P-41

Humidity Cross Correlations in Short Range Ensemble Forecasts

Dora Fohring (German Weather Service)

Variances of humidity and correlations with other parameters (temperature, wind) are highly variable and situation dependent. This makes it difficult to effectively use humidity observations in data assimilation systems which are based on climatological background error covariances like e.g. 3dvar. To overcome this limitation, ensemble data assimilation systems use ensembles of short range forecasts in order to provide flow dependent humidity background error variances and cross-correlations. Thus they should be able to make better use of humidity observations and constrain other correlated variables, like for instance wind vectors, from this data.

Here we take a closer look on the humidity variances and cross correlations as provided by the DWD global pre-operational Ensemble Data Assimilation System in order to assess the prospects of assimilating humidity data. Examples of humidity-temperature and humidity-wind cross-correlations are analysed for different synoptic situations.