Moisture is obviously an important part of the earth atmosphere as it is the main carrier of (latent) heat from the ocean. It is also an important green house gas and a source for clouds and precipitation. Large scale analysis systems, as used in numerical weather prediction (NWP), make an attempt to estimate the state of the atmosphere on a predefined grid in an optimal way. Moisture is part of such systems although emphasis tends to be on wind and temperature. Products from NWP systems are highly relevant for research on the hydrological cycle because NWP systems assimilate a large number of irregularly spaced observations and observation types in a consistent way. The products are available without gaps on a standard grid and with reasonable time resolution. An example of such a system is the global Integrated Forecasting System as used at ECMWF for operational forecasting and for the 40 year re-analysis project (ERA40).

A description of the ECMWF system will be given with emphasis on moisture. Apart from the atmospheric analysis it also has a soil moisture analysis over land. Comparison with independent observations demonstrate that the moisture analysis relies heavily on consistency with the dynamical fields which are analyzed at the same time. Examples from operational analyses compared to radiosondes, buoys and SSMI demonstrate the high level of realism of the analyzed fields. ERA40 data is used to study the different terms in the atmospheric moisture and the land soil moisture budgets. The issue of spinup and its consequences will be discussed.