Assimilation of Ice Thickness Information into a Sea Ice Model

The assimilation of ice-thickness, ice-concentration, and open-water-fraction observations into an ice thickness model offers a number of opportunities and challenges. The opportunities include the chance to correct model calculations with observations of the actual ice cover so that the model state, not only in regards to ice thickness but in other parameters as well, remains closer to the true state. The observations might include just the ice concentration, such as from passive microwave or higher resolution visible and thermal sensors; the mean thickness, such as from radar altimeters; the thickness distribution of thin ice, such as from AVHRR or aircraft measurements; or the full thickness distribution from upward looking sonars mounted on buoys or submarines. All assimilation schemes must use estimates of both the measurement errors as well as the model errors. With limited validation data available, the model error characteristics are difficult to assess and, yet assumptions about the model errors have profound impact on the results of assimilation procedures. This paper will discuss various data assimilation issues and techniques and show results of assimilation procedures on the evolution of the ice thickness distribution in the vicinity of the SHEBA ice camp.