Implementation of a Long-term Snow Water Equivalent Reanalysis over the Extratropical Andes

Abstract:

We describe the implementation and validation of a large-scale, snow water equivalent (SWE) reanalysis dataset over the Andean extratropical region (26°S – 36°S). The reanalysis dataset is based on leveraging information from the MERRA climatological reanalysis, a snow model and Landsat fractional Snow Covered Area (fSCA) data from 1985 to 2014. The assimilation methodology consists of a Particle Batch Smoother, which probabilistically conditions an ensemble prior snowpack estimate derived from MERRA and the snow model on time series of Landsat fSCA data. Uncertainty in precipitation estimates over the Andes proved difficult to quantify, and a statistical approach using snow survey data was devised in order to establish the best possible unbiased prior precipitation flux. Validation results show an increased accuracy of SWE estimates with respect to in-situ snow survey sites. The reanalysis framework is applied to derive climatological characteristics of the snowpack within a high-elevation Andean basin, part of the headwaters of the Maipo River in Central Chile. The large-scale implementation of the reanalysis includes Chilean and Argentinean basins making it of significant regional importance, and will result in spatially and temporally continuous SWE fields for the Landsat 5, 7 and 8 historical record.