Synoptic Mapping of Convective Structure from Undersampled Satellite Observations

Abstract

Climate properties regulated by convection, such as water vapor, cloud cover, and related distributions, are undersampled in asynoptic data from an individual orbiting platform, which must therefore be restricted to time-mean distributions. A procedure is developed to identify small-scale undersampled variance in asynoptic data and reject it, leaving a more accurate representation of large-scale variance that describes the organization of climate properties. The procedure is validated against high-resolution distributions that have been constructed from six satellites simultaneously observing the earth. Observing the high-resolution distributions asynoptically is shown to result in sampling error at large scales that is as great as the large-scale signal present, limiting the usefulness of the raw asynoptic data to time-mean distributions. However, processing the asynoptic data to reject undersampled incoherent variability reduces the error variance to 10% or less, yielding a fairly accurate representation of large-scale coherent variability, which then can be mapped synoptically on periods as short as 2.0 days. Made possible then are studies of how cloud, water vapor, and related distributions are organized by unsteady elements of the general circulation, which cannot be studied in the raw asynoptic data.