Abstract/Summary

Understanding vegetation dynamics provides information on changes in land cover that can directly be related to regional changes in the climate system. In data-sparse regions, i.e. northwest Morocco studies are limited by the availability of comprehensive information on precipitation. We extracted precipitation data of high spatiotemporal resolution (2 km, 1 day) from the Northwest Africa Reanalysis (NwAR) and gridded Normalized Difference Negetation Index (NDVI) of the Moderate Resolution Imaging Spectroradiometer (MODIS) that cover northwest Morocco over ten hydrological years (September 2000 to August 2010). The results are based on a sequence of linear regression analyses. The mean precipitation of different input timeframes is systematically applied as the predicting variables to the mean NDVI of the growing seasons. Results show that 73 % of the variance in mean NDVI is explained by the variance in mean precipitation at the beginning of the growing season (November to the end of December). The results also show that 75 % of the variance in the mean NDVI of agriculturally used areas is explained by the variance in mean precipitation of beginning September to the end of December. Potentially irrigated land cover of low to medium explained variance but of a high seasonal range in NDVI cover about 14 % of the study region. We conclude that a considerable part of agricultural used areas are still potentially rain-fed. The applied methods and especially the re-analysed precipitation data of high spatiotemporal resolution open a new quality of analysis valuable for, e.g. monitoring aspects, policy decisions or regulatory actions.