Impact statement

As the largest consuming water sector (by far), agriculture is a domain where food and water security are strongly interlinked. Sociohydrology has offered post-positivist approaches to understand the dynamics of human–water relationships. This study highlights the contributions of sociohydrology in the agricultural domain. Sociohydrology has been able to describe how agricultural water management actions have often led to counterproductive, often unintended consequences by integrating human behavioural dynamics, interdisciplinary methods and novel data sources. By incorporating concepts from resilience thinking, sociohydrology can emerge as an approach to understand the susceptibility of agricultural systems and agriculturalists to global change and uncertainties, potentially better informing climate-resilient agricultural policy.

Sociohydrological understanding of emergent dynamics in agriculture water systems

Agricultural challenges are complex and context-specific, often resulting from the interplay of factors such as population requirements, governance, water scarcity and climate. In many regions, challenges of increasing food production are linked to agricultural water management (i.e., irrigation management and/or a socioecological imbalance between the approach to agricultural development and ecological limits including available water resources) (Pimentel et al., Reference Pimentel, Berger, Filiberto, Newton, Wolfe, Karabinakis, Clark, Poon, Abbett and Nandagopal2004; Pereira, Reference Pereira2017; Turner et al., Reference Turner, Hejazi, Yonkofski, Kim and Kyle2019). Future irrigation management will likely require increased irrigation coverage and improved irrigation efficiency, through central and decentralised infrastructure/measures, particularly in the face of global change projections. The increasing demands for finite water resources in agriculture have led to a focus on the need for shifting towards demand-side management of water in agriculture instead of supply-driven solutions (Di Baldassarre et al., Reference Di Baldassarre, Wanders, AghaKouchak, Kuil, Rangecroft, Veldkamp, Garcia, van Oel, Breinl and Van Loon2018; Garrick et al., Reference Garrick, Hanemann and Hepburn2020). Such infrastructural irrigation solutions can be complemented by improved water management, through measures like using locally adapted crop varieties (e.g., avoiding water-intensive crops in water-scarce areas), water-smart production methods (e.g., soil management and other techniques to minimise water loss through evapotranspiration) and water harvesting (Oweis and Hachum, Reference Oweis and Hachum2006; Castelli et al., Reference Castelli, Castelli and Bresci2019).

Yet, ill-planned implementation of these interventions can lead to negative externalities and unintended and unexpected impacts on hydrological and social systems (Alam et al., Reference Alam, McClain, Sikka and Pande2022). Human adaptation can lead to a “lock-in” towards unintended or undesired states (Pouladi et al., Reference Pouladi, Nazemi, Pouladi, Nikraftar, Mohammadi, Yousefi, Yu, Afshar, Aubeneau and Sivapalan2022; Prasad et al., Reference Prasad, Damani and Sohoni2022). Irrigation investments can significantly alter the availability and allocation of water flows in a given region, leading to hydrological externalities such as reduced runoff, upstream–downstream impacts, decreased groundwater recharge and baseflows (Calder et al., Reference Calder, Gosain, Rao, Batchelor, Snehalatha and Bishop2008; Bouma et al., Reference Bouma, Biggs and Bouwer2011; Alam et al., Reference Alam, McClain, Sikka and Pande2022). At the same time, societies can respond to these interventions and new hydrological conditions in unpredictable and nonlinear ways (Zhang et al., Reference Zhang, Hu, Tian, Yao and Sivapalan2014; Alam et al., Reference Alam, McClain, Sikka and Pande2022).

For example, improvements in irrigation water-use efficiency have not always resulted in effective or equitable allocation of finite water resources (Perry and Steduto, Reference Perry and Steduto2017; Grafton et al., Reference Grafton, Williams, Perry, Molle, Ringler, Steduto, Udall, Wheeler, Wang and Garrick2018). Rather, such interventions have often led to increased water use instead of the expected reduction (Zhang et al., Reference Zhang, Hu, Tian, Yao and Sivapalan2014; Birkenholtz, Reference Birkenholtz2017). This phenomenon is known as the irrigation efficiency paradox (a sub-phenomenon of the Jevons’ paradox or the rebound effect), which can be explained by the supply–demand cycle – ‘as availability increases, consumption tends to increase’ (Di Baldassarre et al., Reference Di Baldassarre, Sivapalan, Rusca, Cudennec, Garcia, Kreibich, Konar, Mondino, Mård, Pande, Sanderson, Tian, Viglione, Wei, Wei, Yu, Srinivasan and Blöschl2019, Reference Di Baldassarre, Wanders, AghaKouchak, Kuil, Rangecroft, Veldkamp, Garcia, van Oel, Breinl and Van Loon2018).

Sociohydrology explicitly accounts for the changing and adaptive responses of humans and their impact on the environment. Sociohydrological models have investigated many of these externalities, unintended consequences and emerging phenomena (see Table 1 for some examples). For example, Ghoreishi et al. (Reference Ghoreishi, Razavi and Elshorbagy2021a, Reference Ghoreishi, Sheikholeslami, Elshorbagy, Razavi and Belcher2021b) developed and applied an agricultural water demand model to understand and model the phenomenon of the rebound effect in the Bow River basin in Canada where water conservation strategies to reduce water usage have in fact led to increased irrigation. Birkenholtz (Reference Birkenholtz2017) showed that policies supporting drip irrigation to reduce water use paradoxically led to crop intensification and increased groundwater extraction. Ilyas et al. (Reference Ilyas, Manzoor and Muhammad2021) developed a system dynamic sociohydrological model to understand and simulate the phenomenon of irrigation efficiency paradox (Jevons’ paradox for irrigation efficiency), where increasing on-farm irrigation efficiency does not lead to increased water availability at the basin scale. Similarly, Kuil et al. (Reference Kuil, Evans, McCord, Salinas and Blöschl2018) showed how the rebound effect can be captured by developing a sociohydrological model framework which links farmers’ perceptions of water availability to their crop choice and water allocation in Kenya’s Upper Ewaso Ng’iro basin. Pouladi et al. (Reference Pouladi, Nazemi, Pouladi, Nikraftar, Mohammadi, Yousefi, Yu, Afshar, Aubeneau and Sivapalan2022) provided a sociohydrological explanation of the lock-in effect of lake desiccation and soil salinisation of Lake Urmia (Iran), which resulted from interacting anthropogenic and (surface and subsurface) environmental processes, while Prasad et al. (Reference Prasad, Damani and Sohoni2022) explained lock-in in irrigation from groundwater emerging from ‘aspirational’ and ‘vulnerability’ intensification that is mediated by agriculturalists’ perception of income risks in horticulture in Maharashtra (India).

Table 1. Different sociohydrological studies exploring unintended consequences and emergent phenomena

Garrick et al. (Reference Garrick, Stefano, Yu, Jorgensen, O’Donnell, Turley, Aguilar-Barajas, Dai, Leão, Punjabi, Schreiner, Svensson and Wight2019) have pointed out that water reallocations from agriculture to cities have been documented in many regions across the world. Fuelled by growing urban populations, such transfers are shown to have inequitable outcomes both within cities and for rural agriculturalists (Boelens et al., Reference Boelens, Vos, Perreault, Vos, Boelens and Perreault2018). Such transfers may challenge not only food systems but also especially impact smallholder farmers and their associated livelihoods and cultures. Garrick et al. (Reference Garrick, Stefano, Yu, Jorgensen, O’Donnell, Turley, Aguilar-Barajas, Dai, Leão, Punjabi, Schreiner, Svensson and Wight2019) indicated that water-use efficiency measures can either be a driver or be adopted as a result of such rural–urban transfers. Such efforts to increase water-use efficiency have often triggered complex power dynamics with impacts on surface water irrigation management where tail-end and smaller farmers are adversely impacted (Hu et al., Reference Hu, Yang, Han, Yang, Ai, Wang and Ma2017; Linstead, Reference Linstead2018). Such phenomena are important topics to address with sociohydrology. Table 1 provides an overview of such phenomena related to agricultural water studies in sociohydrology.

Not accounting for such externalities and human–water feedbacks can lead to limited understanding of unsustainable outcomes such as drying of reservoirs and wetlands, groundwater depletion, and water, soil and ecosystem deterioration. Additionally, these negative impacts are often mediated by weak financial capital, knowledge, gender and power relations, leading to affect certain and often poor and marginalised populations disproportionately. Often rich or influential farmers with more access to social, financial and biophysical capital capture more advantages, more subsidies and more benefits, thus exacerbating existing inequalities. Thus before investing, it is crucial to consider potential negative externalities and feedbacks to avoid reinforcing existing inequalities and long-term natural resource degradation.

Sociohydrology in addressing climate change

Climate change is affecting the availability of the water resources vital to agriculture via changes to precipitation patterns (quantity, intensity, types and timing), snowmelt timing and rate, stream flows, groundwater recharge, soil moisture, increased temperature and hence atmospheric water demands and evapotranspiration (Jimenez Cisneros et al., Reference Jimenez Cisneros, Oki, Arnell, Benito, Cogley, Doll, Jiang, Mwakalila, Field, Barros, Dokken, Mach, Mastrandrea, Bilir, Chatterjee, Ebi, Estrada, Genova, Gimma, Kissel, Levy, MacCracken, Mastrandrea and White2014; Ficklin and Novick, Reference Ficklin and Novick2017; Masson-Delmotte et al., Reference Masson-Delmotte, Zhai, Pirani, Connors, Péan, Berger, Caud, Chen, Goldfarb, Gomis, Huang, Leitzell, Lonnoy, Matthews, Maycock, Waterfield, Yelekçi, Yu and Zhou2021). A reduction in water available for crops is expected (Cook et al., Reference Cook, Mankin, Marvel, Williams, Smerdon and Anchukaitis2020). The increased frequency of intense precipitation and long dry spells and changes in their timing (Breinl et al., Reference Breinl, Baldassarre, Mazzoleni, Lun and Vico2020) are expected to further increase variability in water availability. Together, these changes are projected to reduce average crop yields (e.g., Hasegawa et al., Reference Hasegawa, Wakatsuki, Ju, Vyas, Nelson, Farrell, Deryng, Meza and Makowski2022) and their year-to-year stability (Challinor et al., Reference Challinor, Watson, Lobell, Howden, Smith and Chhetri2014). Impacts additionally reverberate throughout non-production aspects of local and global food systems (i.e., energy consumption, storage, transportation and food safety). Smallholder farmers can be particularly vulnerable, having limited financial capital, debt cycles and dependence on rainfed agriculture. While climate trends are a challenge for average crop yields and their stability, and hence food security, the impact and extent of such trends and extremes can be exacerbated or alleviated by human decisions and management (Kreibich et al., Reference Kreibich, Schröter, Di Baldassarre, Van Loon, Mazzoleni, Abeshu, Agafonova, AghaKouchak, Aksoy, Alvarez-Garreton, Aznar, Balkhi, Barendrecht, Biancamaria, Bos-Burgering, Bradley, Budiyono, Buytaert, Capewell, Carlson, Cavus, Couasnon, Coxon, Daliakopoulos, de Ruiter, Delus, Erfurt, Esposito, François, Frappart, Freer, Frolova, Gain, Grillakis, Grima, Guzmán, Huning, Ionita, Kharlamov, Khoi, Kieboom, Kireeva, Koutroulis, Lavado-Casimiro, Li, LLasat, Macdonald, Mård, Mathew-Richards, McKenzie, Mejia, Mendiondo, Mens, Mobini, Mohor, Nagavciuc, Ngo-Duc, Nguyen, Nhi, Petrucci, Quan, Quintana-Seguí, Razavi, Ridolfi, Riegel, Sadik, Sairam, Savelli, Sazonov, Sharma, Sörensen, Souza, Stahl, Steinhausen, Stoelzle, Szalińska, Tang, Tian, Tokarczyk, Tovar, Tran, van Huijgevoort, van Vliet, Vorogushyn, Wagener, Wang, Wendt, Wickham, Yang, Zambrano-Bigiarini and Ward2023).

Adaptations to changed growing conditions have been proposed, such as changes in crops and their planting dates, enhanced crop diversification in space and time, drought-resistant crop varieties, water and soil conservation practices, agroecology (i.e., agricultural production drawing on ecological principles and traditional management which may be particularly useful to smallholder farmers) and irrigation application. While these approaches have the potential to reduce the impact of detrimental climatic conditions (e.g., Hasegawa et al., Reference Hasegawa, Wakatsuki, Ju, Vyas, Nelson, Farrell, Deryng, Meza and Makowski2022), they might clash with other needs and preferences, or have unintended consequences.

To exemplify these mechanisms, we consider multiple cases, beginning with irrigation as a measure to climate change adaptation. Irrigation can substantially reduce the impact of projected heat and water stress (e.g., Siebert et al., Reference Siebert, Webber, Zhao and Ewert2017; Luan and Vico, Reference Luan and Vico2021). Model estimates show that deficit irrigation (i.e., irrigation aimed at stabilising yields and maximising water productivity, but not necessarily maximising yields; Zhang and Oweis, Reference Zhang and Oweis1999) could be sustainably expanded in approximately a third of currently rain-fed croplands under a + 3 °C warming (Rosa et al., Reference Rosa, Chiarelli, Sangiorgio, Beltran-Peña, Rulli, D’Odorico and Fung2020). Yet, the choice of water source for irrigation affects and is affected by the sociohydrological system, requiring difficult decisions. For example, crop yield maximisation and risk minimisation emerged as increasingly hard to reconcile under more extreme climatic conditions, according to a model for sizing small on-farm ponds as sustainable source of water for irrigation.

At the regional scale, the choice of source of irrigation water is not just a question of water availability (Bierkens and Wada, Reference Bierkens and Wada2019) but also a problem of collective action. An agent-based model (ABM) showed a single farmer can attain the highest average economic return exploiting a communal resource (groundwater) in a community that is otherwise long-view oriented, that is, privileges renewable water sources (e.g., small farm ponds; Tamburino et al., Reference Tamburino, Di Baldassarre and Vico2020). Nevertheless, the advantage diminished under more extreme climatic conditions. Considering the entire community and the evolution of choices of water source based on previous experiences, an intensification of climatic extremes reduced the fraction of long-view farmers, that is, those relying on renewable water sources instead of nonrenewable groundwater, and caused a worsening of average economic gain and its stability. This is the consequence of nonrational decisions (Di Baldassarre et al., Reference Di Baldassarre, Sivapalan, Rusca, Cudennec, Garcia, Kreibich, Konar, Mondino, Mård, Pande, Sanderson, Tian, Viglione, Wei, Wei, Yu, Srinivasan and Blöschl2019; Schill et al., Reference Schill, Anderies, Lindahl, Folke, Polasky, Cárdenas, Crépin, Janssen, Norberg and Schlüter2019) and experiential lock-ins (Payo et al., Reference Payo, Becker, Otto, Vervoort and Kingsborough2016) and might be exacerbated by policies that undermine the sociocultural underpinnings of collective action in traditional societies (Basel et al., Reference Basel, Hernández Quiroz, Velasco Herrera, Santiago Alonso and Hoogesteger2021).

All in all, climate change and feedbacks between social, hydrological and climate systems (Basel et al., Reference Basel, Hoogesteger and Hellegers2022) could enhance the environmental fragility–economic poverty vicious circle (Cheng et al., Reference Cheng, Dong, Li, Yang, Li and Li2019) and undermine food systems, leading to unintended consequences. These feedbacks and the place for sociohydrological understanding of collective, traditional or individualistic behavioural outcomes are particularly strong when considering smallholder farming. For example, cropping decisions in the Dhidhessa basin (Ethiopia) were influenced by lower sensitivity of crop prices to rainfall variability and farmers’ expectations regarding the same, inducing more climate resilience into the system, unravelling the endogenous role of crop prices resulting from the culture of crop diversification and leading to climate-resilient agriculture (Teweldebrihan et al., Reference Teweldebrihan, Lyu, Pande and McClain2021). In Oaxaca (Mexico), agriculture-dependent Zapotec communities experiencing extended drought self-organised for collective action to implement small-scale managed aquifer recharge (MAR; Basel et al., Reference Basel, Hernández Quiroz, Velasco Herrera, Santiago Alonso and Hoogesteger2021). Here however, if the individual water-use behaviour is not distinguished from the adoption of technologies such as MAR, it may exacerbate the drought conditions due to overuse of saved water due to misperception of water abundance (i.e., another example of the rebound effect) (Kallis, Reference Kallis2010; Gohari et al., Reference Gohari, Eslamian, Mirchi, Abedi-Koupaei, Massah Bavani and Madani2013; Di Baldassarre et al., Reference Di Baldassarre, Sivapalan, Rusca, Cudennec, Garcia, Kreibich, Konar, Mondino, Mård, Pande, Sanderson, Tian, Viglione, Wei, Wei, Yu, Srinivasan and Blöschl2019; Basel et al., Reference Basel, Hernández Quiroz, Velasco Herrera, Santiago Alonso and Hoogesteger2021). Meltwater-dependent irrigated agriculture, which is crucial for livelihood security of mountain communities, also requires adaptive efforts for climate change resilience (Nüsser et al., Reference Nüsser, Dame, Parveen, Kraus, Baghel and Schmidt2019b). Changes in seasonal snow cover can lead to water scarcity; such situations can be compounded by infrastructural interventions such as dams (Nüsser, Reference Nüsser2017; Nüsser et al., Reference Nüsser, Dame, Kraus, Baghel and Schmidt2019a). Collective action or not, adapting to such changes through adoption of ’artificial glaciers’, have to be understood in terms of the benefits perceived by farmers, including perceived reduction in water scarcity, perceived risks of crop failure and the possibility to grow cash crops (Nüsser et al., Reference Nüsser, Dame, Kraus, Baghel and Schmidt2019a).

These examples, among many, clearly show the importance of considering not only whether a specific climate change adaptation is technically feasible, but more importantly whether it is also socially and individually acceptable (i.e., in line with perceived risk, attitudes, abilities and preferences), in how without these measures the human–agricultural water systems may evolve to undesirable states. Sociohydrology recognises that humans are the agents of change, whose actions are mediated by their perceptions of risks, cultural norms, perceptions of risks and ownership of resources. These perceptions may vary from person to person, as a result of which individuals in similar environments respond differently in terms of, for example, their use of water (Daniel et al., Reference Daniel, Pande and Rietveld2022). Moreover, individuals tend to continually develop their unique lenses to focus, categorise and interpret information (Elliott, Reference Elliott2003) and perceive the world differently based on personal experiences and reflective learning (Kuil et al., Reference Kuil, Evans, McCord, Salinas and Blöschl2018). This calls for novel methods based on behavioural theories to improve the understanding of human decision-making that is unique to sociohydrology (Bertassello et al., Reference Bertassello, Levy and Müller2021). Without effective behavioural prediction, efforts to introduce policy and initiatives for innovation to build resilience against global environmental change may turn out to be ineffective (Weersink and Fulton, Reference Weersink and Fulton2020) (Figure 1).

Figure 1. Different case studies related to agriculture investigated by sociohydrological research covered in this review. The size of the bubble indicates the number of publications associated with the particular region (which is mentioned in brackets, if >1).

Incorporating human behaviour into sociohydrological understanding

In earlier sociohydrological works, efforts to endogenise human agency (Pande and Sivapalan, Reference Pande and Sivapalan2017) led to the conception of social state variables such as environmental awareness (van Emmerik et al., Reference van Emmerik, Li, Sivapalan, Pande, Kandasamy, Savenije, Chanan and Vigneswaran2014) or community sensitivity (Elshafei et al., Reference Elshafei, Sivapalan, Tonts and Hipsey2014) (e.g., towards the health of the environment), that are consistent with the values, beliefs and norms sociological theory (Wei et al., Reference Wei, Wei and Western2017). These were incorporated into system dynamics models to explain emergent phenomena of pendulum swings in agricultural systems (Roobavannan et al., Reference Roobavannan, Kandasamy, Pande, Vigneswaran and Sivapalan2017). Other models assumed that humans are rational beings able to evaluate options and maximise their well-being in interpreting rise and dispersal of water-scarce societies (Pande et al., Reference Pande, Ertsen and Sivapalan2014).

Several studies have developed dynamic sociohydrological models based on insights from psychology and sociology and applied them in agricultural contexts. Kuil et al. (Reference Kuil, Evans, McCord, Salinas and Blöschl2018) investigated smallholders’ cropping and water allocation choices by developing a sociohydrological model based on the theory of bounded rationality (which assumes that rational behaviour is bounded by practical considerations like the availability of information and time, Simon, Reference Simon1955). With an application in Kenya’s Upper Ewaso Ng’iro, the model simulated the rebound effect and found that while different perceptions of water availability could lead to different cropping patterns, near-optimal similar cropping patterns would also emerge. Studies have also used agent-based models (ABMs) extensively to capture heterogeneities in socioeconomic characteristics and behaviour of farmers (Pouladi et al., Reference Pouladi, Afshar, Afshar, Molajou and Farahmand2019, Reference Pouladi, Afshar, Molajou and Afshar2020; Kasargodu Anebagilu et al., Reference Kasargodu Anebagilu, Dietrich, Prado-Stuardo, Morales, Winter and Arumi2021; Wens et al., Reference Wens, Veldkamp, Mwangi, Johnson, Lasage, Haer and Aerts2020, Reference Wens, van Loon, Veldkamp and Aerts2022). Agent-based models were developed to explain the catastrophic shrinkage of Lake Urmia (Iran), using questionnaires and field observations based on the theory of planned behaviour (TPB, which consists of attitudes, subjective norms and perceived control shaping a certain behaviour, Ajzen, Reference Ajzen1991) and a data mining technique (Pouladi et al., Reference Pouladi, Afshar, Molajou and Afshar2020) to generate behavioural rules for farmer agents for performing their agricultural practices. More complex psychological approaches towards behavioural change, built on multiple theories including the TPB, have also been used to explain farmers’ socioeconomic characteristics and behavioural perceptions which drive/prevent them towards/from adopting irrigation technologies in drought-prone regions of Maharashtra, India (Hatch et al., Reference Hatch, Daniel and Pande2022).

A further level of complexity is the social structure and its slow evolution as a result of changes in external drivers (e.g., climatic conditions) and internal dynamics (e.g., via adaptation). For instance, social cohesion (via the role of mutualism) emerged as an important force buffering against water availability stresses, in conceptualising a community-managed irrigation system in New Mexico (Gunda et al., Reference Gunda, Turner and Tidwell2018). Collective human actions in response to water scarcity may have even led to the rise (and dispersal) of ancient civilisations via technological progress, economic growth and environmental degradation, arising from intrinsic endogenous processes, as is shown by Pande and Ertsen (Reference Pande and Ertsen2014) for the Indus and Hokoham civilisations. Moreover, the inclusion of institutional arrangements can disentangle social processes into rules (norms), behavioural choices towards these arrangements (whether society abides by them or opposes them) and situational factors influencing decision-making (like social memory) (Konar et al., Reference Konar, Garcia, Sanderson, Yu and Sivapalan2019).

Such sociohydrological studies suggest that human agents such as agriculturalists often tend to not make rational or optimal decisions, particularly while facing uncertainty in terms of biophysical and socioeconomic circumstances (Tittonell et al., Reference Tittonell, Muriuki, Shepherd, Mugendi, Kaizzi, Okeyo, Verchot, Coe and Vanlauwe2010; Richard Eiser et al., Reference Richard Eiser, Bostrom, Burton, Johnston, McClure, Paton, van der Pligt and White2012; Howley et al., Reference Howley, Buckley, O Donoghue and Ryan2015). In particular, their motivations to adapt or adopt agriculture water practices and technologies can extend beyond purely economic (profit) considerations. These extend towards social, psychological and cognitive factors, for example, perception of risk, that may influence their individual behaviour, and cultural or institutional factors that may condition what farmer may perceive to be a normal practice (Konar et al., Reference Konar, Garcia, Sanderson, Yu and Sivapalan2019; Weersink and Fulton, Reference Weersink and Fulton2020; Nair and Thomas, Reference Nair and Thomas2022). This has been demonstrated in examples on low-adoption of wells and irrigation measures (Sampson and Perry, Reference Sampson and Perry2019; Nair and Thomas, Reference Nair and Thomas2022), among others. Moreover, rather than being a binary (yes/no) decision, adoption-related decision-making is a multistage process, starting with awareness (of the existence of the technology/solution), followed by evaluation (of using the technology), adoption and revision/dis-adoption based on further changes in circumstances (Weersink and Fulton, Reference Weersink and Fulton2020). While economic considerations are important during the later stages of the adoption process, social and cognitive factors may be more influential during the earlier stages (of awareness and evaluation).

While these studies have been geared towards understanding human–agricultural systems better, there is scope to consider aspects that could make such systems more resilient against global environmental change (Penny and Goddard, Reference Penny and Goddard2018). The body of knowledge generated by incorporating human behaviour informs us about how agriculturalists may adopt good agricultural interventions. The gap which remains is to identify the set of interventions that would make the system resilient for agriculturalists. Sociohydrological studies have highlighted, for example, that the Murrumbidgee River basin was more resilient due to its diversified economy and therefore was able to transition to a less-water-intensive economy within the basin. However this proved to be difficult for desiccated lake basins (like Aral Sea) whose economy singularly depended on agriculture (Pande et al., Reference Pande, Roobavannan, Kandasamy, Sivapalan, Hombing, Lyu and Rietveld2020). This alludes to ‘diversity’, one desirable property (of many) exemplifying resilient systems. The following section discusses the way forward for understanding what resilience could mean to agriculturalists.

Emerging sociohydrological issues: Integrating resilience thinking into agricultural water and for agriculturalists

Resilience, the capacity of systems to function and develop with change and uncertainty is a vital need for agricultural systems (Morecroft et al., Reference Morecroft, Crick, Duffield and Macgregor2012; Ward, Reference Ward2022). While resilience concepts may be inherent to sociohydrology, there is much scope to incorporate key resilience attributes (diversity, redundancy, connectivity, adaptive learning and inclusivity and equity) more explicitly while conceptually framing sociohydrological problems (Penny and Goddard, Reference Penny and Goddard2018; Rockström et al., Reference Rockström, Norström, Matthews, Biggs, Folke, Harikishun, Huq, Krishnan, Warszawski and Nel2023), potentially improving interpretations of unintended consequences.

Diversity (e.g., biodiversity and diversified livelihoods) increases system flexibility through the ability to respond in multiple ways to systemic changes or external shocks (Rockström et al., Reference Rockström, Norström, Matthews, Biggs, Folke, Harikishun, Huq, Krishnan, Warszawski and Nel2023). Redundancy ensures system functioning and reduces the chances of single-point failure (Rockström et al., Reference Rockström, Norström, Matthews, Biggs, Folke, Harikishun, Huq, Krishnan, Warszawski and Nel2023). Diversity and redundancy work in tandem so that multiple functionally similar system elements respond differently to disturbances to enhance resilience (Penny and Goddard, Reference Penny and Goddard2018). Sociohydrological studies could more explicitly incorporate diversity and redundancy attributes to test hypotheses such as the effect of too little or too much diversity on system resilience (Penny and Goddard, Reference Penny and Goddard2018).

Connectivity relates to how resources, information, species or people interact within systems and higher connectivity can facilitate knowledge transfer, learning and enhance resilience (Penny and Goddard, Reference Penny and Goddard2018; Rockström et al., Reference Rockström, Norström, Matthews, Biggs, Folke, Harikishun, Huq, Krishnan, Warszawski and Nel2023). While sociohydrological models have incorporated connectivity through the flow of capital (Pande and Savenije, Reference Pande and Savenije2016), migration of people (Elshafei et al., Reference Elshafei, Sivapalan, Tonts and Hipsey2014) or pathways of processing information (Kuil et al., Reference Kuil, Evans, McCord, Salinas and Blöschl2018), intra-community connectivity can be further explored through agent-based analyses of social networks (Penny and Goddard, Reference Penny and Goddard2018).

Learning can be both individual and collective (social or institutional) and has been conceptualised as a multilayered iterative process (Medema et al., Reference Medema, Wals and Adamowski2014). Single-loop learning, focussing on correcting errors by changing behaviour, is evident in how Pande and Savenije (Reference Pande and Savenije2016) model smallholder farmers to adjust their actions based on their capital (Medema et al., Reference Medema, Wals and Adamowski2014; Penny and Goddard, Reference Penny and Goddard2018). Double-loop learning, which includes an examination of the underlying values and policies within a system to facilitate such corrections, can be seen in community sensitivity studies where societies change their underlying behaviour from environmentally exploitative to restorative (Elshafei et al., Reference Elshafei, Sivapalan, Tonts and Hipsey2014; Medema et al., Reference Medema, Wals and Adamowski2014; Penny and Goddard, Reference Penny and Goddard2018). Similarly, while learning (generation of new knowledge) and consequent experimentation (by farmers in their decision-making) has been explored in some studies (Kuil et al., Reference Kuil, Evans, McCord, Salinas and Blöschl2018), such dynamics can be further explored in the context of adoption or behavioural change.

Equity and inclusion, via participation (i.e., engaging relevant stakeholders in managing and governing systems), is key to resilience due to its role in collective action and cooperation and can be shaped by power dynamics (Leitch et al., Reference Leitch, Cundill, Schultz and Meek2015; Hahn and Nykvist, Reference Hahn and Nykvist2017). Inclusive and equitable value chains with social safety nets for vulnerable sections can enhance resilience (Rockström et al., Reference Rockström, Norström, Matthews, Biggs, Folke, Harikishun, Huq, Krishnan, Warszawski and Nel2023). Without addressing and transforming historical inequities, rapid shifts from top–down approaches to more holistic ones will not result in increased resilience and successful adaptation. This is applicable to both water used by agriculturalists and in agriculture. Adding power dynamics as inputs that vary water manipulation can lead to understanding alternative outputs that affect people based on their sociodemographics.

Referring to water used by agriculturalists, accessibility refers to the ability of people to access the physical, economic and information means to attain nearby water safely and without discrimination, including the ability of people to participate in water decision-making. Unfortunately, discrimination undermines attempts for universal access. Some people are prevented from accessing water because of their race, ethnicity, nationality, income, gender, ability, age or other social categories. For example, in an investigation of arsenic contamination in Bangladesh, Sultana (Reference Sultana2007) drew attention to the numerous ways through which gender, intersecting with social class, age and geographic location, influenced water access, control and exposure to water. Yet, at least in low-income countries, women often control water reliability in rural areas through water hauling (Zwarteveen, Reference Zwarteveen1997; Meinzen-Dick and Zwarteveen, Reference Meinzen-Dick and Zwarteveen1998; Were et al., Reference Were, Roy and Swallow2008). At least one study found that women leaders invested more in drinking water resources than men (Chattopadhyay and Duflo, Reference Chattopadhyay and Duflo2004). Despite the global advances made in gender equity, women commonly feel excluded and underrepresented in decision-making around water (Haeffner et al., Reference Haeffner, Galvin and Vázquez2017). As water becomes less of a service and more of a commodity to be bought and sold on the open market, households need purchasing power to access water (Bakker, Reference Bakker2001). However, water prices have outpaced inflation (Cardoso and Wichman, Reference Cardoso and Wichman2022). Water affordability refers to the price households pay for water but is also related to the costs of extracting, treating and delivering water since these costs are often passed down to end users through tariffs. Water unaffordability has become more prevalent in high- and low-income countries alike (Mack and Wrase, Reference Mack and Wrase2017; Vanhille et al., Reference Vanhille, Goedemé, Penne, Van Thielen and Storms2018). Water pricing, too, is socially differentiated from rural users sometimes paying more for trucked or kiosk water than urban users relying on pipes (Haeffner et al., Reference Haeffner, Galvin and Vázquez2017; Pihljak et al., Reference Pihljak, Rusca, Alda-Vidal and Schwartz2021). Moreover, a governmental solution such as raising prices to encourage water conservation may have the desired effect for affluent households but increase water insecurity for low-income households (Savelli et al., Reference Savelli, Rusca, Cloke and Di Baldassarre2021). Instead, a sociohydrological perspective has the potential to enable us to identify the areas that are missing or ignored and contribute to advancing current models of measuring water poverty. In the case of agricultural water, while processes of inclusion (or financial exclusion of smallholders based on power dynamics) have been implicitly explored by models (Elshafei et al., Reference Elshafei, Sivapalan, Tonts and Hipsey2014; Pande and Savenije, Reference Pande and Savenije2016), the mechanisms of how participation can transform social preferences into environmental decisions can be understood and incorporated more explicitly in sociohydrology.

Traditional societies may have had certain desirable attributes that lead to climate-resilient outcomes (Berkes et al., Reference Berkes, Folke, Gadgil, Perrings, Mäler, Folke, Holling and Jansson1995; Usher et al., Reference Usher, Jackson, Walker, Durkin, Smallwood, Robinson, Sampson, Adams, Porter and Marriott2021). These are often linked to culture and other informal institutions that govern water use in a sustainable manner. However, culture and institutions are slow moving variables that take generations to change and only influence individual decision-making and collective actions via norms (Pande et al., Reference Pande, Roobavannan, Kandasamy, Sivapalan, Hombing, Lyu and Rietveld2020).

Policies developed and implemented without adequate prediction of behaviour, or including different perspectives (across gender, or of upstream and downstream residents) in a participatory manner, can be ineffective (Baker et al., Reference Baker, Cullen, Debevec and Abebe2015; Chen et al., Reference Chen, Wang, Tian and Sivapalan2016; Weersink and Fulton, Reference Weersink and Fulton2020). Resilient systems are diverse with inbuilt redundancies, but positivist approaches in hydrological sciences have historically been focussed on efficiency which often has to be introduced at the expense of redundancies or diversity in the system (Morecroft et al., Reference Morecroft, Crick, Duffield and Macgregor2012). Sociohydrology, with its post-positivist approaches, can potentially play a significant role in informing such policies by understanding complex system dynamics (including attributes linked to resilience), offering alternatives to sustainable intensification or enhanced irrigation efficiencies (Scott et al., Reference Scott, Vicuña, Blanco-Gutiérrez, Meza and Varela-Ortega2014).

Martínez-Valderrama et al. (Reference Martínez-Valderrama, Olcina, Delacámara, Guirado and Maestre2023) recommend complex policy mixes aimed at reducing the water supply–demand gaps (also considering climate change) via actions at many levels (rural vs. urban; structural vs. nonstructural; local vs. global), highlighting the effectiveness of a "cocktail of solutions” approach over single solutions. Sociohydrology similarly offers multilevel and multiple-source approaches to conceptualise and assess agricultural water and climate policies, such as narrative approaches combined with document analysis and other sources of water system information to describe policy pathways at different governance levels and scales (Yu et al., Reference Yu, Haeffner, Jeong, Pande, Dame, Di Baldassarre, Garcia-Santos, Hermans, Muneepeerakul, Nardi, Sanderson, Tian, Wei, Wessels and Sivapalan2022). Furthermore, more robust and representative sociohydrological modelling incorporating resilience aspects can be used to simulate scenarios arising from different policies with added realism, evaluate potential unintended consequences and test alternative scenarios to iteratively inform policy development.

Conclusions and ways forward

Sociohydrology, as an interdisciplinary field, is placed well to understand and describe the interlinkages between agriculture, water and food security. This review examined how sociohydrology can explain the emergent dynamics of agricultural systems, account for climate change impacts due to/on agriculture and incorporate human behaviour into the description of agricultural water systems. Sociohydrological models have investigated the unintended consequences of water management interventions in agriculture (such as rebound or reservoir effects), highlighting the importance of accounting for such emergent phenomena to avoid unsustainable outcomes, degradation of natural resources or reinforcing existing inequalities. Sociohydrology also highlights why climate adaptation measures need to consider unintended consequences as well as individual behaviour (of individuals and communities), particularly for smallholders affected disproportionately by climate change impacts. Farmers can make apparently nonrational agricultural decisions due to their unique perspectives and experiences, bound by practical considerations. In this regard, sociohydrological models have considered social and psychological insights and have used novel data sources and diverse multi-method approaches to model human behaviour.

Sociohydrology has been used in understanding policy development for agriculture by investigating demand management measures, technology adoption and incentives for behaviour change to improve water-use efficiency and sustainable intensification. However, identified as a gap, this review proposed that sociohydrology can incorporate resilience thinking more explicitly to understand susceptibility of agricultural systems and agriculturalists to global change and uncertainties (in human–water systems) and possible unintended consequences resulting from agricultural policy. By accounting for technical approaches, earth sciences and the complexity of human behaviour, such interdisciplinary understanding can better inform agricultural policy, from conceptualisation (to initially identify and potentially avoid unintended consequences) to appropriate modifications (based on changes in the contexts of water, human behaviour or institutions).

This review thus positioned sociohydrology in how it has been incorporating social–scientific perspectives in agricultural water systems, identified key gaps in terms of desirable properties that are missing in current models to interpret climate resilience and consequently to develop policies which can enable more accessible, equitable, affordable and holistic solutions for improved resilience and successful adaptation.