Observed aerosol characteristics to improve forward-modelled attenuated backscatter in urban areas

Numerical weather prediction (NWP) models often parameterise aerosols to reduce computational needs, while aiming to accurately capture their impact adequately. Increasingly, aerosols are monitored in-situ directly and/or indirectly (e.g. by automatic lidars and ceilometers, ALC). ALC measure the aerosol optical characteristic of attenuated backscatter. This can also be estimated using forward models that combine forecast aerosol and relative humidity to parameterise aerosol physical and optical characteristics. The aerFO is one such forward model, designed to use Met Office NWP model output and parameterisations from the MURK visibility scheme. Given the aerFO-MURK scheme link, assessing the aerFO output could help inform future MURK scheme developments. To identify which parameterised physical and optical aerosol characteristics in the scheme are the most critical in urban settings, aerFO is driven with different in-situ aerosol observations at a background site in central London. Estimated attenuated backscatter is assessed against ALC observations. It is shown that the original MURK scheme parameterisation underestimates the variance of both dry mean volume radius and total number concentration. Representing both the accumulation and coarse mode aerosols in the aerFO reduces the median bias error of estimated attenuated backscatter by 69.1%. Providing more realistic temporal (monthly to hourly) variability of relative mass for different species leads to little improvement, compared to using monthly climatological means. Numerical experiments show that having more realistic estimates of number concentration is more important than providing more accurate values of the dry mean volume radius for the accumulation mode. Hence, improving the parameterisations for number concentration should be a main focus for further development of the MURK scheme. To estimate aerosol attenuated backscatter, the aerFO requires an extinction to backscatter ratio (i.e. the lidar ratio). In addition to forward modelling, the lidar ratio can also be used with ALC attenuated backscatter to calculate aerosol properties estimated in aerosol forecasts. Here, a model is developed that estimates the ratio using in-situ observations of the number size distribution and speciated aerosol masses. The values of lidar ratio derived at the London background site (14–80 sr across selected common lidar wavelengths) compare well to the literature. However, the modelled lidar ratio is unexpectedly correlated to relative humidity. Further, a stronger dependence exists at shorter wavelengths (355 and 532 nm) compared to longer wavelengths (905 and 1064 nm), and is due to the critical relation of lidar wavelength to aerosol size.