Science Seminars

5/10/2017 – Gala Wind; A Radiance Simulator for Proofing of Planetary Passive Imager Retrieval Algorithms

Speaker: Gala Wind (NASA Goddard)
Date: Sunday, May 10, 2020
Time: 4:00 PM
Location: SPSC W120

Seminar Abstract:

Algorithm validation is a continuing issue for atmospheric remote sensing algorithms of every kind. Field campaigns and in-situ ground sites provide means of validation, but can be very limited in location and data amount or impossible altogether (for example when exploring planets other than our own). Existing and proposed remote sensing algorithms can benefit greatly from having a controlled sandbox environment that provides synthetic data that appears equivalent to real instrument-acquired measurements. The Multi-sensor Cloud and Aerosol Retrieval Simulator (MCARS) is such a sandbox. In this presentation we show the MCARS setup that is able to generate radiances equivalent to ones produced by the MODerate resolution Imaging Spectrora-diometer (MODIS) instrument in a standard operational MODIS radiance file format using at-mospheric column data provided by the GEOS-5 model. When we examined the performance of the operational MODIS dark-target aerosol product (MOD04DT) using synthetic granules of bi-omass burning in Brazil, we found that the low bias in MOD04DT aerosol optical depth retrievals when compared to AErosol RObotic NETwork (AERONET) in-situ measurements is due to MOD04DT generating its lookup tables for smoke using a constant value of relative humidity. It is a valid assumption as long as the algorithm does not use wavelengths longer than 1020nm. MOD04DT uses the 2.1um MODIS channel. The three AERONET sites, present the study area, are of the older type that does not have the 1.6um channel. Carbon aerosol single scattering albedo is a very strong function of column relative humidity at wavelengths longer than 1.02µm. We also used synthetic scenes of clouds and aerosols off the coast of Namibia to test the performance of the research-level MODIS above-cloud aerosol properties retrieval algorithm (MOD_ACAERO). The results of executing the MOD_ACAERO algorithm on synthetic data show that the algorithm performs well and achieves closure where the retrieval result matches well against given “truth”. We make some suggestions for quality filtering of data before using it in modeling applications and propose some additional study of the cloud bow scattering angle range in the aerosol model used to create the lookup tables for the MOD_ACAERO retrieval algorithm. While originally developed for Earth atmospheric remote sensing, MCARS could be useful for other planets as well.