Reconstructing streamflow variation of the Baker River from tree-rings in Northern Patagonia since 1765

• We developed a streamflow reconstruction for the Baker River for 1765–2004.
• This study constitutes the southernmost tree-ring streamflow record worldwide.
• The record shows an unprecedented sustained decline since the 1980s.
• This decline is associated to changes in atmospheric circulation (SAM).
• This streamflow decline needs to be considered in water resources planning.

SummaryThe understanding of the long-term variation of large rivers streamflow with a high economic and social relevance is necessary in order to improve the planning and management of water resources in different regions of the world. The Baker River has the highest mean discharge of those draining both slopes of the Andes South of 20°S and it is among the six rivers with the highest mean streamflow in the Pacific domain of South America (1100 m3 s−1 at its outlet). It drains an international basin of 29,000 km2 shared by Chile and Argentina and has a high ecologic and economic value including conservation, tourism, recreational fishing, and projected hydropower. This study reconstructs the austral summer – early fall (January–April) streamflow for the Baker River from Nothofagus pumilio tree-rings for the period 1765–2004. Summer streamflow represents 45.2% of the annual discharge. The regression model for the period (1961–2004) explains 54% of the variance of the Baker River streamflow (R2adj = 0.54). The most significant temporal pattern in the record is the sustained decline since the 1980s (τ = −0.633, p = 1.0144 ∗ 10−5 for the 1985–2004 period), which is unprecedented since 1765. The Correlation of the Baker streamflow with the November–April observed Southern Annular Mode (SAM) is significant (1961–2004, r = −0.55, p < 0.001). The Baker record is also correlated with the available SAM tree-ring reconstruction based on other species when both series are filtered with a 25-year spline and detrended (1765–2004, r = −0.41, p < 0.01), emphasizing SAM as the main climatic forcing of the Baker streamflow. Three of the five summers with the highest streamflow in the entire reconstructed record occurred after the 1950s (1977, 1958 and 1959). The causes of this high streamflow events are not yet clear and cannot be associated with the reported recent increase in the frequency of glacial-lake outburst floods (GLOFs). The decreasing trend in the observed and reconstructed streamflow of the Baker River documented here for the 1980–2004 period is consistent with precipitation decrease associated with the SAM. Conversely, other studies have reported an increase of summer streamflow for a portion of the Baker River for the 1994–2008 period, explained by ice melt associated with temperature increase and glacier retreat and thinning.Future research should consider the development of new tree-ring reconstructions to increase the geographic range and to cover the last 1000 or more years using long-lived species (e.g. Fitzroya cupressoides and Pilgerodendron uviferum). Expanding the network and quality of instrumental weather, streamflow and other monitoring stations as well as the study and modeling of the complex hydrological processes in the Baker basin are necessary. This should be the basis for planning, policy design and decision making regarding water resources in the Baker basin.