Several approaches to divide the continuous world into homogeneous units have been proposed at different scales (e.g., from biomes to species), in different disciplines (e.g., countries in geopolitics or ecosystems in natural sciences) and with different goals (e.g., nature conservation or land use planning). These approaches are usually simplistic and/or very specific. In a framework of global change, as the one in which we are now immerse, a multi-disciplinary and multi-scale approach is required to deal with complex socio-ecological systems and accomplish a multi-goal sustainable management. Accordingly, early proposals (e.g., European Directives as the Water Framework Directive 2000/60/EC or Habitats Directive 92/43/EEC) emphasized the need to integrate ecological habitat mapping with socioeconomic maps when attempting coastal zone management. However, despite the relevance of human activities in coastal areas, the existing estuarine classifications are only based on biological and/or environmental data (e.g., EUNIS by Davies and Moss, 2002; MESH by Coltman et al., 2008; Dutch Ecotope System for Coastal Waters by Bouma et al., 2006; CMECS by Madden et al., 2010; NISB by Mount et al., 2007). To date, as far as we now, the quantitative assessment of the spatial distribution of human activities and uses in estuaries and their cumulative effects have not been integrated in an appropriate methodology capable of recognizing homogenous socio-ecological units.

Throughout history, estuaries have acted as attraction spots for human migration flows, which has intensified the impact over their natural environment, threatening the sustainability of many estuaries around the world. In this situation, an effective management of the tradeoffs between human uses and maintenance of ecosystem integrity requires synthetizing environmental, biological and human data in manageable and compatible units. However, operationalizing an objective methodology to achieve it first requires an unequivocal interpretation of concepts. Although some discrepancy in the terminology can be detected in the literature, this paper uses four key concepts: (i) physiotope, (ii) biotope, (iii) ecotope, and (iv) anthrotope (Figure 1). An ecotope is a spatially explicit and homogenous unit of biotic and abiotic components created by overlapping the biotope (i.e., biotic-homogeneous units that can be environmentally heterogeneous) and the physiotope (i.e., abiotic-homogeneous units representative of biota). The human dimension also has a spatial component, represented by the anthrotope in this research. Mapping activities through both the distribution and intensity of coastal threats is essential for the management of ecosystems by understanding the consequences of cumulative impacts (Halpern et al., 2008).

This conceptual framework (Figure 1) agrees with the recent advancements in the decision support tools for ecosystem-based management, which include the assessment of cumulative impacts of multiple pressures on multiple ecosystem components (Holsman et al., 2016). Many of the approaches used to assess these cumulative impacts are similar to the one developed by Halpern et al. (2008) for the global-scale assessment of human impacts on marine environment (Korpinen and Andersen, 2016). However, their application at a smaller scale such as the estuary require adapting the spatial resolution of the input data and the methods (Borja et al., 2016). Multiple pressures in the estuary includes environmental, ecological and anthropogenic pressures that can be spatially explicit characterized through physiotope, biotope and anthrotope maps, respectively. Likewise, the target ecosystem component of the assessment can be characterized by the biotope-ecotope. In turn, this holistic characterization is required at different spatial scales according to the different meaningful social and ecological scales recognized within the estuary. Regarding the assessment methodologies, those approaches classified as a qualitative or semi-quantitative ecosystem risk assessment of multiple stressor on independent biological subjects (Level 2, Class 2 sensu Holsman et al., 2016), like the one proposed by Halpern et al. (2008), allow optimizing the effort and data/knowledge requirements with the obtained results and response time.

In the context of this study, the basic spatial unit in socio-ecological systems is delimited by the physiotope, which acts as a surrogate for the distribution patterns of biota and human uses. It is described based on those physical and chemical properties of the environment that determine the spatial patterns of biota (i.e., biotope). In this regard, the number and nature of the specific variables that allow explaining the distribution of species and the ecological functioning of the estuary at different spatial scales is the main unsolved question. In general terms, physiotopes in estuaries can be described through a reduced number of variables, mainly those related to salinity, depth, current velocity, sediment composition and water renovation (Ysebaert et al., 2002; Bouma et al., 2006; Lee et al., 2007; Foti et al., 2012).

Therefore, this paper aims at developing a multiscale spatial classification model for estuaries by clearly representing the distribution of environmental components and synthesizing data on the biological components and the anthropogenic drivers of ecological change in an estuary. The hierarchical methodology and resulting maps provide flexible tools for conservation, ecosystems-based management, spatial planning, education, and research.

Multiscale classification models provide a holistic view of estuarine socio-ecological systems avoiding information losses in any of the relevant parts of the system. As opposed to conventional and sectorial approaches of limited uses, the proposed methodology combined biological, environmental and human dimensions in manageable units that can be directly applicable for a wide range of purposes such as ecosystem-based management and conservation (Guarinello et al., 2010; de Jonge et al., 2012). Besides, this classification system can capture the different spatial scales at which ecological processes and human pressures operate within ecosystems, using a versatile hierarchical approach.

An important advance of this approach with respect to existing methods is related with the scale of the application at an estuarine level. A high number of examples are available for the marine environment and broad-spatial scales (Holsman et al., 2016), but few studies have been developed in estuaries (Lonsdale et al., 2018). Improvements of the methods were needed for representing the social complexity (estuaries are one of the most pressured ecosystems) and ecological complexity (estuaries are highly dynamic transitional ecosystems) of estuaries. Specifically, a detailed methodology for characterizing the extent and magnitude of different types of pressures within the estuary was developed in this study (Elliott et al., 2014; Stelzenmüller et al., 2018). Accordingly, the cumulative pressure assessment method developed by Halpern et al. (2008) for the marine environment was adapted here to be applied using the anthrotope map and the ecotope map. Unlike Halpern et al. (2008) and most of the current existing tools, instead of using presence or absence data of target species, it is used the ecotope that provides explicit additional information on the range of environmental characteristics in which such species is present. This is an important point in estuaries because the need to account for the environmental variability in pressure assessment increases at small scales (Borja et al., 2016). Thus future scientific research should be focused on developing weighing scores (sensu Halpern et al., 2008) sensitive to such local factors.

In the proposed methodology, the scale-dependent within-estuary complexity is represented by physiotopes, biotopes, ecotopes and anthrotopes at five nested spatial scales, providing a gradually increasing description of biological and social processes. In turn, this methodology is also susceptible to being nested within other classification systems that operate at a regional scale. Hydrological, morphological, and climatic factors determine these large differences among estuaries (Galván et al., 2010). However, this challenge of incorporating the within-estuary and inter-estuary variability into a single, universal estuarine classification system remains a challenge for the scientific community that should be addressed in the future.

The availability of information is one of the main limitations in the use and application of this type of mapping methodologies, which generally have high data requirements. In this case, the hierarchical approach of the methodology has the advantage of adapting the scale of the study to, among other things, data availability. The higher the level of detail, the more data needed. To meet such increasing data requirements, in our case, complex interpolation techniques and numerical models could be used to distinguish among the various socio-ecological units nested in the estuary at finer-scales. Both techniques allowed characterizing estuarine environmental conditions at high spatial and temporal resolutions, while reducing the costs and efforts in comparison with field surveys and laboratory work (Diaz et al., 2004; Ducrotoy, 2010). Nevertheless, several assumptions and limitations are inherent to the application of these techniques. Specifically, in this study, the variables used for the salinity characterization did not allow detecting small changes in the physiotopes over periods shorter than a year.

There are two other major limitations in this study that could be addressed in future research. The first is that indirect interactions between the human activities and the ecology have not been taken into account in the model (e.g., pressure on predators of target species could reduce density of predators and indirectly increase density of target species). Addressing this task would require an extensive knowledge of the ecology of the target species (e.g., food chain or biotic interactions such as facilitation or competition) (Holsman et al., 2016), in addition to the characterization of the distribution area of interacting species (i.e., their biotope). Therefore, a multispecies biotope map is needed, which can be addressed by using techniques such as joint species distribution models (Wilkinson et al., 2021) or combined approaches (Pellisier et al., 2013). In this context, the role of the present approach is identifying vulnerable components that can be further analyzed considering indirect interactions between human activities and target species through impacts on interacting species. The second limitation is that the proposed approach is a one-way coupled socio-ecological model, which limits the type of questions that the model can answer (Kasperski et al., 2021). Here, outputs from anthrotope model are inputs into the ecotope model with no feedback to the first model. A two-way couple model can be potentially developed based on the physiotope, biotope, ecotope and anthrotope maps. Nevertheless, new linkages between these maps and indicators should be developed and incorporated into the model. For example, variables that link natural and human components, such as ecosystem services, or human variables, such as structures of multilevel governance. Dynamics of human-nature systems are influenced by government policies and, in the context of the present study, by the national and international legislative framework that has given rise to many types of existing estuarine management plans. However, boundaries of management systems and governance mechanisms have not been integrated in the present approach. In the future, they could be incorporated within the model as new additional maps that overlap with the ecotope and anthrotope maps (Lonsdale et al., 2015, 2018), but a further research about how to integrate specific requirements of management plans in this model is needed.

Biotope delimitation is a key point of the methodology that is open to discussion because of its implications in results interpretation and scope. Firstly, the question of how to define the distribution area of the target species, either as a potential or a realized area, must be considered. Secondly, there is a wide range of tools available to define the distribution area, from field data to remote sensing techniques and modeling. Several correlative techniques such as Maxent (Phillips et al., 2006), which was mainly used in this study to delimitate the potential distribution area of Z. noltei, allow predicting suitable environments for species establishment and survival. Other approaches with more data requirements, such as mechanistic and hybrid models, integrate hypotheses about niche, dispersal abilities and biotic interactions that provide more detailed information about patterns of species and other related properties (e.g., productivity or population dynamics). A rigorous exploration of the capabilities of both correlative and mechanistic approaches in the framework of the nested classification system should be carried out. In any case, all these approaches are suitable to be integrated within the proposed ecotope model.

Differences between the realized and potential biotope of Z. noltei can be mainly explained by biotic interactions, that were not taken into account in this study, or by human pressures. In this regard, an innovation of the classification model is that it provides a practical focus to combine, at the same spatial scale, both bio-physical information and human pressures causing impacts on the ecosystem. Integrating both ecotope and anthrotope maps allows an assessment of the ecosystem’s vulnerability and its spatial distribution, which is essential to understand how estuaries are affected by human activities. The anthrotope provides a systematic, transparent, and replicable assessment of the multiple threats that affect ecosystems at different scales relevant to managers. These maps are basic for the design of productive management and conservation actions by identifying key threats or priority ecotopes. Likewise, they can be used to identify areas and species where a better management of human activities could render a higher return on management/conservation investments (Halpern et al., 2008). The future will inevitably be aimed at developing further studies and experimental works to quantify ranges of tolerance of species against different pressure types under varying environmental conditions. New empirical information will allow improving anthrotope maps and impact assessment through the enhancement of predictive capacities, and increasing the wide variety of applications of this type of models.

A sustainable management of estuaries depends on an exhaustive diagnostic procedure able to generate a holistic understanding of the system. In this sense, the interdisciplinary research developed in this study combines natural (ecotope) and human (anthrotope) data to operationalize a spatially explicit analysis of estuaries at different levels of detail and to bridge the gap between social and ecological sciences. Therefore, the resulting nested socio-ecological maps are a powerful spatial planning instrument essential for the design and implementation of estuarine governance, conservation policies and ecosystem-based management approaches. They allow exploring effects of environmental and social changes, such as climate change or transitions to a green economy, on the estuarine ecosystem and their capacity to provide goods and services. In fact, ecosystem services are themselves the expression of complex interactions between the three components included in the classification system: biophysical context (e.g., landscape heterogeneity, biodiversity composition), ecological processes, and human interventions (Bennett et al., 2015). In this way, the proposed methodology can be adapted to understand and manage the trends observed in ecosystem service provision, minimizing their potential future reduced supply.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: Figshare with the identifier http://doi.org/10.6084/m9.figshare.14339312.

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.