How sensitive are clouds to biogenic aerosols? Insights from satellite observations and model simulations 

One of the largest uncertainties in estimating the anthropogenic radiative forcing is related to the impact of atmospheric aerosols on cloud properties. The estimates of radiative forcing due to changes in cloud properties vary significantly between different global climate models, highlighting the need for constraining this forcing by using observations. Currently, one of the least well-understood aerosol components is secondary organic aerosols, most of which are of natural origin, i.e., biogenic SOA (BSOA). Here, we aim to quantify the effects of BSOA on cloud properties by combining field observations of aerosol concentrations and satellite observations of cloud properties. The aerosol particles were measured with a Scanning-Mobility Particle Sizer (SMPS) in Hyytiälä, Finland for the summers 2012-2023. Particles with diameter larger than 100 nm (N100) were considered as a proxy for cloud condensation nuclei. Cloud microphysical properties were obtained from the MODIS Collection 6 Level-2 cloud product (MYD06_L2) at 1 km resolution and averaged over a 1° × 1° region surrounding the site. Based on sensitivity tests and previous studies, the cloud droplet number concentration (CDNC) was derived for only low-level liquid warm clouds over the study region based on the retrieved cloud effective radius (CER) and cloud optical thickness (COT) values. The impact of different linear regression methods, measurement uncertainties, and sampling criteria on cloud susceptibility estimates was thoroughly analyzed. Based on meteorological conditions and aerosol size distributions observed in Hyytiälä, cloud parcel model simulations were performed for comparison. The simulated CDNC–aerosol susceptibility was found to be 0.68, while the observed value was somewhat smaller, 0.37. Using the observed temperature dependence of N100 and CDNC-aerosol susceptibility, we estimated temperature-driven cloud albedo feedback to be −0.68 W m⁻² °C⁻¹ (95 % confidence interval: −0.87 to −0.50 W m⁻² °C⁻¹). The magnitude of this feedback is approximately twice as large as reported in an earlier study which utilized MODIS Level-3 data. This difference highlights the strong sensitivity of estimated cloud susceptibility to satellite data sampling and filtering choices.