As many of us in the Great Lakes community start to don our parkas and break out the snow shovels, we know the splashing waves on our shorelines will soon be replaced with ice. And, with near-record high water levels in the lakes this year, the question of how ice and water levels will affect coastal communities in the months ahead looms large.
The role of ice in the Great Lakes water budget
To start, we know that evaporation plays a major role in water levels by withdrawing water that enters the lakes from precipitation and runoff. So, high evaporation contributes to lower water levels, and low evaporation contributes to higher water levels. (For more on the Great Lakes water budget, check out this infographic.) Traditional thinking is that high ice cover forms a “cap” that leads to decreased evaporation of lake water. However, we now know that the relationship between ice, evaporation, and water levels is more complex than that.
While this assessment on Great Lakes evaporation from Great Lakes Integrated Sciences & Assessments explains that high ice cover is still associated with less evaporation the following spring, it also reports that evaporation rates before winter have an effect on how much ice forms in the first place. Specifically, it explains that high evaporation rates in the fall correspond with high ice cover the following winter. So just as ice cover can influence evaporation, the reverse is true as well – a much different story than the one-way street it was previously thought to be.
A look at 2020 ice cover: NOAA GLERL’s observations & predictions
On January 1st, 2020, the total Great Lakes ice cover was 1.3%. That’s about a third as much ice as around the same time last year, and barely anything compared to early 2018, when it was already about 20%. You’ll see in the figure below that shallow, protected bays tend to freeze first, especially ones that are located in the northern Great Lakes region. So it makes sense that most of the ice so far is in the bays of Lake Superior, followed by northern bays in Lakes Michigan and Huron like Green Bay and Georgian Bay.
GLERL conducts research on ice cover forecasting on two different time scales: short-term (1-5 days) and seasonal. GLERL’s short-term ice forecasting is part of the upgrade to the Great Lakes Operational Forecast System (GLOFS), a set of models currently being transitioned to operations at the National Ocean Service to predict things like currents, water temperature, water levels, and ice. The ice nowcast and forecast products (concentration, thickness, velocity) have been tested for the past several years and will soon become operational (available for the general public).
GLERL’s ice climatologist, Jia Wang, produces an experimental annual projection for Great Lakes ice cover using a statistical model that predicts maximum Great Lakes ice cover percentages for the entire season. This model’s prediction is based on the predicted behaviors of four global-scale air masses: ENSO (El Nino and Southern Oscillation), NAO (North Atlantic Oscillation), PDO (Pacific Decadal Oscillation), and AMO (Atlantic Multidecadal Oscillation). While they’re all pretty far away from the Great Lakes, past research has shown that these air masses — or global teleconnections — heavily influence the year-to-year variability of Great Lakes ice cover.
Based on this experimental model’s results, NOAA GLERL projects this Great Lakes ice cover this winter to be around 47%. That’s almost 9% below the long-term average of 55.7%. Here’s the preliminary projection broken down by lake:
Lake Superior: 54%
Lake Michigan: 41%
Lake Huron: 66%
Lake Erie: 80%
Lake Ontario: 32%
Updated Seasonal Ice Forecast
On 1/24/2020, GLERL researchers reran the experimental ice forecast model. The reason for the revised ice projection was due to significant deviation in the actual teleconnection indices from what was predicted in November. Because the model uses these predicted teleconnections to predict ice cover, it is important to use the most accurate values. The experimental ice forecast model is rerun to get updated values. NOAA GLERL’s research towards improving our capabilities for ice forecasting is ongoing. This research product continues to evolve as we gain understanding of the complex climatic drivers for the Great Lakes Region.
Lake Superior = 46-54%
Lake Michigan = 23-42%
Lake Huron = 47-72%
Lake Ontario = 16-36%
Lake Erie = 67-74%
Why did the 2020 Great Lakes ice cover forecast change? / What factors go into this forecast?
The experimental Great Lakes ice forecast is initially calculated in November and is based on a model run based on the forecasted teleconnection patterns for December, January and February. If the teleconnection values diverge from the forecast (that is, the climate did not act as predicted) then the experimental ice forecast model is updated with the latest information on expected teleconnection indices for the remainder of the winter.
What does the ice coverage range mean?
Because there is uncertainty with this experimental forecast, two different versions of the model are used. Version “a ” uses only 4 teleconnection patterns (NAO, AMO, ENSO, PDO) as variables (inputs). Version “b” also includes observed November lake surface temperature (LST) as a variable. November and December LST were shown to be equally well correlated with ice cover but by using November LST, it enables the forecast to be made a month earlier.
Ice cover is low right now, what could happen that would increase ice coverage?
A lot can still happen as there are many weeks of winter left. Historically, much of the major freezing happens in February. However, if temperature continues to remain abnormally warm, it is unlikely ice cover would reach these values.
Why is it important to continue this research?
NOAA GLERL continues to refine the ice forecast model, active research designed to improve the Great Lakes ice forecast. We plan to improve the forecast skill by adding the cumulative freezing degree days since December 1, and update the forecast every two weeks throughout the ice season.
Predicting Great Lakes water levels
Forecasts of Great Lakes monthly-average water levels are based on computer models, including some from NOAA GLERL, along with more than 150 years of data from past weather and water level conditions. The official 6-month forecast is produced each month through a binational partnership between the U.S. Army Corps of Engineers and Environment and Climate Change Canada.
Want to know more about GLERL’s ice research? Visit our ice cover webpage for current conditions, forecasts, historical data, and more!
Great Lakes ice cover facts since 1973
94.7% ice coverage in 1979 is the maximum on record.
9.5% ice coverage in 2002 is the lowest on record.
11.5% ice coverage in 1998, a strong El Niño year.
The extreme ice cover in 2014 (92.5%) and 2015 (88.8%) were the first consecutive high ice cover years since the late 1970’s.
NOAA Great Lakes Environmental Research Laboratory 