Assimilating ESA-CCI Soil Moisture into the JULES-EMPIRE Data Assimilation System

Abstract:

Land surface models, such as the Joint UK Land Environment Simulator (JULES, the land surface component of the Hadley Centre models) are used in a wide variety of applications, such as climate modelling, flood prediction and crop yield forecasting. However, how best to implement Data Assimilation (DA) for these models remains an open question. At a fundamental level these models are very different from atmospheric models for which traditional DA was developed.

This poster describes the integration of JULES with the EMPIRE framework. EMPIRE (Employing MPI for Researching Ensembles) implements a test bed for ensemble based DA techniques that makes use of MPI message passing to exploit all available processing power. In particular EMPIRE contains several flavours of Particle Filter which show promise for the land surface DA problem. Examples of assimilating soil moisture observations from the ESA CCI data set into JULES are given for a number of sites in Africa. The model ensemble is generated by considering uncertainty in the driving data taken from the TAMSAT operational rainfall product. The results show considerable improvement in the modelled soil moisture and in particular the seasonal timing of the soil wetness.