Clustering Analysis of Autumn Weather Regimes in the Northeast United States

Abstract A k- means clustering method is applied to daily ERA5 500-hPa heights, sea level pressure, and 850-hPa winds, 1979–2008, identify characteristic weather types (WTs) for September–November the northeast United States. The resulting WTs are analyzed in terms of structure, frequency occurrence, typical progressions, precipitation temperature characteristics, relation teleconnections. used make a circulation-based distinction between early late autumn consider shifts seasonality. Seven identified season, representing range trough ridge patterns. largest average values greatest likelihood extremes occur Midwestern Trough Atlantic Ridge extreme temperatures occurs Northeast Ridge. Some strongly associated with phase North Oscillation Pacific–North America pattern, occurrence several changing by more than factor 2. two most common progressions one frequent September, Mid-Atlantic Trough, mid-October–November, Trough. This seasonality allows WT-based season. preliminary trend analysis indicates an increase season later decrease earlier season; that is, shift toward longer period warm patterns shorter, delayed cold