The 21st November 2000 frontal system.

Damian Wilson

This report details radar and aircraft observations, and model predictions of a frontal system which crossed southern England. The leading parts of the front were characterized by very uniform precipitation - it would appear that no convection was occurring. Towards the rear of the system there was more convection and the system ended with a significant amount of supercooled mid level cloud. It is instructive to see whether these features can be picked out by the model. The uniform nature of the system makes it an ideal opportunity to try qualitative verification of the water contents produced by the model.

Observational Data

Radar data, from the Chilbolton radar, for this event can be found via the University of Reading Radar Group website. Aircraft data should follow soon. Note the extreme uniformity of the initial scans. There are no regions of enhanced differential reflectivity (ZDR), which suggests that there is little liquid water present. Later, a plume of high Z appears, which is associated with embedded convection, and some supercooled liquid water, although the animation suggests that this plume is dominated by material falling into it from above. The altostratus at the rear of the front can be observed to contain large amounts of supercooled liquid from the high ZDR values produced by the rapidly growing eccentric crystals. This region appears to have some indication of convective cells present, though restricted in vertical extent.

Model simulation

Click here for postscript version

• Model data has been obtained for the equivalent azimuthal cross section. It shows, as model data always does, a fairly uniform structure within the front. Note that the front is a little late in the model simulation, so should be compared with earlier radar scans. The only convection predicted by the convection scheme is high up, near cloud top. The two precipitation schemes tested, the 3B and 3C schemes, give different ice water contents. Hopefully this can be validated from the aircraft data. Both schemes predict very little liquid water.
• The reflectivities in the rain agree well with the observations, with a preference to the slightly lower values of the 3C scheme. The reflectivities in the ice also agree quite well, again, the lower values of the 3C scheme being slightly more preferred. Note that the observations within the main part of the front contain fairly uniform reflectivity above the melting layer for about 2 kilometres, before a sharp drop - there is some indication of this in the model. Contrast this with the rapid drop in reflectivity above the melting layer at the back of the front - not picked out in the models. This must be associated with intense low level growth, which the model fails to reproduce so well.
• Liquid water content is very dependent upon the strength of the local updraught (a stronger updraught gives more liquid) and the amount of ice falling into the liquid from above). The ice content is also linked to the local updraught, but it is also strongly linked to fall-out and advection (also differential advection and fall out together), so the history of the system is important as well.
• The model correctly suggests the supercooled liquid water at the rear of the front, although it is not picked up as a layer but more as a block. However, there is a drier layer between the liquid water and the 0 deg C isotherm, so the model is attempting to confine the liquid into a limited vertical extent, despite the lack of vertical resolution. The ice content here is much reduced compared to the ice content within the main block of the front, yet the radar sees similar reflectivities. This would be consistent with the larger ice particles grown at the relatively high temperatures and supersaturations.
• Analysis of the model vertical velocities suggests that this region will ice up fairly rapidly in the near future, and is not in an equilibrium state. The liquid will probably be replaced by new liquid cloud further back in the front, so retarding the westerly movement of the precipitation.
• >From the model vertical velocities from this case and the 30th March case I might hypothesize that the observed embedded convection may be linked to strong area-averaged vertical velocities at low levels (just above the melting layer) rather than at more intermediate altitudes. These may be capped by reduced vertical velocities. could we test this hypothesis using other data?