Abstract
The Sea of Okhotsk is well known as one of the most southern sea ice zones in the Northern Hemisphere where all the sea ice melt in summer season. It is estimated that if the current trend of the global warming continues in the future, the sea ice of the Sea of Okhotsk may not be able to survive even in winter season by the middle of 21 st Century. In order to monitor the interannual validation of sea ice extent in this area, two sea ice concentration algorithms, namely the Comiso algorithm and the NASA Team algorithm, were applied to DMSP SSM/I data for sea ice concentration calculation. The both algorithms were revised in 1994 to reduce weather effects resulting from atmospheric water vapor, cloud-liquid water and others. The effects increase the sea ice concentration over the open water. As the humidity of the Sea of Okhotsk is rather higher than most of the other sea ice zones especially in summer season, the weather effects appear more strongly in this area than the others. The interannual validation graph derived from SSM/I data used to have a small peak in each summer under the sea-ice free conditions. Those small peaks were much reduced with the new weather filters introduced to both algorithms. However, the "false sea ice area" were still observed over the open water in the sea ice concentration images derived with both algorithms. The images also suggested the importance of reducing land effects mainly caused by antenna side lobe and mixed-pixels of SSM/I data. In order to reduce the land effects, the authors have introduced a 3X3 filter here we call as land filter. This filter is a kind of conditioned minimum filter which reduces the land effect by using the land mask information stored in the NSIDC SSM⁄I data set. When at least one out of 3X3 pixels was "land", then the ice concentration of the center pixel will be replaced with the minimum value within the 3 X 3 pixels. The land filter was applied to the ice concentration calculated result of the both algorithms. Almost all the false sea ice area not only in summer but also in other seasons were rejected with this filter for the Comiso algorithm. This result proves the use of the land filter for land effect reduction and the use of the new weather filter of the Comiso algorithm for weather effect reduction for the Sea of Okhotsk. As the land effects occurs not only in summer but also in other seasons, the land filter is applicable to any seasons. On the other hand, the NASA Team algorithm with the land Filter did not show as much improvement as the Comiso algorithm. This result suggests that the new weather filter introduced to the NASA Team algorithm was not fitting so well to the Sea of Okhotsk.
