Spatiotemporal variability analysis of vegetation cover status for drought study purposes over North Africa using 8-km NDVI-GIMMS data

Africa is considered as the second driest continent in the world behind Australia, its arid lands cover approximately 60 percent. Droughts that hit the continent in recent decades and land degradations at the margins of deserts, particularly for countries neighboring the Sahara such as Algeria, Morocco, and Tunisia have renewed concerns about desertification progress. For a better understanding to the complete nature of drought and the degree in which human activities and climate changes contribute to its development, it is imperative to determine this phenomenon more accurately. Previous drought assessments had several weaknesses making them less reliable. Indeed, standard measurement methods, based on an unevenly distributed sampling point network, are unrepresentative neither for spatial distribution nor for temporal frequency of desertification. As an alternative to these conventional methods, remote sensing data could offer the needed spatial and temporal coverage. In fact, drought study is possible through monitoring vegetation and/or temperature conditions using vegetation and/or temperatures driven indices. Thus, several indices including the Normalized Difference Vegetation Index (NDVI), Transformed Difference Vegetation Index (TDVI), Soil Adjusted Vegetation Index (SAVI), and Temperature Condition Index (TCI), have been employed using various satellite sensors such as Advanced Very High Resolution Radiometer (AVHRR), Landsat Thematic Mapper (TM), and MODerate Resolution Imaging Spectroradiometer (MODIS) to monitor and analyze drought in various regions of the world. In this study, the AVHRR NDVI-GIMMS data at 8-km spatial resolution were used to study the evolution of the vegetation cover status in North Africa's countries over 25 years. The NDVI-GIMMS data were highly correlated with the rainfall in situ samples collected on different cities over the North Africa's countries as the determination coefficient (R2) was about 0.96. The vegetation cover study was based on a temporal analysis of the annual mean and the annual standard-deviations of NDVI imageries, taken over years from 1982 to 2006, through a descriptive statistics (average, maximum, minimum, and standard-deviation) and a trend analysis using a modified version of Mann-Kendall test. The descriptive statistics analysis has highlighted regions under a constant degraded status of vegetation. It has also allowed mapping the extreme limits of North Africa's areas affected by vegetation degradation. On the other hand, the trend analysis has pointed out the stable regions and those evaluating towards drought or to a wet status. And it has also aided, in combination with the descriptive statistics results, to propose several scenarios of vegetation evolution on North Africa's countries.