Surface Reflectance Modeling for Martian Cloud Optical Depth Retrievals from CRISM Multispectral Image Cubes
Abstract
In order to retrieve the ice optical depth of martian clouds using a radiative transfer model, one needs to have an independent spectral model of the surface reflectance. Presented here is a comparison of three methods used to model the surface reflectance and the effects on the retrieved ice and dust optical depth. In order to reduce the complexity of the radiative transfer modeling, a fixed set of aerosol optical constants and single aerosol size parameters were used. The retrieval uses DISORT subroutines in an atmosphere modeling program tuned for martian atmospheres to create model reflectance spectra which are then compared to CRISM data. The parameters of total optical depth of ice and dust and surface reflectance are methodically adjusted until the chi-squared between model and data is minimized. This is done over all spatial points to create maps of dust and ice optical depth which are compared to MGS-TES and MO-THEMIS results to assess the effectiveness of the particular surface model. Three types of surface models were tested: a spectrally gray surface, a two-element linear combination of spectral endmembers chosen from the data cube itself, and a two-element linear combination of endmembers recovered from the data cube using principle components analysis and target transformation.
- Publication:
-
AAS/Division for Planetary Sciences Meeting Abstracts #45
- Pub Date:
- October 2013
- Bibcode:
- 2013DPS....4531305K