Interannual variability and long-term stability of the north atlantic subtropical high since the 18th century: a binary dendroclimatic reconstruction

Abstract We analyzed the position of the North Atlantic Subtropical High (NASH) as indexed by summertime 1020 hPa pressure contours during 1948–2020 ( n = 73) to assess the utility of using latewood (mid–late summer) radial growth data from climate-sensitive old-growth southeastern USA pine species to reconstruct NASH location since 1771 CE. The linear association between NASH position and latewood growth has limited explanatory strength, with <6% explained variance. Thus, we assessed the association binarily, positing that either dry or wet years, inferred from radial growth, corresponded with a threshold position of NASH based on the sum of NASH latitude and longitude degrees being either less than or greater than 100. Based on validation from a 73-year historical data set, we found that NASH positions: 1) less than 100 degrees were coincident with negative (dry summer) tree-ring chronology z-scores with 75.0% (24/32) accuracy, 2) greater than 100 degrees were coincident with positive (wet summer) chronology z-scores with 63.4% (26/41) accuracy, 3) matched to all z-scores produced an overall accuracy of 68.5% (50/73) and 4) classification accuracy (68.5%) exceeds simulated classification accuracy in 1998/2000 (99.9%) Monte Carlo simulations. Positive (negative) tree-ring chronology z-scores are 5.07 (95% CI: 1.69–16.64) times more (less) likely during years with greater than (less than) 100 NASH latitude and longitude sums and from this relationship we binarily classified and reconstructed the historic location of NASH as either ‘Southeast’ (drier) or ‘Northwest’ (wetter) from 1771–2020 ( n = 250 years). We found that despite considerable interannual variability between NASH positioning where the median length of the Southeast and Northwest phases was <5 years, the reconstructed NASH locations have no significant correlation with time (i.e., year) during 1771–2020 ( r = 0.02, p = 0.75). These results suggest long-term stability regarding the frequency of Northwest and Southeast positioning of NASH over the past quarter millennium, and that alternative reconstruction methodologies, such as a binary approach, may be applicable in instances where linear relationships have minimal explanatory power.