Introduction

An innovative approach to sustainable groundwater management can be found by exploring the contribution of agent-based models (ABMs) in connection with aquifer flow models. Agent-based hydrogeological modeling has the potential to constitute itself as a relevant tool to support decision making in groundwater management (Castilla-Rho et al. 2015; Sánchez-Maroño et al. 2022). These models can provide information on the dynamic interactions between aquifers and their utilization to meet human needs, both for production activities (Aghaie et al. 2021; Anbari and Zarghami 2021; Du et al. 2020) and domestic uses (Al-Amin et al. 2018; Jaxa-Rozen et al. 2019) and in terms of their contribution to sustaining ecosystems (Bakker et al. 2015; Bathmann et al. 2020). An ABM can be described as a computational modeling technique that simulates the actions and interactions of autonomous agents within a system (Axtell and Farmer 2022; Silva and Braga 2020), and their application in different types of interactions between human behavior and other complex systems, including natural systems, continues to grow (Silva and Braga 2020). The agent-based approach has been applied in studies in the area of social sciences and economics since the late 1990s (Axtell and Farmer 2022), but its potential in the field of water resources management (Sánchez-Maroño et al. 2022; Bakarji et al. 2017) and, in particular, its application for groundwater was recently recognized and is still being explored (Castilla-Rho et al. 2019).

The Latin America region presents an interesting combination of challenges to the sustainable management of groundwater resources. Regarding the water resources, the occurrence of large transboundary aquifers (IGRAC and UNESCO-IHP 2015) which may concentrate different interests of usage and conservation, and the perception of groundwater as an abundant and low vulnerability resource can be pointed out as major ones. In terms of other management-related characteristics, the current low availability of reliable and systematic data (Rama et al. 2022) and the still incipient application of irrigation (3.03% of the total area available for agriculture in the region) as stated by the Economic Commission for Latin America and the Caribbean (EC-LAC) in 2013 can be cited.

This paper aims to develop a review of the current scientific literature on the application of ABMs coupled with hydrogeological models for groundwater management, analyzing their strengths and limitations and discussing their potential application for the Latin American region.

Materials and methods

This paper develops a modified systematic literature review based on the PRISMA methodology. PRISMA is a minimum set of evidence-based items for reporting systematic reviews and meta-analyses, which focuses primarily on the reporting of reviews that evaluate the effects of interventions. The modification consisted of the selection process, which involved reading the abstracts of articles obtained from the database search to restrict the analysis to current articles dealing with the topic under study.

Search criteria

The search process was performed using scientific article databases available in the CICCO—Conacyt search engine, an information resource provided by the Paraguayan Council of Science and Technology. This search engine includes commonly used databases such as Clarivate Analytics’ ‘Web of Science’ and Elsevier’s ‘Scopus’ and was selected for providing full text access for most articles. The keywords used in the search were ‘agent-based model’, ‘socio-hydrological model’, ‘socio-hydrology’ and ‘groundwater’ or ‘aquifer’ in different combinations. The search was limited by filters to the following characteristics: only peer-reviewed scientific articles published in the last 10 years (2014–2023) in English. The timeframe limit was established following the work by Sivapalan et al. (2012) and Pande and Sivapalan (2017). The former recognizes socio-hydrology as an emergent new science for the first time, and the latter states the inclusion of the term in the scientific lexicon in 2013.

Sources of information consulted

The search process included 16 databases of the 28 available in the CICCO—Conacyt system. For the selection of the sources of information, all the databases that mentioned only medical sciences in their description were removed from the databases offered. The objective was to restrict the search to engineering and environmental science applications in order to obtain articles focused on the indicated topic of study.

Selection process

The keywords selected and the filters applied allowed the search to be narrowed down to a manageable number of publications; however, many of them contained only one of the words in the search string, and in those cases, the subject of the article was far removed from the objective of this research. To refine the selection, the abstracts of all the articles found were reviewed, and those containing only one of the keywords were left out of the selection.

Data collection

The data collection of the research reviewed took into account the following: objectives and type of use, temporal extension, tools used, model structure, characterization of the input data, and advantages and disadvantages of the methodology. To facilitate the synthesis, the geographical location of the case study (aquifer or basin, region, country) and climate zone were also included. Some of these items are presented in tabular form, and the individual studies are listed in chronological order.

Synthesis

Based on the data collection of the reviewed studies, they were grouped by type of focused problems, climate zone, tools and programming languages used, required input data, main groundwater use studied, and by structure of the coupled model. Advantages and disadvantages of the application of socio-hydrological coupled models in relation to the Latin American context were also taken into account. This grou** aims to identify the relevant characteristics of modeling aquifers as socio-hydrological systems, to establish the advantages and limitations of the hydrogeological modeling methodology coupled with ABMs and to discuss its potential to contribute to sustainable groundwater management in Latin America.

Results

The search was performed using the CICCO system by applying the filters and keywords mentioned, and 146 articles were obtained as a result after eliminating duplicate articles. Among these results, five articles were filtered out of the review because their full text was not accessible through the mentioned system. The requested keyword matches and the articles obtained were then reviewed, and those articles that coincided only with one of the two keywords included in each search were also filtered out. Finally, articles that did not refer to the focused topic were eliminated from the list for review, for example, those that used the term “agent” in its meaning of chemical agent. The execution of this final refinement of the list of articles to be reviewed in detail was supported by reading the abstracts and resulted in 28 articles for the detailed review. Table 1 shows selected characteristics of the articles studied.

Table 1 Selected characteristics of reviewed articles
Full size table

Based on the results of this review, coupled models between ABMs and groundwater flow have been applied mostly to studies in arid and semiarid regions, where conflicts between aquifer use for irrigation and ecosystem sustainability date back a long time and are still relevant today. However, applications of this methodology to diverse ecosystems are also found, such as the Argentine Pampas (Garcia et al. 2019), the Murray-Darling Basin in Australia (Castilla-Rho et al. 2019), the High Plains aquifer region in the central United States of America (Noël and Cai 2017; Hu et al. 2015; Mulligan et al. 2014) and the Punjab district of Pakistan (Bell et al. 2016; Castilla-Rho et al. 2019). This can be interpreted as an indicator of the flexibility of coupled modeling to capture coupled human-natural systems (CHANS) trends and of its potential as a basis for public policy generation related to groundwater management for different types of ecosystems. Under the criteria of the literature review conducted, three articles develo** agent-based coupled models for groundwater in the Latin American region were found (Garcia et al. 2019; Genova and Wei 2023; Lima et al 2015), which is interpreted as an indication that this approach is just in its infancy for studies developed in this region.

Regarding the tools used, two major conceptual groups were identified, which in turn depend on the research objectives—on the one hand, one has fully integrated platforms such as FlowLogo, as seen in Aghaie et al. (2021, 2020a, b) and Castilla-Rho et al. (2015, 2019). On the other hand, various studies apply complex flow modeling environments, which we could identify as the usual hydrogeological approach environments such as MODFLOW (Harbaugh 2005) or MIKE SHE (Garcia et al. 2019), which in turn can be linked to the agent-based model in two ways. In the first, called the tight form, models are coupled at the source code level as indicated by Du et al. (2020); the alternative is to couple them loosely, i.e., with a stand-alone platform for agent modeling (e.g. NetLogo) that is linked to a detailed groundwater model (e.g. MODFLOW) (Jaxa-Rozen et al. 2019; Bloemendal et al. 2018). The latter approach also includes, in several cases, different methods for optimizing the results. While the benefits of an integrated platform as in Castilla-Rho et al. (2015) are evident in terms of reduced technical complexity (Jaxa-Rozen et al. 2019), in cases where the research question involves the study of complex problems in the groundwater system (related to contaminant transport and transformation or involves including interaction between surface and groundwater), a complex hydrogeological modeling environment is needed. In this aspect, considering that the application of this type of model for the Latin American region is just starting to develop, initially exploring questions that can be answered through fully integrated platforms can be considered an interesting and feasible starting point. The latter is strongly related to the next point in this section, which involves the input data sets required for these models.

Modeling groundwater systems by approaching them in an integrated manner as socio-hydrogeological systems has important implications regarding which data are needed and the techniques employed for their collection. The data sets for these models not only include hydrogeologic properties and water level observations in wells as in conventional groundwater models, but must be expanded to include measurements related to aquifer-dependent ecosystems, ecological responses to changes in groundwater discharge, consequences of pum** extraction (not only for the aquifer but also for resource users), social norms, user behavior, and effectiveness of regulations (Huggins et al. 2023). In the studies reviewed, the input data used can be grouped into traditional data for groundwater models (hydrological, hydraulic and meteorological) on the one hand and socioeconomic data on the other hand (net annual income of different types of users, use regulations by authorities, drivers of user decision making, costs related to their economic activity, among others). In turn, the degree of detail to represent users and regulatory agents in ABMs requires different types of data and the application of appropriate methodologies for their collection.

Among the studies included in this review, 23 of them (82%) include irrigation as an activity of the agents to be modeled, and among these, three also mention domestic use. Regarding other groundwater uses, one single article focuses only on domestic use, while five mention other uses or none in particular. The prevalence of studies that include irrigation is consistent with that of studies that focus on arid and semiarid ecosystems, although one condition does not necessarily imply the other. In the case of Latin America, data from the Economic Commission for Latin America and the Caribbean (EC-LAC) (2013) indicate that irrigation is applied to just 3.03% of the total area available for agriculture in the region. This seems to contradict the fact that most countries in the region are among the richest in terms of irrigation potential and have been categorized as good or excellent in terms of areas equipped for irrigation, with the exception of Bolivia and Paraguay (Villa-Cox et al. 2017). At least part of this contradiction can be explained by farmers having to incur private expenses to access irrigation systems (Villa-Cox et al. 2017), but what is relevant in relation to this region is that future pressures such as population increase and climate change could accelerate the currently incipient adoption of irrigation. This would allow taking advantage of existing knowledge in hydrogeological models coupled with ABMs for irrigation users in the early planning of management alternatives considering this type of use.

Discussion

Latin America´s water quantity is undoubtedly impressive, reaching more than 4 times the world average water annual availability per inhabitant and holding around a third of the world´s freshwater reserves (dos Santos et al. 2023). Despite this, groundwater management in Latin America faces several critical challenges that necessitate a comprehensive understanding of the hydrogeological context, socio-economic factors, and institutional frameworks.

Groundwater issues and challenges in Latin America

The following list outlines some of the most relevant problems in groundwater management in the region, connecting them to issues addressed in the reviewed studies. These connections aim to explore which regional recognized problems could benefit from a socio-hydrological modeling approach.

  • Overextraction and depletion. Widespread overextraction imperils aquifer sustainability, significantly affecting water availability for various stakeholders. In this context, Gleeson et al. (2012) presented a global assessment of groundwater depletion, conveying the urgency of sustainable management practices. Moreover, the evolving pressure on groundwater related to the water-food security nexus is starting to be recognized in Latin American countries (Hoogesteger 2022; Mahlknecht et al. 2020; Bierkens and Wada 2019). These issues can be pointed out as emergent in the Latin American region, but they are long-dated problems in other parts of the Global South (Bierkens and Wada 2019). Following this trend, several of the reviewed studies mention groundwater depletion among the problems explored through coupled ABM and hydrogeologic models (Genova and Wei 2023; Aghaie et al. 2021; Anbari and Zarghami 2021; Eyni et al. 2021; Aghaie et al. 2020a, b; Al-Amin et al. 2018; Ohab-Yazdi and Ahmadi 2018; Noël and Cai 2017; Hu et al. 2015).

  • Water quality degradation. Anthropogenic pollutants jeopardize groundwater quality, with profound implications for ecosystems and human health. Regarding this problem, Grondona et al. (2023) developed a scientific literature review underscoring the widespread detection of agrochemicals in aquifers across Latin America and the Caribbean, while dos Santos et al. (2023) investigated and mapped many water quality issues and emerging contaminants throughout the region. Regarding the reviewed studies, water quality was addressed through socio-hydrological approach in the works of Eyni et al. (2021) and Lima et al. (2015).

  • Climate change impacts. Altered climate patterns exacerbate groundwater stress, necessitating adaptive strategies. In this framework, Taylor et al. (2013) explored the interactions between climate change and groundwater resources globally, emphasizing the need for adaptive management strategies. A related relevant study by Döll (2009) assessed global changes in groundwater resources due to climate change, identifying Brazil and central Andes as sensible areas. Among research studies applying socio-hydrological modeling, different climate scenarios were considered by García et al. (2019), and climate change is taken into account in most studies related to drought and groundwater depletion (Noël and Cai 2017; Castilla-Rho et al. 2019).

  • Lack of institutional coordination. Inadequate governance impedes integrated groundwater management, perpetuating conflicts. Related studies are from Foster et al. (2013), who underscored the importance of governance structures in sustainable water management, highlighting the need for collaboration among government, communities, and industries; and from Giordano and Villholth (2007), who analyzed the agricultural groundwater revolution, pointing to opportunities and threats to development while emphasizing the role of institutions. Furthermore, the work of Hoogesteger (2022) highlights the role of policies and interventions as means to advance towards sustainable groundwater use, citing Mexico, Peru and Chile as areas of intensive groundwater extraction for agricultural purposes. As the socio-hydrological approach allows including the interaction of institutional agents with other agents’ types and with the aquifer, strategies addressing institutional and governance problems can be simulated and explored. Examples can be found in Genova and Wei (2023), Du et al. (2022), Sadaf and Jabbar (2022), Anbari and Zarghami (2021), Aghaie et al. (2020a, b), Du et al. (2020), Mehryar et al. (2019), Ohab-Yazdi and Ahmadi (2018) and Bell et al. (2016).

  • Inadequate monitoring and poor data availability. Insufficient data hampers evidence-based decision-making, hindering sustainable groundwater management. In this context, Foster and MacDonald (2014) highlighted the importance of robust monitoring networks for sustainable groundwater management, pointing out to the role of technology in enhancing data collection. The scarce and unevenly distributed availability of reliable and systematic data in Latin America, both hydrogeological and socioeconomic, has been recognized in previous research (Rama et al. 2022; Mahlknecht et al. 2020), and the investment required to obtain them can still be considered an obstacle to the development of coupled models. However, it can also be seen as an opportunity to properly plan which data are required and thus optimize the allocation of resources to obtain them. In addition, another way to envision the possible scarcity of data as an opportunity is to promote data collection strategies through research that requires input from the community connected to the water resource, i.e., science from the stakeholders themselves, as proposed by Huggins et al. (2023) and Herrera (2019). This strategy also has the advantage of contributing to the creation of citizen awareness regarding the environmental system under study (Kimura and Kinchy 2016). It is also relevant to highlight that the activities necessary for the sustainable management of the resource, from contributing to data collection to regulating extraction, are more likely to be effective if they are included in a “bottom-up” approach, i.e., involving direct users and local authorities in their definition (Castilla-Rho et al. 2019).

Relevance of coupled socio-hydrological models for the region

This review allowed the identification of advantages of the coupling of ABMs and hydrogeological models mentioned by previous studies that could encourage their application in Latin America. While the required hydrogeological data might be available for aquifer systems previously characterized for constructing conventional groundwater flow models, the data required for develo** an ABM could represent an obstacle. On this subject, Mahlknecht et al. (2020) mentioned that currently available socio-economic information in the Latin American region is limited due to poorly conceptualized monitoring programs. In this sense, it can be highlighted that population surveys and conventional social research strategies generally require long planning and involve the consumption of resources that are not always available (Castilla-Rho et al. 2019). Counter to the former, coupled models allow analyzing different scenarios and conducting experiments in an “artificial society” of adaptive agents that would be very costly or impossible to carry out in reality (Castilla-Rho et al. 2015). These computational models have the ability to simulate different levels of cooperation, manifestations of cultural values and social norms (Castilla-Rho et al. 2019). Among the disadvantages found by the studied articles that could be relevant regarding the application in the region of interest, it can be mentioned that in several cases, the requirement of data and computational resources may be high and therefore not easily attainable in the context of medium and low-income countries (Eyni et al. 2021). It is also worth pointing out that model calibration could not be carried out in many cases because all scenarios posed are hypothetical (Aghaie et al. 2020a, b, 2021; Noël and Cai 2017); and that model validation may present difficulties for similar reasons as mentioned above (Mehryar et al. 2019).

Based on the findings of this study, the results of this type of model can contribute to establishing sustainable management strategies that take into account the possible activities and decisions of different groups of users whose interests compete with each other. The possibility of modeling these users as adaptive agents allows the study of several scenarios that can represent different policies, restrictions or incentives applied by the local authority in charge of groundwater management, which is included in the model as a regulatory agent. In addition, some of the tools developed and applied in the reviewed works allow modifying certain parameters in real-time so that the model can interact with participants of public hearings (Castilla-Rho et al. 2015). These features might improve the research by capturing possible users’ decisions and feedback on the model, in parallel to allowing local users to visualize the impact of their practices over time, also constituting a potential contribution to environmental education.

As a relevant differentiating element with respect to exclusively hydrogeological modeling, hydrogeological models coupled with ABMs provide the possibility of understanding and modeling the demand to which a groundwater system is subjected in a much more complete way than conventional flow models. This assertion is based on two aspects: (1) the approach is broadened with respect to conventional hydrogeological flow models, expanding the boundaries of the system under study to include not only the water resource as such but also the users and local authorities, in accordance with the very concept of sustainability; and (2) the level of detail is greater because the utilization can be characterized from each user or group of users, and these in turn are able to adapt their decisions to the evolution of the state of the aquifer over time; thus allowing the visualization of the usually hidden resource of groundwater through its interaction with the interests of the different agents (Huggins et al. 2023).

Moreover, this characteristic corresponds to bottom-up strategies, which make it possible to adjust the modeling to the local population of the study site. However, it is important to highlight that the results that can be obtained will largely depend on the precision with which the model can capture the users’ decisions, considering that the willingness to comply with the restrictions imposed by a regulatory agent is site-specific and in general cannot be generalized or extrapolated (Castilla-Rho et al. 2019). This particular capability of these coupled models allows connecting their application with the search for solutions posed from demand management and not exclusively from the increase in water availability, which could be more effective in targeting the sustainability of groundwater resources according to the results of Martínez-Valderrama (2023).

Current and potential applications of coupled ABM-hydrogeological models in Latin America

As examples of socio-hydrological models developed for the Latin American aquifers, the studies by Genova and Wei (2023) and García et al. (2019) can be cited. The former aimed at tracking the evolution of groundwater uses for socio-economic and ecological purposes in the Maipo River Basin in Chile and simulating the response of the groundwater system to water allocation and environmental regulations in terms of economic and ecological outcomes. Key findings of this work were a method to quantify the trajectory of water environmental regulations, providing an independent quantification of these regulations as the social component of the model which allowed to measure it as a function of environmental degradation. The previously mentioned limitations of this study are the capturing of the basin characteristics influencing the water balance, the consideration of only formal environmental regulations, the simplification of ecological outcomes and the neglect of drivers for resource allocation other than agriculture, drinking water and hydropower generation. On the other hand, the model presented by García et al. (2019) studied a shallow aquifer in the Argentine Pampas. The coupled socio-hydrological model aimed to exploring socio-economic and environmental outcomes of unexperienced climate scenarios, hypothetical economic policies and nonstructural technologies such as the adoption of cover crops in winter to manage water-table depth. The hydrogeological submodel for this investigation was a previously developed and calibrated, fully distributed and physically based groundwater model; the ABM coupled to it was also a previously designed land-use decision model including a realistic description of human behavior, decision making and social interaction. Both studies focused on crop irrigation as one of the relevant groundwater extraction drivers, which corresponds to the application trend for the coupled modeling framework observed in the reviewed studies; and both included the exploration of the impact of regulations and policies on economic and environmental outcomes. However, they differ in the selected modeling structure: García et al. (2019) presented a loosely linked coupling of the groundwater and ABM models, while Genova and Wei (2023) worked through an integrated, process-based socio-hydrological model.

In the Latin American region, particularly in South America, there are 29 large transboundary aquifers (IGRAC and UNESCO-IHP 2015). Moreover, the need to develop regional and collaborative management strategies for these collective resources, such as the Guarani Aquifer System (shared between Brazil, Argentina, Paraguay and Uruguay), facing current trends of population growth and climate change is broadly recognized (Espindola et al. 2021; Alcañiz and Berardo 2016). For this type of system, the definition of joint regulations for exploitation by users from different countries, the different socioeconomic drivers and the conservation interests of each country can be represented and explored by means of flow models coupled with ABMs. Mention can also be made of the Copiapó basin in Chile, where agent-based modeling could be applied to study abstraction rights transactions between surface water and groundwater users, or different scenarios for their regulation. Both systems are mentioned as examples of possible applications of the coupled model methodology with ABMs presented in Castilla-Rho et al. (2015). As other case studies that can be considered indicators that the socio-hydrogeological approach is beginning to be relevant in Latin America, the work of Carrión-Mero et al. (2021) can be mentioned, where a conceptual model of the Valdivia River basin, southwest of Ecuador, was developed. Among other results, the conceptualization stage of the functioning of this system allowed the generation of proposals for its sustainable management, which included the establishment of irrigation patterns, the need to define recharge zones and monthly monitoring of water quantity and quality through participatory methods involving water resource users. Considering that the conceptual model is usually the previous step for a computational model, this work could constitute the basis for the second application of the modeling approach of an aquifer as a socio-hydrological system in Latin America. Additionally, the research of Rodrigues-Capítulo et al. (2017) can be identified as another case study in the region to which socio-hydrological modeling could be applied. These authors studied the impact of geomorphological evolution and its consequences for social and economic development in two coastal areas in the province of Buenos Aires, Argentina. In this work, the hydrogeological study of the groundwater system was detailed, and the basis for its connection with the development of high-cost real estate ventures is established by the socioeconomic data collected. The sustainability of groundwater development in this area to supply both the local population and emerging developments could be analyzed through flow modeling coupled with ABMs.

Finally, as a limitation regarding this literature review methodology, the nonstandard way in the description of the various modeling approaches somehow hindered the selection and classification of the studies. It is worth pointing out that the studies following the overview, design and detail (ODD) protocol (Grimm et al. 2010, 2017, 2020), allowed a more straightforward selection, classification and reporting.

Conclusions

From the literature review conducted for this study, it can be stated that the coupled modeling approach between ABM and groundwater flow models has been increasing in its rate of application in the last decade. The vast majority of the reviewed research studies included agriculture as one of the focused drivers for groundwater use, although other uses of the resource were studied conjunctively in some cases. Current modeling structures can be divided into two broad groups—on one hand, coupled complex existing groundwater flow models with an ABM developed for the specific study case; on the other, integrated platforms encompassing the ABM and the groundwater flow model in a tightly coupled manner. Data requirements, computational resources consumption and level of aggregation for the results are consequently different for both mentioned model structures. Regarding the coupled model’s components separately, characteristics differencing the socio-economic component from the hydrogeological one are determined to be the availability of reliable data required for their development and the possibility of proper calibration, which is hampered in the case of the former.

Based on the connection of the problems focused in the reviewed research articles and the recognized Latin American groundwater challenges, it can be interpreted that the study of aquifers focused on as socio-hydrological systems can contribute greatly to the sustainable management of these resources in Latin America. Attending to the heterogeneity of groundwater systems, a thorough understanding of the local characteristics of aquifers encompassing not only hydrogeological parameters but socio-economic characteristics as well is fundamental for the implementation of several of the sustainable development goals (SDGs) (Velis et al. 2017), particularly SDG 6 (clean water and sanitation). Groundwater is a widely distributed but essentially local resource, so its sustainable management is only possible if the needs of direct users and the interests of local communities and authorities are included both in understanding its functioning and in deciding on management strategies. In this sense, socio-hydrological models constitute an emerging strategy not only for understanding these systems but also for studying the environmental, social and economic outcomes of different management approaches. The difficulty of determining, at least in early stages, the impacts of human activity on aquifers and the perception supported by research results (Rama et al. 2022) that groundwater is abundant and of low vulnerability in much of Latin America, can be considered major challenges for sustainable groundwater management in the region. As strategies to address these challenges, several researchers mention early planning, particularly for integrated land use and groundwater management (Foster et al. 2013); to this could be added the need for adequate local regulation, harmonious articulation between countries (considering that large aquifers in the region are transboundary) and cross-cutting attention to environmental education. Hydrogeological models coupled with ABMs can become a fundamental tool within these strategies aimed at the sustainable management of aquifers in Latin America.