Mesoscale analyses and diagnostic parameters for deep convection nowcasting

Abstract We study the contribution of surface data to convection nowcasting over regions of modest orography and under weak synoptic forcing. Hourly mesoscale analyses are performed using the CANARI optimal interpolation analysis scheme, which combines first‐guess fields from the fine mesh (10 km) ALADIN model with hourly routine observations arising from a mesonet of automated ground stations. These analyses then allow the computation of diagnostic parameters that quantify convective instability, low‐level lifting processes and moisture supply: these are the convective available potential energy (CAPE) and the moisture convergence (MOCON). A tuning of the analysis scheme is needed first for it to fit the meso‐β‐scale. Then, the skill of the computed diagnostics for convection nowcasting is evaluated by comparing their fields with radar reflectivities observed between one and four hours after the analysis time. This is done for four selected convective situations. With regard to thunderstorm triggering, results show that this usually happens over areas of persistently high values of CAPE which undergo convergence continuously from four to one hour before the event; on the other hand, areas of persistent divergence are never associated with convective developments. In addition, the proposed criteria allow a significant reduction in the areal extent of predicted thunderstorms (i.e. decreasing the false‐alarm rate) compared with what can be currently done on an operational basis, while maintaining a low non‐detection rate. As to convection monitoring, we find that the organization of convective systems into a reflectivity line is preceded by a similar organization in the MOCON field from one to three hours ahead. Copyright © 2000 Royal Meteorological Society