Clustering Analysis of Autumn Weather Regimes in the Northeast United States

A k-means clustering method is applied to daily ERA5 500 hPa height, MSLP, and 850 hPa wind data to identify distinct Weather Types (WTs) for Sep-Nov from 1979-2018 over the Northeast U.S. The resulting clusters are analyzed in terms of their frequency of occurrence throughout the season, and typical progressions between the WTs. This analysis is able to distinguish between early and late fall weather based on circulation.

Seven WTs were identified for the autumn season. WTs 1 and 2 feature a weak upper-level trough over the Northeast, with a surface low (high) pressure to the north (south) of the region in WT1 (WT2). WTs 3 and 4 have a strong upper-level trough over the Northeast, with a surface low pressure over (northeast of) the region in WT3 (WT4). WTs 5 and 6 feature a strong upper-level ridge over the Northeast with a surface high pressure to the east (over) the region in WT5 (WT6). WT7 shows a weak upper-level ridge over the Northeast with a surface high pressure over the region.

In terms of frequency of occurrence and progression between WTs, three main sequences emerged: 1.) A sequence in September, defined by WT1-WT6-WT1 2.) The second primarly occurs in October, defined by WT1-WT6-WT5-WT1. 3.) Finally, a sequence defined by 4 of the WTs (3, 4, 5, and 7) that occurs most often from mid-October to November. Sequences 1 and 2 both define the "Early Season," while Sequence 3 defines the "Late Season." The early and late season sequences are linked by a common WT, WT5, which is present in both sequences.