Comparing Uncertainty in UTLS Aerosol Direct Radiative Forcing Due to Radiative Transfer Solvers Errors and Variability in Aerosol Optical Properties

The increasing observational evidence of variability in the composition and morphology of Upper Troposphere and Lower Stratosphere (UTLS) aerosols necessitates efficient and reliable methods for assessing their radiative impacts. While computationally expensive numerical radiative transfer solvers offer precise calculations, approximate methods provide a valuable alternative, especially when accounting for variations in aerosol size distribution, composition, and atmospheric column properties. This study quantitatively compares four such methods for calculating the aerosol direct radiative effect of UTLS aerosols: the optically thin layer approximation, the two-stream approximation (as implemented by RRTMG), the successive order of scattering (SOS) approximation, and the radiative kernel method. These methods differ primarily in their handling of multiple scattering and angular representation of scattered radiation.

Each method is evaluated for computational cost, flexibility, and fidelity by applying it to detailed airborne in situ observations of stratospheric aerosols from recent volcanic and wildfire events. Aerosol optical properties are represented through both direct calculations (based on size, morphology, and composition) and a parametric approach (using bulk extinction, scattering phase, and single-scattering albedo). We present the uncertainties associated with each method, attributable to aerosol properties, atmospheric column representation (including surface properties), and the radiative transfer solver. Based on these uncertainties, we discuss the implications of using these approximate methods for comparing satellite observations of aerosols with global chemistry-climate models.