Assessment of the Offshore Migration of Mussel Production Based on an Aquaculture Similarity Index (ASI)

Abstract

Climate change is increasingly affecting the aquaculture sector, particularly in estuarine systems that support high-value production. In the Galician Rías Baixas, where shellfish farming is a cornerstone of the coastal economy, rising sea temperatures, sea-level rise, and changing precipitation patterns pose significant risks to mussel aquaculture. This study presents a spatially explicit Aquaculture Suitability Similarity Index (ASI) designed to identify alternative cultivation areas that replicate the environmental and logistical characteristics of historically successful mussel farms. The ASI integrates a set of environmental variables (water temperature, salinity, biogeochemical quality, current velocity, and wave height) and technical constraints (depth and distance to port), with factor weights derived via expert elicitation using the Delphi method. Results show that most waters are highly similar to current farming areas, suggesting strong potential for spatial expansion or relocation. In contrast, areas near the mouths of the rías and the adjacent continental shelf show lower suitability due to greater oceanic exposure and associated logistical challenges. The ASI provides a robust, transferable tool to inform aquaculture spatial planning and climate adaptation strategies. Its methodological framework can be adapted to other regions and species, supporting evidence-based decision-making for sustainable aquaculture development.

1. Introduction

Climate change poses significant challenges to the aquaculture sector, particularly in estuarine systems such as the Galician Rías Baixas (northwest Iberian Peninsula), where mussel aquaculture, exemplified by Mytilus galloprovincialis, is one of the most significant economic activities in the coastal region [1,2]. Changes in temperature, circulation, stratification, nutrient input, oxygen content, and ocean acidification are altering the environmental conditions [3,4] that have historically promoted mussel production. Together, these factors demand adaptive strategies to ensure the long-term viability of aquaculture systems in a changing climate, which are vital for global food security and regional economies.

Globally, mussel farming is a major aquaculture sector, with production reaching nearly 1.93 million tons in 2022. Europe accounted for over 22% of this total, and Spain contributed nearly 45% of European output and around 10% of global production. The Spain mussel farming market was valued at approximately USD 432.34 million in 2018, increased to USD 616.18 million in 2024, and is projected to reach USD 997.68 million by 2032, reflecting a compound annual growth rate (CAGR) of 5.78% during the forecast (2025–2032) period (https://www.credenceresearch.com/report/spain-mussel-farming-market?utm_source=chatgpt.com#summary, accessed on 22 September 2025). Remarkably, the autonomous region of Galicia generates 97% of Spain’s mussels, with production almost entirely concentrated in the Rías Baixas, highly productive embayments along the Atlantic coast [1,2]. Mussel production in Galicia has maintained strong market demand, with relatively stable prices over recent years, reflecting the economic resilience and continued significance of this sector both domestically and internationally [1]. Prices exhibit some seasonal fluctuations due to harvest cycles, and a substantial portion of production is exported, underscoring the region’s key role in global mussel markets.

The traditional mussel farming method in Galicia relies on floating rafts (bateas) located in protected coastal waters within the rías. This system has demonstrated long-term productivity and strong socio-economic integration. However, the availability of suitable sites in the innermost parts of the rías is increasingly constrained by a combination of environmental, regulatory, and spatial factors. Environmentally, the rías have a limited ecological carrying capacity, and additional rafts could exacerbate eutrophication, degrade the seabed, and reduce water quality. Legally, existing spatial planning frameworks and environmental regulations restrict or prohibit new concessions, with moratoriums in place in many areas. Furthermore, the rías support multiple and often competing uses—fishing, tourism, recreation, and conservation—which can lead to spatial conflicts and concerns over visual saturation. In addition to these structural limitations, changing climatic stressors further complicate mussel farming. Sea-level rise, rising temperatures, and altered precipitation regimes contribute to increased salinity fluctuations, eutrophication, and the occurrence of harmful algal blooms, all of which negatively impact bivalve populations. Elevated water temperatures have also accelerated metabolic rates and altered reproductive cycles, resulting in shifts in species distribution and abundance [3]. Moreover, enhanced nutrient runoff has intensified eutrophication, fostering harmful algal blooms that place further stress on mussel stocks [5]. Since periods of elevated temperature in the Rías Baixas are likely to become more frequent and prolonged under current climate change scenarios, one potential impact of global warming is an increase in these algal blooms (locally known as red tides).

Numerous studies in recent years have highlighted the potential impacts of climate change in the Rías Baixas, particularly changes in ocean water temperature, upwelling intensity, and extreme precipitation events over the last decades of the 20th century and throughout the 21st century, as well as their potential effects on various species [6,7,8,9,10,11]. Beyond species-specific responses, these studies collectively highlight broader ecological consequences and systemic vulnerabilities that affect the region’s marine and coastal environments. Projected alterations in hydrographic conditions, such as increased stratification, reduced salinity, and more frequent extreme weather events, are expected to reshape habitat suitability, productivity, and the spatial distribution of key benthic and pelagic resources. These environmental shifts may lead to significant reductions in aquaculture and shellfish harvesting areas, particularly in the inner parts of the rías. Mussels are expected to be sensitive to projected increases in sea surface temperature and changes in water column stratification, leading to less favorable conditions for growth, especially in the shallow and inner areas of the Rías Baixas [6].

To support the long-term viability of the mussel aquaculture sector, it is essential to identify new, suitable production sites. Expanding mussel farming into deeper, outer waters represents a promising alternative, though it poses new challenges related to environmental exposure, infrastructure demands, and species resilience. Ideally, the relocation of aquaculture zones should be guided by a thorough understanding of species’ resilience to the environmental conditions at prospective sites [12,13]. This includes knowledge of their thermohaline stress thresholds, responses to nutrient concentrations, and tolerance to marine areas with greater exposure to waves, currents and winds. Unfortunately, the available information does not fully encompass this knowledge, as it primarily derives from mesocosm experiments which, although valuable, cannot replicate the full complexity of natural systems [14,15], including long-term ecological interactions, spatial variability, and local adaptation.

The aim of this study is to develop an Aquaculture Suitability Similarity Index (ASI) to assess the feasibility of offshore mussel farm relocation in the Rías Baixas. The proposed approach identifies areas with environmental and logistical characteristics comparable to those of historically successful mussel production sites. It assesses an area’s suitability for mussel farming by comparing it with productive regions throughout the Rías Baixas, rather than focusing solely on its immediate vicinity. The ASI integrates different technical–environmental drivers that affect aquaculture, including water quality (temperature, salinity, and biogeochemical variables), maritime climate (waves and currents), and logistical constraints (depth and proximity to ports). By quantifying similarity to existing production zones, the ASI provides a practical tool to support site selection and promote adaptive aquaculture strategies under changing climate conditions.

2. Study Area

The Rías Baixas consist of four flooded, incised valleys located on the northwestern Iberian Peninsula (NWIP), at the northern boundary of the eastern North Atlantic Upwelling System [16]. From south to north, these estuarine systems are known as the Ría de Vigo, Ría de Pontevedra, Ría de Arousa, and Ría de Muros–Noia, (Figure 1). They exhibit a northeast–southwest orientation and are bordered by steep hills, having formed through postglacial inundation of fluvial valleys as a result of Holocene sea-level rise [16,17,18].

Mean water depths range from 5–10 m in the innermost zones, where the major rivers discharge into the rías, to 40–60 m at the outer (southwestern) mouths. Apart from the Ria de Muros-Noia, the mouths of these rías are partially protected by small islands that form part of the Atlantic Islands of Galicia National Park [17,19]. Key physical and aquaculture-related characteristics of each ria are summarized in

Table 1. Main characteristics of the Galician Rías Baixas, including the number and surface area of mussel farming zones.

Ría	Volume (km3)	Surface (km2)	Length (m)	Mean Mouth Depth (m)	Farming Zones (n)	Aquaculture Area (km2)	Main River (s)	Mean River Discharge (m−3 s−1)
Vigo	3.2	176	31	50 (S), 25 (N)	11	5.6	Verdugo-Oitaven	17
Pontevedra	3.5	141	22	60 (S), 15 (N)	11	4.1	Lérez	25.6
Arousa	5.4	230	25	55 (S), 5 (N)	23	37.4	Ulla & Umia	79.3 & 16.3
Muros-Noia	2.1	125	13	50	3	1.9	Tambre	54.1

.

Hydrodynamically, the Rías Baixas function as partially mixed estuaries exhibiting a classical two-layer estuarine circulation: a surface outflow of freshwater overlain on a bottom inflow of saline oceanic water [20]. This circulation is strongly influenced by wind-driven upwelling and downwelling events [21]. Equatorward winds promote coastal upwelling, intensifying the positive estuarine circulation and enhancing the inflow of cold, nutrient-rich waters at depth [22,23,24]. Upwelling supports high productivity by delivering nutrients to deeper layers, although vertical stratification may limit nutrient availability near the surface. Conversely, poleward winds induce downwelling, which can weaken or reverse circulation, trap surface waters, and facilitate freshwater intrusion into deeper layers [24,25,26]. These dynamics lead to offshore nutrient depletion while maintaining productivity in inner zones due to riverine nutrient input.

The tidal regime is semi-diurnal and mesotidal, characterized by two tides per day with amplitudes ranging from 1.3 m (neap) to 3.4 m (spring). Tidal forcing is the dominant energy source driving short-term water exchange within the rías.

Primary production in the Rías Baixas is governed by nutrient dynamics, with peak rates exceeding 10 g C m⁻² d⁻¹ during spring and summer phytoplankton blooms. Approximately 55% of this primary production is transferred to higher trophic levels [27,28]. Seasonal phytoplankton biomass peaks reach ~8 mg m⁻³ in spring and autumn, decline to <1 mg m⁻³ in winter, and average ~5 mg m⁻³ during summer [29,30].

The region exhibits an oceanic climate with marked seasonality and relatively dry summers [31]. Annual average air temperatures range between 14–15 °C, with seasonal variability from ~10°C in winter to ~20 °C in summer, typically decreasing toward the coast due to maritime influence [32]. Atmospheric circulation is modulated by the seasonal position of the Azores High [33]. In winter, its southward displacement toward northwest Africa, coupled with a low-pressure system over Iceland, favors southwesterly winds. In spring and summer, a northward shift in the Azores High drives persistent NNE winds along the Galician coast, promoting frequent upwelling events [19].

The nutrient enrichment driven by these upwelling events sustains the high biological productivity of the Rías Baixas, supporting diverse marine ecosystems and numerous commercially important species. Among these, Mytilus galloprovincialis (Mediterranean mussel) is particularly notable, with the Rías Baixas contributing approximately 21% of global production [3] as mentioned introduction section. This significant output is primarily attributed to the region’s favorable environmental conditions, including high nutrient availability, moderate temperatures, and relatively sheltered waters.

Mussel farming in the region is predominantly carried out using traditional floating raft systems, which consist of large, buoyant platforms anchored within the rías and equipped with suspended ropes for mussel attachment and growth. These structures are arranged along the margins of the rías and have become a defining feature of the Galician coastal landscape (Figure 1).

Currently, mussel farming occupies approximately 5.6 km² across 13 farming zones in the Ría of Vigo, 4.1 km² in 9 zones in the Ría of Pontevedra, 37.4 km² in 23 zones in the Ría de Arousa, and 1.92 km² in 3 zones in the Ría de Muros–Noia, as summarized in Table 1. The spatial distribution of these zones reflects both the environmental suitability and the historical trajectory of aquaculture development in each ría.

3. Materials and Methods

An Aquaculture Suitability Index (ASI) was developed to evaluate the environmental potential for mussel cultivation across the Galician Rías Baixas. This index integrates a broad range of technical and environmental variables crucial to aquaculture success. Key factors include water quality parameters such as water temperature, salinity, and chlorophyll, as well as hydrodynamic conditions (wave height and currents), topographic constraints (water depth and proximity to ports), and regulatory restrictions (protected areas or marine traffic zones). These variables were selected based on their proven relevance to mussel health, growth performance, operational stability, and economic feasibility.

The ASI was constructed by comparing conditions at each assessment location against those observed in 582 areas where mussel production has historically been successful throughout the Rías Baixas. This comparative approach ensures that site evaluation is grounded in empirical evidence and aligned with the region’s most productive aquaculture zones.

3.1. Databases

3.1.1. Thermohaline Variables

Water temperature (T) and salinity (S) from MOHID were retrieved from MeteoGalicia’s THREDDS server (https://mandeo.meteogalicia.es/thredds/catalogos/DATOS/ARCHIVE/MOHID/catalog_hist.html, accessed on 22 September 2025). MeteoGalicia is the official meteorological agency of the Galician regional government (Spain), offering high-resolution numerical forecasts and climate data. The model outputs used in this study cover the 2014–2024 period at hourly resolution, with a horizontal grid spacing of approximately 0.036° × 0.0025° (approximately 300 m).

3.1.2. Hydrodynamic Variables

Current velocities (C) were obtained from the operational implementation of the MOHID hydrodynamic model (http://www.mohid.com/, accessed on 22 September 2025), which is executed daily by MeteoGalicia. This model is forced by outputs from the ROMS, WRF, and SWAT models. Although MOHID provides a 72-h forecast, only the first 24 h were used in this study to reduce forecast error. Data are available at hourly resolution on a spatial resolution of 0.0036° in longitude and 0.0027° in latitude (~300 m). The latitude increment is effectively constant across the domain. Vertically, with a hybrid vertical coordinate system, time-varying sigma layers in the upper 8 m and fixed layers below. The dataset spans the period from 2014 to 2024 with hourly temporal resolution. Data is available via the MeteoGalicia’s THREDDS server.

Significant wave height (Hs) was retrieved from MeteoGalicia’s SWAN model output, available via the THREDDS server (http://mandeo.meteogalicia.gal/thredds/catalog/modelos/SWAN_HIST/galicia/catalog.html, accessed on 22 September 2025) at hourly resolution for the period 2014–2024. MeteoGalicia runs the third-generation SWAN model operationally [34] configured on an unstructured grid with resolutions down to 75 m in coastal and estuarine zones. The model accounts for bottom friction, shoaling, and depth-induced breaking. The wave spectrum is discretized using a relative frequency resolution of Δf/f = 0.1 across a frequency range of 0.0521 to 1 Hz, and partitioned into 36 directional bins. The full configuration of the model is detailed in Table 1 of [35]. Although each daily forecast extends over four days, only the first 24 h were used here to minimise the accumulation of forecast error.

3.1.3. Biogeochemical Variables

Biogeochemical data, including chlorophyll, dissolved oxygen, and pH, were obtained from the Copernicus Marine Service. Two databases were evaluated: (i) the Atlantic-Iberian Biscay Irish Ocean Biogeochemistry Hindcast (IBI_MULTIYEAR_BGC_005_003; [36]), which provides daily data from 1 January 1993 to 28 December 2021 at a horizontal resolution of 0.083° × 0.083°, with 50 vertical levels, covering a domain between 26–56° N and 19° W–5° E [36]; and (ii) the Atlantic-Iberian Biscay Irish Ocean Biogeochemical Analysis and Forecast (IBI_ANALYSISFORECAST_BGC_005_004; [37]), which offers higher spatial resolution (0.028° × 0.028°) with 48 vertical layers at daily frequency but is only available from 1 January 2020 onward [37]. Given the study’s focus on narrow areas, where fine spatial resolution is essential, the latter database (IBI_ANALYSISFORECAST_BGC_005_004) was selected for analysis.

3.1.4. Topographic Variables

Topographic parameters included water depth and distance to the nearest port, which are proxies for installation and maintenance costs.

Water depth was extracted from the Cartesian grid used by MeteoGalicia’s hydrodynamic model, which has a spatial resolution of approximately 300 m. This locally calibrated bathymetric dataset provides more precise depth estimates than global products such as GEBCO_2024 (~450 m), making it more suitable for regional-scale aquaculture planning. Apart from providing higher spatial resolution, the grid used by MeteoGalicia benefits from local bathymetric surveys, making it more suitable for the present analysis.

Distance to the nearest port was calculated using data from Portos de Galicia (https://portosdegalicia.gal/es/web/portos-de-galicia/locportos, accessed on 22 September 2025), which is the agency responsible for the planning, construction, operation, maintenance, and development of ports under the jurisdiction of the regional government (Xunta de Galicia, Santiago, Spain). In addition, the three largest ports (Vigo, Marin, and Vilagarcía), which are managed by the Spanish Port Authority (https://www.puertos.es/, accessed on 22 September 2025), were also considered. A total of 62 ports of varying sizes and facilities were included in the present analysis.

3.1.5. Administrative Limitations

Regulatory restrictions that may limit aquaculture activities were integrated into the ASI as exclusion layers. The study area is subject to multiple administrative restrictions related to land use, existing occupations, and other maritime activities that may conflict with aquaculture development. In Galicia, raft-based aquaculture is governed by various legal frameworks and spatial planning instruments that define zones where such activities are explicitly prohibited. These include protected natural areas such as national parks and marine reserves, for example, the Atlantic Islands of Galicia National Park, as well as other marine protected areas designated to conserve biodiversity and natural ecosystems. Additionally, restrictions apply to ecologically sensitive zones classified under the Natura 2000 network, such as Special Protection Areas for Birds (SPAs) and Sites of Community Importance (SCIs), where aquaculture operations are heavily regulated or banned.

Navigation zones, including port vicinities, maritime transit corridors, and vessels maneuver areas, were also excluded to prevent interference with maritime traffic and ensure navigation safety. Coastal areas designated for recreational or tourism purposes, such as beaches and leisure-use zones, are also excluded to preserve aesthetic values and maintain water quality. Furthermore, certain areas are reserved exclusively for traditional and artisanal fisheries, where aquaculture infrastructure is limited to prevent conflicts and support the sustainability of fishery resources.

Lastly, environmental protection zones established under Plan de Ordenación del Litoral (POL), Planes de Ordenación del Espacio Marítimo (POEM) and other spatial planning instruments, which impose additional limitations aimed at safeguarding coastal ecosystems and maritime resources, were considered. These restrictions aim to balance the development of aquaculture and the conservation of the marine environment, as well as the sustainability of marine resources in Galicia.

This spatial information was obtained from the official geographic viewer of the regional government (https://mapas.xunta.gal/visores/basico/, accessed on 22 September 2025) and incorporated as exclusion layers in the spatial analysis.

3.2. Methods

To evaluate the spatial suitability for mussel aquaculture in the Rías Baixas, we developed the Aquaculture Suitability Similarity Index (ASI). This composite index quantifies the degree of similarity between environmental conditions at any given location and those observed at established mussel farming sites. The ASI comprises seven sub-indices: Temperature (T_Index), Salinity (S_Index), Biogeochemical Quality (BGQ_Index), Current (C_Index), Wave Height (WH_Index), Water Depth (WD_Index) and Distance to Port (DP_Index). These sub-indices integrate a range of relevant variables, including temperature, salinity, chlorophyll-a concentration (Clh), pH, dissolved oxygen (O₂), current velocity, significant wave height (Hs), water depth, and distance to the nearest port. All variables were interpolated onto a common spatial grid and normalized to a common scale to ensure comparability. The relative importance of each sub-index was then determined through expert elicitation using the Delphi method, enabling the calculation of a weighted suitability index at each grid point. The methodological steps followed in the construction of the ASI are detailed below.

3.2.1. Reducing Data to a Common Grid

Environmental variables used in this study were sourced from diverse datasets, each differing in spatial resolution and grid structure. To enable consistent spatial comparison and analysis, all variables were interpolated onto a unified grid. For this purpose, we adopted the Cartesian grid, with a spatial resolution of approximately 300 m, used for the hydrodynamic model run by MeteoGalicia (see Section 3.1).

Salinity (S), temperature (T), and current speed (C) were vertically averaged over the upper 12 m of the water column, in alignment with the typical length of mussel farming ropes suspended from the rafts [38].

Wave height data (significant wave height, Hs), originally provided on an unstructured grid, was interpolated onto the targeted grid using a Delaunay triangulation-based method for scattered data [39].

Biogeochemical variables, chlorophyll-a (Chl), dissolved oxygen (O₂), and pH, were initially available on a coarser spatial resolution and were interpolated onto the targeted grid using bilinear interpolation [40].

In addition, for each node of the target grid, the distance to the nearest port was computed to support analysis of aquaculture site accessibility and logistical feasibility.

3.2.2. Normalizing Variables

Each environmental or technical variable has distinct magnitudes and units, making normalization essential prior to their integration into a single index (ASI) at each grid point. Normalization procedures were designed to reflect the degree of similarity between each location in the study area and those where mussel aquaculture is currently practiced.

For S, T, C, O₂, pH, and Chl, normalization was based on a statistical measure of distributional similarity known as the Overlapping Percentage (OP) [41]. For each variable, the normalized frequency distributions at every grid point, i $=$ $(i_x,i_y)$, were compared to those at established aquaculture locations, j $=(j_x,j_y)$. The OP was calculated as:

$${OP}_{i,j}\left(\mathrm{\%}\right)=100\times \sum _{k=1}^{N}minimum({\theta }_{i,k},{\theta }_{j,k})$$

(1)

where ${\theta }_{i,k}\mathrm{a}\mathrm{n}\mathrm{d}{\theta }_{j,k}$ represent the normalized frequency of the variable at grid points i and j for the bin k, with N = 20 total bins. The higher OP indicates a greater similarity between point i and aquaculture site j.

The maximum overlap for each point i was then calculated as:

$${OP}_i^{max}=\underset{j}{\max }{OP}_{i,j}$$

(2)

Following procedures in [42,43], these variables were translated into normalized sub- indices on a [0–1] scale, where 1 indicates optimal similarity. In particular, T, S, and C variables were normalized using the criterion described in

Table 2. Normalization criterion for S, T, and C sub-indices based on the maximum overlap percentage (OP) for each grid point. Normalized values were defined on a [0–1] scale, where optimal values correspond to 1 and the least favorable values correspond to 0.

Normalized Value $({\mathit{T}}_{\mathit{i}}^{\mathit{n}}, {\mathit{S}}_{\mathit{i}}^{\mathit{n}} \mathbf{o}\mathbf{r} {\mathit{C}}_{\mathit{i}}^{\mathit{n}})$	${\mathit{O}\mathit{P}}_{\mathit{i}}^{\mathit{m}\mathit{a}\mathit{x}}$(%) (for S, T or C)
0	<50
0.5	(55, 50]
0.55	(60, 55]
0.6	(65, 60]
0.65	(70, 65]
0.75	(75,70]
0.8	(80, 75]
0.85	(85, 80]
0.9	(90, 85]
0.95	(95, 90]
1	≥95

.

Following [44], the three most relevant biogeochemical variables for mussel aquaculture, O₂, pH, and Chl, were further aggregated into a Biogeochemical Quality Index (BGQIndex) using weighted averages:

$${BGC}_i=0.115\times {\widehat{O_2}}_i+0.4\times {\widehat{pH}}_i+0.486\times {\widehat{Chl}}_i$$

(3)

where hatted variables (e.g., $\widehat{O_2}$) correspond to normalized values at grid point i.

Significant wave height (Hs) normalization deviated from the OP approach due to its limitations in capturing tail behavior, which is critical for assessing structural risk and mussels detaching from the ropes. Percentile values (50th, 90th, and 99th) were computed for all current mussel farming sites and their maxima (denoted as ${Hs}_{max}^{50},{Hs}_{max}^{90}\mathrm{o}\mathrm{r}{Hs}_{max}^{99}$) were identified. This approach allows for determining, for each percentile, the most extreme wave conditions currently tolerated by mussel farming infrastructure.

The normalization criterion applied to each percentile, as summarized in

Table 3. Hs normalization criterion for each grid point i was defined on a [0–1] scale, where optimal values correspond to 1 and the least favorable values correspond to 0.

Normalized Value (${\widehat{{\mathit{H}\mathit{s}}^{\mathit{p}}}}_{\mathit{i}}$)	${\mathit{H}\mathit{s}}_{\mathit{i}}^{\mathit{p}}/{\mathit{H}\mathit{s}}_{\mathit{m}\mathit{a}\mathit{x}}^{\mathit{p}}$ (p = 50, 90 or 99)
0	>2
0.5	(1.9, 2]
0.55	(1.8, 1.9]
0.6	(1.7, 1.8]
0.65	(1.6, 1.7]
0.7	(1.5, 1.6]
0.75	(1.4, 1.5]
0.8	(1.3, 1.4]
0.85	(1.2, 1.3]
0.9	(1.1, 1.2]
0.95	(1, 1.1]
1	≤1

, is based on the assumption that the wave conditions at a grid point (i), are considered optimal when the wave height at that location for a given percentile ${Hs}_i^p$ (being p = 50, 90 or 99) is less than or equal to the corresponding maximum value ${Hs}_{max}^p$ observed at established aquaculture locations (j). Suitability decreases as wave energy increases beyond these thresholds.

The Wave Height Index (${WH}_{index}^i$) was then calculated by averaging the normalized values corresponding to every percentile as follows:

$${WH}_{Index}^i={\displaystyle \frac{1}{3}}\left({\widehat{{Hs}^{50}}}_i+{\widehat{{Hs}^{90}}}_i+{\widehat{{Hs}^{99}}}_i\right)$$

(4)

where hatted variables (e.g., $\widehat{{Hs}^{50}}$) correspond to normalized values at grid point i.

The technical factors (depth and distance to port) were treated as static, as they do not vary over time. As with previous variables, normalization was based on existing mussel farming practices in Galician Rías. For water depth, only locations with mean tidal depths greater than 20 m were considered, ensuring sufficient clearance below the typical 12-m cultivation ropes, even under spring low tide conditions, to allow for adequate water circulation beneath the structures. Locations between 20 and 40 m were considered optimal, reflecting the depth range where over 95% of existing mussel rafts are deployed. Suitability decreased progressively for deeper waters, with depths greater than 100 m assigned the lowest value (0) on the normalized scale (

Table 4. Normalization criterion for water depth index (${WD}_{Index}^i$) at each grid point i defined on a [0–1] scale, where optimal values correspond to 1 and the least favorable values correspond to 0.

Normalized Value (${\mathit{W}\mathit{D}}_{\mathit{I}\mathit{n}\mathit{d}\mathit{e}\mathit{x}}^{\mathit{i}}$)	Water Depth (m)
0	>100
0.5	(85, 90]
0.55	(80, 85]
0.6	(75, 80]
0.65	(70, 75]
0.7	(65, 70]
0.75	(60, 65]
0.8	(55, 60]
0.85	(50, 55]
0.9	(45, 50]
0.95	(40, 45]
1	≤40

).

Similarly, the proximity to port facilities was normalized by assuming that logistical efficiency is greatest at shorter distances (

Table 5. Normalization criterion for distance to the nearest port (${DP}_{Index}^i$) at each grid point i defined on a [0–1] scale, where optimal values correspond to 1 and the least favorable values correspond to 0.

Normalized Value (${\mathit{D}\mathit{P}}_{\mathit{I}\mathit{n}\mathit{d}\mathit{e}\mathit{x}}^{\mathit{i}}$)	Distance to the Nearest Port (km)
0	>10
0.5	(9.5, 10]
0.55	(9, 9.5]
0.6	(8.5, 9]
0.65	(8, 8.5]
0.7	(7.5, 8]
0.75	(7, 7.5]
0.8	(6.5, 7]
0.85	(6, 6.5]
0.9	(5.5, 6]
0.95	(5, 5.5]
1	≤5.0

). Sites within 5 km of the nearest port were considered optimal (assigned a value of 1), as this threshold corresponds to the 95th percentile of distances observed for existing mussel farms. Suitability values decreased incrementally with increasing distance, reaching zero for sites located more than 10 km from the nearest port.

3.2.3. Weighing the Different Contributions (Delphi)

The relative contribution of each sub-index to the final ASI was determined using a weighted average. Weights were obtained via the Delphi method [42,43], a structured and anonymous participatory process designed to collect informed judgments from a panel of experts and achieve consensus. This approach is consistent with methodologies previously employed in spatial suitability analyses for offshore wind and wave energy developments [45].

In this study, a panel of 35 experts in aquaculture participated in the Delphi process to assign relative weights to each normalized index. The final ASI score for each grid cell (i) was calculated as:

$${ASI}^i=\sum _{z=1}^M{\overline{w}}_z{I}_z^i$$

(5)

where M is the total number of sub-indices, $I_z^i$ is the index z at grid point i, and ${\overline{w}}_z$ is the mean expert-assigned weight.

The mean and standard deviation of the weight assigned for sub-index z was calculated according to:

$${\overline{w}}_z={\displaystyle \frac{\sum _{k=1}^N{w}_{z,k}}{N}}$$

(6a)

$$\sigma ({\overline{w}}_z)=\sqrt{{\displaystyle \frac{\sum _{k=1}^N{\left|w_{z,k}-{\overline{w}}_z\right|}^2}{N}}}$$

(6b)

where w_z,k is the weight that an individual expert k assigned to the sub-index z, ${\overline{w}}_z$ is the mean weight for that sub-index, and N is the total number of experts (25).

The resulting average weights and standard deviations are presented in

Table 6. Average weight coefficients and standard deviation for each sub-index, based on responses from the 25 experts consulted.

Index	${\overline{\mathit{w}}}_{\mathit{z}}$$\pm \mathit{\sigma }\left({\overline{\mathit{w}}}_{\mathit{z}}\right)$
TIndex	0.151 ± 0.047
SIndex	0.124 ± 0.057
BGQIndex	0.164 ± 0.034
CIndex	0.134 ± 0.044
WHIndex	0.166 ± 0.034
WDIndex	0.138 ± 0.045
DPIndex	0.124 ± 0.052

.

Among the sub-indices, the BGQ_Index (biogeochemical quality) and WH_Index (wave height) received the highest weights, underscoring the central role of biogeochemical and hydrodynamic conditions in determining site suitability. In contrast, DP_Index (distance to port) and S_Index (salinity) were assigned the lowest weights, indicating that while these factors are relevant, they exert less influence on overall suitability compared to other environmental variables. Variables such as T_Index (temperature), C_Index (current), and WD_Index (water depth) received moderate weights (ranging from 0.134 to 0.151), reflecting their consistent but moderate contribution to aquaculture viability. Notably, the relatively low standard deviations across all variables suggest a strong consensus among experts regarding their importance.

The final ASI categories were defined based on these expert-derived weights, using the classification scheme detailed in

Table 7. Classification of the Aquaculture Suitability Similarity Index (ASI).

Categorization of Suitability	ASI
Unsuitable	<0.8
Marginal	[0.8–0.9)
Optimal	[0.9–1.0]

.

3.2.4. Sensitivity Analysis of the ASI

A sensitivity analysis was performed to assess the robustness of the ASI against uncertainty in the expert-derived weights. This analysis aimed to quantify the impact of uncertainties in expert judgment on the final index and was conducted at each grid point using a Monte Carlo approach, following a procedure similar to that proposed by [45,46].

In the first step, each relative weight was perturbed randomly according to:

$${\overline{w}}_z^{i,l}={\overline{w}}_z+R^{i,l}\times \sigma \left({\overline{w}}_z\right)$$

(7)

where ${\overline{w}}_z^{i,l}$ is the mean weight for sub-index z at point i and Montecarlo’s realization number l, ${\overline{w}}_z$ and $\sigma \left({\overline{w}}_z\right)$ are the mean and standard deviation of the weights assigned for sub-index z, which were calculated following Equation (6a,b), and $R^{i,l}$ is a uniformly distributed random number between −1 and +1 that depends on the grid point and the realization.

After the random perturbation, weights were renormalized to ensure that they summed to one, since perturbation can violate the normalization condition, following

$${\overline{w}}_z^{i,l}={\displaystyle \frac{{\overline{w}}_z^{i,l}}{\sum _{z=1}^M{\overline{w}}_z^{i,l}}}$$

(8)

where M is the total number of sub-index. This renormalization enforces the condition

$$\sum _{z=1}^M{\overline{w}}_z^{i,l}=1$$

(9)

In the third step, perturbed ASI was calculated for each Monte Carlo realization l using the perturbed and normalized weights:

$${ASI}^{i,l}=\sum _{z=1}^M{\overline{w}}_z^{i,l}{v}_z^i$$

(10)

where $v_z^i$ is the normalized value of sub-index z at point i. Finally, ASI was calculated by averaging the l realizations following:

$${ASI}^i={\displaystyle \frac{1}{L}}\sum _{l=1}^L{ASI}^{i,l}$$

(11)

where L = 1,000,000 is the total number of Monte Carlo realizations.

4. Results

The spatial distribution of the normalized index used to construct the Aquaculture Suitability Similarity Index (ASI) was spatially mapped across the Rías Baixas: Vigo and Pontevedra rias (Figure 2), Ría de Arousa (Figure 3), and Rías de Muros-Noia (Figure 4).

Water temperature (T_Index, Figure 2a, Figure 3a and Figure 4a), salinity (S_Index, Figure 2b, Figure 3b and Figure 4b), biogeochemical quality (BGQ_index, Figure 2c, Figure 3c and Figure 4c), and current velocities indices (C_Index, Figure 2d, Figure 3d and Figure 4d) exhibited similar spatial patterns across all regions, generally reaching values close to 1. This indicates that environmental conditions closely aligned with those observed at established mussel aquaculture sites (green areas). Slight reduction in index values, typically remaining above 0.8, was observed in outer areas, reflecting still favorable, though marginally lower, suitability (yellow areas).

Among these indices, the BGQ_index showed the greatest spatial variability, with a patchier distribution across all three zones (Figure 2c, Figure 3c and Figure 4c). However, values remain above 0.8, denoting overall favorable water quality conditions. Salinity index showed a localized reduction in the innermost part of the Ría de Arousa (Figure 3b), where the main river flows into the estuary.

Wave height index (WH_Index, Figure 2e, Figure 3e and Figure 4e) reaches maximum values (1) in the inner, sheltered zones of the rías and progressively decreases toward outer areas, where wave exposure increases. Similarly, WD_Index (Figure 2f, Figure 3f and Figure 4f) shows the highest values nearshore, where depths are most suitable for mussel raft anchoring (20–40 m), while values decreased towards the outer part and the central channel of the rías, which is aligned with the longitudinal axis. Notably, WD_Index values ranged between 0.5 and 0.8 along the central channel of the Ría de Arousa, due to its greater depth, while values below 0.5 are observed outside the rías, on the adjacent continental shelf.

The DP_Index (Figure 2g, Figure 3g and Figure 4g) reflects the distribution of the maritime infrastructure. High suitability values are concentrated in the inner part of the rías, where most of the port infrastructure is located. In contrast, the outer areas exhibit markedly lower scores (<0.5), reflecting increased logistical challenges associated with remoteness.

As a result of the combined influence of these indices, the expert-derived weights obtained through the Delphi method, and the averaging of one million Monte Carlo repetitions, the ASI classification revealed high suitability for mussel aquaculture across much of the studied area (Figure 2h, Figure 3h and Figure 4h). In the four Rías Baixas (rias of Vigo and Pontevedra, Figure 2h, Ría de Arousa, Figure 3h, and Ría de Muros, Figure 4h), the optimal zones extend from the inner part to the boundaries of the Atlantic Islands National Park. Suitability decreased sharply beyond the limits of the rías, especially in the adjacent continental shelf, near the boundaries of the Atlantic Islands National Park, and in offshore areas exposed to higher wave energy and greater depths.

Figure 5 presents the distribution of the optimal mussel-farming areas after applying exclusion criteria for areas under administrative constraints, such as navigation channels, vessel-maneuvering zones, and the Atlantic Islands National Park, as well as the footprint of existing aquaculture areas. The resulting map highlights newly identified areas with high suitability (in green), predominantly located in underutilized regions of the middle and outer zones across all four rías.

5. Discussion

Climate change poses an increasing threat to traditional mussel aquaculture activities developed in the Rías Baixas, particularly in the inner zones [6] where cultivation is currently concentrated (Figure 1, green polygons). Rising sea temperatures and the increasing frequency of extreme events, such as marine heatwaves [47,48], are placing mussel populations under significant thermal stress. Although mussel larvae are planktonic and capable of dispersal, their survival and successful recruitment remain strongly influenced by local environmental conditions, including temperature extremes. As sessile adults, mussels cannot relocate to avoid stress, making both life stages vulnerable to these changes, which can negatively affect growth, reproduction, and survival [49]. In this context, exploring alternative cultivation areas within the Rías Baixas that offer greater environmental resilience may contribute to the long-term sustainability of the sector.

The Aquaculture Suitability Similarity Index (ASI) developed in this study provides a spatially explicit tool to support aquaculture planning. The weighting of each factor of ASI was determined through expert consensus using the Delphi method, ensuring the index reflects both scientific understanding and local industry knowledge. The ASI thus facilitates the identification of areas with high potential for either aquaculture expansion or relocation.

Traditional multicriteria evaluation approaches or habitat suitability models, which often rely on species-specific environmental thresholds derived from controlled experiments, e.g., [7,50], provide useful baseline assessments; however, they often overlook the complex interactions among environmental and logistical factors critical to long-term viability [51,52,53]. In addition, establishing suitability limits is challenging and often relies on mesocosm experiments that analyze only a few variables [14,15,49,54]. In contrast, the Aquaculture Suitability Index (ASI) integrates similarity to nearby areas where mussel culture is established with technical feasibility, avoids the challenge of defining arbitrary suitability limits, and captures more holistically the complex interplay of ecological and logistical factors influencing site viability

Results from the ASI classification show that most waters in the Rías Baixas are highly suitable for mussel aquaculture. In particular, the inner zones of the Vigo, Pontevedra, Arousa, and Muros–Noia rias show high ASI values, closely matching the environmental profiles of current production areas. However, suitability decreases significantly beyond the mouths of the rías and the near continental shelf due to increased wave exposure, greater depths, and logistical constraints.

Even when applying a strict threshold (ASI > 0.90), the analysis identifies extensive areas suitable for mussel aquaculture expansion, approximately 160 km² in the rias of Vigo and Pontevedra, 92 km² in the Ría de Arousa, and 71 km² in the Ría de Muros-Noia. These areas substantially exceed the current farmed surfaces, which occupy approximately 10 km², 37 km², and 2 km², respectively. All suitable zones are located at depths greater than 20 m, thereby avoiding conflicts with recreational and port activities near the shore and ensuring adequate clearance beneath mussel ropes (typically 12 m). This clearance promotes adequate water exchange and oxygenation, minimizes sediment resuspension and net fouling, prevents damage from benthic organisms and uneven substrate, accommodates tidal and current-driven movements, and facilitates routine inspection and maintenance [55,56,57].

Previous studies have shown that climate change will significantly alter the thermal suitability of inner and middle zones of the Rías Baixas during the 21st century [6,7,8,9,11]. In particular, Des et al. [6] found that projected changes in water temperature and stratification will negatively affect Mytilus galloprovincialis growth. From current comfort values of 80–100%, they project a sharp decline down to 20–40% by the end of the century in the shallow areas and the inner part of the rías, and to 40–80% in their middle sections, only outer areas are expected to maintain conditions comparable to present-day levels. These projected thermal refuges overlap with the areas identified in this study as highly suitable under current conditions, indicating their potential suitability for the future reallocation of mussel farming activities.

Analysis of the ASI components highlights the relative importance of each factor. Wave height (WH_Index) and BGQ_Index received the highest average weights (0.166 ± 0.034 and 0.164 ± 0.034, respectively), underscoring the critical role of hydrodynamic exposure and water quality in site selection. The significant decline in WH_Index values in the outer part of the rías and the continental shelf was a major factor in their classification as unsuitable, due to increased wave energy and associated exposure that can affect mussel raft stability and operational safety.

Although DP_Index had the lowest assigned weight (0.124 ± 0.052), and WD_Index a moderate one (0.138 ± 0.045), their low values in the outer part of the rías and the continental shelf also reduced overall ASI scores, reflecting the increased cost and logistical difficulty of operating in remote, deeper areas.

Although moderately weighted (0.151 ± 0.047 for T_Index and 0.124 ± 0.057 for S_Index), temperature and salinity exhibited consistently favorable values across most of the study area, thereby contributing positively to the classification of many nearshore regions as highly suitable for aquaculture. Notably, salinity index only showed localized reductions in the innermost part of the Ría de Arousa (Figure 3b), likely due to freshwater inflows from the Ulla River, which has an average discharge of 79.3 m³ s⁻¹ [19]. These inflows, particularly pronounced during winter, T_Index and S_Index play a vital role in delineating suitability, as both parameters are critical for ensuring the physiological tolerance and optimal growth performance of cultured species.

Finally, C_Index, with a weight of 0.134 ± 0.044, showed stable and high values throughout the rías. This index reflects a stable hydrodynamic environment that supports mussel filtration and waste dispersion, reinforcing the high suitability of inner zones.

Importantly, if only thermohaline and biogeochemical variables were considered, both inner and outer zones would appear suitable for relocating the mussel rafts. However, it is the technological and logistical challenges, particularly related to wave exposure and site depth, that constrain offshore expansion. Addressing these constraints will require the development of novel anchoring techniques and floating structures capable of damping current and wave-induced motion [58,59,60,61,62,63,64].

From an integrated perspective, the results suggest that while offshore expansion of mussel farming may be viable in some locations, it remains limited by environmental exposure, logistical challenges, and biological tolerance thresholds. While transitioning toward offshore production could alleviate pressures in overused sheltered areas and help adapt to climate change, it will require addressing multiple techno-economic challenges. The ASI framework developed in this study provides a useful decision-support tool for guiding the spatial reallocation of mussel farming under climate change, offering a transparent and reproducible method to identify alternative sites that mirror the environmental and operational profiles of successful farming areas.

Despite the robustness of the ASI framework and its practical applicability, several limitations should be considered. First, the index is based on present-day environmental conditions and does not account for projected future changes in oceanographic variables under climate change scenarios. Therefore, for long-term planning, coupling the index with oceanographic projections will be necessary. Second, although variable weighting was derived through expert consensus using the Delphi method and the influence of uncertainty was mitigated through one million Monte Carlo repetitions, the process remains inherently subjective. Weight assignments may vary depending on the composition, expertise, or regional focus of the expert panel. Finally, the method is primarily intended to support decision-making for the optimization and reallocation of existing culture sites, rather than to assess the feasibility of introducing non-native species to a given location.

In summary, the ASI developed in this study provides a valuable tool for assessing potential areas for mussel aquaculture under current environmental conditions in the Rías Baixas. By integrating both environmental and technical variables with expert-informed weighting, the ASI offers a geographically detailed and viable framework to guide site selection, particularly in response to climate-driven stressors. The results indicate that while inner zones remain optimal, certain outer areas may also hold potential for expansion, provided that the biological and infrastructural challenges associated with more exposed conditions are addressed.

The future use of ASI could also benefit from including technological innovations which may broaden the range of viable farming sites. For instance, hybrid offshore platforms integrating aquaculture structures with wave energy converters (WECs) are being explored as multifunctional solutions in exposed marine environments. Such systems could simultaneously protect mussel rafts from high-energy wave conditions and provide a renewable source of energy to support farm operations, including monitoring, maintenance, and product handling. Incorporating the potential of these technological advances into spatial planning tools like the ASI would enable a more forward-looking assessment of offshore aquaculture feasibility, aligning site selection not only with environmental and logistical constraints but also with the evolving technological landscape of the sector.

6. Conclusions

This study presents an Aquaculture Suitability Similarity Index (ASI) developed to evaluate the potential for offshore re-allocation of mussel farming in the Rías Baixas, a region of critical economic and cultural importance for bivalve aquaculture. The ASI integrates key environmental (water temperature, salinity, biogeochemical quality, currents, wave exposure) and technical (water depth, distance to port) factors, with variable weights assigned through expert consensus using the Delphi method. Unlike traditional habitat suitability models, the ASI evaluates spatial similarity to currently productive sites, offering a more holistic and operationally grounded approach to site selection.

Results indicated that most of the zones under study are highly suitable for mussel cultivation; however, suitability declines sharply beyond the mouth of the rías, primarily due to increased wave exposure, greater water depth, and logistical limitations. These findings suggest that while offshore expansion is feasible, its success depends on site-specific conditions and careful planning, including the prioritization of areas with favorable environmental and logistical conditions. Realizing the potential of such areas will also require technological innovations and adaptive management to overcome operational challenges.

The ASI offers a robust foundation for spatial planning and future aquaculture development in the face of growing pressure on coastal aquaculture, including climate change, resource competition, coastal saturation, and environmental degradation. Its methodology is transferable and adaptable to other regions and species, supporting a proactive, data-informed approach to sustainable aquaculture planning. Future research should focus on enhancing the ASI by incorporating climate projections and more realistic species-specific physiological thresholds to better anticipate long-term suitability. Empirical validation through field trials or pilot offshore farms would further improve its reliability and relevance for policy and investment decisions.

As the demand for sustainable aquaculture solutions intensifies, tools like the ASI can play a pivotal role in supporting resilient, efficient, and ecologically sound marine spatial planning.