Inter-Research > CR > v23 > n1 > p11-22  
Climate Research

via Mailchimp

CR 23:11-22 (2002)  -  doi:10.3354/cr023011

Automated objective classification of daily circulation patterns for precipitation and temperature downscaling based on optimized fuzzy rules

András Bárdossy1,*, Jiri Stehlík2, Hans-Joachim Caspary3

1Institute of Hydraulic Engineering, University of Stuttgart, Pfaffenwaldring 61, 70550 Stuttgart, Germany
2Department of Physical Geography and Geoecology, Faculty of Science, Charles University in Prague, Albertov 6, 12843 Prague 2, Czech Republic
3Fachhochschule Stuttgart - Hochschule für Technik, Schellingstr. 24, 70174 Stuttgart, Germany

ABSTRACT: A methodology for automated objective circulation pattern (CP) definition and classification is presented based on optimized fuzzy rules. The main goal of the method is to provide a basis (daily classified CPs) for downscaling of precipitation and temperature, which can be done by means of downscaling models with parameters depending on the CP. The CPs are defined using 500 or 700 hPa geopotential height anomalies. Fuzzy rules are defined by the position of high- and low-pressure anomalies. The fuzzy rules are obtained automatically using an optimization of the performance of the classification. For precipitation, the performance of the classification is measured by rainfall frequencies and rainfall amounts conditioned on the CP. So the task is to define wet or dry CPs. For temperature, the deviation from the average long-term annual cycle is used. In this way warm or cold CPs are identified. The performance of the CPs is validated using a split-sampling approach. The presented method produces physically realistic CP definitions. With the help of these definitions the observed (historical) pressure fields can be classified as shown in 2 case studies for regions with different climate conditions: Central Europe (Germany) and the Eastern Mediterranean (Greece).

KEY WORDS: Circulation-pattern classification · Downscaling · Precipitation · Temperature · Fuzzy sets

Full text in pdf format
 Previous article Next article