Numerical simulation of intense precipitation events south of the Alps: Sensitivity to initial conditions and horizontal resolution

Citation
C. Cacciamani et al., Numerical simulation of intense precipitation events south of the Alps: Sensitivity to initial conditions and horizontal resolution, METEOR ATM, 72(2-4), 2000, pp. 147-159
Citations number
29
Categorie Soggetti
Earth Sciences
Journal title
METEOROLOGY AND ATMOSPHERIC PHYSICS
ISSN journal
01777971 → ACNP
Volume
72
Issue
2-4
Year of publication
2000
Pages
147 - 159
Database
ISI
SICI code
0177-7971(2000)72:2-4<147:NSOIPE>2.0.ZU;2-F
Abstract
In this paper we describe the results of several numerical experiments perf ormed with the limited area model LAMBO, based on a 1989 version of the NCE P (National Center for Environmental Prediction) ETA model, operational at ARPA-SMR since 1993. The experiments have been designed to assess the impac t of different horizontal resolutions and initial conditions on the quality and detail of the forecast, especially as regards the precipitation field in the case of severe flood events. For initial conditions we developed a m esoscale data assimilation scheme, based on the nudging technique. The sche me makes use of upper air and surface meteorological observations to modify ECMWF (European Centre for Medium Range Weather Forecast) operational anal yses, used as first-guess fields, in order to better describe smaller scale s features, mainly in the lower troposphere. Three flood cases in the Alpin e and Mediterranean regions have been simulated with LAMBO, using a horizon tal grid spacing of 15 and 5 km and starting either from ECMWF initialised analysis or from the result of our mesoscale analysis procedure. The result s show that increasing the resolution generally improves the forecast, brin ging the precipitation peaks in the flooded areas close to the observed val ues without producing many spurious precipitation patterns. The use of meso scale analysis produces a more realistic representation of precipitation pa tterns giving a further improvement to the forecast of precipitation. Furth ermore, when simulations are started from mesoscale analysis, some model-si mulated thermodynamic indices show greater vertical instability just in the regions where strongest precipitation occurred.