Correcting for biases in satellite retrieved cloud droplet effective radius and its impact on the aerosol first indirect effect
The cloud effective radius, Re, of the drop size distribution derived from passive satellite sensors is a key variable widely used in aerosol-cloud interactions studies and climate model validation. Re is commonly retrieved from passive satellite sensors such as the MODerate resolution Imaging Spectroradiometer (MODIS), based on a visible and near-infrared bi-spectral technique that had its foundation more than a quarter century ago. This technique makes a wide range of assumptions such as 1-D radiative transfer, single-mode droplet size distribution, and cloud horizontal and vertical homogeneity. It is well known that deviations from these assumptions lead to bias in the retrieved Re.
By co-registering multi-angle radiance measurements from the Multi-angle Imaging Spectroradiometer (MISR) with MODIS cloud optical properties at cloud top, we introduced a regional bias-correction approach for MODIS retrieved Re. Our results suggest that estimated monthly mean bias in MODIS Re exhibit large regional dependency, ranging from at least ~ 1 -10 μm (15 - 60%), depending on scene heterogeneity, optical depth, and solar zenith angle. The bias-corrected Re values compare favorably to other independent data sources, including past field observations, global model simulations, and other independent satellite retrieval techniques. The bias estimate changes the interpretation of global Re distributions from MODIS Re products and may further impact studies that use the original MODIS Re products to study, for example, aerosol cloud interactions and cloud microphysical parameterization. The performance of this bias-correction was further evaluated during NASA’s Cloud, Aerosol, Monsoon Processes-Philippines Experiment (CAMP2Ex) that took place in the Philippines from August to October 2019. Multiple research flights during CAMP2Ex were implemented to target the objective of providing the first detailed evaluation of Re retrieved across multiple platforms and techniques in cumulus and congestus cloud region. Our evaluation consists of cross-comparisons between bias-corrected MODIS Re with remote sensing and in-situ retrievals of warm oceanic clouds across all CAMP2Ex research flight segments. We found good agreement between bias-corrected Re, in-situ derived Re and airborne polarimetric Re (differences within 1 - 2 μm), whereas the original MODIS Re and airborne bi-spectral Re showed large overestimates (5 - 7 μm). Since this bias varies with the underlying structure of the cloud field, caution continues to be warranted in studies that use bi-spectral Re retrievals in cumulus cloud fields. We are currently applying the various Re retrieved from different remote sensing techniques to estimate the aerosol first indirect effect in cumulus fields from CAMP2Ex, with an emphasis on the impact of biases and retrieval errors in Re on the estimated strength of aerosol-cloud interactions.