Testing the MODIS satellite retrieval of aerosol fine‐mode fraction

Satellite retrievals of the fine‐mode fraction (FMF) of midvisible aerosol optical depth, τ, are potentially valuable for constraining chemical transport models and for assessing the global distribution of anthropogenic aerosols. Here we compare satellite retrievals of FMF from the Moderate Resolution Imaging Spectroradiometer (MODIS) to suborbital data on the submicrometer fraction (SMF) of τ. SMF is a closely related parameter that is directly measurable by in situ techniques. The primary suborbital method uses in situ profiling of SMF combined with airborne Sun photometry both to validate the in situ estimate of ambient extinction and to take into account the aerosol above the highest flight level. This method is independent of the satellite retrieval and has well‐known accuracy but entails considerable logistical and technical difficulties. An alternate method uses Sun photometer measurements near the surface and an empirical relation between SMF and the Ångström exponent, å, a measure of the wavelength dependence of optical depth or extinction. Eleven primary and fifteen alternate comparisons are examined involving varying mixtures of dust, sea salt, and pollution in the vicinity of Korea and Japan. MODIS ocean retrievals of FMF are shown to be systematically higher than suborbital estimates of SMF by about 0.2. The most significant cause of this discrepancy involves the relationship between å and fine‐mode partitioning; in situ measurements indicate a systematically different relationship from what is assumed in the satellite retrievals. Based on these findings, we recommend: (1) satellite programs should concentrate on retrieving and validating å since an excellent validation program is in place for doing this, and (2) suborbital measurements should be used to derive relationships between å and fine‐mode partitioning to allow interpretation of the satellite data in terms of fine‐mode aerosol optical depth.