PFRs are identified as periods of widespread persistent anomalies as defined by Dole and Gordon (1983) and updated by Miller et al. (2020) using CFSR reanalysis data for DFJ from 1979–2022. All PFRs are aggregated into five groups using k-means clustering of the 500-hPa geopotential height data at the start of each PFR. The teleconnections used in this study are the MJO as defined in Wheeler and Hendon (2004) and the NPJ phase as defined in Winters et al. (2019). Initial work will identify select phases of each index that exhibit increased occurrence frequency in periods leading up to the formation of each composite PRF. These results will be complemented with an assessment of statistical significance through a bootstrap resampling method. The above relationships will be reversed to assess the likelihood of a particular PFR forming at various time lags after a given phase of the MJO or NPJ to translate the initial PRF–teleconnection relationships into ones which have use for forecasting. The third and final component of this study is to compare composites of select phases of the MJO and NPJ that do and do not result in a PFR in an attempt to link the statistical results to improved dynamical understanding of the teleconnection relationships.

