Inverse modeling of soil characteristics from surface soil moisture observations: potential and limitations

Abstract. Land surface models (LSM) are widely used as scientific and operational tools to simulate mass and energy fluxes within the soil vegetation atmosphere continuum for numerous applications in meteorology, hydrology or for geobiochemistry studies. A reliable parameterization of these models is important to improve the simulation skills. Soil moisture is a key variable, linking the water and energy fluxes at the land surface. An appropriate parameterisation of soil hydraulic properties is crucial to obtain reliable simulation of soil water content from a LSM scheme. Parameter inversion techniques have been developed for that purpose to infer model parameters from soil moisture measurements at the local scale. On the other hand, remote sensing methods provide a unique opportunity to estimate surface soil moisture content at different spatial scales and with different temporal frequencies and accuracies. The present paper investigates the potential to use surface soil moisture information to infer soil hydraulic characteristics using uncertain observations. Different approaches to retrieve soil characteristics from surface soil moisture observations is evaluated and the impact on the accuracy of the model predictions is quantified. The results indicate that there is in general potential to improve land surface model parameterisations by assimilating surface soil moisture observations. However, a high accuracy in surface soil moisture estimates is required to obtain reliable estimates of soil characteristics.