Controls in this sense straddle the nature-culture divide, seeking to offer a more balanced analytical hinge between two literatures – STS and postcolonial studies – that do not always endeavor to take their dialogue into greater theoretical depths. Finally, all this can, and should, be applied reflexively. As I was writing these concluding pages, one day an acquaintance asked me what anthropology was about, and how was it different from other social sciences. I took the easy avenue and answered in terms of its “particular” method, ethnography. After I explained what we were supposed to do during fieldwork – basically, “hang out” with people –, he asked, “so you don’t do controlled experiments?” This question is a common one to hear from non-anthropologists, but for the first time it rang differently in my ears. As I hesitatingly replied, “no, we just observe”, I thought of Ian Hacking’s reflections on which I drew here, about knowledge in science being evinced through intervention as much as representation – or, better put, representation through intervention. Like in the experimental field sciences analyzed here, the ethnographer produces new knowledge not through mere contemplation, but by bringing into relation in an active,strawberry gutter system controlled manner the different kinds of knowledges and agencies that she encounters both in the field and in academia. This is an insight that has been around since at least The Gender of the Gift and its “controlled fictions” .

But these controls are also socio-technical in the sense put forth here: they have to do as much with epistemology as with power relations – relations which, in the case of a Brazilian scholar writing for a North-American audience, further intersect with the problematic of coloniality also discussed here. Finally, the “study up” character of this research effort meant that controls coming from my field interlocutors were often as much, if not more, forceful than those coming from academia. The account provided in this dissertation results therefore from a careful balance between these two kinds of demands that were, and will continue to be, made on myself and on this dissertation. In this sense, this entire research effort may be also a way of enhancing transactions between the two networks of relations that the anthropologist mediates in a privileged, sometimes even exclusive, manner: the “field” and the “desk” – indeed, of bringing closer together these two facets of the anthropological métier that have been kept separate for too long.However, growing populations and socio-economic development have required the expansion of irrigated agriculture and urbanization into areas with limited precipitation and inadequate surface water access, forcing a six-fold increase in global groundwater withdrawals over the last century . Over two billion people and more than 40% of the world’s agricultural production systems rely on groundwater as their primary water source, and it now accounts for one-third of the global freshwater supply . This development threatens groundwater resources and is apparent in high rates of aquifer depletion and degradation around the world . Decreasing reliability of surface water and depleted groundwater aquifers in groundwater-dependent regions severely impacts domestic water supplies , food security , and natural ecosystems .

This is especially the case in arid and semiarid regions where poorly monitored and often unregulated pumping has contributed to many negative impacts including lowered groundwater levels , loss of aquifer storage capacity , degraded water quality , seawater intrusion , land subsidence , streamflow depletion , and degradation of groundwater-dependent ecosystems . Managed aquifer recharge includes a suite of methods that are increasingly used to improve the quantity and quality of groundwater . MAR is the intentional diversion, transport, storage, infiltration, and recharge of excess surface water into aquifers during a wet period for subsequent recovery during dry periods or for environmental benefit . Over the last century, MAR projects have been implemented globally for different purposes ranging from flood mitigation , groundwater quality improvements , protection against seawater intrusion , enhancement of environmental flows , to stabilization of drinking water supplies , and other beneficial uses . In rural areas, MAR has been primarily used to increase groundwater security to meet irrigation demand, reduce the lowering of groundwater tables, and improve the quality of irrigation water . Varying water sources have been used for MAR, including river water , stormwater , treated wastewater , high-magnitude flows , and desalinated water . Dillon et al. provide a comprehensive summary of existing MAR methods such as river-bank filtration, infiltration basins, aquifer storage and recovery, soil aquifer treatment, and vadose zone infiltration devices . In recent years, managed aquifer recharge on agricultural land has gained popularity because it can effectively recharge aquifers using high flows from rainfall or snowmelt that could not previously be stored or used before ocean discharge .

Ag-MAR can utilize existing irrigation infrastructure to spread low-contaminant-load water over vast agricultural areas, allowing for large amounts of recharge in short time periods that would overwhelm more localized MAR systems . Because of its larger footprint, Ag-MAR has the potential to address or reverse many of consequences of groundwater overdraft . Identification of suitable groundwater recharge zones is often a first step in the implementation process of a MAR project , but requires the combination and prioritization of biophysical, socio-economic, and environmental criteria . Previous approaches have ranged from costly and time-consuming test drillings and stratigraphy analyses , statistical methods , on-site field investigations , and deterministic modeling . However, the most common approach used in the past 2 decades is an integrated remote sensing and geographic information system -based multi-criteria decision analysis . GIS-based MCDA techniques integrate various thematic layers and/or their individual features using a set of weights, a process for which no standard exists . MCDA approaches are increasingly supported or coupled with numerical modeling to provide more detailed and quantitative assessment of MAR opportunities and impacts . Among the existing studies, Russo et al. used a MODFLOW-2005 model of the Pajaro Valley in California to evaluate MAR project placement and operational parameters of potential infiltration basins. They concluded that combining a GIS-based MCDA analysis with a hydrogeologic model allowed assessing the relative benefits of different MAR scenarios and helped in assuring spatial data quality in the MCDA analysis. Zhang et al. conducted a GIS-based MCDA analysis to determine suitable MAR sites along the West Coast of South Africa. They used MODFLOW and particle tracking to trace the recharged water and concluded that the GIS analysis alone did not necessarily identify the most suitable sites. Likewise, Rahman et al. recommended that mathematical modeling should be combined with MCDA techniques to select optimal MAR locations even if a wide variety of thematic layers are considered in the MCDA. The criteria most frequently considered in site suitability analyses for MAR include intrinsic factors such as hydrogeology, topography, geomorphology, soil type, land use, and climate ,grow strawberry in containers as these represent main constraints on the groundwater recharge process . However, in view of the growing global population, urban sprawl, and water scarcity that affect both irrigation water supply and drinking water supply of groundwater-dependent rural communities, there is increasing recognition that integrated water resources management must consider socio-economic and socio-ecological criteria . The growing field of coupled natural human systems research in general, and socio-hydrology in particular , explicitly recognize the reciprocal interactions and co-evolution of coupled human-water systems. In California’s Central Valley , one of the most productive agricultural regions in the world, agricultural water use and climate-induced reduction of surface water supplies have led to severe groundwater overdraft in the past century. This has diminished access to clean, reliable drinking water in rural communities . This unsustainable and inequitable regional change has disproportionately impacted disadvantaged communities , who provide most of the farm labor to the agricultural sector , and perpetually decreased the economic viability and resilience of these communities to face hydro-climatic change .

With the implementation of new groundwater legislation in California , MAR and especially Ag-MAR could play a central role in optimizing the use of surface water to stabilize depleted groundwater aquifers while addressing critical issues of drinking water supply and quality in rural communities. Targeted Ag-MAR in the well capture zones of rural communities could provide multiple hydrological, socio-economic, and socio-ecological benefits by increasing equitable access to groundwater resources for rural or impoverished communities while supporting the needs of a groundwater-dependent agricultural economy. The goal of this study is to delineate locations for targeted groundwater recharge on agricultural land with the potential to improve the groundwater supply in rural communities. Our study proposes a GIS-based MCDA methodology that combines biophysical and socio-economic data with groundwater modeling and particle tracking to identify and prioritize suitable Ag-MAR locations for multi-benefit recharge. In contrast to previous studies which tend to first focus on-site identification and then on groundwater benefits of recharge locations, our Ag-MAR site selection is spatially constrained to benefit domestic wells located in rural communities. We developed a GIS-based MCDA methodology with the specific objectives: to identify agricultural land parcels suitable for Ag-MAR, to map well capture zones in rural communities using a groundwater model and particle tracking, to estimate the vulnerability of rural communities to changes in groundwater supply, and to prioritize Ag-MAR sites of most benefit to rural communities based on community vulnerability. The methodology framework was developed for California’s southern CV but could be useful to decision-makers and water resources managers in the efficient planning and management of Ag-MAR or MAR efforts worldwide.than agricultural wells . Many domestic wells are compromised due to lowering water tables and contamination with nitrate, metals, and metalloids from agricultural activities and solvents and other chemicals from industrial activities . The four counties spanning the southern CV have some of the lowest median household incomes in the state, $47,518–$53,869 per year compared to the state’s median of $75,277 , and some of the highest poverty rates . Due to the lower income levels generally found in the southern CV, most communities meet the California Proposition 84 definition of a DAC—a community with a median household income of less than 80% of the statewide average household income . In this study, we focus on 288 DACs, hereon called rural communities, located within the valley floor of the southern CV .Delineation of Ag-MAR areas that could improve drinking water supply in rural communities and estimation of community vulnerability to groundwater supply change is a multi-objective and multi-criteria problem that requires an understanding of the regional social-ecological-hydrological system . For this study, a total of 13 biophysical, hydrological, and socio-economic geospatial data were compiled . All data used in this study are from or representative of the 2012–2016 drought period. In addition to the biophysical data frequently used in MAR site selection , this study uses surface water conveyance infrastructure and groundwater flow fields to identify agricultural land parcels suitable for multi-benefit Ag-MAR and well source areas . To estimate the vulnerability of rural communities to change in groundwater supply, pesticide applications, land subsidence, and U.S. Census data are combined. Figure 3 displays the GIS-based MCDA framework. The Soil Agricultural Groundwater Banking Index was used to identify land parcels suitable for Ag-MAR based on soil suitability . SAGBI considers five factors to accommodate groundwater recharge while maintaining soil health, crop growth, recharge efficiency, or groundwater quality . SAGBI ranks soils on a six-class scale ranging from Excellent to Very Poor. For this study, only agricultural areas with soil ranked as Excellent, Good, or Moderately Good were selected as suitable AgMAR locations based on recommendations provided by Dahlke et al. .Land in the southern CV is mainly used for agriculture , with only minimal urban settlements and riparian land cover remaining . The major crops grown in the southern CV are deciduous fruit and nut tree crops , field crops , and vineyards . Because prolonged flooding can cause water logging and anoxic conditions in the root zone, which can promote plant diseases and other pests , implementing Ag-MAR on agricultural land planted with perennial crops can increase the risk of yield loss. However, little research exists on crop tolerance to prolonged flooding conditions created by Ag-MAR.