Abstract
An integrated data set composed of multi-sensor data (VTIR, MSR, MESSR and AVHRR) from MOS-1 and NOAA satellite combined with the observation data of sea surface temperature (SST) by ocean data buoys was developed for the purpose of the study on SST retrieval from satellite data.
The most important aspects for the data set are accuracy and easiness in the comparison of satellite data with sea truth data. The data set was developed by creating sub-images in which the image centers always correspond to the locations of the four ocean data buoys fixed at the oceanic areas around Japan by Japan Meteorological Agency. From the continuons measurement of SST by the buoys, the sea truth of SST can be estimated accurately when a satellite passes over the buoys. As the result, the precise evaluations for the algorithms in SST retrieval from satellite data are expected to be achieved.
Total four kinds of data sets were developed and some of the data sets were preliminary evaluated in order to verify the basic characteristics of MOS-1 data in SST retrieval. A remarkable improvement of the accuracy in SST retrieval with the combination of VTIR and MSR was found from the preliminary evaluation.
