Assimilation of Sentinel-1 backscatter into a land surface model for soil moisture and leaf area index updating: Impact on streamflow simulations

Abstract

<p>Streamflow forecasts suffer from errors in the initial conditions of the catchment-scale soil moisture distribution. In this research, we evaluate the potential of improving streamflow simulations through the assimilation of Sentinel-1 backscatter data into a land surface model. Our modeling setup consists of the Noah-MP land surface model coupled to the HYMAP river routing model and the 'Water Cloud Model' (WCM), which acts as backscatter observation operator, integrated into the NASA Land Information System. The system was set up at 1 km resolution for two contrasting catchments in Belgium: i) the Demer catchment dominated by agriculture and low topographic gradients, and ii) the Ourthe catchment dominated by mixed forests and high topographic gradients. Surface soil moisture and leaf area index (LAI) dynamically simulated by Noah-MP in an open-loop run were used to calibrate the parameters of the WCM using a Bayesian objective function and Sentinel-1 backscatter data processed to 1 km spatial resolution for the period 2015-2021. We present results of a suite of data assimilation experiments obtained from an ensemble Kalman filter that updates both soil moisture and LAI. We tested the use of (i) WCM parameters that were calibrated using backscatter data from all Sentinel-1 orbits simultaneously or using data from each Sentinel-1 orbit separately, (ii) backscatter observations with or without seasonal bias correction, (iii) backscatter observations in VV and VH polarization separately or combined. The different data assimilation experiments are evaluated with leaf area index from optical remote sensing, microwave-based soil moisture retrievals and streamflow measurements.</p><p>Preliminary results indicate substantial differences between the different data assimilation experiments. For the Ourthe catchment, streamflow skill improvement was highest when simultaneously assimilating VV and VH observations without bias correction but using orbit-specific WCM parameters. For soil moisture and LAI, however, the highest skill was obtained by assimilating only VV observations. For the Demer catchment, assimilating observations without seasonal bias correction led to a skill degradation for streamflow while the impact of data assimilation was neutral when applying rescaled observations. Over this agriculturally dominated area, evaluation with soil moisture and LAI generally indicated the highest degradation. Difficulties in the Demer catchment might be related to crop rotation practices typical for the region that causes an interannual variability in backscatter dynamics not well accounted for by a static set of WCM parameters.</p>