Retrieval of Dynamic Changes of Surface Water Extent from Sparse GNSS-R Measurements Using a Model-Driven Approach.

While CYGNSS exhibits a high revisit rate, the sparse, quasi-random tracks make it challenging to utilize data at short timescales for inundation mapping. In order to make use of the high revisit rate of CYGNSS, we implement a model-driven approach in which we compare simulated CYGNSS measurements over a binary water mask with actual CYGNSS measurements. We consider a simple water body like a reservoir that has limited or no vegetation and exhibits dynamic changes in surface water extent within the CYGNSS mission timeframe. We find that our method has the potential to reliably detect changes in surface water extent down to 4 sq. km., corresponding to a 3% increase. Using simple, non-physics-based flood model outputs as inputs to the End-to-End Simulator (E2ES) can lead to the ability to estimate surface water extent over a given area using sparse CYGNSS measurements on sub-weekly timescales.