Great Lakes Dashboard
Operational Data
Research Data
Download Data
Information on Research Based Long Term (Multi-Decadal) Water Level Forecasts

Many studies have been published in the past several decades that use models to assess the impact future climates will have on Great Lakes water levels. Although a thorough understanding of each study will require reading the source material, the dashboard allows us to put these different projections side by side for visual comparison.

The long term water level projections presented here are from four recent studies:


Lofgren et al. developed a new approach for estimating potential evapotranspiration by using an energy budget-based approach instead of the more common method of using air temperature as a proxy for evapotranspiration. Lofgren's base case is 1958-2005. Using two different climate models and two methods (details below) for calculating potential evapotranspiration yielded 5 projections:

  • Base
  • CGCM3 / Delta method
  • CGCM3 / Alternate Energy Adjustment Formulation
  • GFDL20 / Delta method
  • GFDL20 / Alternate Energy Adjustment Formulation

Climate Models Used:

  • Canadian Centre for Climate Modeling and Analysis' Coupled General Circulation Model Version 3 (CGCM3)
  • Geophysical Fluid Dynamics Lab Climate Model Version 2.0 (GFDL20)

Methods used:

  • Delta method (using air temperature as ET proxy)
  • Alternative Energy Adjustment Formulation (energy-budget based approach)

Both methods employ the GLERL Large Basin Runoff Model and were transformed to lake levels using the Coordinated Great Lakes Regulation and Routing Model.


Hayhoe et al. used both high (Special Report on Emissions Scenarios - SRES - A1fi) and low (B1) emissions scenarios to project future climate. Only the high emission scenario was used to plot change in lake levels (Fig 10 in the reference). Since the data was not available in digital form, values were carefully estimated from this graph. GLERL's Large Basin Runoff Model (LBRM, details here) and Advanced Hydrologic Prediction System were used to generate runoff and resulting lake levels. The base period for these projections is 1961-1990. Hayhoe et al developed projections for 3 time periods: 2010-2039, 2040-2069, and 2070-2099.

Angel and Kunkel

This study used global climate model results from the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report, AR4. Three emission scenarios (high, intermediate, and low) were chosen and many runs were completed for each scenario; 565 in all. Between 18 and 23 general circulation models was used for each scenario. GLERL's AHPS was used to determine final lake levels. The base period for this study was 1970-1999. Projected lake levels were reported relative to the average lake level from the base period. Only the extreme emissions scenario projections are reported. The range shown is from 25th (bottom) to 75th (top) percentile.

MacKay and Seglenieks

This paper proposes a new method for estimating future net basin supplies and lake levels based on a bias-correction method that enables the direct use of general circulation model output to run evaporation and runoff models, preserving the land surface-atmosphere feedback loop. The general circulation model used here is the Canadian Regional Climate Model (GLRCM). The base period for this study is 1962-1990. Projections are made for the period 2021-2050. Results for the upper lakes are given in terms of lake level change relative to the base period.