Dendrochronology and links to streamflow

SummaryStreamflow variability on timescales of decades to centuries becomes increasingly important as water managers grapple with shortages imposed by increasing demand and limited supply, and possibly exacerbated by climate change. ► Two applications of dendrochronology to the study of flow variability are illustrated for an existing 1244-yr reconstruction of annual flows of the Colorado River at Lees Ferry, Arizona, USA: (1) identification and climatological interpretation of rare flow events, and (2) assessment of vulnerability of water-supply systems to climatic variability. ► Analysis centers on a sustained drought of the mid-1100s characterized by persistent low flows on both the Colorado and Sacramento Rivers. ► Analysis of geopotential height anomalies during modern joint-droughts suggests more than one mode of circulation might accompany joint-drought in the two basins. ► Monte Carlo simulation is used to demonstrate that a drought as severe as that in the 1100s on the Colorado River might be expected about once in every 4–6 centuries by chance alone given the time-series properties of the modern gaged flows. Application of a river-management model suggests a mid-1100s-style drought, were it to occur today, would drop reservoir levels in Lake Mead to dead-pool within a few decades. ► Uncertainty presents challenges to accurately quantifying severe sustained droughts from streamflow reconstructions, especially early in the tree-ring record. ► Corroboration by multiple proxy records is essential. ► Future improvements are likely to require a combination of methodological advancements and expanded basic data.

Research highlights► The mid-1100s is a model for devastating hydrologic drought in the western US. ► Exact simulations highlight the risk of drought on the Colorado River. ► Persistent drought on the Colorado River could arise from today’s climate variability. ► Recurrence of medieval-period drought would rapidly lower Lake Mead to dead pool.