Statistical analysis of single-track instrument sampling in spaceborne aerosol remote sensing for climate research

We examine likely effects of pixel-wide along-track sampling on climatological means of aerosol optical thickness (AOT) derived from observations with satellite instruments such as CALIOP and APS by sub-sampling AOT retrievals from a wide-swath imaging instrument (MODIS). The advantage of using daily pixel-level aerosol retrievals from MODIS rather than aerosol transport models to assess the results of along-track sampling is that limitations caused by the presence of clouds are implicit in the satellite dataset, so that their seasonal and regional variations are captured coherently. However, imager data can exhibit latitudinal (cross-track) variability of monthly global mean AOTs caused by a scattering-angle dependence. This makes it difficult to separate natural variability from viewing-geometry artifacts complicating direct comparisons of an along-track sub-sample with the full imager data. To work around this problem, we introduce ‘‘latitudinal-track’’ sampling which, by design, captures the cross-track AOT variability of the original imager data. We show that the latitudinal-track standard error of global monthly mean AOTs is much smaller than the longitudinal-track one. This allows us to attribute the difference between the two errors to MODIS viewing-geometry artifacts and obtain an upper limit on AOT errors caused by along-track sampling. We conclude that single-track instruments provide AOT sampling sufficient for climate applications. Since AOT is believed to be the most variable characteristic of tropospheric aerosols, our results imply that pixel-wide along-track coverage also provides adequate statistical representation of the global distribution of aerosol microphysical parameters.