A spatial assessment of hydrologic alteration caused by dams in the Northeastern United States using a Neural Network based daily reservoir operation scheme

Monday, 15 December 2014: 8:45 AM
Nima Ehsani1,2, Charles J Vorosmarty1,2, Balazs M Fekete1,2, Bernice Rosenzweig2 and Zachary D Tessler2, (1)CUNY City College, New York, NY, United States, (2)CUNY Environmental Crossroads, New York, NY, United States
Considering the impacts of dams on natural hydrology and ecosystems, it is important to be able to simulate their behavior and effects in hydrological and ecological models. Overlooking human engineering of river systems may significantly affect modelling results and impact decisions addressing water management issues.

Simulating reservoir operation at the regional and global scale remains a challenge in water resource and environmental science. There are numerous studies that model the operating rules of a single or small cluster of dams based on available observed data or that try to find an optimized set of rules for their operation based on their characteristics and intended purpose. On the other hand, there are few works that consider the operation of dams for regional and global hydrological models. One major problem in modeling dams operation in such large-scale systems is the lack of efficient algorithms for modelling reservoir operation. Depending on site-specific characteristics of the dam, its watershed and its intended purpose, each dam has a specific and optimum operating rule; as a result, effective simulation of their operation is not a trivial task when hundreds and thousands of dams exist in the area of study.

As part of the development of the Northeast Regional Earth System Model (NE-RESM), we are developing an integrated hydrological modeling framework that incorporates various aspects of the coupled human-hydrologic system, from supply to demand, into a single framework. We use an Artificial Neural Network to develop an accurate yet generalized daily operating rule with minimal input requirements that is suitable for use in large scale hydrological models. We implement this reservoir operating scheme into WBMplus and study how dams alter natural hydrology of the Northeastern United States. We also show how climate change impacts the operation of reservoirs and hence availability of water in the region by the end of the 21st century.