Clustering of cyclones in the ARPEGE general circulation model

Abstract

Baroclinic waves give major contributions to the climatology and variability of key climate variables in mid‐ and high‐latitudes. An important aspect of cyclone variability is the seriality (succession of cyclone occurrence). Serial cyclones are clustered in time and are associated with large economic losses in Europe. Cyclone variability and seriality are intimately linked to the low frequency variability. To have a realistic representation of high and mid‐latitude regional climate in general circulation models (GCM), it is of vital importance that the described phenomena are realistically represented. In the present paper we evaluate the ability of the atmospheric GCM ARPEGE to represent various storm track parameters, both with respect to climatology and variability. Cyclone tracks are identified for the Northern Hemisphere during 55 extended winters (1950/51–2004/05) using the NCEP reanalyses by means of an objective tracking method. A corresponding cyclone track data set is derived from a simulation with ARPEGE, forced with observed monthly SSTs from 1950 to 2004. The comparison of the two data sets shows that ARPEGE reproduces the main features of the observed cyclone track fields. The mid‐latitude corridors of cyclones and cyclone variability are well reproduced. The model, however, underestimates the mean number of cyclone occurrence and its variance. In the eastern North Atlantic and Nordic Seas the model simulates 70 and 90% of the observed mean cyclone occurrence and its variance, respectively. ARPEGE also substantially underestimates the clustering of cyclones in the eastern North Atlantic (jet exit region). In addition, the model also fails to simulate the relationship between clustering and low‐frequency flow patterns at the exit of the North Atlantic storm track.

New perspectives on the collective risk of extratropical cyclones

“…Clusters of extratropical cyclones can result in economic losses comparable to those of a US hurricane and, due to the structure of reinsurance contracts, a cluster of events can cost more than a single event with the same total loss (Vitolo et al , ). Climate models have been shown to underestimate clustering (Kvamstø et al , ) and the physical drivers of clustering remains an active area of ongoing research (e.g. Hanley and Caballero, ; Neu et al , ; Pinto et al , ; Blender et al , ).…”

The observed clustering of damaging extratropical cyclones in Europe

“…This suggests we cannot gain the benefits of smaller sampling errors from long integrations of the ECHAM5 climate model at the present time, due to its inability to simulate observed stronger clustering of more severe storms. Kvamsto et al (2008) note significant differences in clustering between a different climate model and observations, though β versus storm severity is not analysed. These two studies suggest climate models have different clustering behaviour from observed; however, they represent a small sample, and analysis of more climate models is needed to make firmer, useful conclusions on climate models' quality of clustering simulations.…”

“…However, previous studies find significant differences between climate models and observed behaviour (e.g. Kvamsto et al, 2008, and the underestimate of clustering for most severe storms in Tables 3a and b of Karremann et al, 2014a). New research by Pinto et al (2014) looks for the underlying mechanisms generating the cyclone families and persistent climate states that produce severe clusters on seasonal timescales.…”

Storm wave clustering around New Zealand and its connection to climatic patterns

“…The I d is a measure of the normal variability of intervals between storm peaks and is the quantity (with time dimension) defined as the ratio of the variance ( ΔT − ⟨ ΔT ⟩) 2 of the time interval ΔT between storm peaks to the mean interval ⟨ ΔT ⟩. The method itself is not innovative, because it has been applied by other authors (Mailier et al ., ; Kvamstø et al ., ; Vitolo et al ., ) in different contexts and using different approaches. However, the way the method has been employed here differs from previous work.…”

“…Several techniques to identify groups of cyclones have been implemented, such as a Bayesian approach (Fawcett and Walshaw, 2008), a running sum of daily cyclone counts e402 V. A. GODOI et al (Pinto et al, 2014), and the calculation of a dispersion statistic based on the Poisson process and cyclone counts (Mailier et al, 2006;Kvamstø et al, 2008;Vitolo et al, 2009). Studying cyclone clusters, however, does not necessarily provide useful information on the formation of SWCs.…”