Christoph Paulik, TU Wien Sebastian Hahn, TU Wien Christoph Reimer, TU Wien Thomas Mistelbauer, EODC (Earth Observation Data Center) Alexander Gruber, TU Wien
The validation of remotely sensed soil moisture data is quickly becoming a routine task. During our work validating the H-SAF and Copernicus soil moisture products we saw the need for a standard toolbox performing common validation tasks. We have developed a suite of open source Python packages for this purpose which covers data reading, temporal matching, scaling, metric calculation and storage of results. It supports common in-situ (ISMN), remotely sensed (ASCAT, CCI, SMOS, SMAP) as well as often used model data (GLDAS, ERA Interim/Land).
To ensure reproducibility of results it also includes implementations of often used scaling methods (CDF matching, mean standard deviation scaling, etc.) as well as common metrics (R, RMSD, triple collocation etc.).
The framework itself is general enough to handle any number of datasets and combinations of metrics and is easily extended by new, user-written components. The validation results are stored as CF-conform netCDF files from which standard plots can be generated.
Additionally it includes a web-interface for interactive exploration of validation results compatible with self-hosted data or data accessible through OpenDAP or WCS.