Abstract
Land degradation is a global environmental issue with social and economic impacts at many scales. Defining land degradation as a long-term decline in ecosystem function and productivity, we may measure it using long-term, remotely sensed NDVI data; deviation from the norm may serve as a proxy assessment of land degradation and improvement if other factors that may be responsible are taken into account.
Few assessments have been published, some focusing on descriptive mapping of land degradation, some based on expert opinion and qualitative assessment that cannot be reproduced. Our quantitative approach aims at reproducible mapping at the global scale that allows continual updating.
Developing existing methods, we use the GIMMS NDVI dataset (1981-2006), improving previously used linear trend and residual analysis on yearly accumulated NDVI by employing harmonic analyses of NDVI time-series (HANTS). HANTS enables correction for phase-shifted productivity cycles (instead of using simple annual cycles), providing optimal estimates of the start-of-season, multiple yearly growing cycles compensation, and non-linear trend analysis to incorporate potential greening to browning trend changes. This reduces the false alarms otherwise generated as well as providing a coherent global stratification of greenness trends, without using a priori land cover information.
de Jong, R