Predicting the variability of river discharges at decadal and longer time scales is critical for anticipating and managing future water supplies. Currently, there is a knowledge gap, in part because of a lack of reliable instrumental streamflow data that typically date back only to the 1950s. Thus, the design and operation of water systems may not adequately take into account persistent climate regimes and associated vulnerability to prolonged dry/wet spells. Knowledge of the duration, severity and near-to-long-term likelihood of dry or wet multi-year sequences can improve the management and planning of water systems and help better adapt to climate and accommodate shifts in environmental regulations and changing demands.
In this Science Talk, Prof. Devineni will present recent innovations in tree-ring-based streamflow reconstructions and their applications for reservoir risk analysis. The talk will include the exhibition of Bayesian statistical machinery necessary to handle large watersheds, multiple proxies and hydrologic records, handling data having unequal lengths, spatial correlation structures, and parameter uncertainties. It will also include addressing reservoir drought risk considering water demands and paleo inflows into the reservoir. Examples from small reservoirs serving large populations (New York Reservoirs) will be used as a case study.
Speaker
Professor Naresh Devineni
