Estimation of real-time brown carbon absorption: An observationally constrained Mie theory-based optimization method

Abstract

Methods to estimate absorption of brown carbon (BrC), a significant fraction of atmospheric absorption, rely on estimating the difference between total measured absorption at near-UV, and that of black carbon (BC). Extrapolation of absorption measured at near-IR wavelengths (assumed exerted by BC alone) use different assumptions of the wavelength dependence of absorption Ångström exponent (AAEBC). Here, we develop an improved method exploiting real-time multi-wavelength absorption and particle count measurements in a Mie based optimization framework, incorporating spectral observational constraints (measured absorption at 880 nm and AAE880-660). An optimization approach, using a Mie model with core-shell and core-gray shell mixing schemes, is used to derive BC size distribution parameters (absorbing core diameter and scattering shell thickness). Goodness of fit (Mie optimization model vs. measurement) was R = 0.77–0.94 (near-IR absorption) and within 4%–30% for BrC estimation. A sensitivity analysis of input parameters (BC geometric standard deviation and refractive index) bounded estimated BrC of 32%. Application to a polluted urban site (Delhi) and a regional background site (Darjeeling) estimated BrC absorption (% contribution) at 370 nm as 18–117 Mm−1 (15%–29%) and 2–12 Mm−1 (5%–21%), respectively. Estimated BrC absorption was larger at the regional background site (Darjeeling) but smaller at the polluted site (Delhi) when compared to constant AAE and two-component approaches. Method efficacy is reinforced through larger estimated BrC absorption at Delhi coinciding with agricultural stubble burning periods in North India. The developed method uses multi-wavelength absorption observational constraints to improve the robustness of BrC estimation.