Something strange is happening across the Great Lakes.
Ice cover is declining, water levels are swinging harder, harmful algal blooms continue to threaten Lake Erie, and the offshore food web in Lake Michigan has been transformed by invasive quagga mussels. At the same time, juvenile silver carp have been found closer to Lake Michigan than ever before.
In this DiscoverTube investigation, we examine NOAA research, Great Lakes ice loss, water-level changes, Lake Erie algae, invasive species, Lake Michigan ecology, silver carp, climate variability and the monitoring systems scientists use to track it all.
More than 30 million people depend on the Great Lakes for drinking water—but the system is changing in ways researchers are still trying to understand.
Subscribe to DiscoverTube for more hidden geology, extreme weather, environmental change and scientific discoveries across America.
The Great Lakes system is facing severe ecological instability driven by climate change and invasive species. Key issues include declining winter ice, volatile water levels, and nutrient runoff. The collapse of keystone species like Diporeia and the threat of silver carp further jeopardize the food web, while budget cuts to monitoring infrastructure hinder essential data collection.
Cliff Steinberg says water levels got so low this summer in this stretch of the river
that he could see waterlogged ghost towns. Look at the middle of North America from above,
and you see five connected basins stretching more than 750 miles from west to east,
holding a shoreline that runs roughly four and a half thousand miles. That's more water than
almost anywhere else on the planet, about five and a half thousand cubic miles of it,
sitting in rock that ice sheets carved out and meltwater filled. Stories like this are what we
dig into on DiscoverTube, where our documentaries go past the headline and keep you updated as new
findings land. The basins the ice left behind. Before you can understand what's changing,
you have to understand the shape of the thing that's changing.
Successive ice sheets ground their way south across central North America,
gouging five enormous depressions out of the bedrock. As the ice pulled back roughly 10,000
years ago, meltwater filled what it left behind. The lakes are, in a very literal sense,
the fingerprint of a glacier. Measured at low water datum, the whole system holds about 5,439
cubic miles of water, slightly less than Lake Baikal in Siberia, which is deeper,
but far smaller in surface area. Spread across the surface, though, nothing on Earth compares.
These five lakes account for about 21% of the world's fresh surface water. But they are not
one uniform body of water, and that difference matters more than almost anything else in this
story. Lake Superior drops to about 1,330 feet at its deepest and holds roughly 2,900 cubic miles.
Lake Erie bottoms out at around 210 feet and holds about 116. Superior alone could fill the
other four lakes and still have room left for three more Lake Eries. Depth governs everything
downstream. A deep, cold basin absorbs heat slowly, stores it for a long time, and mixes
reluctantly. A shallow basin heats fast, mixes fast, and responds to a single hot summer or a
single wet spring almost immediately. That's why the five lakes are changing at five different speeds,
even though they share the same weather. There's also a quirk of geography most people never
register. Lake Michigan and Lake Huron are joined at the Straits of Mackinac, and hydrologically,
they behave as a single lake. The water level rises and falls in both at once.
What makes all of this fragile is how little of it turns over. Less than 1% of the water in the
system leaves it in a given year. On any human timescale, this is effectively a closed inheritance.
What changes in it stays changed, and the surface of that inheritance is shallow enough and cold
enough in winter to freeze. That annual lid of ice is the first thing in the whole system that
measurably started to shift. Winter is the first thing to go. Ice on the Great Lakes is not just
scenery. While it sits there, it does physical work. It shields shorelines from storm surge
and high waves, it caps evaporation, and it delays the moment when sunlight starts warming the water
underneath. Since satellites began watching these lakes, the long-term trend has been a decline of
roughly 5% per decade in ice cover. That number sounds modest until you follow what it does to
the rest of the year. Start with the heat. Incoming solar energy has to melt ice before it can warm
the water beneath it. Take the ice away and the lake begins absorbing that energy weeks earlier.
It stratifies earlier, holds more heat deeper into autumn, and arrives at the next winter
warmer than it otherwise would have. The system partly feeds itself. Now go the other direction.
Open water in winter evaporates. That moisture has to go somewhere, and downwind of the lakes
it comes down as snow. Less ice means more lake effect precipitation and more of the hazardous
winter weather that comes with it. The shoreline consequences are the most visible. Ice acts as
armor along the coast. Without it, winter storms hit exposed sand, bluff, and infrastructure directly,
and coastal erosion accelerates. Add to that the loss of winter recreation and the shifting of
habitat for species that depend on cold, ice-covered water for spawning and survival.
Here's the part that complicates the story. The decline hasn't been a smooth slope. Researchers
describe it instead as abrupt shifts. The ice record holding steady, then stepping down suddenly
and staying there. On Lake Superior and Lake Huron, those shifts lined up with a strong El Nino event
in the late 1990s. But that correlation doesn't fully explain them, and nobody has produced a
mechanism that accounts for the timing on every lake. Ice cover also feeds directly into the rest
of the system. It influences regional weather, water levels, how currents move, how the water
column layers itself by temperature, and when the spring plankton bloom starts. Get the ice wrong,
and you get the following summer wrong, too. Which leaves an obvious problem. To know whether
these sudden steps down are genuinely new behavior or something the lakes have always done, you'd
need a record that goes back further than the satellites. And routine satellite observation
of Great Lakes ice only spans about 50 years. That missing record is exactly what a team in
Ann Arbor went looking for. Reading ice in ship's logbooks. Researchers at the Cooperative
Institute for Great Lakes Research, working with NOAA's lab in Ann Arbor, set out to rebuild that
missing record from whatever evidence had survived. They started with air temperature,
because thermometer records around the basin reach back much further than any ice measurement does.
By tying those readings to the decades of ice data satellites had already supplied,
they built a statistical relationship, then ran it backward to estimate how much
ice each of the five lakes carried, year by year, all the way to 1898.
A reconstruction like that is only as good as its checks. So the team went hunting for
independent accounts of those winters, pulling old newspaper reports of harbors freezing and
ferries halting, and reading logs kept by ship captains who had every professional reason to
record exactly when and where they hit ice. Where those accounts described a hard winter,
the model tended to agree. What came out of it changes how you read the trend. The ice does not
simply track the warming climate in a straight line. It appears locked on to slow, powerful cycles
in the oceans and atmosphere, the kind that turn over across decades rather than seasons,
and steer weather across the continental interior. Those cycles have troughs and crests,
and the reconstruction places Great Lakes ice near the bottom of a trough right now.
That leads to a projection which sounds at first like good news. The researchers wrote that a pause
in the ice decline, or even a moderate and short-lived increase, might be expected as those
cycles swing the other way. Be careful with it. This is a projection, not a measurement,
and the authors were explicit about the distinction that matters. A pause is not a
recovery. The warming trend continues underneath, so a run of colder winters would be the natural
cycle briefly pushing against it rather than canceling it. And when the cycle turns again,
the decline resumes from a lower starting point. One of the co-authors came to the project through
the Great Lakes Summer Fellows Program, having grown up on the Minnesota shore of Lake Superior.
Watching the ice bridge out to the nearby islands form later and later each winter.
That's the same signal the model reconstructed, measured with a childhood instead of a satellite.
For the people who work these lakes, the difference is operational. Shipping companies
plan hauling seasons around ice forecasts, and shoreline towns budget for storm damage
based on whether they expect ice armor in midwinter. If you want more documentaries
that take a headline finding apart and explain what it actually says, subscribe to DiscoverTube.
The Lakes That Can't Hold Still
There's a second reason that ice projection matters, and it runs through evaporation.
Open water in winter loses moisture to the atmosphere, and evaporation is one of the
terms that sets how high the lakes sit. Lake Michigan has spent the past dozen years showing
how unstable that balance has become. It fell to record lows, then climbed to record highs in 2019
and 2020, crossing from one extreme of the instrumental record to the other in under a
decade. This past winter, it dropped below its long-term average again after a dry autumn and
unusually heavy winter evaporation, and the Army Corps' spring outlook had Michigan and Huron
running about a foot below normal into late summer, with Lake St. Clair a few inches under.
Then it reversed. Heavy spring snowmelt and intense rainfall pushed water back into the
basin, and by late summer, the Corps was reporting levels generally above average for the season.
A scientist at the Great Lakes Research Institute described that earlier decline
as not especially steep. The trajectory that concerns them is different.
The system is becoming more variable, swinging further in both directions than planners have
historically had to assume. One caveat on the numbers themselves. Published lake levels are
basin-wide still water averages, so on any given shoreline, a strong wind can pile water higher
or push it lower. The five lakes absorb these swings at wildly different speeds,
and that comes straight back to depth. Lake Erie replaces its entire volume in under three years.
Lake Superior takes about a century and three quarters. Same weather system, two completely
different response times, which is why the shallowest lake is where trouble shows up first.
Green Water in the Shallowest Lake
Every spring, the Maumee River drains a broad stretch of farmland in northwest Ohio,
and empties into the far western end of Lake Erie, carrying phosphorus with it.
Because that western basin is shallow and warms quickly, the phosphorus lands in exactly the
conditions cyanobacteria need. NOAA builds its seasonal bloom forecast around that single
river. Scientists combine discharge data from the U.S. Geological Survey and the National
Weather Service with phosphorus concentrations measured at Heidelberg University's Water
Quality Lab, covering the loading season from early March through the end of July.
The total bioavailable phosphorus delivered in those months is the best predictor anyone has
of how large the bloom will get. The output is a severity index scored on bloom biomass
across the peak 30 days. 3 to 5 is moderate, above 5 is more serious,
and above 7 means extensive surface scum. This season's forecast came in at 3.5,
with a range of 3 to 4.5, and NOAA stated plainly where that uncertainty comes from.
Phosphorus already sitting in the lake from previous years,
and competition from other algal species. The forecast verified.
By late summer, cyanobacteria stretched from Stony Point, Michigan down to just west of
Marblehead, Ohio, reaching into the western basin toward the islands, with a heavy load
of Maumee River sediment visible in Maumee Bay. Toxins were detected below the recreational limit.
The longer-term mechanism behind those blooms has two parts. Heavy rainfall events have grown
more frequent, and when rain arrives hard and fast, soil has less time to absorb it,
so more runoff reaches the river. Layered on that, shifts in agricultural practice increased
the share of phosphorus arriving in the readily bioavailable form algae use most efficiently.
This basin supports roughly a quarter of Canadian agricultural production,
and 7% of American farm output, so that nutrient source is not incidental.
THE FOOD DESERT NOBODY PREDICTED
Down on the bottom of Lake Michigan, in water too deep and dark for anything to grow,
there used to be an animal about half an inch long, called Diaperia.
It was a shrimp-like amphipod, living in the soft sediment across most of the deep basin.
It ate the algae that sank down from the sunlit layer above, converted that into fat,
and then got eaten by nearly everything the fishery depended on.
At its peak, the seafloor held more than 7,000 of them per square meter.
Thousands of animals, packed into a patch of lakebed the size of a small tabletop,
repeated across hundreds of square miles. They were the hinge between the surface and the fish.
Today, they are functionally gone from most of the lake.
That collapse coincided with quagga mussels spreading out of the shallows and into the cold,
deep water. Quaggas filter algae straight out of the water column before it can sink.
Within a few years of that expansion, researchers measured offshore primary production,
the total new algal growth the open lake makes each year, falling by roughly 35%.
So the obvious explanation is starvation. The mussels intercepted the food,
and the animals waiting on the bottom for it went hungry.
The evidence partly supports that.
Sediment core work showed that before the invasion,
Diperia fed almost exclusively on three fatty, high-energy types of diatom.
During the invasion, their diet shifted to smaller, less nutritious diatoms,
and the young ones stopped making it to adulthood.
But the story refuses to close there. Other evidence conflicts with the starvation model
and suggests food limitation may not be the direct link at all.
Diperia disappeared from places where algae remained available,
and persisted in others where it should have vanished.
The exact mechanism by which mussels drove them out is still unknown.
A documented, basin-wide ecological collapse with a cause nobody has pinned down.
What happened above them is clearer.
Prey fish that fed on Diperia — alewife, bloater, deepwater sculpin, lake whitefish —
showed measurable declines in condition, growth rate, or energy density.
In offshore Lake Huron, prey fish biomass fell by the great majority over roughly two decades.
And yet the top of the food web did not fall with it.
Salmon and trout biomass stayed high through the same period,
and lake whitefish remained historically abundant.
The collapse hollowed out the middle rather than the top,
which is part of why it went unnoticed for so long.
There's a visual paradox in all of this.
The water in Lake Michigan is now strikingly clear, and clear water reads as clean water.
That clarity is partly the signature of an offshore food desert.
Fewer particles suspended because there is less life producing them.
What have you noticed changing in the water where you live?
Tell us in the comments because those local observations
are exactly the kind of thing we chase down for DiscoverTube documentaries.
80 miles from Lake Michigan
This past summer, crews from the Illinois Department of Natural Resources
pulled juvenile silver carp out of the Marseilles Pool on the Illinois River System,
about 80 miles downstream of Lake Michigan.
That is the closest this species has ever been recorded to the lake.
The word that matters there is juvenile.
Adult carp wander.
Young fish mean spawning somewhere upstream of where they were caught.
State crews went straight back out with electrofishing boats,
working the shallow habitat where juveniles gather,
trying to establish whether these fish were produced nearby
or carried in from farther down the system.
Silver carp escaped captivity in the American South in the late 1970s
and have been moving up the Mississippi and its tributaries ever since.
They typically reach around 20 pounds, though they can grow to 80.
They breed fast and filter enormous volumes of plankton,
putting them in direct competition with native fish that depend on the same food.
The Alliance for the Great Lakes puts Great Lakes fishing at $5.1 billion
and recreational boating at $16 billion.
The last real choke point on that route is the Brandon Road lock and dam at Joliet, Illinois.
The Army Corps of Engineers is building a layered defense there,
an electric barrier, an underwater bubble curtain,
acoustic deterrents designed to stop fish swimming upstream
while keeping the channel open to barge traffic.
Then the Corps issued a stop-work order and put the project under administrative review.
No reason was given publicly.
Backlash came quickly from governors,
members of Congress and advocacy groups across the basin.
And within weeks, the order was lifted and construction resumed.
The reason for the pause has still not been explained.
That episode landed on a monitoring network already under strain.
The Federal Great Lakes Lab in Ann Arbor lost roughly 40 percent of its staff
over the previous year through layoffs and early retirements.
And a budget proposal would cut about a third from NOAA's funding,
including programs that underpin Great Lakes weather observation,
water quality and maritime safety.
That is a proposal, not an enacted cut.
What's at risk is the measuring apparatus itself.
More than 77 buoys, meteorological towers and radar units
feeding continuous wind, wave and water temperature data,
alongside newer techniques like reading storm waves off fiber-optic cable laid on the lakebed.
As one observing system director put it,
Each lapse makes the next one worse.
And rebuilding isn't just a matter of writing another check.
The Inheritance
Five basins carved by ice,
holding water that effectively doesn't renew itself.
Everything here comes back to that.
The honest answer to what's happening is not a single accelerating collapse,
it's variability.
Ice may pause its decline without recovering.
Levels swing from record low to record high and back within a decade.
Erie's bloom arrives roughly on the schedule the phosphorus predicts.
While out in the deep water, an entire keystone species vanished,
for reasons still unresolved.
Documented change, undetermined cause.
What to watch next is straightforward.
Whether the reconstructed ice record holds up against the coming winters.
Whether Brandon Road gets finished before the carp get closer.
And whether the network measuring all of it survives intact.
If this is the kind of story you want more of, subscribe!
And tell us in the comments,
which of these changes should DiscoverTube investigate in full next?