Land use/Cover dynamics and its implications since the 1960s in the Borana rangelands of Southern Ethiopia

Getachew Haile, Mohammed Assen* and Abule Ebro**

Yabello Pastoral and Dry-land Agricultura Research Centre, Yabello, P.O.Box. 85, Ethiopia
getlews2002@yahoo.com
* Department of Plant Sciences, Haramaya University, P.O. Box 179, Haramaya, Ethiopia
Moh_assen@yahoo.com
** Adami Tulu Agricultural Research Centre, Ziway P. O. Box. 35, Ethiopia
abule_ebro@yahoo.com

Abstract

A land use/land cover dynamics and its implication since the 1960’s was studied in two selected districts of the Borana rangelands (South Ethiopia). The objectives were to examine the spatiotemporal patterns of land use/cover dynamics and quantify the rate and direction of these dynamics. Remotely sensed data including aerial photographs of 1967 and 1987, and ETM⁺ land sat satellite image of 2002 were interpreted. The images were processed using ERDAS imagine 8.6, which were then exported to Arcview GIS for change detection analysis.  

Five major land use/cover patterns namely grass land cover, woody vegetation cover, cultivated land, settlement and bare land were identified. The woody vegetation cover increased by 9% and 15 % in Yabello and Areero districts over 35 years, respectively. Cultivated land, bare land, and settlement increased by 2%, 5% and 3%; and 6%, 7% and 6%, in Yabello, and Areero, respectively. On the contrary, the grass land cover decreased by 8% and 34% in Yabello and Areero districts, respectively. This suggests that the expansion of other land use/cover types were at an expense of grassland cover; the main feed particularly for cattle, resulting in negative effects on local ecology and community. This has forced the local communities to expand cultivation onto marginal semi-arid lands that possibly resulted in the ecological disturbance of grazing land environments.

Keywords: Bare land, classification, GIS, grassland cover, vegetation changes

Introduction

Arid and semi-arid rangelands are characterized by erratic rainfall and high rate of vegetation dynamics (Herlocker 1999; Dahdough-Guebas et al 2002). The dynamics of plant species over time affects the biological diversity and productivity of rangelands, mainly due to continuous and complex interactions of the plant communities with their environment (Snyman 1998; David and Kerper 2003; Solomon et al 2006). In this regard, human interferences and climatic variations form common driving forces in bringing changes to the environment. The final outcome of this process is modification of the rangeland use/cover patterns, probably associated with a general decline in productivity of the rangeland environment.

Worldwide, 3 600 million hectares or 70% of the world’s arid lands, are degraded, and 10 million hectares of arable land deteriorates every year (Essahli and Sokona 2008). Measuring rates of dry-land degradation is a big challenge owing to the complex interactions between fluctuations in rainfall of semi-arid lands and anthropogenic changes (e.g. overgrazing, over exploitation of water sheds, etc) on vegetation cover (Helden 1991; Tucker and Van Praet 1991; Ayana and Oba 2007). Other factors, such as invasions by undesirable species and plant diseases would also make difficult the measurement of degradation of vegetation cover of arid and semi-arid ecosystems (Oba et al 2000; Gemedo et al 2006).

Until very recently, the Borana rangelands of Southern Ethiopia were considered to be one of the best grazing lands in east Africa (Coppock 1994). This was probably related to the low status of degradation patterns, strategic management of available resources by the pastoralists such as avoiding overstocking, unrestricted seasonal movements between dry and wet seasons and separate grazing culture for animals through categorizing by age and sex, lactation period and health conditions (Ayana 2007). In the last few decades, increased grazing pressure and development interventions (e.g. road construction, water developments etc) by government and non-government organizations resulted in degradation of the rangelands (Oba et al 2000) as a result of which a change in land use/cover pattern prevailed in the area. Obviously, this could be partly associated with a change and weakening of the traditional strategic management of the Borana rangelands. 

The development interventions undertaken in the Borana rangelands became unsuccessful, probably due to the lack of adequate information on the natural resource base in general and land use/cover patterns in particular. The availability of such type of information may aid in designing and implementing required development and management plan of resources. A historical analysis of resource base including land use/cover patterns can be obtained through interpretation of series of aerial photos and satellite images as it has been done in different parts of the world (Gete 2000; FAO 2002; Muluneh 2003; Reed and Dougill 2008).  However, there was no any systematic study on land use/cover patterns made for the pastoral zone in general and the Borana rangelands of Ethiopia in particular. Acquiring such types of information can help in understanding historical variations of grazing land resources and thereby contribute to its proper management. In fact, vegetation types form one of the main resources of pastoral area; sound decision on range management depends upon reliable information of the past and present vegetation resources (Herlocker 1999). The purpose of this paper therefore is to determine the temporal changes of land use/cover dynamics and quantify the rate and direction of these dynamics in Borana rangelands, Ethiopia.

Materials and methods 

Description of the study area

The Borana rangelands are located in the southern parts of the Ethiopia lowland. The area extends from 4° to 6° N latitude and 36° to 42° longitude.  Altitude ranges from 1 000 meters above sea level (m.a.s.l) to 1 700 masl and the topography consists of isolated mountains, valleys and depression. The study was conducted in two purposively selected districts of Borana rangelands, South Ethiopia. For convenience of the study, two villages were further selected from each district. These were Eleweya in Yabello and Afura in Areero district, both are situated in the hearts of Borana, and are considered good representatives of the Borana rangelands.

In the selected districts, the pattern of rainfall is bimodal, which is also generally true for the other rangelands of Borana. The mean monthly rainfall ranged from 9 mm to 142 mm in Yabello and from 2 to 114 mm in Areero districts and the mean annual rainfall were 527mm and 412mm in Yabello and Areero districts, respectively. Analysis of long term rainfall data showed a coefficient of variation ranging from 78 to 325%, and 61% to 209% in Yabello and Areero districts, respectively, indicating the presence of high variations in rainfall amount; a typical characteristic of semi-arid climates. In both cases, the main rainy season occurs between March and May with a peak in April and the short rainy season extends from September to November with a peak in October consisting of 59% and 27% of the annual total amount of rainfall, respectively. The mean annual temperature is about 24°C with a mean maximum of 28 °c and a mean minimum of 17 °c. The local people identified four seasons that include the cold dry season (June – August) called adolessa; the short rainy season (September – November) referred to as Bonaa and the long rainy season (March – May) described as ganna (Coppock 1994; Ayana 2007).

The Borana rangeland is dominated by savannah types of vegetation containing a mixture of perennial herbaceous and woody plants (Coppock 1994). These savannah communities vary from grassland to bush encroached areas known for variation in woody and herbaceous materials and marked shifts in composition that occur in response to grazing, browsing, burning and droughts or various combinations of these (Ayana 2007).

The soils of the study area have low fertility, probably due to low inherent fertility of the parent materials. They are approximated to be formed largely from granites and some volcanic and their mixtures (Coppock 1994). The soils are dominated with sandy loam textural classes.  The black clayey soils covered nearly 30% of the area and the others occupied the balance. Spatially, the valley and bottomlands are occupied by cracking and dark-brown colored soils of slight drainage impedance. These soils tend to be deep, reflecting the low or little erosion and more depositions of materials. On other hand, the steeper lands are covered with shallow, well-drained, light colored and eroded soils.

Data source and analysis

Data on land use/cover changes were obtained from the analysis of series of aerial photographs and satellite images. Two sets of aerial photographs (scale 1:50,000) for the 1967 and 1987 were obtained from Ethiopia mapping authority (EMA). In addition, enhanced land-sat thematic map (ETM⁺) for the year 2002 was obtained from the same source.

The aerial photographs were scanned at a resolution of 25μm using VEXCEL Imaging VX4000 Imagining systems from dia-positive transparencies to produce digital database. The images were then rectified using ERDAS IMAGINE 8.6 software. Vector maps of roads, road intersections, bridges, and rivers were overlaid on the imageries for referencing purposes. The resulting images were projected in the Universal Transverse Mercator (UTM). Thereafter, a mosaic was prepared for portions of the landscape that did not appear on the same photo images of the respective years to have a single image of interest area using the MOSAIC function of ERDAS IMAGINE software. Then, land use/cover of the study area was generated using supervised classification in ERDAS Imagine 8.6. The identification and classification using training sites were selected after careful analysis of topographic and aerial photographs. Stereoscopic analysis of the selected aerial photos was used to support field identifications during selection of training sites.  The training regions were shown to contain an adequate number of pixels and to be spectrally separable to avoid misinterpretation of land cover features for those with similar spectral signature. The training sites were taken as polygons that contain unique land-use or land-cover with known properties. After running the maximum likelihood classification, schemes with equal prior probability, the classified image was generated with five major types of land use/cover features (Table 1).

Then, the recognized land use/cover patterns from aerial photographs and satellite images were exported as a polygon into shape files with help of ArcGIS 9.0, which was then open in the arcview3.2 software.

Results and discussion  

A wealth of information was retrieved on land use/land cover changes with the help of interpretation of aerial photographs of 1967 and 1987 and satellite images of 2002. Interpretation of photos and images of these periods indicated the existence of considerable dynamics in the land use/land cover systems of the sample areas of the study districts. For convenience of the study, the 1967 land use/cover pattern was considered as a bottom line in the analysis of dynamics of land use/cover patterns in the study area. The recognized land use/ land cover patterns were presented in Tables 2 and 3.

Grass cover

Grass cover contained the largest land use/cover pattern in 1967, but thereafter showed a significant decrease in both study sites. Grasslands decreased from about 34% (3 945ha) of the total area in 1967 to 28% (3 248ha) in 1987 and 26% (3 017 ha) in 2002 for Yabello sampled site. This land use/cover pattern showed a higher change in Areero decreasing from 57% (12 314 ha) in 1967, 34% (7 500 ha) in 1987 and 23% (5 073 ha) in 2002 of the total area. Therefore, during the analysis period, a considerable decline in grassland cover was observed in Areero than in Yabello site. This may indicate a more important expansion of other land use/cover patterns in the Areero than in Yabello district.

The decrease in grass cover could be associated with a steady increase in the other land use/cover patterns such as wood cover land (Tables 2 and 3). As a result, an increase in woody vegetation cover has been reported elsewhere in response to heavy grazing in rangelands (Coppock 1994; Smit 2004). This will have a significant effect on ecology and socio-economic condition of the area and the consequences of the trends in land use/cover are far reaching.

Grasslands have been the main source of feed for livestock especially cattle in the area. This progressive reduction of grass cover may put serious constraints on animal feed and their productivity. Therefore, the relative increase in cultivated land could be a response to this problem in order to secure the required food demand of the society. This in turn could result in severe soil degradation, as has been shown in the expansion of bare lands of the present study areas.

Woody vegetation cover

The woody land cover, which is mainly composed of species of Acacia, Grewia and Commiphora, showed an increasing pattern in both areas over the analysis period. Of the total land area, the land covered by the woody vegetation in Yabello sampling site was 29% (3 365 ha) in 1967, 31% (3 597 ha) in 1987 and 36% (4 177 ha) in 2002; a consistent increasing pattern throughout the analysis period. Woody land cover pattern increased by about 2% between 1967 and 1987, 5% between 1987 and 2002 and 7% between 2002 and 1967 in Yabello site, giving an average change of 0.2 % per year. This reveals that the greatest change in the Yabello was observed between 2002 and 1987. This probably suggests that once a critical level of grassland degradation is reached, irreversible condition could occur and the rate of woody vegetation expansion could be dominant. It then indicates that a significant grassland deterioration took  place prior to the 1987 period with the slow but steady increase woody vegetation where a clear detection was made n the 1987 aerial photo.

There was a higher change in woody vegetation with 0.43% increase per year in the Areero than in the Yabello site. Of the total land area of the Areero sample site, the land covered by woody vegetation was 23% (4 921 ha) in 1967, 36% (7 941 ha) in 1987 and 38% (8 382 ha) in 2002. This showed an increasing rate of 13% between 1987 and 1967, 2% between 2002 and 1987, and 15% between 2002 and 1967. The greatest change in woody vegetation cover was observed between 1967 and 1987 in Areero site. This indicated a presence of early dominance of woody vegetation in the area than in Yabello. This may be due to low amount of rainfall in the study area. Furthermore, this may not probable suggest, only fluctuation but also the amount of rainfall could be important in shaping local ecology. In general, the overall increase in land covered by woody vegetation might be attributed to a number of factors such as declining pattern in grass cover and hence its replacement by the former land cover system as well as due to a ban on bushfire where by pastoralists used to keep balance between the grass and the tree layers in the past (McCarty et al 2002).

Cultivated land

Coverage of cultivated land increased over the analysis period in both study sites. In the Yabello sites, it covered 5% (580 ha) in 1967, 6% (696 ha) in 1987 and 8% (928 ha) in 2002. In Areero  sites, there was no any cultivated land in 1967 and in 1987 and 6% (1 324 ha) in 2002. The results showed that there was an annual increment of cultivated land by 0.09% with 3% increase over the entire study period in Yabello sites. The increase in Areero site was 0.17% per year with an overall increment of 6% during the whole of the analysis period. The analysis suggested that cultivation started earlier in Yabello than in Areero site. The possible reason is the interventions and awareness of the communities towards cultivations; and because of the emergence of different development interventions by government and non-government organizations in Yabello, which was preferred because of the better infrastructure in the area. The start of early cultivation and increase in the land covered by cultivation in Yabello was also an indication of the interventions undertaken in crop production over years. Much of these interventions could be due to the expansion of annual crops, particularly maize cropping, in meeting food security as perceived by the communities.

Bare land

In Yabello district, the bare land cover in 1967, 1987 and 2002 was 21% (2 436 ha), 22% (2 552 ha) and 16% (1 624 ha), respectively with an average decrease of 0.14% per year (Table 2). This showed a very slight increase between 1967 and 1987, but a declining pattern thereafter which might be related to the increasing pattern of land covered by woody vegetation that hide the visibility of the land from the air. On the other hand, a consistent increasing pattern in land covered by bare ground was observed in Areero district (Table 3). In this part of the study area, the land covered by bare ground consisted of 12% (2 995ha) in 1967, 17% (3 750ha) in 1987 and 18% (3 970ha) in 2002. This suggested the existence of a considerable increase in land covered by bare ground between 1967 and 1987, but a slight increase thereafter in Areero sampling site (Table 3).

Settlement land

Settlement land use/cover pattern showed a consistent (but different rates) increasing pattern in both districts during the study period. Of the total land area, in the Yabello sites, it covered 11% (1 276 ha) in 1967, 13% (1 508 ha) in 1987 and 14% (1 624 ha) in 2002. The overall increase was 0.09% during the study analysis period.

In Areero site, the land covered by settlement land use/cover pattern was 8% (1 828 ha) in 1967, 13% (2 867 ha) in 1987 and 15% (3 309 ha) in 2002. This showed a higher increase in settlement land in Areero than in Yabello site. The increase in settlement land may imply an increase in population size in both study districts. As observed from rate of increase in land use pattern, a higher rate of increase could be expected in Areero than in Yabello site. Also, a relatively high rate of increase in settlement land and hence in population size that occurred between 1967 and 1987 in the Areero sampling site.

Implications of land use/cover dynamics to soil erosion, rangeland and livestock management

Land cover is among the major factors affecting soil erosion (Wishmeier and Smith 1978). Removal of the protective vegetation cover exposes land to impacts of raindrops. This process accelerates detachment and removal of soil particles and its associated consequences (Wishmeier and Smith 1978; de Koning et al 1999). The problem becomes worse in semi-arid lands where with other factors there will not be enough vegetation cover to protect the land both from wind and water erosion processes. In semi-arid lands, such as the present study area, the amount of rain may not allow growing of vegetation to fully cover the land surface. But one single storm may cause severe damage to environment (Wishmeier and Smith 1978; de Koning et al 1999), and particularly becomes worse where the available vegetation cover is disturbed and lack of steady state equilibrium. In the absence of sufficient vegetation cover coupled with high effective temperature in oxidizing available biological matters, such as organic matter, soil particles have little granulation information. Soils with such characteristics are easily entertained by any agents of process of erosion (Snyman 1998). These processes then expose the land to wind erosion in dry and wet erosion in rainy times.

Considering the land use/cover status of the present study area, bare, cultivated and settlement land use /cover patterns could be more vulnerable to soil erosion. As a consequence of increased exposure of land through removal of natural vegetation cover (tables 2 and 3), the proportion of land prone to soil erosion increased from 1967 to 2002 in both districts. The total land area in the Yabello district, exposed land to erosion was 37% in 1967, 41% in 1987 and 38% in 2002;  probably due to the transition stage of the degraded grassland to woody cover land, as already discussed earlier. The canopy cover of woody vegetation observed in 2002 might have protected the land from direct attack of the soil by agents of erosion, thus lowering the vulnerable area to soil erosion from 41% in 1987 to 38% in 2002. On the other hand, in the Areero district, the proportion of land exposed to soil erosion increased from 225 in 1967 through 31% in 1987 to 39% in 2002 in this study area, it was relatively low from the other district both in 1967 and 1987 periods, and reached a maximum in 2002. The low proportion in the earlier periods could be related to the absence of cultivated as well as low proportion of settlement and bare land in the 1967 period. Hence, a high proportion of the land was exposed to erosion through time. This substantiates the existence of spatially variability of soil erosion within a defined ecosystem as stated by various authors (e.g. Snyman 1998).

In the Borana range lands, grazing resources have been owned communally and administered by traditional elders who formulate bylaws about their use with respect to time of grazing, types of animals to graze, etc. In the study area, grazing lands were categorized as (1) warra (used to graze lactating cows, sick and weak animals which return to the encampment every day); (2) forra (used for grazing bulls and non-lactating cows); and (3) calf enclosures. Forra grazing areas were customarily open to all communities. Permanent settlement in these parts was prohibited by traditional administrative units called madda elders. They were regarded as fallback areas for all community members during periods of forage scarcity. Areas designated as warra were normally open to members of the same arda (the lowest level of social organizations that governs resource use and allocation) but can be used by members of different arda under special arrangement usually on reciprocal bases.

Calf enclosures were thorn-fenced pastures that were reserved specifically for use by calves and to a lesser extent by milking cows. The use of calf enclosures was restricted to members of the community that built the fences, usually one or more olla’s (encampment). However, individual Borana households in areas around farms were constructing some enclosures with changes in land cover.

These organizations of grazing systems into warra, forra and calf enclosures emerged in response to seasonal feed shortages and nutritional stresses during periods of forage scarcity. The present patterns of land use/cover dynamics through its influence on the ecological conditions had disrupted these traditional grazing land management systems. The shrinkage and deterioration of grazing lands due to bush encroachment, cultivation and increase in the proportion of bare land had adversely affected traditional pastoral herd management system, which was well suited to the natural resource base of the area. The consequence is that livestock products, such as meat, milk and blood, which provided subsistence for the local community deteriorated from time to time.

The changes in land use/cover had also weakened the traditional management system on resource uses in the area as reported by the communities. As a result, the present condition of the traditional management systems in administrating the rangeland resource use is not as strong as 30 to 40 years back. Consequently, the shortages of grazing resources in the present study area limited the application and enforcement of traditional management systems.

Water is a critical scarce resource and in dry periods, the pastoral communities of the present study area move about 2.5 km to 30 km in search of water for their livestock and/or themselves. Like the other pastoralists elsewhere (Hellend 1997), the Borana pastoral communities employed a variety of strategies for survival in their marginal and risky environment. There were three main water sources in the present study areas. These included (1) occasional water sources, such as natural pools and puddles of rain lasting for few days during wet season; (2) temporal water sources, such as ponds and rivers that can be both natural and artificial; and (3) the permanent, traditional deep-wells that form the pivot of the pastoral life of Borana people.

Deep wells (Ela) formed the main source of water during the dry season followed by natural wells. Pools and puddles were more or less regarded as passage and are accessible to all people. These occur only at the peak of the rains, as a result of which, their use need not be restricted. During this period, the watering frequency of livestock is unlimited and cannot be determined. In the periods of acute water shortages, the frequency of watering cattle were reduced to three days to increase the number of herds that can be supported by the available well water.

The watering frequency for camels was every five day. Thus, the springs, wells and ponds were subjected to peculiar rules and regulations that were administered organized, supervised and instructed by the traditional elders ‘agents known as ponds and well managers.

Conclusions 

• Interpretations of remotely sensed data such as aerials photographs and satellite images were  found to be useful in analyzing historical land use/land covers changes in the present study. The advantage could be more with electronic mapping of geographical information systems (GIS). Applications of these integrated approaches in the Borana rangelands showed the existence of appreciable land use cover dynamics. In the analysis, the identified land use/land cover systems included grass land, woody vegetation, cultivated, settlement and bare land. In our study the Borana rangelands had faced a significant decrease in grass land cover since 1960s; associated with a significant increase in all the other recognized land use/cover systems.  This might have generally resulted in the deterioration of the ecosystem such as increasing the area prone to soil erosion and destruction of original grasses. On the other hand, it was not only coverage of grassland that decreased but also desirable vegetations species types might have been destroyed (e.g. increase in bush and woody vegetation). This directly decreased the productivity and carrying capacity of the area. As a result, local communities shifted to a change in land use pattern (e.g. expansion of cultivated land), seek more income generating activities (e.g. sell of firewood to near by urban centers), and/or rearing of more drought resistance animals. These latter activities could further drive ecological disturbance to irreversible conditions unless proper interventions are made in time.
   

References  

Ayana A 2007 The Dynamics of Savanna Ecosystems and Management in Borana, Southern Ethiopia; PhD Thesis. Norwegian University of life sciences (UMB), Norway.

Ayana A and Oba G 2007 Relating long-term rainfall variability to cattle population dynamics in communal rangelands and a government ranch in southern Ethiopia. Agricultural Systems   94 (3): 715-725

Coppock D L 1994 The Borana plateau of southern Ethiopia: synthesis of pastoral research, development and change, 1980-1991. International Livestock Centre for Africa (ILCA), Addis Ababa, Ethiopia. 393 pp. http://books.google.fr/books?id=TeLRgwQU1HEC&pg=PA193&lpg=PA193&dq=The+Borana+plateau+of+southern+Ethiopia:+synthesis+of+pastoral+research,+development+and+change,+1980-1991&source=bl&ots=0p6Fl7pXkZ&sig=e62KwsKm-jvp7Re-rMgeERgAlOk&hl=fr&ei=jEIBTPm7EcX_4AbMz_HLDg&sa=X&oi=book_result&ct=result&resnum=1&ved=0CB0Q6AEwAA#v=onepage&q&f=false

Dahdough-Guebas F, Kairo J G, Jayatissa L P, Cannicci S and Koedam N 2002 An ordination study to view vegetation structure dynamics in disturbed and undisturbed mangrove forests in Kenya and Sri Lanka. Plant Ecology 161: 123-135

David W and Kerper W 2003 Imagining stereoscopy for determining rangeland stressors in western watersheds. Pollution prevention and new Technology EPA/600/R-01004. 54p

De Koning G, Verburg P, Veldkamp A and Fresco L 1999 Multi-scale modeling of land use change dynamics in Ecuador. Agricultural Systems 61: 77-93

Essahli W and Sokona Y 2008 Policy requirements to combat desertification. In the Proceedings of UNESCO (Cathy Lee and Thomas Schaaf editors). International Scientific Conference on Desertification and drylands Research. Tunisia 19-21 June 2006. Man and the Biosphere series (UNESCO Publishing). 

FAO (Food and Agricultural Organization) 2002 GIS and spatial analysis tools for poverty and food insecurity mapping. Environmental and natural resources, working paper No.7. Rome, Italy. 124p

Gemedo D T, Maass B L and Isselstein J 2006 Rangeland condition and trend in the semi-arid Borana lowlands, Southern Oromia, Ethiopia. African Journal of Range and Forage Science 23: 49-58 

Gete Z 2000 Landscape dynamics and soil erosion processes modelling in the north-western Ethiopia highlands.  Africa studies series A16, University of Berne, Berne.

Helden U 1991 Desertification- time for an Assessment. Ambio 20: 372-383.

Herlocker D J 1999 Rangeland Ecology and Resource Development in Eastern Africa. Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ), Nairobi, Kenya. 393p.

McCarthy N, Kamara A and Kirk A 2002. The effect of environmental variability on livestock and land use management: The Borana plateau, Southern Ethiopia. Socioeconomics and policy research working paper 35. ILRI (International Livestock Research Institute), Nairobi, Kenya and IFPRI (International food policy research institute), Washington, DC, USA. 35p. http://www.ifpri.org/sites/default/files/publications/eptdp75.pdf

Muluneh W 2003 Impact of population pressure on land use and land cover changes, agricultural system and income diversification in west Gurague land, Ethiopia. PhD dissertation, submitted to Norwegian university of science and technology (NTNU). Trondheim. 340p.

Oba G, Post E, Syvertsen P O and Stenseth N C 2000. Bush cover and range condition assessments in relation to landscape and grazing in southern Ethiopia. Landscape Ecology 15: 535-546

Reed M S and Dougill A J  2008 Participatory land degradation assessment. In the Proceedings of UNESCO (Cathy Lee and Thomas Schaaf editors). International Scientific Conference on Desertification and drylands Research. Tunisia 19-21 June 2006. Man and the Biosphere series (UNESCO Publishing) 

Smit G N 2004 An approach to tree thinning to structure southern African savannas for long term restoration from bush encroachment. Journal of Environmental Management 71: 179-191

Snyman H A 1998 Dynamics and sustainable utilization of rangeland ecosystems in arid and semi-arid climates of Southern Africa. Journal of Arid Environments 39: 645-666

Solomon T B, Snyman H A and Smit G N 2006 Rangeland dynamics of southern Ethiopia: (2) assessment of woody vegetation structure in relation to land use and distance from water in semi-arid Borana Rangelands. Journal of environmental management,  85 (2): 443-452

Tucker C J and Van Praet B 1991 Satellite remote sensing of total dry matter production in the Senegalese Sahel. Remote Sensing of Environment 13: 361-374

Wishmeier W H and Smith D D 1978 Predicting rainfall erosion losses. A guide to Conservation planning. Handbook.No.537.USDA, Washington. 

Received 14 September 2009; Accepted 9 May 2010; Published 1 July 2010

Go to top