Warming and acidification threaten glass sponge Aphrocallistes vastus pumping and reef formation

Abstract

The glass sponge Aphrocallistes vastus contributes to the formation of large reefs unique to the Northeast Pacific Ocean. These habitats have tremendous filtration capacity that facilitates flow of carbon between trophic levels. Their sensitivity and resilience to climate change, and thus persistence in the Anthropocene, is unknown. Here we show that ocean acidification and warming, alone and in combination have significant adverse effects on pumping capacity, contribute to irreversible tissue withdrawal, and weaken skeletal strength and stiffness of A. vastus. Within one month sponges exposed to warming (including combined treatment) ceased pumping (50–60%) and exhibited tissue withdrawal (10–25%). Thermal and acidification stress significantly reduced skeletal stiffness, and warming weakened it, potentially curtailing reef formation. Environmental data suggests conditions causing irreversible damage are possible in the field at +0.5 °C above current conditions, indicating that ongoing climate change is a serious and immediate threat to A. vastus, reef dependent communities, and potentially other glass sponges.

Full text

1
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
Warming and acidication threaten
glass sponge Aphrocallistes vastus
pumping and reef formation
A. Stevenson1,2,3 ✉ , S. K. Archer4,5, J. A. Schultz6, A. Dunham4, J. B. Marliave6, P. Martone7 &
C. D. G. Harley1,2
The glass sponge Aphrocallistes vastus contributes to the formation of large reefs unique to the
Northeast Pacic Ocean. These habitats have tremendous ltration capacity that facilitates ow of
carbon between trophic levels. Their sensitivity and resilience to climate change, and thus persistence
in the Anthropocene, is unknown. Here we show that ocean acidication and warming, alone and in
combination have signicant adverse eects on pumping capacity, contribute to irreversible tissue
withdrawal, and weaken skeletal strength and stiness of A. vastus. Within one month sponges
exposed to warming (including combined treatment) ceased pumping (50–60%) and exhibited tissue
withdrawal (10–25%). Thermal and acidication stress signicantly reduced skeletal stiness, and
warming weakened it, potentially curtailing reef formation. Environmental data suggests conditions
causing irreversible damage are possible in the eld at +0.5 °C above current conditions, indicating that
ongoing climate change is a serious and immediate threat to A. vastus, reef dependent communities,
and potentially other glass sponges.
Sponges have an important functional role in ecosystems worldwide and over the entire marine bathymetric gra-
dient where they eciently lter water, link food webs, and facilitate the ow of carbon between trophic levels1–3,
alter the water column and its processes4,5, and provide biogenic habitat6. is is particularly true for the glass
sponge Aphrocallistes vastus (class Hexactinellida), which – along with Heterochone calyx and Farrea occa – form
large biogenic reefs that cover several square kilometres of the seaoor o the west coast of Canada7,8. ese reefs
are built through larval sponges settling atop the fused dead skeletons of previous generations, and therefore
mechanical integrity of the sponge skeleton is vital to reef formation, persistence, and growth. While hexactinel-
lids are widely distributed in the deep sea (>70 m)9, in British Columbia (BC), Canada, they occur as shallow as
20 m and form complex reefs, which are home to a rich community of sh and invertebrate species6,10. ese reefs
process considerable volumes of water (465–47,300 L/m2 per day), twice as fast as the next most intense suspen-
sion feeding community in the ocean (i.e. mussel beds), which could have strong impacts on local and regional
bentho-pelagic coupling, nutrient cycling, and carbon sequestration6,11. ese globally unique glass reefs, prior to
their discovery in 1986, were thought to have been extinct for 40 million years7.
Although glass sponge reefs are now subject to ongoing intensive conservation eorts, responses of these and
other sponge-dominated communities to ongoing environmental change remain largely unknown. Acidication
could be detrimental to invertebrates, such as sponges, because of their inability to compensate for reductions in
extracellular pH12, but responses of non-reef-building marine sponges to acidication in the lab and eld are idio-
syncratic and highly variable13–18. Also, no simple and consistent relationships have been found between tempera-
ture and pumping rate of ciliary suspension feeders19. Drastically dierent and species-specic responses to ocean
acidication and warming make it dicult to predict the vulnerabilities of sponge species that have not been
examined, like glass sponges. However, siliceous sponges, including hexactinellids, have survived pre-historical
mass extinction events caused by ocean acidication20,21. eir prevalence at natural volcanic CO2 seeps in Papua
1Department of Zoology, University of British Columbia, Vancouver, British Columbia, V6T 1Z4, Canada. 2Institute for
the Oceans and Fisheries, University of British Columbia, Vancouver, British Columbia, Canada. 3Marine Evolutionary
Ecology, GEOMAR Helmholtz Centre for Ocean Research Kiel, Dsternbrooker Weg 20, 24105, Kiel, Germany.
4Fisheries and Oceans Canada, Pacic Biological Station, 3190 Hammond Bay Road, Nanaimo, British Columbia, V9T
6N7, Canada. 5Louisiana Universities Marine Consortium, 8124 Highway 56, Chauvin, Louisiana, 70344, USA. 6Ocean
Wise Research Institute, PO Box 3232, Vancouver, British Columbia, V6B3X8, Canada. 7Department of Botany,
University of British Columbia, Vancouver, British Columbia, V6T 1Z4, Canada. ✉e-mail: stevenan@zoology.ubc.ca
OPEN
Content courtesy of Springer Nature, terms of use apply. Rights reserved

2
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
www.nature.com/scientificreports/
New Guinea22, and persistence through rapid regional climate-induced warming along the Antarctic Peninsula23,
suggests that siliceous sponges may do well under conditions of acidication and warming.
Here, we conducted a four-month mesocosm experiment to assess the effects of elevated temperature,
CO2-induced acidication, and their interaction on pumping capacity, tissue withdrawal, and mechanical integ-
rity of the skeletal structure of the reef-building glass sponge A. vastus. For this study, 32 juvenile specimens
were randomly assigned to one of 16 aquaria (n = 2 per aquarium) set to one of four treatment combinations
(4 aquaria per treatment combination): (1) control (ambient temperature = 8.6 C and pH = 7.8), (2) reduced
pH (present-day temperature and projected year 2100 pH = 7.6; OA), (3) elevated temperature (projected year
2100 temperatures 10.4 C ( + 1.8 °C) and present-day pH = 7.8; OW) and (4) elevated temperature and reduced
pH (projected year 2100 temperatures and pH; OAW). Fluorescent dye (Calcein) was injected near the sponge’s
midsection to measure two aspects of pumping capacity: the time it took the dye to be expelled from the oscu-
lum (hereaer ‘minimum residence time’), and the vigor of the dye plume expelled from the oscula (‘pumping
strength’ hereaer). Pumping strength was scored over a gradient of 0–6: ‘weak’ = a diuse plume of dye (score
1–3), and ‘strong’ = dense plume of dye (4–6), ‘none’ = apparent pumping arrest (scored 0). ‘Apparent pumping
arrest’ does not correspond to pumping arrest because a owmeter was not used to record this, it does however
infer that pumping was so weak that it was not observed with the use of a dye. Apparent pumping arrest is a com-
mon behavioral response to exposure to a stressor such as sediment24,25. Onset of tissue withdrawal was noted
when it occurred. To assess sponge skeleton mechanical properties, we measured sponge skeleton fracture force
(breaking point) and modulus (stiness) at the end of the experiment.
Apparent pumping arrest
A greater proportion of sponges exposed to warming and/or acidication treatments ceased pumping than the
control sponges over the course of the experiment (Fig.1a), but there was no signicant eect of acidication,
warming, or their interaction on apparent pumping arrest over time (Table1a). e onset of apparent pumping
arrest was seen as early as two weeks in sponges exposed to warming (including OW and OAW), and the propor-
tion of individuals not ltering remained relatively stable in the OA and OW treatments, but there were uctua-
tions observed in the OAW treatment combination (Fig.1a).
Pumping capacity
Minimum residence times were similar across treatments early in the experiment, but then diverged through time
in response to acidication and warming (Fig.1b). Although minimum residence time remained constant in con-
trol tanks, it declined by 2- to 3.5- fold in sponges exposed to OA, OAW, OW treatment combinations. Individuals
subjected to acidication and warming separately pumped the dye signicantly slower than the control aer four
months (120 days) of exposure to these treatments (Table1b). Treatment interaction dampened this negative
response but not signicantly.
Aer four months of exposure, sponges in OA, OAW, and OW tanks showed reduced pumping strength,
by 2- to 5.5- fold compared to the control (Fig.1c). Strength was signicantly weaker in individuals subjected
to acidication and warming relative to the control (Table1c; Fig.1c). Warmed sponges (OW and OAW) had
depressed pumping strength as early as the rst sampling point, whereas sponges in the acidication only treat-
ment lost pumping strength more gradually (more details Supplementary TableS1). Notably, aer three months,
sponges exposed to elevated temperature alone showed increase in minimum residence time (slowed pumping)
and decrease in pumping strength, but the pumping capacity of individuals in the OAW treatment (exposed to
both acidication and warming) was relatively faster and stronger, similar to the acidied treatment. However, in
the nal (fourth) month pumping capacity of sponges in OAW ultimately worsened, mirroring that of individuals
subjected to warming (Fig.1). ese patterns resulted in a signicant Acidication x Warming x Time interaction
(Table1c).
Tissue withdrawal
e eects of acidication and warming on tissue withdrawal were large in magnitude, but signicance was not
detected (Table1d), potentially as a result of the relatively low sample size (n = 8 per treatment combination). Yet,
trends are alarming and worth detailing: individuals subjected to warming (including OW and OAW treatment
combinations) had earlier onset (by one month) of tissue withdrawal relative to the control and OA treatment
combinations. By the end of the experiment all (100%) sponges in the OA and >75% of sponges in the warmer
(OW and OAW) treatment combinations had signs of tissue withdrawal, a 35–60% increase compared to con-
trol sponges (Fig.1d). e Cox proportional hazards regression model estimated hazard ratio (Exp. Coe. in
Table1d) suggests a threefold increase in the probability of acidied and warmed sponges showing signs of tissue
withdrawal compared to those in the control.
Skeletal breaking force per volume and stiness
Experimental treatment combinations (OA, OW, OAW) reduced the force per volume required to break A. vastus
skeleton (Fig.2a), but only a signicant eect of warming was detected (Table1e). Both acidication and warm-
ing signicantly reduced skeleton modulus (stiness; Fig.2b), meaning the skeleton became more elastic aer
four months exposure to these conditions (Table1f). ere were no signicant Acidication x Warming interac-
tion eects for these material properties.
Warming and acidication pose an immediate threat to sponge ltration and reef
formation
Our results indicate that future acidication, warming, and their combination may have substantial adverse eects
on the pumping capacity, tissue withdrawal, and structural integrity of the glass sponge A. vastus, a species that
contributes to the formation of historically and ecologically important habitats unique to the Pacic Northwest.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

3
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
www.nature.com/scientificreports/
Figure 1. Apparent pumping arrest (a), minimum residence time (b), pumping strength (c), and onset of tissue
withdrawal (d) in the reef-building glass sponge Aphrocallistes vastus exposed to four treatment combinations.
Treatment combinations include: ambient conditions (‘Control’), CO2-induced acidication (‘OA’), increased
seawater temperature (‘OW’), and a combination of both (‘OAW’) for four months. (a) Colour gradient
represents total apparent pumping arrest (dark shade) to strong pumping (light shade). (b) ‘Minimum residence
time’ refers to time (in seconds) taken to expel dye from the oscula aer being injected with a xed volume,
mean values exclude individuals that were not pumping (assigned a pumping strength score of zero). (c)
‘Pumping strength’ is comprised of a score assigned to the volume of the plume expelled from the oscula, mean
values include individuals that were not pumping (score of zero). (d) “Kaplan-Meier survival curve” for the
probability of observing tissue withdrawal in each individual. 95% condence limits are shown (in d) and error
bars represent standard error (SE; in b, c) of the mean (n = 8 per treatment combination).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

4
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
www.nature.com/scientificreports/
Most worryingly, the onset of apparent pumping arrest was quick (occurring within two weeks) for sponges
exposed to elevated temperatures regardless of acidication. Subsequent degradation was observed two weeks
following apparent arrest. e rigid skeleton of glass sponges, like A. vastus, does not permit these animals to
contract like other sponges from class Demospongiae (which possess both protein and silica spicules) in response
to, for example, particle obstruction24,25. To protect themselves from this, glass sponges typically go into tempo-
rary apparent pumping arrest whereby they cease ltering particles from surrounding water to prevent particle
Source of error Estimate Std. Error z/t value p-value
Apparent pumping arrest (Binomial GLMM)
Intercept 4.1732 1.5045 2.774 0.0055
Acidication −2.0661 1.6943 −1.219 0.2227
Warming −3.1130 1.6667 −1.868 0.0618
Acidication × Warming 1.3895 1.9890 0.699 0.4848
Time −0.0212 0.0165 −1.284 0.1990
Acidication × Time 0.0085 0.0194 0.437 0.6621
Warming × Time 0.0076 0.0188 0.404 0.6865
Acidication × Warming × Ti me 0.0101 0.0234 0.432 0.6659
Minimum residence time (LM, log-transformed)
Intercept 3.2466 0.1897 17.114 <0.0001
Acidication 0.0638 0.2637 0.242 0.8093
Warming 0.1749 0.2959 0.591 0.5555
Acidication × Warming −0.1127 0.4268 −0.264 0.7922
Time −0.0043 0.0028 −1.575 0.1178
Acidication × Time 0.0083 0.0040 2.094 0.0383
Warming × Time 0.0106 0.0043 2.448 0.0157
Acidication × Warming × Ti me −0.0120 0.0062 −1.930 0.0559
Pumping strength (Poisson GLMM)
Intercept 1.2967 0.2202 5.890 <0.0001
Acidication 0.1080 0.3088 0.350 0.7265
Warming −0.6554 0.3694 −1.774 0.0760
Acidication × Warming 0.0193 0.5029 0.038 0.9694
Time 0.0001 0.0021 0.069 0.9451
Acidication x Time −0.0095 0.0034 −2.808 0.0050
Warming × Time −0.0093 0.0045 −2.062 0.0392
Acidication × Warming × Ti me 0.0153 0.0061 2.524 0.0116
Probability of tissue withdrawal (Cox proportional hazards test)
Source of error Coecient Exp. Coe.Std. Err z-value p-value
Acidication 1.1961 3.3073 0.6859 1.744 0.0812
Warming 0.9815 2.6684 0.7101 1.382 0.1669
Acidication × Warming −0.8903 0.4105 0.8803 −1.011 0.3119
Breaking force (LM)
Source of error Estimate Std. Error t-value p-value
Intercept 0.1049 0.0155 6.746 <0.0001
Acidication −0.0374 0.0220 −1.703 0.1010
Warming −0.0479 0.0213 −2.248 0.0336
Acidication × Warming 0.0451 0.0306 −1.475 0.1526
Modulus (stiness) (LM)
Intercept 1.1803 0.1581 7.467 <0.0001
Acidication −0.5922 0.2327 −2.545 0.0178
Warming −0.4688 0.2164 −2.166 0.0405
Acidication × Warming 0.5436 0.3178 1.711 0.1000
Table 1. Results of the optimal, most parsimonious Linear and Mixed Models testing for independent and
interactive eects of acidication, warming, and time (where applicable) on Aphrocallistes vastus performance
and mechanical traits: a) apparent pumping arrest, b) minimum residence time, c) pumping strength, as well
as d) the Cox proportional hazards test results for the probability of tissue withdrawal, and e) skeleton breaking
force, and f) skeleton modulus (material stiness). e type of model, distribution, and transformation (where
applicable) used are summarized in brackets. ‘Acidication’ = CO2-induced acidication. Statistically signicant
eects in bold; alpha = 0.05.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

5
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
www.nature.com/scientificreports/
obstruction24. It is possible that treatments caused signicant stress, manifested as apparent pumping arrest lead-
ing to tissue withdrawal, throughout the experiment.
Warming can inuence sponge-feeding behavior by reducing choanocyte chamber density, size, and there-
fore ltration eciency26. Few glass sponges have successfully been kept in recirculating seawater, but for those
that did survive for weeks to months, unusual changes in the structure of choanocyte chambers were noted over
time27, which may help explain tissue withdrawal observed in control sponges. To the best of our knowledge there
have been no studies investigating the eects of acidication and warming on glass sponge ltration, but there are
some examples among demosponges. Similar to our ndings for A. vastus, Rhopaloeides odorabile shows drastic
reductions in pumping rates and feeding eciency in response to warming13. In contrast, Dysidea avara ltra-
tion rates remain unaected by natural changes in temperature28, but in Halichondria panicea rates increased at
higher temperatures19. Overall, available data suggest that ocean warming impacts on ltration capacity could be
species-specic, and the eects of warming will depend on whether a particular species is already near or above
its thermal optimum.
e action potential controlling ltration in glass sponge Rhabdocalyptus dawsoni, is known to function
within a narrow temperature range (7–12 °C)29, but it is unclear if this is the case for other glass sponges. Below
7 °C, sponges are unable to resume ltration aer arrest, and cannot undergo arrest at temperatures above 12 °C,
thereby making them more susceptible to starvation and clogging from sediments25. e ambient and upper limit
of the temperatures examined in the present study were within the physiological tolerance limits of glass sponges,
but signs of distress were still observed under the climatically realistic magnitude of warming used in our experi-
ment. It is possible that prolonged exposure (>2 weeks, as dened by our study) to warming might further restrict
the physiological limits of glass sponges and could cause a decrease in biomass of A. vastus populations as a result
of starvation (marked by apparent pumping arrest).
Periods of prolonged warming have already been observed in the eld, at the collection site of the present
study (Fig.3) and in other Howe Sound bioherms30. Warm periods, dened as temperatures reaching >10.4 C
with no more than 12 hrs of cooling (temperatures < 10.4 C), lasting 6–13 days occurred six times between July
and October, 2016, with ve brief periods of cooling in between warm periods, which corresponded to a weak
temperature anomaly (La Niña year)31. Results suggest that irreversible tissue withdrawal could take place in A.
vastus aer 30 days of exposure to warming (>10.4 C), which could have occurred if it were not for several brief
periods of cooling observed in the summer of 2016. Warming trends pose an immediate stress to glass sponge
reefs, as the addition of 0.5 °C to the 2016 pattern would result in 140 consecutive days of warming, a period
longer in length and warmer than the sponges were exposed to in the present study.
Figure 2. Max breaking force per volume (N/mm3; a) and mean modulus (stiness; MPa; b) withstood by
fresh juvenile Aphrocallistes vastus skeleton. Skeleton samples were tested at the end of the experiment aer four
months exposure to four treatment combinations: ambient conditions (‘Control’), CO2-induced acidication
(‘OA’), increased seawater temperature (‘OW’), and a combination of both (‘OAW’). Error bars represent
standard error (SE) of the mean (n = 8 sponges per treatment combination, with 5 measurements per sponge).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

6
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
www.nature.com/scientificreports/
Responses of glass sponges to ocean acidication have not previously been investigated; responses of species
in other sponge classes are not well known and, similar to responses to warming, appear to be species-specic.
Some species appear to be resistant to acidication: elevated temperatures caused signicant adverse eects on
abundant tropical sponge species, but acidication alone had little eect32. Cliona orientalis, a demosponge, had
increased bioerosion rates under acidied conditions14; demosponge species were present near Mediterranean
CO2 vent sites with pH values as low as 6.6 units16; no eect of acidication was found on the survival rates of
Crella incrustans18; and Mycale grandis showed extraordinary resistance to acidication17. In contrast, increased
mortality in response to acidication was seen in sponges Cliona celata15 and Tethya bergquistae18. Our study
suggests glass sponges are less sensitive to ocean acidication than warming, at least within the range of change
expected for these variables in the coming decades, but are not resilient to long-term exposure to either since both
elevated temperature and acidication ultimately had detrimental outcomes for the sponges.
Importantly, the interactive eect of acidication and warming had mitigating eects on the pumping capacity
of individuals exposed to warming for the rst three months, mirroring the response of acidied sponges. e
oscillations in pumping capacity and apparent arrest throughout the experiment suggest that the interactive eect
of acidication may cause the sponges to intermittently start/stop pumping. Contrary to our work, the interac-
tion between acidication and warming exacerbated the eect of temperature stress in heterotrophic sponge
species32. However, acidication may mitigate these stresses in phototrophic species, reducing mortality, necrosis
and bleaching of tropical sponges32,33. In the nal month of our experiment, individuals subjected to a combi-
nation of elevated CO2 with warmer temperatures performed similarly to the temperature treatment suggesting
that acidication may have a threshold and short-term buering capacity. Because acidication did not dampen
the presence of tissue withdrawal, it may not be able to mediate the eect of temperature and ultimate loss of this
species in the long term.
It must be noted that there have been documented mass mortality of glass sponges (including A. vastus) in
Howe Sound, where the sponges were collected for this study and several glass sponge reefs exist. ese extensive
glass sponge mortalities (including A. vastus) correlate with elevated temperatures reported during the 2009/2010
and 2015/2016 El Niño events30,34, and provide some indication that these sponges are sensitive to elevated tem-
peratures. However, this period of warming was not associated with a decrease in pH35. Furthermore, acidica-
tion independent of warming has been documented in Howe Sound, but not associated with sponge mortalities35.
From these eld observations there can be no conclusions drawn regarding how temperature and acidication
may interact in the eld and how acidication may impact the sponges under natural circumstances. However,
the patterns do qualitatively match the results seen in our experiment suggesting warming is the primary threat
to glass sponges.
e combination of reduced skeletal stiness (under warming and acidication) and strength (from warm-
ing) would be expected to slow or completely curtail reef formation. e fused and three-dimensional skeletal
network, comprised of biosilica and chitin, held together at the joints with low concentrations of calcite, is respon-
sible for the sponges’ rigid body that prevents disaggregation of the skeleton long aer its death, allowing for reef
development36–39. e dictyonine skeleton (fused robust scaolding) is thought to reduce skeletal stiness in
glass sponges like A. vastus, providing natural exibility to minimize stresses posed by hydrodynamic forces in
shallower waters40. Material stiness values (measured as Young’s modulus) from previous work on A. vastus are
slightly higher (2.76–10.04 MPa)40 than those obtained in the present study (control sponges = 1.2 + 0.7 MPa).
Discrepancies might be due to life stage dierences as the present study was conducted on juvenile sponges
(3–8 cm in height). Regardless, warmed and/or acidified sponges were half as stiff as the control sponges.
Alterations to the skeleton, especially in terms of reduced stiness (increased exibility) as presented here, could
reduce feeding eciency, lowering the sponges’ critical water ow threshold, and potentially their distribution,
restricting them to waters with higher food availability. Furthermore, under warming conditions the more brit-
tle (measured as reduced force per volume) skeletons might collapse under the increasing weight of a growing
sponge, which can reach 2–3 m in height36, and/or might not be able to withstand the myriad animals walking
and swimming in and among the sponges40,41. Because we only examined material properties in living tissue, it is
Figure 3. Daily average, maxima, and minima of one year (January 2016 – December 2016) iButton
temperature logger data captured at the collection site, Field of a ousand, o west Bowen Island. Dashed line
represents the maximum temperature sponges experienced in the present study (10.4 °C). e recording iButton
temperature logger was positioned at 23 m. (Data courtesy of Ocean Wise Research Institute).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

7
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
www.nature.com/scientificreports/
unclear how the dead skeleton would be altered by climate change and whether it too would succumb to the fate
of the living skeleton, but it is reasonable to suspect that dierences in skeletal strength apparent during life would
perpetuate aer death. is is critical as dead sponges are important for reef growth as larval glass sponges and
other invertebrates settle and grow on the macerated skeleton36. e unique architecture of glass sponges vital to
reef formation may be vulnerable to climate change.
Implications for associated biodiversity and ecosystem function
Exposure to acidication and warming reduced the feeding eciency (i.e. increased minimum residence time
and decreased pumping strength) of glass sponges, suggesting that the feeding ability of juvenile A. vastus might
be diminished (2–5.5 fold) by the end of the century as a result of climate change. Cascading eects of impaired
pumping on local and regional biogeochemical processes remain unknown, but are likely to be negative. Via their
remarkable ltration capacity, sponges convert large quantities of suspended particles and dissolved organic car-
bon (DOC) into food for other animals1. rough feeding, excretion, and symbiont microbial activity, sponges
are known to chemically transform seawater passing through their structure42. e 19 documented glass sponge
reefs in the Salish Sea, for example, collectively lter 1.04 × 1011L of water each day, representing 1% of the total
water volume in the Strait of Georgia and Howe Sound combined6. By doing so, glass sponges bring microbial
food energy from marine and terrestrial sources into local food webs by feeding on and removing up to 90% of
bacteria from the water11,43. Reduction in this tremendous ltration capacity, as well as the reefs’ eventual phys-
ical decimation, could alter local and regional microbial loop and energy supplied to the benthic community.
Examples of breakdowns in bentho-pelagic coupling exist: sponge populations in Florida Bay have historically
controlled phytoplankton blooms via particle removal and pumping rates44. Devastation of the sponge population
in the area lead to increased toxic blooms in Florida Bay. Reduced skeletal strength could act as a positive feed-
back loop further weakening the sponge infrastructure and making it more prone to damage from inhabitants
(sh and invertebrates) moving about the reef. Habitat loss as a consequence of ocean acidication45 and warm-
ing46 has negative downstream eects on biodiversity in coral reefs, mussel beds, and some macroalgal habitats.
Similarly, we anticipate biodiversity loss in these ancient glass sponge habitats as a result of climate change.
Methods
Collection and husbandry. Juvenile A. vastus, ranging in height from 3 to 8 cm, were randomly selected
from ‘Field of A ousand’ dive site on the west side of Bowen Island (49.396, −123.397) in Howe Sound, British
Columbia, Canada, under collection license XR 321 2017. Sponges were placed in plastic bags with ambient sea-
water (collected at depth) and stored in coolers for transportation to the laboratory at the University of British
Columbia. To ensure longevity, sponges were slow drip acclimatized to their tank chemistry over the course of
one hour by adding 100 mL of water to the collection bag (stored in a cooler) every 10 min from the respective
tank in which an individual sponge was to be housed.
Two sponges were placed in each of sixteen 250 L recirculating seawater aquaria bubbled constantly with
ambient air and equipped with a multistage ltration system, including biological lter (sock ltration, protein
skimmer, and bioballs) and UV sterilizer. e source seawater was obtained locally from 16 m depth in Burrard
Inlet, BC, and coarse ltered by the Vancouver Aquarium. Sponges were held in total darkness with red light
exposure during feeding. White light exposure was kept to a minimum, 1 hr per month or less, to measure pump-
ing activity and observe tissue withdrawal.
Ammonia, nitrite, and nitrate (using API Marine Master Test Kit) were monitored throughout the exper-
iment. Twenty percent water changes were performed when necessary. Water changes were also conducted at
least once per month during cleaning, which was kept at a minimum to avoid stressing the animals with excessive
water movement. Siliconoxide was monitored with Salifert Si Prole Test kit. To supplement silica content in the
water, two drops Sponge Excel Marine High-purity Silica from Brightwell Aquatics were added twice to each tank
throughout the experiment.
e sponges were fed twice daily at xed times (every 12 hrs) to approximate their natural exposure to tidal
rhythm. Each sponge tank received: in the morning, 0.5 mL Reef Nutrition Roti Feast + 0.5 mL Reef Nutrition
Oyster Feast/tank mixed with 10 mL seawater, fed to the sponges using Kent Marine Sea Squirt feeder; in the
evening, two drops of concentrated Sponge Power (Korallen-Zucht Sponge Power) directly added to each tank.
In addition, four times weekly, 0.5 mL Fauna Marine Ultra Min S and 0.5 mL Fauna Marine Ultra Min D mixed
with 10 mL seawater was added to each tank using Kent Marine Sea Squirt feeder. All food was injected near the
water’s surface to prevent contact or movement near the animal.
Experimental setup and water chemistry. e sponges were acclimated in their assigned tanks at
8–9 C for ve days without food; on the sixth day, the sponges were fed and tanks were set to their experimental
temperature and pH over 8 hrs. Experimental treatments were chosen based on conservative future projections
(temperature + 1.8 °C and ΔpH –0.2 units based on year 2100 projections)47. e 16 experimental aquaria were
divided equally into four treatments: (1) control (ambient temperature = 8.6 C and pH = 7.8), (2) reduced pH
(present-day temperature and projected year 2100 and pH = 7.6), (3) elevated temperature (projected year 2100
temperatures 10.4 C ( + 1.8 °C) and present-day pH = 7.8) and (4) elevated temperature and reduced pH (pro-
jected year 2100 temperatures and pH). Because of a leak in the CO2 canister, acidication took place one week
later than other treatments, but for conservative purposes its time span was treated similarly to other treatments
for analyses.
Temperature was maintained using individual chillers connected to each tank. Elevated CO2 concentrations
were achieved using mass ow controllers to bubble an appropriate mixture of compressed CO2 (100% CO2;
Praxair) and ambient air (drawn from outside the building) from an air compressor. Control tanks were bubbled
Content courtesy of Springer Nature, terms of use apply. Rights reserved

8
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
www.nature.com/scientificreports/
with ambient compressed air. Temperature was measured 5x per week using a combination of YSI (YSI Pro 30)
and mercury thermometer. Seawater pH was measured 3–4x per week using Oakton pH 450 (two-point calibra-
tion with saltwater buers AMP and TRIS, pH 6.77 and 8.09 respectively at 25 C). YSI was also used to monitor
salinity.
Water samples for carbonate system parameters were collected bimonthly and stored with 10 μL Mercuric
Chloride 5% (w/v) Aqueous for future analysis. Dissolved inorganic Carbon (DIC) was measured using DIC
Analyzer Model AS-C3 Apollo Sci Tech, according to guidelines of the Standard Operating Procedure 248. ree
replicates of 0.75 mL were taken for each sample. Results were normalized to a Certied Reference Material (CRM
Batch No. 154) supplied by Prof. Andrew Dickson (Scripps Institution of Oceanography). Full carbonate system
parameters were conducted on the control and acidied treatments using CO2SYS49 (Table2).
Sponge pumping and tissue withdrawal. Pumping was monitored on days 16, 31, 50, 71, 92, and 120
from the start of the experiment. Two milliliters of freshly made uorescent Calcein dye (4 g/L; Syndel Laboratories
Ltd), a uorescent derivative of uorescein, was injected with a pipette positioned 0.5 cm from the sponge wall and
halfway down the sponge’s structure to measure pumping capacity, which was quantied by calculating the time
it took the dye to be expelled from the oscula (referred to as ‘minimum residence time’ hereaer), and scoring the
density of the plume expelled from the oscula (‘pumping strength’ hereaer). We calculated minimum residence
time as the amount of time in between dye injection and the emergence of the dye from the sponge osculum. Here,
the time it took for the dye to be expelled from a set distance was calculated not by noting the time the dye appeared
above the oscula (like in the dye front method50) but rather when it appeared at the edge of it (distance = 0 mm from
oscula) therefore a ow rate could not be calculated. We consider minimum residence time a proxy measurement
for pumping rate as we were unable to calculate pumping rate directly in such small specimens. In preliminary trials
we observed that sponges with similar minimum resident times diered signicantly in the shape of the exhalent
dye plume (see examples in Supplementary Videos S1). Consequently, we added a measurement we refer to as
“pumping strength”. Video of the sponges pumping were recorded with Sony Handycam so as to precisely measure
the minimum residence time. Average minimum residence time for each treatment did not include those sponges
that were not pumping. Pumping strength was scored over a gradient of 0–6: ‘weak’ = a diuse (score 1–3) and
‘strong’ = dense (4–6) plume of dye, and ‘none’ = apparent pumping arrest (scored 0). e term ‘Apparent pumping
arrest’ does not correspond to pumping arrest because a owmeter was not used to record this, it does however infer
that pumping was so weak that it was not observed with the use of a dye. Scores were determined by an unbiased
observer. e quantity of dye expelled, speed at which this dye was ejected, and continuous versus pus of dye were
all factors considered when scoring a plume. e presence/absence of tissue withdrawal was monitored daily until
rst signs of tissue withdrawal appeared. Withdrawn tissue was easily distinguished from healthy tissue by its trans-
lucent (colourless) nature whereas healthy tissue maintained its original beige or orange colour (Fig.4). Aer onset,
tissue withdrawal was monitored every two weeks, for the remainder of the experiment.
All sponges survived through to the end of the experiment, except one sponge in the control treatment that
died in the last month of the experiment. e nal time point of this sponge was excluded from the analyses
because its death was deemed to be caused by a microbial infection since the sponge died suddenly (within 24
hrs) despite pumping strongly and with no signs of tissue withdrawal, and developed strings of mucus in that 24
hrs period.
Mechanical properties. Skeleton breaking force per volume and modulus (stiness) was tested using a
standard compression method in a computer-interface tensometer (model 5500 R, Instron Corp., Canton, MA,
USA). Skeleton was selected halfway down the sponge and cut into square pieces (approx. 1 cm2). ese were
placed in the cross-beam of the instrument and a maximum force of 4 N was gradually applied (load rate = 0.25 N/
min; strain rate = 0.35 mm/s) to the skeleton until point of failure. Breaking force was recorded. Because thickness
diered by sponge (1.7–4.6 mm), measurements of sponge skeletal thickness and cross-sectional area were used
to standardize breaking force per volume. e compression surface consisted of a 3 mm diameter puncture probe.
Modulus was calculated by dividing material stress with strain (i.e. the slope of the stress-strain curve produced
during the compressive test). e average of 3–5 replicates was taken for analysis.
Statistical analyses. All statistical analyses were performed in R version 3.6.051 for Mac OS X. For all tests,
signicance was determined at p < 0.05. Data were transformed when necessary (as outlined below) to meet the
pH (+SD)*Te mp. ( +SD)*
(°C) DIC ( + SD)*
(μmol/kgSW) TA** (μmol/
kgSW) pCO2** (ppm)
Control 7.83 + 0.06 8.6 + 0.3 1928 + 32 1873 + 32 919 + 104
OA 7.62 + 0.08 8.6 + 0.3 2002 + 60 1894 + 59 1525 + 234
OW 7.86 + 0.06 10.4 + 0.4 na na na
OAW 7.62 + 0.07 10.3 + 0.4 2002 + 53 1897 + 49 1487 + 186
Table 2. Measured and calculated carbonate chemistry parameters for four treatment combinations: ambient
conditions (‘Control’), CO2-induced acidication (‘OA’), increased seawater temperature (‘OW’), and a
combination of both (‘OAW’). Parameters of carbonate seawater chemistry (total alkalinity (TA) and pCO2)
were calculated from measured dissolved inorganic carbon (DIC), pH, temperature, and salinity values using
CO2SYS. SW – seawater; *Directly measured (n = 8 per treatment); **Calculated.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

9
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
www.nature.com/scientificreports/
assumptions of normality and equal variance. All mixed models were performed with package ‘nlme’52. Visual
inspection of standard model validation graphs was used to verify model assumptions: residuals versus tted val-
ues were used to verify homogeneity; a histogram or Quantile–Quantile (q–q) plot of the residuals for normality;
and residuals versus each explanatory variable to check independence. Fixed and random predictor variables
lacking explanatory power were eliminated via AIC model selection53. Optimal model structures were obtained
rst for random eects and next for xed eects. To account for repeated measures (on individuals) through time,
an autoregressive (AR1) correlation structure was included in the model during the selection process. In all cases,
the most parsimonious model was selected (i.e., the one without interactions and/or fewer terms, least complex
random structure, and removal of correlation structure where necessary).
Generalized linear mixed models (GLMM) were used to analyse apparent pumping arrest (GLMM, binomial
distribution) and pumping strength (GLMM, Poisson distribution). e best model included a random intercept
in which variation around that intercept depended on the sponge (nested within tank). Linear models (LM)
were used on log-transformed minimum residence time, breaking force per volume, and modulus datasets. A
“Kaplan-Meier survival curve” was generated using two pieces of data: status of nal observation and time to
event (here, rst sighting of tissue withdrawal). A Cox proportional hazards regression model was used to test
for the eects of acidication, warming, and their interaction on the probability of observing onset of with-
drawn tissue through time, using package ‘survival’54,55. e Cox proportional hazard assumption was tested
using Schoenfeld residuals with the function cox.zph. Sponges that did not show signs of tissue withdrawal were
censored (considered to not have reached tissue withdrawal).
Received: 12 December 2019; Accepted: 30 April 2020;
Published: xx xx xxxx
References
1. Bell, J. J. Functional roles of marine sponges. Estuar. Coast Shelf Sci. 79, 341–353 (2008).
2. De Goeij, J. M., Van Den Berg, H., Van Oostveen, M. M., Epping, E. H. & Van Duyl, F. C. Major bul dissolved organic carbon
(DOC) removal by encrusting coral reef cavity sponges. Mar. Ecol. Prog. Ser. 357, 139 (2008).
3. McMurray, S. E., Johnson, Z. I., Hunt, D. E., Pawli, J. . & Finelli, C. M. Selective feeding by the giant barrel sponge enhances
foraging eciency. Limnol. Oceanogr. 61(4), 1271–1286 (2016).
4. Diaz, M. C. & ützler, . Sponges: an essential component of Caribbean coral reefs. Bull. Mar. Sci. 69, 535–546 (2001).
5. Archer, S. ., Stevens, J. L., ossi, . E., Matterson, . O. & Layman, C. A. Abiotic conditions drive signicant variability in nutrient
processing by a common Caribbean sponge, Ircinia felix. Limnol. Oceanogr. 62, 1783–1793 (2017).
6. Dunham, A. et al. Assessing condition and ecological role of deep-water biogenic habitats: Glass sponge reefs in the Salish Sea. Mar.
Environ. es. 141, 88–99 (2018).
7. Conway, . W., Barrie, J. V., Austin, W. C. & Luternauer, J. L. Holocene sponge bioherms on the western Canadian continental shelf.
Cont. Shelf. es. 11(8–10), 771–90 (1991).
8. rautter, M., Conway, . W., Barrie, J. V. & Neuweiler, M. Discovery of a “living dinosaur”: globally unique modern hexactinellid
sponge reefs o British Columbia, Canada. Facies 44, 265–82 (2001).
9. Van Soest, . W. M. et al. Global Diversity of Sponges (Porifera). PLoS ONE 7, e35105 (2012).
10. Chu, J. W. F., Maldonado, M., Yahel, G. & Leys, S. P. Glass sponge reefs as a silicon sin. Mar. Ecol. Prog. Ser. 441, 1–14 (2011).
11. ahn, A. S., Yahel, G., Chu, J. W. F., Tunniclie, V. & Leys, S. P. Benthic grazing and carbon sequestration by deep-water glass sponge
reefs. Limnol. Oceanogr. 60(1), 78–88 (2015).
12. Pörtner, H. O. Ecosystem eects of ocean acidication in times of ocean warming: a physiologists view. Mar. Ecol. Prog. Ser. 373,
203–217 (2008).
13. Massaro, A. J. Selective ltration in the tropical marine sponge hopaloeides odorabile: impacts of elevated seawater temperature on
feeding behavior. Independent Study Project (ISP) Collection 774 (2009).
14. Wissha, M., Schönberg, C. H., Form, A. & Freiwald, A. Ocean acidication accelerates reef bioerosion. PLoS One 7(9), e45124
(2012).
Figure 4. Translucent segment in lower half of Aphrocallistes vastus shows withdrawn tissue. Top (beige) half of
sponge is comprised of living, healthy tissue.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

10
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
www.nature.com/scientificreports/
15. Ducworth, A. . & Peterson, B. J. Eects of seawater temperature and pH on the boring rates of the sponge Cliona celata in scallop
shells. Mar. Biol. 160, 27–35 (2013).
16. Go odwin, C., odolfo Metalpa, ., Picton, B. & Hall-Spencer, J. M. Eects of ocean acidication on sponge communities. Mar Ecol.
35, 41–49 (2014).
17. Vicente, J., Silbiger, N. J., Becley, B. A., aczowsi, C. W. & Hill, . T. Impact of high pCO2 and warmer temperatures on the
process of silica biomineralization in the sponge Mycale grandis. ICES J. Mar. Sci. 73(3), 704–714 (2016).
18. Bates, T. E. M. & Bell, J. J. esponses of two temperate sponge species to ocean acidication. New Zeal. J. Mar. Fresh. 52(2), 247–263
(2018).
19. iisgård, H. U., omassen, S., Jaobsen, H., Wees, J. M. & Larsen, P. S. Suspension feeding in marine sponges Halichondria panicea
and Haliclona urceolus: eects of temperature on ltration rate and energy cost of pumping. Mar. Ecol. Prog. Ser. 96(2), 177–188
(1993).
20. Liu, G., Feng, Q. & Gu, S. Extinction pattern and process of siliceous sponge spicules in deep-water during the latest Permian in
South China. Sci. China Ser. D. 51, 1623–1632 (2008).
21. Clarson, M. O. et al. Ocean acidication and the Permo-Triassic mass extinction. Science 348, 229–232 (2015).
22. Morrow, . M. et al. Natural volcanic CO2 seeps reveal future trajectories for host-microbial associations in corals and sponges.
ISME Journal 9, 894–908 (2014).
23. Fillinger, L., Janussen, D., Lundälv, T. & ichter, C. apid glass sponge expansion aer climate-induced Antarctic ice shelf collapse.
Current Biology 23, 1330–1334 (2013).
24. Tompins-MacDonald, G. J. & Leys, S. P. Glass sponges arrest pumping in response to sediment: implications for the physiology of
the hexactinellid conduction system. Marine Biology 154(6), 973 (2008).
25. Grant, N., Matveev, E., ahn, A. S. & Leys, S. P. Suspended sediment causes feeding current arrests in situ in the glass sponge
Aphrocallistes vastus. Mar. Env. es. 137, 111–120 (2018).
26. Massaro, A. J., Weisz, J. B., Hill, M. S. & Webster, N. S. Behavioral and morphological changes caused by thermal stress in the great
barrier reef sponge hopaloeides odorabile. J. Exp. Mar. Biol. Ecol. 416–417, 55–60 (2012).
27. Leys, S. e choanosome of hexactinellid sponges. Invertebr. Biol. 118, 221–235 (1999).
28. ibes, M., Coma, . & Gili, J. . Natural diet and grazing rate of the temperate sponge Dysidea avara (Demospongiae,
Dendroceratida) throughout an annual cycle. Mar. Ecol. Prog. Ser. 176, 179–190 (1999).
29. Leys, S. P. & Meech, . W. Physiology of coordination in sponges. Can. J. Zool. 84, 288–306 (2006).
30. Marliave, J. B., Borden, L. A., Schultz, J. A., Gibbs, D. M. & Dennison, G. J. Formation, persistence, and recovery of glass sponge
reefs: A case study. In Invertebrates-Ecophysiology and Management, InTechOpen: Aquatic Invertebrates. https://doi.org/10.5772/
intechopen.82325 (2018).
31. NOAA/National Weather Service. El Niño/Southern oscillation (ENSO) diagnostic discussion. Source, https://www.cpc.ncep.noaa.
gov/products/analysis_monitoring/enso_disc_mar2015/ensodisc.html (2015)
32. Bennett, H. M. et al. Interactive eects of temperature and pCO2 on sponges: from the cradle to the grave. Glob. Change Biol. 23,
2031–2046 (2017).
33. Bennett, H., Bell, J. J., Davy, S. ., Webster, N. S. & Francis, D. S. Elucidating the sponge stress response; lipids and fatty acids can
facilitate survival under future climate scenarios. Glob. Change Biol. 24(7), 3130–3144 (2018).
34. Marliave, J. B. Cloud Sponge, Aphrocallistes vastus (Porifera: Hexactinellida), fragment healing and reattachment. Can. Field Nat.
129(4), 399–402 (2015).
35. Marliave, J. B., Gibbs, D. M., Young, S. & Borden, L. 2011 climate regime: revealed by seabed biodiversity. In: P.C., Chandler, S.A.,
ing, and J., Boldt (Eds.). State of the physical, biological and selected shery resources of Pacic Canadian marine ecosystems in
2017. Can. Tech. ep. Fish. Aquat. Sci. 3266: viii + 245 p (2018).
36. rautter, M., Conway, . W. & Barrie, J. V. ecent hexactinosidan sponge reefs (silicate mounds) o British Columbia, Canada:
frame-building processes. J. of Paleont. 80, 38–48 (2006).
37. Ehrlich, H. & Worch, H. Sponges as natural composites: from biomimetic potential to development of new biomaterials. In: E.,
Hajdu (Ed.). Porifera esearch: Biodiversity, Innovation & Sustainability, 217–223 (2007).
38. Ehrlich, H. et al. First evidence of the presence of chitin in seletons of marine sponges. Part II. Glass sponges (Hexactinellida:
Porifera). J. Exp. Zool. Part B 308B, 473–483 (2007).
39. E hrlich, H. et al. Calcite reinforced silica–silica joints in the biocomposite seleton of the deep–sea glass sponge. Adv. Funct. Mater.
21, 3473–3481 (2011).
40. Chu, J. W. F. Biological patterns and processes of glass sponge reefs. MSc thesis, University of Alberta (2010).
41. Marliave, J. B., Conway, . W., Gibbs, D. M., Lamb, A. & Gibbs, C. Biodiversity and rocsh recruitment in sponge gardens and
bioherms of southern British Columbia, Canada. Mar. Biol. 156, 2247–2254 (2009).
42. Pawli, J. . & McMurray, S. E. e emerging ecological and biogeochemical importance of sponges on coral reefs. Annu. ev. Mar.
Sci. 12, 3.1–3.23 (2020).
43. ahn, A., Chu, J. W. F. & Leys, S. P. Trophic ecology of glass sponge reefs in the Strait of Georgia, British Columbia. Sci. ep. 8(756),
1–11 (2018).
44. Peterson, B. J., Chester, C. M., Jochem, F. J. & Fourqurean, J. W. Potential role of sponge communities in controlling phytoplanton
blooms in Florida Bay. Mar. Ecol. Prog. Ser. 328, 93–103 (2006).
45. Sunday, J. M. et al. Ocean acidication can mediate biodiversity shis by changing biogenic habitat. Nat. Clim. Change 7, 81–86
(2017).
46. Harley, C. D. G. Climate change, eystone predation, and biodiversity loss. Science 334, 1124–1127 (2011).
47. IPCC. Climate Change 2014: Synthesis eport. Contribution of Woring Groups I, II and III to the Fih Assessment eport of the
Intergovernmental Panel on Climate Change [Core Writing Team, .. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland,
151 pp. (2014).
48. Dicson, A. G., Sabine, C. L. & Christian, J. . Guide to Best Practices for Ocean CO2 Measurements. PICES Special Publication 3,
191 (2007).
49. Pierrot, D., Lewis, E. & Wallace, D. W. . MS Excel Program Developed for CO2 System Calculations. ONL/CDIAC-105a. Carbon
Dioxide Information Analysis Center. Oa idge National Laboratory, U.S. Department of Energy, Tennessee (2006).
50. Morganti, T., Yahel, G., ibes, M. & Coma, . VacuSIP, an improved InEx method for in situ measurement of particulate and
dissolved compounds processed by active suspension feeders. JoVE 114, e54221 (2016).
51.  Core Team. : A language and environment for statistical computing.  Foundation for Statistical Computing, Vienna, Austria.
UL, https://www.-project.org/ (2017).
52. Pinheiro, J., Bates, D., Deboy, S., Sarar, D. &  Core Team. nlme: Linear and Nonlinear Mixed Eects Models.  pacage version
3.1-139, UL, https://CAN.-project.org/pacage=nlme (2019).
53. Zuur, A., Ieno, E. N., Waler, N., Saveliev, A. & Smith, G. Mixed Eects Models and Extensions in Ecology with . Springer (2009).
54. erneau, T. A Pacage for Survival Analysis in S. version 2.38, UL, https://CAN.-project.org/pacage=survival (2015).
55. erneau, T. M. & Grambsch, P. M. Modeling Survival Data: Extending the Cox Model. Springer, New Yor. ISBN 0-387-98784-3
(2000).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

11
SCIENTIFIC REPORTS | (2020) 10:8176 | https://doi.org/10.1038/s41598-020-65220-9
www.nature.com/scientificreports
www.nature.com/scientificreports/
Acknowledgements
is work would not have been possible without the dedicated eorts of Maya Brassard, Mikalyn Trinca Colonel,
Tadhg O Corcora, Emma Foxcro, Josianne Haag, Robert Hechler, Dr. Kyra Janot, Carli Jones, Madison Kaplan,
Lauran Liggan, Dr. Colin MacLeod, and Charlotte Matthews in the lab and Donna Gibbs and Lauran Liggan in
the eld. We thank Sheila Byers, Glen Dennison, and Prof. Sally Leys for their insights, and Dr. Alyssa Gehman
and Dr. Devin Lyons for their comments on earlier stages of the manuscript and statistical analyses. is work
was supported by the MEOPAR and NSERC CREATE Training Our Future Ocean Leaders Program postdoctoral
fellowships to A. Stevenson, a Natural Sciences and Engineering Research Council Discovery Grant and Canada
Foundation for Innovation grant to C. Harley, and the Howe Sound Research and Conservation Group donors of
Ocean Wise Research Institute.
Author contributions
A.S., S.K.A., A.D., J.B.M., C.D.G.H. devised the study. A.S. and C.D.G.H. designed the mesocosm experiments.
A.S. and P.M. designed the biomechanics experiments. A.S. and J.A.S. planned and conducted eld work. A.S.
conducted experiments and laboratory work. A.S. wrote the manuscript. All authors contributed intellectual
input, edited and approved this manuscript.
Competing interests
e authors declare no competing interests.
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/s41598-020-65220-9.
Correspondence and requests for materials should be addressed to A.S.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
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 Cre-
ative Commons license, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not per-
mitted 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 license, visit http://creativecommons.org/licenses/by/4.0/.
© e Author(s) 2020
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