Sensitivity of L-band NWP forward modelling to soil roughness

Abstract

This article investigates the sensitivity of the European Centre for Medium-Range Weather Forecasts (ECMWF) simulated L-band brightness temperatures (TB) in response to different soil roughness parameterizations. To this end, the ECMWF operational conditions during the year 2004 have been used to force the Hydrology Tiled ECMWF Scheme for Surface Exchanges over Land (HTESSEL) coupled to the Community Microwave Emission Model (CMEM). The coupled HTESSEL-CMEM system is then run at five different incident angles (20°, 30°, 40°, 50° and 60°) and for five soil roughness parameterizations available in CMEM. The performance of the simulated TB is analysed at ground point scale over the Surface Monitoring Of Soil Reservoir Experiment (SMOSREX) site in south-west of France. For this particular data set, both ground-based vertical profile of soil moisture and L-band radiometric observations are available for evaluation of the ECMWF forecast over the nearest grid box. In particular, the results show that the simple Choudhury parameterization best fits the observations for both horizontal (H-pol) and vertical (V-pol) polarization and for most of the incidence angles tested. The best forward modelling configuration is at for the V-pol, with coefficient of determination between modelled and observed TB of 82.9% and root mean squared error of 7.9 K. The sensitivity of the L-band TB errors to the empirical soil roughness parameter is also investigated. Strong sensitivity to this parameter is shown, mainly at H-pol for the least rough surfaces. The investigation carried out in this article gives an insight into the soil roughness model to be used in the operational configuration of the CMEM L-band forward operator, for future assimilation of the Soil Moisture and Ocean Salinity (SMOS) satellite data of the European Space Agency.

Acknowledgements

The authors of this article would like to thank the SMOSREX team. The SMOSREX project was co-funded by the ‘Programme National de Télédétection Spatiale’, by the ‘Programme Terre Océan Surface Continentales et Atmosphère’ (CNES) and by participants to the experiment: CESBIO (CNES, CNRS, IRD, UPS), CNRM/GAME (Météo-France, CNRS), INRA and ONERA, all in the framework of the SMOS science preparatory program. Many thanks to M. Drusch for the development of CMEM as in the preparation activities for the assimilation of SMOS data in the Integrated Forecast System of the ECMWF.