Soil moisture is a key component of the hydrological cycle which is relevant for climate modelling, numerical weather prediction and runoff forecasting, as well as for natural and managed ecosystems modelling. Many studies on land-atmosphere interactions have emphasized its important role for the climate system, due to its impact on evapotranspiration, and more particularly on the partitioning of latent and sensible heat fluxes.

In spite of its importance, soil moisture is an unmeasured quantity in most parts of the globe. The few available soil moisture measurements are confined to small areas and are generally conducted in level agricultural fields which are not representative of the average land surface. As a consequence of this lack of data, various derived soil moisture datasets based partly or wholly on model calculations were created for climate studies. These datasets do generally not present a very good agreement with the observations available and do often not agree with one another.

In this study, we use an approach based on water balance calculations for estimating seasonal changes in soil moisture storage for large areas. The feasability of the approach is investigated in a case study for the Mississippi river basin using the very recent ERA-40 reanalysis data of the European Centre for Medium-Range Weather Forecasts in combination with conventional runoff data.