The assessment of global trends in land degradation

Abstract

The motivation for quantitative assessment of land degradation at a global scale is its recognition as an environmental issue of global societal implications. Yet, due to the non-robust definition of “land degradation” and to the paucity of field data, the five global assessments carried out and presented between 1977 and 2003 differ in the selection of measurable attributes of land degradation, in the quality of the data sets, and in their spatial coverage. This resulted in a plethora of degradation estimates ranging 15% to 63% of global degradation and 4% to 74% of dryland degradation. Of these, the figure of 70% degradation (for the drylands only, comprising 41% of global land) has been cited more than the others. Though likely to be overly exaggerated (because it stands for a combination of degradation degree of a land unit and its spatial extent within the mapping unit of which it is a part), this high estimate has apparently served well the globality notion of the dryland degradation syndrome, essential to rallying support for international development assistance under the UNCCD. This thirst for development assistance aimed at “combating desertification” attracted to the UNCCD some 70 non-dryland developing countries (compared to 93 developing dryland country Parties) which experience land degradation that is not included in global assessments of desertification, since only dryland degradation qualifies as “desertification”. The texts of the various assessments, including that of GLASOD as well as the UNCCD definition often trade off “desertification” with “susceptibility” to or “threat” of desertification. This suggests that an assessment of vulnerability to desertification rather than its actual occurrence are of higher credibility and utility for policy- and decision-making. Though soil degradation featured highly in the currently available global degradation assessments, remotely-sensed vegetation attributes not only assess the most valued but threatened ecosystem service, but are also amenable for assessment at the global scale. However, caution is required when using this tool especially in drylands where productivity is tightly linked to rainfall variations. The monitoring required to meet the persistence criterion for qualifying desertification can be also used to detect current desertification trends, which are of relevance for policy-making even more than defining current desertification status. To discern changes of productivity due to state of the land from those due to rainfall features, the ratio of NPP to rainfall (RUE) could be useful were it not negatively correlated with rainfall itself. An alternative method for detecting degradation trends, the Residual NPP Trends (RESTREND) is currently under development. It is based on an analysis of the residuals of the productivity-rainfall relationship throughout a time period for each pixel in the explored region. A statistically significant negative regression of the residuals on time identifies a degradation trend, and the slope stands for its magnitude. To be reliable on a global scale such a remote-sensing approach would serve for guiding field observations required for its own verification.