Summer 2021 eNewsletter

Featured Research: Graduate Research Fellow Update
Forecasts to Improve Multi-Objective Water Level Management on Lake Ontario

Kenji Doering and Kyla Semmendinger, Cornell University

Satellite image of Lake Ontario. Photo Credit: NOAA Great Lakes CoastWatch.

Outflows from Lake Ontario are regulated by the International Joint Commission (IJC) at the Moses Saunders Dam on the St. Lawrence River. In 2017, the IJC implemented a new regulation plan called Plan 2014, with the goal of reintroducing natural variability in lake levels to promote wetland health and services while maintaining or improving other system objectives, such as flood control, navigation, hydropower, and recreation. However, record setting precipitation followed a few months after Plan 2014 was implemented, and Lake Ontario experienced its worst flood event on record. The 2017 record was then broken just two years later during the summer of 2019.

Kenji Doering is a 2020-2021 CIGLR Graduate Research Fellow at Cornell University working with fellow graduate student Kyla Semmendinger (Cornell University), advisor Scott Steinschneider (PhD, Cornell University), and co-advisors Lauren Fry (PhD, NOAA GLERL) and Deborah Lee (NOAA GLERL) to advance the use of subseasonal-to-seasonal hydroclimate forecasts in lake level management. Hydroclimate forecasting on subseasonal-to-seasonal timescales (2 weeks to 3 months) is emerging as a new frontier at the intersection of climate, hydrology, and water resources management. Subseasonal-to-seasonal forecasts could be particularly useful for managing Lake Ontario water levels, as they provide information at critical times of year needed to balance tradeoffs between wetland inundation and flood risk.

“Our results identify the type of flood events that might be reduced or avoided with perfect water supply forecasts,” says Semmendinger. “Additionally, we found that flood reductions on Lake Ontario are possible with forecasts that have moderate improvements in skill, and these improvements do not increase flooding downstream on the St. Lawrence River near Montreal.”

“Our research also evaluates optimal multi-objective management policies for the Lake Ontario system,” says Doering. Together, Kenji and Kyla are expanding their analysis to determine if flood reductions from improved forecasts are possible without compromising other system objectives, like wetland health and navigation, through revised operating rules of the Moses Saunders Dam. This work uses machine learning and state-of-the-art multi-objective optimization to discover new operating policies that can respond to various sources of forecast information. “By improving Lake Ontario flood forecast information, we hope to minimize the impacts of future flood events while preserving the services and benefits of the entire ecosystem,” said Doering.


About the Authors

Kenji Doering is CIGLR Graduate Research Fellowship recipient in 2020-2021 and PhD candidate at Cornell University in Biological and Environmental Engineering. His research focus is on advancing the integration of renewable energies, of which dam and reservoir operations will continue to play a large role.

Kyla Semmendinger is currently a PhD candidate at Cornell University in Biological and Environmental Engineering. This partnership has provided Kyla with an opportunity to collaborate with NOAA GLERL scientists and gain insight into the adaptive management framework of complex water resources.