Validation of a sea ice model using forward simulation of ERS-1 SAR data: a case study in the Beaufort Sea

A sea ice dynamics/thermodynamics model is validated using a one-year time series of data from the C-band synthetic aperture radar (SAR) instrument aboard Earth Resources Satellite 1 (ERS-1). Common validation problems using geophysical parameter estimates produced by operational algorithms are avoided by coupling a simple backscatter model to the physical model to generate a time series of estimated SAR backscatter values, which can be directly compared to the observed time series. Comparison of actual and estimated SAR backscatter shows good agreement during the winter and spring seasons. Differences during the summer and early autumn period are attributed to surficial changes, namely wet snow, meltponds, and young ice covered by frost flowers, that are not included in the ice physics model but affect the C-band backscatter. After adding these effects to the backscatter model, very good agreement is observed throughout the year between the sea ice/backscatter model and the ERS-1 time series, with RMS differences about 10% of the full range of backscatter values. The paper demonstrates the utility of the forward simulation approach for model validation and discusses potential future improvements and applications.