Evaluation systématique de la prévisibilité des dépressions méditerranéennes et impacts sur les territoires
Benjamin Doiteau
Systematic evaluation of the predictability of Mediterranean cyclones and regional impact
Auteur moral
Institut National Polytechnique de Toulouse : INP Toulouse
Auteur moral
Matthieu Plu ; Florian Pantillon
Année de publication
2024
<tbody><tr><td> </td></tr><tr><td align="justify"> </td></tr><tr><td> </td></tr><tr><td> </td></tr><tr><td align="justify"><p align="justify">Cyclones are important components of the weather in the Mediterranean region. On the one hand, the precipitation associated with them provides water for human activities and ecosystems. On the other hand, the most intense cyclones are responsible for severe damage and fatalities in the densely populated Mediterranean basin. Predicting the behaviour of these cyclones is crucial but challenging, and the accuracy of forecasts can vary greatly from case to case. The causes of this variability in forecast quality are not yet fully understood, and the previous studies on the predictability of Mediterranean cyclones are mainly based on the analysis of intense cases, which may not be representative of the broad spectrum of cyclones occurring in the region. Therefore, the aim of this thesis is to provide the first systematic evaluation of the predictability of Mediterranean cyclones.<br />The systematic approach requires a large number of cyclones and forecasts on which the predictability can be evaluated. Cyclones are first tracked on the basis of their relative vorticity in the European Centre for Medium Range Weather Forecasting (ECMWF) fifth generation atmospheric reanalysis (ERA5), which provides homogeneous reference meteorological data over several decades. The same cyclones are then tracked in the ECMWF ensemble reforecasts, which provide forecast data using a fixed model configuration over 20 years. As a result, 1960 Mediterranean cyclones are detected and tracked in both reanalysis and reforecast fields for the period 2001-2021. They form a catalogue in which the track and intensity (central mean sea level pressure) are enriched with a wind severity metric that is adapted to Mediterranean cyclones.<br />The predictability of the 1960 cyclones is first evaluated by quantifying the errors in cyclone location and intensity. The results show that the median location error grows at a constant rate of 40 km per day for the first 72 h lead time, while it grows more slowly for longer lead times. Regarding the intensity error, the bias reaches -0.5 hPa at 12 h lead time, indicating a slight overestimation of intensity compared to ERA5 data. This bias continues to grow at a slower rate of -0.1 hPa per day for longer lead times. The forecast accuracy is then compared between different categories of cyclones. It is shown that motion speed is a critical factor to consider for the predictability of the cyclone location: the slower the better. In particular, the location of the quasi-stationary cyclones in the Gulf of Genoa is well better predicted than in any other category. In terms of intensity, the results reveal that deep and severe cyclones, which mostly occur during winter, are the least predictable at medium-range. Overall, the first systematic evaluation of predictability presented in this thesis shows significant differences between cyclone categories. This approach could be adapted for other basins to facilitate the identification of meteorological contexts that may affect predictability.</p></td></tr></tbody></table>
Accès à la notice sur le site du portail documentaire de Météo-France
Liste complète des notices publiques