<%BANNER%>

Preliminary Assessment of the Effect of High Elephant Density on Ecosystem Components (Grass, Trees, and Large Mammals) ...

Permanent Link: http://ufdc.ufl.edu/UFE0024159/00001

Material Information

Title: Preliminary Assessment of the Effect of High Elephant Density on Ecosystem Components (Grass, Trees, and Large Mammals) on the Chobe Riverfront in Northern Botswana
Physical Description: 1 online resource (62 p.)
Language: english
Creator: Wolf, Andrea
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: botswana, elephants, vegetation
Geography -- Dissertations, Academic -- UF
Genre: Geography thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Elephants are well-known to affect landscapes by converting woodland into shrub or grassland. This is obviously the case along the Chobe riverfront in Chobe National Park, northern Botswana, where dead trees tower over the world?s largest elephant herd while tourists watch transfixed as they move through vast stretches of shrubland on their way to the riverfront. Two types of ecological transects developed specifically for the semi-arid savannas of southern Africa and historical accounts were used to assess change in riverfront ecosystem components over time and space from the herbaceous layer to the woody layer, and the large mammal component. A temporal comparison was made between Chobe National Park in 2007 and 1965, and time was substituted for space to make a spatial comparison between Chobe National Park 1965, 2007, and Bwabwata National Park in Namibia, a physiognomically similar environment, with lower elephant densities. Results demonstrate that the herbaceous layer in Chobe was already degrading by the 1960s because of past land use. The shrub layer has expanded and simplified, and is now dominated by two species (Combretum mossambicense and Capparis tomentosa), and animal composition and densities have changed dramatically. Selective grazers have decreased significantly with the loss of the grass sward, and non-selective browsers such as elephant, impala and kudu now dominate each site analyzed. Results show a simplification in the woody and animal layers, and degradation in the herbaceous layer, leading to questions about sustainability on the system, and whether this park is meeting its goals and objectives as a National Park.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Andrea Wolf.
Thesis: Thesis (M.S.)--University of Florida, 2009.
Local: Adviser: Child, Brian.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024159:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024159/00001

Material Information

Title: Preliminary Assessment of the Effect of High Elephant Density on Ecosystem Components (Grass, Trees, and Large Mammals) on the Chobe Riverfront in Northern Botswana
Physical Description: 1 online resource (62 p.)
Language: english
Creator: Wolf, Andrea
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: botswana, elephants, vegetation
Geography -- Dissertations, Academic -- UF
Genre: Geography thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Elephants are well-known to affect landscapes by converting woodland into shrub or grassland. This is obviously the case along the Chobe riverfront in Chobe National Park, northern Botswana, where dead trees tower over the world?s largest elephant herd while tourists watch transfixed as they move through vast stretches of shrubland on their way to the riverfront. Two types of ecological transects developed specifically for the semi-arid savannas of southern Africa and historical accounts were used to assess change in riverfront ecosystem components over time and space from the herbaceous layer to the woody layer, and the large mammal component. A temporal comparison was made between Chobe National Park in 2007 and 1965, and time was substituted for space to make a spatial comparison between Chobe National Park 1965, 2007, and Bwabwata National Park in Namibia, a physiognomically similar environment, with lower elephant densities. Results demonstrate that the herbaceous layer in Chobe was already degrading by the 1960s because of past land use. The shrub layer has expanded and simplified, and is now dominated by two species (Combretum mossambicense and Capparis tomentosa), and animal composition and densities have changed dramatically. Selective grazers have decreased significantly with the loss of the grass sward, and non-selective browsers such as elephant, impala and kudu now dominate each site analyzed. Results show a simplification in the woody and animal layers, and degradation in the herbaceous layer, leading to questions about sustainability on the system, and whether this park is meeting its goals and objectives as a National Park.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Andrea Wolf.
Thesis: Thesis (M.S.)--University of Florida, 2009.
Local: Adviser: Child, Brian.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024159:00001


This item has the following downloads:


Full Text

PAGE 1

1 PRELIMINARY ASSESSMENT OF THE EFFECT OF HIGH ELEPHANT DENSITY ON ECOSYSTEM COMPONENTS (GRASS, TREES, AND LARGE MAMMALS) ON THE CHOBE RIVERFRONT IN NORTHERN BOTSWANA By ANDREA WOLF A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Andrea Wolf

PAGE 3

3 To my Grandmother

PAGE 4

4 ACKNOWLEDGMENTS I thank my parents and my brother and uncle and aunt, and all my fam ily and friends who supported me through this I also thank my advisors, my fellow graduate students, and those that worked with me in Africa to make this thesis happen. I am also very grateful for my funding sources : Tropical Conservation and Development (TCD) at the University of Florida and the UF Seed grant.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 7 LIST OF FIGURES .............................................................................................................................. 8 ABSTRACT .......................................................................................................................................... 9 CHAPTER 1 PLACEMENT IN GEOGRAPHY ............................................................................................. 11 Introduction ................................................................................................................................. 11 An Integrated Approa ch; Physical Geography, Human Geography, and Other Disciplines ................................................................................................................................ 12 Methodologies: Field Work ................................................................................................ 13 Geographic Technologies: Study Design and Data Analysis ........................................... 14 Space and Time .................................................................................................................... 16 Ecological Theory ................................................................................................................ 17 Conclusion ................................................................................................................................... 21 2 QUANTIFICATION OF CHANGE IN ECOSYSTEM COMPONETS ALONG THE CHOBE RIVERFRONT ............................................................................................................. 23 Study Area ................................................................................................................................... 24 Land Use History ........................................................................................................................ 25 Methodology................................................................................................................................ 27 Data Collection .................................................................................................................... 27 Sampling Method................................................................................................................. 27 Results .......................................................................................................................................... 29 Herbaceous Layer ................................................................................................................ 29 Woody Layer ........................................................................................................................ 30 Riparian s trip ................................................................................................................ 30 Pookoo Flats ................................................................................................................. 3 1 Mixed w ood/ s hrubland ................................................................................................ 32 Baikiaea woodland ....................................................................................................... 32 Animal Composition and Density ...................................................................................... 33 Discussion .................................................................................................................................... 36 Conclusions ................................................................................................................................. 42 APPENDIX DESCRIPTION OF TREE AND SHRUB SPEC IES .............................................. 57

PAGE 6

6 BIBLIOGRAPHY ............................................................................................................................... 58 BIOGRAPHICAL SKETCH ............................................................................................................. 62

PAGE 7

7 LIST OF TABLES Table page 2 1 Number of transects per site and year ................................................................................... 49 2 2 Woody species densities and percentages on Pookoo Flats. ............................................... 53 2 3 Comparison of woody species composition and densities in mixed wood/shrubland betw een Chobe and Bwabwata National Parks. ................................................................... 54 2 4 Comparison of woody species composition and densities in Baikiaea woodland between Cho be and Bwabwata National Parks. ................................................................... 55

PAGE 8

8 LIST OF FIGURES Figure page 2 1 Study site. .............................................................................................................................. 47 2 2 Riverfront vegetation types: Floodplain, riparian strip, Pookoo Flats, sandridge, Baikiaea woodland ................................................................................................................. 47 2 3 Six vegetation types and their associated histories .............................................................. 48 2 4 Comparison of rainfall preceding the 1965 and 2007 vegetation surveys. Government of Botswana; Department of Meteorological Services ........................................................ 48 2 6 Spatial and temporal comparison of bare ground in CNP (1965, 2007) and BNP (2007) in 3 vegetation types using Riney transects .............................................................. 50 2 7 Spatial and temporal comparison of rooted grass in CNP (1965, 2007) and BNP (2007) in 3 vegetation types using Riney transects .............................................................. 51 2 8 Spatial and temporal comparison of aerial grass cover in CNP (1965, 2007) and BNP (2007) in 3 vegetation types using Riney transects .............................................................. 51 2 9 Comparison of riparian strip tree species in CNP (1965, 2007). ........................................ 52 2 10 Comparison of riparian strip tree species in Chobe (1965, 2007) and Bwabwata National Park (2007). ............................................................................................................. 53 2 11 Relative animal densities on Pookoo Flats ........................................................................... 55 2 12 Relative animal densities in mixed wood/shrubland ............................................................ 56 2 13 Relative animal densities in Baikiaea woodland.................................................................. 56

PAGE 9

9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science PRELIMINARY ASSESSMENT OF THE EFFECT OF HIGH ELEPHANT DENSITY ON ECOSYSTEM COMPONENTS (GRASS, TREES, AND LARGE MAMMALS) ON THE CHOBE RIVERFRONT IN NORTHERN BOTSWANA By Andrea Wolf May 200 9 Chair: Brian Child Major: Geography Elephants are well known to affect landscapes by converting woodl and into shrub or grassland. This is obviously the case along the Chobe riverfront in Chobe National Park, northern Botswana, where dead trees tower over the worlds largest elephant herd while tourists watch transfixed as they move through vast stretches of shrubland on their way to the riverfront Two types of ecological transect s developed specifically for the semi arid savannas of southern Africa and historical accounts were used to assess change in riverfront ecosystem components over time and spac e from the herbaceous layer to the woody layer, and the large mammal component A temporal comparison was made between Chobe National Park in 2007 and 1965, and time was substituted for space to make a spatial comparison between Chobe National Park 1965, 2007, and Bwabwata National Park in Namibia, a physiognomically similar environment, with lower elephant densities. Results demonstrate that the herbaceous layer in Chobe wa s already degrading by the 1960s because of past land use. The shrub layer has e xpanded and simplified, and is n ow dominated by two species ( Combretum mossambicense and Capparis tomentosa), and animal composition and densities have changed dramatically. Selective grazers have decreased

PAGE 10

10 significantly with the loss of the grass sward, and non-selective browsers such as elephant, impala and kudu now dominate each site analyzed. Results show a simplification in the woody and animal layers, and degradation in the herbaceous layer, leading to questions about sustainability on the system, and whether this park is meeting its goals and objectives as a National Park.

PAGE 11

11 CHAPTER 1 PLACEMENT IN GEOGRAPHY I ntroduction Geography is often referred to as the mother of the s ciences with astronomy, botany, zoology, geology, meteorology, archaeology and anthropology among its children (Barrows, 1923). This is ironic because in todays world, part icularly in the United States, g eography is frequently marginalized as a discipline, or even considered simply a component of one of the above mentioned fields. In spite of this, through different processes such as explorati on, discover y and cartography, g eography has come to be regarded as the study of our physical environment and humans response to, and impact upon the earth, and is based on the interplay between the environment and human forces (Herbert and Matthews, 2004). This def inition assumes overlap in the humanities and social and physical sciences. While my research is predominantly ecological in nature, it draws from a variety of disciplines and from both human and physical geography. It remains a geographic study because of the methodology employed to collect and analyze the data, and because of how the study fits into several broad geographic theoretical frameworks. I discuss the methodology involved in the data collection in the field and the use of current geographic technologies including geographic information systems (GIS) and remote sensing. I then discuss some basic geographic concepts such as an understanding and comparison of space, area, and region, and how geographic technologies, specifically aerial photogra phy, are used in my study to scale up from area to region. I also look at change over space and time, a staple of geographic theory, and then emphasize more contemporary trends in geographic thought including environmentalism; particularly conservation of biological diversity. Finally, I discuss the human influence on my study site, which is a national park, some of the objectives of national parks generally, and

PAGE 12

12 specifically, how this anthropogenic institution influences the ecological components in my research. An Integrated Approach; Physical Geography, Human Geography, and Other Disciplines Since it is an integrated study, the research can be labeled within a variety of subfields of geography. My study presents a quantitative analysis of vegetation ch ange over time, and utilizes quantitative data from 1965 to determine if there has been a significant change over time, as well as historical qualitative data to determine how and what were and are the drivers of change. The main driver that I look at and correlate vegetation change with is elephants. Animal behavior and population increases fall within the field of zoology, expanding the research from a strictly ecological study. I also used a biogeographical approach by analyzing differences both within and between sites, and then compare the differences in species composition and richness between sites in one area to those in another physiognomically similar region. While the quantitative portion of the study would be considered physical geography, and perhaps more specifically, landscape ecology, the drivers must be understood within a broad context, employing components of human and cultural geography and even those of other disciplines. Although my work is predominantly ecological in nature, the phe nomenon under analysis has developed because of and within the context of social, political, economic, and environmental dimensions. The results of the research must be analyzed and related back to the relevant stakeholders within this integrated sphere. Other drivers and responses to the changes incorporate social, environmental, and economic influences on policies. Management of the study site (Chobe National Park) has determined current and past lands uses and has incorporated some aspects of all the above mentioned facets of an integrated system. Some of these policies include timber extraction and hunting regulations and livestock and agriculture policies, all of which are influenced by the history and social and cultural norms of the local people w ho

PAGE 13

13 previously lived within and now live around the national park. They are also highly linked to the local economy, as well as the national economy of Botswana, which depends on tourism from Chobe National Park (CNP) for a large portion of its citizens l ivelihoods. Therefore, market forces also play a significant role in determining land use at the study site because of strong local ties to the international tourism industry. Methodologies: Field Work Field work is a tradition in geography that is sha red with many other disciplines, particularly for physical geographers, but also for human and cultural geographers (Stoddart and Adams, 2004). My field work is based on two types of vegetation transect methods developed and practiced by ecologists in the semi arid savannas of southern Africa. Together, they provide the required information for a study on vegetation with an emphasis on changes over space and time. Two methods are employed because they complement each other, and provide a comprehensive vi ew of the land cover. The Riney (1963) method is a rapid field technique to assess the effects of fire and herbivory on th e herbaceous layer This method uses a point -step technique and provides the observer with a set of syndromes meant to supply manage rs with an understanding of the landscape before thresholds are crossed and there is an irreversible shift towards degradation of the herbaceous layer in the form of reduced productivity. Tree and animal species are also recorded as well as trend in the h erbaceous and bush layer. The Walker (1976) m ethod is a more comprehensive and time -consuming technique which supplies information about the herbaceous layer through an in -depth analysis o f 1x1 meter plots It also provides information about tree and shrub composition and density, as well as animal species density and diversity. It is the process of carrying out this field work that a thorough understanding of the subject can be gained, in part because i t is provides a holistic view of the study area.

PAGE 14

14 Geographic Technologies : Study Design and Data Analysis Geographic Information Systems (GIS) and Remote Sensing technologies are perhaps the most uniquely geographical tools employed within the discipline to analyze and display data as well as develop and design a study before arriving in the field. These technologies allow Geography to address significant problems of society and the environment using explicitly spatial data, information, evidence and kno wledge (Langley and Barnsley, 2004). They allow users to analyze land covers from afar, and to do so over time, at least as far as the satellite images go back in time. I used satellite images (Landsat 5, 2004) to discern distinct land covers across the mosiaked landscape to develop a sampling frame that would place transects in each vegetation category. This was a very effective method except in the case of fire, which is prolific across the landscape, particularly in Namibia. Several times, a site wa s determined for analysis, and an exact location within that land cover was decided upon randomly, but upon arrival at the site, it was found to be recently burned, precluding it from analysis. Additionally several sites has been recently burned before the image was acquired, so they showed up on the satellite image as burned, irrespective of their true vegetation category. Obviously, there are still constraints to the constantly evolving and improving technologies. I make use of aerial photographs, one form of remote sensing, which provides information about different land covers in the region, about change in land cover over time, and allows me to perform a range of analysis on the images to better interpret the landscape. I have employed aerial photographs to aid in the analysis of the ecological transect data (i.e. grouping vegetation categories). It is necessary for geographers to have a thorough understanding of the landscape, such as that provided by extensive work in the field, to make the mo st effective use of geographic technologies. The collection of ecological data is vital to my work as well as to physical

PAGE 15

15 geography more generally because it is necessary to understand what is happening on the ground before being able to understand and in terpret satellite images or aerial photographs, which display the land cover at significantly coarser resolutions. Although these technologies are generally manipulated by geographers, this emphasizes the need for an integrated approach when working with imagery. I will expand on the current analysis in the future by scaling up from the plot level transect data to the aerial photographs. I also plan to expand that study to the satellite imagery level by working with other ecologists and remote sensors. This has the potential for the development of a regional or even biome -level understanding of the landscape at various scales, and over time, by using aerial photographs which extend back to the 1940s for my study site, by using satellite images which are useful for detecting change and looking at very large areas, and by using the transect data to ground truth both applications. These technologies make valuable contributions to a study on vegetation change over space and time because both GIS and Remote Sensing allow for a prediction of future events based on past events in a spatially explicit way (Langley and Barnsley, 2004). Global positioning systems (GPS) also contributed to my research by allowing each transect site to be recorded specifically for future repeatability and mon itoring. Additionally, all location s can be imported into GIS for analysis and display. As is commonly repeated, a picture is worth a thousand words, and displaying the transects as GPS points on a vegetation map or aerial ph otograph provides a considerable amount of information to other scientists, managers, tourists, or community members interested in the site. The final component of my research that makes it explicitly geographic is in my results. In additional to quanti tatively documenting change in ecosystem components by grouping transects

PAGE 16

16 into vegetation types, I plan to produce a vegetation map showing different categories of land covers and their extents and change over a three step time series. Although not a map in the more tradition sense, such as those used for navigation (Vincent and Whyte, 2004), it does allow for a visual display of the information, highlighting the spatial distribution of land covers and their change over time. Again, this type of imagery c an be exponentially more useful and explanatory than a written analysis for a wide audience. Space and Time Spatial and temporal comparisons are often employed as a means to analyze various physical and humanistic phenomenons in geography. This concept is fairly straightforward and useful in geography because it provides ample information about different processes and their outcomes. Richards, Bithell, and Bravo (2004) explain it simply as underlying mechanisms which vary across space, and cause various processes of change over time (pg. 328). In the case of my study, underlying mechanisms include physical changes in a dynamic ecosystem which evolve because no natural system remains at equilibrium, and the semi arid savanna system is known for its lack of a climax state. Time is relevant because both the physical and human changes happen gradually, but specific drivers combined with time allow the landscape to change. My research looks at different land covers within one ecosystem in a specific area. T he significant changes over time in the ecosystem components determined by analysis of the ecological transect data can be correlated with the drivers which initiated the changes under investigation, in this case from qualitative data from historical recor ds and interviews from people living in the area during the time frame in question. Analysis of different spaces within the landscape then provides additional information about how similar inputs affect various spaces differently. For example, I

PAGE 17

17 found that areas with different soil types exhibit different effects from similar drivers over the same time period. Ecological Theory The concept of space and time leads into the ecological theory upon which this study is based. This is because the study compares ecological transect data from 1965 with data from 2007, and also substitutes space for time by comparing the study site in Chobe National Park with a similar ecosystem in Bwabwata National Park in Namibia. The Bwabwata site is used as a baseline because it has not experienced the same heavy land use pressures which Chobe National Park has experienced, and therefore a comparison with the first time series from CNP can be made based on a continuum from BNP 2007 to CNP 1965 to CNP 2007. The data have been grouped by vegetation category, and are analyzed between sites, with a focus on change in three ecosystem components (the herbaceous layer, the woody layer, and animal species) in regards to the cascade effect from the herbaceous layer to the shrub and tre e layer, and on to animal species density and diversity by vegetation category. Although the cascade affect has largely been described from a top-down perspective, and in relation to predators influence on lower tropic levels, there is evidence that large herbivores have considerable influence on cascade effects within an ecosystem (Pringle et al., 2007). As suggested by Hunter and Price (1992) the cascade effects seen in the Chobe riverfront ecosystem cascade both up and down tropic levels, and are affec ted by the heterogeneity of the landscape. A semi arid savanna is somewhat continuous grass with discontinuous tree and or shrub groupings (Belsky,1990) and include fluctuations between these components based on fire, soil moisture and nutrient levels, an d herbivory, specifically elephants. The mosaic of vegetation types each respond to similar drivers (i.e. cattle grazing, elephants) based on their own inherent

PAGE 18

18 qualities. Results from this study demonstrate the cascade effect in Chobe National Park to very pronounced, although instead of simply moving up or down tropic levels, it is cyclical in nature. Fire and cattle grazing were found to have degraded the herbaceous layer in CNP by 1965 allowing for bush encroachment along the riverfront (Child, 1968). Pringle et al. (2007) found ungulates to have a marginally significant effect on herbaceous cover in an African savanna (Laikiapia District, Kenya), which collaborates evidence from Chobe demonstrating that degradation of the herbaceous layer develope d because of both fire and cattle. Around the same time in the 1960s elephant densities increased dramatically in the park (Campbell, 1990), and they began to exert their influence on the vegetation. They opened the woodland and converted it to shrublan d (Mosugelo et al., 2002). Ungulates have been found to exert a strong top-down effect on tree density (Pringle et al., 2007), and many studies have correlated elephant densities with a decrease or an opening in woodland or grassland (Cumming et al., 1997; Dublin et al., 1990, Skarpe et al., 2004). Along the Chobe riverfront woodland was converted to shrubland, noted particularly on Pookoo Flats and along the sandridge overlooking the Chobe River. On alluvial soils on Pookoo Flats the area has been conve rted to low density stands of Combretum mossambicense and Capparis tomentosa, while the areas which are now mixed wood or shrubland on Kalahari sands are dominated by Combretum mossambicense and Combretum elaeagnoides These Combretum shrub species were r eferenced by Child (1968) as one of several species in the tall bush layer in 1965 and have currently taken over both the tree and shrub layers on the Pookoo Flats and the mixed wood/shrubland sites. Each of these sites supports low tree and shrub densiti es compared with past reports of the sites, and with comparable sites in Bwabwata National Park. The Baikiaea woodland, further from the

PAGE 19

19 riverfront and also based on low -nutrient Kalahari sand has maintained a greater density of trees and shrubs. There i s evidence that a large number of impala became established in the ecotonal areas created by the conversion of woodland to shrubland by elephants in the mid-1980s (Rutina, 2004). While elephants reduced woody cover, impala influenced composition of the w oody layer through selective browsing (Rutina, 2004). The herbaceous layer had already been degraded, and the decrease in tree and shrub richness and abundance appears to have then cascaded to the animal component of the ecosystem. This is clear by a significant shift in mammal species composition and densities on Pookoo Flats and in mixed wood/shrubland, and to a lesser extent in the Baikiaea woodland. The first two sites mentioned above currently have very high relative densities of elephant, impala, b uffalo and kudu compared to those of 1965, and compared to similar sites in Namibia. The Baikiaea woodland also has very high elephant and buffalo densities, and notably demonstrates a complete shift in animal composition from 1965. While grazers such as waterbuck disappeared when the sward collapsed, other mammal species have increased dramatically, specifically non -selective grazers such as elephant, impala, giraffe, and kudu in response a nearly homogenous woody layer. The cascade effect in the Chobe r iverfront ecosystem begets questions of sustainability and conservation of biodiversity, and whether Chobe National Park is achieving its goals and objectives regarding these issues. While it is clear that there has been a substantial increase in large ma mmal density, there appears to be a decrease in mammal species richness, as many species cited in the 1965 data were not recorded in 2007. This leads to questions of sustainability of such large densities of certain species, of whether ecological threshol ds have

PAGE 20

20 been crossed, and if they have if this will eventually lead to a decrease in the carrying capacity of the ecosystem. A system is resilient if it can absorb disturbance, adapt, and maintain its basic functioni ng (Walker et al., 2004). Holling (1973) defines resilience in an ecological system as a measure of the magnitude of disturbance which can be absorbed before a system centered on a locally stable equilibrium flips to another. This can be difficult to measure because change is often successional and not the result of a flipped system, and sometimes a threshold has been crossed and a system has entered another state, but this is not obvious until many years or even decades later allowing degradation to continue unfettered. It is necessary to be aware of ecosystem resilience because a shift can often lead to degradation of an environment in the form of a decrease in productivity upon which humans depend. This can mean different things depending on the purpose of the ecosystem in question, but in Chobe National Park, like in many protected areas in Africa, tourism is one of the main forms of livelihoods for communities surrounding the park, and it is likely to remain so in the future. There is little other industry in the area, and the semi arid biome is prone to climatic va riability (i.e., drought, floods) with low soil nutrients, and therefore does not lend itself to large agricultural yields or high levels of income generation from livestock. The wide variety of wildlife and low yields from other land uses influenced the Government of Botswana to classify the area a national park in 1968 with the hopes of protecting biodiversity, and providing income generation from tourism. Protected areas (PA s) such as Chobe and Bwabwata National Park are de fined by the World Commission on Protected Areas as an area dedicated to the protection and maintenance of biological diversity, and of natural and associated cultural resources, and managed through legal or other effective means. (Baillie et al., 2004). As a category II National Park, CNP is an

PAGE 21

21 area managed for ecosystem protection and recreation. This precludes resource extraction or use within the park and forces communities to rely on tourism revenue from the area, as opposed to any consumptive use of the site. Therefore, it is vital that the area remain a well -frequented tourist destination, with consistent accessibility to a wide range of plant and animal species. This returns to the question of whether CNP has already experienced a loss of biod iversity, or if ecological thresholds have been crossed eventually leading to degradation of the landscape. While, large mammal densities have increased since 1965, animal species richness has decreased, and there is concern that an environmental shock co uld cause a crash to the system as was the case in Tsavo National Park, Kenya. Although Tsavo seems to be recovering ecologically (Chafota, 1998), massive die -offs of elephants and other herbivores such as giraffe and kudu, could severely affect the touri sm industry. Cumming and Brock (1997) found a decrease in diversity of woodland birds and ants in areas where elephants had reduced woody species richness, and the same may hold true for species which distribute seeds, leading to a decrease in diversity a nd abundance of woody species, which could then cascade through to the animal component of the ecosystem. Because the system is so dynamic, it needs to be monitored consistently, and the different layers and components need to be considered spatially and temporally. As protected areas in the region expand and elephants gain access to previously inaccessible sites, it will be necessary for communities and managers of these areas to anticipate changes associated with elephants so as to respond appropriatel y, particularly around water points, which are the most impacted by high herbivore biomass. Conclusion Determining resilience is very difficult in any ecological system, but particularly in semiarid savannas, which have high climatic variability. It is not always clear when a system has

PAGE 22

22 crossed a threshold into another state. Even so, it is necessary to document and monitor changes in ecological systems so that is possible to respond and adapt to these changes, whether that is through management or shifts in usage. National Parks have been set up to maintain biodiversity and provide recreation for people These objectives influence economies and allow for the maintenance of ecosystem services for future generations. Therefore, an analysis of change over space and time contributes to an understanding of the system and its trajectory. Although this study is inherently ecological, it incorporates many different disciplines including zoology, biogeography, history, and economics. It maintains its geographic nature because the tools used (i.e., GIS, satellite imagery, aerial photography) to design the study and analyze the data and because of the basic geographic frameworks it is rooted in such as looking at changes over space and time and scaling up from an area to a region.

PAGE 23

23 CHAPTER 2 QUANTIFICATION OF CHANGE IN ECOSYSTE M COMPONETS ALONG THE CHOBE RIVERF RONT Introduction Even to the untrained eye of a tourist in Botswanas spectacular Chobe National Park (CNP), the Chobe riverfront is inundated by large elephant herds, and most large trees tower as skeletons above an underbrush of hedged shrubs and thickets. The worlds largest known elephant herd (approximately 216,000 animals) is centered on northern Botswana, western Zimbabwe and the Caprivi Strip of Namibia (Blanc et al., 2007). Increasing exponentially, and doubling in under fifteen years, these el ephants are of immense biological (to ecosystem health, environmental service and biological diversity) (Cumming et al., 1997; Rutina, 2004), economic (to tourism, safari hunting, human elephant conflict and local livelihoods) (Jackson et al., 2008) and political (to ivory trade, culling and sustainable use practices) importance (Barnes, 1996). Our intention was to provide repeatable measurements that were affordable financially and in terms of time, to assess changes in the herbaceous, woody and large mammal components of this savanna ecosystem since the colonization and domination of this environment by elephants (Work and Owen -Smith, 1986). A tempo ral comparison was made by repeating and supplementing transects conducted in the mid1960s using methodology developed by Riney (1963). Substituting space for time, we identified the most physiognomically similar nearby environment that is not yet heavi ly impacted by elephant (i.e. Bwabwata National Park (BNP)) west of the Kwando River in the Caprivi Strip, Namibia, and placed a comparable number of transects there. Quantitative data regarding the woody layer were not collected in CNP in 1965, so BNP w as used as an indicator of what woody density and composition may have been in CNP in 1965. To increase repeatability of this method, and particularly our ability to monitor changes in woody vegetation, we complemented each Riney transect with an addition al transect using

PAGE 24

24 methodology developed by Walker (1976). To improve robustness of our conclusions, we triangulated our results with a review of the scientific literature, diaries of hunters and explorers, and with the knowledge of key observers including scientists, managers and local people such as Bushmen. Our intention was to inductively identify and explore key environmental issues, as the foundation for designing more detailed research spatially and methodologically, including the use of remote sens ing and investigations of large mammal, small mammal and invertebrate diversity and sustainability. This study aims to quantify changes in the herbaceous layer, woody layer and large mammal component of the Chobe riverfront ecosystem over time. It is obv ious to us, especially following the quantification outlined in this study, that the Chobe ecosystem is much simplified in terms of plant and animal diversity. However, we still need to determine whether this ecosystem is 1.) less productive than it was before given the high density of some large mammals and the apparent productivity of pioneer plant species, especially the scrub layer and 2.) how resilient this new system is in the face of increasing elephant pressures, and climate variability and change. Models predict that climate change will impact this region more than any other in Africa (Thornton et al., 2006). Study Area The Chobe riverfront (Figure 2 1) is characterized by nutrient rich alluvial soils in areas previously covered in floodplain and by nutrient -poor Kalahari sands elsewhere (Simpson, 1975). As illustrated by the sketch (Figure 2 2) the area is characterized by six vegetation types, some of which have been heavily modified by, among other things, elephants and impala (Rutina, 2004). A fringe of reed beds along the river course has almost entirely disappeared since the 1960s. This is followed by floodplain grassland and then a thin strip of riparian woodland, up to 70 meters in width. Beyond this, and varying in width according to the topography, is grassland on more fertile and alkali alluvial soils (Selous, 1881, Henry, 1966)

PAGE 25

25 which Selous (1881) associated with scattered trees, often Acacia tortilis in 1874. The woody cover has changed at least twice since then, with Dichrostachy s cinerea thickets being associated with cattle trekking in the 1950s (Child, 1968) now been replaced by Capparis tomentosa, a spiny scrambler that can take on a self -supporting tree form. Pookoo Flats is typical of this ecotype. These bushed grasslands grade rapidly into a heavily modified mixed woodland on the slope of the basalt underlain sandridge, with a climax Baikiaea plurijuga (African/Rhodesian teak) dominated Kalahari Woodland on the deep infertile sandy soils on the top of the sandridge (Chi ld and von Richter 1968, Moroka, 1984). See Figure 2 3 We selected the physiognomica ll y similar Bwabwata National Park for comparison with Chobe because BNP is only starting to be damaged by elephants and in places appear s quite similar to what Chobe use d to look like in the 1960s (Child G., pers obs.). Rainfall is similar; Chobe receives 690 mm compared to 600 700 mm in Bwabwata, as is elevation; 910 1,050 m (Mendelsohn and Roberts, 1997) and compared to 9001,100 m respectively. Topographically, the Chobe study area comprises a single sandridge atop a basalt intrusion, whereas however, Bwabwata is characterized by an undulating dune structure and a catenal effect, with deep sands on the ridges and heavier and more fertile inter dune drainage lin es. In making our temporal comparison, we noted that the 2007 transects were done earlier in the dry season (by two to three months) and following better rainy seasons (Figure 2 4 ) than the 1965 surveys. Therefore, if the status of the herbaceous level had not changed, we would expect transect data with higher levels of rooted grass and grass cover than in 1965. Land Use History In 1852, Chapman shot a significant number of ele phants in what is now the southeastern corner of Chobe National Park, but as he traveled northwards into our study area he saw no more elephants (Chapman, 1968) This was confirmed by another famous hunter and explorer, Selous,

PAGE 26

26 who in 1874 recorded approximately 100 elephant s between Kasane and Ngoma (Selous 1881) in our current study area. At this time the area was infested by tsetse fly, and Selous saw no people on the mainland until he reached Linyanti; the only inhabitants of the area were living on what is now Impalila Island (Selous, 1881) The trade r Westbeach first moved cattle onto the Chobe River at Kazungula in 1886. People began to settle what is now the Chobe Enclave, first from Impalila Island and from the 1920s agriculture increased significantly with an influx of people from Gweta and Maka lambabedzi southeast of the Okavango on what is now the main road between Maun and Francistown. There were soon 30,000 cattle in what became the breadbasket of Botswana, but agriculture collapsed with the floods of 1957 and approximately half the popula tion returned to Maun. Bushmen who were living along the Ngwezumba River, hunting and gathering and presumably herding a few cattle, moved into the Enclave so that when Chobe was declared a Game Reserve in 1962 and a National Park in 1968, it was devoid of people. The park was used for logging, with Sussman Brothers constructing a wooden trolley way into present day Zimbabwe in the early to mid 1930s. Chobe Concessions resumed logging for about five years from 1947, producing cross ties for railways. Lik e Sussman Brothers, this forestry venture was also unprofitable, and attempts to make the business work with cattle and rice paddies also failed remnants of paddies and dip tanks are still visible in the Pookoo Flats area. From approximately 1949 to 1962, the riverfront was used as a cattle corridor from Ngamiland to trek cattle to the Zambian Copperbelt, but the reinvasion of tsetse fly forced cattle trekking to the east. The remnants of the invasive woody species Dichrostachys cinerea are presumably a consequence of this period. The park was burned regularly by the Forestry

PAGE 27

27 Department until 1967/68, but by then vegetation along the river could only support creeping ground fires and not the intensity of fires as evidenced by fire scars on trees. Methodology Data Collection Data were collected from 27 sites in BNP ( Namibia ) and 26 sites from CNP (Botswana ). We used the 200-yard Riney step-point method (Riney, 1963) to provide a direct comparison with Childs (1968) earlier work. At each site we a lso used Walkers (1976) method because it is more repeatable and the objective of our large study was to leave behind a baseline survey and to link field methods to remote sensing. Walker developed his method from the original work done by Riney, but des igned it to be more comprehensive and repeatable, and to be adapted to the specific purpose of his study. Thus, 25 1 m2 plots to obtain eight measures of the health of soil surface and grass sward. A variable -width belted transect is used to assess the d ensity and structure of woody species, including damage to it. Both methods include comparable dung plots, which record presence or absence of dung identified by species, and have provided valuable data for very little extra work. Both transect methods w ere developed by experienced southern Africa n ecologist s for use in the highly patchy conditions associated with semi arid savannas Additionally, we classified and counted all the trees in the same one -mile long riparian s trip done by Child (1968) in 1965 to assess what changes have occurred along the riverfront over the past 40 years Sampling Method The locations of the transects were determined by a combination of methods. We replicate the FAO vegetation survey conducted by Dr. Graham Child (FAO Wildlife Ecologist advising the Botswana Government at the time) in 1965, when there was no GPS technology. Therefore, Dr. Graham Childs presence during the field work was invaluable because his me mory of the

PAGE 28

28 previous transect locations in conjunction with his field notes outlining the transect sites allowed for a high degree of repeatability of site location. His input also allowed for standardization of data collection techniques for the Riney tr ansects, which contributes to greater reliability when comparing results. The beginning and endpoints of the 2007 transects sites were recorded on a Garmin GPSMAP 76 CSx for future replication. Dr. Childs 1965 vegetation assessment was based on 19 Riney transects including seven on Po okoo Flats, nine along the sandridge overlooking the Chobe Flats (mixed wood/shrubland), three on Kalahari woodland ( Baikiaea woodland), and an enumeration of tree species along a one mile section of riparian woodland. The 2 007 (Table 1) study consists of seven sites on the Pookoo Flats (two of which were deliberately sites in the best grassland for demonstration purposes) nine sites in mixed wood/shrubland, seven sites on Baikiaea woodland, and three sites on the Chobe Ri ver floodplain (possible because it was the dry season). A 1 mile by 70 meter sample of the riparian strip was also replicated along the Chobe riverfront between the culvert and Sedudu Valley. In Bwabwata National Park site s were select ed so that result s could be compared with the CNP data, but also for the purpose of establishing a baseline survey for long term monitoring. S atellite imagery (Landsat 5, 2004) was used to insure that all major vegetation types in the eastern part of the park were sampled We also identified specific sites (e.g. riparian woodland) that were similar to those measured in Chobe in 1965, which were then assessed using both types of transect. Overall, five vegetation categories were sampled in BNP including two riparian str ip transects, four floodplain sites, seven grassland sites, nine mixed wood/shrubland sites, and five Baikiaea woodland sites.

PAGE 29

29 Results Herbaceous Layer The most repeatable herbaceous data set s from 1965 is the 200 points we classified as having either r ooted grass ( F igure 2 7 ) or bare ground (F igure 2 6 ) (i.e. Riney), as well as whether it had a cover of perennial g rass, annual grass or nothing (F igure 2 8 ). The Pookoo Flats area in CNP demonstrates a significant change in the herbaceous layer in the form of an increase in bare ground and a decrease in rooted grass between 1965 and 2007. Bare ground levels in the mixed wood/shrubland sites in CNP conversely, have decreased and rooted grass has increased very slightly since 1965. However, as mentione d, inland water sources dried up early in the season and forced animals to congregate in larger numbers along the riverfront in 1965 than when we collected data in 2007. Comparable sites in BNP are significantly healthier than CNP 1965 and in 2007. The B aikiaea Woodland sites in CNP also show a significant increase in bare ground although there has not been a significant change in rooted grass. Baikiaea woodland sites in BNP have significantly lower levels of bare ground than in CNP 2007, and higher leve ls of rooted grass than in 2007. A comparison of grass cover (i.e. % of step points with aerial grass cover) shows that Pookoo Flats was dominated by perennial grasses (37% ) in 1965 with a few annuals present (3 % ). By 2007 perennial grass had declined to 1.2%, while annuals increased slightly to 2%. Overall grass cover declined from 40% to 3%. The sparse grass cover in mixed wood/shrubland sites was close to negligible (0.2% and 1.1% respectively). By contrast, similar areas in the lightly stocked BN P are moderately grassed by perennial grasses (18%), with few annuals present (3%). The CNP 1965 data show dominance by perennials (12% ) and a few annuals present (5 % ). I n 1965, the Baikiaea woodland in CNP

PAGE 30

30 had a reasonable grass cover dominated by peren nials (12 % ), but perennials maintain a presence (8 % ) with a similar overall grass cover to CNP in 1965. Woody Layer Riparian s trip T here has not been a significant change in tree density (1965, n=299; 2007, n=324), or in species richness (1965 number of t ree species = 14; 2007 number of tree species = 12) but the species compo sition has changed dramatically ( Figure 2 9 ). Six of the 14 species recorded in 1965 have disappeared including Ziziphus m ucronata, Diosyrus mespliforis, Kigelia africana (p inata) Acacia albida, Acacia galpinii and Acaica eri o loba. Many of these are big, impressive (and palatable) trees. Three species not previous present have colonized the strip Markhamia obtusfolia, Markamia zanzibarica: although one is ( Capparis tomentosa) a vine that grows into the tree layer and can be self -supporting. Perhaps the major change is the replacement of Acacia nigre s cens by Croton megalobotrys as the dominant species. In 1965 Acacia nigrescens then a relatively recent invader based on age structure of the tree population, comprised slightly more than 50% of tree species. This and five other species comprised 88% of species recorded by 2007 a rapidly growing invasive species, Croton megalobotrys, made up 83% of the riparian tree species -o n ly five C. megalobotrys were recorded in 1965. The next most prolific species in 2007 were Markhamia obtusfolia (n=11), a species not recorded in 1965, and Trichilia emetica (n=11), a species present in 1965, and unpalatable to elephants. Additionally, 8 2% of species recorded in 2007 were determined to be in the highest two classes of elephant damage possible. Riparian trees were also recorded in the Bum Hill area of Bwabwata National Park (Figure 2 10) on the west bank of the Kwando River, the site most physiognomically similar to Chobe identified along the river. This data is comprised of four transects (two sites), and a traverse of

PAGE 31

31 undefined area. Although these data are not directly comparable because of differences in collection methods, it does pr ovide information about species composition. Seven of 15 Namibian riparian species (dbh > 5.5) were recorded in Chobe in 1965. Although Acacia nigrescens dominated the area surrounding the transect sites visually, it was not the most recorded species b ecause a campsite had been built around and into many of these trees, therefore precluding it as a transect site. Lonchocarpus capassa was the most common riparian species recorded in the Namibian transects (24). Forty L. nelsii were recorded in Chobe i n 1965. These species are difficult to distinguish, and it is possible that they have been misidentified. Ziziphus mucronata, Garcinia livingstonei, Diospyros mespiliformis, and Acacia erioloba/giraffae were also present in CNP 1965 and BNP. Of these on ly two Garcinia livingstonei and one Acacia erioloba remained in CNP in 2007. It is interesting to note the low incidence of Croton megalobotrys in both CNP 1965 and BNP, and that Markhamia obtusfolia and M. zanzibarica have a substantial presence in CNP 2007, while neither was recorded in CNP 1965 or BNP. This would suggest that these quick growing trees are successional species because they are functionally similar to trees previously recorded in the area (Rutina, 2004). Pookoo Flats Pookoo Flats supp orts 152 woody plants per hectare, characterized as it is by intensively hedged shrubs and large patches of bare ground (Figure 2 2, 2 3 ). In describing the woody layer we combine trees and shrubs because the majority of trees recorded (woody species > 3 m eters) were shrub species even though they exceed 3 meters in height -thus Combretum mossambicense and Capparis tomentosa comprise 73% of the woody layer (39% and 34% respectively). Only seven different tree and shrub species were recorded on Pookoo Flats. See Table 2 2

PAGE 32

32 Mixed w ood/ s hrubland In the heavily utilized mixed woodland ecotype, the density of shrubs (1,373/ha) exceeded that of trees (40/ha) thirty -four fold. By comparison, there are thirteen times as many trees in Bwabwata (534/ha) and about 1.5 times as many shrubs (2,158/ha). As noted, Bwabwata may be naturally more diverse than Chobe, but it also has far fewer elephants and the area burns regularly. In Chobe we recorded 9 species in the tree layer and 16 species in th e shrub layer in CNP. Six species were common to both categories. There were 22 tree species and 32 shrub species in BNP. Many of these species were recorded only once or twice, indicatin g that rare species might be being lost from CNP. In Chobe Comb retum mossambicense (43%) and Combretum elaeagnoides (40%) dominates the woody vegetation (Table 2 3). By contrast, the shrub layer in BNP is dominated by Baphi a massaiensis (29%) which was much more common in CNP in the 1960s (Child, 1968) Terminalia sericea (29%) is the most prolific tree species in BNP, and perhaps indicat ive of a perched aquifer in the ancient dune system as is often the case in Kenya (Tinley, pers com). Baikiaea woodland Baikiaea woodland in CNP is comprised of a tree (195/ha) and shrub (1,565/ha) layer. The density of the shrub layer in physiognomically similar woodland in BNP is similar (1,327/ha) but there are less trees (123/ha). The woody layer in BNP was more diverse than in CNP with 22 tree species compared to 9 in Chobe, and 32 shrub species compared to 16 in Chobe. Baikiaea woodland sites in Chobe are dominated by Baikiaea plurijuga (29%) and Markhamia zanzibarica, (30%), a pioneer species (Table 2 4) In BNP proportions are 53% and 18% respectively. This may reflect past logging practices in Chobe. The composition of the

PAGE 33

33 shrub layer, however, is markedly different. In CNP Combretum elaeagnoides (42%) dominates the shrub layer, whereas Baphia massaiensis (58%), is the dominant shrub in Namibia. Animal Composition and Density Analysis of dung plots in the three main vegetation types (Figures 2 11 through 2 1 3 ) suggest that large mammal composition and densities have changed enormously over time along the Chobe riverfront, although we caution that the absence of some species recorded in 1965 may be exaggerated because we collected data three months earlier than in 1965, a dry year when animals were very dependent on water from the river The most obviou s change demonstrated across sites is th at impala which were highly restricted in 1965 (Child, 1968), are now abundant on Pookoo Fl ats and in mixed wood/shrubland -where only elephant are more prolific. Elephant are the dominant species on all sites excep t Pookoo Flats, and are at much higher densities than in 1965, and roughly four times the density of BNP. Kudu have also increased markedly, presumably reflecting a shift in ecosystem resources from grass to shrubs. Buffalo densities remain similar to th e 1960s with their presence in Baikiaea woodland presumably linked to the fact that they have not all come down to the river by June/July in a wetter season when the field work was conducted. All three vegetation types exhibit a marked decline in speciali st grazing species or species with particular niches in 2007 compared to 1965, but we also recorded some species, particularly unspecialized browsers, that were uncommon or unrecorded then. In most cases, these changes make sense given the habitats of the species and the changes in habitats that we described above. On Pookoo Flats the 2007 transects did not record four large mammals prevalent in 1965 (i. e. wildebeest, puku, waterbuck a nd bushbuck) while giraffe and impala were new The disappearance of the wildebeest population can be attributed to the collapse of the Makgadikgadi population in the late 1960s from where these animals occasionally dispersed (Child, personal

PAGE 34

34 communication). Puku, which were once so prolific on the Flats that Selous name d the area after the sp ecies (Selous, 1881 ), were already thought to have decreased by the 1960s when they made use of the remaining perennial grasses in between the thick bush (Child, 1968) and their subsequent disappearance may be attributed to the col lapse of the grass sward (Figures 2 6 through 2 8 ). The degradation of the grass layer may also explain the disappearance of waterbuck and the decline in warthog density. Although bushbuck were a commonly seen browser on Pookoo Flats in the 1960s, we on ly saw two in this area over three field seasons, and their disappearance is probably related to competition with impala or kudu, and the loss of food species and cover in places formerly characterized as thickets and riverine habitats. Giraffe, by contrast, appear to have responded positively to the habitat changes we have recorded. In two years of intense field work in 1965 1966 o nly five giraffe were recorded betw een Kasane and Ngoma (Child, 1968), but they are now common on Pookoo Flats and in Baikia ea woodland, where they make extensive use of Capparis tomentosa (per obs). Giraffe are reported to disappear from the riverfront during the wet season Kudu, too, have increased and are now extremely common. They browse heavily on Combretum mossambicense on Pookoo Flats, as well as C. tomentosa, which they browse at a different level than giraffe. In t he mixed wood/shrubland sites 14 species were recorded in 1965 compared to six in 2007. Eight species were no longer present (sable, bushbuc k, duiker, eland, waterbuck, wildebeest, tsessebe, baboon) and two appeared much reduced (warthog, zebra). The explanation for the decline or disappearance of wildebeest, waterbuck, warthog and bushbuck reductions is the same as described for Pookoo Flats The relatively earliness of the field season may explain some of the decline or disappearance of zebra, sable, eland, and tsessebe; however, while we saw sable regularly, the same cannot be said for eland and tsessebe which appear to be

PAGE 35

35 far less ab undan t than they used to be ( Child, G., personal communication). Duiker are browsers with small home ranges, and their local decline can be attributed to the heavy use of these habitats by elephant and impala. The substantial increase in impala is consistent with eastward spread of this species in Chobe and is associated with areas heavily impacted by elephants. Buffalo maintain a strong presence, while kudu have increased with large amounts of favorable browse in the area ( Combretum mossambicense ). Similar trends are apparent in Baikiaea woodland sites. Four large mammals recorded in the 1960s were not recorded in 2007 (duiker, zebra, eland, sable), whereas the area has been occupied by impala, kudu, and giraffe. This is consistent with the explanations provided. More specialist large herbivores, especially grazers, have declined because of the large increase in elephants and the decline in perennial grass cover and loss of woody diversity, whereas unspecialized browsers have responded to an increase in l ess palatable, but abundant browse such as Combretum elaeagnoides, Markhamia zanzibarica, Baphia massaiensis and Combretum mossambicense The large increase in elephant in the Baikiaea woodland reflects the increase in the eleph ant population in CNP genera lly elephant passing through the Baikiaea woodland in large numbers on their way to the Chobe River every afternoon. The increase in elephant density to levels consistent with those in the mixed wood/shrubland on the face of the sandridge, the area that in the 1960s supported the highest elephant utilization levels, may also reflect both a change in the woody composition of the Baikiaea woodland and the need for elephant to move much further to get food. Many of the species listed above are palatable to elephant, and may encourage their use of this area, as well as decrease in density of woody species on the sandridge and in mixed wood/shrubland generally.

PAGE 36

36 Because of political unrest including the South African Angolan War and much later formal protect ion (1990) wildlife densities in BNP are much lower and less diverse than CNP, even to the casual observer. Our data shows that wildlife densities in BNP are 19% of CNPs in 1965 and 13% of what they are in 2007. Giraffe and impala, species that colonized the riverfront in Chobe between 1965 and 2007, were not recorded in our transects in Bwabwata (although impala were observed near roadways, which we did not sample because of road effects). In BNP, elephants were the most numerous species, albeit at one quarter the density of CNP. We recorded kudu in Baikiaea transects (and saw them regularly, although at a much lower frequency than CNP), while impala were present in the riparian transects in BNP, the habitat where they are found in the highest densitie s in CNP. Again, inland water pans had not fully dried up when we did our transects, forcing animals to congregate along the Kwando River. Discussion Our data confirm ones immediate impression that the Chobe riverfront has been heavily impacted by ele phants (and impala and kudu), and quantifies these change in three important components of savanna ecosystems the grass layer, woody vegetation and the large mammals. Note also that preliminary work on small mammals and insects in Bwabwata (in a related project) suggest much lower densities and diversities of these components in areas heavily used by elephants (Mfune, personal communication). The decline in grass cover, woody species diversity and large mammal diversity recorded in CNP between 1965 and 2007 is supported by a comparis on with Bwabwata National Park a similar (but not identical) ecosystem (Work and Owen Smith, 1986) with lower levels of past anthropogenic land use and animal pressure (Rice, 1997). The herbaceous layer along the Chobe rive rfront was already overutilized and declining in biomass in the 1960s because of past land use (i. e. cattle grazing, logging, agriculture, fire) and

PAGE 37

37 heavy wildlife (elephant, buffalo, hippo) pressure (Child, 1968). Our data suggests that the grass sward is certainly no better than it was in 1965, and had we conducted our work at the end of the dry season when previous measurements were taken, would presumably have been worse or much worse. While some annual grasses are present on Pookoo Flats and the Ba ikiaea woodland in CNP 1965, by 2007 all three vegetation categories have very little herbaceous cover (<2%), and show signs of bush encroachment, a typical successional step after a shift from perennial to annual species (Kelly and Walker, 1976). By comp arison, similar sites in BNP have considerably more grass, which can be explained by much lower wildlife densities (Rice, 1997), although we also note that fire is more frequent in BNP for anthropogenic reasons (park policy and practice) and because of hig her fuel loads. While BNP maintains peren nial grass cover with less annual grass, annual grasses are beginning to dominate the sward in Baikiaea woodland sites in BNP (Figure 2 8 ), possibly a response to frequent early dry season burning (Riney, 1963). Historically, f ire, especially frequent and early dry season burns lit by the Forestry Department to protect trees played a part in weakening the perennial grass layer in CNP So did use by cattle (1950s) and wild herbivores utilizing the river. This g ives rise to a degradation sequen ce typical of semi arid savanna heavy use and kicking up perennial grass tufts, leading to increasing bare ground and decreasing herbaceous biomass but with a higher component of annual grasses, followed by changes in soil water relationships that favor woody encroachment ( Walker et al., 1981; Child, B., 1986). Similarly, the decreasing vigor of the herbaceous layer opened the riverfront to bush encroachment on Pookoo Flats, first as Dichrostachys cinerea after extensive br owsing by cattle at this site in the 1950s and then as Capparis tomentosa-associated with the heavy utilization by game reported in this paper

PAGE 38

38 Degradation in the herbaceous layer was well established by the 1960s when elephant densities reached level s greater than 2/Km2 (Child, 1968). At the same time, e lephants began to decimate large riverine trees (Child, 1968); Child recorded l arge riparian type trees including Acacia tortilis, Acacia nigrescens, Phyllogeiton discolor (Burchemia discolor today), Croton megalobotrys, and a few Boscias on Pookoo Flats. He noted that the large trees showed extensive elephant damage, and by 2007 these species had been replaced by Croton megalobotrys on both Pookoo Flats and the portion of the riparian strip sampled for this study. Child (1968) noted the dominant species in each site mentioned, but not density, which is why we were unable to make this comparison Several of these have presently disappeared in the transects or are now uncommon, with one or two shrub species Child listed often now dominating the site. For example, Childs 1968 report describes Pookoo Flats as covered in dense bush, specifically Dichrostachys cinerea but also Ziziphus mucronata, and Combretum mossambicense In 2007 the area is domin ated by Combretum mossambicense we recorded only on e D. cinerea, Z. mucronata was not recorded in the transects, and the bush has opened up considerably because of high elephant and other herbivore use. Like Pookoo Flats, the mixed wood/shrubland (sandr idge) sites in CNP shows a disappearance of many tree species previously noted. Child (1968) sites Acacia giraffae /erioloba Acacia nigrescens, Phyllogeiton discolor (currently referred to as Berchemia discolor), Lonchocarpus nelsii, Pterocarpus martinii and Commiphora species as the dominant trees in the area, all of which are palatable to elephant. Of the six species listed, Acacia giraffae Pterocarpus martinii and Commiphora species have disappeared locally, and only Lonchocarpus nelsii maintains any noticeable presence.

PAGE 39

39 Species noted as part of the mixed wood/shrubland (sandridge) shrub layer in 1965 were Combretum mossambicense, Combretum elaeagnoides, Baphia obovata, Bauhinia macrantha, and Commiphora sp. All of these except Bauhinia and Commipho ra still exist, although Baphia is now rare. The two Combretum species are overwhelmingly dominant i n the shrub layer, account ing for 83% of shrub species. Visually, numerous dead trees also stand up over what is now a shrubland (Figure 2 2 2 3 ). Combi ning the tree and shrub layer, Combretum mossambicense dominates the site. The dominant woody species in Bwabwata National Park is Terminalia sericea often in a shrub form, a species that is common on deep sand in CNP but is not prevalent along the Cho be riverfront. This indicates some ecological differences between areas, and clouds our conclusion as to whether the substantially greater number of tree a nd shrub species in BNP compared to CNP are inherent or a function of land management However, Bap hia massaiensis dominates the shrub layer in BNP as it did in CNP in the 1960s which indicated some similarities between sites The abundance of Baphia massaiensis at many Namibia sites including the Baikiaea woodland and mixed wood/shrubland is notewort hy because it is considered fairly high quality browse. Although Child (1968) cited it as a feature of the shrubland on the sandridge in the 60s it was very rare in CNP by 2007, and is completely absent from the Chobe Baikiaea woodland sites. It is prob able that high animal densities in Chobe have already decimated this useful browse, and only left a few re mnants in select areas. Bwabwata National Park by contrast does not have nearly the same levels of animal densities and therefore has much more of th is species remaining.

PAGE 40

40 Dichrostachys cinerea is often associated with cattle. As noted, it once dominated Pookoo Flats in CNP but appeared to die naturally in the late 1960s. It is common in the Baikiaea woodland, in mixed wood/shrubland and even the grassland in BNP, where cattle grazed before the park was declared in 1990. Nevertheless, more palatable species like Z. mucronata and Baphia mass a iensis are being replaced by fast growing, softer, less palatable shrubs such as Combretum mossambicense, Markhamia zanzibarica or a self -supporting liana such as Capparis tomentosa. We have already suggested that past logging may be a contributing factor to less Baikiaea plurijuga in CNP (29%) than BNP (53%). In CNP d ominance of the tree layer is shared with Markhamia zanzibarica, an increaser species functionally similar to other tree species. Combretum elaeagnoides dominat es the shrub layer in CNP, whereas Baphia mass a iensis do minates the shrub layer in BNP and Dichrostachys cinerea is also fairly common in the shrub layer in BNP. Increase in elephant numbers in northern Botswana and additional attractive fauna such as puku, lechwe, sable and roan convinced the Botswana government to protect the area for tourism in the 1960s. Burning and ot her forms of land use were curtailed by 1967/68. Elephant numbers began to increase with continued protection from hunting, and the population in northern Botswana was established based on immigration from Hwange National Park in Zimbabwe (Campbell, 1990) and recruitment. The instillation of artificial boreholes within the park after 1966 may have played a part in augmenting the population. Elephants began to open up the scrub and woodland along the riverfront. Mosulego (2002) found a 5% increase in sh rubland and 15% increase in mixed woodland while woodland decreased by 25% within 8 Km o f the Chobe riverfront between 1962 and 1985. Elephant modified habitat is preferred by

PAGE 41

41 impala (Rutina, 2004) and the change in vegetation categories increased ecotona l area which is conducive to impala browsing. Impala increased significantly by the lat e 1980s (Rutina, 2004), a species which up until the 1960s was almost absent in the area (Child, 1968). As impala entered the riverfront area in high densities in the 1980s there was a 15% increase in shrubland and 11% decrease in woodland within 8 k m of the Chobe riverfront between 1985 and 1998 (Mosugelo et al., 2002). Impala curtail recruitment through extensive browsing which may provide certain species with a comparative advantage in reaching the shrub or tree layer but reduces recruitment generally. The changes in vegetation type near the Chobe riverfront instigated by elephant and realized by impala have had a profound effect on the mammal species in the ar ea. Mosugelo (2002) found that the conversion from woodland to mixed woodland and shrubland accelerated between 1985 and 1998; the conversion to shrubland doubled during the 23 year period between 1962 and 1985, while more than doubling during the 13 year period between 1985 and 1998. Elephant densities were increasing at an increasing rate in the 1980s before the rate of increase slowed down (Rutina, 2004). This is the time when the impala population is said to have increased rapidly (Rutina, 2004), which coincides with the time frame when Mosugelo found the accelerating change in vegetation cover. Because impala browse selectively, they can alter not only the density of the woody layer through reduced recruitment, but also its composition. This is probably accounted for the significant impact on the woody layer leading to a major shift in animal species as well. The conversion from dense shrubland to open shrubland on Pookoo Flats, from woodland to mixed woodland in the case of the Kalahari woodla nd sites, and mixed woodland to shrubland on the sandridge sites demonstrate how inherent differences in location (near or far to

PAGE 42

42 water) and composition (e.g. soil chemistry, moisture, terrain) respond differently to similar drivers, but at the same tim e follow general patters such as one or two shrub species taking over the woody layer, or a general loss of browsers and selective grazers within the animal component of the ecosystem Conclusions We have used a minimal set of transects, spatial and temporal comparisons, related literature and oral histories to quantify changes in herbaceous and woody vegetation and large mammal diversity in a study are on the Chobe riverfront that coincides closely with the primary tourism area. As can be seen from the results, the combination of Riney and Walker transects enabled us to provide meaningful results with two months of field work. The Riney method is much quicker, but is less quantitative and less us eful for measuring changes in woody vegetation. The Walker method is moderately time consuming but quantitative and repeatable and requires the observer to be able to identify species rather than function al components of the ecosystem (e.g. annual versus perennial grass). No park would lose by having a base line done with either method, nevertheless, given the progress being made in simplifying ecological monitoring for communities (Stuart Hill et al 2005) even simpler methods using fi xed point photographs and quicker transects (Stuart Hill, unpublished) should be tested. Since the 1960s, grass cover in the study area has almost certainly decreased (we did not measure species composition), reducing the prevalence of sensitive grazing s pecies (sable, roan, zebra, warthog, waterbuck, puku, lechwe). The diversity of the woody vegetation has been simplified, with large trees and palatable species being replaced by hardier shrubs and bushes, many of which nonetheless are in the highest cate gories of elephant damage. Unlike grazers and habitat specialists (e.g. bushbuck), elephants, kudu, impala and (perhaps) giraffe have thrived,

PAGE 43

43 and buffalo have held their own, presumably relying on floodplains. Our data show that changes were more exten sive in areas of higher fertility (e.g. Pookoo Flats), but that they nevertheless affected all components of the ecosystem (i.e. grass, woody plants and large mammals) even in the infertile Baikiaea woodland. This leaves the situation seen today with low herbaceous biomass, low quality shrub monospecies stands, and extremely high densities of a few animal species along the riverfront. However, although the diversity of the large mammals and woody vegetation has been simplified, we cannot conclude that th e overall biomass of wildlife and availability of browse decreased (this needs more detailed research). Nevertheless, our impression is that wildlife biomass is very high, and that browse is more plentiful than it was in the 1960s even if several of the m ore palatable species have disappeared. So what we have found is not surprising, just severe enough to raise important questions about ecological sustainability and resilience, about whether the park is successful, and about whether it is technical ly or politically possible to resolve the elephant question. Ecologically, our recommendation is to establish a wider baseline of Walker type transects, or a simplification of them, because this will provide quantitative, habitat and species level information that policy makers can use to make or support decisions. This monitoring also needs to be extended spatially, and importantly into areas that have lower elephant densities and where the ecology is still intact such as Moremi game Reserve and Bwabwata National Park. This study also raises important philosophical questions about the performance of protected areas, and the responsibility for monitoring this performance. Ecologically speaking, CNP has certainly conserved the endangered elephant. Howeve r, there have been significant trade offs in terms of grass cover, key ecological zones, large trees, palatable plants, and the

PAGE 44

44 diversity of mammals including species more threatened, if less well known, than elephants like tsessebe. These changes are, mo re than likely, also an indicator of losses of other taxa, as Cumming et al., ( 1997) and Mfune (ibid) found. However, a stated objective of the park was to develop an economic sector, and in this is has been extremely successful. Yet, as with the ecosyst em, there are concerns that the lack of diversity is problematic. Finally, is the issue of what to do with the elephants? To prevent this population of 216,000 elephant growing would require that some 13,000 animals were removed annually. On the one ha nd, culling is politically sensitive. On the other, these animals already support a large hunting and tourism industry, a multi -million dollar elephant economy that could be expanded further, with the greatest opportunities being in leather and leather -va lue adding activities (not ivory as commonly supposed). Alternatively, or additionally, is the option of working with Angola and Zambia to develop policies that enable large areas of sparsely settled land to become elephant economies through a combination of valorizing elephants (and associated species) and ensuring that landholders are the primary beneficiaries indeed, these principles lie behind southern Africas successes in private and community conservation (Suich and Child, 2008). However, to achi eve this, elephants need to be given priority to key ecological zones where humans are currently outcompeting them, such as the floodplains opposite Kasane and the river valleys that connect up the region. In conclusion, the management of parks is usuall y extremely complicated, with CNP being no exception. Given the stated importance of parks as the cornerstone of global conservation, it is surprising how little monitoring of their ecological, economic or social importance takes place (e.g. Cumming, 1997 ). We have used relatively simple and affordable monitoring (i.e. Riney 1963, Walker 1976) to provide quantitative data about the status of the

PAGE 45

45 key components of savanna ecosystems. These systems could easily be integrated into the annual workplans of park wardens. There have been substantial changes at each of the sites along the Chobe riverfront in Chobe National Park between 1965 and 2007 in the herbaceous, woody, and animal components. This is further exemplified by the large disparity between Bw abwata National Park and CNP, which demonstrate many similarities in woody species, especially between CNP 1965 and BNP, but large differences in health of the herbaceous layer, woody species composition and densities in 2007, and animal densities generall y. Is not clear what the simplification of the woody and mammal components of CNP will mean, although it brings up many questions of sustainability. While CNP has experienced significant changes over time, future work will be necessary to determine if th resholds have been crossed, if these thresholds are irreversible, and if there has been a decline in ecosystem resilience because of this. While some feel that elephant impact on vegetation is simply flu ct uation around an equilibrium and reversible (Linds ay, 1990), many others feel that vegetation change by elephants crosses a threshold leading to deterioration of the landscape (Child, 1968). Savanna environments are currently supporting the greatest levels of population growth of any biome on earth (Scholes and Walker, 1993), which makes the changes demonstrated all the more relevant in terms of these environments continued provision of ecosystem services and conservation of biodiversity. Tourism drives the economy in communities around CNP, and the imp lementation of the Kavango Zambezi Transfrontier Conservation Area ( KAZA TFCA ) in 2010 would amplify this situation around the region. It is important for communities and management institutions to be aware of the significant changes that have occurred along the

PAGE 46

46 Chobe riverfront so that changes in other similar environments can be anticipated in the face of similar drivers. Managers of these areas will need to make decisions regarding the instillation of bore holes thereby attracting wildlife and tourists in the short term, and trying to maintain habitat for future tourism, biodive rsity conservation and ecosystem services in the long term. The question of whether CNP is successful in terms of its intended purpose may be useful in determining this.

PAGE 47

47 Figure 2 1. Study site. A.) Southern Africa, Botswana highlighted. B.) Chobe N ational Park, northern Botswana. C.) Transect locations in Chobe National Park. Figure 2 2. Riverfront vegetation types : Floodplain, riparian strip, Pookoo Flats, sandridge, Baikiaea woodland

PAGE 48

48 Figure 2 3 Six vegetation types and their associated histories Figure 2 4 Comparison of rainfall preceding the 1965 and 2007 vegetation surveys Government of Botswana; Department of Meteorological Services

PAGE 49

49 Table 2 1 Number of transects per site and year Transect Numbers per Vegetation Category (n) CNP 1965 CNP 2007 BNP 2007 Riney Walker Riney Walker Riney Walker Pookoo Flats 7 5 5 Pookoo Flats Best Grassland 2 2 Mixed Wood/Shrubland 9 9 9 9 9 Baikiaea Woodland 3 7 7 5 5 Grassland 7 7 Floodplain 3 3 4 4 Riparian Zone Enumeration report Enumeration report 2 2 Figure 2 5. History of elephants and land use in Chobe National Park

PAGE 50

50 Figure 2 6 S patial and temporal comparison of bare ground in CNP (1965, 2007) and BNP (2007) in 3 vegetation types using Riney transects

PAGE 51

51 Figure 2 7 S patial and temporal comparison of rooted grass in CNP (1965, 2007) and BNP (2007) in 3 vegetation types using Riney transects Figure 2 8 S patial and temporal comparison of aerial grass cover in CNP (1965, 2007) and BNP (2007) in 3 vegetation types using Riney transects

PAGE 52

52 Figure 2 9 C omparison of riparian strip tree species in CNP (1965, 2007).

PAGE 53

53 Figure 2 10. C omparison of riparian strip tree species in Chobe (1965, 2007) and Bwabwata National Park (2007). Table 2 2 Woody species densities and percentages on Pookoo Flats. Pookoo Flats Woody s pecies d ensity per ha 1 % of s pecies present Overall woody s pecies Density per ha 1 152.1 100.0 Combretum mossambicense 59.8 39.3 Capparis tomentosa 51.4 33.8 Additional species recorded Total Species = 7 Grewia flavenscens, Croton megalobotrys, Lonchocarpus nelsii, Dichrostachys cinerea, Markhamia zanzibarica

PAGE 54

54 Table 2 3 Comparison of woody species composition and densities in mixed wood/shrubland between Chobe and Bwabwata National Parks. Mixed Wood/Shrubland CNP BNP Trees Shrubs Trees Shrubs Total Density ha 1 39.7 1373.1 533.6 2157.5 Total Number of Species 9 16 22 32 Terminalia sericea Density ha 1 0 0 151.5 57.3 % 0 0 28.4 2.7 Combretum hereroensce Density ha 1 0 Negligible 90.9 162.3 % 0 17.0 7.5 Baphia massaiensis Density ha 1 0 66.2 13.2 615.8 % 0 4.8 2.5 28.5 Dichrostachys cinerea Density ha 1 0 Negligible 44.8 343.7 % 0 8.4 15.9 Combretum mossambicense Density ha 1 0 587.2 4.0 14.3 % 0 42.7 .7 .7 Lonchocarpus nelsii Density ha 1 12.9 Negligible 40.8 47.7 % 32.5 7.7 2.2 Combretum elaeagnoides Density ha 1 0 543.0 0 47.7 % 0 39.5 0 2.2

PAGE 55

55 Table 2 4 Comparison of woody species composition and densities in Baikiaea woodland between Chobe and Bwabwata National Parks. Baikiaea Woodland CNP BNP Trees Shrubs Trees Shrubs Total Density ha 1 194.8 1,565.4 122.5 1326.5 Total Number of Species 8 14 11 15 Baikiaea plurijuga Density ha 1 55.9 14.7 65.2 53.8 % 28.7 .9 53.3 4.1 Markhamia zanzibarica Density ha 1 59.1 143.6 22.2 34.3 % 30.3 9.2 18.1 2.6 Baphia massaiensis Density ha 1 0 165.7 6.8 773.4 % 0 10.6 5.5 58.3 Grewia flavenscens Densit y ha 1 0 3.7 0 63.6 % 0 .2 0 4.8 Combretum elaeagnoides Density ha 1 0 644.6 0 0 % 0 41.2 0 0 Combretum mossambicense Density ha 1 0 117.9 0 83.2 % 0 7.5 0 6.3 Figure 2 1 1 Relative animal densities on Pookoo Flats

PAGE 56

56 Figure 2 1 2 Relative animal densities in mixed wood/shrubland Figure 2 1 3 Relative animal densities in Baikiaea woodland

PAGE 57

57 APPENDIX DESCRIPTION OF TREE AND SHRUB SPECIES Tree/Shrub Species Description of Species Combretum mossambicense shrub or creeper, functionally similar to many Combretums and other species Capparis tomentosa shrub or scrambler, found in overgrazed areas Terminalia sericea tree, grows on sandy soils Combretum hereroensce tree or shrub, grows on many soil types Baphia massaiensis shrub, common on disturbed sites Dichrostachys cinerea shrub or small tree, invasive on disturbed areas, good browse Lonchocarpus nelsii shrub or tree Combretum elaeagnoides shrub, small tree, scrambler, can invade disturbed areas Baikiaea plurijuga deciduous tree, grows on sandy soil, not palatable to elephants Markhamia zanzibarica deciduous shrub or tree, found in riverine, sandy soils and dunes Grewia flavenscens shrub, climber 1Tree Descriptions from Curtis (2005) and Palgrave (1977)

PAGE 58

58 BIBLIOGRAPHY Baillie, J.E.M., Hilton Taylor, C. & Stuart, S. N. (2004) IUCN Red List of Threatened Species. A Global Species Assessment, IUCN. Barnes, J.I. (1996) Changes in the economic use value of elephant in Botswana: The effect of international trade prohibition. Ecol. Economics 18, 215 230. Barrows, H. H. (1923) Geography as human ecology. Annals of the Association of the American Geographer. 13(1), 1 14. Belsky, J. A. (1990) Tree/grass ratios in East African savanna: a comparison of existing models. J. of Biogeography 17, 483 489. Blanc, J.J., Barnes R. F. W., Graig, G. C., Dublin, H. T., Thouless, C. R., Douglas Hamilton, I. & Hart J. A (2007) African Elephant Status Report 2007: An Update from the African Elephant Database IUCN Species Survival Commission. IUCN, Gland, Switzerland and Cambridge, UK. Campbell, A C. (1990) History of elephants in Botswana. The Future of Botswanas Elephants. Proceedings of a symposium organized by the Kalahari Conservation Society and Department of Wildlife and National Parks. Nov. 10, 1990. Gaborone, Botswana. Chafota, J (1998) Effects of changes in elephant densities on the environmen t and other species How much do we know? Cooperative Regional Wildlife Management in Southern Africa. Chapman, J. (1968) Travels in th e Interior of South Africa 18491863; hunting and trading journeys from Natal to Walvis bay and visits to Lake Ngami and Victoria falls. (Edited from the original manuscripts by Tabler, E.C.1971). A. A.Balkema, Cape Town. 2 Vols: 258+244 pp. Child, G. & v on Richter W (1968) Observations on ecology and behavior of le chwe, p uku, a nd w aterbuck al ong t he Chobe River, B otswana. Z. Sangetierk. 34, 275295. Child, G (1968) Report to the Government of Botswana on an Ecological Survey of NorthEastern Botswana, FAO, #TA2563. Rome, Italy. Cumming, D. H. M., Fenton, M. B., Rautenbach, I. L., Taylor, R. D., Cumming, G. S., Cumming, M. S., Dunlop, J. M., Ford, A. G., Hovorka, M. D., Jonston D. S., Kalcounis, M., Mahlangu, Z. & Portfors, C. V. R. (1997) Elephants, woodlands and biodiversity in southern Africa. S. Afr. J. of Science. 93, 231 236. Dublin H T Sinclair A R. E & McGlade J.(1990) Elephants and fire as causes of multiple stable states in the Serengeti Mara woodlands. J. of Animal Eco 49, 1147 64. Gaborone, Botswana. Henry, F.W.T. (1966) Enumeration re port o n the Chobe main forest block Min istry of Agriculture Gaborone Botswana

PAGE 59

59 Holling, C. S. (1973) Resilience and stability of ecological systems. Annual Review of Ecology and Systematics. 4 1 23. Hunter, M. D. & Price P W. (1992) Playing chutes and ladders: Heterogeneity and the relative roles of bottom up and top -down forces in natural communities. Ecology 73(3), 724732. Jackson, T. P., Mosojane, S., Ferreira, S. & van Aarde R. J (2008) Solutions for elephant Loxodonta africana crop raiding in northern Botswana: moving away from symptomati c approaches. Oryx 42, 83 91. Kelly, R. D. & Walker B. H. (1976) The effects of different forms of land use on the ecology of a semi arid region in south -eastern Rhodesia. J. of Eco 64, 553576. Langley, P. A. & Barnsley M. J (2004) The Potential of Geographical Information Systems and Earth Observation. In Unifying Geography: Common Heritage, Shared Future ed. John A. Matthews and David T. Herbert, 83 93London and New York: Routledge. Lindsay, W K (1990) Elephant/Habitat Interactio ns The Future of Botswanas Elephants. Proceedings of a symposium organized by the Kalahari Conservation Society and Department of Wildlife and National Parks. Nov. 10, 1990. Gaborone, Botswana Matthews, J. A. & David T. H. (2004) Unifying Geography: Common Heritage, Shared Future ed. John A. Matthews and David T. Herbert, 8393. London and New York: Routledge. Mendolsohn, J. & Roberts, C. (1997) An environmental profile and atlas of Caprivi. Directorate o f Environmental Affairs, Namibia. Moroka, D. N 1984. Elephants Habitat Relationships in Northern Botswana. DWNP Gaborone, Botswana Mosugelo, D. K ., Moe, S. R., Ringrose, S. & Christian, N (2002) Vegetation changes during a 36year period in northern Chobe National Park, Botswana. Afr. J. of Ec o 40, 232240. Palgrave, K. C (1977) Trees of Southern Africa. C. Struik Publishers. Cape Town, Johannesburg. Pringle, R. M., Young, T. P., Rubenstein, D. I. & McCauley D. J (2007) Herbivore initiated interaction cascades and their modulatio n by productivity in an African savanna. Proceedings of the National Academy of Sciences of the United States of America. 104, 193197. Rice, M (1997) Community-based natural resource management project in the Caprivi Region of Namibia. Cape Town. August 1997. Richards, K., Bithell, M. & Bravo M (2004) Space, Time and Science: Individuals, Emergence and Geographies of Space and Place. In Unifying Geography: Common Heritage, Shared Future ed. John A. Matthews and David T. Herbert, 6282. London and N ew York: Routledge.

PAGE 60

60 Riney, T (1963) A rapid field technique and its application in describing conservation status and trends in semi arid pastoral areas. Afr. Soils 8 159257. Rutina, L.P. (2004) Impalas in elephant impacted woodland: browser driven dynamics of the Chobe riparian zone, northern Botswana. PhD Thesis Agricultural University of Norway. s, Norway. Scholes RJ, & Walker B. H. (1993) An African savanna: S ynthesis of the Nylsvley Study Cambridge, UK: Cambridge University Press. Selous F.K (1881) A Hunter's Wanderings in Africa: being a narrative of nine years spent amongst the game of the far interior of south Africa. (1928 reprint). Macmillan and co., London. 504pp Simpson, C. D. (1975) A detailed vegetation study on the Chobe River in north-eastern Botswana. Kirkia 10, 185227. Skarpe, C., Aarrestead, P.R., Andreassen, H.P., Dhillion, S.S., Dimakatso, T., du Toit, J.T., Halley, D.J., Hytteborn, M., Mari, M., Marokane, W., Masunga, G., Modise, D., Moe, S.R., Mojaphoko, R., Mosugelo, D ., Motsumi, S., Neo Mahupeleng, G., Ramotadima, M., Rutina, L., Sechele, L., Sejoe, T.B., Stokke, S., Swenson, J .E., Taolo, C., Vandewalle, M. & Wegge, P. (2004) The return of the giants: ecological effects of an increasing elephant population. Ambio 33, 276 282. Stoddart, D. R. & William A. M. (2004) Fieldwork and Unity in Geography. In Unifying Geography: Common Heritage, Shared Future ed. John A. Matthews and David T. Herbert, 46 61. London and New York: Routledge. Stuart Hill, G., Diggle, R., Munal i, B., Tagg, J. & Ward, D. (2005) The Event Book System: a community -based natural resource monitoring system from Namibia. Biodiversity and Conservation 14 26112631. Suich, H. & Child B (2008) Evolution & Innovation in Wildlife Conservation, Earthscan, London: 462p Thornton, P K, Jones P G Owiyo T Kruska R L Herrero M Kristjanson P Notenbaert A Bekele, N & Omolo, A with contributions from Orindi V Otiende B. Ochieng A Bhadwal S Anantram K Nair S Kumar V & Kulkar U (2006). Mapping climate vulnerability and poverty in Africa. Report to the Department for International D evelopment, Nairobi Kenya. Pp 171. Vincent, P. & Whyte I (2004) Exploration, Discovery and the Cartographic Tradition. In Unifying Geography: Common Heritage, Shared Future ed. John A. Matthews and David T. Herbert, 21 32. London and New York: Routledge.

PAGE 61

61 Walker, B. H. (1976) An approach to the monitoring of changes in the composition and utilization of woodland and savanna vegetation. S. Afr. J. of Wildlife Research 6 (1), 1 32. Walker, B. H., Holling, C. S., Carpenter, S. R. & Kinzig A S. (2004) Resilience, adaptability and transformability in social ecological systems. Ecology and Society 9 (2), 5. Walker, B. H., Ludwig D. Holling C. S. & Peterman R. M (1981) Stability of semiarid savanna grazing systems. The J. of Ec o 6 9 473498. Work, D. R. & Owen -Smith, N. (1986) Preliminary Management Recommendations for Elephant in Northern Botswana. Resource Ecology Group, Department of Botany. University of the Witwatersrand. May, 1986.

PAGE 62

62 BIOGRAPHICAL SKETCH The author was born in Rochester, Michigan. She received her undergraduate degree from the University of Wisconsin-Madison in 1999. She then went on to be a Peace Corps volunteer in Gabon, where she worked as an environmental educato r. She was also a vo lunteer in Jamaica where she worked as a project coordinator for an environmental NGO. Her interest in s ustainable natural resource us e and development in Africa and the Caribbean then brought her to the Unive rsity of Florida to pursue her m asters degr ee. She plans to continue her education at the University of Florida to the doctorate level.