Hydrological reconstruction from tree-ring multi-proxies over the last two centuries at the Caniapiscau Reservoir, northern Québec, Canada

•We performed hydrological reconstructions from tree-rings in northern boreal region.•We reconstructed annual and summer water supplies back to 1800 for Caniapiscau basin.•The latter half of the 19th century was a period of high water levels.•1920–1950 period is characterized by a low hydrological variability.•Summer water supplies and summer AO index are significantly correlated. The objective of this study was to reconstruct 200years of spring (Qspr), summer (Qsum) and annual (Qann) water supply variability at the Caniapiscau Reservoir in a remote area of northern Quebec. This region which is a key hydropower region of North America lacks long-term hydrological series, and tree-ring proxies are seen here as the best alternative for the extension of the climatic series beyond instrumental records. Thus, ring widths, ring densities and stable isotope ratios (δ13C and δ18O) were used to perform paleohydrological reconstructions. The following reconstruction techniques were evaluated for each variable reconstructed: partial least square (PLS) regression applied to all of the tree-ring series, PLS regression applied to selected tree-ring series, and the best analogue method (BAM) applied to selected tree-ring series. These three reconstructions were then combined in a composite reconstruction. Reconstruction verification shows that the annual and summer water supply reconstruction quality is good. Conversely, the verification tests disqualified our spring water supply reconstruction. The reconstructed long-term water supply variations over the last two centuries are dominated by decadal to sub-decadal fluctuations, including distinct long hydrological periods during which water supplies change in intensity and variability. Annual water supplies (Qann) inversely correlate with both winter and summer indices. Annual water supplies (Qann) inversely correlate with Arctic oscillation (AO) and North Atlantic Oscillation (NAO) winter and summer indices. Also, AO indices also influence Qann to a greater degree than NAO indices. Finally, summer water supplies inversely correlate only with summer indices.