The subtropical thicket biome in the Eastern Cape Province of South Africa has been heavily degraded and transformed due overutilization during the last century. The highly degraded and transformed areas exhibit a significant loss of above ground carbon stocks (AGC) and loss of SOC content. Information about land use /cover change and fragmentation dynamics is a prerequisite for measuring carbon stock changes. The main aim of this study is to assess the trends of land use/cover change, fragmentation dynamics, model the temporal changes of AGC stocks in the Great Fish River Nature Reserve and its environs from 1972 to 2010, quantify and map the spatial distribution of SOC concentrations in the partial subtropical thicket cover in the Great Fish River Nature Reserve and environs (communal rangelands). Multi-temporal analyses based on 1972 Landsat MSS, 1982 and 1992 Landsat TM, 2002 Landsat ETM and 2010 SPOT 5 High Resolution images were used for land use/cover change detection and fragmentation analysis. Object oriented post-classification comparison was applied for land use/cover change detection analysis. Fragmentation dynamics analysis was carried out by computing and analyzing landscape metrics in land use/cover classes. Landscape fragmentation analyses revealed that thicket vegetation has increasingly become fragmented, characterized by smaller less linked patches of intact thicket cover. Landscape metrics for intact thicket and degraded thicket classes reflected fragmentation, as illustrated by the increase in the Number of Patches (NP), Patch Density (PD), Landscape Shape Index (LSI), and a decrease in Mean Patch Size (MPS). The use of remote sensing techniques and landscape metrics was vital for the understanding of the dynamics of land use/cover change and fragmentation. Baseline land use/cover maps produced for 1972, 1982, 1992 2002 and 2010 and fragmentation analyses were then used for analyzing carbon stock changes in the study area. To model the temporal changes of AGC stocks in the Great Fish River Nature Reserve and its environs from 1972 to 2010, a method based on the integration of RS and GIS was employed for the estimation of AGC stocks in a time series. A non-linear regression model was developed using NDVI values generated from SPOT 5 HRG satellite imagery of 2010 as the independent variable and AGC stock estimates from field plots as the dependent variable. The regression model was used to estimate AGC stocks for the entire study area on the 2010 SPOT 5 HRG and also extrapolated to the 1972 Landsat MSS, 1982 and 1992 Landsat TM, and 2002 Landsat ETM. The AGC stocks for the period 1972 -1982, 1982-1992, 1992-200) and 2002-2010 were compared by means of change detection analysis. The comparison of AGC stocks was carried out at subtropical thicket class level. The results showed a decline of AGC stocks in all the classes from 1972 to 2010. Degraded and transformed thicket classes had the highest AGC stock losses. The decline of AGC stocks was attributed to thicket transformation and degradation which were caused by anthropogenic activities. To map and quantify SOC concentration in partial (fractional) thicket vegetation cover, the spectral reflectance of both thicket vegetation and bare-soils was measured in situ. Soil samples were collected from the sampling sites and transported to the laboratory for spectral reflectance and SOC measurements. Thicket vegetation and bare soil reflectance were measured using spectroscopy both in situ and under laboratory conditions. Their respective endmembers were extracted from ASTER imagery using the Pixel Purity Index (PPI). The endmembers were validated with in situ and laboratory thicket and bare-soil reflectance signatures. The spectral unmixing technique was applied to ASTER imagery to discriminate pure pixels of thicket vegetation and bare-soils; a residual spectral image was produced. The Residual Spectral Unmixing (RSU) procedure was applied to the residual spectral image to produce an RSU soil spectrum image. Partial Least Squares Regression (PSLR) model was developed using spectral signatures of a residual soil spectrum image as the independent variable and SOC concentration measured from soil samples as the dependent variable. The PSLR prediction model was used to predict SOC concentration on the RSU soil spectral image. The predicted SOC concentration was then validated with SOC concentration measured from the field plots. A Strong correlation (R2 = 0.82) was obtained between the predicted SOC concentration and the SOC concentration measured from field samples. The PSLR was then used to generate a map of SOC concentration for the Great Fish River Nature Reserve and its environs. Areas with very low SOC concentrations were found in the degraded communal villages, as opposed to the higher SOC values in the protected area. The results confirmed that RS techniques are key to estimating and mapping the spatial distribution of SOC concentration in partial subtropical thicket vegetation. Partial thicket vegetation has a huge influence on the soil spectra; it can influence the prediction of SOC concentration. The use of the RSU approach eliminates partial thicket vegetation cover from bare soil spectra. The residual soil spectrum image contains enough information for the mapping of SOC concentration. The technique has the potential to augment the applicability of airborne imaging spectroscopy for soil studies in the sub-tropical thicket biome and similar environments.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:nmmu/vital:10672 |
| Date | January 2013 |
| Creators | Nyamugama, Adolph |
| Publisher | Nelson Mandela Metropolitan University, Faculty of Science |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Thesis, Doctoral, PhD |
| Format | xiv, 203 leaves : col. ill., 1 map, pdf |
| Rights | Nelson Mandela Metropolitan University |
Page generated in 0.0023 seconds