Using large scale circulation indices to predict the intensity of cold air outbreaks over extended time scales across the southeastern U.S

Extreme cold outbreaks have significant economical and social impacts across the southeastern U.S., particularly in terms of agriculture and energy demand. Recent research has established that these outbreaks are embedded within a distinctive planetary scale circulation regime. In this study relationships are identified between selected large scale circulation indices and the occurrence of different intensities of cold outbreaks across the southeastern U.S. These relationships are exploited to develop empirical models that predict the possibilities of different outbreak intensities over forecasting periods of one week to several months. A 57-yr time series of temperature is constructed from 18 stations to identify cold air outbreaks and quantify their intensity. Daily (PNA, NAO, and AO) and monthly (ENSO 3.4, PDO, and QBO) circulation indices are extracted and combined to characterize aspects of the hemispheric circulation. Summary statistics are used to describe the relationships between circulation anomalies and outbreak intensity. Tree regression models are then developed to predict the long range probabilities of ordinary, strong, and extreme cold air outbreaks from the large scale circulation as summarized by the circulation indices.