A hybrid procedure for classifying synoptic weather types for Louisiana, USA

摘要

A hybrid synoptic classification procedure for classifying Louisiana weather types based on the manual Muller system is presented. This procedure produces a synoptic classification system for Louisiana that harnesses the strengths of both manual and automated classifications, while eliminating the weaknesses. The Muller weather types archive from 1981 to 2001 is used in conjunction with the National Center for Atmospheric Research (NCAR)/National Centers for Environmental Prediction (NCEP) Reanalysis dataset to develop sea level pressure composites for each Muller weather type. The composites are used as seeds in an automated correlation-based algorithm to generate weather types from 1981 to 2001. Results of the automated procedure are compared to the Muller weather type catalog. Despite systematic differences between the two classifications, including the inability of the hybrid procedure to identify frontal placement, the hybrid classification correctly matched the Muller weather type for 52% of the seed days. In addition, the automated catalog captured the seasonal distribution and interannual variability of the Muller types remarkably well. The hybrid synoptic weather classification system applied to weather properties at Shreveport and New Orleans showed significant differences between weather types. The automated procedure does not replicate the Muller weather type classification exactly; however, it is homogenous within itself and has value for describing the variability of surface weather in Louisiana. In fact, it is arguably advantageous for some applications, owing to its objectivity, speed, and reproducibility.