Algal research has been dominated by the use of marine biomass (mainly microalgae) as feedstock in the production of second-generation biofuels, albeit with limited economic success. A promising alternative strategy is the valorisation of seaweed (macroalgae), with the cascaded extraction of its high-value components, as well as lower-value components further downstream, under the ‘biorefinery concept’. The goal of this study was to assess the environmental performance of one such marine biorefinery situated in the UK.
Methods
Attributional life cycle assessment (LCA) was conducted on a hypothetical marine biorefinery coproducing fucoidan, laminarin, protein and alginate/cellulose packaging material (target product), from cultivated Saccharina latissima. The functional unit was the production of 1 kg of packaging material. A total of 6 scenarios were modelled, varying in coproduct management methodology (system expansion, mass allocation or economic allocation) and applied energy mix (standard or green energy). Sensitivity analysis was also conducted, evaluating the systems response to changes in allocation methodology; product market value; biomass composition and transport mode and distance. LCA calculations were performed using OpenLCA (version 1.10.3) software, with background processes modelled using the imported Ecoinvent 3.6 database. Environmental impacts were quantified under ReCiPe methodology at the midpoint level, from the ‘Heirarchist’ (H) perspective.
Results and discussion
The overall global warming impacts ranged from 1.2 to 4.52 kg CO2 eq/kg biopolymer, with the application of economic allocation; 3.58 to 7.06 kg CO2eq/kg with mass allocation and 14.19 to 41.52 kg CO2eq/kg with system expansion — the lower limit representing the instance where green electricity is used and the upper where standard electricity is employed. While implementing the green energy mix resulted in a 67% reduction in global warming impacts, it also incurred a 2–9 fold increase in overall impacts in the categories of terrestrial acidification, human non-carcinogenic toxicity, land-use and terrestrial ecotoxicity. Economic allocation resulted in burden shifting most favourable to the packaging material pathway.
Conclusions
This study demonstrates that the road to environmental optimisation in marine biorefineries is fraught with trade-offs. From the perspective of LCA — and by extension, the eco-design process that LCA is used to inform — when evaluating such product systems, it serves to strike a balance between performance across a broad spectrum of environmental impact categories, along with having consideration for the nature of energy systems incorporated and LCA methodological elements.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
1 Introduction
With the escalating impact of anthropogenic climate change and the limitations of contemporary production methods, transformative technologies capable of meeting the needs of the growing human population must be developed sustainably.
Macroalgae possess a number of advantageous attributes, positioning them as a superior bioresource to traditionally exploited terrestrial crops. As well as having gross primary productivity levels 3 times that of cropland (Duarte et al. 2005; Field et al. 2008; Hughes et al. 2012), seaweeds are a largely untapped source of high-value bioactive, pharmaceutical and nutraceutical compounds such as pigments (carotenoids and chlorophylls); sulphated polysaccharides; essential amino acids and polyunsaturated fatty acids (omega 3 and 6 fatty acids) and middle-to-lower value carbohydrates and hydrocolloids. Unlike agricultural resources, seaweed cultivation does not require the addition of freshwater, artificial fertilisers or arable land use, reducing deforestation and food-crop displacement. Species of macroalgae have also been shown to thrive in water bodies of varying salinity, as well as municipal wastewater environments. Additional benefits include their rapid release of O2 as a by-product of photosynthesis and nutrient (i.e. nitrophen and phosphorous) assimilation helping to combat the effects of phenomena such as ocean anoxia and marine eutrophication.
Anzeige
In light of the many ecosystem services offered by cultivated seaweed, it has been highlighted as a novel, scalable and multifaceted bioresource capable of satisfying the requirements of the growing bio-economy. Furthermore, the full-scale actualisation and ubiquitous practice of seaweed cultivation would directly aid the achievement of global sustainability targets, namely the United Nations Sustainable Development Goals of ‘zero hunger’ (SDG 2), ‘good health and well-being’ (SDG 3), ‘affordable and clean energy’ (SDG 7), ‘Climate action’ (SDG 13) and ‘life below water’ (SDG 14) (Duarte et al. 2021).
The potential benefits offered by seaweed to the global bio-economy are two-fold: (1) Directly as a ‘blue carbon’1 sink and (2) indirectly through the provision of product alternatives (biofuels, biofertilizers and bioplastics) substituting those which would otherwise require the input of fossil fuels, thus reducing carbon output. However, it is because of the transient manner in which carbon is sequestered that the role/value of seaweed within the Blue Economy2 has come into question. In fact, in a meta-analytical study, Krause-Jensen and Duarte (2016) estimate that of the net carbon gained by macroalgae globally (~ 1521 TgC/year), approximately 11% is sequestered with the remaining 89% released back into the environment, via grazing or decomposition. The actualisation of seaweed as a viable carbon mitigative strategy is contingent on its subsequent valorisation, following harvest.
One of the dominant objectives of macroalgal research has been the development of third-generation biofuels and bioproducts, not only for the substitution of fossil fuels but as alternatives to those derived from conventional lignocellulosic biomass — circumventing issues relating to arable land use and deforestation, freshwater and artificial fertiliser requirements. However, with the current state of technology, the large-scale lone derivation of bioproducts from seaweed has been demonstrated to neither be sustainable nor economically feasible (van den Burg et al. 2016; Soleymani and Rosentrater 2017) — though the coproduction of seaweed-based bulk products with other marketable commodities under a biorefinery concept could reduce both the environmental impact and economic cost of the bulk products.
Biorefineries are multistep process systems deriving multiple marketable bio-based products from renewable feedstocks, akin to petroleum refineries where fossil-based products have classically been derived (Cherubini 2010). Within a biorefinery, biomass may be valorised through mechanical/physical, biochemical/catalytic, chemical and thermochemical conversion. The ‘marine biorefining’ concept has been designated a technology readiness level (TRL) of 5–6, below ‘conventional’ starch, sugar crop and wood biorefineries (TRL 9); ‘whole crop’ (TRL 7–8); ‘oleochemical’ (TRL 7–9) and ‘lignocellulosic feedstock biorefineries (TRL 6–8) (de Jong and Jungmeier 2015).
Anzeige
Lifecycle assessment (LCA) is a methodological framework used to systematically evaluate the potential environmental performance of a product, and its lifecycle, spanning the stages of raw material acquisition, through to product use and end-of-life. To date, only a handful of LCAs have been carried out on biorefinery systems incorporating high-value algal products (Langlois et al. 2012; Pérez-López et al. 2014; Helmes et al. 2018; Sadhukhan et al. 2019; Parsons et al. 2019; Zhang et al. 2021; Zhang and Thomsen 2021). Parsons et al. (2019) conducted an LCA, (accompanied by in-depth sensitivity and uncertainty analysis) evaluating the sustainability of a biorefinery coproducing single-cell oils (substitutes for oils in the exotic fats market), 2-phenylethanol (for use in perfumery) and a proteinous yeast extract (for animal feed). Primarily targeting the pharmaceuticals market, Pérez-López et al. (2014) evaluated a series of biorefining schemes producing fucoxanthin, sodium alginate and phenolic compounds (antioxidants), from wild-harvested Sargassum muticum. Zhang et al. (2021) recently performed an ex ante LCA assessing the environmental hotspots in two cascading biorefining systems valorising South African Ecklonia maxima, and producing laminarin, fucoidan and alginate. The evaluated product systems employed novel solvent-free ‘sub-critical water’ and ‘hot water’ methods for laminarin and fucoidan extraction. While the overall water footprint was halved compared to commercially produced alginate, the carbon footprints of the marine biorefining systems were found to be twice as high — a direct consequence of the coal-dominated South African energy mix.
Building on this work, in this study, an attributional LCA of a biorefinery coproducing fucoidan, laminarin, protein and an alginate/cellulose composite packaging material (target product) is presented. The study investigated some key areas of sensitivity including the influence of allocation methodology (system expansion, economic- and mass allocation), energy inputs (UK standard versus green electricity and heating) and various other parameters on the overall environmental footprint of a theoretical/hypothetical pilot-scale seaweed biorefinery stationed in the UK.
2 Methods
2.1 Goal and scope definition
The goal of this attributional LCA study was to assess the potential environmental performance of a pilot-scale seaweed biorefining product system coproducing bioactives, protein and polymeric packaging tray material (i.e. the target product), from brown seaweed (Saccharina latissima) cultivated and bioprocessed in the UK. In addition to identifying hotspots within the modelled product system life cycles, this study also explored the inventory- and methodology-level sensitivities associated with the proposed scheme, reporting LCIA results within ranges of uncertainty. The purpose of this LCA was therefore to inform the process design of seaweed biorefining systems. The LCA was guided by the ISO Standards for LCA, ISO 14040 and ISO 14044.
The assessed product system comprises 9 core foreground processes: seaweed cultivation, water extraction, acid extraction, proteolysis, packaging material production, filtration, drying (SD1–3) and wastewater treatment (Fig. 1). All product system scenarios were modelled in reference to a functional unit defined as the material and energy inputs and outputs associated with the production of 1 kg of raw polymeric packaging material from cultivated Saccharina latissima. In this study, the biopolymeric packaging material was assumed to be the target product; therefore, the results of analyses included in this paper focus on the packaging material production route. The system boundary stops short of further product formation/valorisation (i.e. ‘forming and moulding’ processes that typically follow raw product derivation such as thermoforming, calendaring, extrusion, injection moulding) distribution, use and EOL processes, giving rise to the study’s cradle-to-gate categorisation. In scenarios where system expansion was employed, the avoided impacts associated with the substitution of commercially produced PLA and soybean protein by the biorefinery-derived products were also considered. No cut-off criteria were applied to the foreground system under evaluation. All modelling and LCA calculations were performed with OpenLCA software (version 1.10.3) utilising the datasets available on the imported Ecoinvent (version 3.6) life cycle inventory database. All graphical figures presented were generated using Python-based data visualisation libraries matplotlib (Shopbell et al. 2005) and seaborn (Waskom 2021).
Fig. 1
A system boundary diagram for the modelled seaweed biorefining product system. The core modelled foreground processes include seaweed cultivation, mechanical pre-treatment, water extraction, acid extraction, protein extraction, filtration, packaging production, drying (SD1–3) and wastewater treatment. Packaging polymer is coproduced with protein, laminarin and fucoidan. To the right are the commercial products that would theoretically be substituted by the derived products, i.e. soybean meal (used as a protein source) by our seaweed protein, and polylactic acid (used in biodegradable food containers) by the seaweed packaging polymer. W, wet weight and D, dry weight
×
2.2 System description and lifecycle inventory
Based on the process scheme presented in Fig. 1, 6 scenarios were modelled, with varied electricity mixes (UK standard and UK green electricity mix), heating inputs (i.e. produced by a boiler or burning biomass) and coproduct management methodology (i.e. system expansion, economic allocation and mass allocation):
SS — foreground processes utilising a standard electricity mix, heating provided by a boiler using natural gas and system expansion applied
SG — foreground processes utilising a green electricity mix, heating provided by burning biomass and system expansion applied
ES — foreground processes utilising a standard electricity mix, heating provided by a boiler using natural gas and environmental impacts allocated to each coproduct on an economic basis
EG — foreground processes utilising a green electricity mix, heating provided by burning biomass and environmental impacts allocated to each coproduct on an economic basis
MS — foreground processes utilising a standard electricity mix, heating provided by a boiler using natural gas and environmental impacts allocated to each coproduct on a mass basis
MG — Foreground processes utilising a green electricity mix, heating provided by burning biomass and environmental impacts allocated to each coproduct on a mass basis
In the modelled process scheme, the lifecycle inventory (LCI) data used to model the outlined scenarios is presented in Table 1. The composition of the standard and green electricity mixes modelled is presented in Tables 2 and 3.
Table 1
Lifecycle inventory data showing the material and energy inputs required to produce 1 kg of packaging material from cultivated Saccharina latissima, in the system expansion (SS and SG), mass allocation (MA-PM and MG-PM) and economic allocation (EA-PM and EG-PM) scenarios
Quantity of input per kg packaging material produced
Material/energy inputs
unit
System expansion (SS and SG)
Mass allocation (MS-PM and MG-PM)
Economic allocation (ES-PM and EG-PM)
Ecoinvent 3.6 process
Seaweed cultivation
Ammonium nitrate
kg
1.93E−04
1.51E−04
9.87E−06
Market for nitrogen fertiliser, as N | nitrogen fertiliser, as N | cut off, U-GLO
aThe standard and green energy scenarios respectively use a custom standard and green UK electricity mix, composed of ecoinvent process inputs — see Tables 2 and 3
bThe standard and green energy scenarios respectively use the ecoinvent heating process inputs ‘heat production, natural gas, at boiler modulating > 100 kW | heat, district or industrial, natural gas | cut off, U-Europe without Switzerland’ or ‘heat production, hardwood chips from forest, at furnace 1000kW | heat, district or industrial, other than natural gas | cut off, U-CH’
Table 2
Composition of the standard electricity mix modelled using data sourced from the Digest of United Kingdom Energy Statistics (DUKES), representative of the year 2020 (BEIS 2020)
Electricity source
Share (%)
Ecoinvent 3.6 process
Imports
Imports: France
2.6
Electricity, high voltage, import from FR | electricity, high voltage | cut off, U-GB
Imports: Netherlands
1.2
Electricity, high voltage, import from NL | electricity, high voltage | cut off, U-GB
Imports: Ireland
0.05
Electricity, high voltage, import from IE | electricity, high voltage | cut off, U-GB
Imports: Belgium
1.4
Market for electricity, high voltage | electricity, high voltage | cut off, U-BE
Non-renewables
Nuclear
18.1
Electricity production, nuclear, pressure water reactor | electricity, high voltage | cut off, U-GB
Petroleum
0.6
Electricity production, oil | electricity, high voltage | cut off, U-GB
Natural Gas
34.2
Electricity production, natural gas, combined cycle power plant | electricity, high voltage | cut off, U-GB
Coal
2.5
Electricity production, hard coal | electricity, high voltage | cut off, U-GB
Renewables
Wind onshore
6.5
Electricity production, wind, <1MW turbine, onshore | electricity, high voltage | cut off, U-GB
Wind offshore
7.6
Electricity production, wind, 1–3 MW turbine, offshore | electricity, high voltage | cut off, U-GB
Solar photovoltaics
2.5
Electricity production, photovoltaic, 570kwp open ground installation, multi-Si | electricity, low voltage | cut off, U-GB
Hydro-electricity
1.3
Electricity production, hydro, run-of-river | electricity, high voltage | cut off, U-GB
Landfill + sewage gas
3.3
Heat and power co-generation, biogas, gas engine | electricity, high voltage | cut off, U-GB
Biomass
12.8
Heat and power co-generation, wood chips, 6667 kw | electricity, high voltage | cut off, U-CH
Municipal waste combustion
5.5
Electricity, from municipal waste incineration to generic market for electricity, medium voltage | electricity, medium voltage | cut off, U-GB
Table 3
Composition of the green electricity mix, modelled using data sourced from the UK-based green energy supplier Ecotricity representative of the year 2020/2021 (Ecotricity 2021)
Electricity source
Share (%)
Ecoinvent 3.6 process
Wind (onshore)
14.6
Electricity production, wind, < 1MW turbine, onshore | electricity, high voltage | cut off, U-GB
Wind (offshore)
81.6
Electricity production, wind, 1–3 MW turbine, offshore | electricity, high voltage | cut off, U-GB
Solar photovoltaics
2.4
Electricity production, photovoltaic, 570 kwp open ground installation, multi-Si | electricity, low voltage | cut off, U-GB
Hydro-electricity
1.4
Electricity production, hydro, run-of-river | electricity, high voltage | cut off, U-GB
2.2.1 Seaweed cultivation phase
LCI data used to model the seaweed cultivation phase is summarised in Table 1. The majority of the data used to model the kelp cultivation process was sourced from the seaweed bioprocessing LCA study conducted by Langlois et al. (2012), in which Saccharina latissima plantlets were incubated in indoor artificial ponds during the nursery phase, before being deployed at sea on longline cultivation apparatus. Material inputs constituted a range of minerals and fertilisers promoting plantlet growth, namely ammonium nitrate, sodium phosphate, EDTA, anhydrous boric acid and iron(III) chloride. Electrical energy inputs totalled 341.5 Wh/kg dry seaweed, powering fluorescent and UV lamps, circulation and sand pumps and sparger for water treatment and plantlet cultivation in the nursery. Alvarado-Morales et al. (2013) estimated the barge necessary for longline deployment, and removal requires a total of 30 L of diesel and 30 L of petrol to power a skiff used to monitor the seaweed growth during its 6-month cultivation. It was assumed that the cultivation site was in close proximity to the biorefinery, hence the absence of a vehicle/lorry (and associated fuel required) typically required for the transport of harvested seaweed to the plant. The LCA model excluded foreground infrastructural inputs, as such details regarding longline-farming apparatus; cultivation ropes; maintenance of nursery machinery and the assembly and disassembly of the nursery itself were omitted from the cultivation inventory.
2.2.2 Solid-stream bioprocessing phases
LCI used to model the solid stream bioprocessing processes (i.e. mechanical pre-treatment, water extraction, acid extraction, protein extraction and packaging material production) are summarised in Table 1. It was assumed that the harvested macroalgae were processed at a rate of 5 wet tonnes per hour.
Following mechanical pre-treatment, the biomass undergoes ‘water extraction’ and ‘acid extraction’ (Zhang and Thomsen 2021). During these steps, portions of water or hydrochloric acid are added to the biomass, before it undergoes mechanical pressing — the resultant press water streams, containing water-soluble biomolecules fucoidan and laminarin, flowing through to the liquid extraction portion of the process (see Sect. 2.2.3).
The bulk biomass moves on to the process of ultrasonic-assisted extraction where further addition of water is made. The implosion of bubbles generated by acoustic cavitation essentially creates shockwaves that disrupt the cellulose/alginic acid algal cell walls, releasing cell contents — namely proteins (Ashokkumar et al. 2008; Veillet et al. 2010; Ummat et al. 2020). Following the addition of proteolytic enzymes (proteases), the peptide chains are hydrolysed into activated peptide fragments, giving rise to their bioactive functionality (Admassu et al. 2018). The mixture then undergoes phase separation via centrifugation and pressing with the resultant protein-containing press water stream progressing to drying (SD3) and the residual seaweed biomass (comprising cellulose and alginic acid) stream moving to the final packaging production phase. Twenty per cent sodium carbonate is then added to the residual seaweed biomass stream, neutralising the alginic acid. The final stage comprises a belt-drying step resulting in the formation of 90% DM alginate/cellulose (2:1) polymeric resin (or ‘packaging material’).
2.2.3 Liquid-stream bioprocessing phases
LCI data used to model the liquid stream bioprocessing processes (i.e. filtration and SD1–3) are summarised in Table 1. The ‘water extraction’ and ‘acid extraction’ press-water streams are combined, diluted and filtered through a network of different-sized membranes, resulting in fractions of fucoidan and laminarin separated based on their differing molecular weights (Sterner and Gröndahl 2021, Zhang et al. 2021, Zhang and Thomsen 2021). Excess water is driven off via drying (SD1–3) producing purified products.
2.3 Description of coproduct management procedures
Biorefineries are by definition multi-output systems. In LCA, it is required that the environmental impacts incurred by a multi-output process (i.e. unit processes deriving more than one product flow) are systematically partitioned or ‘allocated’ amongst coproducts. Regarding coproduct management procedures used in LCA, the International Organization for Standardization (ISO) (2018) guidelines dictate the following order of favourability: system expansion > > physical allocation (e.g. mass or energy allocation) > alternative allocation methods (e.g. economic allocation). Methodological choices such as coproduct management have been shown to be highly consequential in the environmental burdens assigned to the coproducts in seaweed bioprocessing systems — a primary methodological source of variability. In light of this, system expansion (by substitution), mass allocation and economic allocation were applied in the various scenarios modelled, to compare their effect on the overall burden distribution.
The crude protein content of soybean meal sourced from the top global producers (Argentina, Brazil, the USA and India) has been determined as 464g/kg (or per kg dry) soybean meal (Ibáñez et al. 2020). Satisfying the ISO recommendations for system expansion, commercially produced soybean meal protein was subject to a 2:1 substitution by the biorefinery-derived protein product. Regarding the alginate/cellulose packaging material production route, a 1:1 mass substitution was applied with commercially produced PLA. Due to the lack of any relevant substitutable products represented in the Ecoinvent database, system expansion in relation to bioactives fucoidan and laminarin was not carried out.
In the scenarios using mass allocation (i.e. MS and MG), environmental emissions were partitioned based on the mass allocation coefficient \({C}_{m}\), calculated per coproduct present in the process stream, which is a ratio between the mass of the ‘subject coproduct’ and all coproducts:
where \({m}_{coproduct 1}\) is the mass of the subject coproduct, and \({m}_{coproduct 2\to n}\) is the mass of all other relevant coproducts.
Similarly, in the scenarios allocating emissions on an economic basis (i.e. ES and EG), the economic allocation coefficient \({C}_{e}\) was calculated as follows:
where ($ · m) is the assigned price ($) per unit of the biomass component, multiplied by the mass (m) of material produced in the process. Product searches were conducted on the e-commerce platform and bulk supplier Alibaba, between 6 Jul 2021 and 7 Jul 2021, with a preference for products from verified suppliers. The average price of biodegradable PLA (mixed with PBAT and corn-starch) (2.39 GBP/kg powder) was used as a proxy for seaweed biopolymer (due to seaweed biopolymers not being commercially available). Protein coproduced in the studied system was intended for use as an animal feed protein supplement, as such the average price of soybean protein (2.37 GBP/kg powder), the traditionally used protein source in animal feed, was used as a proxy. Fucoidan and laminarin are bioactive compounds that would be intended for human consumption so only high-purity products (> 90%) were used to derive average prices of £203.23 and £426.22/kg powdered fucoidan and laminarin. Equations 1 and 2 were adapted from that used by Chen et al. (2010).
2.4 Lifecycle impact assessment
Lifecycle impact assessment (LCIA) is the penultimate stage of an LCA, where the elementary flows associated with the material and energy inputs and outputs catalogued in the LCI are assigned to impact categories and category indicators calculated (‘characterisation’). Environmental impacts were evaluated at the midpoint level of the ReCiPe 2016 impact assessment method preloaded onto the OpenLCA (version 1.10.3) software (Huijbregts et al. 2017). The LCIA was conducted from the Hierarchist (H) perspective, considering potential environmental impacts over a 100-year timeframe. Emissions related to the following impacts categories were quantified: global warming (kg CO2 to air), terrestrial acidification (kg SO2 to air), freshwater eutrophication (kg P to freshwater), human toxicity (cancer) (kg 1,4-DCB to urban air), human toxicity (non-cancer) (kg 1,4-DCB to urban air), terrestrial ecotoxicity (kg 1,4-DCB to industrial soil), freshwater ecotoxicity (kg 1,4-DCB to freshwater), marine ecotoxicity (kg 1,4-DCB to marine water), water use (m3 water consumed), mineral resource scarcity (kg Cu), fossil resource scarcity (kg oil) and land use (m2a crop eq). All impact values were reported in reference to the production of 1 kg of packaging material (the product system functional unit).
2.5 Key assumptions and limitations
The following is a list of the various limitations of this LCA study, as well as the assumptions it was based on:
This LCA study was commissioned at the early stages of product and product-system development — a juncture where the properties of the proposed seaweed biopolymer were little understood. In the absence of material characterisation data, a 1:1 equivalence was drawn between the proposed seaweed biopolymer and PLA which it is assumed to substitute out in the system expansion scenarios (SS & SG).
A decision was made to exclude inputs relating to the construction/maintenance/decommissioning of infrastructure (or ‘capital goods’) from the lifecycle inventory of the modelled foreground unit processes. Within the context of this study, examples of infrastructural inputs would be buildings containing the plantlet hatchery and biorefinery; machinery and tools used in the cultivation, maintenance and harvest of macroalgal biomass, as well as that in the biorefinery, and materials used to form the offshore cultivation apparatus. The focus was solely placed on the consumable product system inputs (i.e. chemicals/materials and energy), as environmental optimisations based on such inputs have been deemed more impactful and plausible than strategies involving improvements to infrastructural inputs, which typically have high investment requirements (Silva et al. 2018).
In the protein extraction phase, enzymes (proteases) are required to catalyse the hydrolysis of the algal peptide chains into activated peptide fragments (giving rise to their bioactive functionality). However, details regarding the type, quantity and life-span/recyclability of proteases used during this process were not specified. Moreover, enzymes are not immediately consumed in the chemical reaction and are largely recyclable, hence have been categorised as ‘infrastructure’ in this study, and omitted from the inventory data.
The final stage of the packaging material pathway comprises a belt-drying step resulting in the formation of alginate/cellulose (2:1) polymeric material. Details regarding ‘forming and moulding’ processes that typically follow polymeric resin production (i.e. thermoforming, calendaring, extrusion, injection moulding) were not specified, nor was there mention of the addition of plasticiser (an additive used to improve polymer flexibility). As such, the ‘gate’ portion of the system boundary was positioned after the production of cellulose: alginate resin.
As with all real-world manufacturing processes, mass losses throughout the seaweed biorefining process (e.g. via the adherence of product to machining equipment) would be inevitable. However, the mass balance on which this LCA model was based does not account for mass losses — due to the studied process largely being theoretical in nature. The final product yields were assumed equivalent to the content of fucoidan, laminarin, protein and carbohydrates (alginate and cellulose) in the feedstock biomass — unrealistically implying (1) a 100% coproduct yield and (2) 100% coproduct purity. It is likely that the real-life execution of the proposed scheme would result in actual coproduct yields, significantly different from the theoretical yield, ultimately influencing the allocation of environmental impacts between the coproducts.
This LCA model does not factor in biogenic exchanges associated with seaweed cultivation or the environmental impacts associated with land-use change.
3 Results and discussion
3.1 Relative contribution analysis
The relative impact contributions made by each of the foreground processes in the baseline standard and green energy system-expansion scenarios (SS and SG) are presented in the form of heatmaps in Fig. 2. The following observations and discussion points regarding process hotspots are made primarily in reference to Fig. 2, with additional insights (i.e. those relating to very specific activities or inputs within the modelled foreground processes) drawn from supplementary information (Fig. S1–5).
Fig. 2
Heatmap showing the relative contribution made by each of the foreground processes modelled in the standard and green energy substitution scenarios (SS and SG) per functional unit, across all ReCiPe (H) midpoint impact categories. Positive values represent harmful emissions to the environment, whilst negative values represent activities providing an environmental benefit/service. SWC, seaweed cultivation; M, mechanical pre-treatment; WW, water extraction; AW, acid extraction; P, protein extraction; PP, packaging production; DF, filtration and SD1–3, drying
×
Commercially and legislatively, GHG emissions tend to be the most widely adopted indicator of environmental health. In the standard energy scenario (SS), the greatest foreground contribution to global warming emissions was made by filtration (24.2%) — the majority (98.8%) ascribed to the treatment of wastewater expelled following the filtration of laminarin and fucoidan and the remainder to electricity. This was closely followed by a 24.3% contribution by ‘water extraction’ — the bulk 97% generated from the heating required to maintain process water at the optimal temperature (40 \(^\circ{\rm C}\)) and 2.67% by the electricity input, and then ‘drying 3’ (20.4%) — dominated by heating (97.1%) and then electrical (0.39%) energy inputs required for the sustained heating and evaporation of water during the drying of the protein product. From the LCI (in Table 1), it can be seen that the volume of liquid added and heated in the ‘water extraction’ phase exceeds that of the ‘acid extraction’ phase, justifying its notably lower GHG emission contribution of 12.3% (Fig. 2). Similarly, the volume of water vaporised in ‘drying 3’ is ~ 5 times that removed in ‘drying 1’ and ‘drying 2’, and this is reflected in their contrast in hues (Fig. 2). The lesser overall impact contributions came from the remaining core processes packaging production (8.3%), cultivation (3.0%), proteolysis (2.7%) and mechanical pre-treatment (0.3%).
In the green energy (SG) scenario, the filtration phase becomes an even more important contributor to total global warming impacts, accounting for 59.8% of CO2 eq emissions — 99.8% of which emanated from the treatment of waste process water. The second highest contribution came from the acid extraction phase (15.7%), the bulk of this (89.6%) attributed to HCl production, and minorly from heating (9.4%).
A markedly different but collective trend was observed amongst the other assessed impact categories. As indicated in Fig. 2, the majority of impacts generated in the SS scenario (43.8–68.6%) were concentrated in the ‘acid extraction’ phase, with the exception of the global warming, fossil resource scarcity and water consumption impact categories. Specifically, the HCl added during the acid extraction phase — aiding the extraction of laminarin from the bulk seaweed biomass — accounted for 89.4–97.8% of the impacts made and 45.78–94.13% of global warming, fossil resource scarcity and water consumption impacts. Future optimised iterations of this product system could explore the use of alternative HCl (or ‘chemical-‘)-free laminarin extraction techniques (e.g. enzyme-assisted, microwave-assisted, ultrasound-assisted, and pressurized solvent extraction) in a bid to reduce the overall emission footprint — although such methods may be comparatively energy intensive.
Amongst the wastewater-producing foreground processes, it is during the ‘filtration’ phase that the greatest quantity of wastewater (66.9kg/FU) is expelled from the product system — with comparatively small quantities produced in ‘drying 1’ (1.8 kg/FU), ‘drying 2’ (2.3 kg/FU), ‘drying 3’ (9.7 kg/FU) and ‘packing production’ (3.8 kg/FU). This is reflected by the dominance of the filtration (DF) phase in the water consumption and global warming categories, in both the standard (SS) and green energy (SG) scenarios.
3.1.1 Impact of green energy mix on alternative metrics
As expected, the transition from the fossil-fuel-based energy mix applied in the standard energy scenario to the renewable-resource–derived energy mix applied in scenario SG results in an overall 67% reduction in potential global warming emissions, as well as an 86% reduction in fossil resource scarcity related impacts (Fig. S3). However, contrary to this trend, in 9 of 13 of the impact categories assessed, the emissions generated by the system using green energy (SG) exceeded that using standard energy (SS), with the most pronounced (i.e. 2–9 fold) disparities occurring in the terrestrial acidification, human non-carcinogenic toxicity, land use and terrestrial ecotoxicity categories. From Fig. S2, it can be seen that this phenomenon was principally attributed to the upstream impacts associated with the biomass-based heating inputs (18.8–97.4%), and this is also reflected in the illumination (yellowing) of the tiles representing processes utilising the most heat (i.e. water extraction, acid extraction, packaging production and drying 3) in the green energy scenario (Fig. 2).
Further evaluation of these results and the background processes assumed in this energy scenario reveals that per kWh of heat generated, the biomass furnace (in the green energy scenarios) incurred a land-use footprint more than 10 times that of the modulating boiler utilised in the standard energy scenarios, i.e. \(4.4\times {10}^{-4}\) versus \(3.4\times {10}^{-5}\) m2a crop eq. Specifically, 82% of land-use impacts were attributed to land occupation related to transport infrastructure, as well as 11% from the land occupation required to farm palm fruit — palm oil being the main constituent in vegetable oil used to lubricate the power-saws used in the acquisition of woodchip from forests. The footprint associated with the biomass furnace was the primary contributor to the 4-fold greater overall land-use impacts observed in the green energy scenario, relative to the scenario using the standard energy mix — presented in Fig. S3.
Concerning terrestrial ecotoxicity, impacts incurred during heat generation via biomass furnace were 31 times greater than that via modulating boiler, i.e. \(5.14\times {10}^{-1}\) kg 1,4-DCB versus \(1.64\times {10}^{-2}\) kg 1,4-DCB, specifically as a result of zinc, cadmium, lead, thallium, barium and mercury emissions stemming from the treatment of wood ash following biomass combustion and wastes derived from copper and lignite mining operations. Again, the footprint associated with the biomass furnace was the dominant contributor to the 4-fold greater overall terrestrial ecotoxicity impacts observed in the green energy scenario, relative to the standard energy scenario — presented in Fig. S3.
These results support some of the identified challenges associated with renewable energy technologies, namely the increases in spatial footprint as a result of land-use change and the intensification of metal mining activities, necessary for infrastructure production/installation and feedstock acquisition, ultimately exacerbating biodiversity loss (Sonter et al. 2020; Spillias et al. 2020).
3.1.2 Comparison with alternative packaging products
Given their importance from a policy perspective, GHG emissions tend to have the most focus when evaluating sustainability and environmental health — popularity reflected in the almost universal consideration of the global warming and climate change impacts in LCA, irrespective of the nature of the technology studied. The overall global warming impact values obtained in the standard energy scenarios using an economic allocation (ES), a mass allocation (MS) and system expansion (SS) were 4.52 kg CO2 eq., 7.06 kg CO2 eq and 41.52 kg CO2 eq. As expected, the green energy scenarios with an economic allocation (EG), a mass allocation (MG) and system expansion (SG) incurred significantly lower overall global warming impacts of 1.25 kg CO2 eq., 3.58 kg CO2 eq and 14.19 kg CO2 eq. The full spectrum of LCIA results, across all 18 impact categories available via the ReCiPe 2016 framework, is provided in the supplementary information (Table S1).
In order to understand the relevance of these results within a commercial context, they were plotted against that of alternative (fossil- and bio-based) polymers, reported in literature (Fig. 3) — references and disaggregated literature values are provided in the supplementary information (Table S2). Note that the results of the system expansion scenarios have been omitted from the figure.
Fig. 3
Comparison of the global warming (kg CO2 eq) impacts incurred across the mass and economic allocation scenarios (i.e. MS, MG, ES and EG) with other available (fossil and renewable) packaging material alternatives. Data was obtained from LCAs conducted from the cradle-to-gate, specifically, assessing the global warming emissions arising from raw material acquisition through to polymeric resin production phases. All values are based on the production of 1 kg of polymeric material and are collated and presented as averages. Error bars represent 1 standard deviation from the mean. MS, mass allocation with standard energy mix; ES, economic allocation with standard energy mix; MS, mass allocation with green energy mix; EG, economic allocation with green energy mix; PLA, polylactic acid; PET, Polyethylene terephthalate; PS, polystyrene; PP, polypropylene; PHA, polyhydroxyalkanoate; PVC, polyvinyl chloride; LDPE/HDPE, low-/high-density polyethylene and TPS, thermoplastic starch
×
The green energy scenario using an economic allocation (EG) proved the most competitive, demonstrating environmental performance comparable to PHA (0.2 \(\pm\) 2.77 kg CO2 eq.), PLA (1.87 \(\pm\) 0.79 kg CO2 eq.), HDPE (0.98 \(\pm\) 1.73 kg CO2 eq.) and PP (1.07 \(\pm\) 1.65 kg CO2 eq.). It also comfortably undercuts the emission footprints for petrochemical polymers LDPE (1.94 \(\pm\) 0.13 kg CO2 eq), PET (2.92 \(\pm\) 0.49 kg CO2 eq), PS (3.23 \(\pm\) 0.51 kg CO2 eq), PVC (1.94 \(\pm\) 0.05 kg CO2 eq) and renewable TPS (2 kg CO2 eq). Scenario using a mass allocation (MG) afforded the second-best emission profile; however, it was comparable only to PS.
Assuming that system boundaries remain fixed, two effective methods of reducing the potential global warming impacts in the 6 scenarios modelled would be the inclusion of (1) biogenic carbon, by way of carbon sequestration during the seaweed cultivation phase (Seghetta et al. 2017; Thomas et al. 2020), and (2) land-use change benefits that arise from the location of the cultivation site offshore — the latter of which is yet to be featured/explored within the context of marine biorefining LCA literature. A notable instance where biogenic carbon and land-use change crediting has been successfully employed within the commercial context was in the LCA conducted on the Brazilian petrochemical company Braskem’s sugarcane-derived HDPE (I’m green™ biobased PE), where the two collectively amounted to an emission saving of − 4.24 kg CO2 eq/kg polymer resin, significantly contributing to the overall net climate change impact of − 3.09 kg CO2 eq/kg polymer resin (Braskem 2016).
The validity of inter/intra-LCA study product comparisons, such as that attempted in Fig. 3, is contingent on adequate homogeneity between LCA methodological elements, and the achievement of full functional equivalence between the products being compared — a prerequisite of the latter being a well-defined functional unit (Furberg et al. 2021).
At the methodological level, even within the constraints of the ISO 14040 and 14044 and ILCD frameworks, LCA practitioners are afforded a considerable degree of freedom, allowing for flexibility with regard to the selection of system modelling approach (attributional, consequential, ex-ante); impact assessment methodology (ReCiPe, IMPACT 2002+, TRACI, CML etc.); system boundary definition (spatiotemporal coverage, cut-off criteria, ‘cradle-to-gate/grave/cradle’, ‘gate-to-gate/grave’ etc.); coproduct management (system expansion or physical/economic allocation); databases used to provide secondary data (Ecoinvent, Agri-footprint, GaBi etc.) and EoL allocation procedure (cut-off method, closed-loop method, substitution method etc.) (Ekvall and Tillman 1997; Nicholson et al. 2009; Pauer et al. 2019). Moreover, environmental modelling frameworks specific to packaging material production, such as the Global Protocol of Packaging Sustainability (The Consumer Goods Forum 2011), have offered little extra in the way of procedural standardisation necessary for cross-study comparability.
This LCA study was conducted at the early stages of the product/system development — in the absence of comprehensive seaweed biopolymer characterisation data. As a consequence, this study suffered poor product function and functional unit definition, with the functional unit simply defined as ‘1 kg of packaging tray material’. An optimal (i.e. ‘performance’-based) functional unit definition would require an understanding/specification of (1) the chemical/mechanical/physical properties of the proposed alginate/cellulose biopolymer; (2) the intended product utility, and hence the required dimension/performance characteristics and (3) obligatory features necessary for legal compliance — e.g. adherence to Directives (EU) 2018/851/EC (packaging waste disposal) and 2018/852/EC (packaging requirements) (European Commission 2018a, b).
Figure 3 assumes functional equivalence based on a mass of 1 kg of each polymer which will likely not be the case. This, combined with a lack of methodological uniformity between the compared studies, and poor system functional unit definition will ultimately influence the relative global warming impact results. As such, the comparisons presented in Fig. 3 are indicative at best and should be used as a rough guide only to inform the early stages of eco-design.
3.2 Sensitivity analysis
3.2.1 Coproduct management methodology
Fig. 4 provides a visual comparison of the way that environmental impacts are distributed between each of the coproduct (packaging material, fucoidan, laminarin and protein) pathways when the methods of economic and mass allocation are employed in each of the four scenarios ES, EG, MS and MG. Global warming impacts were economically allocated between the coproducts in the following descending order: protein (38%) > laminarin (29%) > packaging material (18%) > fucoidan (14%). Mass-allocated global warming impacts differed significantly, resulting in the descending trend: protein (39%) > packaging material (28%) > fucoidan (18%) > laminarin (15%). In all other impact categories (except marine eutrophication and fossil resource scarcity), economic allocation resulted in the bulk of the environmental burdens being shifted to the laminarin production pathway and the minority to the fucoidan pathway. Whereas, mass allocation (with the exception of fossil resource scarcity and water consumption) assigned the bulk of environmental impacts to the packaging material production pathway. Summarily, LCA of targeted algal packaging material produced via the proposed process scheme conducted employing economic allocation methodology appears to be highly favourable, essentially attributing < 25% of the total impacts incurred by the system to the packaging material route — shifting the remaining 75% of burdens to coproducts, whilst claiming the largest share (49%) of ecosystem services offered in marine eutrophication impact category.
Fig. 4
Bar charts illustrating the percentage distribution of overall environmental impacts between coproducts packaging material (PM), fucoidan (Fu), laminarin (La) and protein (Pr) when economic (blue) and mass (green) allocation is used
×
Full expansion of the system boundaries of multifunctional product systems is contingent on the availability of data relating to marginal systems (Mackenzie et al. 2017) — criterion unfulfilled by modelled systems SS and SG, due to the unavailability of data in the Ecoinvent 3.6 database on products that would be appropriately substituted by the seaweed-derived fucoidan and laminarin. This lack of completeness — hence diminished accuracy — in the LCI data of models SS and SG makes the application of system expansion challenging, not only where this study is concerned, but particularly with future systems deriving high-value, niche and novel products, whose commercial equivalents may not yet be represented in commercial LCA databases. In these cases, allocation methodology may be considered more appropriate (depending on the goal and scope of the study).
Moreover, the appeal of the cascading biorefining model lies in the prospect of value addition to harvested marine biomass — and so with considerations for coproduct market value necessitated by economic allocation methodology.
3.2.2 Product market value, seaweed composition and transport
The results of the analysis assessing the sensitivity of the modelled standard energy system expansion, economic allocation and mass allocation scenarios (SS, ES and MS) to changes in seaweed carbohydrate (alginate and cellulose) composition and the seaweed-based packaging material market value are presented in Fig. 5.
Fig. 5
Sensitivity analysis showing the relative influence of various parameters on potential environmental impacts generated in each standard energy scenario. Parameters assessed included biomass carbohydrate (specifically alginate and cellulose) composition and the market value of the packaging polymer. Parameter inputs were varied at 10% increments within ± 50% of the original input/inventory values. Refer to supplementary information (Fig. S6) for sensitivity results generated for impact categories omitted from this figure
×
There are an array of macroeconomic, regulatory, technological and social factors that can influence the demand — and by extension, the market value — of bioplastics, namely crude oil prices (increase favouring bioplastic demand); GDP growth; feedstock prices; taxation/bans on fossil-based products; governmental subsidisation of bioplastics; bioplastic production scale and consumer awareness and sentiment regarding sustainability (Döhler et al. 2022). This parameter was only applicable to the scenarios utilising economic allocation (ES and EG), where the assumed market value of each coproduct directly impacts the proportion of overall impacts allocated to that specific product pathway (in this case, for packaging material). As expected, there was a positive linear correlation between overall environmental impacts and packaging material market price (Fig. 5). However, in all impact categories, the system was considerably less sensitive to product market price, than seaweed carbohydrate composition.
Seaweed composition and biology vary hugely between species but also vary within species at different stages of the growth cycle or as a result of environmental factors, such as salinity and temperature. Rates of growth, quality and composition of harvested seaweed are greatly influenced by the inevitable fluctuations in temperature and solar radiation that occur with seasonal changes — seasonality being one of the bottlenecks to consistent year-round supply, and the basis for its inclusion as a sensitivity parameter (Fig. 5). It was expected that by increasing the cellulose/alginate composition of the seaweed, the quantity of feedstock required to generate 1 kg of packaging material would decrease (with quantities of other inputs to the product system adjusted accordingly), leading to overall reduced environmental impacts and vice versa — rationale supported by the overall negative correlation observed between composition change and total emissions, across all scenarios and impact categories. Furthermore, for many in the impact categories, the system exhibited heightened sensitivity to carbohydrate composition reduction — for example, a 50% composition reduction afforded a > 100%, ~ 100% and ~ 25% increase in global warming impacts, in scenarios SS, ES and MS.
The initial base models assumed proximity between the seaweed cultivation site and biorefining facility — hence the omission of transport inputs, and an assumed transport distance of 0 t km, in the LCI data for the seaweed cultivation phase. In the UK, domestic freight is transported by road (79%), water (13%) and rail (9%) and water (Mazareanu 2021). The distance between the two geographical extremities of Great Britain (Land’s End in Cornwall and John o’ Groats in Caithness, Scotland) is 1407 km by road. Analysis was conducted on the system’s sensitivity to transport distance, as well as the mode of transport — the results for the scenarios SS, ES and MS, in terms of global warming impacts, presented in Fig. 6.
Fig. 6
Graphs showing the sensitivity of global warming potential of the standard energy models (ES, MS and SS) to transport distance between the seaweed cultivation site and biorefining facility, and the mode of transport used (i.e. lorry, rail and ship). Transport distance was varied from 0 to 1500 km
×
Per functional unit of packaging material produced, transport via lorry in the ES scenario incurred emissions at a rate of \(1.00\times {10}^{-4}\) kg CO2 eq/km, \(1.90\times {10}^{-3}\) kg CO2 eq/km in scenario MS and \(1.44\times {10}^{-2}\) kg CO2 eq/km in scenario SS. The MS and SS models were significantly more sensitive than ES, exhibiting a 40% and 52% increase in overall global warming impacts (compared to 4%) at the upper limit of distance travelled. By dramatic contrast, transport via shipping container, at the upper limit, afforded a comparatively negligible overall global impact increase of 0.24%, 2.30% and 2.97% in the ES, MS and SS scenarios.
The trend in overall global warming impact sensitivity observed between the evaluated modes of freight transport (i.e. lorry > rail (diesel) > rail (electric) > container ship) in each model scenario, is largely a function of ‘load factor’ — and minorly (particularly in the case of diesel-powered and electrified rail) due to the nature of fuel consumed. Load factor is defined as the ratio of average load (i.e. 14.7 kg wet harvest seaweed, required per functional unit of packaging material) to the total vehicle freight capacity (European Environment Agency) — therefore, the load factor of a lorry > ship. In simplistic terms, the GHG emissions associated with the container ship input (i.e. life cycle data relating to the production and maintenance of the container ship, the transportation of freight in bulk and the construction of the port) are spread over a freight load larger than what is possible via lorry — reflected in the comparative shallowness of its gradients in Fig. 6.
Although beyond the scope of this study, the results presented in Fig. 6 are applicable to future-modelled product system scenarios where marine feedstock is cultivated and exported to a UK-based biorefinery from one of the top global producers, e.g. China whose shipping distance from the UK is 22,055 km, Indonesia (17,134 km), South Korea (20,294 km) or the Philippines (18,189 km). The principal motivation behind such a value-chain design would be the prospect of cheaper feedstock prices (relative to UK domestic supply) as a result of the high economies of scale boasted by the maturity of Asia’s Seaweed Aquaculture Industry. Crude extrapolation (assuming a shipping distance ranges representative of South Korea–Indonesia, i.e. 20,294–17,134 km) forecast a 2.7–3.1% increase in equivalent CO2 emissions according to the ES scenario, 24.3–28.7% increase with scenario MS and 32.7–38.7% with scenario SS. However, these values should be regarded purely as ballpark indicators, informing the direction of more complex future LCA (and techno-economic) studies evaluating the supply-chain effects on the environmental profile of seaweed-biorefining product systems, utilising LCI data specific to the country of the chosen suppliers.
4 Conclusion
An explorative, attributional LCA was conducted on a marine biorefining product system producing packaging material (from alginate and cellulose), fucoidan, laminarin and protein supplement from cultivated Saccharina latissima. A total of six scenarios were modelled, varying in the nature of the energy mix applied (i.e. UK standard versus UK green heating and electricity) and the coproduct methodology used to resolve the product system multifunctionality (i.e. system expansion (via substitution), mass allocation or economic allocation). In scenarios where allocation methodology was applied (i.e. with economic allocation and mass allocation, using both standard and green energy mixes), the packaging material production pathway was the assumed route of focus, with the fucoidan, laminarin and protein pathways essentially acting as vectors of value addition. The overall global warming impact values obtained for the economic allocation scenarios ranged between 1.25 and 4.52 kg CO2 eq.; mass allocation scenarios, 3.58–7.06 kg CO2 eq and system-expanded scenarios, 14.19–41.52 kg CO2 eq. — with the lower limit representing instances of green energy usage and the upper standard energy.
Process contribution analysis was conducted on the system-expanded scenarios with the standard energy (SS) and green energy (SG). Whilst the transition to a green energy mix in scenario SG afforded an overall 67% reduction in global warming impacts (kg CO2 eq) and 86% reduction in fossil resource scarcity (kg oil), 2–9 fold increases were exhibited in the categories of terrestrial acidification, human non-carcinogenic toxicity, land use and terrestrial ecotoxicity. From the perspective of process design, this trade-off highlights some of the challenges associated with approaching environmental optimisation solely through the lens of carbon emissions.
This study illustrates the advantage LCA has over single-issue environmental assessment tools like carbon footprinting. Through the application of LCA to seaweed biorefining systems and consideration of a multitude of emission types, a broader and more holistic picture of environmental health can be presented. The distribution of environmental impacts between the four coproducts in the mass and economic allocation scenarios was compared. Economic allocation was found to be most favourable to the packaging material pathway, resulting in its allocation of the highest portion of environmental benefits observed in the marine eutrophication category (49%), and no more than 23% of potential environmental burdens in any other category. A case was made for the adoption of allocation methodology over classically recommended system expansion in the resolution of multifunctionality in seaweed biorefining product systems deriving materials not yet adequately represented in commercial LCA databases.
Analysis was also conducted to determine the system sensitivity to various parameters, such as seaweed carbohydrate (alginate and cellulose) composition, the market value of the packaging materials, transport distance between the seaweed cultivation site and biorefinery and the mode of transport itself. Irrespective of the coproduct methodology applied, overall global impacts were consistently found to be most sensitive to transport via lorry, and the least via shipping container. A crude simulation of a potential instance where marine biomass is sourced from the leading global producers China, Indonesia, the Philippines and South Korea, via shipping container, incurred respective increases in overall global warming impacts by 2.7–3.1%, 24.3–28.7% and 32.7–38.7% in the ES, MS and SS scenarios.
There is no ‘one-size fits all’ solution for the environmental optimisation of marine biorefining product systems — rather, a balance needs to be struck between the breadth of emission categories considered, incorporated energy systems and LCA methodological elements.
Declarations
Conflict of interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
As defined by the World Bank, the ‘blue economy’ refers to the sustainable use of the ocean’s resources for economic growth and the improvement of life quality, whilst simultaneously preserving the health of the marine environment (Vierros and De Fontaubert 2017).
