Summary The upper Santa Fe River provides up to 50% of the water supply for the growing population of Santa Fe, NM. Recent droughts have dramatically lowered reservoir levels and raised concern about the future of the water supply, particularly when combined with projections of a warmer and drier future climate. In this study, new and updated tree-ring chronologies are used to reconstruct annual discharge for the upper Santa Fe River and place the short period of gaged flows, 1914–2007, in a long-term context. Principal components analysis and forward stepwise multiple linear regression were used to produce two reconstructions: (1) a better fit, “short reconstruction” (adjusted R2 = 0.62, 1592–2007) and (2) a less robust, “long reconstruction” (adj. R2 = 0.50, 1305–2007). Both reconstructions indicate that recent extreme low flow events (e.g., 2002) are rare (5th percentile) in the long-term records and that the 1950s drought contained the lowest 7-year mean flows over the past 400–700 years. However, longer, multi-decadal dry periods not present in the gaged flows occurred in the past. For example, the 40-year mean for 1544–1583 is estimated at just 86% of the 1914–2007 mean. During extended dry periods in the 16th and 18th centuries the probability that annual flow would not meet the current surface water allocation and instream flow target (7.52 million cubic meters, MCM) was up to 10% greater (78.7% non-exceedence probability) than during the instrumental period. The results indicate that the gaged record does not contain the full range of high and low flows or the variability in the probability distributions of flows present in the long-term record. Therefore current and future water management and planning based on the instrumental period may not adequately buffer against the natural variability in the climate and streamflow systems. This valuable paleo-hydrologic information is in the process of being incorporated into water supply planning for the City of Santa Fe (e.g., modeling future water supply scenarios directly from reconstructed periods of streamflow).
Read full abstract