Cokriging method for spatio-temporal assimilation of multi-scale satellite data

This research extends traditional cokriging method from spatial domain to spatio-temporal domain for assimilating spatial data sets with different temporal sampling frequency and spatial resolution (density). The main advantage of spatio-temporal cokriging lies in the fact that it takes into account the spatial covariance, temporal covariance and spatiotemporal covariance structures in the spatio-temporal data assimilation for better modeling. In comparison with the heuristic STARFM method, our spatio-temporal cokriging create much better assimilation results in terms of accuracy and reliability, because our method takes into account the spatial covariance, temporal covariance and spatio-temporal covariance structures in the spatio-temporal data assimilation. Our spatio-temporal cokriging technique has been successfully applied to assimilate daily MODIS NDVI images (250m) with 30 m spatial resolution Landsat ETM+ NDVI images over the Maumee River Basin.