Procedura integrata di analisi e correzione delle osservazioni radar per la stima di precipitazioni intense in ambiente alpino

Autor: Marra, Francesco
Jazyk: italština
Rok vydání: 2013
Předmět:
Popis: The availability of weather radar precipitation estimations is increasing the knowledge of rainfall systems on spatial and temporal scales that cannot be reached by rain gauge networks; consequently the approach to hydrological and geomorphological modeling is changing from concentrated catchment scale modeling - the only available for operational purposes when rainfal is estimated with rain gauges - to spatially distributed modeling. The use of weather radar in hydrology responds to the need for a description of rainfall spatial distribution integrated over the desired time scale and available on operational areas, a problem that will never be solved using only rain gauges. Nevertheless we have to mention that weather radar based measurements suffer from several sources of uncertainty. Microphysical assumptions are needed to convert reflected power measurements into rain rate and a number of errors, depending both on the instrument and on the rainfall system, decrease raw measurements quality. Despite the scientific and technological development of the last decades, radar quantitative precipitation estimation is still a scientific challenge and new elaboration procedures are needed for operational purposes (Panziera et al., 2011). The objective of the present research is to improve radar quantitative precipitation estimations for intense rainfall events in alpine regions and, in this way, fill the uncertainties concerning hydro-geological and debris-flows risk evaluation and forecasting that depend on the reliability of radar measurements (Berne e Krajewski, 2012; Chen et al., 2013). Heavy hydrometeorological events are often the most interesting for radar hydrometeorology because i) they are difficult to capture with other instruments (Borga et al., 2000, 2008). The work is organized into three themes: i) development of a re-analysis procedure for intense events, ii) characterization of the rainfall fields and analysis of rain gauge sampling for the same events and iii) improvement of the real-time elaboration routine. The study area is Trentino-Alto Adige that it is characterized by alpine topography and climatology where intense summer storms often lead to flash floods and debris flows. The elaboration procedure is focused on intense events re-analyses. The correction chain takes into account the effects of multiple uncertainties and uses physical-based algorithms for the correction of the effects due to i) wet radome attenuation (Marra et al., 2010), ii) beam blockage (Delrieu e Creutin, 1995), iii) attenuation (Delrieu et al., 1997, 1999a,b, 2000; Serrar et al., 2000) and iv) vertical profile of reflectivity (Andrieu e Creutin, 1995a,b). We re-analyzed 7 storm events and we quantified the quality of rainfall estimations by means of statistical indicators. We quantified the effect of the procedure and the contribution of every correction algorithm to the overall performance. Attenuation contributes with more than 50% in the improvement of intense rainfall estimations; nevertheless none of the other corrections can be neglected. The drawback of this algorithm is that it is still not reliable for real-time applications. Debris flows represent an important hydrogeological risk for mountainous regions; several causes contribute in triggering debris flows but strong short storms are one of the most important for this region (Caine, 1980; Guzzetti et al., 2008). We studied the rainfall fields of some storm events that triggered debris flows and measured their dimensional scales and the height distribution. We also quantified the local severity of the precipitation fields over the debris-flows triggering areas and we analyzed the rain gauge sampling of such events. In this way we show that rain gauges sistematically underestimate rainfall severity over the triggering points; this is caused by two main problems: i) the average inter-distance between raingauges is larger than the typical event scale; ii) rain gauges are usually located at altitudes lower than the orographic enhanced precipitation peaks. We improved the real-time elaboration procedure using the knowledge obtained during re-analysis and we implemented the wet radome attenuation correction algorithm. We quantified the performance of the operational procedure revealing important uncertainties in the estimation of convective rainfall. We developed a new algorithm that distinguishes stratiform and convective precipitation and applies the appropriate procedure. We quantified the performance of the new operational product over a 4 months study period obtaining statistical indicators showing that the new procedure performs better than the present one both for stratiform and convective events. An automatic real-time correction for attenuation represents an open challenge. In presence of intense hydrometeorological events the availability of real-time reliable radar quantitative precipitation estimations will improve the hydro-geologic risk prevenction and mitigation techniques.
Databáze: OpenAIRE
Externí odkaz: