Subjectivity in a computer‐assisted synoptic climatology II: Relationships to surface climate

Abstract

Earlier work has demonstrated that the results of a correlation-based (Kirchhofer) synoptic classification are dependent on the way that the internal parameters and data are used. In this study, varying classifications are compared to the climate at one location. The results show that associations between temperature and the synoptic patterns are usually physically interpretable and somewhat stable, whereas precipitation-synoptic type relationships are more difficult to explain and highly variable. We conclude that varying classifications do not produce equivalent synoptic climatologies. Furthermore, we know of no apparent objective or subjective means for determining a 'best' classification. Based on these findings, we discourage the use of correlation-based techniques in all but the most controlled situations.

Winter synoptic weather types in Catalonia (NE Spain) and their linkage with minimum temperature anomalies

“…Some examples can be found in Bailey (1967), Anderberg (1973), Sneath and Sokol (1973), Periago et al (1991) and Lana and Fernàndez Mills (1992). Other authors use automated classifications specifically in synoptic climatology (Kalkstein and Corrigan, 1986;Kalkstein et al, 1987;Yarnal and White, 1987;Yarnal et al, 1988;Davis and Kalkstein, 1990;Serra et al, 1992;Serra, 1995).…”

Linking global circulation model synoptics and precipitation for western North America

“…A number of authors (Key and Crane, 1986;Yarnal and White, 1987;Yarnal et al, 1988;Huth, 1996) have demonstrated that the results of these schemes are highly subjective and that the number of synoptic types and days classified vary with a number of investigator-controlled variables. One of the biggest complaints is that the investigator cannot control the patterns generated.…”

Minimum sample size for synoptic weather type classification. Application to winter period data recorded on the catalan coast (North‐East Spain)

“…Some examples applied to different fields can be found in Bailey (1967), Sneath and Sokol (1973), Anderberg (1973), Yarnal and White (1987), Yarnal et al (1988), Faiers (1988), Davis and Kalkstein (1990), Periago et al (1991), andFernandez Mills (1992).…”

“…The first optimal subsampling studies were taken into account in the 1970s and 1980s (Barry and Perry, 1973;Dixon and Leach, 1978;Key and Crane, 1986, among others). More recently, Yarnal and White (1987) showed that sample size affects the synoptic classification results when a correlation based on the Kirchhofer technique is used. The evaluation of the optimal sample size is important for several reasons.…”