Cirrus formation regimes – Data driven identification and quantification of mineral dust effect

Abstract. The microphysical and radiative properties of cirrus clouds are strongly dependent on the ice nucleation mechanism and origin of the ice crystals. Due to sparse temporal coverage of satellite data and limited observations of ice nucleating particles (INPs) at cirrus levels it is notoriously hard to determine the origin of the ice and the nucleation mechanism of cirrus clouds in satellite observations. In this work we combine three years of satellite observations of cirrus clouds from the DARDAR-N_ice retrieval product with Lagrangian trajectories of reanalysis data of meteorological and aerosol variables calculated 24 h backward in time for each observed cirrus cloud. In a first step, we identify typical cirrus cloud formation regimes by clustering the Lagrangian trajectories and characterize observed microphysical properties for in situ and liquid origin cirrus clouds in midlatitudes and the tropics. On average, in situ cirrus clouds have smaller ice water content (IWC) and lower ice crystal number concentration (N_ice) and a strong negative temperature dependence of N_ice, while liquid origin cirrus have a larger IWC and higher N_ice and a strong positive temperature dependence of IWC. In a second step, we use MERRA2 reanalysis data to quantify the sensitivity of cirrus cloud microphysical properties to a change in the concentration of dust particles that may act as INPs. By identifying similar cirrus cloud formation pathways, we can condition on ice-origin, region, and meteorological dependencies, and quantify the impact of dust particles for different formation regimes. We find that at cloud top median N_ice decreases with increasing dust concentrations for liquid origin cirrus. Specifically, the sensitivities are between 5 % and 11 % per unit increase of dust concentration in logarithmic space in the tropics and between 12 % and 18 % in the mid-latitudes. The decrease in N_ice can be explained by increased heterogeneous ice nucleation in the mixed-phase regime, leading to fewer cloud droplets freezing homogeneously once the cloud enters the cirrus temperatures and glaciates. The resulting fewer, but larger ice crystals are more likely to sediment, leading to reduced IWC, as for example observed for liquid origin cirrus in the mid-latitudes. In contrast, for in situ cirrus in the tropics, we find an increase of N_ice median values of 21 % per unit increase of dust aerosol in logarithmic space. We assume that this is caused by heterogeneous nucleation of ice initiated by dust INPs in INP limited conditions with supersaturations between the heterogeneous and homogeneous freezing thresholds. Such conditions frequently occur at high altitudes, especially in tropical regions at temperatures below 200 K. Our results provide an observational line of evidence that the climate intervention method of seeding cirrus clouds with potent INPs may result in an undesired positive cloud radiative effect (CRE), i.e. a warming effect. Instead of producing fewer but larger ice crystals, which would lead to the desired negative CRE, we show that additional INPs can lead to an increase in N_ice, an effect called overseeding.