Research Project: Improved Agroecosystem Efficiency and Sustainability in a Changing Environment

Location: Sustainable Agricultural Water Systems Research

2025 Annual Report

Objectives
The availability of surface water and groundwater supplies for irrigated agriculture in California are adversely impacted by droughts, groundwater depletion and degradation, and increasing water demands. The overall aim of this project is to increase the efficiency and sustainability of irrigated agriculture, with a special focus on vine and tree orchard crops in the Central Valley, CA. This will be accomplished by: (i) improving irrigation efficiency; (ii) optimizing on-farm strategies for managed aquifer recharge (MAR); and (iii) assessing the long-term impacts on the sustainability of crop production, soil health, and groundwater quantity and quality. Specific objectives and subobjectives for this project are given below. Objective 1: Improve irrigation efficiency in agroecosystems by providing growers with accurate and timely estimates of spatial and temporal variations of crop ET in the field. Sub-objective 1A: Develop and validate an ET modeling framework designed to address the unique and highly structured canopy cover associated with California specialty crops. Sub-objective 1B: Improve irrigation efficiency for woody perennial crops in California by developing techniques to produce near-real-time estimates of ET from satellite data. Sub-objective 1C: Pair ET with quantified soil hydraulic and bio-meteorological properties at field-scale for better information on plant-available water within the soil column. Objective 2: Increase groundwater sustainability for irrigated agriculture by optimizing MAR strategies to capture excess surface water supplies that episodically occur during winter storms and store them into aquifers for later use. Sub-objective 2A: Quantify subsurface heterogeneity in soil properties to identify optimal locations and to better monitor and predict performance of MAR. Sub-objective 2B: Monitor and simulate the performance of MAR sites. Sub-objective 2C: Develop approaches to predict and minimize clogging and pathogen contamination at MAR sites. Sub-objective 2D: Develop and apply a computationally efficient watershed model to predict impacts of MAR on water quantity and quality. Objective 3: Assessment of the long-term impacts of irrigated agriculture on crop production, soil health, and groundwater quantity and quality under changing environmental conditions by monitoring atmospheric and subsurface fluxes and properties of soils and plants. Sub-objective 3A: Use economic analyses to assess impacts of MAR and improved irrigation efficiency on groundwater sustainability and the food-energy-water nexus.

Approach
Objective 1 will be accomplished using a combination of satellite remote sensing data, modeling, and field measurements of micrometeorological (e.g., Eddy covariance towers) and biophysical data during different phenological stages to estimate spatial and temporal variations in evapotranspiration (ET), crop stress, and irrigation requirements in vine and tree orchards crops at multiple sites in the Central Valley of California. The collected soil and bio-meteorological data will be used to validate and refine model estimates of ET from satellite imagery. Improved algorithms will be developed for near real-time ET estimates and their spatial variability in the field that can be used to improve irrigation efficiency. Objective 2 will be addressed by developing, implementing, and overcoming challenges associated with Managed Aquifer Recharge (MAR) technologies. MAR techniques that will be studied include Ag or flood MAR and drywells. Geophysical methods will be employed to identify optimal locations for MAR, to better characterize subsurface heterogeneity, and to monitoring infiltration and recharge behavior. Field sites will be characterized for soil hydraulic properties, and equipped to monitor water inputs, infiltration, recharge, and soil and water quality parameters. Complementary laboratory studies and pore-network modeling studies will be conducted to better infer underlying mechanisms controlling MAR performance, including clogging and pathogen transport and fate parameters. Collected data streams will be used in conjunction with mathematical modeling to inversely determine parameters, design improved MAR strategies that optimize water quantity and quality, and to predict long-term performance of MAR on the sustainability of groundwater and irrigated agriculture. Calibrated models will in turn be used to develop meaningful predictions of risk, management, and future performance at particular sites. A computationally efficient watershed scale model will be developed to rigorously simulate exchange of water and contaminants between surface water, the vadose zone, and groundwater. Numerical experiments will be conducted to test specific hypotheses and generalize results to other sites, water management practices, climatic conditions, and watersheds. Objective 3 involves the extension of Objectives 1 and 2 to include economic analyses to study long-term implications of remote-sensing based irrigation management tools and MAR strategies on the food-energy-water nexus, groundwater sustainability, and the long-term viability of irrigated agriculture. It also includes economic analyses of various land management practices (e.g., land fallowing), policies (e.g., the sustainable groundwater management act), and impacts on endangered species.

Progress Report
This report documents FY 2025 progress for project 2032-13220-002-000D, “Improved Agroecosystem Efficiency and Sustainability in a Changing Environment”, which started March 2022. Significant progress has been made toward Objective 1 to improve irrigation efficiency by enhancing the accuracy and operational utility of satellite-derived evapotranspiration (ET) estimates. Under Sub-objective 1A, ARS researchers have made methodological advancements through the application and testing of an ARS-led ET modeling framework. Specifically, a synthetic approach to redistribute coarse-resolution ET outputs using high-resolution LAI (Leaf Area Index) inputs was utilized to address issues surrounding thermally heterogeneous landscapes, reaffirming the value of Landsat-scale inputs in ET modeling and shedding light on the effects of fallowing land near irrigated landscapes. ARS also advanced the underlying physical understanding of radiation transfer in vineyards, which underpins both ET models and crop process simulations. A one-dimensional radiation absorption modeling study evaluated the limitations of commonly used clumping factor approaches and introduced a novel geometric binomial model that better captures the complexity of radiation interactions in heterogeneous canopies such as vineyards. These findings provide important quantitative guidance for integrating improved radiation transfer calculations into land surface and crop models, ultimately supporting more accurate predictions of ET and related biophysical processes. Also, in advancement of Sub-objective 1A, multiple efforts this year focused on improving the Two-Source Energy Balance (TSEB) modeling framework for better partitioning of soil and canopy heat fluxes. ARS researchers tested a novel field-scale method to directly derive soil and canopy temperatures from satellite land surface temperature (LST), removing the need for iterative solutions. This new approach showed comparable performance to traditional TSEB implementations, suggesting it as a viable operational alternative under suitable canopy conditions. This approach also provides scalability, allowing operational applications in other regions where traditional inputs to the model are unavailable. In support of Sub-objective 1B, ARS researchers advanced the near-real-time applications of OpenET, which now delivers actual ET data at 30-meter spatial resolution and daily temporal resolution with a two-day latency. This effort has been extended through the OpenET Farm and Ranch Management Support application, a user-friendly interface that simplifies access to ET data and allows for greater customization to individual growers’ needs, streamlining agricultural planning. Progress has on Sub-objective 1C in identifying the complexities of Vineyard Irrigation Data Assimilation(VIDA) and ET modeling approaches by expanding VIDA’s capabilities to almond orchards. VIDA, originally developed for vineyard applications, continues to be used operationally to determine springtime soil moisture drawdown and trigger irrigation in wine grape vineyards in collaboration with Gallo Winery. Research continued on Objective 2, which focuses on evaluating managed aquifer recharge (MAR) approaches in California’s Central Valley. In support of Sub-objective 2A, ARS researchers collected and analyzed towed time domain electromagnetic (tTEM) data from a field site in Arbuckle, California, and along an irrigation canal in Yolo County, California. This information is being used by a grower and flood control agency to identify the best locations for MAR. ARS researchers are currently developing and implementing MAR plans for this purpose. For example, ARS has installed and are testing fiber optic distributed temperature sensing in the bottom portion of the canal to monitor the spatial and temporal variability in infiltration, and to find ways to enhance recharge. For Sub-objective 2B, ARS researchers in Davis, California, continued to monitor the evolution of water flow and water quality at a MAR site located about 20 miles southwest of Fresno, California, that employed five drywells. Time-lapse electrical resistivity tomography (TL-ERT) and point sensors (measurement of water content and salinity at depths of about 25, 55, 98, and 177 feet) have been used to monitor water infiltration, redistribution, drainage, and recharge to the water table at about 220 to 250 feet below land surface. The TL-ERT imaging and point sensors revealed that the wetting front reached the water table after about seven months from the start of MAR, but that the vadose zone only slowly drained over that past two years. The majority of infiltrated water was stored in the vadose zone. Approaches are being developed to better quantify water flow behavior using TL-ERT imaging and to separate the coupled influence from subsurface heterogeneity and changes in water content and salinity. One paper was published , and others are currently in development. Measurements of salts, nitrate, pesticides, and geogenic metals were also periodically made during that past two years in the vadose zone (suction lysimeters at depths of 25, 55, 98, and 177 feet) and at the water table (bailer). Salt and nitrate were observed to be flushed, but high levels of pesticides were not found. MAR can alter the solution geochemistry and redox conditions of the vadose zone and groundwater. Results show that in situ metals such as arsenic, uranium, and chromium can be mobilized with MAR in dissolved forms and/or associated with colloids (e.g., iron oxide). Particles and pathogens in source water and soil solutions can adversely impact MAR performance by clogging infiltrating surfaces or causing waterborne disease outbreaks, respectively. Under Sub-objective 2C, research was initiated to better understand, predict, and mitigate clogging and pathogens during MAR. Two papers were completed and published on clogging. One was based on pore-network modeling results, and a second on column studies examining the impact of particle concentration on clogging. A continuum scale model is currently under development to account for these clogging observations. Another continuum scale model is being extended to predict the transport and fate of viruses in groundwater. This model builds on work that accounts for physicochemical conditions on colloid transport, attachment, detachment, and blocking that was published this year. This model includes the roles of nanoscale roughness, charge heterogeneity, microscopic roughness, and hydrodynamics on colloid removal and transport. Significant advancements were made under Sub-objective 2D, which focuses on developing computationally efficient watershed models to assess the impacts of MAR on water quantity and quality. ARS researchers, in collaboration with the University of California, Riverside, developed several versions of integrated modeling systems that simulates overland flow, vadose zone processes, and groundwater flow. These models combine KINEROS2, HYDRUS-1D, and MODFLOW-2005 in a modular and computationally efficient framework. Multiple versions have been tested-at hillslope and watershed scales-for water flow, solute transport, and reactive transport, showing strong accuracy in mass conservation and improved computational performance. A detailed integrated groundwater model was also developed for the Modesto-Turlock-Merced sub-basins, coupling HYDRUS-1D and MODFLOW-2006. The model uses Hydrogeological Recharge Units (HGRUs) to represent zones with similar recharge characteristics. It provides spatially distributed estimates of recharge, groundwater dynamics, and residence/transit times, offering valuable insight for MAR planning and groundwater sustainability. To address poorly quantified recharge processes in the Sierra Nevada foothills, a conceptual box model and the MIKE SHE integrated hydrologic model were developed. These tools simulate diffuse and focused recharge, surface–subsurface interactions, and water movement through complex topography, helping to identify the role of hillslope processes in regional recharge. Furthermore, ARS researchers also made notable progress in applying machine learning and deep learning techniques to understand and predict soil moisture dynamics. A Vision Transformer (ViT)-based model was developed to forecast sub-daily soil moisture across the continental United States using available data. This work enhances our ability to predict runoff, drought stress, and biogeochemical processes, supporting better conservation and land management practices. ARS in collaboration with the University of California, Davis, continue to develop economic optimization models, water availability analyses, and data analyses for reduced water adaptations, including MAR and irrigation systems, accounting for potential variability in water availability per Sub-objective 3A. They have developed a novel remote sensed ET methodology to estimate consumptive water use using Google Earth Engine ET data to integrate economic and hydrologic groundwater models in the Central Valley of California. They have continued to build a large-scale model of the Central Valley, focusing on regional crop and water availability, and water use variations. Models developed can evaluate adaptations to changing water availability, due to overuse, policy, and/or weather changes on croplands in California. Researchers continue to examine MAR, including an assessment of water availability in the Central Valley, and an examination of the costs and benefits of different MAR types (e.g., flood-MAR, recharge basins, drywells, and aquifer storage and recovery). ARS researchers also developed econometric analysis methodology to assess survey data regarding Sustainable Groundwater Management Act implementation in the Central Valley.

Accomplishments
1. Development of a novel, computationally efficient watershed modeling framework. Sustainable watershed resources management increasingly requires computationally efficient and scientifically accurate tools that can capture the complex interactions between surface water, soil, plant, and groundwater processes. ARS scientists in Davis, California, developed a flexible and computationally efficient integrated hydrologic modeling framework to simulate the full continuum of hydrologic and water quality processes from the surface to groundwater. The system couples three widely used state-of-the-art process-based models using sequential coupling, dynamic time stepping, and dimensionality reduction to make the model run faster and use fewer computing resources, addressing a common limitation of many process-based models. The model accurately captures the spatial and temporal variability of hydrologic and hydrochemical processes at both hillslope and watershed scales. Benchmark tests demonstrated the model’s ability to accurately predict key responses, such as runoff, soil water content, groundwater dynamics, solute concentration, and soil erosion, under variable rainfall conditions. This innovative framework provides a robust, scalable tool for integrated watershed analysis, supporting sustainable water management, conservation planning, and ecosystem health assessments.

2. Integrated hydrologic modeling of groundwater flow dynamics and recharge in the San Joaquin Valley, California. California’s Central Valley faces critical groundwater sustainability challenges due to decades of overdraft, recurrent drought, and increasing water demand. These pressures have led to declining groundwater levels, land subsidence, and reduced water reliability. To address these challenges, ARS scientists in Davis, California, developed a high-resolution, integrated hydrologic model to improve understanding of groundwater recharge processes, flow dynamics, groundwater depletion, and sustainable management strategies. The model captures how water moves through the landscapes, soils and into groundwater, accounting for differences in land use and surface conditions across the Valley. The results showed recharge rates ranged from less than 1 mm to over 1460 mm per year. Once water infiltrates, it can take more than a decade to reach the groundwater table with groundwater residence times ranged from 2 to 5000 years. The model provides science-based, spatially explicit insights to help water managers optimize managed aquifer recharge (MAR) strategies, prioritize high-recharge areas, and design targeted interventions to mitigate overdraft and support long-term groundwater sustainability in California’s Central Valley.

3. Synthetic evapotranspiration model to address erroneous evapotranspiration estimation in heterogeneous landscapes. Irrigated agricultural lands neighboring barren landscapes experience advective conditions where hot dry air enters the field, causing model discrepancies with measured values. ARS scientists in Davis, California, tested a modified satellite-based modeling framework, called ALEXI, for estimating evapotranspiration (ET) in a spatially complex, drip-irrigated almond orchard in California. A synthetic satellite product was developed by disaggregating 4-km ET to finer resolutions (2, 1, 0.5 km) using leaf area index data. A disaggregation algorithm (DisALEXI) was then applied using these synthetic inputs and evaluated against ground measurements. The 1-km version produced the best results, especially in areas where coarse satellite pixels blended barren and irrigated land. These results show that spatially refined ALEXI can improve accuracy in regions where native model resolution fails to represent crop-level variability. Additional efforts to address landscape heterogeneity include applying a remote sensing ET modeling framework to delineate irrigation management zones in an almond orchard in California. Results demonstrate that ET and crop water stress indices derived from high-resolution thermal imagery agree well with in situ measurements and can capture spatiotemporal variability in orchard water use and stress.

4. One-dimensional modeling of radiation absorption in vineyards. Specialty crops in California, especially vineyards, have a unique canopy structure that is difficult to model using physical constraints. ARS scientists in Davis, California, evaluated the performance of simplified radiation models in heterogeneous vineyard canopies using field measurements and a high-fidelity 3D radiative transfer model, Helios. Existing models using generic inputs were found to over- or under-predict canopy radiation absorption, especially under varying row orientations and canopy structures. In response, we developed a simple model that accounts for the various unique characteristics of a vine canopy. This model significantly reduced errors across simulated vineyard conditions and showed strong agreement with field observations. The binomial model is suitable for integration into land surface and crop models, offering improved representation of radiation transfer and energy balance in vineyard systems.

5. Field-scale separation of canopy and soil temperatures for Two-Source Energy Balance (TSEB) modeling. Satellite approaches to modeling evapotranspiration (ET) inherently use pixels that contain both canopy and soil components, leading to a mixed signal. ARS scientists in Davis, California, tested a simplified method to partition satellite-derived land surface temperature (LST) data into soil and canopy components at the field scale for application of the TSEB model to improve the estimation evapotranspiration. By assuming spatial homogeneity within vineyard and orchard fields, we simplified the process of deriving soil and canopy temperatures. Comparisons of the original TSEB model and ground measurements showed good agreement, with only small differences in modeled evapotranspiration estimates. The largest discrepancies occurred under low canopy cover, where small differences in soil temperature led to amplified canopy temperature errors. Despite these limitations, the method provides a more direct, computationally efficient approach for ET modeling under appropriate field conditions.

6. Monitored long-term drywell recharge into the deep vadose using geoelectric tomography. Although drywells hold a great deal of promise as a managed aquifer recharge (MAR) technology, there is little information about their long-term efficacy in delivering water through the unsaturated zone to the aquifer. During the historically wet 2023 water year, ARS scientists from Davis, California, monitored the performance of five drywells outside Fresno, California, using geoelectrical tomography. During injection, the MAR water was imaged moving rapidly down through the unsaturated zone towards the underlying aquifer; however, post-injection, the water slowed considerably. Even two years later, much of the MAR water remains suspended in the unsaturated zone, with some water perched on an impermeable clay-rich layer. Additionally, fresher MAR water was observed displacing the more saline water in the unsaturated zone, forcing it down towards the aquifer. For researchers, engineers and water managers, the results illustrate the efficacy and limitations of drywell MAR and geoelectrical surveys for monitoring these operations. Furthermore, they challenge preconceptions of how MAR operations work in practice, which will have important implications for policymakers assessing how to incentivize MAR for landowners and communities.

7. Developed an integrated economic hydrologic model to spatially assess agricultural landscapes. Variation across agricultural landscapes influences how alterations in management practices or cropping systems impact regional targets for water use or water quality outputs. To optimize land use changes to meet different scenarios, ARS scientists in Davis, California, developed a flexible geo-spatial economic framework by combining Soil and Water Assessment Tool model outputs and an agro-economic programming model. This framework accounts for profits, water use, fertilizer needs, and nitrate pollutants from different crops and alternative crop management regimes. It can be used on a regional scale to pinpoint locations that would be the most effective to be included in various programs to meet management goals.

Review Publications
Levers, L.R., Dalzell, B.J., Peterson, J.M. 2025. Optimizing land management for nitrogen reduction: A bio-economic spatial model. Journal of Environmental Management. 377. Article 124702. https://doi.org/10.1016/j.jenvman.2025.124702.
Pradhananga, A., Choi, A., Levers, L.R. 2025. Resident intentions to support organizations that aim to prevent the spread of aquatic invasive species: The role of value orientations, efficacy, and risk perception. Human Dimensions of Wildlife. https://doi.org/10.1080/10871209.2024.2446785.
Hammond, C.B., Kareem, M., Bradford, S.A., Che, D., Sharma, S., Wu, L. 2024. Predicting a wide range of fractal dimensions of salt-induced aggregates in water using a random forest model. Langmuir. 40(45):23606-23615. https://doi.org/10.1021/acs.langmuir.4c01182.
Bradford, S.A., Lin, D. 2024. A theoretical model to predict the influence of physicochemical conditions on colloid transport, attachment, detachment, and blocking in porous media. Journal of Hydrology. 650. Article 132483. https://doi.org/10.1016/j.jhydrol.2024.132483.
Lambert, L., Yao, Y., Levers, L.R. 2025. Optimal cropping patterns and intertemporal groundwater usage under extraction constraints in Oklahoma’s Panhandle. Agricultural Water Management. 313. Article 109472. https://doi.org/10.1016/j.agwat.2025.109472.
Sapkota, A., Roby, M.C., Peddinti, S.R., Fulton, A., Kisekka, I. 2025. Comparative analysis of evapotranspiration (ET), crop water stress index (CWSI), and normalized difference vegetation index (NDVI) to delineate site-specific irrigation management zones in almond orchards. Scientia Horticulturae. 339. Article 113860. https://doi.org/10.1016/j.scienta.2024.113860.
Osterman, G.K., Knight, R. 2025. Measurement of vadose zone water content with direct-push nuclear magnetic resonance logging. Environmental Research Communications. 7(1). Article 011005. https://doi.org/10.1088/2515-7620/ada733.
Arboleda-Zapata, M., Osterman, G.K., Li, X., Sasidharan, S., Dahlke, H.E., Bradford, S.A. 2025. Time-lapse ensemble-based electrical resistivity tomography to monitor water flow from managed aquifer recharge operations. Journal of Hydrology. 659. Article 133282. https://doi.org/10.1016/j.jhydrol.2025.133282.
Ekamparam, A.S., Bradford, S.A., Sasidharan, S. 2025. Pretreatment approaches to minimize clogging during managed aquifer recharge with drywells. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 726(2). Article 137891. https://doi.org/10.1016/j.colsurfa.2025.137891.
Cammalieri, C., Anderson, M.C., Bambach, N., Mcelrone, A.J., Knipper, K.R., Roby, M.C., Kustas, W.P. 2024. Field scale partitioning of Landsat land surface temperature into soil and canopy components for evapotranspiration assessment using a two-source energy balance model. Irrigation Science. https://doi.org/10.1007/s00271-024-00976-w.
Cammalleri, C., Anderson, M.C., Bambach, N., Mcelrone, A.J., Knipper, K.R., Roby, M.C., Ciraolo, G., Decaro, D., Ippolito, M., Corbari, C., Ceppi, A., Mancini, M., Kustas, W.P. 2024. A fully remote sensing-based implementation of the two-source energy balance model: an application over Mediterranean crops. Agricultural Water Management. https://doi.org/10.1016/j.agwat.2024.109207.
Knipper, K.R., Anderson, M.C., Bambach, N., Melton, F., Ellis, Z., Yang, Y., Volk, J., McElrone, A.J., Kustas, W.P., Roby, M.C., Carrara, W., Castro, S., Kilic, A., Fisher, J., Ruhoff, A., Senay, G.B., Morton, C., Saa, S., Allen, R. 2024. A comparative analysis of OpenET for evaluating evapotranspiration in California almond orchards. Agriculture and Forest Meteorology. 355. Article 110146. https://doi.org/10.1016/j.agrformet.2024.110146.
Casillas-Trasvina, A., Meles, M.B., Cui, W., Hatch, T., Bradford, S.A., Harter, T. 2025. Integrated hydrologic modeling of groundwater flow dynamics and recharge in the San Joaquin Valley. Journal of Hydrology. 660(A). Article 133377. https://doi.org/10.1016/j.jhydrol.2025.133377.
Rahman, A., Simunek, J., Bradford, S.A., Ajami, H., Meles, M.B., Chen, L., Szymkiewicz, A., Pawlowicz, M., Acero Triana, J.S., Casillas-Trasvina, A., Beegum, S. 2025. A new externally coupled, physically-based multi-model framework for simulating subsurface and overland flow hydrological processes on hillslopes. Journal of Hydrology. 662(A). Article 133842. https://doi.org/10.1016/j.jhydrol.2025.133842.
Lin, D., Tang, M., Zhang, B., Zhang, X., Bradford, S.A., Hu, L. 2025. Coupled clogging and colloid retention mechanisms in porous media: Insights from Pore-Network modeling. Separation and Purification Technology. 363(1). Article 132055. https://doi.org/10.1016/j.seppur.2025.132055.
Ponce de Leon, M.A., Alfieri, J.G., Prueger, J.H., Hipps, L., Kustas, W.P., Agam, N., Bambach, N., McElrone, A.J., Knipper, K.R., Roby, M.C., Bailey, B. 2025. One-dimensional modeling of radiation absorption by vine canopies: Evaluation of existing model assumptions, and development of an improved generalized model. Agricultural and Forest Meteorology. 373. Article 110706. https://doi.org/10.1016/j.agrformet.2025.110706.
Osterman, G.K., Knight, R. 2025. Measurement of vadose zone water content with direct-push nuclear magnetic resonance logging. Environmental Research Communications. 7(1). Article 011005. https://doi.org/10.1088/2515-7620/ada733.
Tigabu, T.B., Muller, E., Meles, M.B., Dahlke, H.E., Schuler, G., Fohrer, N., Wagner, P.D. 2025. Effects of forest harvesting operations on hydrology: Experiences from the Palatinate Forest Biosphere Reserve. Hydrological Processes. 39(4). Article e70115. https://doi.org/10.1002/hyp.70115.
Knipper, K.R., Bambach, N., Anderson, M.C., Yang, Y., Kustas, W.P., McElrone, A.J., Nocco, M., Torres-Rua, A., Gao, F.N., Hain, C., Castro, S., Crompton, O.V., Saa, S. 2024. Using ALEXI-DisALEXI for estimation of satellite-derived water use in a California almond orchard under spatially heterogeneous conditions. Acta horticulturae. 1409:143-150. https://doi.org/10.17660/ActaHortic.2024.1409.20.
Sasidharan, S., Bradford, S.A. 2025. Assessing drywell designs for managed aquifer recharge via canals and repurposed wells. Scientific Reports. 15. Article 1829. https://doi.org/10.1038/s41598-024-84865-4.