Sensitivity of derived light extinction coefficients to uncertainties in measured aerosol properties used in radiative closure calculations

The major goals of this study are (a) to understand how the microphysical and chemical properties of aerosol particles control light extinction, particularly as a function of aerosol type and (b) to assess the sensitivity of calculated light extinction to uncertainties in these aerosol properties. One technique for probing the relationships between aerosol physical, chemical and optical properties is so-called closure studies, whereby detailed information about the aerosol size distribution, chemical composition and water uptake is used to derive light extinction coefficients that can be compared with direct measurements. Here we use the closure technique to explore how uncertainties in measured aerosol properties propagate through the calculation of the light extinction coefficient and affect the final comparison. In addition to measurement uncertainties there are also uncertainties introduced due to different particle size resolution as well as time resolution of the various measurements needed to calculate closure. For example, chemistry data may be obtained with coarse size resolution on a daily basis, while size distributions and optical properties are often obtained with higher size resolution and more frequently. Preliminary sensitivity studies indicate that uncertainties in the measurement of the size-resolved number fraction of absorbing species, the size-resolved internal versus external mixing state of particle chemical composition, and the measurement of particle size itself are the largest contributors to the overall uncertainty in derived values of the extinction coefficient. We address the sensitivity of closure calculations to the above instrumental and temporal averaging uncertainties using several different datasets from studies conducted in a variety of locations. Results and uncertainties from several field programs were used to provide case studies encompassing an assortment of aerosol types, measurement techniques, and time scales. Analyses of results from the LACE98 field study indicate that the average uncertainties in derived values of the extinction coefficient resulting from propagating assumed uncertainties in the input measurements are about +/-10%. However, different assumptions regarding the size-resolved mixing state of the observed elemental carbon mass fractions resulted in a +/-50% change in derived extinction coefficients. For highly absorbing aerosol (e.g., ACE-Asia), uncertainties in extinction due to uncertainties in the imaginary part of the refractive index (RI) are explored. In general, the ability to achieve closure is significantly impacted by uncertainties in chemistry measurements and the assumptions required for calculating RI. Size distributions measured by optical particle counters (OPCs) also rely on knowledge of the RI, and therefore the chemical composition, of the aerosol. Thus, uncertainties in RI can affect derived values of the extinction coefficient when OPC size distributions are used in closure calculations. Estimated uncertainties in scattering and absorption associated with size measurements using various techniques are presented. The relative importance of uncertainties in aerosol size distribution, chemical composition, and mixing state, for different aerosol types for field measurements conducted in the U.S., Germany, Korea and India is examined.