Fig. 2
From: Atmospheric teleconnection patterns and hydrological whiplashes in the Western U.S.
Dry and wet conditions in western U.S. Basins (1981–2023) and teleconnection influences. (a) Violin plots displaying 5-year statistics of normalized streamflow data for the California Basin over 1981–2023. The central red line marks the historical mean from 1981–2020; plots below this line suggest dry conditions, while those above indicate wet conditions. The extremity of conditions is shown by the distance of each plot from this line. Blue dots represent stations with severe transitions from critically dry to wet conditions (“critical to wet whiplash”), and red dots indicate stations transitioning from wet to critically dry (“wet to critical whiplash”). Scatterplots illustrate the responses of whiplash and non-whiplash stations to changes in Western Pacific Oscillation (ΔWP) values from August of the previous year (e.g., August ΔWP2022–2021 vs. Δ Normalized Indicator2023–2022), highlighting the potential impact of teleconnections on whiplash events. Notably, the R2 value for whiplash stations (0.54) suggests a stronger response than non-whiplash stations (0.1). August was selected for analyzing the influence of the WP based on lag correlation analysis showing the strongest teleconnection signal with subsequent water year streamflow approximately 1–2 months before the water year begins. This timing coincides with late summer atmospheric circulation patterns that often establish conditions for the upcoming wet season. (b) Similar analysis for the Colorado River Basin showing influence of August ΔWP. (c) Analysis for the Columbia River Basin, with scatterplots focused on September Eastern Pacific/Northern Pacific Oscillation (ΔEPO) impacts. EPO in September was specifically analyzed for the Columbia River Basin because this basin shows unique sensitivity to early fall atmospheric circulation patterns over the northeastern Pacific. Our analysis of lag correlations revealed that September EPO configurations have the strongest predictive relationship with Columbia Basin streamflow in the following water year. This timing is critical as it coincides with the initial establishment of winter storm tracks that predominantly affect the northern portion of our study area. (d) As in panel (a) and (b) but for the Great Basin, with scatterplots reflecting the influence of August ΔWP changes on streamflow conditions.
