Snow or rain? - A matter of wet-bulb temperature

Sammanfattning: Accurate precipitation-type forecasts are essential in many areas of our modern society andtherefore there is a need to develop proper working methods for this purpose. Focus of thiswork has been to study important physical processes decisive in deciding both the temperatureof the precipitation particles, hence affecting their phase, and the surrounding air. Two majorlatent heating effects have been emphasized, melting effect and cooling byevaporation/sublimation. Melting of the snow flakes subtracts heat from the surroundings andhence acts as a cooling agent. Phase transformation from solid/liquid into the gas phase alsoneeds heat which here results in a cooling tendency. These two mechanisms may sometimeshave a crucial influence for deciding the correct precipitation-type. The melting effect isdiscussed in a paper about a snow event in Tennessee in USA, and another paper describingan event in Japan showing the influence of the evaporation/sublimation process. In the lattercase the wet-bulb temperature, Tiw as a physical correct discriminator between snow and rainis obtained. A numerical weather prediction model (HIRLAM) is being used to study differentcondensation schemes during three weather situations occurring in Sweden. These areRasch/Kristjánsson condensation scheme, Sundqvist original condensation scheme and amodification of the latter scheme. In the modified Sundqvist condensation scheme the Tiw hasbeen implemented as a limit temperature between snow and rain. The results are showingdifferences between the two main schemes concerning the total precipitation (both snow andrain). Comparisons between Sundqvist condensation scheme and this modified version, calledSundqvist scheme with Tiw show that this latter version creates slightly more snow.Differences between them are largest in dryer areas. Differences in the snow accumulationincrease when the forecast length increases. That makes them harder to be compared to snowanalyses from MESAN (mesoscale analysis) because the analyses is partly based ondifferences in the snow depth and this cannot be directly compared to amount of newly fallensnow especially when surface air temperatures are above freezing. Deviations from the dataanalyses are obtained in both Sundqvist and Sundqvist scheme with Tiw but in some regionsthe latter is in better agreement with measurements. Further work is needed in precipitationtypestudies but the physical correct value with Tiw = 0 ºC as melting temperature used inSundqvist with Tiw scheme is an interesting project for the future in the field of precipitationtypeforecasting.