Cloud obstruction and snow cover in Alpine areas from MODIS products
Snow cover maps provide information of great practical interest for hydrologic purposes: when combined with point values of snow water equivalent (SWE), they enable estimation of the regional snow resource. In this context, Earth observation satellites are an interesting tool for evaluating large scale snow distribution and extension. MODIS (MODerate resolution Imaging Spectroradiometer on board Terra and Aqua satellites) daily Snow Covered Area product has been widely tested and proved to be appropriate for hydrologic applications. However, within a daily map the presence of cloud cover can hide the ground, thus obstructing snow detection. Here, we consider MODIS binary products for daily snow mapping over the Po River basin. Ten years (2003–2012) of MOD10A1 and MYD10A1 snow maps have been analysed and processed with the support of a 500 m resolution Digital Elevation Model (DEM). We first investigate the issue of cloud obstruction, highlighting its dependence on altitude and season. Snow maps seem to suffer the influence of overcast conditions mainly in mountain and during the melting period. Thus, cloud cover highly influences those areas where snow detection is regarded with more interest. In spring, the average percentages of area lying beneath clouds are in the order of 70%, for altitudes over 1000 m a.s.l. Then, starting from previous studies, we propose a cloud removal procedure and we apply it to a wide area, characterized by high geomorphological heterogeneity such as the Po River basin. In conceiving the new procedure, our first target was to preserve the daily temporal resolution of the product. Regional snow and land lines were estimated for detecting snow cover dependence on elevation. In cases when there was not enough information on the same day within the cloud-free areas, we used temporal filters with the aim of reproducing the micro-cycles which characterize the transition altitudes, where snow does not stand continually over the entire winter. In the validation stage, the proposed procedure was compared against others, showing improvements in the performance for our case study. The accuracy is assessed by applying the procedure to clear-sky maps masked with additional cloud cover. The average value is higher than 95% considering 40 days chosen over all seasons. The procedure also has advantages in terms of input data and computational effort requirements.