Important quotes from this document:
“The myriad sea walls, breakwaters, groins, and jetties that line developed shorelines divert offshore, slow down, trap, and otherwise reduce the regional beach sediment supply by longshore currents, and thereby increase erosion rates.” This view is summed up on the cover of one edition of the book “The Beaches are Moving”, by Kaufman and Pilkey, as follows, “these facts provide a new ability to make informed, intelligent decisions for the coast. This is the first book to explain why the shore must move, and how utterly foolhardy we are to armor our coast against the unbeatable force of nature.”Engineers do, however, recognize the limitations of their solutions.
“The Lake Winnipeg Shoreline Handbook”, states, “In most cases a careful review of long term erosion rates, the effects of protective structures on beaches, and the cost of protective structures in relation to the cost of land will indicate that the best course of action is to allow natural erosion to continue.”
Dr. Harvey Thorleifson’s Essay on Shoreline Erosion and Lake Winnipeg
The principal conclusion that I took from the meeting was that it will be very important for there
to be much greater recognition of the distinction between the natural component of shoreline
erosion, and the component that can be attributed to lake level regulation managed by Manitoba
Hydro. To my knowledge, it has not yet been demonstrated whether the latter component is
positive, negative, variable, or negligible.
I have heard verbal claims that shoreline erosion on Lake Winnipeg is entirely due to regulation.
It is this sort of claim that results in a ‘History vs. Hydro’ debate, implying that the cause is one
or the other. I have and will be very quick to side with the ‘History’ side of any debate structured
in this manner, due to our familiarity with the long-term history of Lake Winnipeg basin
expansion, and the several natural mechanisms, primarily uplift, that are driving this expansion.
The most readily available formal measurement of the natural component of shoreline erosion,
that I am aware of, is the Penner and Swedlo study completed in the 1970s, which was based on
surveys extending back to the 1870s, vertical aerial photographs, and ground surveys. Their
summary of pre-regulation natural shoreline retreat rates, typically 0.5 to 5 m/year in the south
basin, was reprinted in ‘The Lake Winnipeg Shoreline Handbook’.
I am not aware of a formal assessment of whether regulation has changed shoreline retreat rates,
although I am aware that many people have opinions on the matter. I can think of two
approaches that could be taken to this question. One would be to measure the rate of shoreline
retreat since regulation, and compare this to the pre-regulation rates. This assessment would
primarily rely on ground surveys and vertical aerial photographs, with shoreline referenced to
cliffs and vegetation limits, rather than the water line that is vulnerable to large short-term
fluctuations. Another approach would be for predictions of shoreline retreat rate to be made on
the basis of pre- and post-regulation lake level regimes. This would best utilize a detailed model
of Lake Winnipeg shoreline processes, which is not yet available. In the absence of a locallybased
model, more generic models could be used. You may be interested to note, however, that,
in the 1992 SEPM Special Publication No. 48, the leading shoreline erosion authorities Pilkey
and Thieler state, “It is doubtful that any existing models can predict shoreline erosion rates with
an accuracy useful to coastal communities.” This does not, however, mean that we should back
off from doing the research that is required to work toward a useful model, especially one
specific to Lake Winnipeg.
There also was discussion regarding current lake level and lake level management operating
rules. What impact short term changes to lake level management would have on shoreline
erosion rates in the short or long term is a topic we are not in a position to evaluate, although I
would think that a deviation of mean lake level, averaged over a few years, from the long term
mean, would to some degree influence shoreline retreat rate.
I also noted recognition that issues of lake level regulation are also relevant to topics such as
wildlife habitat, although we have no role in the topic.
I was asked at the meeting whether one could predict what lake level would have been without
regulation. In my opinion, this analysis could readily be done. For example, one could examine
the relationship between lake level and mean monthly inflow from the four major rivers. Pending
confirmation of a good correlation, one could use mean monthly river inflow data, with
appropriate lags, to estimate what lake level would have been without regulation, whether higher
or lower. It might be found, however, that climatic factors that do not manifest themselves in
river flow data may be significant. This is not, however, an analysis that is within the mandate of
our present work.
On more than one occasion, I have heard the argument that Lake Winnipeg lake level must be
artificially high at present, because very old trees are being felled by shoreline erosion. The
invalid assumption behind this argument is the notion that the natural state of Lake Winnipeg is
stability. These observations can readily be explained by the fact that this lake has been gradually
expanding for 8000 years, so you could have a 5000 year old tree that simply will be taken down
when the lake finally arrives and that tree’s time has come. There is no ‘natural’ level for Lake
Winnipeg that has now been perturbed by human interference. The natural state of this lake is for
it to constantly expand, at least for another few thousand more years.
We believe that lake level rise, largely driven by uplift, is the principal driver of shoreline
erosion. At the meeting, I was, quite justifiably, challenged to explain how a subtle lake level
rise, which we estimate to be 20 cm/century (2 inches in 25 years), in the south basin, could
possibly cause the devastation that property owners have witnessed. I agree that this is difficult
to fathom, but in this area of low land gradients, a slight vertical displacement leads to a large
lateral displacement when the shore profile adjustment is complete.
In order to understand this and every other process at work on Lake Winnipeg, it is essential to
have some insight into how the lake works. I believe that this topic can be understood by anyone,
not just a geologist, so I will now endeavour to lay out our reasoning.
Lake Winnipeg lies at the interface between granites and related rocks to the east, and
sedimentary rocks such as limestones to the west. The granites formed 2 to 3 billion years ago.
The limestones formed about a half billion years ago, when central North America was covered
by a shallow tropical sea in which debris such as shells and corals accumulated, including the
fossils we see in Tyndall Stone. The processes that formed these rocks no longer play a role in
the evolution of Lake Winnipeg.
The Ice Age, also known as the Pleistocene, is geologically very recent. During the peak of the
most recent glacial cycle, between 10,000 and 20,000 years ago, Canada was covered by a
glacier similar to the continental ice sheets that presently cover Greenland and Antarctica. Ice
flow radiated from the Hudson Bay region, and this ice flow scoured out the Lake Winnipeg
basin as we know it.
As the continental ice sheet was reduced in size by climatic change at the end of the Ice Age, the
land that slopes toward Hudson Bay in the Red River Valley filled with water due to the
presence of the ice barrier to the north. This formed Lake Agassiz, which was in existence in an
ever-evolving form between about 11,000 and 8000 years ago. It was in Lake Agassiz that the
clay soils of the Winnipeg region were deposited. When the glacier finally was split into two
remnants that both soon melted, by the formation of icebergs in Hudson Bay, Lake Agassiz
The continental ice sheet was about 4 km thick over Hudson Bay. The surface of the earth
basically floats on the interior of the earth, so accumulation of this ice mass depressed the
surface of the earth by about a km. As the glacier began to wane due to a shift from a positive to
a negative balance between snow accumulation and loss due to melting and formation of
icebergs, its mass was reduced and eventually removed. Removal of this much weight is like
taking a load out of a boat, and the surface of the earth rose. Much of the uplift took place under
the glacier, or soon after its withdrawal.
Several observations indicate that the 8000-year period since deglaciation has not been enough
time for the earth to adjust to removal of the glacier. Around Hudson Bay, there are many marine
shorelines that have been left behind by retreat of the bay due to uplift. The age of these
shorelines can be determined by radiocarbon dating of shells found in the gravels of these fossil
beaches, and in other deposits. The highest shoreline around Hudson Bay dates to about 8000
years, but those closer to the bay date to about 1000 years and younger. This indicates that retreat
of the bay has continued in recent centuries. The fact that the uplift continues today is indicated
by observations such as results from the Churchill tide gauge, where high-quality data collected
since 1940 indicates that sea level at that site is retreating at a rate of about 0.7 m/century.
Allowing for global sea level rise of about 20 cm/century in recent decades, this allows the uplift
rate to be rounded off to about a m per century. The pattern of subtle trends in the strength of
gravity across Canada supports these conclusions and indicates, along with information from the
Great Lakes, that the general trend in uplift is for rates to diminish inland from Hudson Bay in all
Because the rate of uplift diminishes inland from Hudson Bay, a tilting action results. We know
that the Lake Winnipeg region was tilted after the retreat of Lake Agassiz, because the shorelines
of Lake Agassiz, which would have been horizontal at the time of their formation, now rise in
elevation toward the northeast. Hence for at least much of its history, the Lake Winnipeg basin
has been rising, and the north end has been rising more rapidly than the south end.
A clear discussion of the influence of tilting on a large lake requires a review of the natural
mechanisms that control lake level. An open container of water such as a lake undergoes
fluctuations in its level as water is gained and lost. The volume of a lake does not determine lake
level; volume is a result of lake level. Input of water to a lake occurs in the form of river inflow,
direct precipitation, and groundwater discharge from underwater springs. Losses include river
outflow, evaporation, and groundwater recharge as seepage into the lake bottom. If inflow is
greater than losses due to evaporation and groundwater recharge, the lake has a water surplus,
and excess water is evacuated by outflow at the outlet(s). If evaporation and groundwater
recharge together exceed inflow, the lake has a water deficit, and no outflow will occur.
Hence the water budget of a lake is dictated chiefly by climate, with secondary effects related to
groundwater. In the case of a lake with no outflow, a closed basin with a negative water budget,
lake level is purely a result of climate. Examples of closed lakes are Great Salt Lake in Utah and
Devil’s Lake in North Dakota. At present, however, Lake Winnipeg has a large water surplus.
Water primarily derived from the Winnipeg and Saskatchewan Rivers is evacuated by the Nelson
River at a rate of about 60 cubic km per year, a large flux compared to the small volume stored
in the lake, about 300 cubic km. Lake Winnipeg therefore is governed by processes related to a
positive water budget. Secondary, short-term effects on lake level are caused by wind setup, and
to a lesser extent, barometric pressure.
In the case of an outflowing lake with a positive water budget, lake level is controlled by the
combination of climate and outlet geometry. Climate over the drainage basin determines how
much excess water there is to be evacuated. Lake level has to reach at least the elevation of the
lowest point on the topographic barrier around the lake, hence the bed of the outlet stream.
Above this level, an additional depth is required for outflow to be adequate to evacuate the
surplus water. A narrow outlet channel requires more depth than a broad outflow to achieve a
given flow rate. This is called the stage (water level) vs. discharge (volumetric rate of flow)
relationship for the outlet.
Lake Winnipeg has, at least in recent millenia, been an outflowing lake. The mean lake level
therefore is constant at the outlet relative to mean climate of the time, given that a certain depth
is required to evacuate excess water. Tilting of a lake basin causes mean lake level to pivot at the
outlet. Because the outlet of Lake Winnipeg is in the north, uplift of the north end of the lake
progressing at a rate more rapid than the basin to the south has meant lake level rise over the
entire basin, with the rate increasing southward.
Not all lakes rise and expand due to tilting, however. Lake Nipigon has its outlet in the south, so
it is contracting. Lake Superior has its outlet in the middle relative to the pattern of uplift, so it is
rising in the south and falling in the north.
We have collected cores from the bottom sediments of Lake Winnipeg that allow us to sample
the entire sequence of sediments deposited since Lake Agassiz, including the first layer of Lake
Winnipeg sediments that buried the older Lake Agassiz deposits. We have obtained radiocarbon
ages from this procedure that indicate that much of the South Basin of Lake Winnipeg was dry
land 4000 years ago, while Netley Marsh was dry land about 1500 years ago. We also have
radiocarbon dates from rooted tree stumps just below lake level that suggest gradual rise in lake
level over recent centuries. These observations indicate that gradual expansion of Lake Winnipeg
in response to tilting has been continuous throughout post-Lake Agassiz time.
While we place our emphasis on uplift, which has been the dominant influence, at least in the
south, four other process should be mentioned as secondary factors affecting Lake Winnipeg lake
level over the long term: climate, river diversions, basin merging, and outlet downcutting.
Our radiocarbon dating of basal Lake Winnipeg sediments in cores indicates that, unlike the
gradual inundation of the rest of the lake, the inundation of the central South Basin was not
gradual. It seems to have occurred rapidly, as basal ages across this area cluster around 4000
years. This is the time when climate changed rather abruptly from warmer and drier to cooler and
moister, probably raising lake level a few metres. This is the reverse of the trend that we may
presently be experiencing due to the emission of greenhouse gases. In fact, the worst case
scenario for climate change now and in future decades would be a return to the climate of 5000
years ago, when the prairie region was much drier. Climate of the Lake Winnipeg region has
been relatively stable in the past 4000 years, however, so the impact of this climate change
would have been applied rapidly, with control of lake level evolution to the present day returning
to uplift dominance.
Another factor in Lake Winnipeg lake level history was diversion of the Saskatchewan River,
which formerly bypassed Lake Winnipeg in the channel now occupied by the Minago River.
Between 4000 and 5000 years ago, uplift caused diversion of the Saskatchewan River to Lake
Winnipeg. This would have raised lake level on a one-time basis by about a half metre.
At present, Playgreen Lake and Lake Winnipeg are almost functioning as one lake. Strong
northward currents typically flow through Warren Landing, in what could almost be considered a
Narrows rather than a river, feeding the Nelson River to the north. When strong north winds
blow, however, flow at Warren Landing can be to the south. But for a few millenia after Lake
Agassiz, however, what is now Lake Winnipeg was three or more lakes, a South Basin lake
draining through a river in the Narrows to a North Basin Lake, which in turn drained to a
completely separate Playgreen Lake. All of these lakes expanded in response to tilting, and
eventually the North Basin and South Basin lakes merged. Relocation of the outlet for the South
Basin Lake to a point farther north, where uplift is more rapid, would have accelerated lake level
rise in the south. More recently, perhaps about 2000 years ago, Playgreen Lake merged with
Lake Winnipeg, again increasing the rate of lake level rise and lake expansion in the South
Basin, once again renewing the otherwise gradually diminishing rate of rise.
Outlet down-cutting is a factor that seems not to be a significant control on Lake Winnipeg.
Whereas this was the dominant factor in controlling the early history of Lake Superior, the outlet
of Lake Winnipeg at Warren Landing is shallow and broad, and would have been rapidly eroded
to resistant bedrock. Therefore while this could have been a compensating factor offsetting the
rise due to uplift, it seems not to have played a role.
Maps showing the pattern of present uplift on the world and continental scales may be seen on p.
117 of the March 1997 issue of Scientific American, and on the cover of the 4 December 1997
issue of Nature. General models such as these, based to varying degrees on continent-wide
syntheses of radiocarbon-dated marine shorelines, tide gauge trends, lake gauge trends, and
gravity, give a rough estimate for uplift rates of 0.4 m/century at the north end, and 0.2
m/century at the south end of Lake Winnipeg. The difference between these two values implies a
20 cm/century rise in lake level at the south end of Lake Winnipeg.
This prediction can be tested by comparison to available data from Lake Winnipeg. Offshore
from Gimli, at our site 122, the pre-Lake Winnipeg surface lies under 10 m of water and 4 m of
sediment. We have dated the initiation of Lake Winnipeg sedimentation at this site at about 4000
years. A rise of the lake to its present level over the past 4000 years implies a rate averaging 35
cm/century (1400 cm/40 centuries). This would be an average of higher rates earlier in the period
in question, and lower rates at present, perhaps comparable to the current estimate of 20
cm/century. According to Penner and Swedlo, a 40-cm-thick peat bed found 3 m below lake
level near Elk Island was dated at 1060 years for the upper part of the bed and 1660 years for the
lower part. Interpolating between the upper date and present lake level gives an estimate of 28
cm/century (300 cm/10.6 centuries) for lake level rise over the past millennium. Work by Dr.
Erik Nielsen, of the Manitoba Geological Services Branch, on the radiocarbon age of drowned
stumps in the Lake Winnipeg shoreface also indicates a submergence rate of about 20
cm/century over the past 300 years. Hence available data are strongly supportive of the lake level
rise predicted by uplift models.
Even without this sort of data, the experienced eye can quickly see that water levels are rising on
Lake Winnipeg. For example, geologists now agree that barrier islands are a sign of water level
rise. The sandy beach that separates the south end of the lake from Netley Marsh is a barrier
island. Other good examples can be seen on Lake Manitoba, the east coast of the US, Duluth,
Hamilton, and northwestern Europe. The geological model for how barrier islands work is for
there to be erosion on the basin side, and accretion on the lagoon side. In other words, the natural
behaviour for a barrier island is for it to migrate landward like a conveyor belt. One can also
recognize water level rise on Lake Winnipeg in the form of drowned valleys, also known as
estuaries, such as lower Netley Creek and lower Icelandic River.
Even if Hudson Bay is still being uplifted and the Great Lakes are still being tilted, and even if
there is evidence for Lake Winnipeg having expanded in recent millennia, centuries, and
decades, this does not prove that Lake Winnipeg is presently still being tilted. Complexities in
the uplift pattern could have formed in recent time. Lake gauge data, however, have provided
indications of present-day uplift. This takes the form of a gradual increase in the difference
between southern gauges and northern gauges over several decades. We also are investigating
this topic with new approaches. In cooperation with NASA, we have installed two new Global
Positioning System satellite receiving stations at Pinawa and Flin Flon, that will, in combination
with existing stations in Iowa and Churchill, give us measurements of uplift rates. In cooperation
with the US government, we also are doing very sensitive measurements of gravity along a
transect of sites from Iowa to Churchill that will give us an independent check on uplift or
The 1974 Penner and Swedlo report supplemented existing knowledge of shoreline erosion rates
with information from surveys done at intervals of one to a few decades from the 1870s to the
late 1960s. It was found that the shoreline of the South Basin retreated over this period at rates
typically of 0.5 to 5 m per year. An average rate of, for example, 1 m per year could, of course,
represent 10 m in one year and no recession for 9 years. Can this steady rate of shoreline erosion
be explained by a 20 cm/century rise in lake level? Let’s relate the 20 cm/century rise to regional
topographic gradients. At Gimli, the land rises about 25 m within 10 km inland, a gradient of 2.5
m/km. In this case, a 20 cm/century lake level rise would translate to a lateral shift of 0.8 m/year,
similar to actual shoreline erosion rates reported by Penner and Swedlo. From the centre of the
south basin to Netley Creek, the surface under Lake Winnipeg sediments rises to the present land
surface at a rate of about 0.3 m/km. A 20 cm/century lake level rise in this case translates to a
lateral migration of about 6.7 m/year. This estimate is compatible with our data offshore from
Gimli that shows that the south end of the lake has migrated 30 km to its present position in 4000
years, implying an average rate of shoreline retreat of about 7.5 m per year. This agreement is
surprisingly good, considering that we have not yet built a more detailed model that takes into
account the role of climate in expansion of the lake. Penner and Swedlo reported similar retreat
rates over much of the southern shore.
Large increments of basin expansion being driven by a few inches of lake level rise may seem
counter-intuitive. A one-metre rise in lake level happens frequently due to wind setup, and the
water line only moves a few metres. But according to the above reasoning, a one-metre
permanent rise in lake level will drive the shoreline inland 400 metres to the west, and over 3 km
to the south. How can this apparent contradiction be reconciled? The key point is that shoreline
processes have cut a notch at the water line that has a much higher gradient than the surrounding
landscape. Penner and Swedlo indicate that the gradient between the high water and low water
line on Lake Winnipeg typically is about 10%, or 100 m/km. It is this slope that takes up the
short term fluctuations. The steeper nearshore gradient can also be seen on the hydrographic
chart for the south basin. Around Gimli, the offshore gradient is about 3.4 m/km between the
shore and ten feet depth, while farther offshore, the gradient is less than 1 m/km. Along the south
shore, the gradient to ten feet depth averages 1.2 m/km, while farther offshore it is about 0.25
m/km. Hence short term fluctuations are taken up by the high gradient slope at the water line, but
a permanent rise exposes that slope to a sustained increase in wave power. In the case of a onestep
lake level rise, the shoreline would retreat and the shore profile would flatten until wave
power delivered to the shore diminishes to a level that allows a stable coastal position. In the
case of a steady, ongoing rise, a steady retreat of the shore results. Even if a steady rise were to
stop, retreat would continue until equilibrium is reached.
It is useful to compare shoreline erosion on Lake Winnipeg with global trends at sea level, which
are probably best documented in the US. In SEPM Special Publication No. 48, Pilkey and
Thieler present a summary of erosion rates on the US coast. Values of 0.5 to 4 m/year are typical
of the Atlantic and Gulf coasts. There now is consensus that this erosion is driven by the present
global (eustatic) sea level rise of about 20 cm/century, which happens to be similar to the rate of
lake level rise that we estimate for the south end of Lake Winnipeg.
Shoreline protection engineering is an extremely controversial topic, and it is not within our
mandate to make recommendations or to conduct research into design. I can, however, offer a
few thoughts. Engineers tend to recommend engineering, whereas geologists tend to recommend
that you back off and respect the forces of nature. There is no doubt, however, that a Professional
Engineer can design a means to temporarily protect a high value installation that was, through
error or necessity, built too close to a retreating shore.
Geologists, on the other hand, are quick to point out the inadequacies of shoreline engineering.
While erosion may be halted at the water line, it is essential to recognize that erosion continues
underwater to a depth of several metres. Consequently, the shore armour is exposed to
progressively greater wave power. Furthermore, sand must be mobile for shoreline systems to be
in a state of dynamic equilibrium. According to a standard geomorphology textbook by Bloom,
“Extensive reclamation and beach stabilization projects actually endanger barriers, because
unless sand is free to move with changing wave conditions, erosion results.” In SEPM Special
Publication No. 48, Pilkey and Thieler state, “The myriad sea walls, breakwaters, groins, and
jetties that line developed shorelines divert offshore, slow down, trap, and otherwise reduce the
regional beach sediment supply by longshore currents, and thereby increase erosion rates.” This
view is summed up on the cover of one edition of the book “The Beaches are Moving”, by
Kaufman and Pilkey, as follows, “these facts provide a new ability to make informed, intelligent
decisions for the coast. This is the first book to explain why the shore must move, and how
utterly foolhardy we are to armor our coast against the unbeatable force of nature.”Engineers do, however, recognize the limitations of their solutions. For example, “The Lake
Winnipeg Shoreline Handbook”, states, “In most cases a careful review of long term erosion
rates, the effects of protective structures on beaches, and the cost of protective structures in
relation to the cost of land will indicate that the best course of action is to allow natural erosion
to continue.”You may wish to view some of our work at the following Web address:
I hope that these comments will be of use to you and to the members of your association. I know
that I have presented a lot of information in this email, and I don’t doubt that you may wish to
request clarification of some topics. Please do not hesitate to contact me if I may be of more
(formerly) Geological Survey of Canada, 601 Booth Street, Ottawa ON K1A 0E8
Harvey Thorleifson has been Director of the Minnesota Geological Survey, State Geologist of
Minnesota, and Professor in the Department of Geology and Geophysics at the University of
Minnesota since 2003. He is originally from western Manitoba, he did his undergraduate work at
University of Winnipeg, and he completed a Masters thesis in geology on Lake Agassiz history
at University of Manitoba in 1983.
His 1989 geology Ph.D. at University of Colorado in Boulder dealt with Hudson Bay Lowland
Quaternary stratigraphy. His early career research on Lake Agassiz, the Great Lakes, Hudson
Bay, and North American glacial history evolved to work on indicator mineral methods in
mineral exploration, geological and geochemical mapping, regional groundwater investigations,
shoreline erosion, and flooding while at the Geological Survey of Canada from 1986 until 2003.
Harvey is active with the Association of American State Geologists, he is registered as a
Professional Geoscientist in Ontario, he was the 2004-2006 President of the Canadian
Geoscience Council, now renamed the Canadian Federation of Earth Sciences, he was the 2003-
2004 President of the Geological Association of Canada, he is an Associate Editor of the Journal
of Great Lakes Research, and he was a founding member of the OneGeology organizing
committee, a project to accelerate geologic map web accessibility worldwide.
Harvey Thorleifson Ph.D., P.Geo.,
State Geologist of Minnesota;
Director, Minnesota Geological Survey;
Professor, Department of Geology and Geophysics; University of Minnesota;
2642 University Ave W,
St Paul, MN 55114-1057 USA;
Telephone 612-627-4780 ext 224;