Modelled radiative forcing of the direct aerosol effect with multi-observation evaluation

A high-resolution global aerosol model (Oslo CTM2) driven by meteorological data and allowing a comparison with a variety of aerosol observations is used to simulate radiative forcing (RF) of the direct aerosol effect. The model simulates all main aerosol components, including several secondary components such as nitrate and secondary organic carbon. The model reproduces the main chemical composition and size features observed during large aerosol campaigns. Although the chemical composition compares best with ground-based measurement over land for modelled sulphate, no systematic differences are found for other compounds. The modelled aerosol optical depth (AOD) is compared to remote sensed data from AERONET ground and MODIS and MISR satellite retrievals. To gain confidence in the aerosol modelling, we have tested its ability to reproduce daily variability in the aerosol content, and this is performing well in many regions; however, we also identified some locations where model improvements are needed. The annual mean regional pattern of AOD from the aerosol model is broadly similar to the AERONET and the satellite retrievals (mostly within 10–20%). We notice a significant improvement from MODIS Collection 4 to Collection 5 compared to AERONET data. Satellite derived estimates of aerosol radiative effect over ocean for clear sky conditions differs significantly on regional scales (almost up to a facCorrespondence to: G. Myhre ([email protected]) tor two), but also in the global mean. The Oslo CTM2 has an aerosol radiative effect close to the mean of the satellite derived estimates. We derive a radiative forcing (RF) of the direct aerosol effect of −0.35 Wm−2 in our base case. Implementation of a simple approach to consider internal black carbon (BC) mixture results in a total RF of −0.28 Wm−2. Our results highlight the importance of carbonaceous particles, producing stronger individual RF than considered in the recent IPCC estimate; however, net RF is less different. A significant RF from secondary organic aerosols (SOA) is estimated (close to −0.1 Wm−2). The SOA also contributes to a strong domination of secondary aerosol species for the aerosol composition over land. A combination of sensitivity simulations and model evaluation show that the RF is rather robust and unlikely to be much stronger than in our best estimate.