Remote sensing of absorbing aerosols and precipitable water vapor using MFRSR measurements

We present further development of our analysis algorithm for Multi-Filter Rotating Shadowband Radiometer (MFRSR) data. The new additions include techniques allowing us to retrieve spectral aerosol single scattering albedo (SSA) and column amount of precipitable water vapor (PWV). The SSA retrievals employ MFRSR measurements of both direct normal and diffuse horizontal irradiances. We present a sensitivity study indicating dependence of SSA retrievals on optical thickness and other aerosol parameters. Influence on the retrievals of a possible error in separation of the direct and diffuse irradiances by the instrument is discussed. The algorithm has been tested on a long-term dataset from the local MFRSR network at the DOE Atmospheric Radiation Measurement (ARM) Program site in Southern Great Plains (SGP). Our results are compared to AERONET's almucantar retrievals of SSA from CIMEL sun-photometer co-located with the MFRSR at the SGP Central Facility. A constrained variant of the algorithm (assuming zero nitrogen dioxide column values) is used for this comparison and to study the influence of the uncertainty associated with this atmospheric gas on the retrieved aerosol absorption properties. Precipitable water vapor column amounts are determined from the direct normal irradiances in the 940 nm MFRSR spectral channel. HITRAN 2004 spectral database has been used to model the water vapor absorption, while a range of other databases (HITRAN 1996, 2000, ESA) is used in the sensitivity study. The results of the PWV retrievals for SGP's MFRSR network are compared with correlative measurements by Microwave Radiometers (MWR), GPS stations, AERONET, and MODIS satellite product. In the latter case an interpolation technique has been used to determine spatial structure of water vapor field from the network data and to create a 2D dataset comparable with satellite data.