On the Use of SSM/I Measurements in Coastal Regions

Abstract

This study explores the utilization of Special Sensor Microwave/Imager (SSM/I) data in coastal regions where the measured signal consists of radiation received from both land and water surfaces. The problem of mixed land/water measurements (?footprints?) is solved using a high-resolution land?sea mask to infer the fraction of water surface for each measurement. A method to combine high-resolution datasets with SSM/I data is described in section 2, and its error characteristics are investigated. It is then used to derive the fraction of water surface within each SSM/I footprint from a high-resolution land?sea mask. The navigation uncertainty of the SSM/I was identified to be the dominant error source in data fusion. The method was applied to a two-month dataset of F11?SSM/I data for the Baltic Sea region. Based on this dataset the navigation uncertainty of the F11?SSM/I was quantified to be less than 7 km (section 3). In section 4 an algorithm is presented that corrects the coastal SSM/I measurements in such a way that retrieval algorithms designed for homogeneous water surfaces become applicable. However, the correction results in increased noise, 1.0?2.5 K, depending on frequency and polarization. The capability of the correction algorithm is demonstrated by comparing SSM/I estimates of columnar water vapor content with collocated coastal radiosounding measurements, yielding a root-mean-square deviation of 2.53 kg m?2. The method is utilized to derive monthly mean fields of columnar water vapor content over the Baltic Sea. The retrieved fields are compared to results of a numerical weather prediction model. There is a positive bias of the model results of about 1.2 kg m?2 when compared with the SSM/I retrievals as well as with radiosoundings.