GACP Projects
Evaluation of Uncertainties in Satellite Retrievals of Aerosol Using In-Situ Measurements at the Surface
John A. Ogren, PI
Michael H. Bergin, Co-I
David S. Covert, Co-I
Mark J. Rood, Co-I
Patrick J. Sheridan, Co-I
Abstract: Derivation of aerosol radiative forcing from satellite observations requires assumptions about the chemical, microphysical, and optical properties of the particles, because not all the necessary information can be determined from satellites. We propose to assemble the extensive data sets of in-situ, tropospheric aerosol properties (in particular, single-scattering albedo, hemispheric backscatter fraction, and Engstrvm exponent) that have been obtained by our team of investigators and others, and work with the NASA aerosol climatology processing facility to test the sensitivity of candidate satellite data retrieval algorithms to observed variations of aerosol properties. The desired outcome of our investigation is a quantitative estimate of the uncertainty in the satellite-derived aerosol climatologies attributable to assumptions about aerosol properties used in the retrieval algorithms.
Each satellite retrieval of aerosol properties uses different assumptions, either because of different primary observations or different algorithms. These assumptions comprise an "aerosol model", which specifies the unknown aerosol properties needed to calculate the retrieved quantities. In general, the aerosol model specifies the aerosol size distribution and refractive index. Some retrievals may use different aerosol models, depending on geographical region or on the satellite-based observations themselves. In all cases, the aerosol properties specified in the aerosol model can be compared with in-situ measurements of those properties. However, combined observations of aerosol size distribution and refractive index are rare, making it difficult to compare the models directly with measurements.
Our approach begins with a candidate aerosol model, which could be part of a currently operational retrieval algorithm, a planned operational algorithm, or a new algorithm proposed by other members of the NASA Aerosol Radiative Forcing Science Team. Guided by in-situ observations of aerosol size distribution and chemical composition, we will develop a set of perturbed versions of the candidate aerosol model that yield similar frequency distributions of aerosol properties, such as single-scattering albedo, as actually observed. These perturbed versions of the aerosol model can then be used by the NASA aerosol climatology processing facility to evaluate the distribution of retrieved aerosol properties that is consistent with estimated variations in the retrieval algorithm's aerosol model. The statistical nature of our approach requires large data sets that adequately represent the variability of atmospheric aerosols on time scales ranging from diurnal to multi-year. Surface-based measurements from NOAA's worldwide aerosol monitoring network, with sites spanning conditions from polluted continental, to clean marine, to the free troposphere, are one of the few sources of data appropriate for this approach.
