Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Pays Bas → Analysis of soil - vegetation interaction in relation to soil carbon sequestration : a case study in Serowe, Botswana

International Institute for Geo-Information Science and Earth Observation (ITC) 2003

Analysis of soil - vegetation interaction in relation to soil carbon sequestration : a case study in Serowe, Botswana

Ermias Aynekulu

Titre : Analysis of soil - vegetation interaction in relation to soil carbon sequestration : a case study in Serowe, Botswana

Auteur : Ermias Aynekulu

Etablissement de soutenance : University of Twente International Institute for Geo-Information Science and Earth Observation (ITC)

Grade : Master of Science in Geo-Information Science and Earth Observation 2003

Résumé
To quantify the amount of carbon sequestered in the soils, various models have been developed. A closer estimate, however, is achieved through the conventional soil survey in which soil samples are collected from a representative soil pedons, which could be a demanding approach. Although remote sensing is a useful tool in mapping and quantifying biophysical resources, it has a limited application in acquiring information on Soil Organic Carbon (SOC) particularly in dry environments. Since vege- tation is considered as a major source of SOC, this study aimed at evaluating the extent to which the variability in SOC can be explained by above ground woody biomass (as important vegetation vari- able) and to see the possibilities of up-scaling SOC estimation using vegetation indices as a remote sensing options. The study was conducted around Serowe village (26 0 01’37”- 26 0 57’57” East and 22 0 11’46”-22 0 35’31” South), in Botswana, having semi-arid agro-climatic conditions. The project area (267328ha) was stratified in to three main physiographic units, namely : sandveld, escarpment and hardveld. Based on a stratified random sampling design, 75 sample plots were allocated for each stratum optimally. Ankle height basal area was measured in the field and converted to biomass using the existing regression equations developed by Tietema (1993). To convert the fresh weight biomass to dry biomass, moisture content of eight dominant species were determined. From every plot, bulk density (using core sampler), % course fragments and % C (using Walkley-Black method) were de- termined to estimate SOC for the top 30cm. NDVI, SAVI and LAI were estimated form Landsat TM 7(October 2001), ASTER (August, 2001) and MODIS (March, 2002). Pearson correlation was ap- plied to evaluate the relationships between SOC and biomass and vegetation indices. Finally, regres- sion analysis was carried out to evaluate the extent to which woody biomass contributes to the SOC variability and to estimate biomass and SOC from vegetation indices. The results indicate that, bio- mass poorly explains the variation in SOC, R 2 = 17% and 0%, in the sandveld and escarpment, respectively. However, SOC was better explained by biomass in the hardveld (R 2 = 49%). By taking the Arenosol, which covers 78 % of the project area, biomass could explain 13 % of the SOC variability. To explain the remaining SOC variability, clay content for 19 plots were estimated (using Pipette method) and, it was found to explain about 42 %. Of the vegetation indices NDVI has greater correlation with biomass and SOC. However, relatively better estimation was possible in the escarpment (R 2 =0.35 for biomass and 0.23for SOC) where Combretum apiculatum is the dominant species with greater density and biomass. 95 % confidence for average standing biomass, SOC stock and carbon in biomass were estimated to 10.48±1.29 t/ha 20.6± 1.82 t/ha 4.51± 0.55 t/ha respectively. Generally, the SOC was poorly explained by biomass and vegetation indices were found to have application in estimating biomass and SOC in the area. The amount of carbon sequestered in a unit area of soils was found to be relatively low and the possible reasons given were geological and climatic reasons.

Version intégrale (ITC )

Page publiée le 29 janvier 2018