Aerosol Light Scattering Measurements at Cape Grim Tasmania during ACE-1

Mark J. Rood, Christian M. Carrico, Rajendra Shrestha, Mathew J. Skific, Univerisity of Illinois at Urbana-Champaign, John A. Ogren, NOAA-CMDL

As part of ACE-1, total-light scattering and backscattering coefficients (sigma-tsp and sigma-bsp) of ambient aerosol were measured at Cape Grim Baseline Air Pollution Station. Such measurements will provide important parameters for this unperturbed marine site for use in chemical/radiative transfer models that predict the direct radiative effect of aerosols on climate. The controlled relative humidity (RH) nephelometry system (humidograph) measured the dependence of sigma-tsp and sigma-bsp upon RH, particle size (dp), wavelength of light (lambda), and direction of light scattered.

The humidograph employed two TSI Model 3563 nephelometers, a teflon membrane humidification system, and impactors with 10 um and 1 um size cuts. Semi-continuous operation produced a database of over 500 RH scans of sigma-tsp and sigma-bsp. The optical parameters of interest are the increase in sigma-tsp and sigma-bsp due to hygroscopic growth (f(RH)), the backscatter ratio (b), the relative contributions of coarse (dp < 10 um) and fine particles (dp < 1um) to sigma-tsp and sigma-bsp, and the Angstrom exponent (Å).

Results for f(RH), b, Å, and fine to coarse light scattering ratios for the aerosol are presently under analysis. On average, sigma-tsp and sigma-bsp are 4.9e-6 +/- 2.2e-6 m-1 and 5.6e-7 +/- 2.4e-7 m-1 respectively at lambda = 550 nm and dp < 1 um. Low signal-to-noise ratios contribute to significant uncertainties especially in sigma-bsp during particularly clean time periods as sigma-tsp and sigma-bsp at Cape Grim are an order of magnitude lower than has been measured at an anthropogenically perturbed continental site (Bondville, IL). Preliminary results indicate mean f(RH) = 2.47 +/- 0.57 (dp < 1 um, total scatter) with 25% variation with lambda, f(RH) = 2.04 +/- 0.22 (dp < 10 um, total scatter) with 5% variation with lambda, and f(RH) = 1.6 +/- 0.2 (dp < 10 um, backscatter) with 11% variation with lambda. In comparison, this research group has measured f(RH) = 1.5 +/- 0.8 (dp < 1 um, total scatter) with 9% variation with lambda at Bondville, IL, while Charlson et al. (1992) estimated a globally averaged f(RH) = 1.7.