Controls of Large-Scale Meteorological Patterns on Annual Maximum Streamflow across the Conterminous United States

Linking synoptic-scale atmospheric circulations to flood generation remains a critical gap for flood frequency analysis and forecasting. This study presents a continental-scale attribution of annual peak flows (APFs) to their generating large-scale meteorological patterns (LSMPs) across 585 minimally disturbed watersheds in the conterminous United States from 1980 to 2014. We linked APFs to their causative meteorological patterns using nearby rain gauges and identified five primary flood-generating meteorological mechanisms: atmospheric rivers, extratropical cyclones, frontal systems, tropical cyclones, and the North American monsoon. Frontal systems dominated flood generation in 56% of the watersheds, followed by extratropical cyclones at 28% and atmospheric rivers at 16%, with distinct geographic patterns. Notably, 30% of peak flows could not be definitively attributed to an LSMP, reflecting convective systems and transitional conditions beyond synoptic classifications. Precipitation-runoff transformations varied fundamentally among the LSMPs: Atmospheric rivers achieved exceptional flood generation through sustained moisture delivery over multiple days (median of 6 days across all gauges and events) and large accumulations (median of 168 mm). Tropical cyclones generated floods via extreme precipitation intensities (median of 47 mm/day) over short durations (<2 days). Statistical analysis revealed that 63% of the watersheds, mainly in continental interiors, showed no distinction between flood populations associated with the two most frequent LSMPs, validating single-population frequency assumptions. However, coastal regions exhibited mixed populations with strong statistical separation between LSMP-driven APFs. This study identified regions requiring mixed-population approaches for flood frequency analysis.