Snowpack Variations in the Central Andes of Argentina and Chile, 1951–2005: Large-Scale Atmospheric Influences and Implications for Water Resources in the Region

Abstract

The snowpack in the central Andes (30°–37°S) is the primary source for streamflow in central Chile and central-western Argentina, but few published studies are available on snowpack variability in the region. This paper develops the first regional snowpack series (1951–2005) from Chilean and Argentinean snow course records. This series shows a strong regional signal, marked interannual variability, and a positive, though nonsignificant, linear trend. Correlations with local precipitation and temperature records reveal a marked association with conditions in central Chile. High snow accumulation is generally concurrent with El Niño events in the tropical Pacific, but only 5 of the 10 driest years coincided with La Niña events. Evaluation of 500-hPa geopotential height anomaly maps during extreme snow years highlights the crucial significance of tropospheric conditions in the subtropical and southeast Pacific in modulating snowfall. Correlations with gridded SST and SLP data and multiple regressions with large-scale climatic indices corroborate a Pacific ENSO-related influence largely concentrated during the austral winter months. This hampers the predictability of snowpack before the onset of the cold season. Annual and warm-season river discharges on both sides of the cordillera are significantly correlated with the regional snowpack record and show positive linear trends over the 1951–2004 common period, probably related to a greater frequency of above-average snowpacks during recent decades. Future demand and competition for water resources in these highly populated regions will require detailed information about temporal and spatial variations in snow accumulation over the Andes. The results indicate that the relationships between snowpack and atmospheric circulation patterns prior to the winter season are complex, and more detailed analyses are necessary to improve prediction of winter snowfall totals.