Bibliographie complète

ABSTRACT High‐resolution reanalyses offer the potential to improve our understanding of midlatitude cyclones, particularly smaller‐scale systems and those with complex structures. However, previous studies have demonstrated large variations in the frequency and characteristics of Australian midlatitude cyclones between reanalyses when using their native resolution. In this paper we use satellite observations of winds and rainfall in order to evaluate the ability of the ERA‐Interim, JRA55, MERRA and CFSR reanalyses to reproduce Australian east coast cyclones. The MERRA reanalysis produces a large number of erroneous small‐scale lows without cyclonic wind patterns using a simple pressure‐difference‐based cyclone identification and tracking method. Consequently, we recommend the ERA‐Interim reanalysis when using such methods, or applying more complex tracking methods that are able to compensate for these issues.