An Empirical Orthogonal Function Iteration Approach for Obtaining Homogeneous Radiative Fluxes from Satellite Observations

Abstract

Conventional observations of climate parameters are sparse in space and/or time, and the representativeness of such information needs to be optimized. Observations from satellites provide improved spatial coverage over point observations; however, they pose new challenges for obtaining homogeneous coverage. Surface radiative fluxes, the forcing functions of the hydrologic cycle and biogeophysical processes, are now becoming available from global-scale satellite observations. They are derived from independent satellite platforms and sensors that differ in temporal and spatial resolution and in the size of the footprint from which information is derived. Data gaps, degraded spatial resolution near boundaries of geostationary satellites, and different viewing geometries in areas of satellite overlap could result in biased estimates of radiative fluxes. In this study will be discussed issues related to the sources of inhomogeneity in surface radiative fluxes as derived from satellites, development of an approach to obtain homogeneous datasets, and application of the method to the widely used International Satellite Cloud Climatology Project data that currently serve as a source of information for deriving estimates of surface and top-of-the-atmosphere radiative fluxes. Introduced is an empirical orthogonal function (EOF) iteration scheme for homogenizing the fluxes. The scheme is evaluated in several ways, including comparison of the inferred radiative fluxes with ground observations, both before and after the EOF approach is applied. On the average, the latter reduces the RMS error by about 2?3 W m?2.