Predicting the impacts of cloud processing on aerosol properties.

摘要

Cloud processing of aerosol sulfate and optical properties has important effects on the Earth's radiation balance, but computational limitations have inhibited direct microphysical (explicit) simulations in global atmospheric models. Here we describe state-of-the-art parameterizations of explicit behavior using a series of simulation techniques.; A Lagrangian microphysical parcel model was modified to include an efficient multivariate solving algorithm. A wide factorial domain of initial conditions, representing moderately clean to moderately polluted environments, was used to generate explicit predictions for lower troposphere stratiform clouds. Regressions were developed predicting the initial (pre-cloud) dry particulate single scattering coefficient (bspi), initial dry total mass scattering efficiency (MSEi), and sulfate production and scattering changes after 500 and 3000 seconds of cloud processing, respectively. Predictor variables included bulk sulfate and lognormal size parameters for accumulation and coarse modes. Regression bias was assessed for systematically varied coarse modes, and a probabilistic sensitivity analysis was conducted.; The predicted MSEi is within 20% of explicit calculations for coarse mode distributions with small geometric mean radii (rc0.4um) and 40--80% for larger mean radii, and bspi is predicted within 20--40%. Predicted initial scattering was more sensitive to input uncertainties than the other parameters. The regressions were most sensitive to measurement uncertainties associated with the accumulation mode geometric mean radius and coarse mode number concentration. A validation exercise showed MSE enhancements of 1, 6 and 12% were predicted from the regression model compared to previously observed average enhancements of 3, 7, and 19%, respectively. The predicted versus observed particulate scattering coefficients agreed within measurement uncertainty. Application of the regression models to several published aerosol distributions predicted dry MSEi ranging from 1.0 to 1.7 m 2 g-1 and MSE enhancements ranging from 2--50% and 7--100% after 500 and 3000 seconds of cloud processing, respectively. Greater efficiency enhancements were predicted for low-aerosol-mass environments than for polluted regimes. The parameterizations are a new tool for climate modelers to incorporate into existing global models' treatment of aerosol properties.