A Simulation Study of Satellite Emission Computed Tomography

Abstract

Satellite emission computed tomography retrieves the temperature of the atmosphere from radiances observed at multiple viewing angles and frequencies. To the extent that it provides independent information, the use of multiple viewing angles should improve the accuracy of the retrieval. Additionally, the tomographic retrievals should be more horizontally consistent since the fields of view overlap. The present study assesses these capabilities by performing a series of simulation experiments in which two-dimensional temperature fields (X?Z plane) are retrieved. Several limitations cited in previous work (by H. Fleming) are addressed by realistically treating the geometry of the sensor instantaneous field of view and by using appropriate instrumental noise levels. We have used observed atmospheric cross sections and the sensor geometry and simulation codes appropriate for the HIRS2 sensor. It is found that the tomographic approach is superior to the single angle approach in the cases studied when observational noise is 1.5 brightness temperature degrees (K) in each channel. For smaller noise levels (0.75 K) the two approaches are found to be comparable.