The Met. Office's Large eddy Simulation Cloud Resolving Model was run in two dimensions. The initial conditions were based on the temperature and humidity profiles taken by the aircraft.
In the first case study the initial horizontal velocity had a shear of 5m/s/km. A random temperature perturbation between ±0.1K was added in the lowest model layer, which was sufficient to kick off the convective instability in the cloud. Radar reflectivities were calculated from the model ice and rain fields to compare to the observations from the Chilbolton radar. Sensitivity tests were carried out, switching off the Hallett-Mossop (HM) process and varying the number of primary ice nuclei by a factor of ten.
Slanting updraughts appeared in the model between 2km and 4km altitude due to conditional instability in the cloud layer. Above the melting level, liquid water only occurs in the updraughts. The slope and width of the updraughts is similar to embedded features shown in the radar observations, and the maximum radar reflectivity is 33dBZ, compared to about 30dBZ in the observations. The maximum vertical velocity was 1.4m/s, compared to 1.5 m/s in the observations. Realistic values for the supercooled liquid water content were predicted in these updraught regions. The maximum ice number concentration was 65/l, as opposed to 100/l in the observations. Switching off the HM process reduces the maximum ice number concentration by a factor of 20. The number of primary ice nuclei is very uncertain and so a sensitivity test was conducted. Decreasing the number of primary ice nuclei by a factor of 10 increases the maximum ice number concentration to 150/l: due to an increase in the efficiency of HM. The same dynamical framework was used to run a detailed microphysical. This model included explicit treatment of the formation and growth of cloud droplets, the formation of ice crystals by primary nucleation, the growth of ice particles by vapour growth, riming, and aggregation and HM. This model was able to reproduce the observed ZDR well with the low values in the main updraft. The higher values of ZDR outside the main updraft core were well reproduced by the model due to columnar crystals produced by HM.
In some other case studies it was found that vertical windsheer produced Kelvin –Helmholz billows that were responsible for the considerable spatial inhomogeneity of the cloud. The updrafts produced by the billows resulted in supercooled liquid water extending up to –10C . The cloud structure predicted by the model produced a predicted radar reflectivity pattern very similar to that observed The cloud microphysics is found to strongly affect the distribution of ice and liquid water in the cloud. This in turn determines the pattern of latent heat release, which is very important in the cloud dynamics in all the cases studied