Long-Term Land Degradation in the Caucasus

Land degradation, both of forest and in grasslands, is widespread in the Caucasus, and a major environmental challenge. However, land degradation is more difficult to monitor with satellite imagery than, for example, outright land use conversion or forest disturbance. One reason is that understanding land use change and land degradation trends greatly benefits from long-term land use data, but assessing land use prior to the Landsat record is challenging. However, once those technical obstacles are overcome, studying land degradation can greatly advance land use science. The reason is that land degradation patterns in peripheral areas can be greatly affected by demand for resources from the cores of economic activity, especially in countries where ample oil and gas reserves foster rapid growth of their cores, and studying degradation patterns thus allows to investigate core-periphery relationships. This is why we propose here to:

a) develop new remote sensing approaches to monitor degradation that are based on a) the integration of multiple years of Landsat data while accounting for phenology, and b), changes in endmember mixtures, phenology metrics, and the classification likelihood of a given land cover class, in order to map forest and grassland degradation across Armenia, Azerbaijan, and Georgia;

b) apply new methods to rectify 1960s and ‘70s Corona data semi-automatically with structure-from-motion software, and develop new methods to classify Corona data with image segmentation; and

c) examine the effects of economic cores, especially those related to oil and gas development, on land use in the peripheries in Armenia, Azerbaijan, and Georgia. This proposal builds upon a LCLUC project by PI Radeloff on land use change and the effect of wars in the Caucasus that will end in February of 2018, i.e., right when the work proposed here would start. The current project, now in its third year, has been highly successful, and highlighted that while land use conversions are uncommon, land degradation is widespread. The new work proposed here will build upon our findings, and extend them further. We will create wall-to-wall maps of land degradation, examine long-term LCLUC via the analysis of Corona imagery, and advance the understanding of core-periphery relationships and the effects of oil and gas development as drivers of land use patterns. Our project will make extensive use of NASA remote sensing assets, develop new remote sensing methods, and advance land use science via the collaboration of remote sensing specialists, economists, and experts from the Caucasus region.