Atmospheric photolysis rate coefficients during the Polar Sunrise Experiment ALERT2000

Abstract

We measured and modeled UV radiation by three methods to quantify molecular photolysis rate coefficients during the light intensive period of the Polar Sunrise Experiment ALERT2000. UV radiation was measured by either a 2 π filter radiometer or a spectral irradiance sensor, or modeled with the phodis package. The phodis program uses a discrete ordinate radiation transfer method (DISORT) to calculate the actinic radiation field for a model atmosphere. The irradiance measurements must be converted to actinic fluxes in order to calculate photolysis rate coefficients. The conversion between irradiance and actinic flux is non-trivial and requires an assumption about the angular distribution of the diffuse radiation. We contrast two model assumptions. The first model assumes that the diffuse radiation is isotropic; we call this model the isotropic diffuse model (IDM). The second model calculates the angular distribution with DISORT, which we call the anisotropic diffuse model (ADM). Having calculated the actinic flux, we calculate the photolysis rate coefficient of NO 2, J(NO 2). Comparing the IDM and ADM-derived J(NO 2) due to downwelling radiation, we find that the IDM assumption leads to clear-sky photolysis rate coefficients 15% less than the ADM assumption. We show that the IDM method clear-sky results are not consistent with IDM method overcast sky results; therefore, the ADM assumption is preferred. The 2 π filter radiometer directly measures the actinic flux and uses filters appropriate to quantify J(NO 2). The comparison of these three methods for the downwelling component of J(NO 2) demonstrates good agreement; all three methods lie within a 20% range. The 2 π sensor measures 10% higher photolysis rate coefficients than does the ADM method, and the phodis model predicts 10% lower photolysis rate coefficients than does the ADM model. While the correlations between the ADM method and the 2 π sensor are quite compact during homogeneous sky conditions, partially cloudy sky conditions are difficult to model, leading to increased error. The overall standard deviation between the two measurements for all sky conditions is ±10% of the mean (1− σ). Using this full ADM model, we calculate photolysis rate coefficients for many other molecules of atmospheric interest.