Environmental Impact of Human Activities in Marine-Coastal Area: Underwater Wine Cellar as Case Study (Tuscany, Cetacean Sanctuary)

Abstract

A new production activity developing in coastal areas and located in the MPAs are sub-cellars for wine ageing, which combine the results of wine refinement with beautiful bottles decorated with organic concretions. Assessing the associated environmental risks is crucial as wine is a complex mixture of chemical substances that are toxic to marine species if released into the environment. We have assessed the risks associated with the granting of a license to store 2000 bottles in a sunken wine cellar in the Whale Sanctuary (Tuscany). Local risks that could influence the occurrence of offences were assessed to calculate the potential dose of wine released from the cellar. Furthermore, based on the ecotoxicological approach, the effect thresholds (ECx, NOEC, LOEC) were quantified for species from different trophic levels to define the tolerance of the marine ecosystem using the PNEC approach. The results showed that wines with different physicochemical properties developed differently during maturation (0–8 months) and exhibited different ecotoxicity. This led to different PNEC values and, consequently, to different risks of bottle breakage. The main suggestion of our results is that the release of licenses should be based on a case-by-case risk assessment that focuses on both the environmental characteristics of the ecosystem that maintains the cellar and the chemical properties of the wines stored in glass bottles.

1. Introduction

During long-term storage, chemical reactions occur that alter the taste, aroma and palatability of the wine, mainly due to oxidation and other chemical interactions that change the colour, aromatic compounds and concentration [1,2]. Bottle ageing is a common practise in which the wine is stored under reduced conditions, leading to an overall improvement in sensory characteristics. The raw material must be selected based on the organoleptic and physicochemical characteristics of the specific product to ensure an advantage in organoleptic properties at the end of the process. The time required to achieve the desired characteristics depends on the type of wine: “simple” wines require shorter maturation times, while “complex” wines are generally improved by longer maturation times. The reactions that occur depend on the type of wine and the vinification technique used [3]. Consequently, the storage of wine and the environmental conditions during storage are key factors that can significantly influence the quality of the commercial product. In 2010, 68 champagne bottles were recovered in the Baltic Sea from a shipwreck that had hit the Jacquesson champagne houses in 1840 [4]. The marine environment proved to be particularly suitable for the preservation of sparkling wines as the constant temperature of the water at a certain depth, together with the low incidence of light, creates an ideal environment for the preservation and maturation of the organoleptic properties of the wines. It has been shown that bottles salvaged from shipwrecks optimally preserve the wine they contain, both chemically and sensorially. This is due to the environmental conditions at a depth of around 50 m (low temperature, 2–4 °C, low salinity, <10 PSU, slow current movements), which show that this method is suitable for slow maturation [4]. Due to this discovery, the worldwide spread of underwater cellars has become an interesting commercial and touristic trend [5]. This is because, in addition to the organoleptic aspects associated with the maturation process, the products preserved and/or refined underwater have the aesthetic quality of bottles with epiphytic vegetation typical of the marine environment, which enhances the final product and makes it commercially attractive (Figure 1). This characteristic is enhanced by the fact that the natural removal of epiphytes by marine herbivores grazing the solid surfaces underwater is prevented by storing the bottles in cages/baskets. Consequently, epiphytes can colonise the surface of the bottles, while they are inaccessible to herbivorous organisms. Ultimately, this leads to a significant increase in the price of a bottle of sparkling wine with maritime storage, ranging from EUR 68.99 to EUR 99.00, compared to EUR 11.20 to EUR 27.30 for a classic 0.75 litre bottle of sparkling wine (word “cantine”, and “sommerse”, and “shopping” on google, last accessed 31 December 2024).

Recent research on the effects of sea maturation on white, red and rosé wines has been conducted to determine the induced changes in the chemical composition of the tested wines and in the flavour of the commercial products [6,7,8]. Maioli and colleagues (2024) [8] showed that the underwater storage of wines influences the quality of the products depending on the type of wine (red, white and rosé wines).

These productive activities are increasingly attracting the interest and support of local administrations as they represent an opportunity to promote the territory and its agricultural food supply chain while increasing the visibility of typical regional products, especially in the context of culinary tourism. Despite recent interest in the effectiveness of sea cellars on the organoleptic and chemical characteristics of final beverages, there are few assessments of the associated environmental impact of this new human activity. As the commercial interest and importance of this activity increases, so does the potential risk to the environment as the seabeds for this utilisation become more widespread. Although foodstuffs are not chemical substances that are dangerous to human health per se, they can pose a risk if they are released into the environment. To give a concrete example: the presence of ethanol is an element that can have ecotoxic effects on the biotic component. Since the mixture of substances that make up a food is not standard and constant but variable, it must be emphasised that the potential risk to the environment associated with it cannot be theorised a priori but, on the contrary, is an aspect that must be determined on a case-by-case basis. The impact of the release of food on the species present and the risk associated with the event can be assessed by ecotoxicological tests that consider the food as a complex chemical mixture that contaminates the natural seawater. The technique of preservation at sea involves the immersion and anchoring of inert materials (crates/cages) and the temporary storage of bulk wine bottles for human consumption in these inert materials. Accidents and breakage of bottles stored in submerged cellars may occur and the wine could be released into the marine environment where it acts as a pollutant and exerts an ecotoxicological effect on marine species and trophic webs. Although dilution is an important element in the release of wine into the environment and the ecotoxicity of ethanol reported by ECHA is above 100 mg/L in marine species, its PNEC for aquatic marine environments is estimated at 0.79 mg/L (https://echa.europa.eu, accessed on 31 December 2024) and wine represents a complex chemical mixture where risks cannot be excluded a priori. The aim of this study was to assess the risks to the environment associated with the location of a submerged wine cellar in a marine ecosystem of particular interest in terms of the level of protection (Magliano, Tuscany, Italy).

This study assesses the risks associated with the commissioning of the facility, the occurrence of accidents and environmental disasters, and the risks associated with the accidental release of wine due to the accidental breakage of bottles during work activities. The impacts were assessed using a cost–benefit analysis based on a risk assessment and the PEC/PNEC approach used for chemicals. Cellars, bottles, corks and labels were analysed for the release of chemicals of ecotoxicological interest into the marine environment. In addition, red, white and rosé wine, classic sparkling wine and sparkling wine in bloom were analysed to assess their chemical composition before and after eight months of ageing in classic and marine cellars (16 °C) and to assess the change in chemical composition and associated ecotoxicity to marine trophic webs.

2. Materials and Methods

2.1. Study Area

The study area consists of a marine area in the municipality of Magliano, adjacent to the Parco Regionale della Maremma (a terrestrial protected area SIC/ZPS IT51A0014; SIC/ZPS IT51A0015; SIC/ZPS IT51A0016) and within the Santuario Pelagos (protected area for marine mammals, L. 391/2001), close to the IBA098 (Important Bird Area, Monti dell’Uccellina, Stagni della Trappola e Bocca d’Ombrone). According to the literature (Geoscopio, Regional Site of Tuscany, http://www.regione.toscana.it/-/geoscopio, accessed on 31 December 2024), many protected species live in this area, including invertebrates, amphibians, fish, reptiles, mammals, and birds. The protected species/habitats are as follows: 60 species and 15 habitats in IT51A0014; 21 species and 20 habitats in IT51A0015, 26 species and 12 habitats in IT51A0016. The study area has an extension of 400 m², is located 1600 m from the coast, and has an average depth of 14 m on soft sandy soil without the presence of protected species and/or macrophyte meadows. The water volume of the considered marine sector interested in the activity is 5600 m³ (5.6 × 10⁶ L). According to data referring to the year 2020, the study area is characterised by an average temperature of 8.26 °C (2.97–14.68 °C) in January and 25.61 °C (18.29–32.06 °C) in July. The winds reached an average of 9.2 km/h (maximum 17.5 km/h) in July and 12.7 km/h (maximum 27.0 km/h) in January (https://www.ilmeteo.it/portale/archivio-meteo; https://it.wikipedia.org/wiki/Clima_della_Toscana, accessed on 31 December 2024). Based on the characteristics of the study area described here, the greatest risk of the bottles breaking due to natural events is classified as low. Accidental breakage during immersion and recovery of the stored bottles could be considered the main cause of chemical pollution of the marine environment resulting from this activity.

2.2. Tested Wines

Five different types of wine (

Table 1. Main features of tested wines.

	Region	Ageing	Alcohol	Type
RW	Tuscany	Barrel, Boot, Barrels	14.5%	Reserve, DOC
RsW	Tuscany	Steel	14.0%	DOC
WW	Tuscany	Steel	14.0%	DOC
BSW	Lombardy	48 months in bottle	12.0%	DOCG
CSW	Lombardy	18 months on the lees	12.5%	DOCG

) bought at local markets were tested at different times and under different storage conditions to evaluate the changes in their chemical profiles that occur due to the type and duration of storage: red wine (RW, Sangiovese 80%, Ciliegiolo 10%, Cabernet Sauvignon, 10%), white wine (WW, Vermentino 100%), rosé wine (RsW, Sangiovese 70%, Ciliegiolo 15%, Montepulciano 15%), classic sparkling wine (CSW, Pinot Noir 100%) and Charmat sparkling wine (BSW, Chardonnay 90%, Pinot Noir, 10%). The wines were tested as purchased (T₀) and after eight months of storage at 16 ± 2 °C in flooded marine cellars (T₁S) and in classic cellars (T₁C). All tested wines were stored in glass 0.75 L bottles characterised by a similar shape and cork caps. The whole bottles’ content was homogenised to take into some account possible flocculated particles during the ageing, and aliquots were used for the determination of the tested parameters.

2.3. Submerged Marine Cellar Activity

This activity involves the simultaneous immersion of approximately 20 cages with a capacity of 100 bottles per cage, resulting in a total of 2000 bottles at full capacity. The cages are anchored to the unstable floor (−14 m depth, about 16 °C) with hooks and anchors that guarantee tear resistance and resistance to natural wave movements. The depth of marine cellars was authorised by the Municipality in accordance with the specific features of the studied area. Anchorages, hooks and tears due to disturbance from trawling are classified as low risk as the concession area is appropriately marked and difficult to access for activities as it is located between a sunken wreck (raft, 500 m away) and concrete blocks for recolonisation with a deterrent effect for fishing.

2.4. Logic Model Adopted

The logic model used for experiments in this study is reported in Figure 2. An experimental set of bottles for each type of tested wines were analysed directly and placed in both classic and marine cellar as an in vivo trial before the start of cellar activity.

2.5. Physicochemical Analyses on Wines

Chemical analyses were performed following standardised methods used to test wines for commercial purposes in Italy to evaluate starting levels of each different wine type and their changes after eight months of incubation in standard cellars and marine cellars. In particular, pH (UNI EN ISO 10523 [9], pH units, limit of quantification of 0.01 LOQ), reduced sugars (Reg. CEE 2676/90 annex 5, par. 3.2, g/L; 0.5 g/L LOQ), SO₄ total (Reg. CEE 2676/90 annex 25, par. 2.2.3.3, mg/L, 1 mg/L LOQ), SO₄ free (Reg. CEE 2676/90 annex 25, par. 2.3.3.1, mg/L, 1 mg/L LOQ), alcohol (Reg. CEE 2676/90 annex 5 + Reg. CEE 2676/1990 annex 3 + Reg. CE 1493/1999 annex 2, par. 3, % volume, 0.1 %vol. LOQ), total acidity (DM 12/03/1986 SO GU n 161 14/07/1986 Met II, g/L, 0.1 g/L LOQ), acetic acid (OIV-OENO 621-2019, g/L, 0.01 g/L LOQ), L-malic acid (OIV-OENO 599-2018, g/L, 0.05 g/L LOQ), and L-lactic acid (OIV-OENO 598-2018, g/L, 0.1 g/L LOQ) were measured following standardised methods.

These changes over time may determine the variation in the chemical composition and/or the concentration of bioactive molecules in the wine mixture, which could determine a variation in the overall chemical fingerprint spectrum of the wine. For this reason, to quantitatively evaluate the total amount of the chemical change that occurred during ageing, the wines were analysed to evaluate the changes in the infrared spectrum after five and eight months of ageing. Some bottles for each wine type of specific commercial interest (CSW, BSW) were extracted after five months of ageing to determine wine maturation. The analyses were carried out using a μFT-IR infrared detector with the Fourier transform technique to determine the percentage of total change at the end of the ageing process. Analyses were performed using a Nicolet iN10 MX instrument (Thermo Fischer Scientific^®, Waltham, MA, USA) and a liquid nitrogen cooled detector (MCT-A) operating under conditions of reflection, transmission and attenuated total reflectance (ATR, germanium–iridium crystal) with a detection limit of 10 μm. The software used for the integration of the spectra obtained from the samples is the Omnic™ Picta™ software v1.7.192. The composition of the sample was verified using a statistically significant number of spectra (n = 15). The average sample of the original product (before ageing) was stored in the library as a reference for determining the percentage match on the same sample at the end of the refinement process.

2.6. Assessment of General Risks for the Environment

The ecological risk assessment was carried out in three steps. First, we excluded this activity from those covered by Directive 2001/42/EC of the European Parliament and of the Council of 27 June 2001 on the assessment of the effects of certain plans and programs on the environment (EU 2001/42/EC). In Italy, the settlements excluded from the cited EU legislation are regulated by regional laws. Secondly, the regional law of Tuscany was considered to assess the environmental risks related to the immersion of cellar structures and wine bottles in the marine environment (BURT, Resolution No. 613 of 18 May 2020). Accordingly, we carried out a semi-quantitative evaluation of the direct and indirect impacts of the intervention by using a risk assessment approach based on a two-tier matrix. The risk factors are categorised into main groups: unforeseen events (G1) and impacts during cellar works (G2). In both cases, eight sub-factors of exposure to the marine environment were identified and assessed, which relate to the environment and local structures and may influence their relevance. For each risk sub-factor, the impact was assessed using a risk-based approach (risk, R) that combines the potential harm (severity, S) with the probability of its occurrence (occurrence, O) to obtain the associated risk (R) according to the model R = S × O. Both the severity and the probability of occurrence were categorised as low (1), medium (2) or high (3) for each risk factor based on the information collected in the study area. The risks were categorised as follows: very low (VL, R < 3), low (L, 3 ≤ R < 6), high (H, 6 ≤ R < 7), very high (VH, ≥7). The sub-factors resulting from the evaluation matrix with a higher risk than the VL value (R ≥ 3) were the subject of specific, in-depth analyses.

2.7. Risk Assessment-Pollution

A specific risk assessment was performed for the risk sub-factors associated with lines G1-4 (cage and/or content placement activities), G1-5 (cage and/or content utilisation activities) and G2-7 (bottle breakage contamination). This risk analysis was divided into two steps: ecotoxicological risks due to the release of chemicals from the cellar materials and ecotoxicological risks related to chemical contamination due to the release of wine from the bottles.

2.7.1. Ecotoxicological Analyses on Cellar Materials

The ecotoxicological determinations were carried out in the laboratory on the inert materials immersed in a weight/volume ratio (1:4) in accordance with the provisions of Ministerial Decree 173/2016 for ecotoxicological tests on solid matrix. The species battery chosen for the evaluation complies with the provisions of Ministerial Decree 173/2016, and the type of solid phase test was accordingly replaced by the type of liquid phase test for elutriate. The inert materials to be evaluated are glass bottles, metal cages and rivets to fix the incoherent substrate.

The tests were carried out with untreated inert material (not immersed), which is assumed to have the greatest impact on the environment. Indeed, the immersion time determines the accumulation of a continuous layer of epiphytes on the exposed surface of the submerged materials, reducing the exchange between the material itself and the water column and, consequently, the release of chemical substances from the materials into the environment (Figure 1).

The ecotoxicological tests were performed on three species at different trophic levels: bacteria (A. fischeri; UNI EN ISO 11348-3 [10]), algae (P. tricornutum; UNI EN ISO 10253 [11]), and herbivores (P. lividus; EPA/600/R-95-136/S. 15 + ISPRA 11/2017 [12,13]), according to standardised methods. In addition, the values for pH (UNI EN ISO 10523 [9]), oxygen (UNI EN ISO 5814 [14]) and salinity (APAT CNR IRSA 2070 [15]) were measured.

2.7.2. Ecotoxicological Analyses on Wines

Even though wine is perceived by humans as a beverage, it is a chemical substance that could have a toxic effect on marine life once it enters the marine environment. The wines were tested at T₀ and after eight months of ageing in both traditional and submerged cellars at 16 °C. The ecotoxicological effects on three species at different trophic levels were assessed using the methods described above to determine the ECx, LOEC and NOEC values and to calculate the PNEC values used for risk assessment in the event of accidental degradation of bottles posing a potential hazard to the marine food web.

2.7.3. Risk Calculation

In this study, weights were assigned based on the in-depth knowledge of the study area and the type of activity performed, assuming that activity will be managed during the working period as described previously. The risk assessments were performed based on two different approaches:

The weight-of-evidence approach for solids and their elutriates. The weight-of-evidence (WOE) approach is often used to weigh the impact on the trophic web [16,17]. The data collected on tested species were used to determine the ecotoxicological risk to the environment of environmental matrices such as water and sediment and solid waste or materials (e.g., cigarette butts) using the WOE approach. The results of the tests performed on elutriates from cellar structures, bottles and labels are integrated using Sediqualsoft^® v.1. to assess the ecotoxicological risks. Risks were evaluated on pristine materials (glass of the bottles, galvanised iron of the cages and rivets for fastening to the movable base) because the fouling growth that forms on the outer layer of the bottles and cages decreases after a few months of storage, and it is considered a factor able to reduce the release from the surface of used materials. The risk of perforation of the lid due to abrasion by herbivores is also estimated to be low as the cage does not allow for direct contact between the bottles and the animals grazing on a solid substrate and the experiments carried out in the laboratory under high-stress conditions by the grazing organisms did not show any impact on the integrity of the cork.

The risk characterisation ratio for wine release pollution: The environmental risks were assessed based on the estimated PEC value and the calculated PNEC value. The ecotoxicological results of the tested natural and aged wines (EC₁₀, EC₅₀, NOEC and LOEC) were used to assess the PNEC value. The PNEC estimation was based on the conceptual process described in the REACH regulation for chemical substances, assuming a production context <10 tonnes/year and calculating with limited data (acute and/or chronic) according to the deterministic approach based on the EC₅₀. The PEC was calculated considering the volume of substance released by the simultaneous breakage of 11% of the total amount of bottles immersed at full capacity in the submerged cellar (220 bottles out of a total of 2000; this corresponds to 5.5 times the value of 2%, which is considered unlikely). Estimated PEC without water renewal was estimated to be 3 × 10⁻⁵%. Calculations were performed for the most sensitive species per each wine type based on ecotoxicological tests performed using an appropriate correction factor for the specific case study (AF = 100). The PEC/PNEC is referred to as the Risk Characterisation Ratio (RCR). The RCR is a pure number and a value of RCR < 1 was considered “safe” for the environment [18]. All calculations were performed under the assumption that the system is closed and not open, as is the case in reality; the estimates obtained in this study are therefore extremely conservative. The software Biorender^® was used to perform graphical representations of collected data.

3. Results

3.1. Physicochemical Features of Tested Wines

The results of the FT-IR spectral analysis show a significant change in the spectral fingerprint when comparing the initial wines with the wines stored for five and eight months. As a result, the ecotoxicological profiles had to be determined for both ageing periods. Ageing led to an average change in spectral match in the range of 3–5% after five months and a 5–8% spectral mismatch after eight months. The results obtained by these analyses supported the decision to perform all other determination using T₀ and T₁C-S wines aged for eight months.

The results of the most important physicochemical properties of the wines tested are shown in Figure 3 (values determined at T₀) and Figure 4 (percentage deviation of the values determined at T₀ after eight months of incubation in classic and maritime cellars).

To summarise, it can be said that the reduced sugar values at T₀ showed significant differences in almost all the wines tested. The lowest values were found in RW (0.5 g/L); WW and RsW showed similar values (0.9 and 1.5 g/L, respectively), while BSW (5.5 g/L) and CSW (17 g/L) showed higher values than the others. Ageing can significantly influence the reduced sugar content in all wines tested, resulting in a significant reduction, except for RW.

The pH values were highest in RW T₀ (3.7) and lowest in CSW T₀ (pH = 3.1). In terms of total acidity, the wines tested in this study varied between 5.0 and 6.0 g/L at T₀ and showed different variation behaviour due to ageing and depending on the type of cellar. In terms of alcohol content, the wines tested were between 11.1 (CSW) and 14.5 (RW) per cent by volume and showed very little variation after eight months of ageing in both types of cellars tested. The values for free sulphur (free SO₂) were between 6 mg/L (BSW) and 28 mg/L (CSW). The ratios between total sulphur and free sulphur were as follows: 2.5 (RW), 5.0 (CSW), 5.8 (WW), 8.6 (RsW), 10.3 (BSW). In terms of total sulphur content, the wines tested in the classic cellars showed a significant increase in free SO₂ after ageing under both conditions (between 6.7 WW and 50 RW % of initial values). In contrast, WW and BSW increased by 10.3–7.5% in the marine cellars. Only the CSW wine showed a decrease of −2.9% of the initial value in the classic cellars and −4.5% in the marine cellars.

3.2. General Risks for the Environment

Figure 5 shows the results of the semi-quantitative evaluation of the direct and indirect effects of the intervention using a risk assessment approach based on a two-level matrix. For the unforeseen events (G1) and the impacts during the cellar works (G2), the sub-factors for rows G1-4 (insertion of cages and/or contents), G1-5 (use of cages and/or contents) and G2-7 (contamination due to bottle rupture) were above 3 and required a more detailed evaluation of the specific risk assessment; G1-4 and G1-5 were scored 4.5 (low), while G2-7 was scored 6 (high). On this basis, a more detailed ecotoxicological evaluation of the risks associated with the release of chemicals from the cellar structure and the backs of the bottles was carried out.

3.3. Results on Risk Assessment Related to Pollution

3.3.1. Ecotoxicological Risks Associated with Cellar Structures

Ecotoxicological tests were performed on inert materials such as cages, substrate anchors and glass (bottles) to determine the associated ecotoxicological risk. The results on elutriates showed that elutriates had no ecotoxicological effect on the species A. fischeri, cellar cages and substrate anchors (I% = −1.86 ± 0.96% after 15 min and −1.75 ± 2.04% after 30 min of exposure). The glass and the label used also had no effect on bacteria under the test conditions (I% = −1.50 ± 0.78% after 15 min and −1.01 ± 2.11% after 30 min of exposure). Algae showed no effect (Iµ % = −13.39 ± 0.82% after 72 h and inhibition of growth rate, µ of 0.09 ± 0.001%) on cages and floor anchors and Iµ % of −4.80 ± 1.29% and µ of 0.08 ± 0.001%) on glass and labels. For echinoderms, a significant effect was observed on both cages and floor anchors. Elutriates showed an effect of 32.6% (Abbott’s corrected data, 38.67 ± 2.31% of the effect without correction), while glass and labels showed an effect of 0 (Abbott’s corrected data, 4.00 ± 2.00% of the effect without correction).

Nevertheless, the WOE approach calculated using Sediqualsoft^® for the elutriates showed in a HQ of the tested battery values of 0.55 for the base cages and floor anchors and of 0 for glass and labels with a risk categorised as “absent”. At the end of this evaluation, the G1 risks for the specific case could be categorised as “low”.

3.3.2. Ecotoxicological Risks Associated with Wine Release

The ecotoxicological results of the wines tested for bacteria are shown in Figure 6 and grouped by concentration tested, ageing (T₀, T₁) and cellar type (C and S). The results show that 0.6% of the tested wine dissolved in seawater can affect almost all exposed bacteria, with 100% of the total population, with small differences depending on the type of wine. On the contrary, the type of wine becomes significant at lower concentrations of the tested wine, ranging between 60 and 100% of inhibition at a concentration of 0.4%. A concentration of 0.01% of T₀ wines can affect the bacteria between 5 and 30%. Interestingly, all wines showed 100% ecotoxicity to bacteria after storage in classical cellars at a concentration of 0.6%, while lower concentrations showed broader responses and 0.01% became ineffective for a large proportion of the samples tested. The situation is similar with wines stored in sea cellars. Here, too, the effect on A. fischeri is lower at higher concentrations, with hormesis occurring at 0.01% and 40% of the samples being inhibited even at 0.6%. The ecotoxicological responses to all species tested are summarised in

Table 2. Ecotoxicological responses and risk assessment for tested wines. Notes: EC₅₀ (Effective concentration 50%), were estimated using TRAP software; LCL = Lower Concentration Limit; UCL = Upper Concentration Limit; Risk was evaluated as PEC/PNEC. PEC was estimated considering a dilution volume equal to the considered marine sector; PNEC was calculated based on the most sensitive species for each tested wine using a correction factor of 100. In bold, the effect used to determine the PNEC. EC₅₀ and their intervals of confidence are expressed as % of wine (v/v).

	A. fischeri			P. tricornutum			P. lividus			Risk Evaluation
	EC50 (I% 15′)	95% LCL	95% UCL	EC50 (Iµ mean)	95% LCL	95% UCL	EC50 (anom%)	95% LCL	95% UCL	PNEC	RCR
RW T0	0.24	0.10	0.38	0.30	0.27	0.35	0.012	0.005	0.645	0.0001	0.24
WWT0	0.24	0.17	0.32	0.04	0.04	0.04	2.801	1.705	7.308	0.0004	0.08
RsWT0	0.25	0.19	0.31	0.62	0.35	1.07	1.289	0.957	1.621	0.0025	0.01
CSWT0	0.31	0.24	0.38	0.03	0.03	0.03	0.517	0.305	0.692	0.0003	0.10
BSWT0	0.25	0.19	0.31	0.24	0.22	0.26	0.498	0.319	0.528	0.0024	0.01
RWT1C	0.27	0.01	0.54	0.31	0.25	0.38	0.004	0.002	0.096	0.0000	0.68
WWT1C	0.21	0.11	0.73	0.06	0.00	0.13	0.767	0.545	0.936	0.0006	0.05
RsWT1C	0.25	0.22	0.29	0.04	0.04	0.05	0.012	0.006	0.064	0.0001	0.24
CSWT1C	0.32	0.29	0.35	0.12	0.05	0.20	0.529	0.307	0.648	0.0012	0.02
BSWT1C	0.23	0.20	0.27	0.06	nc	0.11	0.858	0.520	0.981	0.0006	0.05
RWT1S	0.63	0.38	0.88	0.27	0.19	0.35	0.788	0.516	0.982	0.0027	0.01
WWT1S	0.31	0.16	0.45	0.05	nc	0.11	0.803	0.500	1.193	0.0005	0.06
RsWT1S	0.25	0.21	0.28	0.04	0.01	0.08	0.769	0.665	1.248	0.0004	0.07
CSWT1S	0.31	0.17	0.45	0.15	0.01	0.29	0.541	0.276	0.724	0.0015	0.02
BSWT1S	0.23	0.19	0.28	0.16	0.02	0.31	0.768	0.578	1.001	0.0028	0.01

(a comparison between T₀, T₁C and T₁S effects), and the associated risks evaluations. In Table 2, for both tested species and wines, EC₅₀, (half maximal effective concentration), Upper Control Limit (95% UCL; +3 sigma), and Lower Control Limit (95% LCL; −3 sigma) are reported. In bold, the most sensitive effect is highlighted. The PNEC (predicted no-effect concentration) and RCR (Risk Characterisation Ratio) are also reported.

The results are used to calculate the PEC/PNEC risk ratio. In particular, based on the risk scenario calculated on the basis of an estimate of bottle breakage of 11% of the total number of bottles stored in the cellar (PEC value of 0.003%) and a PNEC value based on the species shown to be more sensitive to contamination based on the ecotoxicological results (P. lividus), the results showed a risk of less than <1 for all wines tested, which is considered a threshold for environmental safety.

In particular, the results showed a risk lower than 1 and ranging between 0.01 and 0.68. It is interesting to observe that RW is associated with the highest environmental risks.

4. Discussion

4.1. General Feature Concerning Ageing in Marine Cellars

The wines tested in this study represent a local selection of market commercially available products of complex grape varieties (Southern Tuscany wines).

Regarding pH and total acids, Maioli and colleagues [8] showed that the pH and total acids of the wines tested (Alabana, white wine, Sangiovese, rosé wine, and Merlot, red wine) did not change significantly during the six-month ageing period, neither during classic cellar ageing nor during underwater ageing. The values reported in the literature for the wines tested were between 3.20 and 3.36 and were comparable to the values for WW, RsW and BSW, while CSW in this study had a lower pH value (3.1) and RW a significantly higher value than the red wine (Merlot) reported in the literature. In addition, all wines tested in this study showed a significant, albeit small, increase in pH at the end of the eight-month ageing period in both types of cellars. A similar behaviour was found by Maioli and colleagues (2024) [8] for total acids, who reported the highest values in Albana (WW, 6.78 g/L) and the lowest values in Rosé (Sangiovese, 5.82 g/L), with no significant changes as a function of ageing. On the contrary, our results show a different behaviour after eight months of ageing depending on the type of cellar and the type of wine.

According to the literature [19], dry wines contain less than 5 g/L of reduced sugar, while abocado wines contain between 5 and 15 g/L. The wines tested were categorised as dry wines at the beginning, except for the abocado wines, which were from BSW. At the end of the maturation period, almost all wines were categorised as dry. RW stored in marine cellars deviated from the norm in almost all parameters tested, indicating problems with the cork used to seal the bottle in a marine environment.

A comparison with the literature shows that the reduced sugar measured in BSW T₀ corresponds to the values reported by Maioli and colleagues [8] for Alabana white wine (4.43 g/L), while the red wine analysed in this study has the same values compared to Merlot (0.50 vs. 0.54 g/L). The rosé wine (RsW) tested in this study showed higher values than the Sangiovese reported in the literature (<LOQ). The literature has shown that a six-month maturation in both classical and maritime cellars cannot significantly reduce the sugar content. On the contrary, our results indicate a significant decrease in reduced sugar after eight months of ageing, which can be observed in both cellars and in almost all the wines tested, with the sole exception of the RW, which shows an increase of about 29%.

With regard to free SO₂, values similar to those in this study were found in the literature for RW (16 mg/L versus 14 mg/L for Merlot in Maioli and colleagues [8], RsW (10 g/L versus 12 mg/L for Sangiovese in Maioli and colleagues [8] and WW (12 mg/L versus 16 mg/L for Albana in Maioli and colleagues [8]. In contrast, BSW was significantly lower and CSW significantly higher than reported in the cited literature.

Changes in the total amount of SO₂ can lead to changes in the sensory properties of the wine [20]. Mercanti and colleagues [7] showed that marine cellars caused a significant decrease in total sulphur (SO₂) in Merlot wines over time, reaching values that averaged 80 mg/L after five months of ageing, resulting in similar values to the wines tested in this study (T₀), with the exception that CSW had higher values (140 mg/L, T₀). The decrease in total SO₂ could be due to the occurrence of oxidation processes [21]. In the literature, it was found that Sangiovese wines stored in marine cellars showed a more significant reduction in total sulphur content after five months of ageing than classic wines and that the values were lower compared to Merlot wines [7].

Despite the rapid development of this human activity [22,23], there are very few results from experiments on the effects of maturation in sea cellars on the physicochemical properties and flavour of wines, which makes it difficult to compare our results with the literature. Our results showed that the tested RW is not suitable for maturation in underwater cellars, as the RWT₁S was not up to standard at the end of the process, while other wine varieties showed a better and much more stable performance. Based on the existing literature, the wines tested in this study showed a significant change during maturation in marine cellars, which was comparable and much better than in the literature [8,20,21]. Birkić and colleagues [6], who analysed Malvazija istarska bijela (Vitis vinifera L.) wines from the Istria and Kvarner regions aged in underwater springs, found that neither the total concentrations of polyphenols nor the antioxidant capacity had changed significantly compared to the samples aged in wine cellars, although the amounts of the antioxidants naringenin and myriketin had increased significantly. These differences could be due to the fact that the results reported in the literature for studies investigating the effects of ageing on the physicochemical properties of wines were carried out with different grape varieties and shorter ageing times, but also to the different winemaking process, which could determine effects on the polymerisation and condensation of phenolic compounds [6] and other chemicals and the different performance observed.

4.2. Regulatory Framework and Perspectives

In Italy, the interventions for the placement of inert materials at sea are regulated by the resolution of 18 May 2020, No. 613 of the Tuscany Region, which applicates the Ministerial Decree 173/2016 and its technical annexes in the regional territory of Tuscany. Specifically, underwater “cellars” are metal baskets that contain the product (wine, sparkling wine and the like in glass containers) and are placed on the seabed. Legally, the dumping of inert substances into the sea is regulated by Resolution No. 613 of 18 May 2020 of the Tuscany Region entitled “Procedure for granting authorisations pursuant to Article 17(1)(e) and (f) of Regional Law No. 80/2015 to ensure the coordination of technical assessment measures for the restoration and balance of the coastal area”, which applies to the territory of the Tuscany Region. Article 3, “Types of interventions subject to authorisation”, point 2, “Immersion in the sea of inert materials, inorganic geological materials and artefacts intended exclusively for exploitation, provided that their compatibility and environmental safety are demonstrated”, regulates the technical documentation to be attached to the application for authorisation. A simplified procedure is provided for the immersion of artefacts for marine use. The proposed intervention is a case of particular interest. As far as the current legal framework is concerned, the operation to be authorised is the dumping of inert materials at sea. It concerns cages and baskets to be anchored to the seabed and glass bottles for the storage of wine to be matured underwater. However, this type of productive activity must not ignore the risk associated with the release of food contained in inert materials into the sea. Although inert materials are generally considered reasonably safe under normal conditions, the release of chemicals with potential biological effects into the sea when glass bottles break must be assessed to rule out specific risks to the environment. Although the risk of accidental release is low, it cannot be considered negligible and must be assessed accordingly.

According to the literature, a more detailed case-specific evaluation should be performed to define general and specific risks to the marine environment [24]. This aspect is of particular importance in the case of marine cellars, as these activities are spreading rapidly in the Mediterranean and along the coastline.

4.3. General and Specific Risks for the Environment

Sea cellars are a widespread commercial phenomenon that quickly makes the coastal areas of the Mediterranean interesting [22]. In Italy, since 2008 (Tenuta del Paguro, Ravenna, IT4070026-ZSC-Relitto della piattaforma Paguro, Adriatic Sea), a diffuse number of such activities have been authorised and in operation, including some in marine protected areas (e.g., Bisson, Portofino MPA, Tyrrhenian Sea). This activity has developed strongly in the last 5 years as it effectively combines tourism and culinary experiences and has created a new branch of the tourism market called “wine aquatourism”. This innovation could have a positive impact on wine- and food-tourism-related businesses as it represents an opportunity for policy makers to boost regional tourism development [6,23]. These precedents, which are already in operation, show that the risk of accidental bottle rupture during installation and operation can be considered low and rare, even during exceptional meteorological events. Nevertheless, the increasing prevalence of these activities also in marine protected areas [23] should be supported by a case-by-case risk assessment for the receiving environment to ensure the safety of such human activities for marine ecosystems.

In this study, the risk assessment was conducted based on a two-tier matrix to evaluate the contribution of each sub-factor that may influence the overall risk using a semi-quantitative approach. In the case studied, the greatest risk was found to be associated with the transfer of materials and the release of substances in the bottles placed on the seabed during the colonisation and recovery of the commercial product. A more in-depth study on the potential impact of the release of chemicals from the cellar structures and the breakage of wine bottles was analysed.

The weight-of-evidence (WOE) approach used in this study to determine the ecotoxicological risks to the marine environment from the release of chemicals from the basement structure allowed us to exclude the occurrence of significant features for the specific activity in the area studied. This approach is in line with the literature and is considered robust for this type of assessment [16,17]. Based on REACH and ECHA guidance [25], the risks associated with chemicals released from cylinders should be better determined by applying the PEC/PNEC approach. It is important to note that due to the close relationship between PEC and PNEC to the receiving environment or the chemical properties of the wines tested, the risk assessment performed in this study is strictly bound to the case study investigated and cannot be transferred to other areas or conditions [26]. It is known in the literature that the use of a battery of tested species allows for a much more informative risk assessment and can make better-dimensioned decisions on a more solid basis [17,27]. Our results confirmed that, even if the higher tested concentration (0.6% v/v) can significantly affect the tested species (100% effect in almost all tested wines and ageing conditions), lower concentrations showed significantly different ecotoxicological behaviour in relation to wine type and ageing mode. Algae were significantly more sensitive than echinoderms and bacteria in all tested wines stored in marine cellars. Calculating the risks to the environment based on the most sensitive species resulted in a much more cautious approach, which made it possible to better compensate for the underestimation of the potential threat that the in vitro test poses to the ecosystem as a whole. In addition, the environmental impact estimates made in this study are precautionary. Indeed, the risk assessments assumed that the affected area is “closed” and not open when estimating the dilution of chemicals in the water, and the actual dilution potential of the open ocean system was not calculated in this study. These assumptions were made in order to choose the most prudent state for the environment, as the reality, characterised by a greater dilution than the estimated one, certainly has a greater environmental resilience to releases and/or breaches. These assumptions were made to better protect a marine area that is already heavily affected by human impacts and activities and is in an international marine protected area for marine mammals.

5. Conclusions

The widespread of this practice use is not only related to the effectiveness of maturation in sea cellars on the sensory, flavour, aromatic and nutritional aspects [6] of the final product, but also to additional aspects such as the natural design created on the bottles by maturation in sea cellars, which can be considered a piece of “natural art” that increases the value of the final product.

According to Italian law, the demonstration of compatibility and environmental compatibility is a “conditio sine qua non” for the authorisation of the intervention when inert substances are discharged into the sea (Ministerial Decree 173/2016 and Regional Law Tuscany, No. 613 of 18 May 2020). In general, a significant effect of underwater storage of wines depending on the type of wine has been found in the literature [8], and our study shows a significant difference in the characteristics of the wine and the associated toxicity depending on the type of wine and the type of cellar chosen for storage.

From a commercial point of view, our experiment showed that RW was better stored in Cv. In all the other tested cases, marine cellars produced a suitable product for commercialisation. Our results show that the chemical composition of the wine in the bottle can vary by up to 10% during the ageing process, resulting in different ecotoxicity. The changes in ecotoxicity are associated with higher responses to larval stages of the tested herbivores (P. lividus), indicating a high potential risk to marine trophic webs and food filtering species, including protected species. However, the risk assessments carried out, which took into account the particular characteristics of the area studied, the specific nature of the human activity (structure, size and location of the marine cellars) and the specific impact on trophic webs due to wine pollution in the marine environment, showed that the risks are below the critical threshold of 1 and the activity considered can be categorised as safe for the environment. Nevertheless, the results also suggest that a case-specific assessment should be carried out to evaluate local factors of particular interest that could alter the risk assessment and that could be considered before the authorisation for these activities is released.