The development of a gridded weather typing classification scheme

Abstract

ABSTRACTGridded reanalysis datasets have proven quite useful for a broad range of synoptic climatological analyses, especially those utilizing a map‐pattern classification approach. However, their use in broad‐scale, surface weather typing classifications and applications have not yet been explored. This research details the development of such a gridded weather typing classification (GWTC) scheme using North American Regional Reanalysis data for 1979–2010 for the continental United States. Utilizing eight‐times daily observations of six surface variables, the GWTC categorizes the daily surface weather of 2070 locations into one of 11 discrete weather types – nine core types and two transitional types (TRs) – that remain consistent throughout the domain. Due to the use of an automated deseasonalized z‐score initial typing procedure, the character of each type is both geographically and seasonally relative, allowing each core weather type to occur at every location, at any time of the year. Diagnostic statistics reveal a high degree of spatial cohesion among the weather types classified at neighbouring locations, along with an effective partitioning of climate variability into these 11 weather types. Daily maps of the spatial distribution of GWTC weather types across the United States correspond well to traditional surface weather maps, and comparisons of the GWTC with the Spatial Synoptic Classification are also favourable. While the potential future utility of the classification is expected to be primarily for the resultant calendars of daily weather types at specific locations, the automation of the methodology allows the classification to be easily repeatable, and therefore, easily transportable to other locations, atmospheric levels and datasets, including general circulation models. The enhanced spatial resolution of the GWTC may allow for new applications of surface weather typing classifications in mountainous and rural areas not well represented by weather stations.