An analysis of New Zealand synoptic types and their use in defining weather regimes

A new set of 12 daily weather types for the New Zealand region has been der ived from the 40-year NCEP/NCAR reanalysis dataset. Cluster analysis of the monthly frequencies of these patterns has led to the definition of three ' regimes', characterized by (1) frequent troughs crossing the country, Oil h ighs to the north with strong zonal flow to the south of the New Zealand, a nd (iii) blocking patterns with highs more prominent in the south. Blocking regimes are more frequent in summer and autumn and are associated with above-normal temperatures, less precipitation in the southwest of the country and more precipitation in the northeast. The Zonal regime, which br ings below-normal precipitation to the northeast and milder conditions in t he south, is less common in summer. The Trough regime is less frequent in a utumn and is linked to cooler temperatures in the west and above-normal pre cipitation over the entire country. The monthly frequencies of individual synoptic types are only weakly relate d to the Southern Oscillation Index (SOI) and other indices of the hemisphe ric-scale flow, with variance reductions from regression equations ranging from between 3 and 33%. Similar predictions of weighted combinations of the synoptic types corresponding to each regime gave mean variance reductions 70-80% higher. The correct regime could be specified from the hemispheric i ndices by the use of discriminant analysis for 56% of the 474 months in the dependent dataset. The monthly frequency of synoptic types was also tested as an alternative t o monthly mean 1000 hPa patterns in the selection of analogues over the New Zealand region. The success of the selection process was measured by the v ariance of temperature and rainfall patterns over homogeneous regions for s ets of two, four and eight analogues. The mean 1000 hPa patterns scored bet ter in nearly all cases, but gave variance reductions of only 13-19% over r andomly chosen analogues. Copyright (C) 2000 Royal Meteorological Society.