A thermodynamic model for estimating sea and lake ice thickness with optical satellite data

Abstract

Sea ice is a very important indicator and an effective modulator of regional and global climate change. Current remote sensing techniques provide an unprecedented opportunity to monitor the cryosphere routinely with relatively high spatial and temporal resolutions. In this paper, we introduce a thermodynamic model to estimate sea and lake ice thickness with optical (visible, near‐infrared, and infrared) satellite data. Comparisons of nighttime ice thickness retrievals to ice thickness measurements from upward looking submarine sonar show that this thermodynamic model is capable of retrieving ice thickness up to 2.8 m. The mean absolute error is 0.18 m for samples with a mean ice thickness of 1.62 m, i.e., an 11% mean absolute error. Comparisons with in situ Canadian stations and moored upward looking sonar measurements show similar results. Sensitivity studies indicate that the largest errors come from uncertainties in surface albedo and downward solar radiation flux estimates from satellite data, followed by uncertainties in snow depth and cloud fractional coverage. Due to the relatively large uncertainties in current satellite retrievals of surface albedo and surface downward shortwave radiation flux, the current model is not recommended for use with daytime data. For nighttime data, the model is capable of resolving regional and seasonal variations in ice thickness and is useful for climatological analysis.