Verification of an objective cyclone climatology for the North Atlantic

Abstract

A numerical cyclone indentification scheme is tested with respect to its ability to reproduce positions and intensities of deep low pressure systems over the North Atlantic. The scheme uses 1000 hPa geopotential height analyses. First, data are computed for a regular grid, with input data adapted to a uniform spectral resolution (T63 in our case). The height minima are then determined from an interpolated field compiled from a bicubic spline interpolation. A comparison is performed using the positions and depths of extreme cyclones (central pressure below 950 hPa) determined from weather charts for 1980-1991. Corresponding lows from the numerical scheme were determined as the ones with minimum distance from the hand-analysed positions. The median value of the distances for the 102 events considered is about 100 km. Large distances are associated with extended lows incorporating separate minima, part of which are missing either in the analyses or in the charts. Central pressures of the numerically identified lows were estimated by a reduction to sea-level pressure, using a standard value for temperature. The resulting pressures are systematically less extreme than in the maps, with a mean difference of 5 hPa. The most likely reason is the smoothing effect of the analysis scheme. The number of extreme cyclones determined for the period 1980 to 1993 is found to be increasing, which is in accordance with other authors. Inclusion of weaker systems (central pressures lower than 990 hPa) yields no evidence for a trend.

Keywords

• cyclone indentification scheme
• north atlantic
• bicubic spline
• weather charts
• central pressure
• zyklonen-identifikationsprogramm
• nordatlantik
• bikubische splines
• wetterkarten
• kerndruck