In addition, the number of households that rely on a single water supply source is only 22 on average, and more communities have access to surface water as drinking water supply source.About half of the final Ag-MAR parcels are associated with extremely and very highly vulnerable communities; most are located within Tulare County . A total of 1,334 parcels are associated with communities with high and moderate vulnerability, while 150 Ag-MAR parcel are associated with low vulnerability communities. Among the Ag-MAR parcels associated with extremely, very highly, and highly vulnerable communities, about 68% of the parcels have excellent soil suitability and are planted with vineyards . Likewise, Ag-MAR parcels associated with moderate and low vulnerability communities have predominantly excellent soil suitability and the majority of the parcels are either vineyards, planted with field crops, or idle . There are fewer Ag-MAR parcels surrounding rural communities in the southern part of the study area where large urban centers provide less opportunity to implement Ag-MAR as a mitigation strategy. Figure 7a also indicates that there are some communities, particularly in the western and southwestern part of the study area, where no suitable Ag-MAR parcels could be found within the well capture zones. In fact, suitable Ag-MAR parcels could only be identified for about half of the 288 communities,vertical hydroponic nft system leaving 139 communities without nearby Ag-MAR sites. For these communities, Ag-MAR potential was mainly diminished by soil suitability and lack of surface water conveyance infrastructure . Of the 149 communities for which suitable Ag-MAR parcels could be identified, 88 communities had at least 10 associated Ag-MAR parcels, and 60 communities had at least 20 associated parcels.
There were 61 communities with less than 10 associated parcels, 14 of which had only one associated parcel.Although a wide variety of decision support tools are available for general surface and groundwater management and drinking water quality in California ; none of these tools provide information on mitigation or remediation options for chronic groundwater overdraft or contamination. This study is the first effort to systematically explore the potential for targeted Ag-MAR to directly improve the drinking water supply from groundwater in rural communities. In past decades, MAR has been used to achieve varying objectives , however, implementation of MAR is often limited by challenges of recharge water availability , locating suitable groundwater recharge zones, regulatory constraints, and funding obstacles . Ag-MAR overcomes many of these challenges due to low capital cost and permitting requirements , and with appropriate planning can be used to provide multiple benefits to a region including stabilized domestic and agricultural water supply, flood control, and climate change mitigation . However, Ag-MAR implementation in the southern CV might be constrained by the existing surface water conveyance capacity, which Hanak et al. deemed inadequate for capturing and moving high flows to suitable recharge locations. Conveyance capacity data were not available for this analysis, but according to Hanak et al. represents one of the major limitations for MAR implementation. In this study, almost 3,000 land parcels suitable for Ag-MAR ranging in size from 0.2 to 260 ha have been located within the well capture zones of rural communities. Of the 288 rural communities included in this analysis, 253 communities rely on groundwater as their main source of drinking water. However, suitable Ag-MAR parcels could only be identified within the capture zones of 149 of the 288 communities, 144 of which are reliant on groundwater for their drinking water supply. Most of the communities for which no nearby AgMAR parcels could be identified are located near large urban areas or near the CV rim, where topography and a lack of conveyance infrastructure prohibit Ag-MAR. A complex political and socio-economic environment around water governance in the region has historically prevented more inclusive water management but for these communities, other types of MAR , well head treatment, or incorporation into nearby public water supply systems might be the only options to improve the quantity and quality of drinking water supplies.
For reference, 118 of the 288 communities studied have no access to public water supply sources but 56% of these communities are within the boundaries of existing public water supply systems.MAR site selection studies using GIS-based MCDA approaches have been developed in many regions across the world . The majority of these studies use slope, land use, geology and soil type as the main criteria for identifying MAR sites . Similarly, our study uses soil characteristics and land use as the main criteria to determine Ag-MAR site suitability, but differs from earlier studies in that we refine suitable sites by linking the GIS analysis with deterministic groundwater modeling and particle tracking to only select sites with potential to benefit the drinking water supply in rural communities. The integration of groundwater modeling and particle tracking also ensured the inclusion of climate and hydrogeological data in the analysis. However, the groundwater modeling also introduced uncertainty in the estimated well capture zones, due to the spatio-temporal resolution of the model and because a quasisteady-state groundwater flow field was used for the particle tracking. The generalized groundwater flow field likely does not capture local spatio-temporal dynamics in the flow field caused by seasonal pumping, which can change or reverse some of the flow directions depicted in Figure 5. These seasonal dynamics should be considered in the final selection of Ag-MAR locations using field-level studies. In addition, in groundwater-dependent regions where an integrated surface water-groundwater model is not available, well capture zones may need to be derived from field observations. The Ag-MAR locations identified in this study relied on the integration of regionally specific data for the southern CV, but the methodology can be applied to other groundwater-dependent regions. To implement the Ag-MAR site suitability analysis, regional soil or geomorphology data can be used instead of SAGBI, and land use and surface water hydrology can be inferred from air photographs and satellite images.
Similarly, data descriptive of the socio-economic status of rural communities in groundwater-dependent regions or adverse environmental effects of human activities and groundwater overdraft on rural populations can be substituted with locally available demographic data or remote sensing data , respectively. In regions where little geologic or physiographic data exists, nft hydroponic system growing availability of high-resolution remote sensing data of land surface and subsurface characteristics may be useful . Many previous MAR site suitability studies were conducted to inform sustainable groundwater management , to serve as guidelines and decision support for farmers and policy makers , or to raise general interest for MAR development . However, as showcased in this study, GIS-based MCDA can also be used to identify priority areas for intervention or disaster management if site suitability analysis is combined with vulnerability analysis . This combination can be particularly useful in water resources management because the outputs can provide easily interpretable visual information, help refine the spatial focus of the problem, support priority development, and allow for assessment of different management scenarios before field-level investigations begin.To date, few MAR site suitability studies have conducted a sensitivity analysis or validation of recommended sites . Previous MAR suitability assessment studies have used indirect methods to validate MAR locations , while few have used numerical models and in situ observations . With this study, we propose to guide selection of suitable MAR sites by ensuring quantifiable benefits to groundwater levels, storage, water quality, and land subsidence. Although water management agencies maintain multiple MAR basins in the southern CV, most of these facilities have not been implemented to benefit the domestic water supply to rural communities. The Tulare Irrigation District has a 42 ha MAR basin located south of the Okieville community that has been operational since the 1940 . The recharge basin overlays the capture zone of the community’s southern groundwater wells. Its location was accurately identified by this study as suitable Ag-MAR location . Data from Okieville domestic wells show groundwater quality improvements from MAR, including lower nitrate, uranium and arsenic concentrations, which are well below the groundwater concentrations of nearby communities . These indicate that our methodology has positively identified locations where recharge can improve the drinking water supply of rural communities in a region of our study area. Although many studies have used GIS-based MCDA for MAR suitability studies, there is no consensus on appropriate criteria, weights, and methods as these are generally dependent on the study objective, data availability, and local experience .
The assignment of weights to each thematic layer or feature is one of the most subjective factors of MCDA and thus, one of the main sources of uncertainty . To address this issue, AHP is increasingly used to convert subjective assessments of relative importance into a set of weights , though sometimes the relative importance of themes may not be discernable . In this study, local experts in hydrology and human ecology similarly recommended the use of equal weights for thematic layers in both the site suitability and community vulnerability analyses. However, future iterations of these analyses will require the active involvement of local stakeholders , a process that may benefit greatly from the integration of AHP into the GIS-based MCDA . One main difficulty when estimating suitable recharge areas is the spatial and temporal variability of the physical system. We acknowledge that our analysis mainly uses land surface characteristics to determine suitable Ag-MAR sites, while subsurface characteristics were not directly included. Other factors not accounted for in our analysis include water availability, water quality, unsaturated zone transport, and willingness of landowners to flood agricultural land. Although robust quality control measures were taken, the accuracy of our results relies on the integrity of input data. Issues of accuracy and completeness of proprietary, hand-digitized, or self-reported data are inevitable, hence field-level studies of local surface and subsurface characteristics should be completed as part of project scoping and pilot testing. They are also essential to assess soil surface conditions, the presence of potential unprotected wellheads, capacity of connected surface water conveyance systems, feasible Ag-MAR water application amounts , and cropping and agro-chemical application history to determine potential legacy contaminant loading in the unsaturated zone that could be mobilized by recharge . Although nitrate loading to groundwater has been assessed at larger scales in California’s CV , parcel-level data on fertilizer application rates and nitrogen removal by crops is not publicly available, preventing the assessment of legacy nitrate loading in the unsaturated zone. Future improvements of this methodology should include the addition of contaminant transport modeling or site-specific simulation of drinking water contaminants to address this gap. Climate projections and impacts on surface water availability for recharge require further investigation . As shown by Bachand et al. , despite its semiarid climate, the southern CV faces frequent flood risks. Along the Kings River, flows have exceeded the flood stage almost once every 7 years in the last 4 decades, creating total losses exceeding $1.2 billion . Kocis & Dahlke showed that excess surface water from high flows occur on average every 4.7 out of 10 years with total amounts reaching up to 1.6 km3 between November and April in years when high flows are available. Water scarcity is expected to increase as the southern CV experiences more frequent and longer droughts and more frequent extreme events during wet years . Integrated water management solutions like Ag-MAR are urgently needed to stabilize groundwater supplies in the region.As the world strives to reduce greenhouse gas emissions, natural forest regrowth and active tree planting are frequently proposed as mitigation pathways to sequester carbon through increases in above- and below-ground biomass and soil organic matter . This reforestation process requires land. Although the portion of the Earth’s land surface used for agriculture continues to expand , a growing literature uses remotely sensed biophysical and land cover data to map the distribution of so-called “abandoned agricultural land” in high, middle, and low income countries. Goga et al. and Yin et al. estimate a great deal of abandoned agricultural land, whereas Crawford et al. and Potapov et al. indicate far less and even diminishing amounts. This estimation discrepancy may be due to differences in land cover measurement, but we argue that it also likely arises from the omission of information about landholding status and the complex landholder decision-making process.