Vulnerability and adaptation of US shellfisheries to ocean acidification

Abstract

Ocean acidification is a global, long-term problem whose ultimate solution requires carbon dioxide reduction at a scope and scale that will take decades to accomplish successfully. Until that is achieved, feasible and locally relevant adaptation and mitigation measures are needed. To help to prioritize societal responses to ocean acidification, we present a spatially explicit, multi-disciplinary vulnerability analysis of coastal human communities in the United States. We focus our analysis on shelled mollusc harvests, which are likely to be harmed by ocean acidification. Our results highlight US regions most vulnerable to ocean acidification (and why), important knowledge and information gaps, and opportunities to adapt through local actions. The research illustrates the benefits of integrating natural and social sciences to identify actions and other opportunities while policy, stakeholders and scientists are still in relatively early stages of developing research plans and responses to ocean acidification.

Full text

NATURE CLIMATE CHANGE | VOL 5 | MARCH 2015 | www.nature.com/natureclimatechange 1
The ocean has absorbed about 25% of anthropogenic
atmospheric CO2 emissions, progressively increasing dis-
solved CO2, and lowering seawater pH and carbonate ion
levels1. On top of this progressive global change in oceanic car-
bon conditions, local factors such as eutrophication2,3, upwelling
of CO2-enriched waters4 and river discharge5 temporarily increase
anthropogenic ocean acidication (OA)6 in coastal waters7–9. Ocean
acidication could primarily aect human communities by chang-
ing marine resource availability1. Studies have shown that, in gen-
eral, shelled molluscs are particularly sensitive to these changes in
marine chemistry10–12. Shelled molluscs comprise some of the most
lucrative and sustainable sheries in the United States13. Ocean
acidication has already cost the oyster industry in the US Pacic
Northwest nearly $110 million, and directly or indirectly jeopard-
ized about 3,200 jobs13. e emergence of real, economically meas-
urable human impacts from OA has sparked a search for regional
responses that can be implemented immediately, while we work
towards the ultimate global solution: a reduction of atmospheric
CO2 emissions. Yet there is little understanding about which loca-
tions and people will be impacted by OA, to what degree, and why,
and what can be done to reduce the risks.
Here, we present the rst local-level vulnerability assessment
for ocean acidication for an entire nation, adapting a well-estab-
lished framework and focusing on shelled mollusc harvests in the
United States; for other evaluations of OA social vulnerability, see
Vulnerability and adaptation of US shellfisheries
to ocean acidification
Julia A. Ekstrom*†1, Lisa Suatoni2, Sarah R. Cooley3, Linwood H. Pendleton4,5, George G. Waldbusser6,
Josh E. Cinner7, Jessica Ritter8, Chris Langdon9, Ruben van Hooidonk10, Dwight Gledhill11,
Katharine Wellman12, Michael W. Beck13, Luke M. Brander14, Dan Rittschof15, Carolyn Doherty†15,
Peter Edwards16 and Rosimeiry Portela17
Ocean acidification is a global, long-term problem whose ultimate solution requires carbon dioxide reduction at a scope and
scale that will take decades to accomplish successfully. Until that is achieved, feasible and locally relevant adaptation and
mitigation measures are needed.To help to prioritize societal responses to ocean acidification, we present a spatially explicit,
multi disciplinary vulnerability analysis of coastal human communities in the United States.We focus our analysis on shelled
mollusc harvests, which are likely to be harmed by ocean acidification.Our results highlight US regions most vulnerable to
ocean acidification (and why), important knowledge and information gaps, and opportunities to adapt through local actions. The
research illustrates the benefits of integrating natural and social sciences to identify actions and other opportunities while policy,
stakeholders and scientists are still in relatively early stages of developing research plans and responses to ocean acidification.
refs14–16. We explored three key dimensions—exposure, sensitivity
and adaptive capacity (Fig.1, Supplementary Fig. S1)—to assess
the spatial distribution of vulnerable people and places to OA. e
underlying assumption guiding this assessment is that addressing
existing vulnerability can reduce future vulnerability to OA, some-
times called ‘human-security vulnerability’15.
Exposure of marine ecosystems addresses acidication driven
by global atmospheric CO2 and amplied by local factors in coastal
waters. We divided the coastal waters around the United States into
existing National Estuary Research Reserve System bioregions17
(Supplementary Fig.S7), and for each bioregion, examined: (1) pro-
jected changes to ocean chemistry based on a reduction in aragonite
saturation state (ΩAr) (Supplementary Fig.S2), and (2) the preva-
lence of key local ampliers of OA, including upwelling, eutrophi-
cation and input of river water with low-aragonite saturation
state [AU: OK?], for each bioregion (Supplementary Figs S4–S6).
Aragonite saturation state (ΩAr) is a measure of the thermodynamic
stability of this mineral form of calcium carbonate that is used by
bivalve larvae and other molluscs, which is also commonly used to
track OA1. Declining ΩAr makes it more dicult and energetically
costly for larval bivalves to build shells even before ΩAr becomes
corrosive [AU: is it ΩAr that becomes corrosive, or should this
be OA?], and ΩAr seems to be the important variable for the most
sensitive early stage of bivalve larvae18. We evaluated relative expo-
sure to anthropogenic OA as the time [AU: i.e. ‘time until’, or ‘the
1Natural Resources Defense Council, 111Sutter Street, SanFrancisco,California 94104, USA; 2Natural Resources Defense Council, 40West 20th Street,
New York, New York10011,USA; 3Ocean Conservancy, 1300 19th Street NW, Washington DC20036,USA; 4Nicholas Institute, Duke University, Durham,
North Carolina27708,USA; 5University of Western Brittany Brest, 29238 Brest, France; 6College of Earth, Ocean, and Atmospheric Sciences, Oregon
State University, Burt 200, Corvallis, Oregon 97331USA; 7ARC Centre of Excellence Coral Reef Studies, James Cook University, Townsville, Queensland,
Australia; 8US Senate Commerce Committee, WashingtonDC, USA; 9Department of Marine Biology and Ecology, Rosenstiel School of Marine &
Atmospheric Science, University of Miami, Florida 33149,USA; 10NOAA Atlantic Oceanographic and Meteorological Laboratory, Miami, Florida33149,
USA; 11NOAA Ocean Acidification Program, Silver Spring, Maryland 20910,USA; 12Northern Economics, Seattle, Washington98107,USA; 13The Nature
Conservancy, Santa Cruz, California 95060, USA; 14Independent Consultant, Hong Kong [AUTHOR: full street address including zip/postcode?] 15Duke
Marine Lab, Duke University, Beaufort, North Carolina 28516,USA; 16NOAA Habitat Conservation Restoration Center, Silver Spring, Maryland20910,USA;
17Conservation International, Arlington Virginia 22202,USA. †Present address: Policy Institute for Energy, Environment, and the Economy, University of
California at Davis, 1605 Tilia Street 100, Davis 95616, California, USA (J.A.E.); Oce of Marine Conservation, US State Department, Washington DC,
USA (C.D.). *e-mail: jaekstrom@gmail.com
PERSPECTIVE
PUBLISHED ONLINE: XX FEBRUARY 2015 | DOI: 10.1038/NCLIMATE2508

2 NATURE CLIMATE CHANGE | VOL 5 | MARCH 2015 | www.nature.com/natureclimatechange
extent of time for which’?] mean annual surface seawater exceeds
an empirically informed absolute ΩAr threshold for several spe-
cies of bivalve larvae. is indicator for disruption to the biologi-
cal processes of calcication and development in larval molluscs
was favoured over alternatives (for example time until the historic
range of ΩAr is exceeded) because the biological mechanism was
clear19 and empirical evidence exists20. For comparison purposes,
the Supplementary Information includes the time until the historic
range of ΩAr is exceeded (Supplementary Fig. S3), but below we
document the outcomes based on the ΩAr threshold projections and
local ampliers of OA.
Sensitivity of social systems was evaluated at the scale of ‘clus-
ters of coastal counties’ around the United States, using three indi-
cators of community dependence on shellsh, adapted from the
National Marine Fisheries Service’s shing community vulnerabil-
ity and resilience index21: (1)the 10-year median landed value of
shellsh (including both wild and aquaculture harvests); (2) the
10-year median proportional contribution of shellsh to total
value of commercial landings; and (3)the 5-year median number
of licences (representing jobs) supported by shelled mollusc shing
(Supplementary Information). Sensitivity indicators were re-scaled
and combined into a single index (Supplementary Information and
Supplementary Fig.S8).
Adaptive capacity of social systems to cope with and adapt to
OA is represented by three classes of indicators: status of state gov-
ernment climate and OA policies, local employment alternatives
and availability of science. We examined a total of six indicators
representing adaptive capacity that are derived largely from the
broader economic and policy landscape, yet are directly relevant
for dealing with the threat of OA (Supplementary Fig. S9). is
is a deliberate departure from studies conducted at broader and
ner geographic scales that use general demographic indicators
(see Supplementary Information). We assessed ‘potential govern-
ment support for adaptation’ through measures of: (1) the status
of state legislative action on OA and (2)the status of state climate
adaptation planning. ese indicators reect social organization
and assets at the state jurisdictional level that could be used by
communities to adapt to, cope with, or avoid the impacts of lost
shellsh harvests. We examined aspects of employment alterna-
tives through: (3) the diversity of shelled mollusc harvests, sug-
gesting potential alternative shellsh that could be harvested and
(4) the diversity of non-shellsh-related employment industries.
ese reect the likelihood of job alternatives for shellsh har-
vesters and those in the aquaculture industry. Finally, we captured
‘access to and availability of science’ through (5) a score for marine
laboratoriesdeveloped totake into account the high local inuence
that such laboratories can have as well as the potential contribution
beyond their immediate vicinity. For each county cluster, a metric
based on the number ofuniversity marine laboratories(on-campus
and satellite laboratories) in that county cluster was averaged with
a metric based on the total number of university marine laborato-
ries in that state (see Supplementary Information for more infor-
mation) and (6)Sea Grant state budgets normalized by shoreline
length. ese indicators represent the availability of local scientic
capacity, the potential for troubleshooting assistance, and the pos-
sibility of access to a range of tools and data products, such as avail-
able early warning information. We attributed each county cluster
(as used in Sensitivity) to each variable score of the six indicators.
We then combined into a single index by averaging re-scaled (0–1)
overall component scores for sensitivity and adaptive capacity
(Supplementary Information Fig.S9). Coincidence of high marine
ecosystem exposure to OA with high sensitivity and low adaptive
capacity of social systems reveals the areas at highest overall vul-
nerability to OA.
Places vulnerable to ocean acidification
Our results show that 16out of 23 bioregions around the United
States are exposed to rapid OA (reaching ΩAr 1.5by 2050) or at
least one amplier (Fig.2; Supplementary TableS1); 10 regions are
exposed to two or more threats of acidication (note that Alaska
and Hawaii are missing local amplier data; Fig. 2). e marine
ecosystems and shelled molluscs around the Pacic Northwest
and Southern Alaska are expected to be exposed soonest to ris-
ing global OA, followed by the north-central West Coast and the
Gulf of Maine in the northeast United States. Communities highly
reliant on shelled molluscs in these bioregions are at risk from
OA either now or in the coming decades. In addition, pockets of
marine ecosystems along the East and Gulf Coasts will experi-
ence acidication earlier than global projections indicate, owing
to the presence of local ampliers such as coastal eutrophication,
upwelling and discharge of low-ΩAr river water (see Supplementary
FigsS4–S6, Supplementary TableS1).e inclusion of local ampli-
ers reveals more coastline segments around the United States that
are exposed to acidication risk than when basing exposure solely
on global models.
Combining sensitivity and adaptive capacity reveals that the
most socially vulnerable communities are spread along the US East
Coast and Gulf of Mexico (Fig.2), yet the sources of high social
vulnerability are very dierent between these two regions (see
Supplementary Information for breakdown separated by sensitivity
and adaptive capacity, FigsS8 and S9). Specically, the East Coast
is dominated by high levels of sensitivity, or economic depend-
ence, from strong use of shellsh resources. For example, south-
ern Massachusetts measures as having the highest sensitivity. is
county cluster ranks in the top four for all three sensitivity indica-
tors (Supplementary Fig.S8), meaning that this area has the highest
mollusc harvest revenues of any coastal area in the United States,
second highest number of licences and fourth highest proportion
of seafood revenues coming from molluscs. In contrast, the Gulf
of Mexico region is socially vulnerable from low adaptive capacity,
owing to social factors such as low political engagement in OA and
climate change, low diversity of shellsh shery harvest and rela-
tively low science accessibility (Supplementary Fig.S9).
Importantly, our visually combined overall vulnerability analy-
sis reveals that a number of socially vulnerable communities lie
adjacent to water bodies that are exposed to a high rate of OA
or at least one local amplier, indicating that these places could
be at high overall vulnerability to OA (Fig. 2). e areas that are
exposed to OA (including local ampliers) and high and medium–
high social vulnerability coincide include southern Massachusetts,
Rhode Island, Connecticut, New Jersey and portions around the
Overall vulnerability
Marine ecosystem exposure
Marine ecosystems exposed to
ocean acidification (OA)
Social vulnerability
Sensitivity
Local societal
importance
of shellfish
Adaptive
capacity
Assets available
to help prepare
for or avoid
impacts of OA
Figure 1 | Conceptual framework structuring the analysis of vulnerability
to ocean acidification. Vulnerability analyses can focus on three key
dimensions (exposure, sensitivity and adaptive capacity): (1) the extent
and degree to which assets are exposed to the hazard of concern; (2)the
sensitivity of people to the exposure; and (3) the adaptive capacity
of people to prepare for and mitigate the exposure’s impacts. These
three dimensions together provide a relative view of a place’s overall
vulnerability. Adapted conceptual model components from refs 16,52–55.
PERSPECTIVE NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2508

NATURE CLIMATE CHANGE | VOL 5 | MARCH 2015 | www.nature.com/natureclimatechange 3
Chesapeake Bay, the Carolinas, and areas across the Gulf of Mexico
(Fig. 2b–d). Interestingly, global ocean models that project the
advance of OA, primarily as a result of atmospheric CO2, do not
reveal these areas as exposed to global OA until aer 2099, based on
our study’s ΩAr threshold (Table1). e marine ecosystem exposure
in the areas located along the Atlantic coast and the Gulf of Mexico
is from low-ΩAr conditions caused primarily by the addition of river
water and eutrophication, local factors that have only more recently
been considered major ampliers of nearshore acidication6,7. ese
coastal processes are likely to tip coastal oceans past organism
thresholds as atmospheric CO2 uptake continues in the future (see
ref. 22). Although the Pacic Northwest, northern California and
Maine exhibit only medium and medium–low social vulnerability
(Fig.2a,b), these areas are particularly economically sensitive and
lie adjacent to marine ecosystems highly exposed to global OA23,24
(sensitivity, Supplementary Fig.S8). is prole of relatively high
dependency and high exposure in these three regions has already
activated signicant research and local action/engagement among
local scientists, government and shellsh growers (see for example
refs 25,26). is engagement has driven up adaptive capacity (based
on our study’s indicators) in these areas, which reduces their social
vulnerability relative to other regions across the United States. In
comparison, the lower level of OA-related action in other regions
such as the Gulf of Mexico (Fig.2d), Massachusetts (Fig.2b) and
Mid-Atlantic (Figs 2c,d) with high overall vulnerability proles
might be partly because their marine ecosystem exposure is domi-
nated by the presence of local OA ampliers rather than global OA
(Supplementary Fig.S2, Supplementary TableS1). At the same time,
some of these areas (for example Maryland) do have strong advo-
cates for addressing water quality which could provide an oppor-
tunity to address locally driven acidication as awareness of the
issue grows.
Alaska
California
Oregon
Washington
Maine
Mass.
Conn.
Texas
Hawaiian Islands
Louisiana
Gulf of
Maine
Gulf of Mexico
Pacific Ocean
Pacific Ocean
Pacific Ocean
Pacific Ocean
Atlantic
Ocean
Mississippi
Alabama Georgia
Florida
Virginia
North Carolina
South
Carolina
Puget Sound
Chesapeake
Bay
CANADA
UNITED STATES
MEXICO
f
e
ab
c
d
abc
d
f
e
U
U
U
U
nd
nd
Atlantic
Ocean
E(1/7),R(2/9)
E(2/8),R(2/7)
R(1/6)
R(3/4)
E(5/8) E(3/6)
E(1/5)
E(1/10)
E(1/5)
E(1/11)
R(7/9) R(2/12)
R(4/7)
R(4/8)
Social vulnerability (land)
Highest SV (top 20%)
Medium high
Medium SV (middle 20%)
Medium low
Lowest SV (bottom 20%)
Marine ecosystem exposure (water)
Year hits threshold
After 2099
2006–2030
2031–2050
2051–2070
2071–2099
U: Upwelling is strong
River drainage low saturation state
and high annual discharge volume
nd: No data available for E or R
Highly eutrophic estuaries presentE:
R:
Local amplifiers
Figure2 | Overall vulnerability of places to ocean acidification. Scores of relative social vulnerability are shown on land (by coastal county cluster) and
the type and degree of severity of OA and local amplifiers to which coastal marine bioregions are exposed, mapped by ocean bioregion: (a) contiguous US
West Coast; (b) Northeast; (c) Chesapeake Bay; (d) Gulf of Mexico, and Florida and Georgia’s coast; (e) Hawaii Islands; and (f) Alaska. Social vulnerability
(red tones) is represented with darker colours where it is relatively high. Exposure (purple tones) is indicated by the year at which sublethal thresholds
for bivalve larvae are predicted to be reached, based on climate model projections using the RCP8.5 CO2 emission scenario27. Exposure to this global OA
pressure is higher in regions reaching this threshold sooner. Additionally, the presence and degree of exposure to local amplifiers of OA are indicated for
each bioregion: E(x/y)marks bioregions [AU: OK?] in which highly eutrophic estuaries are documented, x is the number of estuaries scored as high, and y
is the total number evaluated in each bioregion (source: ref. 56), locations of highly eutrophic estuaries are marked with a star; R(x/y) marks bioregions in
which sampled river water draining into bioregion scored [AU: this description is not clear grammatically: should it be ‘bioregions in which... water was
scored’, or is something missing here? Also, does ‘scoring in the top quintile’ here mean top quintile of discharge volume only? Please clarify phrasing]
based on very low saturation state and high annual discharge volume (top quintile, calculated by authors from US Geological Survey57), x is the number
of rivers scoring in the top quintile of those evaluated, and y is the total number evaluated in this study. Approximate locations of river outflows of those
rivers scoring in the top quintile are marked with a delta [AU: a yellow triangle?]; and U marks bioregions where upwelling is very strong in at least part of
the bioregion (source: ref. 58).
PERSPECTIVE
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2508

4 NATURE CLIMATE CHANGE | VOL 5 | MARCH 2015 | www.nature.com/natureclimatechange
Robustness of analysis
To examine the robustness of these spatial patterns of vulnerability,
we varied the index aggregation methodology and the selection of
indicators. To test the dierence in index aggregation methods for
social vulnerability, we compared the output of adding and multi-
plying sensitivity and adaptive capacity indices and found little dif-
ference; the same set of county clusters made up the top 10 most
socially vulnerable places using either aggregation method.
To explore the eect of indicator selection on adaptive capac-
ity (and thus social vulnerability), we compared a set of commonly
used generic indicators for adaptive capacity relating to income,
poverty, education and age with the set of threat-specic indi-
cators developed for this study (see Table 3 and Supplementary
FigsS10and S11). Using the generic capacity measures to calculate
social vulnerability, we found that six of the same county clusters
measured within the top 10 highest socially vulnerability places in
the United States as those found using the threat-specic indicators
(see Supplementary Information for analysis and maps). is is con-
siderable overlap given that the two sets of variables indicate entirely
dierent notions of adaptive capacity. Because the sensitivity indica-
tors were developed and vetted by sheries social science research-
ers21 and alternative potentially appropriate data were not available
nationwide, we did not have a useful comparison for this element
from which to draw.
To explore the criterion for ΩAr, we examined one alternative
for disruption of biological processes with respect to rising atmos-
pheric CO2: the time until average surface waters move outside
the present range of ΩAr (that is, exceeding a historic envelope)27.
e map generated by this ‘historic envelope’ approach shows that
southern areas experience potential OA exposure earlier, which
is nearly an inverse pattern to our chosen criterion of a chemical
threshold when calcication and development of larval molluscs
may decrease (Supplementary Fig. S3). is dierence in pat-
terns is because natural variability is much smaller in southern
regions, although evidence of greater sensitivity in populations
of bivalves that live in tropical and subtropical waters is lacking.
is discrepancy underscores the need for targeted research inte-
grating a physiological, ecological and evolutionary perspective on
the potential and limitations of strong local biological adaptation
to dierent carbonate regimes for commercially valuable shelled
mollusc populations.
Overall, we found that variable selection has stronger eects
than aggregation methods, which provides high condence in our
aggregation methods for social vulnerability. e dierences found
in variable selection identify research needs relating to what factors
underlie vulnerability on the ground that are relevant to OA; this
conversation has only just begun.
Opportunities to reduce vulnerability to ocean acidification
Social–environmental syntheses, including vulnerability analyses,
can help to identify opportunities for actionable solutions to address
the potential impacts of ocean acidication. Our analysis reveals
where and why the overall vulnerability from OA varies among
the many coastal areas of the United States, and thus identies
opportunities to reduce harm.
One way to tackle OA is by reducing marine ecosystem exposure
to it. Several portions of the east coast are highly exposed to OA
from high levels of eutrophication (Fig.2b–d). In addition to releas-
ing extra dissolved CO2 and enhancing acidication, eutrophication
can also decrease seawater’s ability to buer further acidication3.
People in these regions are uniquely positioned to reduce expo-
sure to OA through regional actions by curtailing eutrophication
(as compared, for example, with regions exposed to upwelling).
Although a signicant challenge, reducing nutrient loading to the
coastal zone in these areas could provide multiple benets, mak-
ing it a no-regrets option. Reducing eutrophication can decrease
hypoxia and harmful algal blooms, in addition to reducing risk
from fossil-fuel-derived OA at the local and regional level. Policy
Table 1 | Indicators of drivers and amplifiers of ocean acidification, and the criterion for each used in this study.
Factors causing and amplifying OA
(reducing ΩAr)
Indicator Scoring scale Criterion for ranking the risk factor
as ‘high’
Rising atmospheric CO2 reduces ΩAr
causing chronic stress to shelled
mollusc larvae
Projected year that surface water will
reach 1.5ΩAr (ref. 27)
Continuous scale from current year
to 2099
1.5ΩAr threshold reached by 2050
Eutrophication increases pCO2 locally
via respiration, leading to reduced ΩAr
Degree of eutrophication56 Eutrophication scored on a five-point
scale: low to high
Presence of a high-scoring eutrophic
estuary in bioregion
River water can reduce ΩAr locally in
coastal waters
Combined metric of river’s aragonite
saturation state and annual
discharge volume
Rivers scored on a five-point scale:
low to high
Presence of high scoring river (for
low aragonite saturation and high
discharge volume) in bioregion
Significant seasonal upwelling
delivers water rich in CO2 to shallow
waters, leading to reduced ΩAr
Degree of upwelling58 Coastal zones scored on a five-point
scale: low to high
Presence of high upwelling zone
in bioregion
Table 2 | Indicators representing ‘sensitivity’ (people’s dependency) on organisms expected to be aected by ocean acidification
(in this study, shelled molluscs).
Indicator or measure Source Raw format Processing for subindex
Landed value
(median of 10 years)
Regional fisheries databases (ACCSP,
GulfBase, PacFIN), and States of
Alaska and Hawaii
US dollars, annual Calculated median for years
2003–2012
Winsorized the top 10%
Percentage of shellfish by value [AU:
i.e. as percentage of all fish caught?]
(median of 10 years)
For each year: shelled molluscs
value/total commercial landed value
Divided landed value of shellfish by
landed value of all fish
Winsorized the top 10%
Number of licences as proxy for jobs
(median over 5 years)
Number of commercial
licences, annual
Winsorized the top 10%
All indicators are in units of county clusters.
[AU: Please indicate where Table 2 should be cited in the text.]
PERSPECTIVE NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2508

NATURE CLIMATE CHANGE | VOL 5 | MARCH 2015 | www.nature.com/natureclimatechange 5
instruments to reduce eutrophication exist in the United States28
and can be leveraged to facilitate eorts to reduce OA8.
Another important way to combat the eects of OA will be
by reducing social vulnerability. In regions where high sensitiv-
ity (one component of social vulnerability) arises from the struc-
ture of the shing industry, an entirely dierent approach to
adaptation may be more appropriate than those geared to reduce
marine ecosystem exposure. For example, where shery harvest
portfolios are dominated by a single species, such as in the Gulf
of Mexico where mollusc production is limited to the eastern oys-
ter (Crassostrea virginica), diversication of the species harvested
might be a benecial strategy.
A further way to reduce social vulnerability may be by increas-
ing adaptive capacity of people and regions. Access and availability
to science already has helped shellsh aquaculturists in the Pacic
Northwest to identify and avoid some of the consequences of OA20.
Working with local scientists, hatcheries have implemented several
strategies to adapt and mitigate OA eects on bivalve seed produc-
tion. rough local industry–research partnerships in the Pacic
Northwest, implementation of real-time monitoring of saturation
state, chemical buering of water, changes in timing of seasonal seed
production and use of selectively bred lines of oyster broodstock, this
collaboration has prevented collapse of the regional oyster industry.
In every case, when developing a broader array of adaptation
strategies, it is critical to work directly with the coastal communities
in each region so they can develop context-appropriate and feasi-
ble adaptation options. Targeted projects to develop local adapta-
tion plans may even require developing further regionally relevant
indicators of adaptive capacity and community resilience that this
nationwide study does not capture. In fact, zooming in to assess par-
ticular regions at a higher resolution would enable regional stake-
holders to provide input into a possible dierent set of variables that
denes vulnerability in their particular region based on values and
social or economic context.
Barriers to and path forward for addressing OA
is study oers the rst nationwide vulnerability assessment of
the spatial distribution of local vulnerability from OA focusing on a
valuable marine resource. But it is just a rst step to understanding
where and how humans and marine resources are at highest risk to
OA and its local ampliers. Another key nding of this assessment is
that signicant gaps in the scientic understanding of coastal ocean
carbonate dynamics, organismal response and people’s depend-
ence on impacted organisms limit our ability to develop a full suite
of options to prepare for, mitigate and adapt to the threats posed
by OA, and these can be considered in a structured way using the
framework (Fig.3). e types of gaps identied—as commonly clas-
sied in information science and other disciplines29,30—range from
data inaccessibility to knowledge deciencies.
Marine ecosystem exposure. Key gaps remain in understanding
how global and local processes interact to drive nearshore OA,
and how this will aect marine organisms and ecological systems.
Recent studies suggest that the biogeochemical interaction between
global OA and local ampliers is additive3,22,31; however, most ocean
models used to project future OA cannot adequately resolve these
processes, which are also increasingly aected by human activity7,32.
Even though direct measurements incorporate an ever-growing
global network of monitoring instruments, they are oen located
oshore and remain too sparse in space and time to resolve the
dynamics of seawater chemistry near shore, where most shellsh
live. Historically, OA monitoring has focused on oshore regions,
where long-term, high-accuracy and precise measurements enabled
detection and attribution of the rising atmospheric CO2 acidica-
tion signal. But many commercially and nutritionally important
organisms live in the coastal zone where they experience the com-
bined eects of multiple processes that alter the carbonate chemis-
try7. is results in greatly variable ‘carbonate weather’ for a given
location33. Characterizing this variation, including modelling how
rising atmospheric CO2 will increase the frequency, duration and
severity of extreme events [AU:OK?], would provide a fuller picture
of how OA is unfolding within the dynamic coastal waters.
To improve our understanding of which marine ecosystems
and organisms are most susceptible to ocean acidication, addi-
tional information on the ΩAr thresholds below which reproduc-
tion and survival are disrupted is needed. In the US context, the
Table 3 | Threat-specific indicators used to assess capacity of fishing communities to deal with impacts of ocean acidification.
Group Indicator Source Raw format Processing for subindex
Access to scientific
knowledge
Budget of Sea Grant
programmes
National Sea Grant State-level total funds of
budget (state and federal
contributions combined, 2013)
• Re-scaled (0–1)
• Attributed normalized
scores to each
county cluster
Number of university marine
laboratories
Direct count from registries
and Internet
Latitude/longitude location
of laboratories
• Combined score of
laboratories per state/
shoreline length and labs
per county cluster
Employment alternatives Shelled mollusc diversity Regional fisheries databases
(ACCSP, GulfBase, PacFIN),
and States of Alaska
and Hawaii
Ratio of landing revenues for
each taxon by county cluster
• Calculated Shannon
Weiner Diversity Index
Economic diversity ACS Census Proportion of county
population employed in
each industry
• Calculated Shannon
Weiner Diversity Index
for county clusters
Political action Legislative action for OA Keyword searches on
legislature websites and
follow-up calls
Established five-point scale
for state’s legislative progress
on OA
• Re-scaled 0–1
• Attributed score to
county clusters
Climate adaptation planning Georgetown Law School
Climate programme website
Status of climate adaptation
plan for state
• Re-scaled 0–1
• Attributed score to
country clusters
See Supplementary Information for discussion and presentation of alternative indicators and measures.
PERSPECTIVE
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2508

6 NATURE CLIMATE CHANGE | VOL 5 | MARCH 2015 | www.nature.com/natureclimatechange
concentration of value in a limited number of shellsh species
means that the identication of biologically susceptible and resist-
ant species and populations is both prudent and feasible. Based
on total landed value from 2003to 2012, approximately 95% of
shelled-mollusc revenues in the United States come from only
10 species (and 80% from ve).ese species include sea scallop
(52.9%), eastern oyster (11.3%), Pacic geoduck (5.8%), Pacic
oyster (5.2%) and six species of clam (that range from 5% to
2.6% of total value)34. ere is some evidence of local biological
adaptation of other marine taxa to varying carbonate chemistry
regimes35–37. is potential genetic variation, if present, could be
documented to aid in the development of resistant strains of cul-
tured or other organisms.
Social vulnerability. Our study also revealed large gaps in infor-
mation about mollusc-dependent communities to inform measures
of social vulnerability. We do not have high-resolution nationwide
data on the full cultural and societal signicance of shelled mol-
luscs. Even data on the contributions of shellsh to human nutri-
tion, shoreline protection, and water ltration were inadequate
nationwide. Incorporation of these other ecosystem services pro-
vided by molluscs could alter the social vulnerability landscape. For
the commercial sheries data that we did obtain, condentiality
constraints forced us to aggregate our analysis into county clusters,
preventing county-specic or port-level analyses of social vulner-
ability that might have revealed more spatial heterogeneity. We also
lack social science data that describe use at species-, human com-
munity-, port- or household levels. We lack data on the value chain
that links threatened organisms to harvesters, processors and end-
users. Finally, empirically tested adaptive capacity measures could
contribute to a more rigorous evaluation of social vulnerability.
is includes data on scientic spending and infrastructure directly
relevant to end-users, as well as social and demographic data that
are reective of end-users (for this study, shing and aquaculture
communities) and not the general population (for example generic
indicators quantifying education and income).
Beyond helping in prioritizing and developing adaptation strate-
gies, social science is also useful to inform and guide planning for
social adaptation and mitigation. As with climate change adapta-
tion, preparing for and adapting to the impacts of OA is a social
process1,38,39. Implementation does not occur automatically once
strategies are developed, but instead must oen overcome a suite
of institutional (including legal), political, psychological and other
types of barriers40. As learned from climate change initiatives, the
‘soer side’ of adaptation (such as coordination among stakehold-
ers, industry and scientists) is the rst step towards preparing for
a threat like OA41. Despite its fundamental importance, this type of
eort is oen overlooked and remains underfunded. Social science
can also help practitioners even in early stages of adaptation g-
ure out how to engage public and policy-makers eectively in OA
issues42–44. Farther along in adaptation processes, social science can
inform the development of strategies by accounting for social val-
ues45,46 and existing property rights in use and norms47,48 and even
helping to work out what type of information is salient for and
trusted by decision-makers49,50. Although important for reducing its
risks, social science relevant for understanding OA has been mini-
mal thus far. A budget assessment conducted by the Interagency
Working Group on Ocean Acidication reported that federal
research in scal year 2011 allocated $270,000of Federal funds for
social science research related to OA, which represents 0.9% of the
entire OA spending for that year’s budget51.
Conclusions
As with other global environmental changes, acidication of the
oceans is a complex and seemingly overwhelming problem. Here we
have focused only on OA (and nearshore ampliers) as the threat to
coastal species. Although other stressors also threaten coastal eco-
systems, our single-threat assessment allows us to tease out where
OA in isolation could hit people and organisms the hardest, which
can inform research agendas and decision-making geared speci-
cally to address OA. A vulnerability framework helps to structure
our thinking about the ways in which ocean acidication will aect
Nearshore
projections
What is the ΩAr
threshold for
each species?
Where are the
shellfish beds?
What is the ecological and
economic productivity
of beds?
How does atmospheric
CO2-driven OA cumulate
with local drivers?
Community and
household reliance
on shellfish
What is ‘successful’
adaptation for
shellfish-dependent
communities?
Community- and
household-level
attributes that improve
capacity deal with OA
Importance of shellfish
to community
well-being
e
Types of gap
Knowledge
Information
Data
Data access
Relative eort
needed to fill
High
Low
Overall vulnerability
Marine ecosystem exposure
Marine ecosystems exposed to
ocean acidification (OA)
Social vulnerability
Sensitivity
Local societal
importance
of shellfish
Adaptive
capacity
Assets available
to help prepare
for or avoid
impacts of OA
Figure 3 | Sample of gaps in knowledge related to OA vulnerability, information and data organized around components of the framework. Dierent
types of gaps are classified by the level of eort that is required to fill them (gaining knowledge is the most challenging, whereas data access tends to be
the most straightforward).
PERSPECTIVE NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2508

NATURE CLIMATE CHANGE | VOL 5 | MARCH 2015 | www.nature.com/natureclimatechange 7
ecosystems and people. e framework also helps to identify and
organize the opportunities and challenges in dealing with these
problems. But this study is the beginning; adaptation to OA and
other global environmental change is an iterative process that
requires both top-down and bottom-up processes. Our analysis of
OA as it relates to [AU: OK?] US shelled mollusc sheries makes
clear just how much the pieces of the OA puzzle vary around the
country. Marine ecosystem exposure, economic dependence and
social capacity to adapt create a mosaic of vulnerability nation-
wide. An even more diverse set of strategies may be needed to help
shellsh-dependent coastal communities adapt to OA. Rather than
create and apply a nationwide solution, decision-makers and other
stakeholders will have to work with shing and aquaculture com-
munities to develop tailored locally and socially relevant strategies.
Meaningful adaptation to OA will require planning and action at all
levels, including regional and local levels, which can be supported
with resources, monitoring, coordination and guidance at the
national level.
Over the past decade, scientists’ understanding of ocean
acidication has matured, awareness has risen and political action
has grown. e next step is to develop targeted eorts tailored to
reducing social and ecological vulnerabilities and addressing local
needs. Tools like this framework can oer a holistic view of the
problem and shed light on where in the social–ecological system to
begin searching for locally appropriate solutions.
Received 22 August 2014; accepted 19 December 2014; published
online xx February 2015.
References
1. IPCC. Climate Change 2014: Impacts, Adaptation, and Vulnerability Part B:
Regional Aspects. (eds Field, C. B. et al.) (Cambridge Univ. Press, 2014).
2. Waldbusser, G.G., Voigt, E.P., Bergschneider, H., Green, M.A. & Newell, R.I. E.
Long-term trends in Chesapeake Bay pH and eects on biocalcication in the
Eastern Oyster Crassostrea virginica. Estuar. Coasts 34, 221–231 (2011).
3. Cai, W-J. etal. Acidication of subsurface coastal waters enhanced by
eutrophication. Nature Geosci. 4, 766–770 (2011).
4. Feely, R.A., Sabine, C.L., Hernandez-Ayon, J.M., Ianson, D. & Hales, B.
Evidence for upwelling of corrosive ‘acidied’ water onto the continental shelf.
Science 320, 1490–1492 (2008).
5. Salisbury, J., Green, M., Hunt, C.W. & Campbell, J. Coastal acidication by
rivers: a threat to shellsh? EOS Trans. Am. Geophys. Union 89, 513–528 (2008).
6. IPCC. Report of the IPCC Workshop on Impacts of Ocean Acidication on
Marine Biology and Ecosystems, 164 (Carnegie Inst., 2011).
7. Duarte, C.M. etal. Is ocean acidication an open-ocean syndrome?
Understanding anthropogenic impacts on seawater pH. Estuar. Coasts
36, 221–236 (2013).
8. Kelly, R.P. etal. Mitigating local causes of ocean acidication with existing
laws. Science 332, 1036–1037 (2011).
9. Waldbusser, G.G. & Salisbury, J.E. Ocean acidication in the coastal zone from
an organism’s perspective: multiple system parameters, frequency domains,
and habitats. Annu. Rev. Mar. Sci. 6, 221–247 (2014). [AU: please check
added details].
10. Gazeau, F., Parker, L.M., Comeau, S. & etal. Impacts of ocean acidication on
marine shelled molluscs. Mar. Biol. 160, 2207–2245 (2013).
11. Parker, L.M. etal. Predicting the response of molluscs to the impact of ocean
acidication. Biology 2, 651–692 (2013).
12. Kroeker, K.J. etal. Impacts of ocean acidication on marine organisms:
quantifying sensitivities and interaction with warming. Glob. Change Biol.
19, 1884–1896 (2013).
13. Washington State Blue Ribbon Panel on Ocean Acidication. Ocean Acidication:
From Knowledge to Action. Washington State’s Strategic Response. Available at
https://fortress.wa.gov/ecy/publications/publications/1201015.pdf (2012).
14. Cooley, S.R., Lucey, N., Kite-Powell, H. & Doney, S.C. Nutrition and income
from molluscs today imply vulnerability to ocean acidication tomorrow.
Fish Fisher. 13, 182–215 (2012).
15. Mathis, J.T. etal. Ocean acidication risk assessment for Alaska’s shery sector.
Prog. Oceanogr. (in the press).
16. Hilmi, N. etal. Exposure of Mediterranean countries to ocean acidication.
Wat e r 6, 1719–1744 (2014).
17. National Estuary Research Reserve System. Bioregions
http://www.nerrs.noaa.gov/ (2009).
18. Waldbusser, G.G. etal. A developmental and energetic basis linking
larval oyster shell formation to ocean acidication. Geophys. Res. Lett.
40, 2171–2176 (2013).
19. Waldbusser, G.G. etal. Saturation-state sensitivity of marine bivalve larvae
to ocean acidication. Nature Clim. Change, http://dx.doi.org/10.1038/
nclimate2479 (in the press). [PRODUCTION: UPDATE?].
20. Barton, A., Hales, B., Waldbusser, G.G., Langdon, C. & Feely, R.A. e Pacic
oyster, Crassostrea gigas, shows negative correlation to naturally elevated
carbon dioxide levels: Implications for near-term ocean acidication eects.
Limnol. Oceanogr. 57, 698–710 (2012).
21. Jepson, M. & Colburn, L.L. Development of Social Indicators of Fishing
Community Vulnerability and Resilience in the US Southeast and
Northeast Regions. NOAA Technical Memorandum NMFS-F/SPO-129
(US Dept Commerce, 2013).
22. Feely, R.A. etal. e combined eects of ocean acidication, mixing,
and respiration on pH and carbonate saturation in an urbanized estuary.
Estuar. Coast. Shelf Sci. 88, 442–449 (2010).
23. Gruber, N. etal. Rapid progression of ocean acidication in the California
Current system. Science 337, 220–223 (2012).
24. Hauri, C. etal. Spatiotemporal variability and long-term trends of ocean
acidication in the California Current system. Biogeosci. 10, 193–216 (2013).
25. th Maine Legislature. in Legislative Document No. 1602 (Maine, 2014).
[AUTHOR: what shoul ‘th’ say? Delete?].
26. Veneziano, S. in Boothbay Register (Maine, 2014).
27. van Hooidonk, R.J., Maynard, J.A., Manzello, D. & Planes, S. Opposite
latitudinal gradients in projected ocean acidication and bleaching impacts on
coral reefs. Glob. Change Biol. 103–112, (2014).
28. Bricker, S. etal. Eects of nutrient enrichment in the nation’s estuaries:
A decade of change. Harmful Algae 8, 21–32 (2008).
29. Zins, C. Conceptual approaches for dening data, information, and knowledge.
J.Am. Soc. Inform. Sci. 58, 479–493 (2007).
30. Boisot, M. & Canals, A. Data, information and knowledge: Have we got it
right? J.Evol. Econ. 14, 43–67 (2004).
31. Harris, K.E., DeGrandpre, M.D. & Hales, B. Aragonite saturation state
dynamics in a coastal upwelling zone. Geophys. Res. Lett. 40, 1–6 (2013).
32. Doney, S.C. e growing human footprint on coastal and open-ocean
syndrome? Understanding anthropogenic impacts on seawater pH. Science
328, 1512–1516 (2010).
33. Newton, J.A., Feely, R.A., Jewett, E.B., Williamson, P. & Mathis, J.T.
Global Ocean Acidication Observing Network: Requirements and Governance
Plan (Global Ocean Acidication Observing Network (GOA-ON), 2014).
34. NOAA Oce of Science and Technology. NMFS Commercial Fisheries
Statistics (2003–2012) http://www.st.nmfs.noaa.gov/commercial-sheries/
commercial-landings/annual-landings-with-group-subtotals/index, 2014).
35. Pespeni, M.H. etal. Evolutionary change during experimental ocean
acidication. Proc. Natl Acad. Sci. USA 110, 6937–6942 (2013).
36. Sunday, J.M. etal. Evolution in an acidifying ocean. Trends Ecol. Evol.
29, 117–125 (2014).
37. Hofmann, G.E. etal. Exploring local adaptation and the ocean acidication
seascape: studies in the California Current large marine ecosystem.
Biogeosci. Discuss. 10, 11825–11856 (2013).
38. Adger, W.N. Social capital, collective action and adaptation to climate change.
Econ. Geogr. 79, 387–404 (2003).
39. Wolf, J. Climate Change Adaptation as a Social Process Vol. 42 (Springer, 2011).
40. Moser, S. & Ekstrom, J.A. A framework to diagnose barriers to climate change
adaptation. Proc. Natl Acad. Sci. USA 107, 22026–22031 (2010).
41. Moser, S.C. & Ekstrom, J.A. Identifying and Overcoming Barriers to
Climate Change Adaptation in San Francisco Bay: Results from Case Studies.
CEC-500-2012-034 (California Energy Commission, 2012).
42. Kahan, D.M. Fixing the communications failure. Nature
463, 296–297 (2010).
43. Maibach, E., Roser-Renouf, C. & Leiserowitz, A. Global Warming’s Six Americas
2009: An Audience Segmentation Analysis (Yale Project on Climate Change,
George Mason Univ. Center for Climate Change Communication, 2009).
44. Peters, R.G., Covello, V.T. & McCallum, D.B. e determinants of trust and
credibility in environmental risk communication: an empirical study. Risk Anal.
17, 43–54 (1997).
45. Adger, W.N. etal. Are there social limits to adaptation to climate change?
Clim. Change 93, 335–354 (2009).
46. Adger, W.N., Barnett, J., Brown, K., Marshall, N. & O’Brien, K. Cultural
dimensions of climate change impacts and adaptation. Nature Clim. Change
3, 112–117 (2013).
47. Kelly, P.M. & Adger, W.N. eory and practice in assessing vulnerability to
climate change and facilitating adaptation. Clim. Change 47, 325–352 (2000).
48. Moser, S.C., Kasperson, R.E., Yohe, G. & Agyeman, J. Adaptation to climate
change in the Northeast United States: opportunities, processes, constraints.
Mitig. Adapt. Strateg. Glob. Change 13, 643–659 (2008).
PERSPECTIVE
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2508

8 NATURE CLIMATE CHANGE | VOL 5 | MARCH 2015 | www.nature.com/natureclimatechange
49. Vogel, C., Moser, S.C., Kasperson, R.E. & Dabelko, G.D. Linking vulnerability,
adaptation, and resilience science to practice: Pathways, players, and
partnerships. Glob. Environ. Change 17, 349–364 (2007).
50. Dilling, L. & Lemos, M.C. Creating usable science: opportunities and
constraints for climate knowledge and their implications for science policy.
Glob. Environ. Change 21, 680–689 (2010).
51. Interagency Working Group on Ocean Acidication (Oce of the President,
Washington DC, 2013). [AUTHOR: is there a report title or further details?].
52. Cutter, S.L., Boru, B.J. & Shirley, W.L. Social vulnerability to environmental
hazards. Social Sci. Q. 84, 242–261 (2003).
53. Marshall, N. etal. A framework for social adaptation to climate change:
sustaining tropical coastal communities and industries, 36 (IUCN, 2010).
54. Cardona, O. etal. in IPCC Special Report of Working Groups I and II:
Managing the Risks of Extreme Events and Disasters to Advance Climate Change
Adaptation (eds Field, C. etal.) 65–108 (Cambridge Univ. Press, 2012).
55. Turner, B.L. I. etal. A framework for vulnerability analysis in sustainability
science. Proc. Natl Acad. Sci. USA 100, 8074–8079 (2003).
56. Bricker, S. etal. Eects of Nutrient Enrichment in the Nation’s Estuaries: A
Decade of Change, 328 (National Centers for Coastal Ocean Science, 2007).
57. US Geological Survey. USGS National Water Information System (NWIS)
database. Accessed online April 2014at http://waterdata.usgs.gov/nwis (2014).
58. Hoekstra, J.M. etal. Upwelling Presence by Marine Province
(Univ. California Press, 2010).
Acknowledgements
is work was supported by the National Socio-Environmental Synthesis Center
(SESYNC) under funding received from the National Science Foundation DBI-1052875.
Support for R.v.H. to generate model projections was provided by NOAA’s Coral Reef
Conservation Program. We thank the institutions and individuals that provided data (see
Supplementary Information for full details), and W. McClintock and his laboratory for
use of SeaSketch.org to enable collaborative discussions of spatial data and analysis. We
are grateful for the contributions and advice provided by E. Jewett throughout the project.
Author contributions
All authors provided input into data analysis and research design, and participated in at
least one SESYNC workshop; J.A.E. led the draing of the text with main contributions
from L.S., S.R.C., L.H.P., G.G.W. and J.E.C.; R.v.H. contributed projections of ocean
acidication; J.A.E., L.S., S.R.C., J.R. and C.D. collected the data; J.A.E. carried out data
analysis and mapping.
Additional information
Supplementary information is available in the online version of the paper. Reprints
and permissions information is available online at www.nature.com/reprints.
Correspondence should be addressed to J.A.E.
Competing financial interests
e authors declare no competing nancial interests. [AUTHORS: OK?]
PERSPECTIVE NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2508