Water Consumption and the Water Footprint in Aquaculture: A Review

Abstract

Aquaculture is a rapidly growing industry that contributes to the growing global demand for food. Numerous studies have investigated the necessity of increasing food production while reducing its negative effects on the environment. Aquaculture involves the cultivation of aquatic animals such as fish, shrimp, and mollusks that require water for their growth and maintenance in various types of aquaculture operations, such as recirculated aquaculture systems (RASs), ponds, and sea cages. This study investigates mainly life cycle assessment (LCA) in relation to water consumption, the water footprint (WF) and water budgeting approaches in aquaculture. In addition, it contributes to the expansion of knowledge and understanding of the different methodologies used, production practices, types of water (freshwater, marine or brackish) and direct or indirect water consumption in intensive, semi-intensive and extensive types of aquaculture. Notably, this study focuses on water consumption and does not include water indices that account for all the water used in a system, regardless of whether it is returned to the sourced watershed and is therefore available for other uses. Approximately 15% of the reviewed studies focus on the fish processing stage of the production chain, which emphasizes the need for more research on this stage. The species of carp, tilapia, shrimp, and catfish are the most frequently studied aquatic animals in relation to water consumption in aquaculture. Research on water consumption patterns can contribute to the development of a more water-efficient aquaculture system that is essential for promoting sustainable practices.

1. Introduction

The demand for food production and protein is expected to increase significantly over the next few decades due to changes in diet and population growth, which is anticipated to exceed 9 billion by 2050 [1,2]. Because of the problems that humanity faces, such as climate change, population growth, and environmental degradation, sustainable aquaculture production has an impact on global food security [3]. Quantifying freshwater consumption for various food products and production methods has become a priority because of the need to ensure both food and water security. According to Troell et al. [4], aquaculture is crucial because of the potential benefits for all 17 SDGs, with the most evident ones addressing hunger elimination and health promotion (SDGs 2 and 3); responsible production and consumption to improve the environmental sustainability of the land, water, climate, and oceans (SDGs 6, 12, 13, 14, and 15); and reductions in poverty and inequalities and improved livelihoods (SDGs 1, 5, 8, and 10).

An estimated 179 million tonnes of fish were produced worldwide in 2018, of which 82 million tonnes were produced in aquaculture, while China is considered a significant fish producer, with a contribution of 35% to global fish production in the year 2018 [5]. One of the fastest-growing markets for livestock feed worldwide is aquaculture, and fish consumption constitutes an option for supplying the growing human population with environmentally sustainable protein sources. Between 1961 and 2017, the average annual growth rate of global food and fish consumption was 3.1%, which was higher than the 2.1% annual growth rate of all other foods containing animal protein, such as dairy, meat, and milk, and nearly twice as fast as the global population was growing (1.6%) during the same period [5]. Moreover, the demand for aquatic food products is anticipated to increase globally, and estimates of population and income by 2050 indicate that more than 500 megatons (Mt) of meat will be required annually for human consumption [3]. The worldwide per capita food consumption of fish and seafood from 1961 to 2013 increased from 9.96 kg/capita/year to 19.86 kg/capita/year [6].

Systems for raising livestock that produce meat and other animal products rely heavily on natural resources. According to Mekonnen and Hoekstra [7], 29% of the global agricultural sector’s water footprint (WF) is used in the production of animal feed, whereas beef accounts for one third of the WF worldwide. Only in recent years has the scientific community begun to pay considerable attention to the WF of aquaculture.

Investigating the sustainability of aquafeed production has also emerged as an important issue. Concerns regarding sustainability exist for capture fisheries, which are essential for the production of fishmeal and fish oil because, due to the depletion of many marine fish stocks, fewer fishmeal volumes are available for fish farming, which increases the cost of fishmeal and makes alternatives more affordable [2]. Due to cost concerns, as well as the knowledge that wild fish are a finite resource, efforts are being made to identify substitute feed ingredients that provide lipids and protein, including microalgae, yeasts, insects, and plants [8,9].

Aquaculture consumes water directly and indirectly. Direct water use in aquaculture can be categorized into consumptive water use, which refers to the removal of water from the system that is not returned, such as the sum of the reductions in stream flow, groundwater and water incorporated into the farmed biomass, and total water use, which includes all sources of water input and output, such as runoff, precipitation, seepage and any water added by management practices [10]. Consumptive use occurs when freshwater is used and cannot be released into the same watershed because of evaporation, product integration, or discharge into the sea or other watersheds [11]. The portion of water that is not returned and either evaporates or is incorporated into products and/or organisms is considered consumed and no longer available, whereas water withdrawal refers to the diversion of water for human use from streams, rivers, or aquifers; however, some of the water is returned and may then be recycled or restored to the environment [12,13]. Freshwater in the aquaculture production system itself is related to the direct water footprint (WF), whereas the indirect WF refers to the amount of water used throughout the entire supply chain of aquaculture operations, including the production of feed, energy, and other inputs [14].

Vasquez-Mejia et al. [15] reviewed quantitative water use (volume of water input into a system irrespective of whether it is returned to the sourced watershed) and focused on finfish aquaculture utilizing life cycle assessment (LCA). This review included LCA impact factors such as water dependence, which accounts for total water use (the total volume of water used, diverted, flowed or pumped from a river, divided by fish growth recorded during a reference period) [15]. Bohnes et al. [16] performed a critical review of the LCA of aquaculture systems, whereas Philis et al. [17] compared the status of the LCA of salmonid aquaculture production systems. Furthermore, Ghamkhar et al. [18] analyzed particular procedures (energy, feed and infrastructure) and specific impact categories (water use, land use, and eutrophication potential) of the LCA of aquaculture. The aforementioned reviews are focused on LCA approaches and investigate water use regardless of whether it is consumed and no longer available for other purposes.

According to Boyd [19], the primary concern should be consumptive water use by freshwater aquaculture, as this consumption diminishes the amount of water available for other advantageous uses and enables estimates of the economic value of water. The amount of freshwater consumed in the production chains of aquaculture processes is a field that has begun to be explored more thoroughly in recent years. Furthermore, the identification and assessment of the water-consuming procedures in brackish and marine aquaculture are important for proper water management practices, and further research in this area is needed.

Thus, this study reviews the literature, focusing, for the first time, solely on the aspects of aquaculture water consumption—namely, water that is not returned to the sourced watershed and is not available for other uses—including studies based on life cycle assessments (LCAs), as well as water budgeting/hydrologic analysis approaches and water footprint estimations. It aims to (a) summarize the main findings of the studies performed on the investigation of water consumption utilizing water budgets, LCA and water footprint accounting in different types of aquaculture systems (freshwater/brackish/marine), (b) analyze the aquaculture supply chain stages, management practices and highlight the main related methodological concerns in relation to the water consumption patterns, and (c) suggest future research activities.

2. Materials and Methods

Studies concerning the water consumption of aquaculture have been summarized, and an effort has been made to identify the gaps in the field’s evolution by analyzing the differences in methodological approaches, spatial scales, water types, technologies, process steps, species, system boundaries, and aquaculture practices.

Articles were identified and selected using the online search engines Scopus, Web of Science, Google Scholar and Google. The keywords and the search operators used were a combination of “water footprint” OR “water consumption”, OR “consumptive water”, OR “water productivity” AND “aquaculture”. The timespan ranged from 2004 to May 2024. An analysis of 34 studies (peer-reviewed articles) based on the water consumption aspects of aquaculture systems was performed in this review.

The exclusion criteria included studies that (1) were not relevant/did not include water consumption or WF estimations; (2) were literature reviews, book chapters, thesis/dissertations, or conference papers; (3) did not focus on aquaculture; (4) were outside of the specified time frame; (5) dealt with aquaponics/integrated aquaculture–agriculture; (6) dealt with the WF of food consumption (diet WF); (7) were not published in English; and 8) were not accessible. The database and cross-reference search identified 257 works after the removal of duplicate records; 223 works were rejected after the cut-off criteria were applied, and as a result, 34 papers are examined. The methodology for selecting the studies included in the present review is presented in Figure 1.

Notably, the current review is focused mainly on the quantitative aspects of aquaculture practices, and it does not thoroughly examine aspects of water quality, such as water degradation. Moreover, the present study reviews papers focused on the WF of aquaculture food production and is not oriented towards studies on the WF of consumption due to different diets that include aquaculture products [8,20,21,22,23,24,25]. This review examines all aquaculture species, such as fish, shrimp, and mollusks, and the different applied methodologies.

Figure 1. PRISMA flowchart for the selection methodology of the studies under review. Source: [26].

Figure 1. PRISMA flowchart for the selection methodology of the studies under review. Source: [26].

34 works were chosen using the methodology shown in Figure 1, and their main characteristics such as the examined aquaculture species, the region of the study, the estimated indexes, the applied methodologies and the supply chain stage are analyzed in

Table 1. List of articles selected for review.

ID	Study	Species	Region	Indexes	WP 5/ CWP 6	Methodology	Aquaculture Type	Supply Chain Stage	Water Type
1	Malcorps et al. [27]	Shrimp	Global	WF 1		WFN 7	NR 11	Feed production	NR
2	Mohanty et al. [28]	Pacific white shrimp	India	WF, CWU 2, CWUI 3	✓ 13	HA 8	ponds	Fish production	brackish water
3	Guzmán-Luna et al. [14]	Tilapia	Mexico	WF		WFN	extensive, semi-intensive and intensive	Feed and fish production, Fish processing	freshwater
4	Song et al. [29]	24 farmed fish	China	WF		WFN	pond, open waters, paddy field, and industrial farming system	Feed and fish production	marine and freshwater
5	Jiang et al. [30]	Fish, shrimp, and bivalves	Global	WF	✓	WFN	NR	Feed production	marine and freshwater
6	Yuan et al. [31]	22 popularly farmed fish	China	WF		WFN	ponds; lakes, reservoirs and rivers; rice fields; industrialized systems	Feed production	marine and freshwater
7	Pahlow et al. [32]	39 major fish and crustacean	Global	WF	✓	WFN	NR	Feed production	marine and freshwater
8	Pérez-Rincón et al. [33]	Tilapia, cachama, trout	Colombia	WF		WFN	ponds	Feed and fish production, Fish processing	freshwater
9	Newton et al. [34]	Atlantic salmon	Scotland	CWU		LCA 9	marine net pens	Feed and fish production, fish primary processing	marine
10	Henriksson et al. [35]	Tilapia	Egypt	FWC 14		WFN and LCA	ponds	Feed and fish production	freshwater
11	Henriksson et al. [36]	Eight species	Indonesia	FWC		LCA	ponds and cages	Feed and fish production	marine, brackish and freshwater
12	Cooney et al. [37]	Eurasian perch	Ireland	WU AWARE 4		LCA	RAS and ponds	Feed and fish production	freshwater
13	Féon et al. [2]	Trout	France	WU AWARE		LCA	ponds	Feed and fish production	freshwater
14	Petroski et al. [38]	Nile tilapia	Brazil	WC12		LCA (ReCipe)	cages	Feed and fish production	freshwater
15	Viglia et al. [39]	Farmed catfish	USA	FD 10		LCA (ReCipe)	ponds	Feed and fish production, Fish processing	freshwater
16	Konstantinidis et al. [40]	Seabass and meagre	Greece	WC		LCA (ReCiPe)	marine cages	Feed and fish production, Fish processing	marine
17	Haslawati et al. [41]	Giant Freshwater Prawn	Malaysia	WC		LCA (ReCiPe)	ponds	Fish production	freshwater
18	Mohanty et al. [42]	Shrimp	India	WF, CWU, CWUI	✓	HA	ponds	Fish production	brackish water
19	Pattusamy et al. [43]	Shrimp	India	CWUI	✓	HA	ponds	Fish production	brackish water
20	Mohanty et al. [44]	Carp polyculture	India	CWU, CWUI	✓	HA	ponds	Fish production	freshwater
21	Das et al. [45]	Carp polyculture	India	CWU	✓	HA	large outdoor concrete tanks	Fish production	freshwater
22	Adhikari et al. [46]	Freshwater fish	India	CWUI		HA	ponds	Fish production	freshwater
23	Mohanty et al. [47]	Carp-prawn polyculture	India	CWU, CWUI	✓	HA	ponds	Fish production	freshwater
24	Tucker et al. [48]	Ictalurid catfish	USA	CWUI		HA	ponds	Feed and fish production	freshwater
25	Das et al. [49]	Carp polyculture	India	CWU, CWUI	✓	HA	ponds	Fish production	freshwater
26	Sharma et al. [50]	Carp polyculture	India	CWUI	✓	HA	ponds	Feed and fish production	freshwater
27	Mohanty et al. [51]	Carp-prawn polyculture	India	CWU, CWUI	✓	HA	ponds	Fish production	freshwater
28	Mohanty et al. [52]	Black Tiger Shrimp Penaeus monodon	India	CWU, CWUI	✓	HA	ponds	Fish production	brackish water
29	Mohanty [53]	Carp-prawn polyculture	India	CWU, CWUI	✓	HA	ponds	Fish production	freshwater
30	Gephart et al. [54]	Chinese aquaculture	China	WF		WFN	semi-closed and open water systems	Feed and fish production	marine, brackish and freshwater
31	Lima et al. [55]	Nile tilapia fingerling	Brazil	WC		HA (on-site measurement)	low-salinity biofloc system	Fish production	low-salinity water
32	Konstantinidis et al. [56]	Sea Bass	Greece and Albania	WC		LCA (ReCiPe)	marine cages	Feed and fish production	marine
33	Troell et al. [8]	Marine aquaculture species	Global	WF		WFN	NR	Feed production	marine
34	Boyd [19]	Channel Catfish	Alabama	CWU, CWUI		HA	ponds	Fish production	freshwater

Notes: ¹ WF: water footprint; ² CWU: consumptive water use; ³ CWUI: consumptive water use index; ⁴ WU AWARE: water use using the AWARE methodology ⁵ WP: water productivity; ⁶ CWP: consumptive water productivity; ⁷ WFN: water footprint network; ⁸ HA: hydrologic analysis; ⁹ LCA: life cycle assessment; ¹⁰ FD: freshwater depletion; ¹¹ NR: not reported, ¹² WC: water consumption, ¹³ ✓: Included in the study, ¹⁴ FWC: freshwater consumption.

.

3. Results

3.1. Analysis of Findings

An LCA was used in 29.4% of the reviewed studies to analyze the environmental performance of aquaculture, including its water consumption throughout different stages of the supply chain, whereas 26.5% of the papers used the WFN approach to estimate the WF associated with feed and/or fish production and processing. The remaining studies (44.1%) used a hydrologic analysis approach that applies a water balance accounting for inland aquaculture (Figure 2).

As shown in Figure 3, studies on the analysis of water consumption, which have focused mainly on water budgeting, LCA estimations and the WF in aquaculture, have focused primarily on the last ten years, even though the time frame that concerns the findings of the current review is for the period of 2004–2024. Notably, the inquiry was performed until May 2024.

3.1.1. Studies Based on the Water Footprint Network Approach

The water footprint (WF) of a product, such as fish, refers to the volume of freshwater used throughout the entire production process of that product, including all stages of the supply chain [57]. According to Hoekstra et al. [57], the blue WF specifically refers to the consumption of blue water resources, which includes surface water and groundwater, throughout the supply chain of a product. This measurement accounts for the loss of water from the available water body in a catchment area due to factors such as water evaporation, discharge into different watersheds or the sea, or incorporation into a product within the same period. The green WF is related to the consumption of green water resources (rainwater, as it does not become runoff) [57]. Given ambient water quality standards and natural background concentrations, the gray WF is the volume of freshwater needed to assimilate the load of pollutants [57]. Agriculture uses green water extensively, as it is needed to grow plant-based ingredients for aquafeed. Globally, green water accounted for more than 85% of the total water consumption by agricultural crops from 1971 to 2000 [58]. The studies described below measured WF indicators via the WFN approach, and most of the research utilized focused on the blue, green and gray WFs of the supply chain stage of feed production in aquaculture.

The global WF related to aquaculture feed production was estimated for the first time by Pahlow et al. [32] in 2015. This study obtained data for the feed ingredients from Mekonnen and Hoekstra [59,60], Mekonnen and Hoekstra [7] and Van Oel and Hoekstra [61] and examined the WF of terrestrial feed ingredients. These findings have been used in various papers to estimate diet-related water consumption. Fish and crustaceans fed commercial aquafeed were computed to have green, blue, and gray production weighted average feed WFs of 1629 m³ t⁻¹, 179 m³ t⁻¹, and 166 m³ t⁻¹, respectively, whereas for the year 2008, the estimated total WF of commercial aquafeed worldwide was between 31 and 35 km³. The findings of the study [32] revealed that grass carp, Nile tilapia, common carp, Atlantic salmon, and whiteleg shrimp accounted for the top five contributors to the overall WF of commercial feed, totaling 18.2 km³. The investigation of alternative diets revealed the possibility of additional strain on freshwater resources if fish oil and fish meal are substituted for components of terrestrial feed, which may increase the WF of feed production and therefore the WF of aquaculture production. Their findings demonstrate the need to consider freshwater consumption and aquafeed-related pollution for the aquaculture industry to expand sustainably. While this study [32] provides a thorough investigation of the indirect WF (of commercial feed), it does not examine water consumption throughout the entire aquaculture production supply chain.

In a study by Yuan et al. [31], 22 species of fish that are commonly raised in China were analyzed for average growth and feed parameters to determine the feed-associated WFs during the farming period. The WFs that were examined for every species of fish were restricted to the feeds that the fish consumed during their growth period. They estimated a production-weighted average WF of 3.11 L g⁻¹ and average blue, green and gray components of 0.74, 1.93 and 0.44 L g⁻¹, respectively. The WF of marine fish was 1.49 L g⁻¹ lower than that of freshwater fish, which was 3.16 L g⁻¹, since wild fish used as feed for marine aquaculture are considered to consume negligible amounts of water. Rice, soybeans, and wheat bran, which require a large amount of water, are the main crops used as crop feeds in freshwater aquaculture. A comparison of the WF values with those of the study conducted by Pahlow et al. [32] (which used the same database) revealed variations from one fish species to another, and Yuan et al. computed higher WFs (with the exception of the species flounder, grass carp, Asian swamp eel, red drum and grouper). These deviations may have occurred because of differences in methodologies and in the feed ingredients of farmed fish. Their findings show that because different fish species have large feed input variances and diverse WFs, replacing terrestrial animal food products with farmed fish in China might not lead to a decrease in the amount of consumed water.

In another study conducted in China, Song et al. [29] analyzed 198 common feed components and measured the WF originating from evaporation loss, feed consumption, and pollutant discharges of 24 farmed fish species. Their estimations included the feed-related WFs of the 22 fish species examined in the study by Yuan et al. [31]. The WFs of fish farming, as well as economic and protein production, were measured in the baseline year of 2014. In addition, with 2020 as the target year, 29 scenarios were created to represent the species composition of fish raised in aquaculture. They performed optimizations for all predefined scenarios to minimize WF, prevent reductions in the economic and protein outputs of China’s fish farming system, and reduce the strain on wild fishery captures. In this study, gray water was calculated using nitrogen as a proxy. The production-weighted average WF of farmed fish was 5.51 m³ kg⁻¹, which is greater than that reported by Yuan et al. [31] and Pahlow et al. [32], who focused on the feed production stage. Freshwater-farmed fish presented a production-weighted WF of 5.02 m³ kg⁻¹, whereas the WF of mariculture fish was 16.08 m³ kg⁻¹ (with a significant amount of gray WF). According to their study, blue water is consumed during fish farming because of evaporation, and, for example, compared with the water equivalents of feed consumption, which were 4.00 m³ kg⁻¹ and 3.90 m³ kg⁻¹, the average evaporation losses of raising tilapia and sea bass were higher (5.21 m³ kg⁻¹ and 4.58 m³ kg⁻¹, respectively).

The WF studies under review in the present study broadly consider that the amount of water consumed in the process of producing animal byproducts such as bone or blood meal might be minimal, since it is already calculated for the production of food for humans, and that the water consumption to produce fish meal and fish oil can also be regarded as negligibly small. Based on these hypotheses, Pahlow et al. [32] considered the aforementioned feed ingredients to have a negligible WF and that the types and quantities of plant-based components are the primary determinants of feed-associated water consumption. According to Verdegem et al. [13], water use for fish oil and fish meal production is minimal compared to the amount of water used for crop production, and the types and quantities of plant ingredients used in the feed are the primary determinants of feed-associated water use. Malcorps et al. [27] obtained data on the global average freshwater demand for the processing of fish meal and fish oil from Chatvijitkul et al. [62].

Malcorps et al. [27] gradually replaced fishmeal with plant ingredients in shrimp feed and evaluated the effects on terrestrial and marine resources such as land, fish, nitrogen, freshwater, and phosphorus. Their findings showed that a complete substitution of 20–30% of fishmeal totals might result in an increase in the demand for freshwater. Data were obtained from Chatvijitkul et al. [62] and Mekonnen and Hoekstra [59]. In the alternative plant scenario, gray water increased as a result of the increase in the amount of fertilizer-demanding crops for the needed ingredients, whereas blue water decreased as a result of the high inclusion rate of pea protein concentrate. Changes in the inclusion of specific ingredients affected the total WF, with variations in the green, blue, and gray components. The research findings for the total WF of Litopenaeus vannamei under different scenarios (fishmeal 20%) were strikingly comparable to those for Penaeus monodon (fishmeal 24%) (both close to 1600 m³ MT⁻¹), which were similar to the findings of Pahlow et al. [32] (fishmeal inclusion of approximately 25%).

The outcomes of the research conducted in Mexico by Guzmán-Luna et al. [14] revealed that when comparing the WF evaluations of tilapia filet protein, two and four times more freshwater is needed compared to the WFs of beef and pork protein, respectively. Because of the continuous effluent loads from the ponds, tilapia filets use more freshwater than beef, pork, and poultry do and furthermore contaminate more water than other terrestrial animals do. They estimated the energy, land and blue, green and gray WFs of semi-intensive, extensive, and intensive tilapia aquaculture. The results of the study [14] revealed that the blue WF of the intensive system was 14 times greater than the blue WF of the extensive system and 4.5 times greater than the blue WF of the semi-intensive system, mainly because of the intensive system’s reliance on high refreshment rates (refreshment rates of 250% of the pond water per day) and higher stocking density. They argued that water exchange rates must be reduced to reduce blue WFs. While green WFs are negligible in extensive systems (with no need for aquafeed), they are relatively large for tilapia produced in semi-intensive and intensive systems. The gray WFs of tilapia were also high (greater than those associated with meat production). The use of aquafeed and water refreshments determines the total WFs during the fattening phase of semi-intensive and intensive systems, and according to their findings, that stage adds the greatest amount to the WF. Guzmán-Luna et al. [14] estimated WFs for extensive, semi-intensive and intensive systems of tilapia as blue WFs of 927 m³ t⁻¹, green WFs of 5 m³ t⁻¹, gray WFs of 398 m³ t⁻¹, blue WFs of 2909 m³ t⁻¹, green WFs of 7827 m³ t⁻¹, gray WFs of 1873 m³ t⁻¹, and blue WFs of 13,027 m³ t⁻¹, green WFs of 7831 m³ t⁻¹, and gray WFs of 1873 m³ t⁻¹, respectively, whereas Pérez-Rincón et al. [33] estimated a WF of 5486 m³ t⁻¹ for the aquaculture of tilapia in Colombia.

Pérez-Rincón et al. [33] computed the direct and indirect WFs among the three species that are produced most frequently in Colombia: tilapia (Oreochromis mossambicus), trout (Oncorhynchus mykiss), and cachama (Piaractus brachypomus). The WFs of tilapia, cachama and trout cultures were 5486 m³ t⁻¹, 6193 m³ t⁻¹, and 19,854 m³ t⁻¹, respectively. The concentrated feed had the highest overall WF for tilapia, followed by the blue WF, which is related to the volume of water that remains stored in the ponds. For trout and cachama, on the other hand, the gray WF was the highest because fish excretions and leftover feed contain high concentrations of solids and nutrients. The direct WF is the primary component of the total WF in all three species, accounting for 63% of the total WF in cachama, 50% of the total WF in tilapia, and 85% of that in trout, according to an integrated analysis of the data. The gray WF was the primary component of the direct WF in the trout and cachama cultures, whereas the blue WF was the main constituent in the tilapia culture. The gray WFs in all the cases were linked to nutrients in the water, which were present as leftover food and fish waste. Despite the low effluent concentrations, the production of trout had the highest total and direct WFs. During the trout fattening stage, if the ponds’ input flow was equal to the theoretical amount (18 L s⁻¹), the total WF would drop to 5841 m³ t⁻¹, demonstrating the importance of flow rate optimization for this crop.

Jiang et al. [30] formulated a food–energy–water–carbon (FEWC) index ranging from 0 to 100 to assess the sustainability of aquaculture globally. These results indicate that the sustainability of aquaculture is generally low. They estimated that the water consumption due to aquaculture production was approximately 122.6 km³ in 2018. Because aquafeed was the only variable considered, the WFs may have been underestimated. The FEWC index results showed that developing nations typically have relatively high water/energy intensities (water/energy consumption per unit of aquaculture production), probably because these countries tend to have low feed conversion ratios, inefficient farming technologies, and resource-intensive species. Norway is a developed country with a subsustainability score for food and water that is more than ten and two times greater than that of Egypt, respectively, even though the volume of aquaculture production in both countries is comparable. This disparity may be explained primarily by factors such as feed technology, management and farm species, and different culture environments (freshwater vs. marine), as aquafeed input is typically less necessary in nature-based mariculture.

Gephart et al. [25] computed the WF of marine fish protein consumption and assessed the potential freshwater savings from substituting with terrestrial protein without considering the WF generated from the feed resources for marine aquaculture, which include agricultural compounds. Troell et al. [8] complemented their approach and investigated the WF of the crops used to feed marine aquaculture, and their estimations resulted in a total WF of 8 km³ yr⁻¹.

3.1.2. Life Cycle Assessment Studies

The life cycle assessment (LCA) methodology is widely employed to assess the environmental effects of a product, process, or service over its full life cycle—from cradle to grave. LCA investigates water consumption in aquaculture by evaluating the entire life cycle of aquaculture systems, from feed production to fish harvesting and processing. LCA provides a comprehensive understanding of the water consumed in aquaculture practices and helps in the evaluation of the potential environmental impacts, helping stakeholders make informed decisions to enhance sustainability.

Newton et al. [34] utilized a midpoint CML2001 with a focus on acidification potential (AP), global warming potential (GWP), eutrophication potential, ozone depletion potential, consumptive water use, photochemical oxidation potential, and land use in farmed Scottish salmon aquaculture. In addition to eutrophication potential because of direct nitrogenous emissions into the marine ecosystem, more than 90% of the impact on the farm gate was attributable to feed. The total amount of water required was estimated based on crop rainfall requirements and assumed to be provided either by irrigation or precipitation. The consumptive water use in m³ t⁻¹ was obtained by extrapolating rainfall to the given yields per hectare. Mekonnen’s and Hoekstra’s [60] data on crop water requirements were used to validate the estimates in the study. The overall effects of feed production are anticipated to increase with the increasing substitution of marine ingredients with refined vegetable ingredients, as vegetable ingredients account for more than 99.7% of each category in the cases of both consumptive water use and land use.

Henriksson et al. [35] investigated best management practices in the aquaculture of Nile tilapia and explored five LCA impact categories—eutrophication, global warming, acidification, freshwater consumption (FWC) and land use—using CMLCA v5.2 for model construction. The FWC in industrial processes was derived from Ecoinvent v2.2, whereas the FWC in agriculture was defined by the blue WF according to Mekonnen and Hoekstra [60]. Water scarcity indices and water degradation were not considered when accounting for FWC; instead, it was defined as the amount of FW consumed via evaporation. The study area presented high pond evaporation rates (average of 556 l m⁻² month⁻¹). According to their findings, FWC was primarily caused by pond evaporation due to Egypt’s warm and arid climate combined with long grow-out times, followed by the consumption of irrigation water in agriculture (accounting for 7–12% of the overall FWC).

Henriksson et al. [36] quantified the environmental impacts and socioeconomic indicators for six different projections of Indonesian aquaculture growth through 2030 by investigating ten production systems and eight species. The amount of fresh water lost to evaporation in ponds was computed using climate data across Indonesia, and the evaporation rate was assumed to average 317 ± 76 l m⁻² month⁻¹. Regarding freshwater consumption, only consumptive use was considered, which was limited to the use of freshwater that was rendered unusable for other purposes due to pond evaporation (not cages) or dilution with seawater. With respect to freshwater consumption and land use, groupers performed better than did any other species, as they were primarily fed wild-caught fish being raised in marine cages; however, in the other impact categories, they performed the worst. The impact category with the highest overall dispersion between outcomes was freshwater consumption, which also showed the greatest variation across production systems. Irrigation and pond evaporation dominated freshwater consumption in freshwater systems. Brackish water ponds (milkfish and shrimp) were found to be the primary consumers of fresh water (the dilution of marine water into brackish water is a water-consuming procedure). Caged tilapia and pangasius farms used marginally less water than pond farms did because feed production accounted for more than 80% of freshwater consumption in these systems.

Two of the reviewed studies used the AWARE approach [2,37]. In Boulay et al. [63], according to the new WF framework outlined in the ISO 14046 standard [64], a life cycle assessment (LCA) was used to evaluate freshwater-related impacts. The suggested approach, known as AWARE, is predicated on measuring the relative amount of water that is still available in each area after human and aquatic ecosystem demands have been satisfied. Water consumption is the portion of water withdrawn that is no longer available for the users of the originating river basin as a result of product integration, evapo(transpi)ration, or discharge into other basins or the sea. Cooney et al. [37] performed an LCA of RAS for perch production and examined the impact categories of acidification potential (AP), global warming potential (GWP), freshwater and marine ecotoxicity potential (FAETP and MAETP), eutrophication potential (EP), net primary production use (NPPU), cumulative energy demand (CED), and water use. Feed production accounted for 43% of the water use (AWARE) category, with fish production coming in second at 22% and electricity production at 20%. The average volume of water needed to produce one kilogram of fish was 1.49 m³. Féon et al. [2] investigated pan-sized trout production using insect-based meal rather than fish-based meal and compared a baseline scenario with three other scenarios: one with a baseline consisting of zero mealworm meal and two with 15% and 30% of fish meal replaced with mealworm meal. The large amount of water that was consumed directly from the river, which was required in some processes, such as egg production (which did not vary across scenarios), had a significant effect (on AWARE). The proportion of mealworm meal in the feed decreased the impact on AWARE, in contrast to other impact factors investigated. The study’s findings suggest that including mealworm meal in fish feed may not produce many environmental benefits. Nonetheless, mealworms were supplied with food from agricultural coproducts (such as beet pulp and wheat bran) and specially prepared feed that included dietary supplements and agricultural coproducts with effects on the environment.

Haslawati et al. [41] performed a cradle-to-farm LCA to assess the potential environmental impacts (terrestrial ecotoxicity, global warming, terrestrial acidification, freshwater eutrophication, human carcinogenic activity, water consumption and human noncarcinogenic toxicity) of the giant freshwater prawn GFP in Malaysia. Pond preparation, farming, stocking, and harvesting were the four main methods of iterative farming. The software SimaPro 9.3.0.3 was used for the impact analysis, with background data from the databases Ecoinvent 3.0 and the ReCiPe 2016 midpoint method. In the ReCipe 2016 method [65], every impact related to water is derived from water consumption, which refers to the utilization of water in a manner in which the resource is incorporated into products, evaporated, or transferred to other watersheds or to the sea. No discernible impact on water consumption was observed in this study.

The environmental hotspots in seabass and meagre farming associated with fish feeds of different granulations were studied by Konstantinidis et al. [40]. The amounts of raw materials, heat, and energy required during the production of fish feeds were considered. The amounts of energy, fuel and feed required for each size class, which are required to produce one ton of meagre or seabass, were determined using an LCA of Greek cage farms. They used the ReCiPe 2016 method [65] with 18 impact categories. They considered three feed size classes with various compositions, or “formulas”, comprising varying proportions of wheat, fishmeal, fish oil, vitamins, soy, etc. In particular, in seabass and meagre, fry production affected the impact category “water consumption” by 85.4% and 66.2%, respectively. Due to the extensive water requirements of hatcheries, fry production is the main factor affecting water consumption. When meagres and seabass were compared, meagres showed a noticeably smaller impact on all eighteen environmental impact indicators.

The assessment of environmental and energy performance throughout the life cycle of Nile tilapia production for harvesting at varying weights was investigated by Petroski et al. [38]. The impact categories included global warming, area occupation, energy demand, water consumption, acidification and eutrophication. The water volume required for the production of inputs (such as irrigated agriculture/manufacturing processes) and fingerlings, as well as the evaporated water from the lake’s cage, was considered when calculating water consumption. Although the smaller fish required more fingerlings than the larger ones did to produce 1 t of Nile tilapia, the smaller fish had a better feed conversion rate. The fish rearing and fattening processes were primarily responsible for acidification and eutrophication, whereas feed production dominated the impacts of area occupation, energy demand, water consumption and global warming.

Viglia et al. [39] investigated the water and energy demands in American catfish aquaculture systems from the cradle to the processing/packaging gates and conducted in-depth interviews with stakeholders. For the estimation of indirect energy and water consumption, SimaPro software (v 9.0.0.30), the Ecoinvent database, and the ReCiPe midpoint (H) method were used. The freshwater depletion potential impact category was calculated (freshwater depletion according to Simapro [66] is the amount of fresh water consumed). The water embodied in the agricultural production of feed ingredients accounted for the vast amount of water used in feed production (98%). Ninety percent of the total embodied water (1038.2 L kg⁻¹) concerned direct water use (evaporation, water replenishment, and pond refilling after drawdown) in the hatchery stage (933.3 L kg⁻¹). Τhe largest percentage (59%) concerned direct water inputs for filling ponds after harvesting and evaporation, whereas the water demand for feed production was 40%.

3.1.3. Hydrologic Analysis/Water Budgeting Studies

A standard hydrologic analysis can be used to illustrate the primary direct water losses from ponds, such as seepage, evaporation, and water exchange. In some studies included in this review, the hydrological water balance equation, inflow = outflow ± change in volume (ΔV), was utilized to obtain accurate estimates of the amount of water used in ponds. In most of the papers, the total water use (probable inflows to ponds) was estimated as the sum of the initial water filling, the additions or regulated inflows during management, the precipitation, the groundwater seepage (Si), and the runoff, whereas the consumptive water use (possible outflows) was computed as the sum of the intentional discharge or regulated discharge, the evaporation, the overflow, the transpiration, the seepage, and the water content in the harvested biomass. The estimated consumptive water use index (CWUI), which indicates the volume of water used per unit of production, is displayed as follows: CWUI = CWU (m³)/production (kg).

Two studies conducted in India investigated the shrimp species Litopenaeus vannamei [28,42]. Mohanty et al. [28,42] investigated the effects of stocking density and the water-saving approach on sediment quality and water, production performance and growth of a shrimp species (Litopenaeus vannamei) and examined how water conservation strategies and stocking density affect the growth, productivity, and quality of water and sediment. Mohanty et al. [28,42] examined the effects of rearing Litopenaeus vannamei at three densities using three replicate treatments. They investigated various aspects of water budgeting to reduce water waste and production costs and improve water productivity and WF, and performed a hydrological water balance study to quantify the total water requirement (TWR), CWU and CWUI. The consumptive water use efficiency (WUEc) was computed by dividing the shrimp production in kg ha⁻¹ by the CWU. The WF was measured in cubic meters (m³) of consumptive water per ton of shrimp produced. All four aspects of water loss—water contained in harvested biomass, evapotranspiration and/or evaporation, contaminated water, and water that is not returned to the original location—were considered. Seepage and percolation losses, on the other hand, were excluded from the estimations of the WF because they are not considered losses to the catchment and can be recycled within the same area. In L. vannamei culture, an optimum stocking density of 50 post-larvae/m² is perceived as a way to improve productivity, CWUI, WUEc and total WF. The investigations conducted at various stocking densities suggest that the TWU increases with increasing water exchange. Low to moderate water exchange farming systems reduce the amount of sediment and effluent outputs, enhance water use efficiency, and maintain water quality appropriate for shrimp growth. They concluded that a crucial step in enhancing aquaculture performance is to conduct a hydrological water balance study to ascertain density-dependent water use in ponds.

Pattusamy et al. [43] and Mohanty et al. [52] examined shrimp species Penaeus vannamei and Penaeus monodon in India, respectively. They estimated the CWU and CWUI, but they did not include WF computations. Pattusamy et al. [43] conducted a 90-day study to determine how much water is needed for the culture of Penaeus vannamei in inland saline groundwater, and three distinct stocking densities were used in the earthen grow-out ponds. A stocking density of 45 per m² and a water budget of 6.53 m³ kg⁻¹ of shrimp production could be optimal, considering the total amount of water used and the amount of shrimp produced from the earthen grow-out ponds using inland saline groundwater. In the study by Mohanty et al. [52], lower rates of water exchange demonstrated noticeably better crop performance, productivity, and water quality compared to zero water exchange. Additionally, the results under three different feed management protocols showed that the total water use (TWU) and CWUI increased with feed input, while it was also noted that the yield and the growth performance increased with the length of the refeeding period.

Carp polycultures in India were investigated by Mohanty et al. [44], Das et al. [49], Sharma et al. [50] and Das et al. [45]. The consumptive water use per kg of fish in the treatments used by Das et al. [49] ranged from 1.66 to 1.96 m³ kg⁻¹. A substantial portion of consumptive water use (CWU) was contributed by seepage and evaporation losses. Sharma et al. [50] estimated the total water requirements in a semi-intensive polyculture system to be 10.3 m³ kg⁻¹ fish (or 6.5 m³ kg⁻¹ fish if the seepage losses were considered green water), of which 7.6 m³ kg⁻¹ was the system-associated demand. Mohanty et al. [44] reported that TWU increased with increasing water exchange at various stocking densities. Furthermore, the CWUI ranged from 5.61 to 6.38 m³ kg⁻¹ fish. According to all these studies, the amount of water needed for carp production varies based on the culture conditions. According to Mohanty et al. [44], moderate/reasonable rates of water exchange significantly improved crop performance overall and water use efficiency (WUE) compared with high water exchange at relatively high densities and no water exchange at relatively low densities.

Large outdoor concrete tanks (50 m²) were used to test the effects of different water exchange practices (0%, 40%, 60%, 80%, and 100% in total in four phases over three months) on fingerling production in Indian major carp in the study by Das et al. [45]. When calculating TWU, consideration was given to the initial water amount in the tanks as well as the following water replenishment for evaporation loss (in control) and exchange (in treatments) supplied during the rearing stage, whereas the computation of CWU took into account the water volume lost from the tank (exchange and evaporation) in the raising stage. Water loss from seepage in tank systems was minimal and was not considered in the study. However, the consumptive water uses included the water exchanged on a regular basis and the volume of water added to replenish evaporation loss. Due to its higher water productivity as measured in terms of number and biomass production of fingerlings per unit (m⁻³) of TWU and CWU, the study suggests 80% as the optimal water exchange rate in concrete tank rearing systems for the production of fingerlings of Indian major carps.

The amount of water consumed for carp culture was estimated by Adhikari et al. [46] at the Central Institute of Freshwater Aquaculture in Orissa, India, where field research was conducted in six ponds. According to the study, evaporation and seepage are important factors influencing water loss in ponds. The amount of water that evaporated from the ponds was determined by weather conditions such as humidity, dry weather, and the depth and size of the ponds. The soil type and the method of estimation were the primary causes of the variation in seepage. Location-specific differences existed in terms of seepage loss and evaporation in their study.

Three studies focused on carp–prawn polycultures in India (Mohanty et al. [47], Mohanty [53], and Mohanty et al. [51]). Mohanty et al. [47] examined the effects of feed restriction on water productivity, sediment loading, and compensatory growth (CG). The impact of feed restriction on sediment loading, water productivity, and the compensatory growth (CG) performance of Indian major carps was investigated by Mohanty [53], while the computed CWUI ranged from 5.43 to 6.58 m³ kg⁻¹. They concluded that the rate of sedimentation increased with increasing apparent feed conversion ratios, while the water exchange requirement, CWUI, and TWU increased with increasing feed input. In the study by Mohanty et al. [47], for various water management practices, the computed CWUIs (m³ kg⁻¹ biomass) were 6.62 (no water exchange), 9.31 (10% water exchange on a monthly basis) and 7.08 (10% water exchange depending on the water quality variables). Evaporation loss contributed significantly to CWU, followed by seepage loss.

Tucker et al. [48] explored catfish aquaculture in the USA. In order to replicate groundwater withdrawal for the foodfish grow-out stage of ictalurid catfish production in northwest Mississippi, they employed a hydrological model., using a 50-year daily record of evaporation and precipitation. The total consumptive water use index was calculated as ~2.7 m³ kg⁻¹, when ground water computations used for producing fingerlings and water used to produce grain-based feedstuffs were combined with simulated ground water use for the most effective set of water conservation techniques in food fish grow-out ponds. They argued that when ponds are not emptied on a yearly basis, the volume of wastewater decreases significantly. Moreover, extending the time between pond emptying gives the biological and chemical processes in the pond more time to eliminate waterborne wastes. According to this study, reducing CWU (through appropriate siting, reusing water, and capturing rainfall) or increasing fish production per hectare are two ways to improve the water use index.

3.1.4. Water Productivity

The net benefits from forestry, crop, livestock, fishery, and mixed agricultural systems divided by the amount of water required for the production of those benefits refer to water productivity (WP) [67]. Economic water productivity refers to the value obtained per unit of water used, whereas physical water productivity refers to the ratio of agricultural output to the amount of water required.

In studies conducted on the hydrological analysis/water budgeting method to assess the water management efficiency, the net total water productivity (NTWP), gross total water productivity (GTWP), and net consumptive water productivity (NCWP) were calculated (USD m⁻³) [42,44,47,51,53]. They investigated the WP variations under different feed management systems and different water management protocols.

According to Mohanty et al. [53] reducing feed input and utilizing the compensatory growth response are effective strategies for raising WP and profitability in aquaculture operations. Due to the limited feed input, refeeding and cyclic food deprivation supported the preservation of water quality, which reduces input costs and total water use, while increases production efficiency.

Das et al. [49] evaluated four cropping patterns of carps in India: intercropping of minor carps and Indian major carps; single stock–multiple harvests; and multiple stocks–multiple harvests. They concluded that in terms of water productivity, after the intercrop of minor carps and the Indian major carp method, the multiple stock–multiple harvests method was the most water-productive cropping pattern.

Pahlow et al. [32] in their study computed the economic consumptive (blue + green) water productivity [USD per m³], which is determined by the ratio of the unit value [USD t⁻¹] to the total green and blue WF [m³ t⁻¹]. According to their study, compared to most marine fishes, crustaceans, and diadromous fishes, freshwater fishes have lower economic blue and green water productivity (exceptions are milkfish, mullet, and red drum because of their relatively low monetary value and their comparatively high WF values). They concluded that enhancing productivity (increased yield per consumed and/or contaminated unit of water) and maximizing feed composition offer opportunities to decrease the WF in aquaculture. This allows for both limiting pressure on freshwater resources and encouraging the growth and well-being of the species raised in aquaculture systems.

4. Discussion

Figure 4 presents the percentages of the reviewed studies based on the type of aquaculture water (freshwater/brackish/marine), the stage of the supply chain (feed production/fish production/fish processing), the water system (closed/semi-closed/open), and the culture technique (monoculture/polyculture). The studies that address the fish processing stage represent approximately 15%, a fact that highlights the need for further research on this relevant stage of the production chain. Freshwater aquaculture dominated the studies under review (70.6%), similar to the monoculture method (79.4%).

The distributions of the aquatic animals under investigation are shown in Figure 5. The most common crops studied are carp (40.6%), tilapia (34.4%), shrimp (28.1%) and catfish (25%).

The CWUI values for the shrimp and carp species in the reviewed studies are shown in Figure 6. The variance in the values is due mainly to the various management strategies and the spatiotemporal factors that affect the water balance of the examined aquaculture systems.

Studies utilizing hydrologic analysis methodologies have investigated different management practices, such as various stocking densities, intensity levels, water exchange and cropping patterns, and their effects on the CWUI values of the examined species.

The majority of brackish water ponds are located in river deltas, where fresh water has historically been lost to the ocean, raising doubts about the alternative destiny of the water [36,54]. Henriksson et al. [36] argued that future studies should determine the scales at which redirecting river flows into brackish water ponds begins to negatively affect local ecosystems and, furthermore, should investigate area-based management and zoning as possible measures to minimize adverse consequences. Pattusamy et al. [43] argued that saline groundwater utilization has enormous potential in the aquaculture sector due to it considering the factors of stocking density, fish species, the quantity of water required and water productivity. The management of brackish water pond farming needs further investigation.

Regarding the WF evaluations, despite the use of the same database, differences were observed between fish species when the WF values from the studies by Yuan et al. [31] and Pahlow et al. [32] were compared. The WF evaluations seem to be affected by the approach taken and the management practices, as, according to the authors, the variations may have occurred due to differences in methodologies and in the feed ingredients of farmed fish. According to Pahlow et al. [32], increasing productivity—a relatively high yield per unit of water used or contaminated—and improving feed composition represent opportunities to decrease the WF of aquaculture. Potential substitutes for soybean, which is currently the most popular ingredient in terrestrial aquafeed, will be crucial in this regard [32,68].

5. Conclusions

According to the studies included in this review, WFN studies are applied mostly at the country/global scale, and most of the papers examine multiple species. On the other hand, LCA research, which is focused on impacts, is used mainly at the farm level or regional scale. Water budgeting/hydrologic analysis is mostly applied to aquaculture ponds. The species of carp (40.6%), tilapia (34.4%), shrimp (28.1%), and catfish (25%) are the most frequently studied crops.

Water consumption varies in space and time (for example, evaporation is dependent mainly on climate conditions). Therefore, an examination at the farm/regional scale is valuable, especially in aquaculture systems such as ponds. The global perspective of the volumetric approach of WFN is also valuable because it provides the possibility to examine the hidden water embodied in the products and to track the water trade among countries. Furthermore, WFs can be used in demand-side research to estimate diet-related water consumption. The different methodologies applied make comparisons of water consumption challenging. In LCA, green water consumption of agricultural systems is usually sidelined, because it is taken into account as a consequence of the land use change. However, in aquaculture, the inclusion of green water consumption is meaningful, especially when agricultural ingredients are used in aquafeed formulations.

In pond management, water accounting and quality monitoring are crucial for improving aquaculture performance and productivity. Improvements in management practices, such as reducing unnecessary water exchange, water reuse or proper management of the stocking density of species, to increase aquaculture sustainability, and water productivity should be further investigated. The demand is increasing, and research on water productivity can provide insights into the proper water management of aquaculture requirements in the context of producing more crops per unit volume of water.

Many of the studies focused only on the indirect water consumed in feed production or examined only direct water consumption. Additionally, differences in the processes that are part of the system boundary have been noted in LCA studies. Only 15% of the studies included in this review addressed the fish processing stage, indicating the necessity for additional research on this pertinent stage of the production chain. Future research should focus on water consumption across the whole supply chain from cradle to grave to obtain more comprehensive and comparable results.

Freshwater aquaculture and the monoculture method dominated the reviewed studies (70.6% and 79.4%, respectively), and further research is needed in the water consumption patterns of brackish and marine aquaculture and the polyculture method, as well.

According to Pahlow et al. [32], the source of the feed’s components used in commercial aquafeed is not recorded in official statistics, despite the fact that aquaculture producing countries rely on imports to varying degrees [32,68]. More comprehensive databases are needed to reduce the uncertainties of studies in that field.

Different inputs such as feed, energy, fertilizers, water refreshments and so on are needed for semi-intensive, extensive and intensive systems, which results in different environmental impacts. RASs reuse and purify culture water to a large extent, thereby using a lot less water than traditional systems. They are used to produce a wide variety of freshwater and saline species. RASs do not experience water seepage, whereas there is considerable water loss in semi-closed agricultural systems such as aquaculture ponds, as a result of evaporation, seepage, and water exchange [54]. RASs have great potential to increase water productivity, and according to the studies included in this review, water consumption in RASs has not received sufficient research attention. Therefore, this topic should be further examined in the future.

The expansion of aquaculture and the resulting increase in demand for plant ingredients in aquafeed may impact the availability of agricultural resources such as freshwater. Therefore, the quantitative impact should be explored in relation to the optimal constitution of aquafeed products (in terms of sustainability and fish health) and to alternative solutions for water-intensive ingredients. Trade-offs exist between various feed ingredients, particularly in regard to replacing marine ingredients with vegetable-based ingredients. Although marine ingredients do not consume water, their availability is extremely restricted. Resource demand calculations are crucial for understanding the environmental impact of aquafeed production. According to Yuan et al. [31], the main elements that could eventually lead to a decrease in the WF are the feeding rates, feed components, fry and harvest fish weights.

A review on the papers focused mainly on the investigation of the water consumption patterns on aquaponics and integrated aquaculture–agriculture would be recommended as future research.

Research on water consumption and the WF of the aquaculture sector has garnered attention only in recent years. The various methodologies, fish species, technologies, management practices, water types, aquaculture systems and spatiotemporal scales make investigating water consumption in the aquaculture sector a challenging task, and further research is essential for promoting sustainable practices in the future.