Estimation of Uncertainties of Retrieved Aerosol Parameters in Version 3 of the AERONET Inversion Algorithm.

A new technique to estimate uncertainties of retrieved aerosol parameters was developed in Version 3 of the AERONET aerosol retrieval algorithm. It uses the variability in retrieved aerosol parameters as a proxy for retrievals uncertainty. The variability in retrieved aerosol parameters is generated by perturbing observations in inversion inputs assuming that the perturbations in measurements and other inputs are due to uncertainties (biases, systematic errors) in aerosol optical depth (AOD), radiometric calibration of sky radiances, spectral solar irradiance, and surface reflectance (BRDF). Each bias assumes three values: positive, negative, and zero (non-biased). The values and the signs of biases of the same input type are assumed spectrally independent. The radiometric calibration and solar spectrum irradiance uncertainties are combined in one bias because both of them affect the magnitude of sky radiances. Thus there are three distinct biases each assuming three values which makes 27 distinct combinations in perturbing the inversion inputs. For each set of Sun photometer observations 27 inputs corresponding to 27 combinations of biases are produced. All of these inputs are then inverted with the AERONET operational retrieval algorithm and the following statistics of the inversion results is generated: average value, standard deviation, minimal and maximal values. These statistics describe the variability in retrieved aerosol parameters due to uncertainty in inversion inputs for each individual inversion. From these statistics only the standard deviation (termed U27) is used as a proxy for the estimated uncertainty. This technique was applied to the AERONET almucantar retrieval database and statistics of estimated uncertainties of all inversion parameters were analyzed as a function of AOD for different aerosol types.