Total cloud top radiative cooling (left) saturates with liquid water path whereas the maximum cloud top radiative cooling (right) does not.

Left: Cloud-mean correlation is not evident. Right: Local correlation with activation fraction is strong.

We show that aerosol-induced invigoration cannot occur by additional freezing of condensate alone. Substantial and near-immediate unloading of condensate must occur simultaneously which is unrealistic. Our more realistic assumptions suggest that lofting of additional condensate in polluted storms will usually weaken, not strengthen, invigoration, despite additional freezing. If invigoration occurs, it will likely be through additional low-level condensation and a reduction in supersaturation.

We simulated a wintertime orographic storm over the Sierra Nevada Mountains with RAMS and found that allowing the model to vary ice habit in time and space resulted in 37% more accumulated ice precipitation than simulating the storm with a fixed ice habit. We found that the variability of various ice processes with different habits effectively maximized ice growth in the variable habit simulation when compared to any single-habit simulation, highlighting the importance of accounting for time and space variation of ice properties in these types of storms.”

We show that the concentration of Aitken mode particles can have a significant control on the Arctic mixed-phase cloud properties when the accumulation mode concentration is low. Low accumulation mode concentration is common in the remote Arctic.

Following on from the Igel et al. (2018) paper, we examined the sensitivity of sea breeze properties to environmental factors with a focus this time on sea breezes associated with cloud development. Many similar results were found but one notable contribution of this paper is the analysis of updraft speeds. Different controls are found for deep and shallow convective regimes as shown by the figure where the maximum updraft speed (w) is a proxy for the deep convective updrafts and other quantiles characterize the shallow convective updrafts.

I developed a bulk microphysics scheme from a bin microphysics scheme called the Arbitrary Moment Predictor (AMP). AMP is unique in that it can predict any set of two or three distribution moments. It is also unique in that simulations with AMP can be directly compared to simulations with the bin scheme upon which it is built for a truly apples-to-apples comparison of the two scheme types. In this study I find that AMP compares most favorably to a bin scheme for a different set of predicted moments for cloud droplet condensation, evaporation, and collision-coalescence. On average, predicting the 3rd and 4th moments minimizes the error for a double-moment AMP configuration. This is in contrast to a typical bulk scheme configuration that predicts the 0th and 3rd moments.

Bin and bulk microphysics schemes often result in starkly different simulations, but the difference aren’t necessarily tied to the fundamental structural nature of each scheme. Here we found that a bulk scheme produced a strongly splitting supercell whereas a bin scheme produced a weakly splitting storm. The degree of splitting could be flipped simply by modifying the large ice fall speeds. The study highlights the difficulties in comparing bulk and bin schemes.

In double-moment bulk schemes, an assumption must be made regarding the shape parameter in the gamma PDFs used to approximate all hydrometeor size distributions. Here we show that the choice of the rain shape parameter can have a large influence on the total accumulated precipitation. The results call for better treatment of size distibution shapes.

Sea breeze properties are known to be sensitive to many different meteorological and land surface properties, but previous studies have only been able to assess the importance of one to three of these factors at a time. In this study, we assess simultaneously the relative importance of eleven environmental properties for sea breeze properties and aerosol redistribution using a combination of models simulations and statistical emulation of the model output. The figures (Figs. 5 and 9 in the paper) show that wind speed and the sea-air temperature difference, perhaps unsurprisingly, are important for controlling these properties. Other important factors, such as the soil moisture content, are more unexpected. The results of the study highlight future avenues of research and identify ways in which models could be improved for better forecasts of sea breezes and air quality.

Sources of aerosol particles are scarce in the Arctic. Here we show that the free troposphere may be an underappreciated source of particles. Observations from the ASCOS campaign show that high concentrations of particles may often be in contact with the boundary layer top (left). Model simulations show that when present, these particles can easily be entrained into the Arctic boundary layer. This transport is aided by clouds that may be present. A unique aspect of Arctic stratocumulus clouds is that they often extend above the boundary layer top. Aerosol particles then may be activated to cloud droplets above the boundary layer top, transported by cloud motions to cloud base, sediment out of the cloud and be regenerated when the drops evaporate below cloud base. The simulations systematically varied the boundary layer aerosol concentration (AER) and above boundary layer aerosol concentration gradient (GR).

A portion of Fig. 5. A comparison of condensation (left column) and evaporation (right column) rates from bin and bulk schemes. Each colored line represents a distribution of the log(rate ratios) for a different pair of simulations. Raw comparisons are shown in the top row. It can be seen that the two schemes often do not agree. When differences in rates due to distribution shape are accounted for, the agreement becomes markedly better (bottom row). This result points to a need for better assumptions regarding the assumed PDF shape in bulk microphysics schemes.

Some bulk microphysics schemes use the cloud droplet number concentration to diagnose the shape parameter of the gamma PDF (dashed lines). The relationship between these two quantities is not unique, and more research needs to be done in order to better diagnose the shape parameter. This paper suggests that diagnosing this parameter based on both the droplet concentration and the aerosol concentration would lead to improvements. Figure 7a from the paper is shown above. Different colors show results from simulations with different aerosol concentration.

Uncertainty about the cloud droplet shape parameter can have big impacts on the properties of simulated clouds. Figure 2 from the paper (above) shows that the cloud fraction is impacted more by the choice of shape parameter than by the aerosol concentration in simulations of shallow cumulus clouds. The top row shows liquid water path, and the bottom row shows vertically integrated droplet concentration. Focus though on the percentage of the domain covered in colors to see the change in cloud fraction.

Panels from Figures 8, 11, and 12 showing the role of latent heating in frontogenesis, frontal stability, and the frontal slope. The values in the first two panels are combined mathematically to arrive at the values in the third panel. The results indicate that one impact of latent heating is to weaken the warm front in the sense that the slope is reduced and thus weaker vertical velocities may be expected.

Fig. 13. Summary schematic showing the changes to the microphysical properties of the mixed-phase portion of the warm-frontal cloud and the total precipitation. Vapor deposition onto pristine ice, snow and aggregates (PSA) increases (as indicated by the change in arrow thickness) while riming of the mixed-phase species (GH) decreases. These two trends in the growth of ice mass cancel one another resulting in little change in the total ice mass and little change in rain production through melting. Differences in size and number of hydrometeors between the clean and polluted scenarios indicate qualitative changes.