Evaluating extreme precipitation in gridded datasets with a novel station database in Morocco.

Morocco is a large country with complex terrain and many sparsely populated regions. With a semi-arid climate, it is highly vulnerable to floods driven by extreme precipitation, whose distribution is highly variable in space and time. Yet, this topic has received little attention. The limited availability of data has so far been the major obstacle to pursue such research in Morocco. Public gridded datasets offer good opportunities to overcome this problem. However, the use of such data should be handled with care, especially when applying extreme value theory. The present work aims at addressing this issue. First, we introduce and analyse a comprehensive set of 120 daily precipitation series which we assembled from different stakeholders in Morocco. Then, we perform quality control of the data and use extreme value statistics to infer trends and large return levels. Finally, we assess the accuracy of nine gridded satellite-based and reanalysis daily precipitation datasets using the station data. These results are intended as a first step towards a comprehensive understanding of extreme precipitation in Morocco, and can help select gridded datasets for future hydrometeorological research. [ABSTRACT FROM AUTHOR]