory cohorts are quantitatively enough to replace their elders in the overstory, they are of a d ifferent profile or species composition distribution. In other words, the forest may remain, but it may look physiognomically different. This could have broader implications depending on the nature of the changes. For example, if there is a loss of the bes t charcoal species this could reduce its value as a resource, or if there is a loss of a certain tree from a biodiversity point of view it affects habitat. 36 CHAPTER 4 : RESULTS 4 .1 Analysis of Species Composition and Diversity 4.1.1 Species composition and richness A total of 8077 individual s and 122 species were identified across the three landscapes under this study. The main tr ee genera for Miombo woodlands. Total stem density estimations indicate 541 stems/ha in Liwonde FR, 200 stems/ha in Ntchisi and 3 11 stems/ha in Perekezi. Liwonde recorded 5053 individual s and 99 species, where Brachystegia bussei , Diplorhynchus condylocarpon and Julbernardia globiflora dominate as indicated in figure 6 . Figure 6 : Tree species composition in Liwonde FR 18% 13% 12% 6% 6% 5% 5% 5% 5% 3% 3% 3% 2% 2% 2% 2% 2% 2% 1% 1% Brachystegia bussei Diplorhynchus condylocarpon Julbernardia globiflora Brachystegia boehmii Brachystegia longifolia Pericopsis angolensis Uapaca kirkiana Brachystegia manga Dalbergia nitidula Pseudolachnostylis maprouneifolia Burkea africana Brachystegia microphylla Dalbergiella nyasae Parinari curatellifolia Bridelia cathartica Combretum molle Bauhinia petersiana Pterocarpus angolensis 37 In Ntchisi FR, 1603 individual trees and 61 species were recorded . Brachystegia ( boehmii, manga and spiciformis ) species dominates the forest reserves as depicted in figure 7 . Figure 7 : Tree s pecies composition i n Ntchisi FR A total of 2021 individual trees and 50 species were recorded in Perekezi. Figure 8 shows Brachystegia boehmii, Julbernadia paniculata and Uapaca kirkiana as the dominan t species . 17% 13% 9% 7% 7% 6% 6% 5% 4% 4% 3% 3% 3% 3% 3% 2% 2% 1% 1% 1% Brachystegia boehmii Brachystegia manga Brachystegia spiciformis Brachystegia bussei Julbernadia paniculata Diplorhynchus condylocarpon Uapaca kirkiana Brachystegia utilis Newtonia buchananii Bauhinia petersiana Markhamia obtusifolia Diospyros kirkii Combretum collinum Julbernardia globiflora Pseudolachnostylis maprouneifolia Multidentia crassa Friesoldielsia obovata Faurea rochetiana 38 Figure 8 : Tree s pecies compositio n in Perekezi FR 4 . 1. 2 Species Diversity Table 2 : B iodiversity indices and growth parameters for tree species in three study sites Parameter Liwonde Ntchisi Perekezi Species richness 99 61 50 Shannon - Wiener Diversity Index (H') 3.3323 2.7618 2.4342 Simpson's Index of Diversity (D) 0.0803 0.1716 0.1921 Species evenness ( ) 0.7236 0.6718 0.6222 Table 2 above shows species ri chness, Shannon Diversity , Simpson's Index of Diversity and Evenness indices that were calculated in the three FRs. A Shannon - 15% 12% 12% 12% 7% 7% 5% 5% 5% 4% 2% 2% 2% 2% 2% 1% 1% 1% 1% 1% Brachystegia boehmii Julbernadia paniculata Uapaca kirkiana Brachystegia longifolia Brachystegia manga Brachystegia microphylla Brachystegia utilis Brachystegia spiciformis Brachystegia taxifolia Combretum molle Faurea rochetiana Monotes africanus Brachystegia floribunda Parinari curatellifolia Faurea saligna Dalbergia nitidula Ochna schweinfurthiana Multidentia crassa Isoberlinia angolensis 39 ranging from 2.4 to 3.3 across t he three reserves was recorded . Index of dominance recorded for each site: 0.08 for Liwonde, 0.17 for Ntchisi and 0.19 for Perekezi. Species evenness recorded the following v alues: Liwonde 0.72, Ntchisi 0.67 and Perekezi 0.62. 4 . 1. 3 C arbon estimation Table 3 : Kachamba allometric equation 1 summary of calculations for carbon Table 3 shows carbon estimates in each FR. In Liwonde FR carbon is estimated to be 51.2 tC/ha, while Ntchisi FR has a record of 66.3 tC/ha and Perekezi FR recorded 6 5.3 t C /ha respectively . Forest Reserve Carbon Stock (tC/ha) Area (ha) AGB BGB Total Liwonde 26,889.18 35 16.2 51.2 Ntchisi 9,410.52 45.2 21.1 66.3 Perekezi 15,600.14 43.8 21.5 65.3 34 4.2 Analysis of Size Cla ss and Species regeneration 4.2.1 Analysis of size class Figure 9 : Individual s ize class distribution in Liwonde FR Figure 10 : Individual size cl ass distribution in Ntchisi FR 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Number of individuals Size classes 0 100 200 300 400 500 600 Number of individuals Size classes 35 Figure 11 : Individual size class distribution in Perekezi FR F ig ure s ( 9 , 10 and 1 1 ) show size class distribution for seedlings, small young trees to large mature trees , for the study sites. The resu lts reveal a consistent reverse J - shaped distribution with decreasing density and with increase in DBH in all figures. classes recorded very few trees, which may be a consequence of illegal selective logging. 4.2.2 Species regeneration Figures 1 2 to 17 are current profiles of the FRs based on the older and larger trees. The figures s how a distribution of dominant tree species in the mature cohorts which are undisputedly different from the species that are dominant in the young cohorts in all the three study sites. 0 100 200 300 400 500 600 700 Number of individuals Size classes 36 Figure 12 : Mature trees cohort abundance in Liwonde FR Figure 13 : Young trees cohort abundance in Liwonde FR 0 50 100 150 200 250 300 350 Number of individuals Species 0 50 100 150 200 250 300 350 400 Number of individuals Species 37 Figure 14 : Mature trees cohort abundance in Ntchisi FR Figure 15 : Young trees cohort abundance in Ntchisi FR 0 20 40 60 80 100 120 140 Number of Individuals Species 0 20 40 60 80 100 120 Number of individuals Species 38 Figure 16 : Mature trees cohort abundance in Perekezi FR Figure 17 : Young trees cohort abundance in Perekezi FR 0 20 40 60 80 100 120 140 160 Number of individuals Species 0 20 40 60 80 100 120 140 160 Number of individuals Species 39 CHAPTER 5 : DISCUSSION 5 .1 Analysis of Species Composition and Diversity Data from t he Forest Inventory 5.1.1 Species composition and richness Miombo woodlands characterize the three FRs with regards to species composition. Brachystegia spp and Uapaca kirkiana are the most dominant in all FRs . Brachystegia is fundamentally dominant becaus e of the presence of more species (9) than any other species in the Miombo woodlands. On the other hand, large pockets of disturbance generated Uapaca kirkiana which is a pioneer species. There were other dominant species such as Diplorhynchus condylocarpo n and Julbenardia globiflora in Liwonde and Ntchisi FRs ; Julbenardia paniculata and Combretum molle in Perekezi FR . The presence of many species in the three FRs suggests high species composition and richness , which could be attributed to presence of many seedlings and small trees . Increased rate of anthropogenic disturbance through tree felling for firewood and charcoal production could be the main cause for more seedling development . Chidumayo (1989a; Frost, 1996; Käll, 2006) reported that anthropogenic a ctivities, climate change, edaphic conditions and other factors like fire pl ay big role in the dynamics of Miombo Woodlands. 5 . 1. 2 Species Diversity The study revealed high values of Shannon - Wiener Index of , ranging from 2.4 to 3.3 and eve nness ( E ) , ranging from 0.62 to 0.72 in all the three Miombo woodlands (Table 2). An Magurran 1988 ; Barbour et al. , 1999). A maximum evenness of (1.0) arises when al l sp ecies are equally abundant (Alatalo, 1981). The forests under study are in general high diverse Miombo with 40 pockets or patches of disturbance. A high diversity of species is important to maintain the ecological stability and nutrient cycling (Käll, 2006) . High species diversity could be attributed to two factors; firstly, many tributaries and streams that empty rich organic content and mineral resources utilized by the species for growth and production (Giliba et al. , 2011; Ndah, et al. , 2013; Kamangadazi e t al. , 2016). Secondly, presence of undisturbed woodlands across the three landscapes as reported by Chidumayo (1987 ) that tree spe cies diversity remains high in Miombo once the woodland is left to recover without subjecting it to charcoal production and o ther disturbances. Tree genera that contribute to high species diversity include: Brachystegia spp, Julbenardia spp, Isoberlinia angolensis, Uapaca kirkiana, Diplorhynchus condylocarpon, Combretum spp among others. Zimudzi et al., (2013) high diversity is mainly due to the diversity of habitats like river valleys, anthills, streams, hills and slopes for species establishment. On the other hand, the index of d ominance is low, this correlated to ectomycorrhizae in their roots that facilitate up - taking of nutr ients in nutrient poor soil (Campbell, 1996). Rainfall distribution contributes significantly to the differences observed. Perekezi and Ntchisi FRs are wet Miombo s that receive more than 1000 mm of rainfall annually while Liwonde a dry Miombo receives less than 1000 mm of rainfall annually (White, 1983). 5 . 1. 3 C arbon estimation Estimated of carbon stocks in these FRs did not record very high values as compared to other countries such as Tanzania and Zambia. This is due to increased cases of disturbances ei ther natural or anthropogenic causes. Carbon stocks estimated in this study is higher than that reported in Zambia by Kalaba (2013); (43.6 tC/ha) and in Tanzania as reported by Shirima et al. (2011), 25.7 tC/ha and Munishi et al. (2010), 21 tC/ha . 41 5 . 2 Anal ysis of Size class and regeneration 5.2.1 Analysis of size class The overall size class distribution across the three FRs in small to mature Miombo woodlands, of which according to Kalaba, (2013) inverse J - shaped size class distr ibution showing more trees in the smaller size classes is an indicator of a steady and expanding population. Increased stem densities in lower DBH size classes as compared to high DBH size classes coupled with J shape of the diameter distribution curve ind icate good rates of active regeneration and recruitment (Philip, 1983; Isango, 2007). Zimudzi et al., (2013) further there is active regeneration and recruitme nt in the woodlands. Th is may lead to suggest that t he continuity of inverse J - shaped size classes signify increased seedling populations hence indicating a steady and expanding population in intact mature woodlands. Luoga et al. (2003) argue that high pro portion of stems in the smallest size classes shows that there is ongoing seedling recruitment and/or harvesting recovery . The decline in stem density in older age class could be a consequence of disturbance due to charcoal production. Low stocking of larg er diameter trees classes comes as a result of disturbance for timber, charcoaling, pole cutting and possibly due to frequent fires (Zahabu, 2008). 5.2.2 Species regeneration Ernst (1988) suggested that seed dormancy may be rare in many woody species of s outhern African Miombo woodland. Therefore, banks of stunted seedlings in the grass layer might be the major source of regeneration in Miombo woodland (Chidumayo 1989). 42 Nevertheless, t aking into account the J - curve figures of size class distributions abov e, it is suffici ent to propose that , while the forest is sustaining and trees in the younger understory cohorts are quantitatively enough to replace their elder s in the overstory, they are of varying species composition distribution. For instance, in Liwon de FR by looking at the younger cohorts dominated by Diplorhynchus condylocarpon and Julbenardia globiflora are different from mature cohorts where Brachystegia ( bussei and boehmii ) dominate. Likewise, the mature cohorts of Ntchisi FR are dominated by Bra chystegia ( boehmii and bussei ) unlike the young cohorts where Brachystegia ( manga and spiciformis ) are dominant. As for Perekezi FR Brachystegia boehmii and Julbernadia paniculata are dominant in mature cohorts whereas the young cohorts are dominated by Br achystegia longifolia and Uapaca kirkiana . The differences can be drawn because of several factors including : small number populations amongst some species in all the FRs maybe due to seedling shoot die - back caused by overgrazing and uncontrolled fires fro m charcoal production. Fire is a major constraint to sustained growth of seedlings of woody savanna plant (Chidumayo, 1991). In this case, loss of tree species could reduce its value as a resource, or if there is a loss of a certain tree from a biodiversit y point of view it affects habitat. Trapnell (1959) explained that seedlings of Miombo woodland canopy species die back because of drought and nutritional stress during the long dry season. In addition, intensive grazing in Miombo keeps grass short and red uces regeneration of young woody plants (Malmer and Nyberg, 2008). A dditionally, opening of the canopy through illegal logging might have created a conducive environment for growth of light dependent species . 43 CHAPTER 6: CONCLUSION The study has two facet s, firstly, the aim was to characterize the woodlands by looking at common parameters that are used to describe the Miombo woodlands, which include: species composition, diversity, dominance and carbon estimation. The study reveal s that there is high speci es composition coupled with high species diversity and evenness in the Miombo woodlands of Liwonde, Ntchisi and Perekezi FRs. O n the other hand , there is low species dominance across all FRs . Carbon calculations do not significantly vary from carbon values from other neighboring countries such as Tanzania and Zambia. The second part of the study, which is also the main focus for this study was to analyze the distribution of size classes in each FR, where an expected normal reverse J - curve size class distri bution has been discovered. This implies that the number of seedlings and young trees are abundant as compared to mature trees, suggesting that seedlings are sufficient in number to replace the older trees. Nonetheless, it is important to emphasize that de spite the potential of seedlings to replace the mature trees, and the demography of each FR is going to change . In essence, a different species profile will take over. This may be a consequence of overexploitation through charcoal production and overgrazin g or some other factors that this study may have overlooked. These factors have direct negative impact on biomass and carbon accumulation for individual tree species type as well as diameter class distribution, which are crucial to the sustainability of th e Miombo woodlands under study. Therefore, it is sufficing to c arry out a detailed analysis of spatial disaggregate dynamics of the Miombo woodlands. 44 APPENDICES 45 APPENDI X A : FIELD INVENTORY TOOLS Appendix 1.1: Data Tabulation Shee ts Figure 18 : Recording Standard Form for Sample Cluster 46 Appendix 1.2: Basic Equipment for Field Inventory Table 4 : Field inventory equipment ITEM PRIMARY USE QUANTITY/TEAM Diameter tape Tree dia meters 2 50 m tape Plot layout, tree crown diameters 2 GPS device Boundaries, plot layout, tree location 1 Increment borer Wood density, tree age 1 Clinometer Slope, tree height 1 Spherical Densiometer Stand density 1 Compass Navigation and cardinal directions 1 Data recorder or data sheets Record measurement data Variable Chalk Mark trees in plot already measured Variable Aluminum tags Identify trees in permanent plots Variable Sample bags Returning vegetation and soil samples to laboratory Varia ble Machete Cut vegetation and determine wood decay class 2 Calipers Measuring diameters of lying dead wood 1 Ranging rods Positioning plots 3 Flagging tape Marking trees and boundaries 12 Marking chalk Marking trees 12 Aluminum tree markers Marking trees 200 47 Appendix 1.3: layout of the nested plot cluster Figure 19 : Cluster plots Appendix 1.4: Design of each nested fixed radius plot Figure 20 : Concentric plot 48 APPENDIX B : TREE SPECIES C OMPOSITION IN THREE FRS Table 5 : Species composition in three FRs Species Liwonde Ntchisi Perekezi Acacia galpinii 1 22 Albizia versicolor 6 7 5 Allophylus africana 7 8 Annona senegalensis 32 6 Antidesma venosum 7 13 Apody tes dimidiata 2 Azanza garckeana 6 Bauhinia petersiana 58 45 Bobgunia madagascariensis 5 Boscia salicifolia 13 1 Brachystegia bussei 1206 630 809 Brackenridgea zanguebarica 8 Breonardia salicina 3 Bridelia micrantha 75 10 5 Burkea afri cana 97 1 Byrsocarpus orientalis 172 12 Carissa edulis 2 Cassine aethiopica 13 Catunaregam spinosa 151 8 Colophospermum mopane 1 Combretum molle 149 58 61 Commiphora africana 2 1 Crossopteryx febrifuga 12 Croton macrostachys 1 1 Cuss onia arborea 10 2 2 Dalbergia nitidula 159 21 26 Dalbergiella nyasae 73 Dichrostachys cinerea 39 4 24 Diospyros kirkii 40 50 8 Diplorhynchus condylocarpon 388 74 6 Dombeya rotundifolius 24 3 7 Ekebergia benguelensis 1 Elephantorrhiza goetzei 7 Erythrina abyssinica 2 4 Erythrophleum suaveolens 3 Euclea crapes 6 1 Euphorbia matabalensis 36 Fagara charybeum 3 49 Faidherbia albida 1 Faurea saligna 24 29 57 Ficus sycomorus 8 16 Flacourtia indica 97 5 3 Flueggea v irosa 2 Friesoldielsia obovata 19 Garcinia huillensis 2 8 Gardenia jovis - stonantos 1 2 Hexabux 4 Hippocratea parviflora 2 Holarrhena pubescens 17 Hugonia Orientalis 6 5 Hymenocordia acida 6 Isoberlinia angolensis 4 59 Julbernadia paniculata 118 177 Julbernardia globiflora 362 Kirkia acuminata 1 Lannea discolor 89 15 15 Lecaniodiscus fraxinifolius 1 3 Lonchocarpus capassa 15 Margaritaria discoidea 26 1 Markhamia obtusifolia 46 Maytenus heterophylla 1 Monodora j unodii 5 Monotes africanus 15 9 29 Multidentia crassa 9 21 20 Mundulea sericea 4 4 Newtonia buchananii 1 Ochna schweinfurthiana 102 27 42 Olax obtusifolia 197 9 Oncoba spinosa 6 Ormocarpum kirkii 37 1 Ozoroa insigns 24 3 Ozoroa reticula ta Parinari curatellifolia 69 11 29 Paullinia pinnata 1 Pavetta schumanniana 50 Pericopsis angolensis 161 1 8 Phyllantis Nummunilatolius 21 Phyllanthus africana 13 Pittosporum viridiflorum 9 Pleurostylia africana 3 1 Protea gaguedii 3 8 5 50 Pseudolachnostylis maprouneifolia 100 31 7 Psorospermum febrifugum 49 20 Psychotria Mahonii 67 Psydra Glnde 5 Psydrax Livida 22 Pteleopsis myrtifolia 9 Pterocarpus angolensis 52 7 Rauvolfia caffra 1 Rhus natalen sis 2 164 Rothmania engleriana 1 12 6 Rytinia Monatha 33 Schrebera alata 6 15 4 Sclerocarya birrea 1 Securidaca longepedunculata 7 Senna singueana 16 4 Sepium ellipticum 8 Steganotaenia araliacea 10 6 10 Sterculia quinqueloba 1 Stereo spermum kunthianum 14 6 Strychnos innocua 80 3 Strychnos madagascariensis 3 Syzigium cordatum 2 1 22 Tapiphyllum africana 1 Tarena Neurophylla 6 Terminalia stenostachya 34 4 Tncalysiea Suohera 4 Trema orientalis 13 Tricalysia Soutahua rae 1 Turraea nilotica 25 Uapaca kirkiana 204 74 222 Uapace Robyngu 26 Vanguelopsis canthium 7 Vangueria infausta 41 Vapaca Robynsu 4 Vernonia myriantha 8 2 Vitex doniana 36 2 Xeroderris stuhlmannii 8 Ximenia caffra 101 11 Xylop ia parviflora 9 31 Zanha africana 33 1 Ziziphus mucronata 2 Total 5053 1603 2021 51 REFERENCES 52 REFERENCES Alatalo R.V. 1981. Problems in the measurement of evenness in ecology. Vol. 37, No. 2. pp. 199 - 204. Backéus, I ., Petterson, B., Stromquist, L., Ruffo, C., 2006. Tree communities and structural dynamics in Miombo (Brachystegia_Julbernardia) woodland, Tanzania. Forest Ecol. Manage. 230, 171 178. Barbour M, Burk JH, Pitts WD, Gillians FS, Schwartz MW 1999. Terrestria l Ecology. Chicago, Illinois: Addson Wesley Longman, Inc. Begon, M., J. L. Harper and C. R. Townsend. 1990. Ecology: Individuals, Populations and Communities. Blackwell Science, MA, USA. Campbell B. 1996. The Miombo in transition: woodlands and welfare in Africa. Center for International Forestry Research. Malaysia. Campbell, B.M., Angelsen, A., Cunningham, A., Katerere, Y., Sitoe, A., Wunder, S., 2007. Miombo Woodlands Opportunities and Barriers to Sustainable Forest Management. Available: http://www.cifor.org/Miombo/docs/Campbell_BarriersandOpportunities.pdf Chamshama, S.A.O., Mugasha, A.G., Zahabu, E., 2004. Stand biomass and volume estimation for Miombo woodlands at Kitulangalo, Morogoro, Tanzania. Southern Afr. For. J. 200, 59 70. Chidumayo, E.N., 1987. Woodland structure, destruction and conservation in the Copperbelt area of Zambia. Biol. Conserv. 40, 89 100. Chidumayo E.N. 1989a. Land use, deforestation and refore station in the Zambian Copperbelt. Land Degradation and Rehabilitation, 1: 209 - 216 Chidumayo, E.N., 1991. Woody biomass structure and utilisation for charcoal production in a Zambian Miombo woodland. Bioresour. Technol. 37, 43 52. Chidumayo, E.N., 1991. Se edling development of the Miombo woodland tree Julbernardia globiflora. IAVS; Opulus Press Uppsala, Sweden. Journal of Vegetation Science 2: 21 - 26 Chidumayo, E.N. 1992a. Seedling ecology of two Miombo woodland trees. Vegetation 103, 51 - 58. Chidumayo, E.N . 1993. Responses of Miombo to Harvesting: Ecology and Management. Stockholm Environment Institute, Stockholm, p. 110. 53 Chidumayo, E.N., 1997. Miombo Ecology and Management: An Introduction. IT Publications in association with the Stockholm Environment Inst itute, London. Chidumayo, E.N. and Frost, P. 1996. Population biology of Miombo trees. In: Campbell (Ed.), The Miombo in Transition: Woodlands and Welfare in Africa. Center for International Forestry Research (CIFOR), Bogor, pp 57 71. Chidumayo, E.N.; Gumb o, D.J. 2010. The Dry Forests and Woodlands of Africa: Managing for Products and Services; Earthscan: London, UK. Chidumayo, E. N. 2002. Changes in Miombo woodland structure under different land tenure and use systems in central Zambia. Journal of biogeogr aphy, 29(12), 1619 - 1626. Coder, K.D. 2013. Tree Diversity: New Solutions to Current Homogeny & Loss. Warnell School, University of Georgia, Athens, Georgia, USA. Day, M., Baldauf, C., Rutishauser E. and Sunderland T.C.H. 2013. Relationships between tree sp ecies diversity and above - ground biomass in Central African rainforests: implications for REDD. Environmental Conservation 41 (1): 64 72. Desanker, P. V., Frost, P. G. H., Justice, C. O., & Scholes, R. J. 1997. The Miombo Network: Framework for a terrestri al transect study of land - use and land - cover change in the Miombo ecosystems of Central Africa. IGBP Global change report. Ernst, W., 1988. Seed and seedling ecology of Brachystegia spiciformis, a predominant tree component in Miombo woodlands in south cen tral Africa. Forest Ecology and Management, 25, 195 - 210. Frost, P., 1996. The ecology of Miombo woodlands. In: Campbell, B. (Ed.), The Miombo in Transition: Woodlands and Welfare in Africa. CIFOR, Bogor, pp. 11 57. Giliba, R.A., Boon, E.K., Kayombo, C.J., Musamba, E.B., Kashindye, A.M., Shayo, P.F., 2011. Species composition, richness and diversity in Miombo woodland of Bereku Forest Reserve, Tanzania. J. Biodivers. 2, 1 7. Gonçalves, F.M.P., Revermann, R., Gomes, A.L., Aidar M.P.M., Finckh, M. and Juergen s, N 2017. Tree Species Diversity and Composition of Miombo Woodlands in South - Central Angola: A Chronosequence of Forest Recovery after Shifting Cultivation. Hindawi, International Journal of Forestry Research, D 6202093, 13 pages https://doi.org/10.1155/2017/6202093 Grundy I. M., Campbell, B. M. & Frost E G. H. 1994. Spatial pattern, regeneration and growth rates of Brachystegia spiciformis and Julbernardia globiflora . Kluwer Academic Publishers. Printe d in Belgium. Vegetatio 115: 101 - 107, 54 Haase, J., Castagneyrol, B., Cornelissen, J. H. C., Ghazoul, J., Kattge, J., Koricheva, J., ... & Jactel, H. 2015. Contrasting effects of tree diversity on young tree growth and resistance to insect herbivores across t hree biodiversity experiments. Oikos, 124(12), 1674 - 1685. Ifo, S. A., Moutsambote, J. M., Koubouana, F., Yoka, J., Ndzai, S. F., Bouetou - Kadilamio, L. N. O.,& Mbemba, M. 2016. Tree Species Diversity, Richness, and Similarity in Intact and Degraded Forest i n the Tropical Rainforest of the Congo Basin: Case of the Forest of Likouala in the Republic of Congo. International Journal of Forestry Research, 2016. a case stu Finnish Forest Research Institute no. 50, pp. 43 56. Kachamba DJ, Eid T, Gobakken T., 2016. Above - and Belowground Biomass Models for Trees in Miombo Woodlands of Malawi. Fo rests, 2016; 7(2). Kalaba, F.K., Quinn, C.H., Dougill, A.J., Vinya, R., 2013. Floristic composition, species diversity and carbon storage in charcoal and agriculture fallows and management implications in Miombo woodlands of Zambia. For. Ecol. Manage. 304, 99 109. Käll K. 2006. The role of fire in the Miombo forest - And the adaptation of the Community - based forest management to meet local needs. Sodertons Hogskola University College Kamangadazi, F. Mwabumba, L. Missanjo, E. and Phiri, F. 2016. Selective Ha rvesting Impact on Natural Regeneration, Tree Species Richness and Diversity in Forest Co - Management Block in Liwonde Forest Reserve, Malawi. International Journal of Scientific Research in Environmental Sciences, 4(2), pp. 0047 - 0054, 2016 Lumbwe, F. C. 20 10. Modeling avifauna responses to Miombo woodland degradation in Serenje District, central province, Zambia [Ph.D. thesis], University of Zambia. Luoga, E.J., Witkowski, E.T.F., Balkwill, K., 2004. Regeneration by coppicing (resprouting) of Miombo (Africa n savanna) trees in relation to land use. F orest Ecol. Manage. 189, 23 35. Magurran, A.E. 1988 Ecological Diversity and Its Measurement. Great Britain Uni versity Press, Cambridge. Malmer, A. and Nyberg, G. 2008. Forest and water relations in Miombo woodlan ds: need for understanding of complex stand management. Working Papers of the Finnish Forest Research Institute 98: 70 86 May, L. 2013. Effects of forest management and land use on regeneration in REDD+ villages, southeastern Tanzania (Doctoral dissertatio n). 55 Mbow, C., Skole, D., Dieng, M., Kwesha, J.C., Landing Mane, D., El Gamri, M., Von Vordzogbe, V., and Virji, H. 2012. Challenges and Prospects for REDD+ in Africa: Desk Review of REDD+ Implementation in Africa. GLP Report No. 5. GLP - IPO, Copenhagen. Mi ssanjo, E., Kamanga - Thole, G., Mtambo, C., & Chisinga, O. 2014. Evaluation of natural regeneration and tree species diversity in Miombo woodlands in Malawi. Journal of Biodiversity Management and Forestry, 3(3), 4. Missanjo, E., Kamanga - Thole, G. and Ndema A. 2015. Biomass and Carbon Stock Estimation for Miombo Woodland in Selected Part of Chongoni Forest Reserve, Dedza, Malawi. International Journal of Forestry and Horticulture (IJFH). Volume 1, Issue 1, 2015, PP 12 - 17. www.arcjournals.org Munishi, P.K.T., Mringi, S., Shirima, D.D., Linda, S.K., 2010. The role of the Miombo woodlands of the southern highlands of Tanzania as carbon sinks. JENE 2, 261 269. Mwakalukwa, E. E., Meilby, H., & Treue, T. 2014. Floristic c omposition, structure, and species associations of dry Miombo woodland in Tanzania. Isrn Biodiversity, 2014. Parrotta, J.A., Wildburger C. and Mansourian, S. (eds.), 2012. Understanding Relationships between Biodiversity, Carbon, Forests and People: The Ke y to Achieving REDD+ Objectives. A Global Assessment Report. Prepared by the Global Forest Expert Panel on Biodiversity, Forest Management, and REDD+. IUFRO World Series Volume 31. Vienna. 161 p. Peters, C.M., 1994. Sustainable Harvest of Non - Timber Plant Resources in Tropical Moist Forest: An Ecological Primer. Biodiversity Support Programme, Washington, DC. Piiroinen T, Roininen H, Valkonen S. 2008. Regeneration of Miombo woodlands: Effects of herbivory, management and competition. Finnish Forest Research Institute 98: 46 - 51. Phillip, M.S. 1983. Measuring Trees and Forests. Aberdeen University Press, Aberdeen. Phillips, O. L., & Gentry, A. H. 1994. Increasing turnover through time in tropical forests. Science, 263(5149), 954 - 958. REDD Research and Developm ent Center. 2012. REDD - plus Cookbook: how to measure and monitor forest carbon. REDD Research and Development Center Ribeiro N.S., Matos C.N., Moura I.R., Washington - Allen R.A. and Ribeiro A.I. 2013. Monitoring vegetation dynamics and carbon stock density in Miombo woodlands. Carbon Balance and Management, 8:11. Licensee BioMed Central Ltd. Schwartz, M.W., Caro, T.M., 2003. Effect of selective logging on tree and understory regeneration in Miombo woodland in western Tanzania. A frican Journal of Ecology, Afr. J. Ecol., 41, 75 - 82 56 Shirima, D.D., Munishi, P.K.T., Lewis, S.L., Burgess, N.D., Marshall, A.R., Balmford, A., Swetnam, R.D., Zahabu, E.M., 2011. Carbon storage, structure and composition of Miombo woodlands in tern Arc Mountains. Afr. J. Ecol. 49, 332 342. Shirima, D.D., Totland, Ø., Munishi, P.K., Moe, S.R., 2015a. Does the abundance of dominant trees affect diversity of a widespread tropical woodland ecosystem in Tanzania? J. Trop. Ecol. 31, 345 359. Skole D. and Samek, J. 2016. Final Technical Report. Forest Inventory: Liwonde, Ntchisi & Perekezi Forest Reserves. USAID PERFORM Project, Michigan State University Smith, H. 2016. The charcoal sector in Southern Malawi: a livelihoods perspective (Doctoral disserta tion, University of Southampton). Syampungani, S., 2009. Vegetation Change Analysis and Ecological Recovery of the Copperbelt Miombo Woodland of Zambia. University of Stellenbosch, Stellenbosch. Trapnell, C.G., 1959. Ecological results of woodland burning experiments in northern Rhodesia. J. Ecol. 47, 129 168. Vicharnakorn, P., Shrestha R.P., Nagai M., Salam A.P. and Kiratiprayoon, S. 2014. Carbon Stock Assessment Using Remote Sensing and Forest Inventory Data in Savannakhet, Lao PDR. Remote Sens. 2014, 6, 5452 - 5479; doi:10.3390/rs6065452. www.mdpi.com/journal/remotesensing White, F. 1983. The vegetation of Africa. UNESCO, Paris. Pp 356 + maps. Zahabu, E. 2008. Sinks and sources: a strategy to involv e forest communities in Tanzania in global climate policy. Dissertation submitted in University of Twente, Netherlands. Zimudzi C., Mapaura A., Chapano C. and Duri W. 2013. Woody species composition, structure and diversity of Mazowe Botanical Reserve, Zim babwe. Journal of Biodiversity and Environmental Sciences (JBES). ISSN: 2220 - 6663 (Print) 2222 - 3045 (Online). Vol. 3, No. 6, p. 17 - 29