WCD-22. Exploring low cloud and aerosol interactions in geostationary satellite observations and CESM: causal relationships and timescales

Abstract
The impact of atmospheric aerosol perturbations on cloud albedo is the most uncertain anthropogenic forcing of the climate system, primarily due to the unknown magnitude of cloud water and cloud fraction (CF) adjustments in response to these perturbations. Such adjustments are commonly inferred from linear regression analysis between cloud water path (CWP) and cloud droplet number concentration (Nd), and between CF and Nd. However, it is important to recognize that correlation alone does not imply causation. Quantifying the causal relationships between CWP and Nd, as well as CF and Nd, and their respective timescales remains challenging given the intricate interplay between meteorological conditions, aerosol and cloud properties, and other co-variabilities in natural systems. This leads to large uncertainty in the magnitude and sign of CWP and CF adjustments in response to aerosol-induced perturbations, making it difficult to evaluate and identify deficient physical processes in climate models. In this study, we perform wavelet coherence analysis on time series of CWP and Nd obtained from the Spinning Enhanced Visible and InfraRed Imager (SEVIRI) onboard geostationary satellite Meteosat-11. Our analysis is applied to 2ºx2º scenes over the Eastern North Atlantic region with low horizontal wind. With this analysis, we identify time periods where CWP and Nd exhibit significant coherence at time frequencies ranging from 30 minutes to 8 hours. Within the selected time periods, we examine the phase of the wavelet cross-spectrum between the CWP and Nd time series to determine the relative lag between the two variables. Our analysis enables us to classify scenarios into four distinct phase lag categories at different time frequencies: CWP positively/negatively leading Nd; Nd positively/negatively leading CWP. The four scenarios represent four distinct physical causal relationships between CWP and Nd. We quantify the magnitude and trend of aerosol-induced CWP adjustments in each scenario and their respective timescales. A comparable analysis is conducted for CF and Nd. We further extend our analysis to the Community Earth System Model (CESM) to efficiently assess the representation of aerosol cloud interactions in the model.