The Multi-Dimensional Challenge of Validating Remote-Sensing Aerosol-Type Retrievals

Abstract:

In addition to aerosol optical depth (AOD), aerosol type is required globally for climate-forcing calculations, constraining aerosol transport models, and other applications. However, validating satellite aerosol type retrievals is much more challenging than testing AOD results, because aerosol type is a more complex quantity, and ground-truth data are far less numerous and generally not as robust. We employ a combination of assessment relative to climatological expectation, statistical comparisons with surface-based observations, and near-coincident field campaign measurements, to evaluate MISR aerosol-type retrieval results. Although the retrievals are not constrained by a priori expectations, there is general regional coherence in the dominant retrieved aerosol types, indicating consistency in the retrieval process. Comparisons with expectation, on a regional, seasonal basis, demonstrate qualitative consistency with regard to particle size (three-to-five bins), shape (spherical vs. non-spherical), and single-scattering albedo (SSA; two-to-four bins) when mid-visible AOD exceeds about 0.15 or 0.2. Statistical comparisons with surface-based sun and sky-scanning photometer retrievals provide both qualitative and quantitative illustration of retrieval sensitivity, identifying strengths and limitations of the MISR Standard Version 22 aerosol product, and pointing to specific areas where improvements could be made. Field campaign results offer the most detailed and robust aerosol-type constraints. They allow us, with the help of the MISR Research aerosol retrieval algorithm, to test the limits of the MISR data information content, which in specific cases substantially exceeds the general sensitivity. This presentation will briefly review the statistical techniques employed and summarize the key MISR aerosol-type retrieval validation results of this work.