Subsurface heatwaves in lakes

Abstract

Lake heatwaves (extreme hot water events) can substantially disrupt aquatic ecosystems. Although surface heatwaves are well studied, their vertical structures within lakes remain largely unexplored. Here we analyse the characteristics of subsurface lake heatwaves (extreme hot events occurring below the surface) using a spatiotemporal modelling framework. Our findings reveal that subsurface heatwaves are frequent, often longer lasting but less intense than surface events. Deep-water heatwaves (bottom heatwaves) have increased in frequency (7.2 days decade⁻¹), duration (2.1 days decade⁻¹) and intensity (0.2 °C days decade⁻¹) over the past 40 years. Moreover, vertically compounding heatwaves, where extreme heat occurs simultaneously at the surface and bottom, have risen by 3.3 days decade⁻¹. By the end of the century, changes in heatwave patterns, particularly under high emissions, are projected to intensify. These findings highlight the need for subsurface monitoring to fully understand and predict the ecological impacts of lake heatwaves.

Full text

Nature Climate Change | Volume 15 | May 2025 | 554–559 554
nature climate change
Article
https://doi.org/10.1038/s41558-025-02314-0
Subsurface heatwaves in lakes
R. Iestyn Woolway 1 , Miraj B. Kayastha 2,3, Yan Tong 4, Lian Feng 4,
Haoran Shi 1 & Pengfei Xue 2,3,5
Lake heatwaves (extreme hot water events) can substantially disrupt
aquatic ecosystems. Although surface heatwaves are well studied, their
vertical structures within lakes remain largely unexplored. Here we analyse
the characteristics of subsurface lake heatwaves (extreme hot events
occurring below the surface) using a spatiotemporal modelling framework.
Our ndings reveal that subsurface heatwaves are frequent, often longer
lasting but less intense than surface events. Deep-water heatwaves (bottom
heatwaves) have increased in frequency (7.2 days decade−1), duration
(2.1 days decade−1) and intensity (0.2 °C days decade−1) over the past
40 years. Moreover, vertically compounding heatwaves, where extreme
heat occurs simultaneously at the surface and bottom, have risen by
3.3 days decade−1. By the end of the century, changes in heatwave patterns,
particularly under high emissions, are projected to intensify. These ndings
highlight the need for subsurface monitoring to fully understand and
predict the ecological impacts of lake heatwaves.
Lake heatwaves, prolonged periods of anomalously warm water events,
have recently become distinguishable features of lake temperature
variability. Studies have demonstrated that climate change has driven
a notable increase in the frequency, duration and intensity of these hot
extremes
1
. Model projections suggest that, during the twenty-first
century, lake heatwaves are likely to intensify, become longer lasting
and their occurrence frequency is expected to increase
1–3
. The rise
of unprecedented temperatures during a lake heatwave can benefit
some aquatic species by expanding their thermal habitat
4
, but can
be detrimental for others, particularly to those that live in regions
close to their thermal limit5–7. Quantifying changes in lake heatwaves
is thus critically important to anticipate the likely impact of climatic
warming on lakes.
Previous studies investigating the impacts of climate change on
lake heatwaves have focused on surface conditions1,2. Most notably, lake
heatwaves have been defined, like marine heatwaves8–11, as periods in
which surface water temperatures increase above a seasonally varying
90th percentile threshold
1
. This one-dimensional (temporal) approach
for describing a lake heatwave is important, not only for understand-
ing how extreme lake surface conditions respond to climate change,
but also for anticipating the likely impact of these extreme heat events
on aquatic organisms that cannot or will not move. However, mobile
aquatic species can respond to environmental disruptions, such as
extreme surface temperatures, by relocating to favourable habitats
12–14
.
In stratifying systems, bottom waters are often cooler than the lake sur-
face and, if other environmental factors are favourable, many aquatic
species could migrate to these deeper layers to escape surface thermal
stress15. Moreover, the thermal response of lakes to climate change can
differ considerably between surface and bottom waters, with the latter
often, although not always16,17, experiencing a somewhat muted climatic
response18–20. In turn, cooler water at depth could provide a potential
thermal refuge for aquatic species as surface heatwaves become more
common and intense
1,14
. However, unlike in marine systems
21–24
, the
vertical dimension of lake heatwaves has not yet been considered, thus
limiting our understanding of how lake environments below the water
surface are responding to a more extreme world.
This study aims to fill this knowledge gap by investigating the
depth to which thermal anomalies associated with lake surface heat-
waves penetrate, as well as exploring how this has changed over the
historic period (1980–2022) and will probably change in the future
Received: 14 May 2024
Accepted: 11 March 2025
Published online: 10 April 2025
Check for updates
1School of Ocean Sciences, Bangor University, Menai Bridge, UK. 2Great Lakes Research Center, Michigan Technological University, Houghton, MI, USA.
3Department of Civil, Environmental and Geospatial Engineering, Michigan Technological University, Houghton, MI, USA. 4School of Environmental
Science and Engineering, Southern University of Science and Technology, Shenzhen, China. 5Environmental Science Division, Argonne National
Laboratory, Lemont, IL, USA. e-mail: iestyn.woolway@bangor.ac.uk; pexue@mtu.edu
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change | Volume 15 | May 2025 | 554–559 555
Article https://doi.org/10.1038/s41558-025-02314-0
escape depth has increased ~1.5-fold, from 24.7 to 38.2 m, at a rate
of 4.2 ± 3.19 m decade
−1
(Extended Data Fig. 2). However, the escape
depth has remained unchanged in some lakes (Supplementary Fig. 7
and Supplementary Table 5). The interannual variability in the escape
depth is positively correlated (R2 = 0.7) with the intensity of lake surface
heatwaves (Extended Data Supplementary Fig. 3a and Supplementary
Fig. 8a–c). On average, the escape depth is also positively related to the
depth of the upper mixed layer (Methods) (Extended Data Fig. 3b and
Supplementary Fig. 8d).
Subsurface and lake bottom heatwaves
Subsurface heatwaves, instances of extreme hot thermal events at dif-
ferent depths relative to a depth-specific climatological seasonal cycle,
follow a clear vertical pattern of variability in stratified lakes (Fig. 2,
Supplementary Fig. 2 and Extended Data Fig. 4). The vertical structure
of subsurface heatwaves can differ seasonally (Extended Data Figs. 5
and 6), typically occurring more frequently, being more intense and
longer lasting during the warmer seasons (for example, April to June
and July to September in the Northern Hemisphere). Annually, the total
number of heatwave days is least at the surface and increases with depth
(Fig. 2a, Supplementary Fig. 2a,b and Extended Data Fig. 4b). The aver-
age duration of heatwaves is also shorter at the surface and increases
with depth (Fig. 2b, Supplementary Fig. 2c,d and Extended Data Fig. 4c).
In contrast, our analysis suggests that the average intensity follows the
opposite pattern (Fig. 2c, Supplementary Fig. 2e,f and Extended Data
Fig. 4d). Regarding their cumulative intensity, which is influenced by
both the duration and intensity of lake heatwaves, there are less notable
relative changes with lake depth (Fig. 2d). In the Great Lakes, where the
rate of change in lake heatwave intensity with depth is far greater than
the change in duration, a decreasing pattern in cumulative intensity
from the surface to deep waters is estimated (Extended Data Fig. 4e).
These large systems also experience considerable spatial variations
in subsurface heatwave characteristics (Extended Data Fig. 4). Our
analysis illustrates that the vertical structure of subsurface heatwaves
has changed across all stratified lakes in our study (Fig. 2e,h).
Lake bottom heatwaves are defined as extreme hot events that
occur in the deepest regions of lakes. On average, the occurrence,
duration and cumulative intensity of bottom heatwaves are greater
than those of lake surface heatwaves (Extended Data Fig. 7). From
1980 to 2022, our analysis suggests that bottom heatwaves have
increased in frequency (7.2 ± 0.6 days decade−1), average duration
(2.1 ± 0.3 days decade
−1
), average intensity (0.2 ± 0.01 days decade
−1
)
and cumulative intensity (5.0 ± 0.5 days decade
−1
). Similar changes
are calculated for the Great Lakes (Extended Data Fig. 8), which have
also experienced an increase (1.9 ± 0.4% decade
−1
) in the percentage
coverage of lake bottom heatwaves since the 1980s. Our analysis based
on individual lakes also indicates an increase in the average intensity
(mean = 0.05 °C; range = −0.3–0.5 °C) of bottom heatwaves (Supple-
mentary Fig. 3e). However, this is not always the case; in certain lakes
and depths, a decreasing pattern in the occurrence days and cumulative
intensity of bottom heatwaves is observed (Supplementary Fig. 3a,c,g).
It might be assumed that bottom heatwaves occur in lakes due to
the presence of a surface heatwave. However, our analysis illustrates
that bottom heatwaves can occur without a lake surface heatwave (Sup-
plementary Figs. 9 and 10 and Supplementary Table 6), the frequency
of which has increased since 1980 (4.3 ± 0.3 days decade
−1
) (Supple-
mentary Fig. 9). However, for deep lakes, it is important to note that
the level of synchrony between surface and subsurface temperature
anomalies varies with depth, meaning that there can be a lag (10 days)
between surface and bottom heatwaves, suggesting some causality;
that is, that bottom heatwaves reflect past surface heatwave conditions
(Extended Data Fig. 9).
Our simulations project considerable changes to subsurface and
lake bottom heatwaves by the end of the twenty-first century (Extended
Data Fig. 10). Future changes to lake heatwaves are prominent at the
(2080–2099). This study also introduces (1) the concept of sub surface
heatwaves, defined as periods in which depth-specific water tempera-
tures reach extreme levels and (2) the presence of a vertical thermal
escape, instances where aquatic species could theoretically move to
deeper water to escape the thermal stress of lake surface heatwaves.
By focussing on key metrics used to describe the severity of lake
heatwaves—their average duration, average intensity and cumula-
tive intensity—this study evaluates how these extreme events have
changed. A multimodel approach is followed to investigate changes
in lake surface and subsurface heatwaves. To quantify these changes,
simulated water temperature profiles available from ISIMIP2b25 were
investigated, as well as lake temperatures simulated from a suite of
independently developed one-dimensional lake models. Moreover,
to capture changes in some of the largest lakes of the world, and to
investigate within-lake variations in heatwaves, the outputs from a
three-dimensional (3D) model of the Laurentian Great Lakes of North
America26 was investigated.
A vertical thermal escape from lake surface
heatwaves
Over the past four decades, lake surface heatwaves have occurred
widely and exhibited a marked increase (Fig. 1 and Supplementary
Figs. 1–4). These extreme hot events have occurred more frequently
(7.8 ± 0.5 annual days decade
−1
) and experienced an increase in their
average duration (2.1 ± 0.1 days decade−1). They have also become
stronger with an increase in their average (0.4 ± 0.02 °C decade
−1
) and
cumulative (5.8 ± 0.4 °C days decade
−1
) intensity (Fig. 1a–d and Supple-
mentary Fig. 1). These lake surface heatwave metrics have undergone
even greater change in some individual lakes (Supplementary Fig. 3 and
Supplementary Tables 1–3), with prominent alterations in the Lauren-
tian Great Lakes (Supplementary Fig. 4). Lake surface heatwaves in the
Great Lakes have occurred more frequently (7.5 ± 1.03 annual days
decade−1), become longer lasting (1.4 ± 0.36 days decade−1) and
experienced an increase in their average (0.01 ± 0.03 °C decade
−1
) and
cumulative (3.7 ± 1.05 °C days decade−1) intensities.
Our study identified the presence of a vertical thermal escape
in lakes, described as instances where motile aquatic species could
theoretically move to deeper water to escape surface thermal stress.
Several lake surface heatwaves have occurred at times in which a verti-
cal thermal escape could be found in deeper regions (where tempera-
tures at depth were below the 90th percentile temperature threshold
of lake surface heatwaves). From 1980 to 2022, our analysis illustrates
an increase (4.3 ± 0.3 days decade−1) in the annual total number of
lake surface heatwaves occurring without a vertical thermal escape
(Fig. 1e), with even greater changes in some lakes (Supplementary Fig. 5
and Supplementary Table 4). The numbers of no-escape events in the
studied lakes are greater (17.5 days) in water bodies that do not experi-
ence thermal stratification, compared to those that seasonally stratify
(7.0 days) (Fig. 1f,g and Supplementary Fig. 6). This is because vertically
mixed lakes lack the thermal gradient needed to separate warm surface
waters from cooler, deeper layers. Consequently, during heatwaves, the
entire water column warms uniformly, preventing the establishment
of cooler refuges. In mixed systems, the number of no-escape events
increases linearly with surface heatwave days (Extended Data Fig. 1a–d).
In stratifying lakes, the relationship is variable and is related to the
duration of thermal stratification (Extended Data Fig. 1d). Our analysis
also suggests that changes to the number of no-escape events during
the historic period (1980–1999 versus 2000–2022) have been greater
in mixed (25.1 days) than in stratified (8.0 days) systems, influenced by
the change in the frequency of surface heatwaves as well as a change in
the proportion of stratified days (Extended Data Fig. 1e,g).
When a vertical thermal refuge exists in stratified lakes, our analy-
sis illustrates that the vertical distance that aquatic species must travel
to escape a lake surface heatwave (the escape depth) has increased
(0.9 ± 0.05 m decade−1) (Fig. 1i). In the Laurentian Great Lakes, the
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change | Volume 15 | May 2025 | 554–559 556
Article https://doi.org/10.1038/s41558-025-02314-0
lake surface, with the magnitude of projected change then decreasing
with depth. Considerable changes are also projected in the number of
no-escape events, increasing by a median of between 17.3 and 92.5 days
(representative concentration pathways RCP 2.6 and 8.5, respectively)
and the escape depth by between 2.1 and 3.2 m (Extended Data Fig. 10e,f).
In terms of lake bottom heatwaves, their occurrence (30.8–136.8 days),
average duration (10.8–81.2 days), average intensity (0.5 to 1.3 °C) and
cumulative intensity (27.9–334.5 °C × days) are projected to increase
in the future relative to the historic period (Extended Data Fig. 10).
Similarly, future heatwave projections for the Laurentian Great Lakes
are consistent with global trends, showing considerable increases
across all depths with large spatiotemporal variability among and within
lakes, particularly under RCP 8.5 (Supplementary Figs. 11 and 12).
Vertically compounding heatwaves
Instances where lake heatwaves occur at either the water surface or
at depth can influence ecosystem functioning. However, it could be
argued that instances where these events co-occur, which we define
as vertically compounding heatwaves, could have a greater impact.
We estimate that vertically compounding heatwaves, which are most
common during the coldest times of the year when many seasonally
stratifying lakes mix (Supplementary Fig. 13) and most frequently occur
in mixed systems (Fig. 3a), have increased in frequency (3.3 ± 0.3 days
decade
−1
) since 1980 (Fig. 3b). Similarly, in the Great Lakes, the per
-
centage of lake area experiencing these vertically compounding
extremes has also increased (1.2 ± 0.13% decade
−1
) (Supplementary
Fig. 14). Furthermore, among other individual lakes studied, 47 out
of 48 lakes experienced a positive change in vertically compound-
ing heatwaves (Supplementary Table 7). By the end of this century,
vertically compounding heatwaves will become increasingly com
-
mon in lakes, increasing by a median of between 16.8 and 119.1 days
(RCP 2.6 and 8.5, respectively), relative to historic conditions. Vertically
compounding heatwaves will become increasingly frequent in lakes
that remain vertically mixed (Fig. 3c–e).
1985 1995 2005
Year
Year
Year
Year Year Year
2015
0
10
20
30
40
50
Number of days
a
7.8 ± 0.5 days decade–1
Days
1985 1995 2005 2015
0
5
10
15
Average duration (days)
b
2.1 ± 0.1 days decade–1
1985 1995 2005 2015
0
0.5
1
1.5
2
2.5
3
3.5
Average intensity (°C)
c
0.4 ± 0.02 °C decade–1
1985 1995 2005 2015
0
10
20
30
40
50
Cumulative intensity (°C x days)
d
5.8 ± 0.4 °C days decade–1
1985 1995 2005 2015
0
10
20
30
Number of no-escape events (days)
e
4.3 ± 0.3 days decade–1
0 5 10 15 20 25 30
Number of no-escape events (days)
0
10
20
30
Number of no-escape events (days)
0 5 10 15 20
∆ No-escape events
1985 1995 2005 2015
2
4
6
8
Escape depth (m)
i
0.9 ± 0.05 m decade–1
0 1 2 3 4 5 6 7
Escape depth (m)
0 5 10
Escape depth (m)
0 1 2 3 4 5
∆ escape depth (m)
050 0 5 10 15 0 2 0 20 40
0 20 40
0 5 10
Mixed
Stratified
f g h
j
k
l
1980–2022 average
1980–2022 average
Days °C °C × days
Days
m
Fig. 1 | Lake surface heatwaves under climate change. a–d, Simulated changes
in occurrence (a), duration (b), intensity (c) and cumulative intensity (d) of lake
surface heatwaves from 1980 to 2022. e–h, Changes in no-escape events for
aquatic species, including days without thermal escape (e), spatial variability
in events (f), histogram comparisons between mixed and stratifying lakes (g)
and overall change in events (2000–2022 relative to 1980–1999) (h). i–l, The
vertical escape depth: temporal variability (i), spatial variability (j), histogram
for stratified lakes (k) and overall change (l). Boxplots (n = 16,455) summarize
changes (blue/red: first/second periods), with the central mark indicating the
median and the bottom and top edges of the boxes indicating the 25th and 75th
percentiles, respectively. The whiskers extend to the most extreme data points
not considered outliers; shaded regions in the time series represent standard
deviation and the solid line represents the mean; trends are shown in the top-left
corners. Basemaps in f, h, j and l made with Natural Earth.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change | Volume 15 | May 2025 | 554–559 557
Article https://doi.org/10.1038/s41558-025-02314-0
Discussion
Lake ecosystems are undergoing transformative changes within a warm-
ing world. Our investigation further explores the thermal response
of lakes to climate change, focussing on the dynamics of heatwaves
across a depth gradient. Our investigation identified several findings
relating to (1) the decreasing potential of a vertical thermal escape
from surface heatwaves, (2) an increase in the vertical distance that
species should travel to escape a surface heatwave when a thermal
refuge exists, (3) considerable variability in the vertical structure of
subsurface heatwaves across lakes, (4) the increased occurrence of
bottom heatwaves with and without extreme surface conditions and
(5) an increase in the frequency of vertically compounding heatwaves.
Previous studies have reported that lake surface heatwaves have
increased in intensity and duration in recent decades1–3. The extreme
water temperatures that species must endure during a surface heat-
wave can often lead to severe consequences5,7. However, previous
studies have also described that motile aquatic species can respond
to environmental disruptions, such as increasing temperatures, by
relocating to favourable habitats
14
. Studies have suggested that many
aquatic species will need to migrate to cooler water at higher elevation
or latitude this century to maintain a preferred thermal habitat13,27. How-
ever, aquatic species could also escape the thermal stress of surface
heatwaves by migrating to deeper regions within a lake. Our investiga-
tion demonstrates that, while there is often the potential for aquatic
species to travel vertically within a lake to reach cooler water, the pro-
portion of lake surface heatwaves without a thermal refuge in deeper
water has increased. This vertical expansion of lake surface heatwaves
highlights the dynamic nature of these extreme heat events, prompting
aquatic organisms to adjust their distribution patterns. As the effects of
lake surface heatwaves reach increasingly deeper water, a reduction in
sufficient habitat can result in changes to species abundance and range.
Our understanding of lake heatwaves is dominated by extreme
conditions at the lake surface. However, our study demonstrates that
the subsurface layers of lakes can often exhibit heatwaves that are
10 15 20 25 30 35
Number of days
–10
–8
–6
–4
–2
0
Depth (m)
a
5 10 15 20 25
Average duration (days)
–10
–8
–6
–4
–2
0
b
0.5 1 1.5 2
Average intensity (°C)
–10
–8
–6
–4
–2
0
c
10 20 30 40
Cumulative intensity (°C x days)
–10
–8
–6
–4
–2
0
d
1990 2000
Year
Year
Year
Year
2010 2020
–10
–8
–6
–4
–2
0
Depth (m)
e
10
20
30
40
50
Number of days
1990 2000 2010 2020
–10
–8
–6
–4
–2
0
f
5
10
15
20
25
Average duration (days)
1990 2000 2010 2020
–10
–8
–6
–4
–2
0
Depth (m)
g
0.5
1
1.5
2
2.5
Average intensity (°C)
1990 2000 2010 2020
–10
–8
–6
–4
–2
0h
10
20
30
40
Cumulative intensity (°C days)
Median
MAD
Fig. 2 | Subsurface evolution of lake heatwaves. a–d, The vertical variability of
each lake heatwave metric from 1980 to 2022, with the central line representing
the median and the shaded region representing the median absolute deviation
(MAD). Shown as the vertical profiles of total heatwave days (a), average
heatwave duration (b), average heatwave intensity (c) and heatwave cumulative
intensity (d). e–h, The interannual variability in each lake heatwave metric at
different depths (negative depths are below the lake surface), showing the
changes with depth in total heatwave days (e), average heatwave duration (f),
average heatwave intensity (g) and heatwave cumulative intensity (h). Here we
compare only the vertical profiles of each heatwave metric across lakes that are
shallower or equal to 10 m.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change | Volume 15 | May 2025 | 554–559 558
Article https://doi.org/10.1038/s41558-025-02314-0
longer lasting and more severe. The extended duration and higher
cumulative intensity of heatwaves in deeper water presents a nuanced
aspect of lake thermal dynamics. Moreover, a greater intensity of
bottom heatwaves implies prolonged exposure to elevated tempera-
tures in deeper waters, potentially posing challenges for demersal
organisms or bottom-dwelling fish inhabiting these regions. Another
important consideration, which was not investigated in this study, is
the potential for lake bottom heatwaves to influence the temperature
of lake sediments, with knock-on impacts on aspects such as GHG
production17. Studies in the marine environment have demonstrated
that sediment heatwaves often co-occur with pelagic heatwaves28. This
suggests that extremes in the overlying water column can propagate
into sediments. Future research should explore this connection in
freshwater systems to better understand the full ecological conse-
quences of lake heatwaves, including the effects on sediment-dwelling
species, nutrient cycling and biogeochemical processes.
Our analysis illustrated that subsurface heatwaves could occur
independently of surface heatwaves as a result of the thermal struc-
ture of lakes, which can lead to the accumulation of heat in deeper
layers, even in the absence of extreme surface warming. Subsurface
heatwaves can arise without a surface heatwave because of time lags
between surface and subsurface temperature responses. In the Great
Lakes, this delay is partly caused by a pronounced difference in tem-
perature seasonality at varying depths. While surface temperatures
in the Great Lakes peak during the summer, deeper waters experience
their maximum temperatures in the autumn. It is also important to
note that surface heatwaves could strengthen the thermal stability
of lakes and subsequently lead to a shoaling of the upper mixed layer.
This increased stability could limit vertical mixing which can, in turn,
restrict the transfer of heat to deeper waters and potentially hinder
the development of subsurface heatwaves. However, our findings sug-
gest that, while surface heatwaves could stabilize the thermocline and
initially restrict heat transfer, prolonged periods of elevated surface
temperatures could lead to the penetration of heat into deeper layers,
facilitating subsurface heatwaves. To fully understand the implications
of surface heatwaves on thermocline stability and the subsequent
development of subsurface heatwaves, further research is needed.
Our study highlights that it is critical for resource managers
to consider shifts in lake temperature extremes, particularly below the
surface. We encourage that our analysis be expanded upon for indi-
vidual or groups of species by incorporating additional considerations,
including horizontal movements, physiology, additional essential
habitat properties and other restrictions on species distributions such
as the need to be near shore or specific breeding or nursing grounds.
When assessing the ecological impacts of subsurface or vertically com-
pounding heatwaves, the significance of these events will vary depend-
ing on the specific biological context. For organisms that occupy the
entire water column, simultaneous surface and bottom heatwaves are
0 5 10 15 20
Concurrent surface and bottom heatwaves (days)
–50 0 50 100 150 200 250 300
Concurrent surface and bottom
heatwaves (days, anomaly)
1980 1985 1990 1995 2000
Year
2005 2010 2015 2020
0
5
10
15
20
25
30
Concurrent surface and bottom
heatwaves (days)
3.3 ± 0.3 days decade–1
b
0 100 200 300 400
Concurrent surface and bottom
heatwaves (days, anomaly)
RCP 2.6
RCP 6.0
RCP 8.5
d
0
10
20
Days
Mixed
Mixed
Mixed
Stratified
Stratified
Mixed
Stratified
Mixed
Stratified
Stratified
0 100 200 300
Days (anomaly)
a
c e
Fig. 3 | Vertically compounding heatwaves. a, Simulated co-occurrence of
lake surface and bottom heatwaves during 1980–2022. b, Temporal changes
in concurrent heatwaves, with global summaries for mixed and stratified lakes
(inset). c, Projected change in co-occurrence frequency by 2080–2099 under
RCP 8.5, relative to 1980–2022. d, Global summaries for RCP 2.6, 6.0 and 8.5.
e, Comparison of future changes between mixed and stratified lakes. Solid
lines in time series represent global averages; shaded regions show standard
deviation. Boxplots (n = 16,455) summarize variability, with the central mark
indicating the median and the bottom and top edges of the boxes indicating the
25th and 75th percentiles, respectively. The whiskers extend to the most extreme
data points not considered outliers, with trends shown in b upper left. Basemaps
in a and c made with Natural Earth.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change | Volume 15 | May 2025 | 554–559 559
Article https://doi.org/10.1038/s41558-025-02314-0
particularly critical, as they can alter the thermal conditions across the
entire habitat. For species that primarily inhabit surface waters, the
absence of deep-water thermal refuges during surface heatwaves is of
concern, as these events create extreme conditions with no possibil-
ity for escape to cooler depths. It is also important to note that, while
some studies in the oceans have shown that marine heatwaves can
have a detrimental impact on the ecosystem23, others have not shown
a dominant effect29. Thus, when anticipating the response of aquatic
species to subsurface lake heatwaves, it is important to consider that
not all species will respond in the same way and that single-species
responses5 do not suggest a net ecological effect.
The vertical expansion of heatwaves and the disparities in occur-
rence, duration and cumulative intensity provide valuable insights
into how climate change manifests in the thermal profiles of lakes.
These findings contribute to a more comprehensive understanding of
the multidimensional impacts of climate change. In terms of manage-
ment and conservation, our investigation advocates for an integrated
approach that considers the diverse responses of surface and subsur-
face ecosystems to changes in extreme temperature. Protecting and
restoring thermal refuges in the lower layers of lakes may be crucial for
preserving biodiversity and ensuring the long-term sustainability of
these ecosystems. Our analysis offers a new perspective on the verti-
cal imprint of lake heatwaves, expanding our understanding of these
extreme events and their potential impacts.
Online content
Any methods, additional references, Nature Portfolio reporting sum-
maries, source data, extended data, supplementary information,
acknowledgements, peer review information; details of author contri-
butions and competing interests; and statements of data and code avail-
ability are available at https://doi.org/10.1038/s41558-025-02314-0.
References
1. Woolway, R. I. et al. Lake heatwaves under climate change. Nature
589, 402–407 (2021).
2. Wang, X. et al. Climate change drives rapid warming and
increasing heatwaves of lakes. Sci. Bull. 68, 1574–1584 (2023).
3. Zhang, K. & Yao, Y. Lake heatwaves and cold-spells across
Qinghai-Tibet Plateau under climate change. J. Geophys. Res. 128,
16 (2023).
4. Cline, T. J. et al. Climate change expands the spatial extent and
duration of preferred thermal habitat for Lake Superior ishes.
PLoS ONE 8, e62279 (2013).
5. Till, A. et al. Fish die-os are concurrent with thermal extremes in
north temperate lakes. Nat. Clim. Change 9, 637–641 (2019).
6. Kelly, S., Moore, T. N. & de Eyto, E. et al. Warming winters threaten
peripheral Arctic charr populations of Europe. Clim. Change 163,
599–618 (2020).
7. Tye, S. P. et al. Climate warming ampliies the frequency of ish
mass mortality events across north temperate lakes. Limnol.
Oceanogr. Lett. 7, 510–519 (2022).
8. Hobday, A. J. et al. A hierarchical approach to deining marine
heatwaves. Prog. Oceanogr. 141, 227–238 (2016).
9. Frölicher, T. L., Fischer, E. M. & Gruber, N. Marine heatwaves under
global warming. Nature 560, 360–364 (2018).
10. Oliver, E. C. J. et al. Longer and more frequent marine heatwaves
over the past century. Nat. Commun. 9, 12 (2018).
11. Laukötter, C., Zscheischler, J. & Frölicher, T. L. High-impact
marine heatwaves attributable to human-induced global
warming. Science 369, 1621 (2020).
12. Comte, L. & Grenouillet, G. Do stream ish track climate change?
Assessing distribution shifts in recent decades. Ecography 36,
1236–1246 (2013).
13. Woolway, R. I. & Maberly, S. C. Climate velocity in inland standing
waters. Nat. Clim. Change 10, 1124–1129 (2020).
14. Kraemer, B. M. et al. Climate change drives widespread shifts in
lake thermal habitat. Nat. Clim. Change 11, 521–529 (2021).
15. Jane, S. F. et al. Concurrent warming and browning eliminate
cold-water ish habitat in many temperate lakes. Proc. Natl Acad.
Sci. USA 121, e2306906120 (2024).
16. Anderson, E. J. et al. Seasonal overturn and stratiication
changes drive deep-water warming in one of Earth’s largest lakes.
Nat. Commun. 12, 1688 (2021).
17. Jansen, J. et al. Global increase in methane production under
future warming of lake bottom waters. Glob. Change Biol. 28,
5427–5440 (2022).
18. Winslow, L. A. et al. Small lakes show muted climate change
signal in deepwater temperatures. Geophys. Res. Lett. 42,
355–361 (2015).
19. Bartosiewicz, M. et al. Hot tops, cold bottoms: synergistic climate
warming and shielding eects increase carbon burial in lakes.
Limnol. Oceanogr. Lett. 4, 132–144 (2019).
20. Pilla, R. M., Williamson, C. E. & Adamovich, B. V. et al. Deeper
waters are changing less consistently than surface waters in a
global analysis of 102 lakes. Sci. Rep. 10, 20514 (2020).
21. Scannell, H. et al. Subsurface evolution and persistence of
marine heatwaves in the Northeast Paciic. Geophys. Res. Lett. 47,
e2020GL090548 (2020).
22. Amaya, D. J., Jacox, M. G. & Alexander, M. A. et al. Bottom
marine heatwaves along the continental shelves of North
America. Nat. Commun. 14, 1038 (2023).
23. Fragkopoulou, E., Sen Gupta, A. & Costello, M. J. et al. Marine
biodiversity exposed to prolonged and intense subsurface
heatwaves. Nat. Clim. Change 13, 1114–1121 (2023).
24. Zhang, Y., Du, Y. & Feng, M. et al. Vertical structures of marine
heatwaves. Nat. Commun. 14, 6483 (2023).
25. Golub, M. et al. A framework for ensemble modelling of climate
change impacts on lakes worldwide: the ISIMIP Lake Sector.
Geosci. Model Dev. 15, 4597–4623 (2022).
26. Xue, P., Ye, X., Pal, J. S., Chu, P. Y., Kayastha, M. B. & Huang, C.
Climate projections over the Great Lakes Region: using two-way
coupling of a regional climate model with a 3-D lake model.
Geosci. Model Dev. 15, 4425–4446 (2022).
27. Comte, L. & Olden, J. D. Climatic vulnerability of the world’s
freshwater and marine ishes. Nat. Clim. Change 7, 718–722 (2017).
28. Tassone, S. J. & Pace, M. L. Increased frequency of sediment
heatwaves in a Virginia seagrass meadow. Estuaries Coasts 47,
656–669 (2024).
29. Fredston, A. L., Cheung, W. W. L. & Frölicher, T. L. et al. Marine
heatwaves are not a dominant driver of change in demersal
ishes. Nature 621, 324–329 (2023).
Publisher’s note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional ailiations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format,
as long as you give appropriate credit to the original author(s) and the
source, provide a link to the Creative Commons licence, and indicate
if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons licence, unless
indicated otherwise in a credit line to the material. If material is not
included in the article’s Creative Commons licence and your intended
use is not permitted by statutory regulation or exceeds the permitted
use, you will need to obtain permission directly from the copyright
holder. To view a copy of this licence, visit http://creativecommons.
org/licenses/by/4.0/.
© The Author(s) 2025
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Methods
Large-scale simulations of lake water temperature
To simulate the occurrence of surface and subsurface lake heatwaves,
we analysed daily lake temperatures provided by the ISIMIP2b lake
sector. The simulations used in this study include water temperature
simulations for 16,455 representative lakes worldwide, specifically
those situated between 60° S and <71° N. An in-depth explanation
of the ISIMIP lake sector is given by ref. 25. The ISIMIP2b lake sector
simulations have been used previously to investigate the impact of
climate change on lake heat budgets
30
, the timing and duration of sum-
mer stratification31 and for attributing the anthropogenic influence
on lake ice and surface water temperature variations32. All ISIMIP lake
sector projections investigated in this study were simulated via a lake
model (see below) driven by an ensemble of bias-corrected climate
projections—GFDL-ESM2M, HadGEM2-ES, IPSL-CM5A-LR and MIROC5—
for historic and future periods, the latter under different RCPs. The
data used to drive the lake models in ISIMIP2b included projections of
air temperature at 2 m, wind speed at 10 m, surface solar and thermal
radiation and specific humidity, which were available at a daily reso-
lution. In this study, to obtain historic to contemporary (1980–2022)
lake temperature projections, we combine the historical simulations
(1980–2005) with the ‘future’ projections (2006–2022) under RCP 8.5
(refs. 11,33,34). Future (2080–2099) lake temperature projections were
investigated under RCP 2.6 (low-emission scenario), 6.0 (medium-high)
and 8.5 (high). These pathways encompass a range of potential future
global radiative forcing from anthropogenic GHGs and aerosols, and
results span a range of potential impacts on lake temperature.
The 16,455 representative lakes investigated in this study were
simulated at a 0.5° × 0.5° grid resolution (the spatial resolution of the
climate projections) with the SimStrat-UoG model. The dataset used to
describe the size distribution of all lakes within each 0.5° grid has a hori-
zontal resolution of 30 arcsec (refs. 35,36) and includes all known lakes
equal or greater than this size threshold. Given the one-dimensional
nature of the SimStrat-UoG model, we considered only lakes with a
depth of <60 m, removing any deeper representative lakes from the
ISIMIP2b global lake sector from the analysis. Furthermore, we included
only lakes that experienced at least 2 months of ice-free conditions
each year. In this study, we standardized the vertical resolution of
the water temperature profiles generated by SimStrat-UoG through
linear interpolation
37
. For the depth range from the surface down to
20 m, we applied a resolution of 0.1 m to capture fine-scale variations.
Between depths of 20 and 50 m, the resolution was set to 0.5 m to
balance detail with computational efficiency. For depths >50 m, we
used a coarser resolution of 1 m to reflect the generally more uniform
temperature patterns in this deeper range. This approach ensures a
consistent and accurate representation of the temperature profile
across different lakes.
For these representative lakes, we compare only vertical profiles
of each lake heatwave metric (see below) across sites that are shal-
lower or equal to 10 m in mean depth. These relatively shallow systems
contribute to >95% of the number of studied sites, which is consistent
with previous studies that have demonstrated that shallow lakes domi-
nate the global distribution of lake types. For example, the HydroLakes
dataset of ref. 38 suggests that ~98% of the 1.42 million lakes worldwide
are ≤10 m in mean depth. While our study encompasses a broad range of
lakes with varying depths, it is important to note that ~77% of the lakes
have a depth of 10 m. This depth distribution implies that most of our
findings, particularly those concerning deeper layers, are primarily
representative of lakes with a maximum depth of ~10 m.
Selection of individual lakes
To evaluate subsurface heatwaves in individual lakes—providing a
more localized perspective than the broader ISIMIP global simulations
described above—we undertook a comprehensive analysis involv-
ing 53 lakes, each with finer-scale vertical temperature profiles. This
approach allows for a granular understanding of thermal dynamics
at a scale that reflects individual lake characteristics rather than
regional averages. Among these lakes, we included the f ive Laurentian
Great Lakes. Owing to their vast surface area and substantial depth,
temperatures in the Great Lakes were simulated using a sophisticated
3D model, which captures the complex interactions between various
lake layers and provides a detailed depiction of thermal stratification
and mixing processes. This high-resolution modelling is crucial for
understanding the unique thermal responses of these large, deep lakes.
In addition, we examined 42 lakes predominantly located in Europe and
North America, for which temperature profiles were simulated using
three distinct one-dimensional lake models. These models, while less
complex than the 3D approach, are still capable of capturing essential
vertical temperature variations and heatwave dynamics. The choice of
these models was influenced by the availability of historical data and
the specific thermal characteristics of these lakes. To further broaden
the geographic scope and improve the spatial representation of our
dataset, we included six additional lakes, some of which are situated
on the Tibetan Plateau. These lakes are characterized by unique cli-
matic and hydrological conditions, which may considerably influence
their thermal behaviour. For these lakes, temperature profiles were
simulated using lake-specific FLake models. The FLake model is well
suited for representing thermal profiles in diverse environments,
including high-altitude and remote regions, thereby enhancing the
overall robustness of our analysis. By incorporating these varied mod-
elling approaches and expanding our dataset to include lakes from
different geographic and environmental settings, we aim to provide a
more comprehensive understanding of subsurface heatwaves across a
range of lake types and locations. This multifaceted approach ensures
that our findings are not only reflective of individual lake characteris-
tics but also contribute to a broader understanding of thermal dynam-
ics in freshwater lakes worldwide. Each of these additional modelling
approaches is described in detail within the following sections.
Three-dimensional model-simulated lake temperature
profiles for the Great Lakes
The water temperatures of the Great Lakes were simulated using
the second version of the Great Lakes–Atmosphere Regional Model
(GLARM)26, which is a two-way 3D lake ice–atmosphere coupled
climate modelling system designed for the Great Lakes region. GLARM
is built upon a two-way coupling between the fourth version of the
International Centre for Theoretical Physics regional climate model
(RegCM4) which simulates land and atmospheric processes39 and the
finite volume community ocean model (FVCOM) which simulates the
3D lake dynamics, thermal dynamics and ice dynamics of the Great
Lakes
40–42
. In the two-way coupled framework of GLARM, the Great
Lakes surface temperature and ice coverage are dynamically calculated
by FVCOM and are provided to RegCM4 as the lower boundary condi-
tion for the atmosphere over the Great Lakes. In turn, RegCM4 calcu-
lates and provides the surface meteorological forcing fields required
by FVCOM. This two-way coupling between RegCM4 and FVCOM is
achieved through the OASIS3-MCT coupler43, and lake hydrodynamic
conditions in FVCOM are configured to evolve and freely interact with
atmospheric conditions over the entire simulation.
GLARM was run for our historical study period (1980–2022) by
prescribing the lateral atmospheric boundary conditions for GLARM
through a combination of ERA-Interim
44
and ERA5 (ref. 45) climate
reanalysis data from the European Centre for Medium-Range Weather
Forecasts. The lateral atmospheric boundary conditions included the
6-hourly pressure and wind components, air temperature and mixing
ratio at all vertical model levels. Future (2080–2099) lake temperature
projections were investigated under RCP 8.5 (high emission) with the
lateral atmospheric boundary conditions provided by three general cir-
culation models (GCMs)—IPSL-CM5A-MR, MPI-ECM-MR and GISS-E2-H.
The future projections presented in this paper are the ensemble average
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
of the three downscaled projections. The three GCMs were selected
out of 19 GCMs on the basis of their performance in reproducing
the observed climate and projecting future warming trends over
North America
26
. GLARM modelling offers climate change projections
that incorporate both the Great Lakes basin and the changes within
the five Great Lakes, by integrating a two-way interaction between a
regional climate model and a 3D lake model. The configuration and per-
formance of the model, in replicating historical lake conditions, select-
ing GCM for dynamical downscaling and projecting future climates,
are comprehensively detailed in ref. 46 and ref. 26. The comprehensive
comparisons of GLARM-simulated lake temperature against mooring
observation
47
temperatures and satellite-derived temperatures are
presented in Supplementary Fig. 15. The Great Lakes simulations are
available from ref. 48.
Ensemble model-simulated lake temperature profiles for
42 individual lakes
The ISIMIP2b lake sector provides lake-specific simulations from 42
lakes with detailed bathymetry and validation data. Lake temperatures
for these lakes were simulated by a suite of independently developed
lake models: (1) general lake model, (2) general ocean turbulence model
and (3) SimStrat. We conducted an analysis of these simulations with
the objective of studying individual lakes at a fine scale, ensuring a
diverse representation across climates and lake types. For all sets of
simulations, lake heatwave metrics were calculated independently
for each lake–climate model combination and then averaged across
the lake–climate model ensemble.
FLake-simulated lake temperature profiles for six individual
lakes
Focusing on the additional six lakes, we adopted the FLake model to
simulate characteristics of lake vertical heatwaves. FLake is a heat
transfer model that parameterizes the vertical temperature profiles of
two-layer water, including a vertically uniform upper layer and a stably
stratified lower layer consistent with self-similarity theory49–51. When
the lake surface is covered by ice and snow, the model accounts for the
presence of these layers. Owing to the balance between computational
efficiency and simulation performance, FLake has been widely used
for accurately reproducing surface temperatures
52
and lake mixing
regimes53,54 at regional and global scales.
The climate forcing variables used to drive the FLake model were
obtained from the hourly grid ERA5-Land product (1981–2020), with
data extracted from grid cells located at the designated lake centres.
Detailed initialization instructions for lake characteristics and other
model parameters required by FLake are outlined in our previous
work52. In this study, to ensure the accurate simulation of vertical
thermal conditions, we first compiled vertical temperature observa-
tions from six lakes not included in the 42 lakes provided in ISIMIP2b
local simulations. These in situ observations were then leveraged to
determine the optimal model settings within FLake for individual lakes.
The parameters subjected to calibration in this study include snow
accumulation rate (kg (m
−2
s
−1
); ranging from 0.0000001 to 0.5), scale
and offset of lake depth (m; ranging from 0.75 to 1.25 for scale and 0–20
for offset), scale and offset of wind speed (U, m s−1; ranging from 0.75
to 1.25 for range and 0 to 8 for offset), albedo for snow and white ice
(ranging from 0.6 to 0.8), albedo for melting snow and blue ice (ranging
from 0.1 to 0.6), light attenuation coefficient (Kd, m
−1
; ranging from
0.1 to 3), scale of solar radiation (ranging from 0.7 to 1.3) and scale of
surface air temperature (ranging from 0.9 to 1.1).
Note that the calibration objective aimed to satisfy several error
criteria, with all errors <2 °C and the minimum average of several error
criteria serving as the selection criterion of optimal parameter settings.
The error criteria selected for evaluation include simulation errors at
various depths (full depth, surface layer, one-third depth, two-thirds
depth and bottom) and across seasons (spring, summer, autumn and
winter). The error metric used is the median absolute error (MAE) of
daily mean temperature simulations. The comprehensive evaluation
results of simulation performance for the six lakes are presented in
Supplementary Table 8 and Supplementary Fig. 16. These simulations
are available from ref. 53.
Lake heatwave metrics
Using the daily simulated lake water temperatures, the occurrence,
average duration, average intensity and cumulative intensity of lake
heatwaves were estimated following the methods described by ref. 1.
Specifically, using the R package heatwave55, lake heatwaves were
defined as when daily lake temperatures were above a local and sea-
sonally varying 90th percentile threshold for at least five consecutive
days. Climatological mean of lake temperature was calculated for each
calendar day using the daily temperatures within an 11-day window
centred on the date across all years within a climatological period,
here defined as 1991–2020 and smoothed by applying a 31-day moving
average. In addition, two events with a break of <2 days were considered
as a single event. The average duration of lake heatwaves is defined as
the difference between the start and end dates of a specific heatwave;
the average intensity is the average temperature anomaly relative to
seasonal climatology averaged during an event; and the cumulative
intensity is calculated as the sum of the temperature anomaly dur-
ing the total duration of each lake heatwave. Lake heatwaves were
defined exclusively during the ice-free season. Our primary focus was
on examining changes in heatwave metrics on an annual (ice-free) scale,
acknowledging that ecologically critical events occur throughout the
year in lakes and are not limited to the warmer (for example, summer)
season. However, lake heatwave characteristics were also compared
across different seasons: January–March, April–June, July–September
and October–December. For each seasonal comparison, only lakes that
experienced a minimum of two ice-free months within the respective
season were included. Subsurface heatwaves were calculated relative
to a subsurface lake temperature climatology, with this climatology
calculated from the temperature time-series data at specific depths.
For example, a bottom lake heatwave was calculated relative to the
bottom water temperature climatology.
The thermal escape depth was estimated as the depth at which water
temperatures decrease below the 90th percentile threshold of a lake
surface heatwave. The escape depth is crucial for anticipating how sur-
face heatwaves might affect lake ecosystems. Surface heatwaves, which
create extreme temperatures in the upper layers of a lake, can prompt
aquatic species to seek cooler refuges at depth. The thermal escape
depth, therefore, represents the boundary below which temperatures
remain conducive to aquatic life, offering a vital refuge during periods
of intense surface heating. The occurrence of vertically compounding
heatwaves, defined as times in which a lake surface and bottom heatwave
occurred at the same time, was also calculated. The global heatwave
metrics calculated from the ISIMIP projections are available from ref. 56.
Lake stratification and mixing
Understanding lake stratification is essential for analysing vertical
thermal dynamics and the occurrence of subsurface heatwaves in lakes.
In this study, a lake is defined as stratified if there is a temperature dif-
ference of >1 °C between the surface and the bottom water57–59. Lakes
that do not meet this criterion are classified as mixed. In mixed lakes,
where the water column is relatively homogeneous and temperature
gradients between surface and bottom layers are minimal, surface
heatwaves typically reflect the thermal conditions throughout the
entire lake. Consequently, extreme temperature variations are gener-
ally uniform across the water column. In contrast, stratified lakes usu-
ally exhibit distinct thermal layers during the summer, including the
epilimnion, thermocline and hypolimnion. In these lakes, the thermal
characteristics of the surface and bottom layers can be markedly dif-
ferent. Although the thermocline reduces the likelihood of extreme
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
surface temperatures directly affecting the hypolimnion, long-term
trends and prolonged surface heatwaves can eventually contribute
to increased thermal stress in deeper layers over time. This study
compares subsurface heatwaves between stratified and mixed lakes
to understand these dynamics better. Additionally, in this study the
depth of the upper mixed layer was calculated by using Lake Analyzer
60
.
For the Great Lakes, which are very deep dimictic lakes, the mixed
layer depth was calculated as the shallowest depth where the density
exceeded the surface density by 0.1 kg m−3 (ref. 61).
Analysis
To explore potential temporal synchrony between the warming of the
surface and subsurface layers, Pearson’s correlation coefficients (R)
and lag between surface and subsurface intensities were estimated
on an event-by-event basis with P values as a measure of statistical
significance (P < 0.05). Correlations were calculated if a surface event
overlaps with a subsurface event within a 5-day margin at the start or
end of each event.
Reporting summary
Further information on research design is available in the Nature
Portfolio Reporting Summary linked to this article.
Data availability
ISIMIP local lake simulations are available from https://doi.org/
10.48364/ISIMIP.563533. ISIMIP global lake simulations are found at
https://doi.org/10.48364/ISIMIP.931371. Simulations of lake tempera-
ture profiles over individual lakes generated in this study are available
at ref. 53. The mooring observations for the Laurentian Great Lakes
used in the model evaluation are available from https://www.glerl.noaa.
gov/data/#watertemp. The heatwave metrics calculated for the Great
Lakes are available at ref. 48. Global heatwave metrics calculated from
the ISIMIP projections are available at ref. 56.
Code availability
Code used to reproduce the analysis and figures generated for this
paper are available at ref. 56. The source code for the FLake model is
accessible at http://www.flake.igb-berlin.de/.
References
30. Vanderkelen, I., Van Lipzig, N. P., Lawrence, D. M., Droppers, B.,
Golub, M. & Gosling, S. N. et al. Global heat uptake by inland
waters. Geophys. Res. Lett. 47, e2020GL087867 (2020).
31. Woolway, R. I., Sharma, S. & Weyhenmeyer, G. A. et al.
Phenological shifts in lake stratiication under climate change.
Nat. Commun. 12, 2318 (2021).
32. Grant, L. et al. Attribution of global lake systems change to
anthropogenic forcing. Nat. Geosci. 14, 849–854 (2021).
33. Oliver, E ., Benthuysen, J. & Bindo, N. et al. The unprecedented
2015/16 Tasman Sea marine heatwave. Nat. Commun. 8, 16101 (2017).
34. Woolway, R. I. Severe lake heatwaves attributable to human-
induced global warming. Geophys. Res. Lett. 49, e2021GL097031
(2022).
35. Kourzeneva, E. External data for lake parameterization in
numerical weather prediction and climate modelling. Boreal
Environ. Res. 15, 165–177 (2010).
36. Choulga, M. et al. Estimation of the mean depth of boreal lakes for
use in numerical weather prediction and climate modelling. Tellus
A 66, 21295 (2014).
37. North, R. P. & Livingstone, D. M. Comparison of linear and cubic
spline methods of interpolating lake water column proiles.
Limnol. Oceanogr. Methods 11, 213–224 (2013).
38. Messager, M. L. et al. Estimating the volume and age of water
stored in global lakes using a geo-statistical approach. Nat.
Commun. 7, 13603 (2016).
39. Giorgi, F. et al. RegCM4: model description and preliminary tests
over multiple CORDEX domains. Clim. Res. 52, 7–29 (2012).
40. Chen, C. et al. An unstructured grid, inite-volume coastal ocean
model (FVCOM) System. Oceanography 19, 78–89 (2015).
41. Chen, C. et al. An Unstructured-Grid, Finite-Volume Community
Ocean Model: FVCOM User Manual (Massachusetts Institute of
Technology, 2012).
42. Anderson, E. J. et al. Ice forecasting in the next-generation Great
Lakes Operational Forecast System (GLOFS). J. Mar. Sci. Eng. 6,
123 (2018).
43. Craig, A., Valcke, S. & Coquart, L. Development and performance
of a new version of the OASIS coupler, OASIS3-MCT_3.0. Geosci.
Model Dev. 10, 3297–3308 (2017).
44. Dee, D. P. et al. The ERA-Interim reanalysis: coniguration and
performance of the data assimilation system. Q. J. R. Meteorol.
Soc. 137, 553–597 (2011).
45. Hersbach, H. et al. The ERA5 global reanalysis. Q. J. R. Meteorol.
Soc. 146, 1999–2049 (2020).
46. Xue, P. et al. Improving the simulation of large lakes in regional
climate modeling: two-way lake–atmosphere coupling with a 3D
hydrodynamic model of the Great Lakes. J. Clim. 30, 1605–1627
(2017).
47. Austin, J. & Elmer, C. An archive of Lake Superior temperature and
current measurements, 2005–2020. Limnol. Oceanogr. Lett. 8,
595–602 (2023).
48. Kayastha, M. B. & Xue, P. Great Lakes dataset for ‘Subsurface
heatwaves in lakes’. Zenodo https://doi.org/10.5281/zenodo.
14790088 (2025).
49. Mironov, D. et al. Implementation of the lake parameterisation
scheme FLake into the numerical weather prediction model
COSMO. Boreal Environ. Res. 15, 218–230 (2010).
50. Mironov, D. V. Parameterization of Lakes in Numerical Weather
Prediction: Description of a Lake Model (DWD Oenbach, 2008).
51. Kirillin, G. Modeling the impact of global warming on water
temperature and seasonal mixing regimes in small temperate
lakes. Boreal Environ. Res. 15, 279–293 (2010).
52. Tong, Y., Feng, L., Wang, X., Pi, X., Xu, W. & Woolway, R. I. Global
lakes are warming slower than surface air temperature due to
accelerated evaporation. Nat. Water 1, 929–940 (2023).
53. Tong, Y. Local FLake simulations of lake temperature proiles for
‘Subsurface heatwaves in lakes’. Zenodo https://doi.org/10.5281/
zenodo.14776366 (2025).
54. Woolway, R. I. & Merchant, C. J. Worldwide alteration of lake
mixing regimes in response to climate change. Nat. Geosci. 12,
271–276 (2019).
55. Schlegel, R. W. & Smit, A. J. heatwaveR: a central algorithm for the
detection of heatwaves and cols-spells. J. Open Source Softw. 3,
821 (2018).
56. Woolway, R. I. Subsurface heatwaves in lakes global dataset and
code. Zenodo https://doi.org/10.5281/zenodo.14966110 (2025).
57. Stefan, H. G. et al. Simulated long term temperature and
dissolved oxygen characteristics of lakes in the north-central
United States and associated ish habitat limits. Limnol. Oceanogr.
41, 1124–1135 (1996).
58. Woolway, R. I. et al. A novel method for estimating the onset of
thermal stratiication in lakes from surface water measurements.
Water Resour. Res. 50, 5131–5140 (2014).
59. Read, J. S. et al. Simulating 2368 temperate lakes reveals weak
coherence in stratiication phenology. Ecol. Model. 291, 142–150
(2014).
60. Read, J. S. et al. Derivation of lake mixing and stratiication indices
from high-resolution lake buoy data. Environ. Model. Softw. 26,
1325–1336 (2011).
61. Wilson, H. L. et al. Variability in epilimnion depth estimations in
lakes. Hydrol. Earth Syst. Sci. 24, 5559–5577 (2020).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Acknowledgements
L.F. was supported by the National Natural Science Foundation of
China (42425604) and the Natural Science Foundation of Guangdong
Province (2023B1515120061). R.I.W. was supported by a UKRI Natural
Environment Research Council (NERC) Independent Research
Fellowship (NE/T011246/1) and a NERC grant reference number
NE/X019071/1, UK EO Climate Information Service. This is contribution
number 125 of the Great Lakes Research Center at Michigan
Technological University. The Michigan Tech high-performance
computing cluster, Superior, was used in obtaining the GLARM
modelling results presented in this publication. GLARM simulations
and analyses were also supported by the Center for Climate-driven
Hazard Adaptation, Resilience and Mitigation, funded by the US
Department of Energy, Oice of Science, Oice of Biological and
Environmental Research (grant number DE-SC0024446). PX was
supported by the U.S. National Science Foundation under Award
Number 2438826. Y.T. was supported by the Postdoctoral Fellowship
Programme (grade C) of China Postdoctoral Science Foundation
(grant number GZC20240637).
Author contributions
R.I.W. conceived the idea for this study. R.I.W., M.K. and P.X. designed
the methodology. R.I.W. analysed the large-scale simulations, Y.T. and
L.F. performed the one-dimensional local-scale simulations and
M.K. and P.X. performed the 3D simulations for the Great Lakes. R.I.W.,
H.S., Y.T., M.K. and P.X. analysed the data with input from the other
authors. R.I.W. led the writing of the paper. All authors contributed
critically to the drafts and gave inal approval for publication.
Competing interests
The authors declare no competing interests.
Additional information
Extended data is available for this paper at
https://doi.org/10.1038/s41558-025-02314-0.
Supplementary information The online version contains supplementary
material available at https://doi.org/10.1038/s41558-025-02314-0.
Correspondence and requests for materials should be addressed to
R. Iestyn Woolway or Pengfei Xue.
Peer review information Nature Climate Change thanks Yan Du, Xinze
Wang and the other, anonymous, reviewer(s) for their contribution to
the peer review of this work.
Reprints and permissions information is available at
www.nature.com/reprints.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Extended Data Fig. 1 | Drivers of no escape events. (a) Correlation between the
inter-annual variability in the number of no escape events and the occurrence
of surface heatwaves. (b) Average correlation in mixed and stratifying lakes.
(c) Relationship between the average number of no escape events and lake surface
heatwaves. (d) For stratifying lakes, we show the relationship between the
average number of no escape events, the number of surface heatwave days, and
the proportion of stratified ice-free days. (e) Changes in the number of no escape
events between 1980-1999 and 2000-2022. (f) Relationship between the change
in the number of no escape events and the number of surface heatwave days.
(g) For stratifying lakes, we show the relationship between the change in the
number of no escape events, the number of surface heatwave days, and the
proportion of stratified ice-free days. In each box plot (n = 16,455), the central
mark indicates the median, and the bottom and top edges of the boxes indicate
the 25th and 75th percentiles, respectively. The whiskers extend to the most
extreme data points not considered outliers. Basemap in a made with Natural Earth.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Extended Data Fig. 2 | Changes in the escape depth in the Great Lakes. Shown in panels a-b are the (a) daily and (b) annual changes in escape depth in the Great Lakes.
Panel c shows the spatial variability across the Great Lakes in terms of temporal differences (2000-2022 relative to 1980-1999) in the escape depth. Basemap in c made
with Natural Earth.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Extended Data Fig. 3 | Relationship between escape depth and surface
heatwave intensity. (a) Correlation (R2) between the inter-annual variability in
the escape depth and the intensity of lake surface heatwaves. (b) The relationship
between the average (1980-2022) escape depth, the average intensity of lake
surface heatwaves, and the depth of the upper mixed layer in lakes. Basemap in a
made with Natural Earth.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Extended Data Fig. 4 | Subsurface heatwaves in the Great Lakes. Shown are
the spatial variability across the Great Lakes at surface, bottom, 10 m, 50 m, and
100 m depths for (a1-a5) water temperature, (b1-b5) occurrence, (c1-c5) average
duration, (d1-d5) average intensity, and (e1-e5) average cumulative intensity
of lake heatwaves. We also show a lake-wide averaged change in (a6) water
temperature, (b6) occurrence, (c6) average duration, (d6) average intensity, and
(e6) cumulative intensity of lake heatwaves with depth. The values are the annual
means during 1980-2022. Basemaps in a–e made with Natural Earth.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Extended Data Fig. 5 | Seasonality of subsurface heatwaves. Shown in panels
a-d are the vertical variability of each lake heatwave metric, namely their (a)
occurrence, (b) average duration, (c) average intensity, and (d) cumulative
intensity, from 1980 to 2022 during different seasons. The central line in each
profile represents the median and the shaded region represents the median
absolute deviation. In panels e-h we show the seasonal variability in each of the
lake bottom heatwave characteristics, namely their (e) occurrence, (f) average
duration, (g) average intensity, and (h) cumulative intensity during the historic
period (1980-2022). In each box plot (n = 16,455), the central mark indicates the
median, and the bottom and top edges of the boxes indicate the 25th and 75th
percentiles, respectively. The whiskers extend to the most extreme data points
not considered outliers.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Extended Data Fig. 6 | Seasonality of subsurface heatwave characteristics in
the Great Lakes. Characteristics are shown for heatwave events that span a single
season (a1-d1) and two seasons (a2-d2). We show the seasonal variability in each
of the lake heatwave characteristics, namely their (a) average occurrence,
(b) average duration, (c) average intensity, and (d) average cumulative intensity.
The values are the annual means during 1980-2022.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Extended Data Fig. 7 | Lake bottom heatwaves under climate change.
Simulated (a-c) occurrence, (d-f) duration, (e-g) intensity, (h-j) cumulative
intensity of bottom heatwaves (1980-2022). Panels (a), (d), (e) and (h) show
the historic average of each heatwave metric (1980-2022), with the global
distribution of each bottom heatwave metric (blue) compared against those of
lake surface heatwaves (red) in (b), (e), (f), and (i). In each histogram, the green
square represents the median. Shown in (c), (f), (g) and (j) are the simulated
changes in bottom heatwave characteristics. The solid line represents the global
average, and the shaded region represents the standard deviation. Trends for
each time series are shown. Basemaps in a,d,e,h made with Natural Earth.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Extended Data Fig. 8 | Evolution of lake bottom heatwaves in the Great
Lakes. Shown are the (a) spatial variability across the Great Lakes for the annual
mean occurrence and the simulated changes in the (b) occurrence, (c) average
duration, (d) average intensity, and (e) average cumulative intensity of lake
bottom heatwaves from 1980 to 2022 in the Great Lakes. Panel (f) shows the
simulated change in the percentage of the Great Lakes area experiencing a
bottom heatwave from 1980 to 2022. Basemap in a made with Natural Earth.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Extended Data Fig. 9 | Relationship between surface and bottom heatwaves
in the Great Lakes. (a) simulated changes in the percentage of the Great Lakes
area experiencing a bottom heatwave without a surface heatwave. We also show
the change in the average lagged correlation between surface and subsurface
heatwave intensity with depth (b) and the spatial variability of the average
lagged correlation between surface and subsurface heatwave intensity (c).
Panel (d) shows the statistically significant average lag identified at different
depths between surface and subsurface heatwave intensities. Basemaps in c
made with Natural Earth.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

Nature Climate Change
Article https://doi.org/10.1038/s41558-025-02314-0
Extended Data Fig. 10 | Future changes in subsurface and lake bottom
heatwaves. (a-d) Vertical variability in lake heat wave metrics by 2080–2099
under RCPs 2.6, 6.0, and 8.5, relative to 1980-2022. Profiles include stratifying
lakes that are less or equal to 10 m in depth, with medians shown as central lines
and shaded regions as median absolute deviation (MAD). Panels (e–j) depict
end-of-century changes in (e) no-escape days, (f) vertical escape distance,
and (g–j) occurrence, duration, intensity, and cumulative intensity of bottom
heatwaves (all lakes, including >10 m depth). In each box plot (n = 166,455),
the central mark indicates the median, and the bottom and top edges of the
boxes indicate the 25th and 75th percentiles, respectively. The whiskers extend
to the most extreme data points not considered outliers.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1
nature portfolio | reporting summary April 2023
Corresponding author(s): RI Woolway
Last updated by author(s): Mar 4, 2025
Reporting Summary
Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency
in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.
Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.
n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement
A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly
The statistical test(s) used AND whether they are one- or two-sided
Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested
A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons
A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient)
AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals)
For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted
Give P values as exact values whenever suitable.
For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings
For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes
Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated
Our web collection on statistics for biologists contains articles on many of the points above.
Software and code
Policy information about availability of computer code
Data collection We used ISIMIP Local lake simulations (https://data.isimip.org/10.48364/ISIMIP.563533) and ISIMIP Global lake simulations (https://
data.isimip.org/10.48364/ISIMIP.931371) for investigating historic and future lake conditions. We used the Freshwater Lake model (FLake) for
simulating some lake temperature profiles. The source code of FLake is available here: http://www.flake.igb-berlin.de/.
Data analysis We used R version 3.5.2. No custom algorithms or software were used.
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and
reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.
Data
Policy information about availability of data
All manuscripts must include a data availability statement. This statement should provide the following information, where applicable:
- Accession codes, unique identifiers, or web links for publicly available datasets
- A description of any restrictions on data availability
- For clinical datasets or third party data, please ensure that the statement adheres to our policy
ISIMIP Local lake simulations: https://data.isimip.org/10.48364/ISIMIP.563533; ISIMIP Global lake simulations: https://data.isimip.org/10.48364/ISIMIP.931371;
simulations of lake temperature profiles over individual lakes generated in this study are available by ref. 53. The mooring observations for the Laurentian Great
Content courtesy of Springer Nature, terms of use apply. Rights reserved

2
nature portfolio | reporting summary April 2023
Lakes used in the model evaluation are available from https://www.glerl.noaa.gov/data/#watertemp. The heatwave metrics calculated for the Great Lakes ae
available from ref. 48. Global heatwave metrics calculated from the ISIMIP projections re available from ref. 56.
Research involving human participants, their data, or biological material
Policy information about studies with human participants or human data. See also policy information about sex, gender (identity/presentation),
and sexual orientation and race, ethnicity and racism.
Reporting on sex and gender N/A
Reporting on race, ethnicity, or
other socially relevant
groupings
N/A
Population characteristics N/A
Recruitment N/A
Ethics oversight N/A
Note that full information on the approval of the study protocol must also be provided in the manuscript.
Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.
Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf
Ecological, evolutionary & environmental sciences study design
All studies must disclose on these points even when the disclosure is negative.
Study description We calculated the characteristics of subsurface heatwaves in lakes worldwide.
Research sample We calculated the heatwave metrics for 16,455 lakes worldwide.
Sampling strategy The study sites are selected based on the availability of model derived lake temperature projections. The forcing data and lake
models used are commonly used in lake studies.
Data collection We modelled lake temperature profiles in this study.
Timing and spatial scale 1980-2099 at daily resolution
Data exclusions No data were excluded.
Reproducibility The forcing data, lake model, and modelled lake temperature profiles are publcly available.
Randomization Not relevant as all simulation results were used in the analysis.
Blinding Not relevant as all simulation results were used in the analysis.
Did the study involve field work? Yes No
Reporting for specific materials, systems and methods
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material,
system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

3
nature portfolio | reporting summary April 2023
Materials & experimental systems
n/a Involved in the study
Antibodies
Eukaryotic cell lines
Palaeontology and archaeology
Animals and other organisms
Clinical data
Dual use research of concern
Plants
Methods
n/a Involved in the study
ChIP-seq
Flow cytometry
MRI-based neuroimaging
Novel plant genotypes N/A
Seed stocks N/A
Authentication N/A
Plants
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com