Modelling low-frequency rainfall events using airflow indices, weather patterns and frontal frequencies

Low-frequency, high-magnitude daily rainfall amounts recorded at several si tes in central and southern England were related to the prevailing Lamb Wea ther Types (LWTs), daily weather front and airflow data. Three statisticall y distinct weather-type clusters were indentified and used to construct a s implified frontal model of daily precipitation occurrence/amount. The model was calibrated against station data for the period 1970-1990 and used to r econstruct observed daily precipitation between 1875 and 1969 given the his toric sequence of LWTs. Although the model reproduced the incidence of low- frequency, high-magnitude events, it failed to capture variations in mean w et day probabilities and wet/dry spell persistence. This inability was attr ibuted to the general limitations of the weather classification methodology , which did not capture all aspects of the precipitation regime with equal levels of proficiency. Therefore, the prospects for downscaling high-resolu tion precipitation series directly from indices of mean sea-lever pressure rather than via weather patterns was discussed. Preliminary results indicat e that relationships can be established between mean daily precipitation oc currence and airflow indices such as vorticity and strength of air how. How ever, further research is required to establish the value of such indices f or modelling low-frequency, high-magnitude precipitation events. (C) 1998 P ublished by Elsevier Science B.V. All rights reserved.