Over the last decades, the need for assessment of the state of the natural environment has become a primary concern. Marine habitats, under increasing human pressure, are declining at an accelerating rate (Claudet and Fraschetti, 2010). Change is now affecting so many compartments and levels of the ecosystems that, while easily perceived, the phenomenon is challenging to quantify (Sala et al., 2000). To address this need, diverse biotic indices have been developed based on different target organisms or habitats deemed particularly sensitive to alterations of the surrounding environment (Birk et al., 2012). In addition, there is a growing trend to develop novel indices that consider the resources required to make an ecosystem function (Rigo et al., 2020, 2021), the ecological complexity (Paoli et al., 2016) or the environmental DNA (Pawlowski et al., 2018). Indices typically provide measures of the structure, function or some particular characteristics of marine communities that show a predictable response to anthropogenic disturbances. Regardless of the wide range of indices available, there is to date no rule for the proper selection of an index or a combination of indices that should be used in the assessment of the environmental quality (Borja et al., 2015).

Across the European seas, the use of biotic indices to define the ecological quality of the marine environment has been strongly supported by two consecutive EU directives: the Water Framework Directive (WFD, Directive 2000/60/EC) and the Marine Strategy Framework Directive (MSFD, Directive 2008/56/EC). The WFD suggests the use of four Biological Quality Elements (BQEs) (i.e., phytoplankton, aquatic flora, benthic invertebrates, and fish) through which a measure of water quality can be obtained and ranked, on the basis of their abundance, diversity, biomass, distribution and/or cover. The MSFD describes the quality of the habitat on the basis of 11 qualitative descriptors (Annex 1) (Borja et al., 2013). The overlapping of the two European directives has generated some confusion in the distinction between ‘water quality’ and ‘habitat quality’: the two compartments might not be equivalent, as water naturally has a more rapid turnover with respect to the whole habitat. In practice, however, the borderline between water quality and habitat quality is hard to trace and indices initially created for the WFD in 2000 were then adopted for the MSFD in 2008 and are still widely used today. All the indices considered in the present paper may be taken as responsive to the requirements of both directives. In both the MSFD and WFD, the good status of the environment may correspond to the so-called reference condition from which the change is measured, in the absence of pre-disturbance historic data or data referring to analogous pristine areas (Smit et al., 2021).

Marine Protected Areas (MPAs), where the goal of marine conservation dovetails with the regulated use of marine resources (i.e., fishing) (Grorud-Colvert et al., 2021), have gained further relevance in this context as they are often taken as a reference point, and improved environmental conditions are expected to be found there (Bianchi et al., 2022; Fraschetti et al., 2022). Most of the priority habitats and valuable species with conservation priority considered in the EU directives and typically included in MPAs, are concentrated in coastal areas (Vassallo et al., 2020), where they are also exposed to multiple pressures (Azzola et al., 2023). The alteration of the natural environment due to anthropogenic and natural disturbances may be more evident if the flora and fauna are investigated together (Smit et al., 2021). The quality of the marine environment is often assessed on the basis of a single index and/or the status of a single indicator, which can be an oversimplification. The combination of several metrics has been proved to be more accurate than the use of single metrics in habitat assessment (Pinto et al., 2009; Ruitton et al., 2014; Rastorgueff et al., 2015; Sartoretto et al., 2017; Thibaut et al., 2017; Enrichetti et al., 2019; Astruch et al., 2022). The challenge is to understand the complexity of coastal systems by considering all their components without trying to oversimplify them (Paoli et al., 2016).

The present study aims to compare 11 biotic indices with reference to three habitats/biotic components (i.e., reefs, seagrass, and fish) of the Marine Protected Area (MPA) of Capo Carbonara (SE Sardinia, Italy). Although data have been collected in the frame of distinct projects with different specific objectives focused on single descriptors each time, the resulting large dataset offers a unique opportunity to compare the performance of a variety of indices. We used the graphical approach RESQUE (REsilience and QUality of Ecosystem), which provides a single and comprehensive measure of the status of the environment by integrating several metrics (Oprandi et al., 2021), here applied to different habitats for the first time.

To reliably assess the state of the marine environment, considering as many habitats and biotic components as possible, is desirable, as single indicators/indices may provide only partial answers (Rombouts et al., 2013; Smit et al., 2021). The dataset collated for the present study has enabled us to consider three out of the four BQEs mentioned in the WFD and five out of the 11 qualitative descriptors listed in the MSFD. Being based on a significant proportion of the biological indicators suggested by the abovementioned EU directives, the analysis we carried out can be considered a useful alternative when more specific assessments with purposely planned sampling designs cannot be applied.

Indices are indeed suitable and easy to use tools for management purposes (Gatti et al., 2012), but when possible analytical approaches such as BACI and beyond BACI (Underwood, 1994; Smith, 2002; Christie et al., 2019) should be preferred to get a solid assessment of environmental status or of the impacts occurred in a certain area. These approaches, however, require data collected specifically according to dedicated protocols, which was not the case for the present study.

We utilized data collected for different purposes, with specific and distinct sampling designs, which certainly represents one of the limitations of our study. Of course, some important habitats/biotic components (e.g., soft bottom) were not taken into account, but the main aim of our work was not to explore all the marine habitats included within the MPA, nor to verify the efficacy of the protection regime, but rather to see whether a number of components of the marine environment that are customarily considered by environmental agencies and administrations were in agreement in terms of state of the environment. Even with a dedicated study, sampling all the habitats/biotic components of the MPA would have required a huge effort in terms of funds, time, and people, and therefore be virtually impossible to implement. The monitoring activities that are carried out more or less regularly in the MPA enabled us to dispose of a large dataset covering most of the environmental features of the study area.

The problem is not only the availability of data but also how to assess the whole to achieve an unequivocal measure of the ecosystem status, avoiding loss of information. The risk, when considering many different or conflicting responses together, is to fail to interpret them and thus simplify the final result (Gatti et al., 2012; Gatti et al., 2015; Thibaut et al., 2017; Mancini et al., 2020). The RESQUE approach had already revealed its potential in comparing different metrics related to the same ecosystem (seagrass meadows), whether it changed over time or space (Oprandi et al., 2021, 2022). Here, it also proved to be effective in the comparison of indices relating to different habitats and biological components, clearly showing that all the indices were highly congruent. In contrast, previous Mediterranean studies comparing different indices in soft bottom habitats provided results that were not conclusive or were even contradictory (Occhipinti Ambrogi et al., 2009; Reizopoulou et al., 2014; Sánchez-Moyano et al., 2017; Magni et al., 2023). Peleg et al. (2023) found only “some consistent patterns” when using nine health indicators to assess the status of rocky reefs in New Zealand.

The available reference classifications of indices are often inadequate when the geographical context is different from the one in which they were originally created (Pinto et al., 2009). The method we used does not actually require the use of reference classifications, but is based on the calculation of EQRs, and thus on the identification of a reference value for each index, which represents the best obtainable value for that index in the investigated area (Oprandi et al., 2021). Having a large available dataset is a key asset in order to identify reference values consistent with local scale patterns (Smit et al., 2021). In the case of Capo Carbonara MPA, it was possible to successfully apply the RESQUE approach thanks to the availability of a considerable amount of data which, being collected in independent studies, were designed to maximize information in the study area.

For simplicity, in the present work, we have referred to the method of Gobert et al. (2009) for the identification of the reference values, which is widely used by the Italian environmental agencies in their monitoring plans. This approach, however, implies that particularly low EQR values can only be observed when there is great variability in the data. Thus, values close to 1 could also be observed in the case of conditions with little variability but low environmental quality. This was never the case for Capo Carbonara MPA: all indices, taken individually, presented ecological status between good and high, and the data used to calculate them showed high variability.

The choice of reference is a hotly-debated issue that has no univocal answers (Muxika et al., 2007; Borja et al., 2012). Standardizing (i.e., bring all values in the range 0-1) remains the correct thing to do (Legendre and Legendre, 2012). However, a better definition of the references and related EQR calculation would certainly be desirable; to this end, expert judgement is a key tool to include. Future studies may explore further this issue to find more effective ways of providing a reference for the EQR.

We assumed that an ideal hendecagon, which is obtained when all indices values reach their maximum (i.e., are equal to 1), was the benchmark describing the best conditions for the Capo Carbonara MPA, corresponding to a high environmental status. Area and circularity values of the polygon obtained from RESQUE analysis corresponded to 73% and 97%, respectively, of those of the ideal hendecagon: thus, all the indices considered not only provided strong suggestive evidence that support the good status of Capo Carbonara MPA, but also – and even more importantly – that they are fully consistent.

NIS and Biodiversity are, by far, the descriptors that contributed the most to the overall good status of the MPA. Paradoxically, NIS also represents the only descriptor that should not be directly affected by the protection. There is evidence, however, that habitat maintenance contributes, to some extent, to limiting NIS establishment (Bernardeau-Esteller et al., 2020; Houngnandan et al., 2022). The general healthy status of the benthic communities within the MPA possibly manages to keep allochthonous species at a rather low level (Piazzi et al., 2021a). This is particularly true with regard to Posidonia oceanica meadows, as NIS substitutes are known to settle mostly on dead matte areas rather than within dense canopies of living shoots (Oprandi et al., 2014; Montefalcone et al., 2015; Piazzi et al., 2016). In Capo Carbonara MPA, the values of the Conservation Index, which measures meadow degradation (by comparing the cover of dead matte with that of living P. oceanica), indicated a reduced state of conservation in some areas inside the meadow, which likely facilitated the establishment of NIS. In fact, the mean value of ALEX_S was slightly lower than ALEX_R. The reduced abundance of macroalgal NIS could be the result of successive colonization stages that led to the achievement of a balance between native and NIS assemblages (Boudouresque and Verlaque, 2012). It is likely that the general healthy status of the benthic communities manages to limit the expansion of NIS in the study area (Piazzi et al., 2021a).

According to CARLIT, more than 50% of the rocky coastline of the Capo Carbonara AMP is characterized by species with a medium to medium-high level of sensitivity (Ballesteros et al., 2007; Atzori et al., 2020), thus indicating good water quality. Similarly, the ESCA index, which is known to be sensitive to human disturbances (Piazzi et al., 2021b), displayed high values. These outcomes underline how proper management can positively affect both shallow macroalgal communities and circalittoral coralligenous assemblages.

The benefits of protection on fish biomass and abundance are well known (Guidetti et al., 2014), and can be perceived over a time span of 5 to 20 years (Edgar et al., 2009; Babcock et al., 2010), while the reserve effect on fish diversity is still debated, as it may depend on the time since the enforcement of the protection and other factors not yet fully elucidated. A recent study showed that a stabilizing effect on fish biodiversity is apparent within 4 years since the establishment of an MPA (Pettersen et al., 2022): Capo Carbonara MPA was established about 20 years ago.

COARSE was created in the frame of the MSFD, for the specific purpose of providing information on seafloor integrity (Gatti et al., 2015). The index consists of three distinct parts concerning bioconstruction, biodiversity, and the three-dimensional structure of the coralligenous habitat. These are able to detect alterations in the environmental quality that are not only related to human pressure. In particular, the diversity of the intermediate layer and the conditions of the upper layer seem to be sentinels of climate change (Piazzi et al., 2017b). Recent comparative studies reported significant changes in the deep reef communities of Capo Carbonara MPA over the last 20 years (Azzola et al., 2022), including a local mass mortality event of gorgonians following summer heatwaves (Piazzi et al., 2021b). Boundaries of MPAs do not constitute a physical barrier against sea warming. However, the high COARSE value we found, may indicate high resilience of the coralligenous reefs of Capo Carbonara MPA.

CI and LIMA indices, adopted to assess seafloor integrity, reported comparatively lower values. However, both indices consider benthic associations that have a wider distribution and bathymetric range than coralligenous reefs and are, therefore, potentially exposed to greater risk of impact and a greater natural variability. CI in particular may have been affected by leisure boating, as signs of anchoring were reported during routine monitoring activities on P. oceanica meadows (Montefalcone and Oprandi, 2020). With regard to LIMA, it should be emphasized that the formulation of this index takes the topography of the seabed into account when assessing habitat complexity. This can be to the detriment of the biological part at those sites where, notwithstanding equal presence of species, bottom rugosity is lower.

Indices of exploited fish and food webs (Exploited fish and HLP) showed similar trends as they focus mostly on the same species: approximately 75% of high-level predators are actually commercially exploited. The average biomasses of exploited fish and HLP we estimated in Capo Carbonara MPA are comparable to those found in other regional-scale studies where the MPA management was deemed sufficient to obtain measurable effects on fish (Di Franco et al., 2009; Guidetti et al., 2014; Guidetti et al., 2019; Rojo et al., 2021).

Besides the comparison of different indices, the use of RESQUE provided some valuable insights into the way of interpreting the differences between water quality, defined according to the WFD, and habitat quality, defined according to the MSFD. This type of analysis strengthened our results, because apparently both Good Ecological Status and Good Environmental Status were achieved in the Capo Carbonara MPA. The differences, in terms of requirements and goals, between the two European directives have long been discussed (Borja et al., 2010), but the present study evidences for the first time that they are congruent in their assessments of the status of marine ecosystems.

The necessity of using a combination of several indicators to define the quality of the marine environment has been emphasized more than once. To date, there is no risk of failing to achieve this objective given the number of existing indices. The question is no longer just how (structural approach vs. functional approach) to assess what (water quality or habitat quality), but more importantly, with which metrics/indices and how many (Borja et al., 2015).

Although our study has been carried out at a local scale, the approach used here offers a starting point for further work in the near future, as biological indicators are being implemented at broader scales (Fraschetti et al., 2022). We consider that our approach may represent an example to be followed in other areas where monitoring data are available, in comparative studies involving both protected and unprotected environments.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AO: conceptualization, investigation, data analysis, and writing original draft – review & editing. FA: writing – review & editing. AA: investigation and data analysis. CNB: conceptualization, data analysis, and writing – review & editing. NC: investigation and data analysis. LC: data analysis. ED: investigation and writing – review & editing. FF: investigation and data analysis. MG: investigation. PG: investigation and writing – review & editing. CM: data analysis and writing – review & editing. LP: investigation, data analysis, and writing – review & editing. FP: data analysis and writing – review & editing. MM: investigation, supervision, and project administration. All authors contributed to the article and approved the submitted version.