• DOI:10.1127/0941-2948/2009/0393
• Corpus ID: 67845213

Serial clustering of intense European storms

@article{Vitolo2009SerialCO,
  title={Serial clustering of intense European storms},
  author={Renato Vitolo and David B. Stephenson and Ian M. Cook and Kirsten Mitchell-Wallace},
  journal={Meteorologische Zeitschrift},
  year={2009},
  volume={18},
  pages={411-424}
}

• R. Vitolo, D. Stephenson, +1 author K. Mitchell-Wallace
• Published 1 April 2009
• Environmental Science
• Meteorologische Zeitschrift

This study has investigated how the clustering of wintertime extra-tropical cyclones depends on the vorticity intensity of the cyclones, and the sampling time period over which cyclone transits are counted. Clustering is characterized by the dispersion (ratio of the variance and the mean) of the counts of eastward transits of cyclone tracks obtained by objective tracking of 850 hPa vorticity features in NCEP-NCAR reanalyses. The counts are aggregated over non-overlapping time periods lasting from 4 days up to 6 month long OctoberMarch winters over the period 1950–2003. Clustering is found to be largest in the exit region of the North Atlantic storm track (i.e. over NE Atlantic and NW Europe). Furthermore, it increases considerably for the intense cyclones, for example, the dispersion of the 3-monthly counts near Berlin increases from 1.45 for all cyclones to 1.80 for the 25 % most intense cyclones. The dispersion also increases quasi-linearly with the logarithm of the length of the aggregation period, for example, near Berlin the dispersion is 1.08, 1.33, and 1.45 for weekly, monthly, and 3-monthly totals, respectively. The increases and the sampling uncertainties in dispersion can be reproduced using a simple Poisson regression model with a time-varying rate that depends on large-scale teleconnection indices such as the North Atlantic Oscillation, the East Atlantic Pattern, the Scandinavian pattern, and the East Atlantic/West Russia pattern. Increased dispersion for intense cyclones is found to be due to the rate becoming more dependent on the indices for such cyclones, whereas increased dispersion for longer aggregation periods is related to the small amounts of intraseasonal persistence in the indices. Increased clustering with cyclone intensity and aggregation period has important implications for the accurate modelling of aggregate insurance losses. Zusammenfassung