Global opportunities for mariculture development to promote human nutrition

Defining mariculture opportunity

For a country to tackle the nutritional deficiencies of its population through mariculture development, it should have three main characteristics, expressed herein through three scores that we compile for each nation in our analysis. First, the country should have a demonstrated need for the macro- and micronutrients that seafood can provide. As described above, seafood can be an efficient and important source of not just calories, but also protein, healthy fatty acids, zinc, vitamin A, and iron (Kawarazuka & Béné, 2011; Béné et al., 2015). On the other hand, countries that are well-nourished will not necessarily benefit (nutritionally) from adding more fish to the diet. We refer to a country’s relative deficiencies in these key nutrients as the country’s nutritional opportunity.

There should also be good evidence within the country of a cultural predisposition to seafood consumption. Clearly, increases in mariculture production will be most directly important for alleviating nutritional deficiencies if seafood accounts for a large proportion of a country’s diet. For this reason, we also include seafood reliance—calculated as the relative contribution of seafood to total diet—as a core enabling factor for mariculture opportunity.

Finally, a country’s mariculture production should be economically viable in order to sustainably provide a nutritional solution. Many combined mariculture/development projects fail to be sustainable because of a lack of scalability or long-term economic feasibility (Béné et al., 2016; Little, Newton & Beveridge, 2016). Hence, our third score for each country is a measure of this economic opportunity, constructed from each country’s current aquaculture production and seafood trade data, as well as proxies for the value of the seafood production sector and latent economic development potential.

Mariculture opportunity metrics

We compiled raw data for economic opportunity, nutritional opportunity, and seafood reliance from two publicly available databases (Fig. 1). We endeavored to limit our metrics to those that are directly relevant to the economic development of mariculture and the alleviation of nutritional deficiencies through seafood production. The resulting set of raw data includes five economic opportunity metrics, six nutritional opportunity metrics, and six seafood reliance metrics by country. To facilitate global comparisons, these 17 raw metrics were normalized and then combined into the three opportunity metrics and a final mariculture opportunity metric (Fig. 1).

Nutritional opportunity and seafood reliance scores were calculated using raw metrics from the Harvard GENuS database (Smith et al., 2016, https://dataverse.harvard.edu/dataverse/GENuS). GENuS models comprehensive country-specific diet and nutrient supply information by extrapolating from the Food and Agriculture Organization of the United Nations’ (FAO) food balance sheets, household surveys, and production data. GENuS estimates per capita nutrient consumption across hundreds of food categories. For the purposes of this study, we utilized data on the average daily per capita intake by country of five essential nutrients that can be obtained from seafood: protein, vitamin A, zinc, iron, and polyunsaturated fatty acids (PUFAs). Separately, we also collected each country’s average Dietary Energy Supply Adequacy, an FAO measure of the basic adequacy of total caloric intake relative to a sufficient diet (Eqs. (1) and (2)). These data—energy adequacy plus the daily per capita intake of five nutrients—comprise our six nutritional opportunity metrics. Together, these measures provide a synthesis of the average nutritional status of each country, specific to those nutrients that mariculture products can provide.

Our raw seafood reliance data were also drawn from the GENuS database. Because GENuS provides per capita nutrient intakes by food-group, we were able to sum per capita intakes from all FAO marine harvest categories (pelagic fish, demersal fish, other marine fish, crustaceans, and mollusks) to calculate total nutrient and calorie intakes obtained from seafood. We divided these seafood-specific intake values by total per capita intake values to calculate the percent of each nutrient obtained from seafood products. Six of these percentage values—for calories, protein, vitamin A, zinc, iron, and PUFAs—comprise our six seafood reliance metrics. Having both average nutritional status (nutritional opportunity score) and seafood reliance allows our scoring system to identify countries where increased mariculture production may have the greatest chance to directly address nutritional deficiencies, and where vulnerability to potential declines in wild-caught fisheries is highest.

The third dimension of mariculture opportunity is economic opportunity. Economic metrics were drawn from FishStatJ (http://www.fao.org/fishery/statistics/software/fishstatj/en), a freely available software used to access data from the Fisheries and Aquaculture division of FAO. FishStatJ provides panel data on fisheries and aquaculture production and trade by country, species, and commodity type. Selecting only the most recent year for which all metrics are available (2011), and excluding all commodity categories not for direct human consumption (e.g., fish meal or fish oil), these data were analyzed to produce the five economic opportunity metrics for each country: (1) production ratio, (2) trade balance in terms of quantity, (3) trade balance in terms of value, (4) GDP per capita, and (5) willingness to pay for seafood.

We define a country’s production ratio as its total aquaculture production divided by its total marine fisheries production in metric tons. Both production metrics were drawn directly from FAO reported data. This measure serves as a proxy for relative importance of two sectors that share infrastructure and markets. The logic is that countries with active fishing sectors should have both capital and management institutions that could also be functional to production and regulation of mariculture. The balance of fisheries and aquaculture production determines the opportunity for mariculture to utilize that shared infrastructure. The more skewed the production ratio is toward fisheries, the more potential there is to take advantage of these overlaps through the further development of a mariculture sector. While an indirect proxy for infrastructure, production ratio was chosen because of its generality across multiple types of potential mariculture production and its consistency across countries.

Two of our economic opportunity metrics measure trade balance in quantity and value. In our study, trade balance describes each country’s total volume or value of exports of seafood products (not just mariculture) divided by its imports. Trade balance measured in this way is a proxy for how a country balances supply and demand in the global seafood market. Trade imbalances reveal how countries compensate for their domestic seafood demand: a trade imbalance in which imports outweigh exports implies an opportunity to satisfy excess demand with augmented domestic mariculture production. Because seafood products vary so widely in their value relative to their volume, this trade balance signal could manifest in either metric, hence our inclusion of both quantity and value metrics.

The metric for GDP per capita is included as a proxy for latent economic opportunity. Countries with low per capita GDP have a need for economic development that may be partially pursued through mariculture. In this way, lower GDP per capita corresponds to higher economic opportunity scores in our analysis.

Willingness to pay for seafood, our final economic opportunity metric, is defined as a country’s total value of seafood imports divided by its GDP. This measure serves as a proxy for seafood value in each country. Although an imperfect metric, as it combines high-volume, low-value seafood with high-value niche products, willingness to pay still reflects overall expenditure on foreign seafood production. In our analysis, a higher willingness to pay corresponds to a greater opportunity to capture that willingness to pay with mariculture products produced domestically. In the calculation of opportunity scores in the next section, we use the reciprocal value of willingness to pay so that its ordering aligns with the other economic metrics (a lower value of the metric corresponds to a higher economic opportunity).

Together, these economic measures provide the essential information to describe a given country’s current mariculture production status relative to other nations. Furthermore, by utilizing FAO data, this set of economic metrics provides the ability to contrast countries while reducing potential sources of inconsistency and bias that might arise from using disparate sources, while at the same time being readily amenable to update as new data become available. Each individual metric provides one perspective on the enabling conditions for economic development of mariculture. Based on our economic opportunity metrics, a country with a high economic opportunity is one with existing seafood industry infrastructure, a seafood trade balance that could benefit from increased domestic production, a demonstrated value of seafood in the country, and a relatively low per capita GDP.

Calculation of mariculture opportunity scores

The 17 metrics were combined into three opportunity scores and one final mariculture opportunity score (Fig. 1). All raw metrics were normalized to a zero to one scale to allow comparison across categories of metrics. For each metric, X_i was scaled by dividing by its 80th percentile value across countries (Eq. (1)). (1)${\displaystyle X_i=\frac{X_{raw,i}}{P_{80}\left[X_{raw}\right]}.}$

This scaling was chosen to reduce the influence of large single-metric outliers on the ability to distinguish between nations. The choice had little effect on the final ranks of nations compared to dividing by the 90th percentile or simply the maximum value for each metric (see Table A2; final country ranks between three alternate scaling choices were significantly concordant; Kendall’s W = 0.925, p < 0.05).

The three scores–nutritional opportunity, seafood reliance, and economic opportunity–for each nation were determined by calculating a mean across the set of normalized metrics associated with that score (Eqs. (2)–(4), Fig. 1). No weighting was done in the calculation of aggregated scores because our emphasis is on countries’ relative positions. There was no definitive rationale for weighting any metric more heavily than any other, and doing so might unnecessarily complicate the interpretation of our scoring system and results. While we did not choose to use weighted scores, our methodology remains flexible to that extension.

Equations (2)–(5) describe score calculation. Country i’s economic opportunity score (Eq. (2)) was defined as the mean of its normalized metrics for production ratio, trade balance in value, trade balance in quantity, willingness to pay, and GDP per capita. The country’s nutritional opportunity score is the mean of its normalized intakes of protein, vitamin A, zinc, iron, and PUFAs, as well as its normalized FAO energy adequacy. Because the raw dietary nutritional supplies do not scale to zero–no country’s diet consists of zero calories–the set of nutritional opportunity scores were further rescaled by subtracting the minimum country score and dividing by the range across all nutritional opportunity scores. Finally, a country’s seafood reliance score is the mean of its normalized metrics for protein, vitamin A, zinc, iron, fatty acids, and calories derived from seafood. Each of the three metrics was ordered such that a higher score (closer to 1) corresponds to higher opportunity in that dimension, as described in the previous section. (2)${\displaystyle Eco{n}_i=1-Average\left(ProdRati{o}_i+Trade{Q}_i+Trade{V}_i+\frac{1}{{WTP}^i}+GDPp{c}_i\right).}$ (3)${\displaystyle Nutr{i}_i=1-Average\left(Protei{n}_i+Vit{A}_i+Zin{c}_i+Iro{n}_i+PUF{A}_i+Energ{y}_i\right).}$ (4)${\displaystyle Relianc{e}_i=Average\left(ProtSe{a}_i+VitASe{a}_i+ZincSe{a}_i+IronSe{a}_i+PUFASe{a}_i+CalSe{a}_i\right).}$ (5)${\displaystyle Opportunit{y}_i=Average\left(Eco{n}_i+Nutr{i}_i+Relianc{e}_i\right).}$

A final mariculture opportunity score for each country was calculated by averaging its economic opportunity, nutrition opportunity, and seafood reliance scores (Eq. (5)).

Non-coastal nations (N = 49) were excluded from the analysis before score calculation, because we were focused on the potential for local mariculture to address in-country nutritional needs. We likewise removed countries missing entire categories of data (e.g., missing all nutritional data). To avoid overly biasing our sample against data-limited countries, we retained countries with missing aquaculture production (N = 23), dietary energy supply adequacy (N = 10), and GDP (N = 4) values and simply omitted those individual variables from the countries’ score calculations. Sensitivity analyses on nations with complete data revealed that the effect on countries’ opportunity scores of single missing metrics was minimal, so gap-filling procedures (and their associated uncertainty) were not deemed necessary (Table A3). The final sample consists of 117 coastal nations. All raw and normalized metrics and scores for each country are available in the Supplemental Information.

Limitations of the analysis

The metrics described in the section above have two main advantages: (i) the data required is readily available through standard and credible data sources, and (ii) the calculation of each metric is straightforward and amenable to further refinement. However, the metrics also have two important limitations that warrant mention, and both are fully explored in the Appendix.

First, our metrics are constructed with cross-sectional data. This feature impedes our ability to explicitly examine past trends or predict future trajectories. But any cross-sectional approach suffers from the same limitations, and we are forced to rely on the assumption that the data reflects the actual realities of every country for that specific point in time.

Particular to our case, the limiting factor is that data for nutritional intakes (GeNUS) are only available for 2011. If a panel dataset on nutritional intake were to be available, it may prove beneficial to work with trends by country, rather than single observations for specific years. Unfortunately, to the best of our knowledge, these data are unavailable. We therefore use cross-sectional (2011) data in our main analysis, but in the Appendix we provide an extension to test the sensitivity of our results to the economic data, which are available in time series.

Secondly, it is possible that our seafood reliance score reflects preferences rather than pure reliance. For example, rich countries showing high seafood reliance could be merely a result of the average consumer’s preferences and may have little to do with the overall seafood availability or its price. If the index reflects preferences rather than pure reliance, the interpretation of the seafood reliance may be altered. This is an important issue if we are trying to establish the potential of mariculture in solving nutritional needs, but at the same time it is intrinsic to how these aggregate measures are collected across nations. We address this limitation in the Appendix by extending our main analysis and sorting countries in the sample based on their income level as reported by the World Bank (Fig. A3), as well as their geographical characteristics (Continental or Island; Fig. A2). Breaking down our results in this manner allows us identify the relative influence of income and country type on seafood reliance.