A43B-0274
The resolution-dependence of satellite-based cloud retrievals: First results from ASTER and MODIS observations

Thursday, 17 December 2015
Poster Hall (Moscone South)
Frank Werner, Joint Center for Earth Systems Technology, Baltimore, MD, United States, Gala Wind, Science Systems and Applications, Inc., Lanham, MD, United States, Zhibo Zhang, University of Maryland Baltimore County, Baltimore, MD, United States, Steven E Platnick, NASA Goddard Space Flight Center, Greenbelt, MD, United States and Larry Di Girolamo, University of Illinois, Urbana, IL, United States
Abstract:
The spatial resolution dependence of retrieved optical and microphysical cloud properties of marine shallow convective water clouds is presented using data from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), as well as the Moderate-resolution Imaging Spectroradiometer (MODIS) onboard the scientific research satellite Terra. Both instruments are characterized by vastly different spatial resolutions of 15m (ASTER) and 1000m (MODIS), respectively.

Cloud optical thickness (τ) and effective droplet radius (reff) are derived by means of the Cross-platform HIgh resolution Multi-instrument AtmosphEric Retrieval Algorithms (CHIMAERA) system which yields MODIS-like cloud property retrievals via a shared-core architecture. The retrieval algorithm employs a standard bi-spectral retrieval scheme with two reflectances (ρ) in the visible to near-infrared spectral wavelength range (VNIR, 0.86μm) and shortwave infrared spectral wavelength range (SWIR, 2.1μm), respectively.

For an exemplary granule the high-resolution ρ sampled by the ASTER instrument are aggregated from 15m to an increasingly coarse spatial resolution between (30-1000m). Subsequently, retrieved τ and reff from aggregated ρ are compared to the mean of the high-resolution cloud properties within the aggregated pixels. The differences in retrieved τ and reff are related to the sub-pixel covariance of ρ in the VNIR and SWIR band, as well as the inhomogeneity index (i.e., the ratio of standard deviation to mean value of ρ in the VNIR). This analysis highlights the impact of sub-pixel inhomogeneity and plane-parallel assumptions in the cloud property retrieval.

CHIMAERA also allows for a comparison of ASTER and MODIS retrievals without introducing biases due to individual instrument algorithms. Retrieved τ and reff from the 1000m aggregated ρ sampled by ASTER are compared to the retrieved cloud properties provided by MODIS. The presented results highlight the different sensitivities of the ASTER and MODIS instruments and possible implications on data interpretation.