Title:
Exploring biases in model/measurement comparisons using surface in-situ data - Results from the AEROCOM INSITU Project
Abstract:
In 2015, AeroCom, an open international collaboration of scientists seeking to improve global aerosol models, initiated a project comparing model output to in-situ, surface-based measurements of aerosol optical properties. The model/measurement comparison project, called INSITU, aims to evaluate the performance of a suite of AeroCom aerosol models with surface in-situ aerosol optical data in order to inform iterative improvements to model aerosol modules. Quality-assured, in-situ aerosol absorption and scattering data from over 60 surface sites is being used to evaluate simulated aerosol optical properties from 14 global models. Previous results have suggested that, in general, the modelled aerosol is larger and darker than what is observed.
Current work is focused on assessing biases in the model/measurement comparisons as a function of aerosol type and season. For example, the mean normalized bias (MNB) for aerosol extinction appears to be positive in polar regions (i.e., models are higher than measurements), but negative for continental and marine sites. In contrast the MNB for aerosol absorption is negative for polar regional sites (i.e., models are lower than measurements) but positive for continental and marine sites. Marine site measurements are most highly correlated with model simulations of marine aerosol, while observations in continental Europe tend to be least correlated.
Finally, the impact of water on the model/measurement comparisons is beginning to be evaluated. The in-situ measurements are made at low RH (<40%) which the observational community refers to as ‘dry’, while the model simulations are actually dry (RH=0%). Preliminary results from a limited set of models suggest that, at RH=40%, aerosol water may contribute 20% or more of simulated aerosol scattering.