Available via license: CC BY-NC 4.0
Content may be subject to copyright.
ECOLOGY Copyright © 2019
The Authors, some
rights reserved;
exclusive licensee
American Association
for the Advancement
of Science. No claim to
original U.S. Government
Works. Distributed
under a Creative
Commons Attribution
NonCommercial
License 4.0 (CC BY-NC).
Disease epidemic and a marine heat wave are
associated with the continental-scale collapse of a
pivotal predator (Pycnopodia helianthoides)
C. D. Harvell
1
*
†
, D. Montecino-Latorre
2
*, J. M. Caldwell
3
, J. M. Burt
4,5
, K. Bosley
6
, A. Keller
7
,
S. F. Heron
8,9,10
, A. K. Salomon
4,5
, L. Lee
4,5
, O. Pontier
5
, C. Pattengill-Semmens
11
, J. K. Gaydos
12
Multihost infectious disease outbreaks have endangered wildlife, causing extinction of frogs and endemic birds,
and widespread declines of bats, corals, and abalone. Since 2013, a sea star wasting disease has affected >20 sea
star species from Mexico to Alaska. The common, predatory sunflower star (Pycnopodia helianthoides), shown to
be highly susceptible to sea star wasting disease, has been extirpated across most of its range. Diver surveys
conducted in shallow nearshore waters (n= 10,956; 2006–2017) from California to Alaska and deep offshore
(55 to 1280 m) trawl surveys from California to Washington (n= 8968; 2004–2016) reveal 80 to 100% declines
across a ~3000-km range. Furthermore, timing of peak declines in nearshore waters coincided with anomalously
warm sea surface temperatures. The rapid, widespread decline of this pivotal subtidal predator threatens its per-
sistence and may have large ecosystem-level consequences.
INTRODUCTION
Host-pathogen theorypredicts that multihost pathogens can cause ex-
treme population impacts, including extinction of susceptible species
if they are continuously infected from reservoir species (1,2). For ex-
ample, introduced multihost pathogens such as avian malaria and
avian pox have driven multiple native Hawaiian bird species to extinc-
tion (3). Similarly, the Batrachochytrium dendrobatidis pandemic may
have caused hundreds of species extinctions worldwide and decimated
more than 38 amphibian species in Central America (4,5). In another
example, spillover of shared pathogens from domesticated bees to wild
bumblebees(e.g., deformed wing virus and Nosema ceranae)isdriving
declines in the wild populations (6). These pathogen-associated im-
pacts are further exacerbated in a changing climate (7–9).
Since 2013, sea star wasting disease (SSWD) has caused massive,
ongoing mortality from Mexico to Alaska (known as the Northeast
Pacific SSWD event). In particular, the epidemic phase of the North-
east Pacific SSWD event (2013–2015) was notably different from pre-
vious events elsewhere in terms of its geographic extent, persistence,
involvement of multiple species, symptoms in reproductive stars, and
the extremely rapid progression of disease to death (10–15). More than
20 asteroid species have been affected in what is currently the largest
documented epizootic of a noncommercial marine taxon (13,14,16).
Diseased sea stars develop progressively worse dermal lesions (13,14),
arms detach from the central disc, gonads spilled from fully reproduc-
tive stars and individuals die, often leaving white piles of ossicles and
disconnected limbs (fig. S1). Sea star mortalities during the first years
of the Northeast Pacific SSWD event (2013–2015) were linked to a sea
star–associated densovirus (SSaDV; family Parvoviridae), based on me-
tagenomic analysis of bacteria and viruses in field samples, the experi-
mental generation of disease in sea stars challenged with nonheated
viral-sized material, the correlation between SSWD progression and
SSaDV loads, and higher SSaDV prevalence in symptomatic stars
(13). Further support for densovirus involvement in the Northeast
Pacific SSWD event includes experimental infection and morbidity
in Pycnopodia helianthoides (17), the relationship between asteroid
densovirus load and SSWD in P. helianthoides,andexperimental
transmission of SSWD disease to asymptomatic individuals through
exposure to a viral-sized agent (0.22 mm) from SSWD symptomatic
stars (17,18).
While the Northeast Pacific SSWD event caused severe reduc-
tions of the keystone intertidal ochre sea star across its entire west-
coast range (14,15,19–21), the impacts on subtidal species are less
wellknown.Becauseitwasthemostabundantsubtidalstarand
also the most susceptible in early reports, initial studies quantified
devastating declines of P. helianthoides,inWashingtonandBritish
Columbia, linked with high prevalence of wasting disease (22,23),
but these observations are limited to the initial years, specific regions,
and shallow depths accessible to divers. Hence, the current status of
P. helianthoides in the Northeast Pacific and at all depths is un-
known. This species is an important predator of sea urchins, and ur-
chin populations released from top-down predatory control can
expand and threaten kelp forests and biodiversity (24,25). In many
locations, P. helianthoides is the apex subtidal predator when other
urchin predators are absent (22,26). In these locations, its decline has
triggered a trophic cascade, causing urchin populations to explode and
kelp to rapidly diminish (22). If P. helianthoides reductions occur
throughout its range at the magnitude reported in local surveys, and
at greater depths, this disease could not only threaten the long-term
persistence of this species but also have wide-ranging cascading eco-
system effects.
1
Department of Ecology and Evolutionary Biology, Cornell University, Ithaca, NY
14853, USA.
2
One Health Institute, School of Veterinary Medicine, University of
California, Davis, CA 95616, USA.
3
Department of Biology, Stanford University,
Stanford, CA 94040, USA.
4
School of Resource and Environmental Management,
Simon Fraser University, Burnaby, BC V5A 1S6, Canada.
5
Hakai Institute, Heriot
Bay, BC V0P 1H0, Canada.
6
Fishery Resource Analysis and Monitoring Division,
Northwest Fisheries Science Center, National Marine Fisheries Service, National
Oceanic and Atmospheric Administration (NOAA), 2032 SE OSU Drive, Newport,
OR 97365, USA.
7
Fishery Resource Analysis and Monitoring Division, Northwest
Fisheries Sc ience Center, National Ma rine Fisheries Service, N OAA, 2725 Montlake
Boulevard East, Seattle, WA 98112, USA.
8
NOAA Coral Reef Watch, College Park,
MD 20740, USA.
9
ReefSense Pty Ltd., Townsville, Queensland, Australia.
10
Marine
Geophysical Laboratory, Physics, College of Science and Technology, James Cook
University, Townsville, Queensland, Australia.
11
Reef Environmental Education
Foundation (REEF), Key Largo, FL 33037, USA.
12
The SeaDoc Society, Karen C.
Drayer Wildlife Health Center–Orcas Island Office, University of California, Davis,
942 Deer Harbor Road, Eastsound, WA 98245, USA.
*These authors contributed equally to the work and supervision of the project.
†Corresponding author. Email: cdh5@cornell.edu
SCIENCE ADVANCES |RESEARCH ARTICLE
Harvell et al., Sci. Adv. 2019; 5: eaau7042 30 January 2019 1of8
on January 31, 2019http://advances.sciencemag.org/Downloaded from
Increasingly warm or anomalous temperatures are being shown
to influence the prevalence and severity of marine infectious diseases
[e.g., (8)]. Experimental and field studies support a role for tempera-
ture in SSWD morbidity. Current evidence suggests that at warmer
temperatures, Pisaster ochraceus have a higher risk of infection and
progression to mortality (10,14,15,27). If this relationship applies to
other sea star species, then we expect that declines of P. helianthoides
populations will be associated with warmer temperature exposure.
Here, we investigated the current status of P. helianthoides in both
shallow nearshore and deep offshore waters, from California to British
Columbia (~3000 km), using data from over a decade of complemen-
tary survey methods that cover from pre- to post-outbreak periods
of the Northeast Pacific SSWD event. We report the rapid collapse of
P. helianthoides populations along most of its range after the onset of
the Northeast Pacific SSWD event, and at all depths, confirming the
lack of a deep-water refuge for this species. Furthermore, to explore
the hypothesis that warmer waters may be linked to the decline, we
assessed the relationship between P. helianthoides abundance in
shallow nearshore waters and sea surface temperature (SST). We de-
tected a negative association between P. helianthoides abundance
and anomalously warm SST. Potential ecosystem impacts of this
decimation are discussed.
RESULTS
To assess P. helianthoides decline in deep offshore waters, we esti-
mated the average yearly biomass (kg/10 ha) collected in 8968 bottom
trawls (55 to 1280 m depth) conducted from California to Washington
between 2004 and 2016. Deep-water trawl surveys showed fluctuat-
ing but constant P. helianthoides biomass up to 2011–2012 and an
unprecedented decline after the onset of the Northeast Pacific SSWD
event. In California and Oregon, the average biomass decreased 100%
during 2013–2015 (from 2.78 and 1.73 kg/10 ha, respectively). In
Washington, average biomass declined 99.2% (from 3.11 to 0.02 kg/10 ha)
during this period. In 2016, no P. helianthoides were collected across
the 1264-ha area covered by 692 trawl surveys. The collapse in biomass
collection occurred 1 year earlier in California compared with other
regions (2014; Fig. 1).
To quantify the impact of the Northeast Pacific SSWD event on
P. helianthoides in shallow nearshore waters, we analyzed changes
in abundance reported in 10,956 roving-diver surveys conducted be-
tween 2006 and 2017 from California to Alaska (sparse survey cov-
erage for Alaska not plotted). Abundance categories (ACs; 0 to 4
corresponding to 0, 1, 2 to 10, 11 to 100, and >100 individuals, re-
spectively) were analyzed within regional jurisdictions as an annual
abundance score (28). Furthermore, we report nearshore biomass of
P. helianthoides (kg/10 m
2
) along the coast of central British Columbia
using annual subtidal belt transect surveys conducted between 3 and
18 m depth in 2010–2011 and again between 2013 and 2017.
In the years before the Northeast Pacific SSWD event onset, ACs
2 (2 to 10 stars) and 3 (11 to 100 stars) were the most commonly
reported (64 to 80% of the reports), but since 2014 (after the epi-
demic onset), at least 60% of surveys across the study area and up to
100% in California and Oregon report declines to ACs 0 and 1 (Fig. 2).
Belt transect surveys in central British Columbia also showed that
P. helianthoides biomass declined by ~96% (from 0.57 to 0.93 kg/10 m
2
in 2010–2014, pre-Northeast Pacific SSWD event) to virtually zero
(0.01 to 0.07 kg/10 m
2
) in 2015–2017. The abundance score from the
shallow nearshore waters surveys revealed fluctuating but regionally
constant P. helianthoides abundance during 2006–2013 and a con-
sistent continental-wide decline after the onset of the Northeast Pa-
cific SSWD event (Fig. 2).
To assess the relationship between P. helianthoides declines at shal-
low nearshore waters and SST, we modeled the ACs reported for this
species in the roving-diver surveys as a function of a satellite-derived
SST anomaly metric and days since the SST metric was observed. Be-
tween 2013 and 2015, SST anomalies (departures from the climatolo-
gically expected temperature calculated from 1985 to 2012) were
warm at all locations, but their magnitude and duration varied with
latitude (Fig. 3). In California, 2014 was anomalously warm, increas-
ing to extreme warming throughout 2015 with a peak anomaly of 4°C.
In Washington, 2013 was anomalously warm during July to October
but otherwise followed the seasonal climatology. In July 2014, a pro-
longed warm anomaly began, increasing to an extreme 2.5°C anomaly
through 2015, coinciding with the long residence of the heat wave in
the northeast Pacific Ocean (29). Central British Columbia followed a
similar pattern seen in Washington, with the arrival of the anomalous-
ly warm waters in the fall of 2015 (30). Oregon was more variable than
other locations, likely because of periodic cold upwelling events.
On the basis of ordinal regression models of P. helianthoides abun-
dance and biologically relevant SST anomaly metrics, we provide
evidence that P. helianthoides declines in shallow nearshore waters
were associated with the maximum temperature anomaly exposure
from within 60 days before each survey (tables S1 and S2). Our
selected model indicates that with every 1°C increase in the maxi-
mum temperature anomaly, we would expect a 6% increase in the
log odds of observing a low AC compared with all higher ACs (AC 0
versus 1 to 4, AC 1 versus 2 to 4, etc.), when all other variables are
held constant (table S2). We evaluated the goodness of fit between
the model and data using Nagelkerke’spseudoR
2
(31)andesti-
mated that our model explained 68.5% of the variance in sea star
ACs compared to a null model. To investigate the role of the max-
imum temperature anomaly exposure from within 60 days before
the survey, we also calculated a pseudo R
2
for a model that included
all covariates other than this variable. This model explained 30.5%
ofthevarianceinP. helianthoides ACs, suggesting that this SST
anomaly metric alone explains ~38% of the variance.
DISCUSSION
Using longitudinal data from complementary survey methods, we show
(i) consistent or slightly increasing populations of P. helianthoides
across most of its natural range in the decade before the Northeast
Pacific SSWD event in shallow nearshore waters; (ii) consistent popu-
lations of P. helianthoides across most of its natural range in the dec-
ade before the Northeast Pacific SSWD event in deep offshore waters;
(iii) a continental-scale collapse of this major ocean predator in these
habitats associated temporally and spatially to the Northeast Pacific
SSWD event and, therefore, to the multihost SSaDV; and (iv) an as-
sociation between P. helianthoides declines in shallow nearshore
waters and period of anomalously warm nearshore waters.
Our statistical analysis provides evidence for anomalous tem-
perature as a key facilitator of the disease-related declines in the shal-
low nearshore waters, explaining more than a third of the variance by
itself. These results align with field and experimental evidence of the
interacting roles of temperature and SSWD in P. ochraceus morbid-
ity and mortality, where infected stars exposed to warmer tempera-
tures died at a faster rate (10,14,15,27). Previous studies suggest that
SCIENCE ADVANCES |RESEARCH ARTICLE
Harvell et al., Sci. Adv. 2019; 5: eaau7042 30 January 2019 2of8
on January 31, 2019http://advances.sciencemag.org/Downloaded from
high water temperatures are associated with lower coelomic fluid vo-
lumes, higher metabolic demands, and metabolic stress in asteroids
(10,32–34), making the case for how a viral epidemic could be ex-
acerbated in an invertebrate with a limited immune response capabil-
ity (35). Although warming waters likely accelerated and increased the
scale of disease-induced morbidity, P. helianthoides mortality still
occurred at high levels in colder temperatures of British Columbia.
This finding supports a facilitating role of anomalously warm tem-
perature for disease morbidity as previously reported in intertidal
P. ochraceus during the Northeast Pacific SSWD event (15).
Our data document the widespread decline of P. helianthoides at
depths beyond shallow nearshore waters, confirming the lack of a
deep-water refuge for this species. Available data support that the
Northeast Pacific SSWD event is the most parsimonious explanation
forthiscollapse.Thisspecieswasidentifiedasthemostsusceptibleto
SSWD (13,14), the observed widespread declines occurred right after
the onset of this event, and they followed the timing and spatial pat-
tern of the declines observed in nearshore P. helianthoides popula-
tions and intertidal P. ochraceus populations (15). Moreover,
necrotic stars with autotomizing arms were observed by one of the
coauthors (K. Bosley) in the trawls at the initiation of the Northeast
Pacific SSWD event. The limitations of deep offshore water tempera-
ture data prevented an analysis of the role of water temperature in
exacerbating P. helianthoides declines at deep offshore waters, as this
Fig. 1. Continental collapse of a pivotal predator: Deep offshoresurveys. Mean biomass of sunflower star in 8968 deep offshore trawls (55 to 1280 m) from (A) Washington,
(B) Oregon, and (C) California from 2004 to 2016 with 95% confidence interval in light blue. Gray line marks the year 2013 for comparison of SSWD initiation across regions.
Yellow circles depict the 2013–2016 trawl locations. The trawls per jurisdiction per year are shown in the top of each plot.
SCIENCE ADVANCES |RESEARCH ARTICLE
Harvell et al., Sci. Adv. 2019; 5: eaau7042 30 January 2019 3of8
on January 31, 2019http://advances.sciencemag.org/Downloaded from
Fig. 2. Continental collapse of a pivotal predator: Shallow nearshore surveys. (Ato D) Percentage of shallow nearshore ACs of sunflower star (P. helianthoides)
reported in roving-diver surveys from southern California to southern British Columbia, Canada, from 2006 to 2017 (blue scale bars, right axis). Black line, annual
abundance score (left axis); red line, annual mean of the maximum temperature anomaly 60 days before each survey (whiskers, 95% confidence interval; left axis).
(A) British Columbia. (B) Washington. (C) Oregon. (D) California. (E) Mean biomass (kg/10 m
2
) in belt transect surveys in central British Columbia, with 95% confidence
interval in light blue. Yellow circles depict the 2013–2017 locations. The red rectangle depicts the area where the belt transect surveys were conducted. The surveys per
jurisdiction per year are shown in the top of each plot. For other details, see Fig. 1.
SCIENCE ADVANCES |RESEARCH ARTICLE
Harvell et al., Sci. Adv. 2019; 5: eaau7042 30 January 2019 4of8
on January 31, 2019http://advances.sciencemag.org/Downloaded from
information is collected once annually. However, new findings show
that anomalous warm temperatures associated with recent marine
heat waves were broadly detected at about 140 m depth in the North-
east Pacific beginning in mid-2014 (36), which may have contributed
to SSWD morbidity at these depths.
The abrupt decline of P. helianthoides after the onset of the North-
east Pacific SSWD event occurred in shallow and deep waters regard-
less of the abundance before the collapse. Similarly, Miner et al.(15)
did not detect a relationship between the degree of population decline
and pre-outbreak intertidal P. ochraceus density. The current popula-
tion status of P. helianthoides raises important questions about the
potential for recovery and persistence of this species. With SSWD,
reservoir species could be a continuous source of SSaDV to remaining
P. helianthoides given that asymptomatic star species within the range
of P. helianthoides have tested positive for SSaDV genetic material
(17). In its current status and with new bouts of mortality recorded
by divers in August 2018, continuous infection from reservoir popula-
tions or small stochastic disturbances could cause the restricted rem-
nant populations of P. helianthoides to vanish (1,2,37,38).
Cascading effects of the P. helianthoides loss are expected across
its range and will likely change the shallow water seascape in some
locations and threaten biodiversity through the indirect loss of kelp
(22,25,26,39). P. helianthoides was the highest biomass subtidal as-
teroid across most of its range before the Northeast Pacific SSWD
event (39). Loss or absence of this major predator has already been
associated with elevated densities of green (Strongylocentrotus
droebachiensis), red (Mesocentrotus franciscanus), and purple urch-
ins (Strongylocentrotus purpuratus) across their range (22,23,26,39),
even in regions with multiple urchin predators (40). Associated kelp re-
ductions have been reported following the outbreak (22,39). Examples
from other widespread marine diseases—the near extirpation of the in-
tertidal sea star Heliaster kubiniji from Gulf of California (40), the mass
mortality of the urchin Diadema antillarum from Caribbean reefs (41),
and the withering syndrome–influenced endangerment of multiple
California abalone species (42)—demonstratehowthelossofkeyspecies
can drive community effects that influence marine ecosystem processes.
SSWD, the anomalously warm water, P. helianthoides declines, and
subsequent urchin explosions (fig. S2) have been described as the “per-
fect storm.”This “storm”could result not only in trophic cascades and
reduced kelp beds (22) but also in abalone and urchin starvation (43).
We encourage scientific review of P. helianthoides recovery, assessment
of the probability for endangerment, and, more generally, expanded
surveillance ofthe consequences stemming from complex interactions
that emerging infectious diseases and warming ocean temperatures can
have in shaping ecosystems, even in the deep ocean.
MATERIALS AND METHODS
Study area
This study includes data from across the northeast Pacific, encom-
passing most of the natural range of the endemic sunflower sea star
P. helianthoides. We analyzed data from shallow nearshore roving-
diver surveys collected in California, Oregon, and Washington (USA)
and British Columbia (Canada), in addition to deep offshore trawl
surveys conducted from California to Washington.
Deep offshore surveys
The National Marine Fisheries Service [National Oceanic and Atmo-
spheric Administration (NOAA)] conducted 8968 individual bottom
Fig. 3. Ocean temperature anomaly averaged over the roving-diver survey
locations for the three initial years of the epidemic. Blue, 2013; green, 2014;
red, 2015. BC, British Columbia; WA, Washington; OR, Oregon; CA, California.
SCIENCE ADVANCES |RESEARCH ARTICLE
Harvell et al., Sci. Adv. 2019; 5: eaau7042 30 January 2019 5of8
on January 31, 2019http://advances.sciencemag.org/Downloaded from
trawls along the coasts of California, Oregon, and Washington be-
tween 2004 and 2016. These bottom trawls quantify the biomass of
P. helianthoides (kg) as well as the total area swept by the trawl (ha).
We estimated the mean kg/10 ha per state per year using bottom
trawls conducted between 55 and 1280 m depth. The 95% confidence
intervals of the yearly means were calculated.
Shallow nearshore surveys
Trained and tested recreational scuba divers searching for P. helianthoides
on the coast of Washington, Oregon, and California (USA) and north-
ern British Columbia (Canada) conducted 10,956 roving-diver surveys
between 2006 and 2017. After the surveys were completed, the abun-
dance of P. helianthoides was ranked between 1 and 4 corresponding
to estimated ACs of 1, 2 to 10, 11 to 100, or >100 individuals sighted,
respectively. This information was submitted to the Reef Environmental
Education Foundation (REEF) Volunteer Fish Survey Project Database
(28). With this information, we calculated P. helianthoides abundance
score as previously explained (23,44) per state and per year. Those
surveys that did not report P. helianthoides presence were assigned 0
abundance (category 0). Furthermore, we calculated the proportion that
each AC, including the category 0, was reported per state and per year. The
surveys were conducted in kelp forest, rock/shale reefs, open ocean, sea
grass beds (Phyllospadix and Zostera spp.), pinnacles, bull kelp beds
(Nereocystis sp.), cobblestone/boulder fields, and walls over 10 feet high.
The abundance and size of P. helianthoides were recorded on an-
nual scuba surveys using belt transects at 11 rocky reef sites located
on the central coast of British Columbia between 2010 and 2017 (3 to
15 m depth). Belt transects (30 m × 2 m, n= 6 per site) were con-
ducted at all 11 sites in 2013–2017, whereas 8 and 4 sites were sur-
veyed in 2011 and 2010, respectively (belt transects, 10 m × 2 m; n=3
to 9 per site). P. helianthoides biomass was calculated using a length-
to-biomass regression (45) and summed across transects at a site to
yield kg/10 m
2
.P. helianthoides biomass for the central coast region
(range from 51°24.612′N to 52°4.242′N) was calculated as the mean
across all site-year combinations. The 95% confidence intervals of
these yearly means were calculated.
Sea surface temperature
Satellite SST data were obtained at 0.05° (~5 km) daily resolution
from the CoralTemp product by NOAA Coral Reef Watch (NOAA
Coral Reef Watch 2018). Time series of SST were acquired at each
location of the roving-diver surveys or at the nearest neighboring
satellite pixel near the coastal boundary. SST anomalies describe
the variation in temperature from expected values at a given time
of year and location and were determined using the Coral Reef
Watch approach based on monthly climatologies (46). Where SST
is warmer (cooler) than expected, the SST anomaly is positive (neg-
ative). For each survey, maximum values of SST and SST anomaly
from several periods immediately prior were extracted (30, 60, 90,
180, and 360 days) for comparison with survey data. Jurisdictional
(California, Oregon, Washington, and British Columbia) SST anom-
aly summaries for each year were calculated by spatially averaging
60-day-prior maximum values from survey locations for that year
(Fig. 2). Jurisdictional SST and SST anomaly time series (Fig. 3
and fig. S3) were averaged across all survey locations.
Statistical analysis
We estimated the relationship between sea star abundance re-
ported in the shallow nearshore roving-diver surveys and SST by
fitting a hierarchical ordinal regression model with a probit link
function
PðYi≤jÞ¼qjb1ðSSTmetriciÞb2ðDays:SSTmetriciÞ
b313ðYear20072017iÞmðLatitudeiÞ
gðMonthiÞ
mðLatitudeiÞeNð0;s2
mÞ
gðMonthiÞeNð0;s2
gÞ
i=1,…, I surveys,
j=1,…,j−1 abundance categories,
where the cumulative probability of the ith survey falling in the jth AC
or below is modeled as a function of the following: q
j
threshold pa-
rameters across ACs, which provide a separate intercept for each
category j, an SST anomaly metric (SSTmetric
i
), days since the SST
anomaly metric was observed (Days.SSTmetric
i
), year (Year
2007–2017
),
month (Month
i
), and latitude (Latitude
i
).Yearwasincludedasafixed
effect to determine the year when sea star abundances collapsed (years
that were statistically significantly different from the 2006 baseline).
Month and latitude were included as random effects to account for ad-
ditional variation over time and space. s
2m
and s
2g
corresponded to the
variance of the distribution of month and latitude random effects, re-
spectively. Our data met all model assumptions: (i) the response variable
was measured on an ordinal scale; (ii) the predictor variables were con-
tinuous or categorical; (iii) there was no multicollinearity among predic-
tor variables, which we assessed with correlation tests for correlations
between two predictors and visually for correlations among three pre-
dictors; and (iv) there were proportional odds between each AC as in-
dicated by nearly identical effects among generalized logistic regression
models comparing each AC split individually (slopes < 2). We fit 10
candidate models that included the year, latitude, and month covariates
and one of the following SST metrics: the maximum SST in the 30, 60,
90, 180, or 360 days prior to each roving diver survey; or the maximum
anomalous SST in the 30, 60, 90, 180, or 360 days prior to each roving-
diver survey. We compared the AIC value of the candidate models with
and without the covariate “days since the SST metric was observed,”and
then selected the model with the lowest AIC value (tables S1 and S2). We
assessed convergence of models by inspecting the maximum absolute gra-
dient of the log-likelihood function and the magnitude of the Hessian.
Each model was empirically identifiable by ensuring that the condition
number of the Hessian measure was no larger than 10
4
(47). We evalu-
ated variance explained by the final model using Nagelkerke’spseudoR
2
(31). Nagelkerke’spseudoR
2
is a commonly used statistic to measure
goodness of fit that is calculated by comparing likelihood ratios between
a full model and an intercept model. We conducted this analysis in R
statistical software v3.4.3 (48)usingtheclmmfunctionofthe“ordinal”
package (47) for the ordinal regression model and the nagelkerke function
in the “rcompanion”package to calculate pseudo R
2
values (49).
SUPPLEMENTARY MATERIALS
Supplementary material for this article is available at http://advances.sciencemag.org/cgi/
content/full/5/1/eaau7042/DC1
Table S1. Summary of results of the candidate hierarchical ordinal regression models.
Table S2. Parameter estimates, SEs, and 95% confidence interval of the selected ordinal model
linking the reporting of ACs 0 to 4 in the shallow nearshore roving-diver surveys and
maximum temperature anomalies from within 60 days before each survey.
Fig. S1. Massive decline of P. helianthoides over 20 days between 9 and 29 October 2013.
SCIENCE ADVANCES |RESEARCH ARTICLE
Harvell et al., Sci. Adv. 2019; 5: eaau7042 30 January 2019 6of8
on January 31, 2019http://advances.sciencemag.org/Downloaded from
Fig. S2. Annual SST records during 2013, 2014, and 2015 by jurisdiction for British Columbia,
Washington, Oregon, and California.
Fig. S3. Sackinaw Rock before and after development of green urchin barrens following
decimation of P. helianthoides.
REFERENCES AND NOTES
1. H. McCallum, Disease and the dynamics of extinction. Philos. Trans. R. Soc. Lond. B Biol. Sci.
367, 2828–2839 (2012).
2. H. McCallum, A. Dobson, Detecting disease and parasite threats to endangered species
and ecosystems. Trends Ecol. Evol. 10, 190–194 (1995).
3. C. van Riper III, S. G. van Riper, W. R. Hansen, Epizootiology and effect of Avian Pox on
Hawaiian forest birds. Auk 119, 929–942 (2002).
4. J. Alroy, Current extinction rates for reptiles and amphibians. Proc. Natl. Acad. Sci. U.S.A.
112, 13003–13008 (2015).
5. K. R. Lips, F. Brem, R. Brenes, J. D. Reeve, R. A. Alford, J. Voyles, C. Carey, L. Livo, A. P. Pess ier,
J. P. Collins, Emerging infectious disease and the loss of biodiversity in a Neotropical
amphibian community. Proc. Natl. Acad. Sci. U.S.A. 103,3165–3170 (2006).
6. M. A. Fürst, D. P. McMahon, J. L. Osborne, R. J. Paxton, M. J. F. Brown, Disease associations
between honeybees and bumblebees as a threat to wild pollinators. Nature 506,
364–366 (2014).
7. J. Bosch, L. M. Carrascal, L. Durán, S. Walker, M. C. Fisher, Climate change and outbreaks
of amphibian chytridiomycosis in a montane area of Central Spain; is there a link?
Proc. Biol. Sci. 274, 253–260 (2007).
8. C. D. Harvell, C. E. Mitchell, J. R. Ward, S. Altizer, A. P. Dobson, R. S. Ostfeld, M. D. Samuel,
Climate warming and disease risks for terrestrial and marine biota. Science 296,
2158–2162 (2002).
9. S. Altizer, R. S. Ostfeld, P. T. J. Johnson, S. Kutz, C. D. Harvell, Climate change and
infectious diseases: From evidence to a predictive framework. Science 341, 514–519
(2013).
10. A. E. Bates, B. J. Hilton, C. D. G. Harley, Effects of temperature, season and locality on
wasting disease in the keystone predatory sea star Pisaster ochraceus.Dis. Aquat. Organ.
86, 245–251 (2009).
11. G. L. Eckert, J. M. Engle, D. J. Kushner, Sea star disease and population declines at the
Channel Islands, in Proceedings of the Fifth California Islands Symposium, Santa Barbara,
CA, 29 March to 1 April 1999 (US Minerals Management Service, 2000), pp. 390–393.
12. B. A. Menge, Coexistence between the seastars Asterias vulgaris and A. forbesi in a
heterogeneous environment: A non-equilibrium explanation. Oecologia 41, 245–272
(1979).
13. I. Hewson, J. B. Button, B. M. Gudenkauf, B. Miner, A. L. Newton, J. K. Gaydos, J. Wynne,
C.L.Groves,G.Hendler,M.Murray,S.Fradkin,M.Breitbart,E.Fahsbender,
K.D.Lafferty,A.M.Kilpatrick,C.M.Miner,P.Raimondi,L.Lahner,C.S.Friedman,
S.Daniels,M.Haulena,J.Marliave,C.A.Burge,M.E.Eisenlord,C.D.Harvell,
Densovirus associated with sea-star wasting disease and mass mortality. Proc. Natl.
Acad. Sci. U.S.A. 111, 17278–17283 (2014).
14. M. E. Eisenlord, M. L. Groner, R. M. Yoshioka, J. Elliott, J. Maynard, S. Fradkin, M. Turner,
K. Pyne, N. Rivlin, R. van Hooidonk, C. D. Harvell, Ochre star mortality during the
2014 wasting disease epizootic: Role of population size structure and temperature.
Philos. Trans. R. Soc. Lond. B Biol. Sci. 371, 20150212 (2016).
15. C. M. Miner, J. L. Burnaford, R. F. Ambrose, L. Antrim, H. Bohlmann, C. A. Blanchette,
J. M. Engle, S. C. Fradkin, R. Gaddam, C. D. G. Harley, B. G. Miner, S. N. Murray,
J. R. Smith, S. G. Whitaker, P. T. Raimondi, Large-scale impacts of sea star wasting
disease (SSWD) on intertidal sea stars and implications for recovery. PL OS ONE 13,
e0192870 (2018).
16. Multi-Agency Rock Intertidal Network, Sea Star Wasting Syndrome | MARINe (2016);
www.eeb.ucsc.edu/pacificrockyintertidal/data-products/sea-star-wasting/index.html.
17. I. Hewson, K. S. I. Bistolas, E. M. Quijano Cardé, J. B. Button, P. J. Foster, J. M. Flanzenbaum,
J. Kocian, C. K. Lewis, Investigating the complex association between viral ecology,
environment, and northeast pacific sea star wasting. Front. Mar. Sci. 5, 77 (2018).
18. C. Bucci, M. Francoeur, J. McGreal, R. Smolowitz, V. Zazueta-Novoa, G. M. Wessel,
M. Gomez-Chiarri, Sea star wasting disease in Asterias forbesi along the Atlantic Coast of
North America. PLOS ONE 12, e0188523 (2017).
19. B. A. Menge, E. B. Cerny-Chipman, A. Johnson, J. Sullivan, S. Gravem, F. Chan, Sea star
wasting disease in the keystone predator Pisaster ochraceus in Oregon: Insights into
differential population impacts, recovery, predation rate, and temperature effects from
long-term research. PLOS ONE 11, e0153994 (2016).
20. M. M. Moritsch, P. T. Raimondi, Reduction and recovery of keystone predation pressure
after disease-related mass mortality. Ecol. Evol. 8, 3952–3964 (2018).
21. L. M. Schiebelhut, J. B. Puritz, M. N. Dawson, Decimation by sea star wasting disease
and rapid genetic change in a keystone species, Pisaster ochraceus.Proc. Natl.
Acad. Sci. U.S.A. 115, 7069–7074 (2018).
22. J. A. Schultz, R. N. Cloutier, I. M. Côté, Evidence for a trophic cascade on rocky reefs
following sea star mass mortality in British Columbia. PeerJ 4, e1980 (2016).
23. D. Montecino-Latorre, M. E. Eisenlord, M. Turner, R. Yoshioka, C. D. Harvell,
C. V. Pattengill-Semmens, J. D. Nichols, J. K. Gaydos, Devastating transboundary
impacts of sea star wasting disease on subtidal asteroids. PLOS ONE 11, e0163190
(2016).
24. J. A. Estes, J. F. Palmisano, Sea otters: Their role in structuring nearshore communities.
Science 185, 1058–1060 (1974).
25. S. D. Ling, R. E. Scheibling, A. Rassweiler, C. R. Johnson, N. Shears, S. D. Connell,
A. K. Salomon, K. M. Norderhaug, A. Pérez-Matus, J. C. Hernández, S. Clemente,
L. K. Blamey, B. Hereu, E. Ballesteros, E. Sala, J. Garrabou, E. Cebrian, M. Zabala, D. Fujita,
L. E. Johnson, Global regime shift dynamics of catastrophic sea urchin overgrazing.
Philos. Trans. R. Soc. Lond. B Biol. Sci. 370, 20130269 (2015).
26. C. Bonaviri, M. Graham, P. Gianguzza, N. T. Shears, Warmer temperatures reduce the
influence of an important keystone predator. J. Anim. Ecol. 86, 490–500 (2017).
27. W. T. Kohl, T. I. McClure, B. G. Miner, Decreased temperature facilitates short-term sea
star wasting disease survival in the keystone intertidal sea star Pisaster ochraceus.
PLOS ONE 11, e0153670 (2016).
28. C. Pattengill-Semmens, Reef Environmental Education Foundation volunteer survey
project database; www.reef.org.
29. L. M. Cavole, A. M. Demko, R. E. Diner, A. Giddings, I. Koester, C. M. L. S. Pagniello,
M.-L. Paulsen, A . Ramirez-Vald ez, S. M. Schwenck, N. K. Yen, M. E. Zill, P. J. S. Franks,
Biological impacts of the 2013–2015 warm-water anomaly in the Northeast Pacific:
Winners, losers, and the future. Oceanography 29, 273–285 (2016).
30. B. P. V. Hunt, J. M. Jackson, A. A. Hare, K. Wang, Hakai oceanography program: Central
Coast and Northern Strait of Georgia time series, in State of the Physical, Biological
and Selected Fishery Resources of Pacific Canadian Marine Ecosystems in 2015 (Canadian
Technical Report of Fisheries and Aquatic Sciences 3179, Fisheries & Oceans Canada,
Institute of Ocean Sciences, 2016).
31. N. J. D. Nagelkerke, A note on a general definition of the coefficient of determination.
Biometrika 78, 691–692 (1991).
32. S. Pincebourde, E. Sanford, B. Helmuth, An intertidal sea star adjusts thermal inertia to
avoid extreme body temperatures. Am. Nat. 174, 890–897 (2009).
33. C. J. Monaco, D. S. Wethey, B. Helmuth, A Dynamic Energy Budget (DEB) model for the
keystone predator Pisaster ochraceus.PLOS ONE 9, e104658 (2014).
34. E. K. Fly, C. J. Monaco, S. Pincebourde, A. Tullis, The influence of intertidal location and
temperature on the metabolic cost of emersion in Pisaster ochraceus.J. Exp. Mar. Biol.
Ecol. 422–423,20–28 (2012).
35. L. E. Fuess, M. E. Eisenlord, C. J. Closek, A. M. Tracy, R. Mauntz, S. Gignoux-Wolfsohn,
M. M. Moritsch, R. Yoshioka, C. A. Burge, C. D. Harvell, C. S. Friedman, I. Hewson,
P. K. Hershberger, S. B. Roberts, Up in Arms: Immune and nervous system response to sea
star wasting disease. PLOS ONE 10, e0133053 (2015).
36. J. M. Jackson, G. C. Johnson, H. V. Dosser, T. Ross, Warming from recent marine heatwave
lingers in deep British Columbia fjord. Geophys. Res. Lett. 45, 9757–9764 (2018).
37. F. De Castro, B. Bolker, Mechanisms of disease-induced extinction. Ecol. Lett. 8, 117–126
(2005).
38. D.T.Haydon,S.Cleaveland,L.H.Taylor,M.K.Laurenson,Identifyingreservoirsof
infection:Aconceptualandpracticalchallenge.Emerg. Infect. Dis. 8, 1468–1473
(2002).
39. J. M. Burt, M. T. Tinker, D. K. Okamoto, K. W. Demes, K. Holmes, A. K. Salomon, Sudden
collapse of a mesopredator reveals its complementary role in mediating rocky reef
regime shifts. Proc. Biol. Sci. 285, 20180553 (2018).
40. M. L. Dungan, T. E. Miller, D. A. Thomson, Catastrophic decline of a top carnivore in the
Gulf of California rocky intertidal zone. Science 216, 989–991 (1982).
41. H. A. Lessios, Diadema antillarum populations in Panama twenty years following mass
mortality. Coral Reefs 24, 125–127 (2005).
42. L. M. Crosson, C. S. Friedman, Withering syndrome susceptibility of Northeastern Pacific
abalones: A complex relationship with phylogeny and thermal experience. J. Invertebr.
Pathol. 151,91–101 (2018).
43. C. Catton, L. Rogers-Bennett, A. Amrhein, “Perfect storm”decimates northern California
kelp forests. California Department of Fish and Wildlife Marine Management News
(2016); https://cdfwmarine.wordpress.com/2016/03/30/perfect-storm-decimates-kelp/.
44. C. V. Pattengill-Semmens, B. X. Semmens, Reef Environmental Education Foundation,
Conservation and management applications of the REEF volunteer fish monitoring
program. Environ. Monit. Assess. 81,43–50 (2003).
45. L. C. Lee, J. C. Watson, R. Trebilco, A. K. Salomon, Indirect effects and prey behavior
mediate interactions between an endangered prey and recovering predator. Ecosphere 7,
e01604 (2016).
46. S. F. Heron, G. Liu, C. M. Eakin, W. J. Skirving, F. E. Muller-Karger, M. Vega-Rodriguez,
Jacqueline L. De La Cour, T. F. R. Burgess, A. E. Strong, E. F. Geiger, L. S. Guild, S. Lynds,
Climatology Development for NOAA Coral Reef Watch’s 5-km Product Suite (NOAA
Technical Report NESDIS 145, NOAA NESDIS, 2015).
SCIENCE ADVANCES |RESEARCH ARTICLE
Harvell et al., Sci. Adv. 2019; 5: eaau7042 30 January 2019 7of8
on January 31, 2019http://advances.sciencemag.org/Downloaded from
47. R. H. B. Christensen, ordinal: Regression Models for Ordinal Data.R package version 2018. 4–19
(2018).
48. R Core Team, R: A Language and Environment for Statistical Computing (R Foundation
for Statistical Computing, 2015); http://www.R-project.org.
49. S. Mangiafico, rcompanion: Functions to support extension education program evaluation.
R package version 1.5.0. The Comprehensive R Archive Network (2017).
Acknowledgments: We thank REEF surveyors for contributing data. Funding: This work
was supported by the SeaDoc Society, private donors, Seattle Aquarium, NSF RCN
OCE 1215977, NOAA Coral Reef Conservation Program, and Ocean Remote Sensingprogram.
British Columbia monitoring was funded by an NSERC Discovery, Canadian Foundation
for Innovation, and Tula Foundation grant to A.K.S. as well as NSERC postgraduate and
Hakai fellowships to J.M.B. and L.L. The sci entific results and conclusions, as well as any views
or opinions expressed herein, are those of the author(s) and do not necessarily reflect
the views of NOAA or the Department of Commerce. Author contributions: C.D.H., D.M.-L.,
and J.K.G. conceived and planned the paper, with contributions from J.M.C., S.F.H., J.M.B.,
A.K.S., A.K., and K.B. J.M.C., S.F.H., and D.M.-L. conducted statistical analyses with input
from J.M.B., A.K.S., and C.D.H. C.D.H., J.K.G., D.M.-L., J.M.C., S.F.H., J.M.B., A.K.S., A.K., and K.B.
contributed to writing the paper. J.M.B., L.L., O.P., K.B., A.K.S., and C.P.-S. collected data.
Competing interests: The authors declare that they have no competing interests. Data
and materials availability: Supplementary material is available for this paper and includes
a summary of the candidate hierarchical ordinal regression models, a summary of the
final model, and supplementary figures. All data needed to assess the main results in the
paper are in the Figshare repository (DOI: 10.6084/m9.figshare.7300409), while the script
to conduct the ordinal regression is located in https://github.com/jms5151/SSWD.
Additional data related to this paper may be requested from the authors.
Submitted 7 July 2018
Accepted 17 December 2018
Published 30 January 2019
10.1126/sciadv.aau7042
Citation: C. D. Harvell, D. Montecino-Latorre, J. M. Caldwell , J. M. Burt, K. Bosley, A. Keller,
S. F. Heron, A. K. Salomo n, L. Lee, O. Pontier, C. Pattengill-Semmens, J. K. Gaydos, Disease
epidemic and a marine heat wave are associated with the conti nental-scale collapse of a
pivotal predator (Pycnopodia helianthoides). Sci. Adv. 5, eaau7042 (2019).
SCIENCE ADVANCES |RESEARCH ARTICLE
Harvell et al., Sci. Adv. 2019; 5: eaau7042 30 January 2019 8of8
on January 31, 2019http://advances.sciencemag.org/Downloaded from
)Pycnopodia helianthoidesof a pivotal predator (
Disease epidemic and a marine heat wave are associated with the continental-scale collapse
Pontier, C. Pattengill-Semmens and J. K. Gaydos
C. D. Harvell, D. Montecino-Latorre, J. M. Caldwell, J. M. Burt, K. Bosley, A. Keller, S. F. Heron, A. K. Salomon, L. Lee, O.
DOI: 10.1126/sciadv.aau7042
(1), eaau7042.5Sci Adv
ARTICLE TOOLS http://advances.sciencemag.org/content/5/1/eaau7042
MATERIALS
SUPPLEMENTARY http://advances.sciencemag.org/content/suppl/2019/01/28/5.1.eaau7042.DC1
REFERENCES http://advances.sciencemag.org/content/5/1/eaau7042#BIBL
This article cites 40 articles, 8 of which you can access for free
PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions
Terms of ServiceUse of this article is subject to the
registered trademark of AAAS. is aScience Advances Association for the Advancement of Science. No claim to original U.S. Government Works. The title
York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive licensee American
(ISSN 2375-2548) is published by the American Association for the Advancement of Science, 1200 NewScience Advances
on January 31, 2019http://advances.sciencemag.org/Downloaded from