Spatio-temporal prediction of level 3 data for NASA's earth observing system
The first Earth Observing System (EOS) satellites was launched in 1998 and generated massive amounts of atmospheric data in both space and time. We explore the statistical issues relating to (optimal) processing of the resulting Level 2 data. We consider an approach to constructing Level 3 data products from Level 2 data that uses the spatio-temporal dependence of the data. We discuss the impact of global-grid-system choice and of spatial resolution on the error characteristics of the Level 3 data product. Data from the Total Ozone Mapping Spectromoter (TOMS) will be used for illustration.