First Retrievals of Cloud Condensation 1 Nuclei (CCN) Proxies from PACE/SPEXone: Validation and Global Distribution

Cloud condensation nuclei (CCN) are critical for understanding aerosol-cloud interactions (ACI), yet the scarcity of direct CCN measurements limits global characterization. This study presents the first CCN retrievals from the SPEXone multi-angle polarimeter (MAP) aboard the PACE satellite. We utilize a neural network (NN)-based algorithm (CCN-NN) to retrieve column-integrated and height-dependent CCN proxies. The algorithm is trained using synthetic SPEXone measurements generated from a three-mode aerosol representation including fine, coarse soluble, and coarse insoluble (dust) modes. The primary CCN proxy, defined as the column number of soluble aerosol particles exceeding an activation threshold radius (NCCN), is retrieved for multiple activation threshold radii, both for ambient and dry aerosols. The algorithm also retrieves aerosol layer height (ALH), boundary layer fraction of total column aerosols fBL, and boundary layer contribution of NCCN (NCCNBL). The algorithm is applied to one year of PACE SPEXone observations. Ambient and dry NCCN retrievals are evaluated against collocated AERONET observations, while ALH and fBL are assessed using EarthCARE ATLID vertical extinction profiles. Ambient NCCN retrievals show good agreement with AERONET. About 81% of ocean and 68% of land retrievals fall within ±0.2 absolute error in log10 space (FAE(ϵ = 0.2)). For dry NCCN, FAE(ϵ = 0.2) is approximately 70% and 56% over the ocean and land, respectively, when the threshold radius is 0.075 μm. About 84% ocean and 82% land ALH retrievals satisfy the GCOS accuracy requirement. fBL shows large retrieval uncertainties. The validation of NCCNBL retrievals remains challenging due to limited validation data. Compared with the physics-based RemoTAP algorithm, the CCN-NN algorithm shows improved CCN proxy retrieval accuracy. The algorithm is further applied globally to generate annual mean distributions of NCCN, NCCNBL, and ALH. The global distributions show high NCCN and NCCNBL values over key anthropogenic and biomass-burning aerosol emission regions. Higher NCCN values are also observed over the Southern Ocean, particularly during austral spring and summer.