Preview only show first 10 pages with watermark. For full document please download

Analysis Of Soil Organic Carbon And Vegetation Cover Trends Along The Botswana Kalahari Transect

Determination of trends in soil organic carbon (SOC) and vegetation cover along savanna ecosystem moisture gradients is critical to the understanding of ecosystem functioning and global change. Field results from 57 sites along the Botswana Kalahari

   EMBED


Share

Transcript

    Analysis of soil organic carbon and vegetation covertrends along the Botswana Kalahari Transect SusanRingrose*, WilmaMatheson† & CornelisVanderpost* *Department of Environmental Sciences, University of Botswana, P/Bag 0022, Gaborone, Botswana †Westwood International School, P. O. Box 2446, Gaborone, Botswana  Determination of trends in soil organic carbon (SOC) and vegetation coveralong savanna ecosystem moisture gradients is critical to the understanding of ecosystem functioning and global change. Field results from 57 sites along theBotswana Kalahari Transect (BKT) showed general increases in both SOCand vegetation cover components along the temperature/moisture gradient.However, details in both SOC and woody cover trends revealed an area of relatively declining values in the central Kalahari. Image classification using TM single band data as input values confirmed the existence of ananomalously low woody cover area in the north-central Kalahari, within anarea mapped as northern Kalahari tree and shrub savanna. It is postulatedthat the occurrence of such a zone of low shrubs in a gradient of otherwiseincreasing tree height and structural complexity may result mainly fromedaphic factors, because of the co-incidence of near surface calcrete.However, the effects of fire and heavy grazing also influence the extent of lowabove-ground biomass, at least locally. Attempts were made to examine theextent to which soil organic carbon and vegetation components can bepredicted using Thematic Mapper (TM) imagery, to increase potentially thenumber of sites in the BKT database. As a result of correlation and regressionanalyses, it was determined that the imagery-based SOC indexes of theliterature were not useful in the Kalahari environment because of theextremely low SOC values encountered. Results of multiple regressionanalysis confirmed that woody vegetation cover could be predicted mainlythrough the use of TM3 single band radiance values throughout the BKTeven during a non-drought period. Woody vegetation cover could bepredicted to a lesser extent by the Transformed Normalised VegetationIndex. As high correlation coefficients were found between woody vegetationcover and SOC, the density of woody cover as derived from TM3 data maybe taken as a surrogate for relative SOC abundance. Because this method isindirect, field checking will be required to verify results.©1998 Academic Press Limited Keywords: Botswana Kalahari Transect; soil organic carbon; moisturegradient; vegetation anomalies; thematic Mapper imagery; woody covercomponents Journal of Arid Environments  (1998) 38 : 379–3960140–1963/98/030379+18 $25.00/0/ae970344©1998 Academic Press Limited  Introduction Global change studies include the determination of how soil organic carbon (SOC)content changes across climatic gradients in response to differing vegetation covertypes (Walker & Menaut, 1991). This provides background information on the overallcarbon budget (Scholes & Walker, 1993) and the measurement of sources and sinkscritical to climate change impact analysis (Graves & Reavey, 1996). Normally carboncycling is considered to be impacted by different land uses of which losses by burningare probably the most critical (Crutzen & Andreae, 1990; Houghton, 1991). Despiteextensive fires which periodically occur in the Kalahari environment, recent workunder the Botswana Country Studies programme has estimated the Botswana Kalaharias a major carbon sink (Mathale, 1997). Although humus accumulation tends to beminimal under dry savanna conditions due to relatively low leaf densities, a significantcontribution may be derived from below-ground biomass which is greater than above-ground biomass in certain areas due to fires (Scholes & Walker, 1993).If increased levels of CO 2 can be said to stimulate plant production even over theshort-term, then levels of newly derived organic carbon may be increasing in Kalaharisoils (Graves & Reavey, 1996). However, the viability of the sink may be altered byland use or land management changes. Global change and terrestrial ecosystem workin Australia indicates that CO 2 probably has a minimal effect on the tree:grass ratioand that grazing intensity and fire frequency are much more important (Will Steffen,pers. comm., 1997). This also appears to be true in Botswana, where, for instance, anincrease in Acacia  spp. or woody weed dominated areas as a result of cattle postestablishment has already been recognized (Ringrose & Matheson, 1991; Ringrose et al  ., 1996). A number of inter-related biogeochemical and human factors are known toinfluence woody species distribution in savanna environments (e.g. Skarpe, 1990;Perkins & Thomas, 1993). However these changes have not previously beenconsidered in terms of the moisture–temperature gradient in southern Africa.Here attempts are made to quantify trends in SOC and vegetation covercomponents primarily by sampling along the Botswana Kalahari Transect (BKT)moisture–temperature gradient. As the characterization of SOC and vegetativecomponents in the field is essential but time consuming it is important also to establishthe extent to which SOC and vegetative parameters can be predicted, for instance byusing Thematic Mapper imagery. Little widely published literature is available on thevariability of SOC and vegetation cover components along the Kalahari Transect inBotswana. Few literature examples which specifically deal with changes in spectralresponse resulting from response from natural vegetation and SOC changes along aclimatic gradient are available, especially in Africa (e.g. Matheson & Ringrose,1994 a  , b  ). Recent studies have suggested that critical aspects of vegetation cover can bemapped using satellite imagery along semi-arid vegetation–moisture gradients (Mathe-son, 1994; Ringrose et al  ., 1994; Shoshany et al  ., 1995). Use of remotely sensed datacan also lead to estimates of net primary production (Goward & Dye, 1987) andbiomass (Elvridge & Lyon, 1985), sometimes through the use of  vegetation indices (Tucker & Sellers, 1986; Choudhury, 1987; Justice et al  ., 1991). In addition, soilsurface reflectance measurement using fine resolution imagery is regarded as apotentially accurate and therefore efficient method for estimating SOC content (Baumgarder et al  ., 1985; Frazier & Cheng, 1989; Henderson et al  ., 1989; Wilcox et al  ., 1994). The objectives of the research reported here are primarily to determine present daylevels and trends of SOC and vegetation cover components along the BotswanaKalahari Transect (BKT) in relation to known vegetation belts. A secondary objectiveis to determine the extent to which satellite (Thematic Mapper) data can be used topredict SOC levels and vegetation cover components in the Botswana Kalahari. Thiswork is also intended to establish a preliminary soil–vegetation cover database along S. RINGROSE ET AL. 380  the BKT and to provide a measure of the value of Landsat derived values formonitoring environmental change (e.g. McKeon et al  ., 1992). Study area  The study area focuses on a 1000km south-west to north-east transect acrossBotswana. The Botswana Kalahari Transect (BKT) extends through semi-arid/subtropical vegetation zones approximately orthogonal to the main isohyets (White,1983). Whereas the BKT is part of a larger Kalahari transect and as such contributesto the International Geosphere Biosphere Project (IGBP) global set of transects (Koch et al  ., 1995), the present work is restricted to Botswana only. Rainfall along thegradient varies from less than 200mm in the south-west (CV>45%) to over 650mmin the north-east (CV<35%) and falls during the summer months (October toMarch). Potential evapo-transpiration rates vary from >2000mm year –1 in the south-west to between 1000–1500mm year –1 in the north-east (Hulme, 1996). Recentanalysis of actual evapo-transpiration using 1-month Meteosat data indicated a rate of 1mm day –1 for south-west Botswana and 3mm day –1 for the north-east (Peters,1995).Vegetation communities over Botswana have previously been mapped by Weare & Yalala (1971). Along the line of the transect these are characterized initially by arid shrub savanna in the south-west (Table 1; Fig. 1). This zone is divided into moresoutherly arid shrub savanna, southern Kalahari bush savanna and more northerlycentral Kalahari bush savanna (Fig. 2). Further north, tree savanna predominates andis referred to as the northern Kalahari tree and bush savanna (Fig. 3). This extendsnorthwards into the broadleaf  Mophane belt, which merges north-eastwards into thedry deciduous Chobe forest (Fig. 4). The overall trend is one of increasing woodycover, increasing tree over shrub cover and increasing density (and biodiversity) of savanna species from south-west to north-east. The predominant land use varies alongthe gradient from mixed wildlife with smallstock grazing in the south, to more intensivecattle grazing and browsing in the centre, to mainly commercial forestry and wildlifein the north. Fires are known to be widespread, especially in a dry season followingheavy rains when the fuel load is high.Kalahari soils along the BKT comprise arenosols mapped by the Soil Survey andAdvisory Services Project (1991) at 1:1,000,000 and 1:250,000. The arenosols arecharacterized by high fine sand percentages (average 62%) in both surface and lowerhorizons. The organic carbon percentages vary (on average) from 0·20% in upperhorizons to 0·08% in the subsoil. The sandy soils have an average infiltration rate of 33cm h –1 (ranging from 54·3 to 18·5cm h –1 ), a porosity of around 40% and availablemoisture content of between 5–10% by volume (Joshua, 1981). While the Kalahariarenosols are relatively uniform on the surface, considerable subsurface heterogeneityprevails as the sands are interbedded with layers of calcretes and silcretes of varyingthickness (Thomas & Shaw, 1991). The effect of the proximity of calcretes inparticular on arenosol pH and ground-water infiltration and storage on woody coverstructure and distribution requires further investigation ( cf  . Ringrose et al  ., 1997 c  ). Field work  Field work took place in the wet season (December–March) 1994/1995 along the1000km transect. The wet season was chosen to provide information on amounts of live herbaceous cover and to ensure identification of prevalent tree/shrub species. Field SOIL AND VEGETATION, KALAHARI TRANSECT381  work also corresponded to the time of the satellite overpass. Fifty-seven sites wererecorded in the field using access afforded by the road network mainly based on offsetsalong the BKT gradient. Each site was located using Magellan 7000AX GlobalPositioning System. Soil samples and vegetation cover records were obtained for eachsite. Field work consisted of pacing 90  90m transects along which all species wereidentified and their canopy cover assessed (Ringrose & Matheson, 1990, 1992).Quadrats thrown at 30-m intervals formed the basis for quantifying live and deadherbaceous cover and plant litter. Soil type, colour and condition were also recorded. Two soil samples were taken from each site 45m along the first transect or in thenearest open area. The uppermost sample was taken directly from the surface afterremoval of debris. A second sample was taken at a depth of approximately 20cm. ae . Main vegetation zones in Botswana and prevalent species (from field data and Weare & Yalala, 1971)  Vegetation zoneWoody speciesHerbaceous speciesArid shrub savanna Acacia haematoxylon, A. mellifera,Eragrostis lehmanniana,Boscia albitrunca, RhigozumE. heteromera, E. biflora,trichotonumStipagrostis amabilis,Aristidia stipitata, Schmidtia kalaharensis  Southern Kalahari Acacia erioloba, A. mellifera,Aristidia uniplumis, Eragrostis  bush savanna A. hebeclada, Boscia albitrunca,latimanniana, Schmidtia Grewia flava, G. retinervis,bulbosa, Antephora pubescens,Dichrostachys cinerea, ZiziphusAristida meridionalis mucronata  Central Kalahari Acacia erioloba, A. mellifera,Aristidia uniplumis, Eragrostis  bush savanna A. hebeclada, A. fleckii,latimanniana, Schmidtia Terminalia sericea, Lonchocarpusbulbosa, Antephora pubescens,nelsii, Boscia albitruna, GrewiaAristida meridionalis flava, G. retinervis, Dichrostachys cinerea, Ziziphus mucronata  Northern Kalahari Burkea africana, PeltophorumAristidia uniplumis, tree and bush africanum, Terminalia sericea,A. meridionalis, Eragostis  savanna Croton spp. , Lonchocarpus nelsii,pallens, E. superba, Antephora Combretum spp., A. fleckii,pubescens, Heteropogon A. luederitzii, A. mellifera,contortus A. tortilis  Mophane Colophospermum mophane,Aristidia meridionalis, Eragostis  woodlands A. erubescens, Rhus tenuinervis,pallens, Antephora pubescens Ochna pulchra, Ximenia caffra,Commiphora spp.Chobe dry Baikiaea plurijuga, Bauhinia spp., Schmidtia bulbosa, Aristidia  deciduous forest Ricinodendron rautanenii,uniplumis, A. meridionalis,Pterocarpus angolensis, BurkeaEragostis pallens,africana, Erythrophleum africanum,E. lehmanniana, Chloris Lonchocarpus capassa, Terminaliavirgata sericea  S. RINGROSE ET AL. 382  28 ° 18 ° Chobe district20 ° 22 ° 24 ° 26 ° 28 ° 26 ° 24 ° 22 ° 20 ° 6554Ghanzi districtNgamiland district5Central districtSouth EastdistrictKgatlengdistrictSouthern districtKweneng districtKgalagadi district210 200km3North Eastdistrict Satellite imagery and soil analysis Forty Landsat Thematic Mapper (TM) scenes taken from the wet season, 1994–95were mosaiced into an image of Botswana by GIMs (Pty) Ltd. The data comprised afully processed 800 Mb file projected to the UTM co-ordinate system (Zone 35) usingthe Clarke 1880 spheroid. Pixel size was degraded to 75m. The Landsat TM datacomprised one visible band (TM3 centered on 0·660 µ m), one near infrared band(TM4 centered on 0·830 µ m) and one mid-infrared band (TM5 centered on1·650 µ m). Spectral data from each field site were obtained using the Inquire Cursorfunction in ERDAS Imagine 8·2 and taken as an average of nine pixels around the site(e.g. Ringrose & Matheson, 1990). Figure 1. Main vegetation zones encountered along the Botswana Kalahari Transect (afterWeare & Yalala, 1971). (1)=Arid shrub savanna; (2)=Southern Kalahari bush savanna;(3)=Central Kalahari bush savanna; (4)=Northern Kalahari tree savanna; (5)=Morphanetree and bush savanna; (6)=dry deciduous forest. SOIL AND VEGETATION, KALAHARI TRANSECT383