Vegetation dynamics in Southern Africa from NOAA-AVHRR and SPOT-VGT time series
(2008)
Files
Swinnen.pdf
Restricted Access - Adobe PDF
- 11.21 MB
Details
- Authors
- Supervisors
- Lambin, Eric
- Abstract
- Understanding vegetation dynamics in Southern Africa in relation to climatic variability supports better management of natural resources. This study investigates the response of vegetation to rainfall variability in Southern Africa, based on a long-term time series of satellite images from SPOT-VGT and NOAA-AVHRR sensors. Of course, the differences in sensor-platform combination will hinder direct combination of these datasets. Therefore, the approach was, to first use VGT pre-processing as a benchmark for AVHRR pre-processing. Combination of the two NDVI datasets was improved afterwards by degrading the spectrally more accurate dataset, i.e. to decrease the dynamical range of the better' sensor rather than to upgrade the data from the oldest sensor to the new one. Comparison of the two NDVI datasets showed that though the values are similar, the discrepancy between the NDVI values of the sensors is spatially dependent. Applying empirical adjustment functions that account for the difference in spectral band definition improved the similarity of the two datasets. The performance of this adjustment was spatially dependent, and the one performing consistently best for the entire area was applied on the dataset. The resulting NDVI dataset was then used to study the spatial and temporal patterns of ecosystem dynamics in Southern Africa, in response to rainfall variability. To do so, a method was proposed based on wavelet coherency between the time series. Both the strength and the time lag of the vegetation response to rainfall were investigated, as well as the variability of these indicators over time. Further, we investigated the dependency of the relationship between NDVI and rainfall on mean annual rainfall, soil type, vegetation and topography. Among these, mean annual rainfall and topography, both at the site and at the watershed level, have the largest impact on the NDVI response to rainfall. The large influence of the topography at the watershed level on the strength and the response time between vegetation growth and rainfall has clear implications for the analysis of trends in vegetation change.
- Affiliations
Citations
Swinnen, E. (2008). Vegetation dynamics in Southern Africa from NOAA-AVHRR and SPOT-VGT time series. https://hdl.handle.net/2078.5/124071