War, Drought, and Phenology: Changes in the land surface phenology of Afghanistan since 1982

摘要

Land surface phenology (LSP) is the study of the spatio-temporal patterns of the vegetated land surface as observed by synoptic sensors at spatial scales relevant to atmospheric boundary layer processes. LSP provides an important approach to change detection in terrestrial ecosystems. In recent years, global warming has been used to explain many regional changes. However, anthropogenic influences on the land surface phenology, such as those due to institutional change, are both direct and significant. Afghanistan offers a unique case of a semi-arid country with multiple institutional changes over the last 20 years that occurred within the span of the satellite record. The PAL NDVI dataset has a relatively low spatial resolution of 8km but a high temporal resolution with a nearly cloud-free composite image available for every 10 days since 1982. We use this long image time series to describe the phenological development of the land surface. Phenological characteristics such as start and end of the growing season and maximum greenness differ as a function of land cover and vegetation type and are highly influenced by climatic extremes such as drought. In recent years considerable effort has been made to link NDVI variability to changes in temperature, precipitation, and atmospheric concentrations of greenhouse gases. However, other changes such as physical deterioration of irrigation structures, land use changes from wheat to poppy crops, and the destruction and recovery of permanent crops such as pistachios and almond orchards can significantly change the tempo of LSP. Here, we focus on the agricultural regions in Afghanistan. Daily temperature data are provided by the NCEP/NCAR Reanalyses at a spatial resolution of 2°×2° lat/lon. These data are used instead of ground data to maintain consistency in meteorological variables. We apply the statistical framework to each study period and compare the average NDVI across periods, and calculate the trends within in each study period. Furthermore, we discuss the significant changes in LSP in the region around Qandahar. Results reveal both the strengths and the limitations of LSP modeling in a terrestrial environment characterized by both high interannual and spatial variability as well as by socioeconomic turbulence.