Reconstruction of sea-surface temperature using AMSR data

This paper presents a model of the Data Interpolating Empirical Orthogonal Functions technique (DINEOF) which allow to reconstruct the sea-surface temperature using AMSR-E data. Simulating results of the model by Monte-Carlo, we analyze the error of reconstruct result, and then choose the reasonable EOF modes and the number of iteration. Considering the correlation of the spatial position of sea-surface temperature (SST), we complete the reconstruction of SST for the region of West-North Pacific Ocean using different modes and different iterative counts, and an example is given correspondingly. This reconstructing method is useful for the fusion of the multi-instrument or multi-time-state data, and which is important for the research and application of the oceanography.