Improved parameterization of the soil emission in L-MEB

launch early 2008. When operational, it will provide global microwave brightness temperature observations at L-band, in dual polarisation and under a range of viewing angles, both, over land and water surfaces. For the purpose of relating the observed quantities to soil moisture, the L-band Microwave Emission of the Biosphere (L-MEB) model has been developed, specifically for its use for L-band emission over land surfaces. The core of L-MEB is the t-w approach which takes into account vegetation cover effects on the soil moisture signal. L-MEB is the result of an exhaustive review of "state of the art" modelling approaches in the field of passive microwaves, with the objective of being accurate while remaining simple enough for operational use at global scale. It is the forward model used in the SMOS level 2 processor, in order to produce geophysical products, e.g. soil moisture (SM) and vegetation characteristics. Based on tests made against several experimental data sets acquired over bare soil fields (SMOSREX at the Fauga site, near Toulouse, PORTOS and EMIRAD data sets acquired at the INRA Avignon site, MELBEX campaign in the semi-arid Mediterranean environment near Valencia, etc), improved parametrisations of the soil component in the L-MEB model have been developed. In particular, new calibrations of the HR (parametrizing the intensity of the surface roughness effects), Q (parametrizing depolarization effects), NRV and NRH parameters (parametrizing the dependence of emissivity on incidence angles) have been computed. The use of a new physical routine (Mironov et al., 2008), better accounting for the effects of free water in dry and sandy soils in the computation of the soil dielectric permittivity was also evaluated. In parallel, a very large synthetic data sets computed by coupling a SVAT model and a radiative transfer model, was built to evaluate the effects of soil structure and texture (in terms of sand and clay fraction) on the soil effective temperature. This paper will make a synthesis of the main results obtained during the evaluation of the Level-2 algorithm over these different experimental and synthetic data sets over bare soil surfaces.