Analysis of Sites’ Suitability for Floating Photovoltaic Plants on a National Scale and Assessment of the Decarbonization Potential

Abstract

Given the current global context, in which emissions’ reduction to mitigate climate change is a primary concern, the use of new clean technologies is being explored. Floating photovoltaics (FPV), given their many advantages, such as increased efficiency and water savings, were examined here originally to estimate on a national scale their potential contribution to decarbonisation. Thus, our study assesses whether Italy is a suitable territory for hosting FPV plants on bodies of water. The analysis consisted of two phases: a selection of suitable bodies of water and a subsequent prioritization using scores. To perform these, predisposing factors were first determined. In parallel, quarry lakes on Italian plains were identified because they could be redeveloped by installing FPV plants. Other analyses, moreover, allowed us to estimate that, in the best scenario, there could be up to 507 plants larger than one hectare, which could come to satisfy almost 3% of the annual electricity demand and could save more than 4.6 million tons of CO₂ in one year. These results allow us to conclude that it is indeed possible to use this technology in Italy, marking a big step in terms of innovation.

1. Introduction

The transition towards the use of renewable energies is becoming increasingly necessary in a context where the massive use of fossil fuels, even more favoured by population and economic growth, is inevitably impacting global warming and climate change [1]. The last decade, in fact, has seen such rapid growth in the use of renewable energies (hydroelectric, geothermal, wind, solar, biomass, marine) that, in 2021, 38% of the world’s electricity was generated from clean sources, exceeding 36% of coal [2].

One of the most important measures to counteract a phenomenon as extensive as climate change, in fact, lies in the decarbonisation of the global economy, which results in the generation of low-carbon energy [3]. Climate change, in fact, has an anthropogenic origin and, in particular, the main cause is the burning of fossil fuels to meet energy demands [4].

The goal of decreasing carbon dioxide emissions is one of the cornerstones of the Paris Agreement, which prescribes the achievement of climate neutrality by 2050 and aims to limit temperature increase to 1.5 °C or 2 °C by the end of the century. To achieve this, the European Union (EU) has planned to reduce carbon dioxide (CO₂) emissions by 40% by 2030, compared to 1990 levels [5].

This increased use is generated in a European framework that aims, with the 2023 revision of the Renewable Energy Directive, to achieve a renewable energy target of 42.5% of the European Union’s total energy consumption by 2030, also thanks to an EU renewable energy financing mechanism.

In particular, solar photovoltaic (PV) plays a central role in the global energy transition because of its sustainability, versatility, scalability, and rapidly decreasing costs [6], and it is expected that solar energy will be the dominant technology, as the International Energy Agency (IEA) has determined that solar energy will have to grow annually by an average of 24% to reach the Paris Agreement target of zero net carbon emissions by 2050 [7]. From 2010 to 2019, global photovoltaic capacity grew from 40 GW to 580 GW [8] and, in 2023, global energy capacity from renewable sources increased by 50% with respect to 2022: this growth is necessary but not yet sufficient to reach the target set at the 28th Conference of the Parties (COP28) to triple global renewable energy capacity by 2030 [9]. In general, 2023 was a decisive year, and Italy was among the top 10 photovoltaic markets by added capacity in the world, placing eighth in the ranking, dominated by China [10].

Photovoltaic systems are the most common and widespread application of solar energy, while floating photovoltaic (FPV) systems are one of the most promising [11].

FPV is a type of installation of photovoltaic systems that, instead of being placed on land or buildings, are placed on water bodies. The installations are equipped with mooring and anchoring lines that prevent the system from moving or tipping over. Nylon ropes are often used to allow the system to move if the depth of the water body changes and thus allow it to adapt to any variations due to the water balance of the body of water [7]. These innovative systems have technical features that make them advantageous compared to traditional ground-based systems: since they are built entirely on water, land consumption and impact on the environment and the landscape are minimal, positive externalities are triggered, water is saved as they limit evaporation, and efficiency is increased as water and air beneath the panels reduce their average temperature and prevent them from overheating. In this way, moreover, the degradation of PV modules is prevented since it is accelerated by elevated temperatures [6]. In addition, their installation, maintenance, and decommissioning are facilitated and these plants can be built on sites already exploited by human activity, such as hydropower plants or flooded quarries [7]. Finally, a study in 2021 [12] showed that the changes produced by the installations on water bodies are comparable in magnitude to those that climate change would cause, which is why the construction of FPV plants can be considered a form of mitigation of some of the effects of global warming.

A recent study estimated that installing floating photovoltaic power plants in Africa could produce 20–100% of the electricity expected by hydroelectric plants, with the difference that the energy output would be less variable and more resistant to hydrological changes [13]. Another study demonstrating the great potential of FPV produced the following result: covering 10% of the median surface area (1991–2020) of more than one million lakes worldwide, both natural and man-made, with floating photovoltaic systems would, on average, satisfy 16% of nations’ demand. In addition, energy demand would be fully met in 3% of countries. From this, we can deduce the great contribution that this new technology can make to the decarbonisation process: in 2021, it was estimated that globally the use of floating photovoltaic systems would lead to a total annual reduction of 0.45 billion tonnes of carbon dioxide [3].

Also starting from the two studies cited above, carried out as strategic-level reflections at the global and continental scales and showing the great potential of FPV in terms of production and resistance to hydrological changes, our goal was to verify its operational and dimensional practicability on a smaller scale, at the national level, in Italy, a country that could have the economic and technical conditions to trigger its extensive use; nowadays, FPV systems are poorly spread in the territory [14].

In addition, we would like to highlight that, in the literature, it is pointed out that the use of PV systems around the world shows a disparity, as they are concentrated more in economically advanced nations. Accelerating decarbonisation, on the other hand, would require taking a global perspective on PV deployment and devising a strategy that prioritizes ‘high-potential regions’, i.e., those with high carbon and solar resource intensity [15]. From this point of view, although Italy can be considered among the advanced nations, it can also be included among those countries that could be particularly beneficial to the decarbonisation process because it is characterised by the heaviness of the electricity production from fossil fuels [16] and the abundance of the solar resource. These two conditions strongly suggest that Italy should be significant as a ‘high-potential nation’, to be included among those in which a strategy based on FPV must be a priority, with the goal of triggering similar studies in countries that can contribute even more to decreasing emissions. Italy, moreover, thanks to its fabric of manufacturing Small and Medium-sized Enterprises (SME), could be a nation capable of triggering innovation for this sector, useful for spreading this technology to other countries.

To summarise, this review aimed to address on a national scale the following questions:

• Is Italy a suitable territory for hosting FPV plants on its water bodies?
• Are there already exploited or abandoned water bodies in Italy where FPV plants could be installed?
• What would be the contributions of potential FPV plants in terms of electricity production and decarbonisation?

2. Materials and Methods

This work had the entire Italian territory as its study area, in order to identify all those water bodies that, following selection and hierarchisation phases, can be considered suitable for hosting floating photovoltaic systems.

This study area had a total surface area of 302,070 km² [17], divided into twenty regions. For this reason, it was necessary to aggregate and integrate different databases in order to obtain a coverage of the territory as completely as possible.

The main data source for water bodies in Italy is the DBPrior10k, a database containing the coverage of road and rail routes, hydrography, and administrative boundaries of the national territory, at a scale of 1:10,000. It is a project, ‘Priority Layers of National Interest’, implemented under the State-Regions-Local Authorities Agreement on Geographic Information Systems (IntesaGIS), starting in 2003, in three phases (2003, 2006, 2007). The fact that DBPrior10k is a homogeneous product with national coverage, even if, in some cases, it is less up-to-date, makes it an excellent basis for conducting a study that is not on a regional but on a national scale.

Within the scope of this work, the road routes’ information layer and the hydrography one were used. The road system layer consisted of tabular, point, and linear elements, continuous from intersection to intersection (dynamic segmentation). The hydrography information layer consisted of tabular, polygonal, point, and linear elements, broken up at attribute changes (physical segmentation) [18]. In this case, the elaborations were carried out only on the polygonal elements, i.e., the bodies of water, divided into the seven categories shown in

Table 1. Nature of bodies of water [18].

Nature
0	Not coded
1	Lake
2	Pond/swamp
3	Peat bog
4	Lagoon/valley
5	Artificial basin
6	Other

.

Only in the case of the Emilia-Romagna Region, however, was there a lack of a portion of data, so the DBPrior10k was integrated with its Regional Topographic DataBase (DBTR).

The complete set of water bodies on Italian territory, shown in Figure 1, consists of 81,253 elements, with a total surface area of 3226.2 km².

2.1. Predisposing Factors

In order to be able to select and subsequently hierarchise those bodies of water that could be considered suitable for FPV systems, predisposing factors for the suitability of a site were identified in this study: restrictions of the site itself, the prevalent use of the body of water, solar radiation, wind, distance to electrical substations, and accessibility of the site. With this set of factors, the suitability of a site was designed to be assessed not only based on the radiation received but also according to its practicability and possible conflicts with other site uses. Accordingly, the analysis can be considered at the same time comprehensive and precautionary. All computations were made by using ArcGIS Pro 3.2.0.

Concerning the first predisposing factor, to start selecting those sites that would definitely be suitable, the precautionary criterion of excluding bodies of water that are located within protected areas was adopted. The data containing the location and characteristics of protected areas in Italy were taken from the World Database on Protected Areas (WDPA) [19]. The bodies of water within or intersecting these protected areas are 17,384 and occupy a total area of 2802.6 km². Their exclusion resulted in a loss in terms of the number of bodies of water of 21.4% and a loss in terms of surface area of 86.9%. This considerable loss in terms of surface area is because most of the Italian large lakes are located in protected areas. If they are excluded a priori, in fact, the loss in terms of surface area will be 28.7%. Subsequent analyses were therefore only carried out on the ‘unconstrained’ bodies of water, which are 63,869 and occupy an area of 423.7 km².

The database of large dams in Italy [20] was used to obtain information on the prevailing use of water bodies. ‘Large dams’ in Italy are structures exceeding 15 metres in height or capable of containing a volume of water exceeding 1 million cubic metres [21]. There are currently more than 500 large dams, but 310 were considered in this study, i.e., those that are not in protected areas and are not out of operation. Possible uses were hydroelectric, industrial, irrigation, flood control, drinking, various, and no use. The subset of bodies of water, henceforth referred to as ‘bodies of water-large dams’, consisted of 152 elements, occupying a total area of 96.6 km². There were 85 ‘bodies of water-large dams’ that had hydroelectric as their predominant use and they were significant because of the great potential of FPV–hydroelectric coupling. The idea of coupling in the energy sector is being more and more explored, as also demonstrated, for instance, by a recent study about Coupled Hydrogen-Electricity Energy Storage Systems (CHEESS) in Port Integrated Multi-Energy Systems (PIMES), in which higher levels of efficiency are reached thanks to the complementarity of hydrogen and electricity [22].

The National Agency for New Technologies, Energy and Sustainable Economic Development (ENEA) provides solar radiation tables [23]. They contain, for the period 2006–2022 and 243 locations in Italy, monthly and annual averages of the daily values of three types of irradiation:

• direct normal irradiation (DNI);
• horizontal diffuse irradiation (Diff);
• global horizontal irradiation (GHI).

From interpolation of the annual mean values with the kriging method, the radiation maps were obtained. In Figure 2, the GHI map is shown.

The Italian Wind Atlas (AEOLIAN) [24] was used for wind velocity data and to obtain the values of distance to the primary electrical substations. The location of Italian electricity substations is sensitive information, so it is not disseminated. Concerning secondary substations, therefore, it was assumed that there was definitely at least one of them in an inhabited area because they are strategically located within communities in order to make power distribution over shorter distances more efficient, since energy losses are reduced [25]. For this reason, National Institute of Statistics (ISTAT) data, updated to 2021, of Italian localities (excluding ‘scattered houses’) were used [26]. The distance between the perimeter of each body of water and the centroid of the nearest Italian locality was then calculated, as this represented the point at which a secondary substation was most likely to be found.

In order to determine whether a primary or secondary substation would be used, it was necessary to establish the voltage level at which the electrical current, generated by the potential FPV plants, would be delivered. To achieve this, these thresholds were used:

• For extensions of less than 2 hectares or power outputs of less than 100 kW, the current will be supplied at low voltage.
• For extensions between 2 and 12 hectares or for power outputs between 100 kW and 10 MW, it will be supplied at medium voltage. In these two cases, the distance between the body of water and the nearest secondary substation is considered when assessing the suitability of a site.
• If the size of the installation exceeds 12 hectares or if its power exceeds 10 MW, the current will be supplied at high voltage and a primary substation will be required.

As far as the size of FPV plants is concerned, three scenarios were assumed: an intermediate one, where the potential plant is expected to cover 10% of the surface of the water body, and two ‘extreme’ ones, where the coverage is set at 5% and 15%.

These ‘extreme’ coverage values refer to the thresholds set by the Spanish Government in the Royal Decree regulating the granting of permits to install floating photovoltaic plants in public water basins, whose trophic status determines the choice of coverage percentage [27].

As far as power P [kW] was concerned, it was derived as follows:

$$P={\displaystyle \frac{ArH\eta }{h_{eq}}}$$

(1)

where A is the area occupied by the photovoltaic modules [m²], r is the yield of the modules themselves, H is the average annual solar radiation (in this case GHI) [kWh/m²], and η (performance ratio) is a constant that takes into account the different energy losses that can occur and is usually set equal to 0.75 [28]. With regard to the values of A and r, the three scenarios shown in

Table 2. Scenarios considered for the calculation of E.

Scenario	A	r 1
“Worst”	Coverage set at 5%	16.5%
“Intermediate”	Coverage set at 10%	20.2%
“Best”	Coverage set at 15%	24%

¹ These yield values were derived considering that floating photovoltaic panels have a 10% higher yield than ‘traditional’ photovoltaic panels, whose yield value varies between 15% and 22% [31].

were considered. These scenarios included the uncertainty associated with the variability that potential installations may have in terms of size and yield. As far as the uncertainty related to the H value was concerned, it was not included in the scenarios since its value was significantly reduced by the fact that this average value was calculated over a multi-year period, running from 2006 to 2022. As an analysis published in 2025 shows, in fact, the inter-annual variability of GHI decreases as the period considered increases [29]. Coherently, based on the available literature, for Italy it is reasonable to consider an inter-annual GHI variability of 3% decreasing to certainly less than 1% for multi-year periods [30].

About the accessibility of the site, it was necessary that the body of water can be reached by the appropriate means for it to be considered suitable for the installation of FPV equipment, so there must be a section of road around the site itself. Information and the location of road segments on Italian territory are also provided in the DBPrior10k, which lists 482,173 road segments. Of the highway stretches, however, only those that were in the vicinity of a toll booth were selected.

2.2. Quarry Lakes

Brownfields are a ‘latent resource in sustainable urban land management’ [32]. Instead of continuing to be polluted and unused sites, in fact, they could be ‘reused’ and ‘recycled’ in a ‘circular land management’ perspective [33]. In particular, sustainable land use in post-mining areas has also aroused interest recently. These areas, in fact, concern local communities and cause socio-economic problems. They degrade landscapes and, sometimes, even land, water bodies, and air [34].

Quarry lakes, in particular, originate due to the surfacing of the water table as a result of mining activities. They are quite widespread in Italy but are associated with various possible impacts, mainly related to changes in the morphology of the water table or its contamination with hazardous pollutants [35]. For such reasons, they are a current and common problem, studied in the literature in order both to limit transfiguration, degradation, and pollution of the landscape and of several environmental values and to see them as common goods and territorial resources [36]. Their environmental restoration is, therefore, a topic of great interest, and there are several ways to implement it. One of these could be the installation of FPV plants on the quarry lakes: this would produce renewable energy and, at the same time, control and protect the exposed water table from possible dangers or misuse.

Quarry lakes have specific fragilities: often they are hydrologically isolated from the rest of the surface basin, and they are affected by stratification during summer and in cold winters. Again, they are not always included in quality monitoring campaigns carried out by authorities in charge (as, for instance, required in Italy by the European Water Framework Directive) [37].

In this frame, considering that the reuse, transformation, and maintenance costs of quarries for local administrations could be too high, identifying remunerative new kinds of uses for such sites could be a win–win option, showing many positive externalities. Also considering, in addition, that, not rarely, soil consumption of traditional solar ground-based systems is not seen positively by citizens worried about agricultural uses and landscape protection.

An FPV plant built on a quarry lake could play an important role in improving and preserving water quality, triggering fewer doubts about landscape impact, since it gives new life to an abandoned and run-down situation. For instance, quarries situated in alluvial plains are often below the water table, and that implies an eventual change in land use, piezometric alterations of water table morphology, and cases of mixing between the shallow aquifer and deep aquifers, bringing chemical changes or a bad contamination situation [38]. In such fragile contexts, if quarry lakes are not monitored or supervised, the risk of becoming illegal waste dumping sites or places subject to incompatible uses could be not negligible. On the contrary, an FPV plant, needing permanent surveillance, could protect the water body from pollutants and degradation, thanks to the role of limiting evaporation and stratification [39].

This is the reason why, in parallel with the definition of predisposing factors, quarry lakes present on the Italian plains were identified.

Since there is no national catalogue in Italy, the method used to identify quarry lakes was an application of Machine Learning (ML) [40]. Both the DBTR of the Emilia-Romagna Region and the map of inactive quarries of the Tuscany Region, which contain examples of quarry lakes, were used as input to train a model. The most important characteristics deduced from the observation of the input data were the proximity of these lakes to large river courses and major infrastructures such as motorways or major railways (this was due to the great need for the supply of inert materials), their presence only in areas with certain lithotypes, and the lack of a recurring shape, so that a shape parameter, given by the ratio of the surface area to the perimeter of the lakes, was calculated to characterise them. By means of the shape parameter, the label assigned to the lithotypes, and the scores associated with the distances between the quarry lakes and the nearest rivers and highways, automatic training was carried out with the MATLAB Classification Learner app (MATLAB R2024b). After the generation of synthetic ‘quarry lake’ elements in the training dataset, which was very ‘unbalanced’, the model identified 2585 quarry lakes among those found on the Italian plains.

It is interesting to underline that the number of quarry lakes identified represents an important field of work, on which the vast literature investigates possible ways of redevelopment, with difficulties in identifying positive economic conditions including irrigation [41,42], recreational [43], ecosystem services [44], thermal energy [35], ecological [45], and others. Despite several attempts and studies, quarry lakes are often still an unsolved problem worldwide.

3. Results

Once the predisposing factors were defined, the bodies of water outside the protected areas were ranked by assigning a score from 1, for the least suitable sites, to 5 for the best sites.

First, a score was assigned for each predisposing factor, using the following criteria: the score increased as the solar radiation increased, and the score decreased as the distance to the nearest substation and stretch of road increased. Only the ‘bodies of water-large dams’ were assigned a score according to the prevailing type of use. The specific thresholds used are reported in Appendix A.

Subsequently, each body of water in each scenario was associated with an overall score, S, calculated as a weighted average of the scores s_i associated with the individual predisposing factors:

$$S={\displaystyle \frac{\sum \left(s_i{p}_i\right)}{\sum p_i}}$$

(2)

The weights $p_i$ associated with the different predisposing factors are a function of the importance of each predisposing factor for the overall suitability of a site: 10 for solar radiation, 4 for wind, and 8 for the remaining ones. Solar radiation was assigned the maximum weight, and accordingly it was used as a benchmark for assigning the other weights because solar irradiance is the most significant environmental factor affecting solar PV output, as cited by a recent study [6]. Consequently, wind was given the lowest weight since, in Italy, on average, it does not reach high speeds or, in any case, speeds that could compromise the installed panels. The other predisposing factors, lastly, were assigned a weight between these two extreme values, but still high, as they are relevant factors (even if less so than solar radiation) in assessing the suitability of a site to host FPV plants. These values led to satisfactory results for the purposes of this work given the wide scale considered, i.e., the national one. Of course, they must be tuned or re-assessed and analysed when applied on lower scales or in different geographical environments.

Figure 3 and Figure 4 and

Table 3. Number of elements for each overall score in the ‘Intermediate’ scenario.

	Overall Score
	1	2	3	4	5
No. of bodies of water	49	4833	44,554	14,353	80
No. of ‘bodies of water-large dams’	0	8	64	78	2

, concerning the ‘Intermediate scenario’, show respectively the map of the water bodies hierarchized by the overall score, the number of elements for each overall score, and the respective bar chart.

The assigning of the overall score revealed that the ‘best’ bodies of water, i.e., those with a score greater than or equal to 3, were significantly better than the bodies of water assessed as unsuitable, i.e., with a score of 1 or 2, as shown in Figure 5. This could be the first indication that Italy could be a suitable territory, given the characteristics of its water bodies, to host floating photovoltaic plants.

To confirm this, however, two factors must be considered: the producibility of the potential FPV plants and the economic income that could be obtained. To calculate them, only plants with a minimum size of one hectare, in the three scenarios mentioned above, were considered.

3.1. Producibility and Income

From the value of the total annual energy produced, E, it was possible to quantify both:

• The economic income that would be generated by placing the electricity produced by these FPV plants on the market, as a first approximation, knowing the National Single Price (PUN). The average annual base-load PUN of Italy in 2023 was 127.24 EUR/MWh. Consider that the average annual peak-load and off-peak working day PUN values for the same year are 138.16 EUR/MWh and 126.14 EUR/MWh, respectively [46].
• The percentage of Italian electricity demand that the production of these plants could satisfy. Italian electricity consumption in 2023, equal to 306.1 billion kWh, was taken as a reference [47].

The results obtained in the ‘Best’ scenario, for the bodies of water only, are reported in

Table 4. Results for the bodies of water in the ‘Best’ scenario.

	Surface [ha]
	1–3	>3
No. of bodies of water	341	166
Total surface area [ha]	547.44	3182.03
Annual E [MWh]	1,468,777.38	8,828,011.05
Annual income [M EUR]	186.89	1123.28
% demand	0.48	2.88

. For the sake of completeness, all results obtained in each scenario for both water bodies and ‘bodies of water-large dams’ are reported in Appendix B.

3.2. Decarbonization

Considering only installations larger than 3 hectares and knowing that each kWh produced by a photovoltaic source saves approximately 0.53 kg of CO₂ [48], the amounts of carbon dioxide that could be saved in one year are reported in

Table 5. Contribution to decarbonization in terms of ‘saved’ CO₂.

		Saved CO2 [t]
	Worst	Intermediate	Best
Bodies of water	895,989.59	2,460,984.85	4,678,845.86
‘Bodies of water-large dams’	417,435.99	1,144,093.59	2,137,861.99

.

4. Discussion

This work dealt with FPV, a technology that is still emerging, but which appears extremely promising in terms of potential contribution to decarbonization, large-scale renewable energy production, and zero land consumption. This is an approach that requires strategic technical and economic choices that consider the real potential size of the application of the technology in order to appropriately calibrate the strategies of business actors in developing investment methods and timing of the implementation of the interventions.

The paper proposes an approach at the national scale using Italy as a use case. It is, in fact, a country with a strong manufacturing sector, second in Europe for industrial production [49], and therefore with a high energy demand. At the same time, it is located in the centre of the Mediterranean, with an extremely favourable position with respect to the availability of radiation, compared to many other countries and regions. There are numerous bodies of water, several of which are related to active or abandoned industrial sites and therefore subject to potentially favourable use and reuse.

The census of water bodies was carried out through the mosaic of technical cartography and topographic databases produced by the regions at a scale of 1:10,000 or higher. This resulted in a properly detailed and homogeneous reading of the territory. For these potential sites, the predisposing factors towards the use aimed at floating photovoltaics were considered, yielding a hierarchy that excluded sites where feasibility was restricted or prohibitive and instead indicated the greater or lesser aptitude for the others.

A specific focus was on reservoirs intended for hydroelectric production (and extremely suitable for integrated use with FPV) and on disused quarry lakes, thus obtaining an interesting recovery, both industrial and environmental.

The work was carried out considering three scenarios that are more or less favourable to the development of the technology.

The results obtained are encouraging. First of all, in the best scenario, the number of water bodies with a small portion such as 5%, 10%, or 15% of the surface area larger than one hectare was 507 water bodies or 96 so-called ‘bodies of water-large dams’, where a hydropower plant is already active. Secondly, the total annual electricity and income that could be obtained was significant, considering that this is a new technology and that it exploits sites already present on Italian territory, which could also reduce the amount of energy imported from abroad. In 2023, 83.2% of electricity demand was met by domestic production for consumption and the remaining 16.8% was met by net imports from abroad [50]. In this regard, it should be noted that 5% of the electricity consumed in Italy comes from nuclear sources and is totally imported since there are no longer any active nuclear power plants in the country following the 1987 referendum [51].

The share of Italy’s electricity demand that would be met with the energy produced by these potential new plants alone is also an interesting result: In the best-case scenario, it comes to almost 3%. But even the values of the other scenarios, which may seem apparently less relevant, are important in an energy transition context, in which the aim is to depend increasingly on renewable sources and to cut emissions.

Finally, the amount of carbon dioxide that could be saved each year, in the best-case scenario, amounts to 4.6 million tonnes (and proportionally reduced for the other scenarios).

The results obtained, net of the large selection of bodies of water carried out, show that Italy is an area where floating photovoltaics could find application as there are several sites deemed potentially suitable to host this type of installation. In order to actually assess their feasibility, however, some aspects must first be investigated, ranging from an analysis of the limitations of this study, to the potential ecological and social impacts of FPV plants, up to considerations about economic initiative.

Concerning the limitations to be considered about this study, we would like to point out two aspects that are derived essentially from the fact that the study was conducted on a national scale: The first one is the lack of knowledge of the exact distance between the water bodies and the secondary substations, since such sensitive data are not available. The second one is the non-calibration of the weights $p_i$ in the evaluation of the overall scores. Since it was not possible to make comparisons with similar installations and since there are few FPV installations in Italy, it was not possible to set a target against which the weights could be optimised on the basis of operational experience.

From these two aspects, therefore, we also derived possible future developments of this study. The first concerns carrying out analyses at a more detailed scale, on specific sites, in order to estimate the exact distances between water bodies and secondary substations. The second concerns the determination of the weights $p_i$ used to calculate the overall scores of water bodies, which could be calibrated based on business factors that make a site more or less suitable (i.e., logistics, authorisation process, alternatives, etc.). Using these factors, fictitious examples of water bodies with a given score could also be created, based on which existing ones could be ranked, using Machine Learning.

Regarding the potential ecological impact, it is well known that PV systems in general have negligible effects on air and noise pollution during operation [52]. More specifically, there are several recent studies dealing with the effects that FPV plants may have on bodies of water. In particular, the first results of a study conducted in the Baywa r.e. plant in Zwolle (the Netherlands) showed that there were no negative effects on the environment around the installation on the flora, the fauna, or on the water quality. Still, on the contrary, positive effects on the biodiversity of the aquatic system [53] and on water quality control [54] may occur. In general, it must be considered that the effects depend on the level of water surface cover [55], as large covers can change the functioning of ecosystems [12]. Therefore, the assumption adopted in this study of a maximum cover of 15% was considered to be quite conservative, but, in general, any negative effects could be mitigated by changing the FPV design [55]. More generally, it is fair to underline that we are still in the stage of ‘infancy of knowledge’ in this field, so further research and case-by-case study of impacts could be necessary in order to have a safer and more complete assessment of this topic [12].

With regard to the potential social impacts of the installation of FPV plants in Italy, two aspects must be considered. On the one hand, the benefits arising from the installation of FPV plants, in addition to the contributions in terms of energy and decarbonisation, should be considered. A few examples of these benefits are the increase in energy independence, the diversification of the energy supply, and work and economic opportunities for local communities. The photovoltaic sector, in fact, generates wide impacts in terms of direct, indirect, and induced employment [52]. On the other hand, there is the problem of social acceptance, both in terms of the landscape impact of the installations and possible conflict with other uses of the water bodies, such as recreational, tourist, and sports uses. Social acceptance, in fact, is fundamental for FPV developers to avoid public resistance actions and resulting delays [52]. In this regard, Italy has adopted the ‘Decreto Aree Idonee’ in force since 2 July 2024, which regulates the identification of surfaces and areas suitable for the installation of renewable energy plants [56]. Based on this act, Italian Regions have to detail criteria to be established to develop renewable energies from the point of view of land uses, and our study could be useful to include water bodies in such regulations. In this regard, we can add that, in the frame of our study, the social impact was taken into account, as moderate coverage of water bodies (up to 15% of the surface area) was considered and the precautionary criterion of excluding all sites that fall within or intersect protected areas was adopted. In addition, a particular focus was placed on hydroelectric dams and quarry lakes, which are sites where FPV installations are less likely to create problems of social acceptance and, indeed, in the case of quarry lakes, could lead to redevelopment of the sites.

The introduction of this new technology in Italy, to conclude, could mark a significant step in terms of innovation, renewable production, and decarbonisation, which are not only national but also European and global goals. For these reasons, when small-scale assessments are positive, it seems logical to use this type of plant in Italy, and, to do so, it should be necessary to encourage, on the one hand, an economic initiative in this sector and, on the other, the creation by competent authorities of authorisation processes that favour their use. The results obtained can contribute to foster both aspects in need of initiative, considering that renewables are also developing strongly in Italy: in 2023, the share of Italian electricity generated from clean sources exceeded the global average, counting mainly on hydroelectric (14%) and solar and wind sources (21%) [16].

As far as the economic initiative is concerned, certainly, economic operators should investigate the income obtainable from the plants and their technical feasibility by carrying out site-specific analyses. This study, in fact, was performed on a national scale (and therein also lies its relevance) to assess whether it is actually possible to install FPV plants in Italy. Once this is assessed, a more detailed scale would have to be used to carry out analyses on specific suitable bodies of water. This would also make it possible to assess the ecological and social impacts of the installations and to survey secondary cabins within the surrounding area, thus overcoming most of the mentioned limitations of this study. In this frame, economic initiative could be favoured by the fact that quarry lakes and dams represent sites on which a start-up phase could be initiated to trigger the growth of FPV technology. This study, therefore, is to be considered as a methodological proposal for assessing the suitability of sites on a national scale. It provides a knowledge framework that should be followed by an economic and authorisation initiative in the field of FPV plants, which entails carrying out site-specific analyses and decision-making.

5. Conclusions

This study has made it possible to encouragingly state that Italy is a suitable territory to host floating photovoltaic plants on water bodies, which constitute a resource and also an element of safeguarding the territory.

The method used consists essentially of two phases: an initial selection and subsequent prioritisation of the bodies of water present in the territory. Fundamental was the identification of some predisposing factors for the suitability of a site: possible constraints of the site, the prevalent use (only for those bodies of water that are also large Italian dams), solar radiation received, wind, distance from electrical substations, and accessibility. At the same time, the identification of existing quarry lakes in the Italian lowlands, which are bodies of water of particular interest for the purposes of this study, was carried out using Machine Learning techniques.

The results in terms of potential producibility and income showed that implementing floating photovoltaic plants in Italy is not only possible (even in abandoned or already exploited sites) but could make a significant contribution in terms of decarbonisation and in response to ongoing climate change.

The method used, in fact, is relevant not only because it was carried out on a national scale (in a constrained study area like Italy) but because it fits perfectly into a European and global context in which solutions that limit negative environmental consequences as much as possible are sought and encouraged.

Some limitations of the study are recognised: firstly, the lack of data concerning secondary substations; secondly, the idea of optimising the weights used in scoring based on business factors.

On the basis of the results and limitations highlighted, on the one hand, economic and policy initiative in this field could be encouraged, starting with the opportunity represented by the quarry lakes to be redevolped and ‘water bodies-large dams’ on which further power production could be added; on the other hand, researchers and developers are suggested to continue future developments of such study and to conduct site-specific analyses.