Aquaculture has grown rapidly over the last three decades expanding at an average annual growth rate of 5.8% (2005–2014), down from 8.8% achieved between 1980 and 2010. The sector now produces 44% of total food fish production. Increasing demand and consumption from a growing global population are driving further expansion of both inland and marine aquaculture (i.e., mariculture, including marine species farmed on land). However, the growth of mariculture is dependent on the availability of suitable farming areas for new facilities, particularly for open farming practices that rely on the natural oceanic environmental parameters such as temperature, oxygen, chlorophyll etc. In this study, we estimated the marine areas within the exclusive economic zones of all countries that were suitable for potential open ocean mariculture activities. To this end, we quantify the environmental niche and inferred the global habitat suitability index (HSI) of the 102 most farmed marine species using four species distribution models. The average weighted HSI across the four models suggests that 72,000,000 km2 of ocean are to be environmentally suitable to farm one or more species. About 92% of the predicted area (66,000,000 km2) is environmentally suitable for farming finfish, 43% (31,000,000 km2) for molluscs and 54% (39,000,000 km2) for crustaceans. These predictions do not consider technological feasibility that can limit crustaceans farming in open waters. Suitable mariculture areas along the Atlantic coast of South America and West Africa appear to be most under-utilized for farming. Our results suggest that factors other than environmental considerations such as the lack of socio-economic and technological capacity, as well as aqua feed supply are currently limiting the potential for mariculture expansion in many areas.
Citation: Oyinlola MA, Reygondeau G, Wabnitz CCC, Troell M, Cheung WWL (2018) Global estimation of areas with suitable environmental conditions for mariculture species. PLoS ONE 13(1): e0191086. https://doi.org/10.1371/journal.pone.0191086
Editor: Luciano Bosso, Universita degli Studi di Napoli Federico II, ITALY
Received: July 17, 2017; Accepted: December 1, 2017; Published: January 19, 2018
Copyright: © 2018 Oyinlola et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Mariculture locations and production data are available from the Sea Around Us global mariculture database (http://www.seaaroundus.org/). Species ecological information from FishBase (http://www.fishbase.org/) and the Encyclopaedia of life (http://eol.org/). Species occurrence data from Ocean Biogeographic Information System (OBIS, http://www.iobis.org/), Global Biodiversity Information Facility (GBIF, http://www.gbif.org/), FishBase (http://www.fishbase.org/) and the International Union for the Conservation of Nature (IUCN, http://www.iucnredlist.org/technical-documents/spatial-data). Annual climatology for the period 1955-2012 for temperature, salinity, dissolved oxygen concentration and silicate concentration were obtained from the World Ocean Atlas 2013 (http://www.nodc.noaa.gov/OC5/woa13/). Euphotic depth and chlorophyll-a concentration annual climatology for the period from 1998 to 2012 were downloaded from the Ocean Colour website (http://oceancolor.gsfc.nasa.gov). Ocean current velocity data (1992-2002) from the Estimating the Circulation and Climate of the Ocean (ECCO) Project (http://www.ecco-group.org). Surface and bottom pH values were extracted from the Geophysical Fluid Dynamics Laboratory Earth System Model (GFDL-ESM-2G) and averaged over the period 1970-2000.
Funding: This study is a contribution by the Nippon Foundation- Nereus Program (http://www.nereusprogram.org/), an interdisciplinary ocean research program to WC. W. Cheung also acknowledges funding support from the Natural Sciences and Engineering Research Council of Canada.
Competing interests: The authors have declared that no competing interests exist.
Fish and other aquatic organisms, particularly from the marine environment, significantly contribute to the nutritional security and well-being of human society. Many coastal communities in tropical developing countries are highly dependent on fish as an important source of nutrients [1–3]. Given projected increases in the world's population over the next decades , together with increased inclusion of seafood in diets , global demand for seafood is expected to increase . Although, there may be some room for expansion of capture fisheries if fish stocks are rebuilt , production from capture fisheries has reached its maximum capacity in most parts of the global oceans (Fig 1) [7, 8]. Thus, increasing food production from aquaculture has been considered a main solution to meet the rising demand for seafood [9–11].
Data from Sea Around Us (2010) and FAO (2012).
Aquaculture has grown rapidly over the last three decades, expanding at an average annual growth rate of 5.8% (2005–2014) , down from an average of 8.8% between 1980 and 2010 . It is the fastest growing agro-food sector in the world . In 2014, aquaculture contributed about 73.4 million tonnes (44%) to total food fish production, with about 27.6 million tonnes (37.6%) of this production from aquaculture in marine and brackish environment, also known as mariculture  (Fig 1). It is important to emphasise that more than 60% of this production stems from the farming of molluscs and crustaceans. Mariculture has expanded by 9.3% in production since 1990 [15, 16]. Currently, 112 countries and territories produce seafood in the marine environment  with earnings reaching 65.4 billion USD in 2013 from the mariculture sector  and representing 43.5% of the total aquaculture income. Given mariculture’s significant contribution to seafood supply and the economy, various attempts are being made to estimate the potential to expand aquaculture in the ocean . Information on area potentially suitable for mariculture could be useful for planning ocean based activities (e.g. energy production, shipping, marine protected areas) [18, 19].
A range of environmental and social-economic factors influence the sustainable development of mariculture. Most mariculture involves growing fish or invertebrates in nets or cages that are submerged, allowing for free water exchange with the surrounding marine environment [20, 21]. Farmed species’ survival and growth rates are directly influenced by natural environmental conditions  that consequently affect the suitability of an area to farm these species. However, a number of other factors play an important role in determining the actual production capacity and its sustainability. These include but are not limited to: the environmental carrying capacity of the farming area , feed ingredients, e.g. the demand for forage fish in aquafeeds [8, 24, 25] or plant proteins , the development of closed life cycles farming techniques to reduce the need for wild-caught juveniles and broodstock for restocking [27, 28], other technological developments that reduce the environmental impacts of mariculture ; and the quality of governance to ensure sustainable mariculture practices .
An important first step towards better describing the ocean environmental spatial extent for sustainable mariculture production potential, is the identification of marine areas that offer suitable environmental conditions for currently farmed species. Since most ocean-based mariculture operations use open facilities in which farmed organisms are exposed to natural environmental conditions, the physical and chemical properties of the waters may thus affect the organisms’ growth and survivorship, and consequently, their potential for mariculture production. Here, suitable environmental conditions refer to the area of the ocean that can support the physiological needs of farmed species for sustainable mariculture production . Environmental preferences of (farmed) marine species can be approximated and mapped using species distribution models (SDMs). Based on the environmental niche theory , this modelling approach consists of quantitatively describing the relationship between a species' observed occurrence records and various parameters that describe its environment. Such relationship can be developed using historical occurrence records of the species in both the natural and farmed environments. SDMs can be applied to predict species’ distribution in the past, present and future [33, 34].
This study aims to predict the spatial extent of area that is suitable environmentally for mariculture. We applied four SDMs: Ecological Niche Factor Analysis (ENFA) , the Non- Parametric Probabilistic Ecological Niche (NPPEN) , Maximum Entropy (MAXENT)  and Surface Range Envelope (SRE) , to quantify the environmental niche of the presently important farmed species and project their habitat suitability index (HSI) over the global ocean based on current environmental conditions. We focused on coastal and open water farming systems only. Based on the predicted HSI, we calculated the total area of the world’s exclusive economic zones (EEZ) that is suitable for farming marine species. We examined the variation among models and compare it with the mean predictions across models to highlight where predictions were most robust to variations. We then evaluated model projection uncertainties and estimated the total area that would be suitable for mariculture. Finally, we discussed the implications of our results for future mariculture development.
Material and methods
1. Biotic data
We obtained a list of farmed species from the Sea Around Us mariculture database (SAU, www.seaaroundus.org). The database is derived largely from the Food and Agricultural Organisation (FAO) database, with augmented information from national statistics to subdivide annual mariculture production by sub-national units (e.g., provinces, states), in addition to countries and taxa, for the period 1950 to 2010. We extracted the species’ names of all fish and invertebrates reported in the database (307 in total). Records that are not reported at the species level (i.e., with genus and species specified) were excluded from our analysis. Following the minimum occurrence data requirements for SDM , we only retained species that occurred in more than seven sub-national units. This is to ensure a greater model accuracy with higher numbers of occurrence locations. Thus, this study focused on a total of 102 species (57 Chordata, 15 Arthropods, 29 Mollusca and 1 Echinodermata) (S1 Table). Although seaweed mariculture contributes substantially to global production, we did not include seaweed in this study because of the lack of geospatial data on seaweed mariculture locations. Species’ ecological information, including maximum depth, minimum depth, trophic level, preferred biome  and habitat, were obtained from FishBase (http://www.fishbase.org/) and the Encyclopaedia of life (http://eol.org/).
To obtain representative spatial distribution of each farmed species and quantify their environmental niche, we developed two databases. The first one consisted of natural occurrence records (i.e., from the wild) for all 102 species from a number of open source databases: Ocean Biogeographic Information System (OBIS, http://www.iobis.org/), Global Biodiversity Information Facility (GBIF, http://www.gbif.org/), FishBase (http://www.fishbase.org/) and the International Union for the Conservation of Nature (IUCN, http://www.iucnredlist.org/technical-documents/spatial-data). For each species, we removed duplicate records of occurrences among databases and records for which geographic information was not available.
Secondly, we developed a georeferenced occurrence database exclusively for mariculture (i.e., coastal and open ocean farmed). Using all sub-national units where farms were recorded in the Sea Around Us mariculture database, we visually identified any mariculture installations (pens, cages and lines) based on satellite photos available from Google Earth (http://www.google.com/earth/). The coordinates were extracted for each installation by using the Google Earth placemark tool .
All species’ natural occurrence and mariculture location records were converted to a binary database of presence or absence and rasterized on a regular spatial grid of 0.5° latitude by 0.5° longitude over the global ocean.
2. Environmental data
We collected eight environmental parameters: temperature, dissolved oxygen concentration, chlorophyll-a concentration, salinity, pH, silicate concentration, current velocity and euphotic depth. Annual climatology for the period 1955–2012 for temperature, salinity, dissolved oxygen concentration and silicate concentration were obtained from the World Ocean Atlas 2013 (http://www.nodc.noaa.gov/OC5/woa13/). Euphotic depth and chlorophyll-a concentration annual climatology for the period from 1998 to 2012 were downloaded from the Ocean Colour website (http://oceancolor.gsfc.nasa.gov). We obtained 10-year averaged ocean current velocity data (1992–2002) from the Estimating the Circulation and Climate of the Ocean (ECCO) Project (http://www.ecco-group.org). Surface and bottom pH values were extracted from the Geophysical Fluid Dynamics Laboratory Earth System Model (GFDL-ESM-2G) and averaged over the period 1970–2000. All environmental data were interpolated using bilinear methods  over the global ocean (189.75°W to 179.75°E and 89.75°N to 89.75°S) on a regular spatial grid of 0.5° latitude by 0.5o longitude (the same as occurrence rasterized data) and for two vertical layers: surface (0-10m) and sea bottom depth where available.
3. Modelling habitat suitability
We predicted the habitat suitability for each farmed species on 0.5o by 0.5o grid of the global ocean using species distribution models (SDMs). Firstly, we harmonised the biotic and environmental data based on the regular spatial grid coordinates with data on occurrences in natural and mariculture environments. Secondly, we determined the most important environmental parameters to model the farmed marine species’ distribution using the eigenvalue diagram implemented in Ecological Niche Factor Analysis (ENFA) . The diagram was constructed based on the departure of the ecological niche from the mean habitat for each species, thus identifying the species’ preference for particular environmental parameters among the whole set of parameters (S1 Fig). We selected the most important set of environmental parameters by identifying the direction in which the ‘specialisation’ was highest. Specialisation is a measure of the narrowness of the niche, i.e., the higher the specialisation, the more restricted the niche. We then used these selected parameters for all SDM in this study. Finally, we predicted the habitat suitability for each species (mariculture and natural occurrence) using four SDMs. A multiple-model approach was used to explore the variations and uncertainty of predictions from different SDMs [37, 46].
The four SDMs predict species distributions based on different algorithms and assumptions. The first SDM, ENFA, is an analysis that uses multivariate statistics to assess species’ habitat selection by providing the realised niche measurement within the available hyper-space from two estimates: the marginality, that identifies a species’ preference for given environmental conditions, and its specialisation (i.e., the species’ sensitivity to variations in its optimum environment) .The lower the marginality, the less the niche deviates from available conditions; the higher the specialisation the more restricted the niche . The second model, the Non- Parametric Probabilistic Ecological Niche (NPPEN) , is based on a simplification of the Multiple Response Permutation Procedures (MRPP) using the Generalised Mahalanobis distance. The other two SDMs, Maximum Entropy (MAXENT)  and Surface Range Envelope (SRE) , use various numerical procedures. MAXENT estimates the ratio of a species’ presence site to the study area and then calculates the probability of occurrence through a logistic transformation . SRE is an environmental envelope model that identifies cells, which have environmental values that fall within the range of values measured from the presence data . All the SDMs use presence-only data to determine a species’ environmental distribution. We applied each SDM to predict a species’ habitat suitability index (HSI)—an index that scales from 0 to 1 to indicate the environmental suitability of the selected environmental conditions in each spatial cell for each studied species.
4. Model testing
We tested the robustness of the SDM outputs by comparing the predicted HSI with reference records of species occurrences. Mariculture and natural occurrence data were analysed separately and divided into two datasets: 75% of all records were used for training purposes to develop each SDM and calculate each species’ HSI. The remaining 25% of the data were used for model evaluation. Specifically, the Area Under the Receiver Operating Characteristic Curve (AUC) of each set of model predictions was calculated using the ROCR package in R . The AUC values range from 0 to 1, with 0.5 indicating that the model is no better than a random sample of values and 1 indicating that the model has high predictive power.
We examined the correlation between the HSI predictions from the natural and mariculture occurrence data. For each species, we fit a linear model to predicted HSI from mariculture and natural occurrences as dependent and independent variables, respectively. We cross-validated this statistical relationship by applying a generalised linear model to the combined species HSI by assuming binomial error distribution. We calculated the degree of overdispersion for proportional data from residue deviance and then fitted a quasibinomial distribution to account for the overdispersion . We set a minimum predicted HSI for each species as the value below which a species was considered to be absent. This approach used species-specific minimum HSI thresholds instead of a fixed threshold for all species, as used by Jones and Cheung . The thresholds were identified by quantifying a species’ “prevalence” (i.e., the fraction of cells at which the species is present)  and represent the minimum HSI threshold for that species. This tests mariculture location HSI (presence) versus natural occurrence HSI (absence). The predictive values of habitat suitability for mariculture below the prevalence value were assigned 0, and predicted HSI values higher than the prevalence value were assigned 1.
5. Identifying potential mariculture area
The first criterion that we used to define potential mariculture area was the suitability of the environment for farmed species. For an area to be suitable for mariculture, it needs to meet the minimum environmental conditions for the growth and survival of farmed species. Thus, for each farmed species, its potential mariculture area must have an HSI above its minimum threshold (i.e., prevalence).
Secondly, we assumed that mariculture operations would not go beyond the area of sovereign nations’ jurisdiction, and that water current and ocean management would be gruelling operations. Thus, we constrained the potential mariculture area to be within countries’ Exclusive Economic Zone (EEZ). We also assumed that mariculture could only be found in spatial cells with a current velocity between 10 cm.s-1 and 100 cm.s-1 . Low current would result in more rapid food depletion (less particulate organic matter flow) and less efficient production , less waste removal (feed and organic waste), and high benthic impact in finfish aquaculture . In contrast, strong currents can damage farm structures and holding facilities  and affect the growth of farmed fish through skeletal malformations . Furthermore, as most marine protected areas (MPAs) do not allow mariculture activities to take place, they were not considered as suitable future mariculture areas (Table 1).
We calculated the total suitable mariculture area for the 102-farmed species considered for each model and derived an overall multi-model weighted average (each model contributed to the final output based on its AUC value). We also calculated the number of species that were predicted to be suitable for mariculture in each spatial cell. We then compared the spatial distribution of potential mariculture area and species richness of the 20-farmed species with the highest cumulative production from 1950 to 2010.
1. Prediction of HSI
The predicted mariculture and natural habitat suitability index generally agree with observed occurrences. Almost all predicted species distributions for all species and models had AUC values greater than 0.7 (Fig 2). The AUC values varied among models, with predictions from SRE having a median AUC across species < 0.8, and predictions from MAXENT scoring the highest median AUC values. The 25th and 75th percentiles of the estimated prevalence value for each species ranged from 0.30 to 0.62.
(A) AUC for the habitat suitability index (HSI) for natural occurrences of farmed species across SDM; (B) AUC for mariculture location HSI across SDMs. The horizontal lines represent median values. The upper and lower boundaries of the box represent the upper and lower quartiles of the data. ENFA- Ecological Niche Factor Analysis, MAXENT- Maximum Entropy, NPPEN- Non- Parametric Probabilistic Ecological Niche and SRE- Surface Range Envelope.
2. Potential mariculture area and species richness
Overall, we found a significant and positive relationship between the environmental niche of mariculture locations and that of farmed species’ natural occurrences. Across all farmed species, we found a significant and positive linear relationship between mariculture occurrence-based (y) and natural occurrence-based (x) HSI (y = 0.27x + 0.54, p < 0.001) (Fig 3A). For each individual species, the adjusted R-square values of the regression between mariculture and natural-based HSI were above 0.50, with a mean of 0.66 (Fig 3A). Species that showed the strongest relationship between the mariculture-based and natural-based HSI were Red drum (Sciaenops ocellatus), Peruvian scallop (Argopecten purpuratus) and Pacific cupped oyster (Crassostrea gigas) (R2 = 0.90, 0.88 and 0.86, respectively).
(A) Regression of global predicted mariculture HSI and natural occurrence HSI (p<0.001). (B) Histogram of adjusted R2 of individual species’ regression analysis with a mean value of 0.66.
The predicted environmentally suitable areas for mariculture with high agreement among the four SDMs are found between 66.5o N and 66.5o S (Fig 4). Based on the weighted average predictions from the four SDMs, the total suitable mariculture area for the 102 species was estimated at 72 million km2 (Fig 5A). Predictions from ENFA result in the largest suitable marine area (107.0 million km2) while NPPEN yielded the lowest value (91.5 million km2) (Fig 6). Sixty-six million km2 of this area is suitable for finfish, 39 million km2 for crustaceans and 31 million km2 for mollusc. This included areas currently used for mariculture purposes. Areas that were predicted as unsuitable for the mariculture of any species include Antarctica and pockets around the Arctic region.
Blue—high agreement (4 models), Yellow—moderate agreement (3 models), Green—low agreement (2 models) and Red–very low agreement (1 model).
(A) Total predicted suitable marine areas for mariculture in blue and unsuitable marine areas in red based on an average of four different species distribution models; (B) Comparison between present numbers of species farmed in different countries with potential numbers of farmed species based on model outputs.
ENFA predicted the highest area with 107 million km2 while the Multi-model predicted 72 million km2.
Mariculture species richness was highest in potential mariculture areas between 25° N and 25° S (S4 Fig) and include the southwestern Atlantic coast (species richness = 30–45 among models) and West Africa (species richness = 35–40 among models). Other notable areas with high mariculture species richness included the Gulf of Mexico, the Caribbean Sea, the East China Sea, the Yellow Sea, the Sea of Japan and the Banda Sea off the coast of Timor-Leste. However, despite results showing high species richness for these potential mariculture areas, most of the species were not being farmed in these regions.
Specifically, we found a large potential area for finfish mariculture in the tropics, between 20°N and 20°S (mariculture species richness between 5 and 30 spp.) (Fig 7). In contrast, molluscs were predicted to be the dominant mariculture group in the temperate areas (23.0° to 66.5° in both the northern and southern hemispheres), with mariculture species richness of 3 to 15 spp. Large potential areas for crustacean mariculture were predicted in waters within 23° N to 25° N and 23° S to 25° S, with predicted mariculture species richness of 2 to 12 spp. and 3 to 11 spp., respectively.
3. Species and country-level comparisons
Predicted potential mariculture area varied across the most important farmed species. Milkfish (Chanos chanos) was predicted to have the largest suitable farming area (21 million km2), while hooded oyster (Saccostrea cucullata) was predicted to have the smallest area (0.2 million km2) amongst the 20-species selected for comparison (Fig 8A). We found a minimum of 10% difference between the number of countries currently practising mariculture and those with the potential for marine farming production. For example, predicted mariculture potential area was 16% larger than the area where mariculture was reported to have occurred for Atlantic salmon (Salmo salar). For Cobia (Rachycentron canadum) and Pacific cupped oyster (Crassostrea gigas), reported mariculture area was 80% and 46% below their predicted spatial extent, respectively.
1. Overall model performance
By using an ensemble of SDMs , we were able to explore the structural uncertainty of the predicted mariculture species distributions and environmental suitability. Projections from ENFA and SRE showed considerable variation in comparison with MAXENT and NPPEN. MAXENT and NPPEN had the highest AUC values. For the former approach, this may be due to the model’s weighting ability of input variables , thereby preventing over-fitting of the model by determining how much the likelihood should be penalised . NPPEN measures the correlation between species’ occurrence and distribution, thus avoiding random quantile classes selection to allow direct estimation of species occurrence . The multi-SDM approach explored the uncertainty due to discrepancies among models thereby capturing the full range of potential suitable marine areas for mariculture.
The relevance of the set of environmental factors identified by the modelling algorithms applied here to best predict mariculture areas are supported by empirical knowledge [58–61]. For example, the physiology of farmed species is affected by changes in temperature . Optimal temperature determination for farmed species growth is important practice in mariculture . In addition, oxygen is an important factor affecting the growth of fishes , and dissolved oxygen level is an important consideration in determining the holding capacity of fish farms [65, 66]. Salinity influences the growth some of farmed species, with salinity levels above 25 Practical Salinity Unit (PSU) having shown sometimes to result in high mortality and retarded growth in cultured scallop (pecten maximus) [67, 68] while such salinity level could yield better growth performance in seabream . Moreover, primary production and suspended particulate organic matter or seston, indicated by chlorophyll a concentrations and silicate, respectively, are food sources for molluscs [70, 71]. Thus, their concentration is expected to affect the growth of farmed molluscs.
2. Mariculture opportunities and limitations
Overall, our findings suggest that the global ocean environmentally suitable area for mariculture is much larger than the area in which mariculture is currently practised. Particularly, most offshore areas considered environmentally suitable are not being used for farming activities. Our results therefore suggest that the lack of environmentally suitable area for mariculture is not the main limiting factor for the expansion of mariculture  in most regions of the world. Instead, other factors such as the socio-economics of producing countries, including capacity and political instability; technology, its availability and cost effectiveness ; trades; aqua feed availability [73, 74]; aquaculture development-related policies [75, 76] and competition for space within an EEZ for instance; shipping, oil and gas, as well as tourism all play major roles in the development of mariculture operations and their future expansion.
Our conclusion about the area suitable for mariculture and its limitation for its utilization for mariculture operations agree with a related study that employed different methods to estimate the global potential area for mariculture . Specifically, Gentry et al.  applied more constraints than our study in estimating available ocean area for mariculture, resulting in a lower estimate. For instance, our study does not consider water depth as a limiting factor to define environmentally suitable areas for finfish, as the use of submersible cages is widespread for a number of farmed species (e.g., Salmon, Cobia) . In this case feed development, socio-economic and technological factors may represent a constraint to farming rather than water depth itself . The use of multiple approaches to predict suitable mariculture area should be encouraged so that the uncertainties associated with such predictions can be better characterised.
Regionally, our results show that the difference between the predicted environmentally suitable area for mariculture and the extent of current mariculture activities is largest in Africa, the Caribbean and along the Atlantic coast of South America. These regions are predicted hotspots for mariculture species richness (60% of the 102 species), yet actual mariculture operations appear to be relatively limited, accounting for only 1.3% of global mariculture production [15, 16]. Factors such as poor economic conditions, lack of supporting infrastructure , political instability, limited foreign investment in the sector  and inadequate value chain linkages  in many countries of these regions may have impeded mariculture development.
China, in contrast, is presently using the largest extent of its suitable area for marine aquaculture. The country has a long history of aquaculture, going back 2500 years . Owing to the late 70s economic reform initiatives, China’s aquaculture industry has benefited from open market policies. In addition, the fact that China accounts for one- quarter of the global fish demand has made the country’s market suitable for expansion. However, farm expansion might exert more pressure on wild capture fisheries as China’s aquaculture industry accounts for one- third of global fishmeal production .
Concerns regarding the broader environmental sustainability of mariculture may limit the sector’s expansion. Particularly, the expansion of farming carnivorous species will increase the demand for fish resources in feeds, adding to the stress on forage fisheries and also other non-selective trawling . Fishmeal and oil are reduction fisheries products and the two most essential inputs into aqua feed since they are highly digestible and essential sources of amino acids [83, 84]. While successful partial replacement of fishmeal with plant and other sources (e.g. insects, yeast and algae) has been tested and used in aqua feed [85, 86], there are still challenges remaining for using fully plant-based feed source particularly for high value species such as Atlantic salmon . High quality plant based ingredients used in feeds must also in the long run be replaced by innovative resources not competing with human food.
Affordable and efficient technology can also contribute to the sustainable expansion of mariculture. Significant advances have been achieved in land-based systems designed to reduce nutrient load discharges and farmed species escapes, as well as improve disease management [87–89]. In coastal finfish and shellfish mariculture systems, a variety of cages are being developed to withstand high wave actions and reduces escapes in offshore areas [21, 90]; advances that will be necessary given predictions of increased storm activity in the future [91, 92]. For crustaceans, such advances are lagging behind and a number of significant hurdles will need to be overcome for these to become operational. In addition, integrated multi-trophic aquaculture systems (IMTA) where lower trophic level species, such as seaweed and bivalves, are farmed together with finfish  maximise inorganic and organic nutrient waste recycling, thereby reducing the environmental footprint of mariculture farms. While these systems hold much promise, large commercial IMTA farms are still uncommon.
Strong environmental governance is also needed to regulate and ensure the sustainable development of mariculture. Aquaculture activities are regulated by law in many countries (e.g., Canada, China, Norway and the Philippines ). Effective monitoring and enforcement are imperative if mariculture expansion is to be sustainable in the future. However, the monitoring and enforcement varies considerably among countries . Countries such as the U.S. , Australia , Canada and the E.U.  have also adopted formal codes of conduct whose main objective is to promote the responsible development and management of aquaculture with appropriate sanctions.
The high correlation for individual species’ predicted HSI between mariculture location and natural occurrence in this study, suggests that mariculture farms are sensitive to changes in environmental conditions driven by climate change or other anthropogenic activities such as pollution [35, 36, 98–100]. Ocean warming may drive environmental conditions beyond the suitable range for mariculture and will cause thermal stress for a number of currently farmed species (e.g., cod, oysters). However, these rising temperatures may extend the the growing season for some species and may provide opportunities to farm new species, or species that are currently economically marginal in the affected areas. Also, shellfish aquaculture is sensitive to ocean acidification [101–103], as lower carbonate saturation state in the water can make it more difficult for calcifying invertebrates to produce shells. Carbon emissions may thus have a substantial impact on the distribution and diversity of potentially suitable farm sites for currently farmed species.
In this study, we identified a large global environmentally suitable area for mariculture. We suggest that other non-environmental factors such as technological, economic and social constraints play a fundamental role in determining mariculture production in these regions. While, our approach is useful in defining areas on a broad scale, more detailed studies on the ‘ecological suitability’ (i.e., in terms of carrying capacity) will be required to further constraint the prediction of suitable area for mariculture development. Also, these currently suitable areas might transition to being unsuitable in the future due to human activities such as pollution, coastal zone activities, and climate change. Future studies should include other human uses of marine areas e.g. ship routes, wind farms etc. and their potential competition with mariculture to further characterize the suitable areas for farming. Moreover, given the importance of seaweed farming in various regions, future studies should also collate data and information about seaweed mariculture locations and extend all analyses to seaweed. It will also be important to investigate and address the main constraints in sustainable mariculture development to help develop pathways that ensure the continued contribution of mariculture to global seafood production.
S1 Fig. ENFA biplot with the x-axis (marginality) and y-axis (specialisation).
The white dot within the dark area represents the centre of used area while the light area is the available niche. The arrows are projections of oceanic parameters based on mariculture locations of the species (A) Pacific cupped oyster (Crassostrea gigas) (B) Cobia (Rachycentron canadum) (C) Atlantic salmon (Salmo salar) (D) Giant tiger shrimp (Penaeus monodon).
S2 Fig. Positive linear relationship between mariculture occurrence based (y) and natural occurrence based (x) HSI (y = 0.27x + 0.54, p < 0.001).
S3 Fig. The linear regression between predicted mariculture location and natural occurrence habitat suitability.
(A) Pacific cupped oyster (Crassostrea gigas) y = 0.01773x + 0.7906, R2 = 0.8607, p < 0.0001.
(B) Cobia (Rachycentron canadum) y = 0.04762x + 0.5525, R2 = 0.5951, p < 0.0001.
(C) Atlantic salmon (Salmo salar) y = 0.03001x + 0.7027, R2 = 0.8118, p < 0.0001.
(D) Giant tiger shrimp (Penaeus monodon) y = 0.02476x + 0.7012, R2 = 0.7383, p < 0.0001.
S4 Fig. Latitudinal predicted species richness.
(A) global mariculture species richness (B) Finfish (C) Crustacean (D) Molluscs
This study is a contribution by the Nippon Foundation- the University of British Columbia Nereus Program, an interdisciplinary ocean research program.
- 1. Golden C, Allison Edward H., Cheung WWL, Dey Madan M., Halpern Benjamin S., McCauley Douglas J., et al. Fall in fish catch threatens human health. Natural. 2016;534(7607):317–20.
- 2. Roos N, Wahab MAC, Thilsted SH. The role of fish in food-based strategies to combat vitamin A and mineral deficiencies in developing countries. The journal of Nutrition. 2007;137(4):1106–9. pmid:17374688
- 3. Tacon AGJ, Metian M. Fish Matters: Importance of Aquatic Foods in Human Nutrition and Global Food Supply. Reviews in Fisheries Science. 2013;21(1):22–38.
- 4. UN. World Economic and Social Survey: Sustainable Development Challenges. Department of Economic and Social Affairs. 2013.
- 5. Delgado CL. Fish to 2020: Supply and demand in changing global markets. WorldFish. 2003;62.
- 6. Costello C, Ovando D, Hilborn R, Gaines SD, Deschenes O, Lester SE. Status and solutions for the world’s unassessed fisheries. Science. 2012;338(6160):517–20.
- 7. Watson RA, Pitcher TJ, Jennings S. Plenty more fish in the sea? Fish and Fisheries. 2017:18(1):105–13.
- 8. Pauly D, Zeller D. Catch reconstructions reveal that global marine fisheries catches are higher than reported and declining. Nat Commun. 2016;7:10244. pmid:26784963
- 9. Naylor RL, Goldburg RJ, Primavera JH, Kautsky N, Beveridge MC, Clay J, et al. Effect of aquaculture on world fish supplies. Nature. 2000;405(6790):1017–24. pmid:10890435
- 10. Beveridge MC, Thilsted SH, Phillips MJ, Metian M, Troell M, Hall SJ. Meeting the food and nutrition needs of the poor: the role of fish and the opportunities and challenges emerging from the rise of aquaculture. J Fish Biol. 2013;83(4):1067–84. pmid:24090563
- 11. Belton B, Thilsted SH. Fisheries in transition: Food and nutrition security implications for the global South. Global Food Security. 2014;3(1):59–66.
- 12. FAO. The State of World Fisheries and Aquaculture. Contributing to food security and nutrition for all. Rome. 200 pp. 2016.
- 13. FAO. The state of world fisheries and aquaculture. Food & Agriculture Org. 2012.
- 14. Toufique KA, Belton B. Is Aquaculture Pro-Poor? Empirical Evidence of Impacts on Fish Consumption in Bangladesh. World Development. 2014;64:609–20.
- 15. Campbell B, Pauly D. Mariculture: A global analysis of production trends since 1950. Marine Policy. 2013;39:94–100.
- 16. FAO. The State of World Fisheries and Aquaculture. Opportunities and challenges. Food & Agriculture Org Rome. 2014.
- 17. Gentry RR, Froehlich HE, Grimm D, Kareiva P, Parke M, Rust M, et al. Mapping the global potential for marine aquaculture. Nature Ecology & Evolution. 2017;1(9):1317–24.
- 18. Edgar GJ, Stuart-Smith RD, Willis TJ, Kininmonth S, Baker SC, Banks S, et al. Global conservation outcomes depend on marine protected areas with five key features. Nature. 2014;506(7487):216–20. pmid:24499817
- 19. Venter O, Fuller RA, Segan DB, Carwardine J, Brooks T, Butchart SH, et al. Targeting global protected area expansion for imperiled biodiversity. PLoS Biol. 2014;12(6):e1001891. pmid:24960185
- 20. Islam MS. Nitrogen and phosphorus budget in coastal and marine cage aquaculture and impacts of effluent loading on ecosystem: review and analysis towards model development. Marine Pollution Bulletin. 2005;50(1):48–61. pmid:15664033
- 21. Benetti DD, Benetti GI, Rivera JA, Sardenberg B, O’Hanlon B. Site selection criteria for open ocean aquaculture. Marine Technology Society Journal. 2010;44(3):22–35.
- 22. Pillay TVR. Aquaculture and the Environment. John Wiley & Sons. 2008.
- 23. Byron C, Link J, Costa-Pierce B, Bengtson D. Calculating ecological carrying capacity of shellfish aquaculture using mass-balance modeling: Narragansett Bay, Rhode Island. Ecological Modelling. 2011;222(10):1743–55.
- 24. Tacon AGJ, Metian M. Fishing for Aquaculture: Non-Food Use of Small Pelagic Forage Fish—A Global Perspective. Reviews in Fisheries Science. 2009;17(3):305–17.
- 25. Cashion T, Le Manach F, Zeller D, Pauly D. Most fish destined for fishmeal production are food-grade fish. Fish and Fisheries. 2017.
- 26. Troell M, Naylor RL, Metian M, Beveridge M, Tyedmers PH, Folke C, et al. Does aquaculture add resilience to the global food system? Proc Natl Acad Sci U S A. 2014;111(37):13257–63. pmid:25136111
- 27. Hermansen O, Dreyer B. Capture-based aquaculture-sustainable value adding to capture fisheries? 2008.
- 28. Ottolenghi F, Silvestri C, Giordano P, Lovatelli A, New MB. Capture-based aquaculture: the fattening of eels, groupers, tunas and yellowtails. FAO. 2004.
- 29. Troell M, Joyce A, Chopin T, Neori A, Buschmann AH, Fang J-G. Ecological engineering in aquaculture—Potential for integrated multi-trophic aquaculture (IMTA) in marine offshore systems. Aquaculture. 2009;297(1–4):1–9.
- 30. Hofherr J, Natale F, Trujillo P. Is lack of space a limiting factor for the development of aquaculture in EU coastal areas? Ocean & Coastal Management. 2015;116:27–36.
- 31. Tidwell JH, Allan GL. Fish as food: aquaculture's contribution. EMBO reports. 2001;2(11):958–63. pmid:11713181
- 32. Hutchinson GE. Homage to Santa Rosalia or why are there so many kinds of animals? The American Naturalist. 1959;93(870):145–59.
- 33. Guisan A, Zimmermann NE. Predictive habitat distribution models in ecology. Ecological modelling, 135(2), 147–186. 135. 2000;2:147–86.
- 34. Elith J, Leathwick JR. Species Distribution Models: Ecological Explanation and Prediction Across Space and Time. Annual Review of Ecology, Evolution, and Systematics. 2009;40(1):677–97.
- 35. Cheung WWL, Brodeur RD, Okey TA, Pauly D. Projecting future changes in distributions of pelagic fish species of Northeast Pacific shelf seas. Progress in Oceanography. 2015;130:19–31.
- 36. Cheung WWL, Lam VWY, Sarmiento JL, Kearney K, Watson REG, Zeller D, et al. Large-scale redistribution of maximum fisheries catch potential in the global ocean under climate change. Global Change Biology. 2010;16(1):24–35.
- 37. Jones MC, Cheung WWL. Multi-model ensemble projections of climate change effects on global marine biodiversity. ICES Journal of Marine Science. 2015;72(3):741–52.
- 38. Basille M, Calenge C, Marboutin É, Andersen R, Gaillard J-M. Assessing habitat selection using multivariate statistics: Some refinements of the ecological-niche factor analysis. Ecological Modelling. 2008;211(1–2):233–40.
- 39. Beaugrand G, Lenoir S, Ibañez F, Manté C. A new model to assess the probability of occurrence of a species, based on presence-only data. Marine Ecology Progress Series. 2011;424:175–90.
- 40. Phillips SJ, Anderson RP, Schapire RE. Maximum entropy modeling of species geographic distributions. Ecological Modelling. 2006;190(3–4):231–59.
- 41. Busby J. BIOCLIM-a bioclimate analysis and prediction system. Plant Protection Quarterly (Australia). 1991.
- 42. Elith J, Graham CH, Anderson R. P., Dudı´k M., Ferrier S, Guisan A, Hijmans RJ, Huettmann F, Leathwick J. R., Lehmann A., et al. Novel methods improve prediction of species’ distributions from occurrence data. Ecography. 2006;29:129–51.
- 43. Reygondeau G, Longhurst A, Martinez E, Beaugrand G, Antoine D, Maury O. Dynamic biogeochemical provinces in the global ocean. Global Biogeochemical Cycles. 2013;27(4):1046–58.
- 44. Trujillo P, Piroddi C, Jacquet J. Fish farms at sea: the ground truth from Google Earth. PLoS One. 2012;7(2):e30546. pmid:22347383
- 45. Legendre P, Legendre L. Numerical Ecology Developments in Environmental Modelling. 1998;24.
- 46. Jones MC, Dye SR, Pinnegar Jk, Warren R, Cheung WWL. Applying distribution model projections for an uncertain future: the case of the Pacific oyster in UK waters. Aquatic Conservation: Marine and Freshwater Ecosystems. 2013; 23(5): 710–22.
- 47. Hirzel AH, Hausser J, Chessel D, Perrin N. Ecological‐niche factor analysis: how to compute habitat‐suitability maps without absence data? Ecology. 2002;83(7):2027–36.
- 48. Elith J, Phillips SJ, Hastie T, Dudík M, Chee YE, Yates CJ. A statistical explanation of MaxEnt for ecologists. Diversity and Distributions. 2011;17(1):43–57.
- 49. Hernandez PA, Graham CH, Master LL, Albert DL. The effect of sample size and species characteristics on performance of different species distribution modeling methods. Ecography. 2006;29(5):773–85.
- 50. Sing T, Sander O, Beerenwinkel N, Lengauer T. The ROCR package. Technical Report, http://rocr.bioinf.mpi-sb.mpg.de. 2007.
- 51. Crawley MJ. Statistics: an introduction using R, 2005. Wiley, ISBN 0, 470(02298), 1. 2005.
- 52. Phillips SJ, Dudík M, Elith J, Graham CH, Lehmann A, Leathwick J, et al. Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. Ecological Applications. 2009;19(1):181–97. pmid:19323182
- 53. Kapetsky JM, Aguilar-Manjarrez J, Jenness J. A global assessment of potential for offshore mariculture development from a spatial perspective. FAO. 2013.
- 54. Ferreira JG, Hawkins AJS, Bricker SB. Management of productivity, environmental effects and profitability of shellfish aquaculture—the Farm Aquaculture Resource Management (FARM) model. Aquaculture. 2007;264(1–4):160–74.
- 55. Borja Á, Rodríguez JG, Black K, Bodoy A, Emblow C, Fernandes TF, et al. Assessing the suitability of a range of benthic indices in the evaluation of environmental impact of fin and shellfish aquaculture located in sites across Europe. Aquaculture. 2009;293(3–4):231–40.
- 56. Chatain B. Abnormal swimbladder development and lordosis in sea bass (Dicentrarchus labrax) and sea bream (Sparus auratus). Aquaculture. 1994;119(4):371–9.
- 57. Phillips SJ, Dudík M. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography. 2008;31(2):161–75.
- 58. Ross LG, QM EM, Beveridge MCM. The application of geographical information systems to site selection for coastal aquaculture: an example based on salmonid cage culture. Aquaculture. 1993;112(2):165–78.
- 59. Karthik M, Suri J, Saharan N, Biradar RS. Brackish water aquaculture site selection in Palghar Taluk, Thane district of Maharashtra, India, using the techniques of remote sensing and geographical information system. Aquacultural Engineering. 2005;32(2):285–302.
- 60. Longdill PC, Healy TR, Black KP. An integrated GIS approach for sustainable aquaculture management area site selection. Ocean & Coastal Management. 2008;51(8–9):612–24.
- 61. McDaid Kapetsky J, Aguilar-Manjarrez J, Jenness J, Dean A, Salim A. A global assessment of offshore mariculture potential from a spatial perspective. FAO, Roma (Italia). 2013.
- 62. Cochrane K, De Young C, Soto D, Bahri T. Climate change implications for fisheries and aquaculture. FAO Fisheries and aquaculture technical paper. 2009;530(212).
- 63. Imsland AK, Sunde LM, Folkvord A, Stefansson SO. The interaction of temperature and fish size on growth of juvenile turbot. Journal of fish biology. 1996;49(5):926–40.
- 64. Mallya YJ. The effects of dissolved oxygen on fish growth in aquaculture., pp30. The United Nations University fisheries training program, Final project. 2007:pp30.
- 65. Merino GE, Piedrahita RH, Conklin DE. The effect of fish stocking density on the growth of California halibut (Paralichthys californicus) juveniles. Aquaculture. 2007;265(1–4):176–86.
- 66. Besson M, Vandeputte M, van Arendonk JAM, Aubin J, de Boer IJM, Quillet E, et al. Influence of water temperature on the economic value of growth rate in fish farming: The case of sea bass (Dicentrarchus labrax) cage farming in the Mediterranean. Aquaculture. 2016;462:47–55.
- 67. Christophersen G, Strand Ø. Effect of reduced salinity on the great scallop (Pecten maximus) spat at two rearing temperatures. Aquaculture. Aquaculture. 2003;215(1):79–92.
- 68. Duncan PF, Brand AR, Strand Ø, Foucher E. The European Scallop Fisheries for Pecten maximus, Aequipecten opercularis, Chlamys islandica, and Mimachlamys varia. 2016;40:781–858.
- 69. Tandler A, Anav FA, Choshniak I. The effect of salinity on growth rate, survival and swimbladder inflation in gilthead seabream, Sparus aurata, larvae. Aquaculture. Aquaculture. 1995;135(4):343–53.
- 70. Pilditch CA, Grant J, Bryan KR. Seston supply to sea scallops (Placopecten magellanicus) in suspended culture. Canadian Journal of Fisheries and Aquatic Sciences. 2001;58(2):241–53.
- 71. Newell RI. Ecosystem influences of natural and cultivated populations of suspension-feeding bivalve molluscs: a review. Journal of Shellfish Research. 2004;23(1):51–62.
- 72. Theodorou JA. Current and Future Technological Trends of European Seabass-Seabream Culture. Reviews in Fisheries Science. 2010;10(3–4):529–43.
- 73. Tacon AGJ, Metian M. Global overview on the use of fish meal and fish oil in industrially compounded aquafeeds: Trends and future prospects. Aquaculture. 2008;285(1–4):146–58.
- 74. Naylor RL, Hardy RW, Bureau DP, Chiu A, Elliott M, Farrell AP, et al. Feeding aquaculture in an era of finite resources. Proc Natl Acad Sci U S A. 2009;106(36):15103–10. pmid:19805247
- 75. Burbridge P, Hendrick V, Roth E, Rosenthal H. Social and economic policy issues relevant to marine aquaculture., 17(4), 194–206. Journal of Applied Ichthyology. 2001.
- 76. Broitman BR, Halpern BS, Gelcich S, Lardies MA, Vargas CA, Vásquez-Lavín F, et al. Dynamic Interactions among Boundaries and the Expansion of Sustainable Aquaculture. Frontiers in Marine Science. 2017;4.
- 77. Brown N. Flatfish Farming Systems in the Atlantic Region. Reviews in Fisheries Science. 2010;10(3–4):403–19.
- 78. Troell M, Jonell M, Henriksson PJG. Ocean space for seafood. Nat Ecol Evol. 2017;1(9):1224–5. pmid:29046558
- 79. Akinrotimi OA, Gabriel UU, Owhonda NK, Onunkwo DN, Opara JY, Anyanwu PE, et al. Formulating an environmentally friendly fish feed for sustainable aquaculture development in Nigeria. Agriculture journal. 2007;2(5):600–12.
- 80. Ridler NB, Hishamunda N. Promotion of sustainable commercial aquaculture in sub-Saharan Africa. Policy framework, Food & Agriculture Org. 2001;401.
- 81. Hishamunda N, Subasinghe RP. Aquaculture development in China: the role of public sector policies (No. 427). Food & Agriculture Org. 2003.
- 82. Cao L, Naylor R, Henriksson P, Leadbitter D, Metian M, Troell M, et al. China's aquaculture and the world's wild fisheries. Science. 2015;347(6218):133–5. pmid:25574011
- 83. Gómez-Requeni P, Mingarro M, Calduch-Giner JA, Médale F, Martin SAM, Houlihan DF, et al. Protein growth performance, amino acid utilisation and somatotropic axis responsiveness to fish meal replacement by plant protein sources in gilthead sea bream (Sparus aurata). Aquaculture. 2004;232(1–4):493–510.
- 84. Moe H, Dempster T, Sunde LM, Winther U, Fredheim A. Technological solutions and operational measures to prevent escapes of Atlantic cod (Gadus morhua) from sea cages. Aquaculture Research. 2007;38(1):91–9.
- 85. Gatlin DM, Barrows FT, Brown P, Dabrowski K, Gaylord TG, Hardy RW, et al. Expanding the utilization of sustainable plant products in aquafeeds: a review. Aquaculture Research. 2007;38(6):551–79.
- 86. Turchini GM, Ng WK, Tocher DRE. Fish oil replacement and alternative lipid sources in aquaculture feeds. CRC Press Chicago. 2010.
- 87. Martins CIM, Eding EH, Verdegem MCJ, Heinsbroek LTN, Schneider O, Blancheton JP, et al. New developments in recirculating aquaculture systems in Europe: A perspective on environmental sustainability. Aquacultural Engineering. 2010;43(3):83–93.
- 88. Tal Y, Schreier HJ, Sowers KR, Stubblefield JD, Place AR, Zohar Y. Environmentally sustainable land-based marine aquaculture. Aquaculture. 2009;286(1–2):28–35.
- 89. van Rijn J. Waste treatment in recirculating aquaculture systems. Aquacultural Engineering. 2013;53:49–56.
- 90. Halwart M, Soto D, Arthur JRE. Cage aquaculture: regional reviews and global overview. Food & Agriculture Org. 2007.
- 91. Callaway R, Shinn AP, Grenfell SE, Bron JE, Burnell G, Cook EJ, et al. Review of climate change impacts on marine aquaculture in the UK and Ireland. Aquatic Conservation: Marine and Freshwater Ecosystems. 2012;22(3):389–421.
- 92. Pickering TD, Ponia B, Hair CA, Southgate PC, Poloczanska ES, Patrona LD, et al. Vulnerability of aquaculture in the tropical Pacific to climate change. Vulnerability of Tropical Pacific Fisheries and Aquaculture to Climate Change. Noumea, New Caledonia: Secretariat of the Pacific Community. 2011:647–732.
- 93. Van Houtte A. Establishing legal, institutional and regulatory framework for aquaculture development and management. In Technichal Proceeding of the Conference on Aquaculture in the Third Millenium, Bangkok, Thailand. 2000:20–5.
- 94. Buschmann AH, Cabello F, Young K, Carvajal J, Varela DA, Henríquez L. Salmon aquaculture and coastal ecosystem health in Chile: Analysis of regulations, environmental impacts and bioremediation systems. Ocean & Coastal Management. 2009;52(5):243–9.
- 95. NMFS, NOAA, Commerce UDo. Implementation Plan for the Code of Conduct for Responsible Fisheries. http://www.nmfs.noaa.gov/ia/resources/publications/ccrf/nmfs_imp_plan.pdf.2012.
- 96. NSW DoPI. Australian Aquaculture Code of conduct. http://www.dpi.nsw.gov.au/__data/assets/pdf_file/0005/446765/Appendix-7-Australian-Aquaculture-Code-of-conduct.pdf. 2012.
- 97. Kristensen V. Code of conduct for European aquaculture. The producers' approach. Cahiers Options Méditerranéennes (CIHEAM). 2000.
- 98. Easterling W., Aggarwal P, Batima P, Brander K, Bruinsma J, Erda L, et al. Food, Fibre, and Forest Products 3. 2007.
- 99. Parry ML, Rosenzweig C, Iglesias A, Livermore M, Fischer G. Effects of climate change on global food production under SRES emissions and socio-economic scenarios. Global Environmental Change. 2004;14(1):53–67.
- 100. Merino G, Barange M, Blanchard JL, Harle J, Holmes R, Allen I, et al. Can marine fisheries and aquaculture meet fish demand from a growing human population in a changing climate? Global Environmental Change. 2012;22(4):795–806.
- 101. Evans W, Mathis JT, Ramsay J, Hetrick J. On the Frontline: Tracking Ocean Acidification in an Alaskan Shellfish Hatchery. PLoS One. 2015;10(7):e0130384. pmid:26131723
- 102. Ferreira JG, Hawkins AJS, Monteiro P, Moore H, Service M, Pascoe PL, et al. Integrated assessment of ecosystem-scale carrying capacity in shellfish growing areas. Aquaculture. 2008;275(1–4):138–51.
- 103. Clements JC, Chopin T. Ocean acidification and marine aquaculture in North America: potential impacts and mitigation strategies. Reviews in Aquaculture. 2016.