A single observation of the microwave brightness temperature over sea ice cannot determine the composition of the ice surface in terms of the three constituents open water, first-year ice, and multiyear ice. The ambiguity can be removed by interpreting the time series of brightness temperatures at a fixed location in the context of a model of how the composition evolves in time. A model is discussed which includes the effects of ice growth, melting, deformation, and aging. The aging process accounts for the change of first-year ice into multiyear ice at the beginning of the cold season. The model reproduces the annual cycle of brightness temperature. A procedure, based on Kalman filtering, is presented for estimating the fractional areas of the three surface types. The procedure combines the model of ice evolution with the time series of brightness temperatures. The estimation errors are in the range 2--12%. These methods can be used with existing passive microwave data to construct a climatology of sea ice based on the three-type classification. It is also possible to use the observations to improve model parameters such as the rate of melting and the rate of creation of open water by mechanical processes. ¿ American Geophysical Union 1988