A new parameterization for simulating global ice-nucleating particle concentrations based on long-term measurements with a network of expansion chambers

The most important milestone in a mixed-phase cloud’s life is the initiation of ice. If we cannot sufficiently capture this process in weather and climate models then it is unlikely that the cloud’s properties and evolution will be accurately represented. Here we focus on improving the representation of ice-nucleating particles (INPs), which are the fundamental link between aerosols and primary ice production in mixed-phase clouds.

We use a newly collated dataset of 20 campaigns from across the northern hemisphere using Portable Ice Nucleation Experiment (PINE) instruments together with nudged simulations of the UK Earth System Model, which includes a modal aerosol microphysics model. We use 30,000 collocated PINE measurements and output from the UK Earth System Model to derive a new parameterization that links the full dust size distribution to an INP concentration. We base the functional shape of the parameterization assuming the presence of a mineral component and a biogenic ice-nucleating component, which is consistent with recent understanding.

The new parameterization reproduces 80% of the 30,000 PINE measurements within a factor of 2 and 96% within a factor of 10, and UKESM simulations correctly represent many of the short term synoptic events seen in the PINE time series. The new parameterization also performs considerably better than alternative parameterizations. The analysis shows that the INP concentrations are correlated with the dust surface area but cannot be explained by the mineral component (K-feldspar) alone. The new parameterization is consistent with the spread of laboratory derived activity for mineral soils that contain biogenic material, suggesting a biogenic ice-nucleating component associated with dust is prevalent throughout the northern hemisphere.