Measuring accurately Sea Surface temperature is important for many marine applications and monitoring the global climate system. Many instruments are used for the measuring the SST. The SST delivered from satellite have the advantages that are a broad scope and consistent detection. But SST products show the different value because of different of retrieval algorithm and sensor. To reduce the uncertain, SST data ensemble is carried out using the Bayesian model averaging(BMA). BMA is method of the weighted average using the posterior probability distribution. And the means and variances of the posterior probabilities are estimated using Expectation-Maximization(EM) algorithm. The estimated mean of the posterior probability is used as the weight for the weighted average. SST data of Aqua/MODIS, Terra/MODIS and NOAA/AVHRR was used as ensemble member. SST data Envisat/AATSR was used as reference data for calculating the posterior probability and validation data. To make the monthly ensemble SST, their provided monthly SST data was used. one-leave-out-cross validation that is one of the statistical validation method is used for validating the ensemble SST. The 12 cases, except for the data of one month per the case, was made and excepted month was used validation period. And we compared with the ensemble mean and median. As the result, ensemble BMA showed the lowest RMSE.
Despite the same purpose, each satellite product has different value because of its inescapable uncertainty. Also the satellite products have been calculated for a long time, and the kinds of the products are various and enormous. So the efforts for reducing the uncertainty and dealing with enormous data will be necessary. In this paper, we create an ensemble Sea Surface Temperature (SST) using MODIS Aqua, MODIS Terra and COMS (Communication Ocean and Meteorological Satellite). We used Bayesian Model Averaging (BMA) as ensemble method. The principle of the BMA is synthesizing the conditional probability density function (PDF) using posterior probability as weight. The posterior probability is estimated using EM algorithm. The BMA PDF is obtained by weighted average. As the result, the ensemble SST showed the lowest RMSE and MAE, which proves the applicability of BMA for satellite data ensemble. As future work, parallel processing techniques using Hadoop framework will be adopted for more efficient computation of very big satellite data.