Duration: 10/01/2025 to 09/30/2030

Administrative Advisor(s):

NIFA Reps:

Non-Technical Summary

The Issue & Its Importance: Managing water resources effectively is crucial for preventing flooding, pollution, and water shortages. Despite decades of research, current watershed models struggle to accurately predict water quality and hydrologic changes, particularly under extreme weather conditions. Traditional models often lack real-time data integration and fail to capture the complex interactions of land use, climate, and water flow. This limits their ability to guide conservation efforts like placing Best Management Practices (BMPs) in the most effective locations. More advanced models and data collection techniques, such as sensor technology are needed to support better decision-making and resource management.

Our Goal & Objectives: This project aims to develop next-generation watershed models that integrate Artificial Intelligence (AI), real-time monitoring, and stakeholder input to improve water quality, flood prevention, and conservation planning. Our key objectives include:

1. Developing AI-enhanced watershed models for improved accuracy.
2. Using real-time sensor data to enhance predictions.
3. Optimizing conservation strategies (BMPs) based on scientific and economic factors.
4. Engaging stakeholders through participatory model development and citizen science.

Target Audiences & Benefits: Farmers, policymakers, scientists, and communities will benefit from better data, improved water management strategies, and enhanced environmental resilience.

How Our Activities Lead to Outcomes: By combining AI, real-time monitoring, and stakeholder participation, we create smarter, more adaptable models that help decision-makers act before problems escalate. This ensures sustainable water management for future generations.

Statement of Issues and Justification

Despite significant progress in understanding hydrology, geomorphology, and biogeochemistry processes over the past 40 years, gaps remain in predicting watershed outcomes and meeting diverse water quality and management goals under dynamic weather conditions. Current models, while improved, still fall short in accurate and scalable representations of complex biophysical processes, especially when considering the temporal and spatial variability of complex dry and wet cycles within watersheds. These shortfalls currently constrain efforts to inform the design and optimal placement of Best Management Practices (BMPs) in agricultural and urban watersheds to meet goals related to water quantity, water quality, and climate resilience at the watershed scale. Additionally, the incorporation of real-time, high-frequency monitoring data into models remains limited, with spatial representativeness often uncertain and compromised.

There is a growing need for next-generation watershed models that integrate hybrid approaches that combine artificial intelligence (AI) and machine learning (ML) approaches, statistical methods, process-based understanding of physical processes, and stakeholder-driven datasets to better inform decision-making, adoption, and implementation of various management practices at different scales. These models must support the interactive effects of hydrology, geomorphology, and biogeochemistry, address “critical source areas,” and improve the scalability and reliability of water quality outcomes. At the same time, limited financial resources demand that BMPs be strategically placed, considering both biophysical and socio-economic factors to maximize their effectiveness and adoption. Through this collaborative multi-state research project our team will address these issues and focus on the following specific objectives

1. Develop Next-Generation Hybrid Watershed Models

2. Advance Sensing and High Spatio-Temporal Resolution Monitoring

3. Optimize BMP Placement for Multiple Outcomes

4. Engage Stakeholders for Participatory Model Development

These challenges underscore the necessity for collaboration between multistate projects and research institutions. By combining regional expertise, diverse datasets, and interdisciplinary approaches, such collaboration can enhance model development, improve the design and implementation of Best Management Practices (BMPs), and facilitate the integration of real-time monitoring data. This cooperative framework will better inform watershed-scale strategies to achieve water quantity, quality, and climate resilience goals, ultimately advancing both scientific understanding and practical decision-making in water resource management.

Related, Current and Previous Work

Previous multistate projects evaluated and developed models and monitoring approaches to predict the effectiveness of BMPs and their implementation at multiple spatial scales. The last multistate project (S-1089) focused more specifically on the following objectives:

 Develop tools that utilize both monitoring and modeling to better inform targeted BMP implementation,

1. Advance water quantity and quality models for mixed-use watersheds,
2. Test advanced/new monitoring technologies to detect water quality issues, and
3. Conduct integrated assessment of uncertainty and sensitivity analysis of monitoring and modeling approaches both individually and in combination.

Through our efforts over the past five years, we have focused on widely used and supported models to apply them in mixed-use (rural and urban) watersheds. Close integration of new computational and monitoring technologies with stakeholder participation is critically needed to implement practices in agricultural and urban lands that mitigate adverse anthropogenic impacts. In this proposal, we embrace evolving AI and ML techniques to enhance predictive accuracy, automate model calibration, and enable real-time data assimilation beyond what is possible with traditional models. While data collection efforts in our prior project included high-frequency sensors, here we expand those approaches to spatial optimization and the use of newly developed and existing commercial sensors to measure additional nutrients and contaminants that are beyond their marketed capabilities. We leverage the extension efforts of many of our members to more fully integrate stakeholder participation in assessment and decision-making for BMP optimization, planning, and watershed management. Moreover, this multistate effort will allow us to collectively and collaboratively work on these issues and create processes, strategies, and tools that can be replicated across land uses and ecoregions.

Objectives

1. Develop Next-Generation Hybrid Watershed Models
   Comments: Our goals include: 1. Incorporate high-frequency sensor networks, IoT data, and citizen science contributions to improve spatial accuracy and real-time forecasting (provided by Objective 2). 2. Combine process-based models with AI and ML to enhance predictive accuracy, automate calibration, and reduce computational demands. 3. Identify Critical Source Areas using advanced modeling techniques to detect pollutant hotspots and optimize BMP placement for targeted water quality improvements (supporting Objective 3). 4. Develop biogeochemical process modules that integrate wetting and drying cycle effects on pollutant fate and transport.
2. Advance Sensing and High Spatio-Temporal Resolution Monitoring
   Comments: Our goals for this objective are to: 1. Establish methods for new high-frequency sensor installation, maintenance, calibration, rapid identification of malfunctions, and missing data gap-filling. 2. Develop new and enhance current sensors beyond their marketed capabilities to detect and measure additional nutrients and contaminants. 3. Expand current sensor capabilities to short-range spatial resolution to track the fate of pollutants in BMPs and extract biogeochemical process kinetics during hydrological events. 4. Optimize sensor placement to ensure comprehensive spatial coverage while minimizing redundancy. 5. Use high-frequency data to identify and predict watershed water quality traits through pattern recognition in historical and real-time data. 6. Use high-frequency sensor data to calibrate and validate hydrological and water quality models (also in Objective 1).
3. Optimize BMP Placement for Multiple Outcomes
   Comments: The primary tasks for this objective include: 1. Develop and refine modeling tools that identify areas suitable for BMP placement in both agricultural and urban watersheds 2. Collect data to more accurately represent individual practice functions and constraints 3. Conduct economic data analysis on life-cycle of each practice 4. Develop stakeholder-informed decision support through prioritization frameworks. While much of this work is site-specific, our collective efforts will focus on generating repeatable strategies and templates that are transferrable.
4. Engage Stakeholders for Participatory Model Development
   Comments: The primary tasks for this objective include: 1. Engage the community to guide model development by identifying key groups, strategies, and effective collaborative structures. 2. Train the next-generation workforce by developing protocols and by providing training in user-friendly tools. 3. Organize field demonstrations to showcase the applicability of models and technology in practical settings. 4. Enhance modeler awareness of the utility of volunteer monitoring data for improving watershed modeling efforts. 5. Develop and advance new streamlined processes and tools to enhance communication and feedback loops between watershed modelers and stakeholders.

Methods

Objective 1: Develop Next-Generation Hybrid Watershed Models:

Watershed modeling has evolved as a critical tool in understanding hydrologic and water quality dynamics, offering valuable insights for water resource management. The adaptability of traditional models like SWAT, HSPF and DRAINMOD has made them instrumental in policy development, infrastructure planning, and regulatory compliance. These models also help identify effective BMPs for mitigating pollution hotspots and environmental degradation (Mohamoud et al. 2010).

 

Despite their strengths, traditional models present significant limitations that hinder their effectiveness in modern watershed management. Their high dependence on high-quality, site-specific input data often limits their applicability, particularly in data-scarce regions (Hawkins et al. 2015). The extensive calibration and validation processes required for these models demand substantial computational resources and technical expertise, making them time-consuming and costly (Ahmad et al. 2010). Additionally, traditional models struggle to capture dynamic environmental interactions, relying on static assumptions rather than adaptive learning mechanisms (Zhang 2025). As a result, they fail to effectively predict non-linear watershed responses under variable climate and land-use conditions (Muharemi et al. 2019).

 

The emergence of Artificial Intelligence (AI) and Machine Learning (ML) presents a transformative opportunity to overcome the limitations of traditional watershed models. AI-enhanced models address these gaps by integrating high-resolution, real-time data from remote sensing, IoT sensors, and big data analytics. Deep learning and transfer learning techniques further improve predictive robustness, allowing AI models to adapt to different watershed conditions without extensive retraining. Hybrid models—which combine physical process simulations with AI/ML-driven analytics—provide a scalable, data-driven approach to watershed modeling. By incorporating spatially explicit remote sensing data, AI-driven hydrologic forecasts, and multi-fidelity modeling frameworks, these hybrid systems offer a new frontier for sustainable watershed management. AI-driven models can process large datasets from satellite imagery, sensor networks, and climate records, filling spatial and temporal data gaps more efficiently than traditional models. AI-powered real-time monitoring systems allow decision-makers to detect pollution events, adjust watershed management strategies, and improve adaptive conservation planning.

 

To ensure accessibility and facilitate decision-making, a web-based decision support tool will be developed to provide interactive visualization and real-time scenario testing (Objective 4). This platform will allow stakeholders—including policymakers, researchers, and water managers—to explore model outputs, assess potential management strategies, and simulate the effects of conservation practices on water quality and hydrology. Additionally, stakeholder engagement and capacity-building efforts will be a priority. Workshops and training programs will be conducted to familiarize stakeholders with the model’s capabilities, ensuring effective adoption for real-world watershed management applications. To maximize accessibility, the framework will be open-source and cloud-based, promoting collaboration among researchers, government agencies, and industry professionals while allowing for ongoing model improvements based on user feedback.

 

While the above approach will provide greatly enhanced methods to existing models and modeling approaches, this objective will also explore the development of new biogeochemical process modules that will consider the effect of wetting and drying cycles that naturally occur in soils. Evidence of the disproportional impact of these new biogeochemical ‘hot moments’ that occur during hydrological events (details in Objective 2) suggests that current models may have to be updated to better take into account these emerging aspects. We will use new sensing approaches and high-frequency data described in Objective 2 to extract time-series of biogeochemical processes. We will use AI and ML techniques described above to correlate sensing data with observed process kinetics to detect and identify drivers of these new ‘hot moments’. This will be done in an effort to better predict these new ‘hot moments’ patterns and incorporate them as new modules in existing models.

 

Our expanded vision of next-generation modeling must be supported by high-quality data collection to accurately predict hydrological, physical, and biogeochemical processes at multiple scales. While the goals of Objective 1 are focused on modeling creation and application, we recognize they are connected and supported by the other objectives in this proposal; specifically, our goals include:

 

1. Incorporate high-frequency sensor networks, IoT data, and citizen science contributions to improve spatial accuracy and real-time forecasting (provided by Objective 2).
2. Combine process-based models with AI and ML to enhance predictive accuracy, automate calibration, and reduce computational demands.
3. Identify Critical Source Areas using advanced modeling techniques to detect pollutant hotspots and optimize BMP placement for targeted water quality improvements (supporting Objective 3).
4. Develop biogeochemical process modules that integrate wetting and drying cycle effects on pollutant fate and transport.

 

Objective 2: Advance Sensing and High Spatio-Temporal Resolution Monitoring

The increasing availability and publicly available deployment of advanced water quality sensors present unprecedented opportunities to enhance our understanding and management of aquatic ecosystems. The proliferation of sensors capable of measuring a wide range of water quality constituents—such as nutrients, turbidity, dissolved oxygen, and contaminants—corresponds to a true revolution in environmental monitoring, with major changes that range from enabling the discovery of overlooked processes to the largely improving management via real-time wireless communication and the Internet of Things (IoT).

 

It is now clear to all researchers that most of the nutrient loads from a watershed occur during a few hydrological events (Rode et al. 2016). However, the value of ensuring complete water quality data with very missing time points is necessary and worth the cost when specific advantages it may offer are unclear. Having access to high-frequency data during hydrological events has unleashed a new area of research based on the analysis of the relative behavior of hydro- and chemographs. This approach has led to the idea that it might be possible to identify the relative location and dominance of transport compared to biogeochemical processes based on the interpretation of concentration and flow patterns over time (Bieroza et al. 2018).

 

Are biogeochemical process kinetics sensitive to and possibly triggered by the flashy nature of hydrology? There is increasing evidence that alternating wetting and drying cycles in soils may trigger disproportionately high biogeochemical response upon rewetting of the soil during hydrological events (Maxwell et al. 2019; McGuire et al. 2023; Sauers et al. 2025). If confirmed, these ‘hot moments’ of a new kind must be studied as they profoundly affect our current modeling approaches and may actually be prevalent in watersheds and must be considered in new watershed hybrid models (Objective 1).

 

The unveiling of these new ‘hot moments’ requires the ability to track the fate of nutrients along their flow paths in or above the soil, so as to extract the process kinetics and the drivers of disproportional biogeochemical activity. New sensors and new sensing approaches must be developed and applied to increase the number of parameters measured, to compute time series of biogeochemical process kinetics, and to obtain time series of as many indicators as possible which can be used as predictors of the process kinetics in hybrid models.

 

Additionally, the incorporation of continuous, sensor-derived data streams has the potential to improve accuracy and reduce uncertainty and allowing models to capture fine-scale variations in hydrological and existing biogeochemical processes. By strategically deploying sensors in anticipated critical source areas, locations that disproportionately contribute to pollutant loads, we can better capture these dynamic water chemistry traits. Coupling high temporal resolution data into the hybrid models described in Objective 1 also allows us to predict pollutant transport under varying hydrologic conditions and identify locations where siting BMPs may be most impactful (Objective 3). This integration enhances our ability to model nutrient transport, pollutant attenuation, and ecosystem responses to anthropogenic and climatic stressors, thereby informing more precise and responsive management strategies.

Complementing sensor-derived datasets with citizen science monitoring further enriches spatial representativeness by increasing data density across diverse land-use settings. While community-driven water quality assessments are often less precise, they provide localized and spatially representative insights that are often unavailable through conventional monitoring networks. When appropriately integrated using ML algorithms and geostatistical interpolation methods, these data sources can enhance the spatial resolution of biogeochemical and watershed models, bridging gaps in monitoring coverage and improving model accuracy.

To overcome existing limitations and capitalize on emerging opportunities, the tasks in this objective will focus on the development and deployment of sensor networks for continuous environmental monitoring in critical areas. By integrating remote sensing, IoT-enabled sensors, and citizen science platforms, we can enhance the spatial and temporal resolution of water quality assessments. This multi-scale approach ensures that monitoring efforts are both precise and scalable, allowing for real-time decision-making at local, regional, and watershed levels. By creating adaptive monitoring systems that dynamically respond to environmental conditions, we can inform decision making in real-time while also capturing hot moments in time where rapid changes occur. We will leverage AI/ML tools to both analyze complex datasets and support data quality control. Our goals for this objective are to:

 

1. Establish methods for new high-frequency sensor installation, maintenance, calibration, rapid identification of malfunctions, and missing data gap-filling
2. Develop new and enhance current sensors beyond their marketed capabilities to detect and measure additional nutrients and contaminants
3. Expand current sensor capabilities to short-range spatial resolution to track the fate of pollutants in BMPs and extract biogeochemical process kinetics during hydrological events
4. Optimize sensor placement to ensure comprehensive spatial coverage while minimizing redundancy
5. Use high-frequency data to identify and predict watershed water quality traits through pattern recognition in historical and real-time data
6. Use high-frequency sensor data to calibrate and validate hydrological and water quality models (also in Objective 1)

 

Objective 3: Optimize BMP Placement for Multiple Outcomes

Strategic BMP placement can maximize environmental benefits, improve cost-effectiveness, and ensure that conservation efforts align with broader watershed management goals. One of the key challenges in BMP placement is the mismatch between the scale at which pollutants move through the landscape and the scale at which BMPs operate (Lewandowski and Cates, 2023). Pollutant transport, particularly for nutrients like N and P, occurs over varying spatial and temporal scales, influenced by factors such as soil type, topographic position, hydrology, and land management. In contrast, BMPs function at much smaller scales, with their effectiveness dependent on localized processes such as infiltration, retention, and microbial activity. This spatiotemporal disconnect makes it difficult to quantify the effectiveness of BMPs and can create a lag between implementation and observable water quality improvements (Van Meter and Basu 2017).

Additionally, the spatial scale at which water quality improvements can be achieved and measured does not align with the placement of individual BMPs. While a single bioswale, wetland, or cover crop has been shown to be effective at the field or neighborhood level, detecting measurable changes at the watershed outlet requires large-scale implementation across multiple sites (Lewandowski and Cates 2023). While these changes are hard to detect through point-based measurements, advances in sensor technologies are allowing high-resolution time series to be collected at multiple points along flow paths within BMPs and at sub-watershed scales (Objective 2). There is great potential to strategically position these within the landscape to verify models that can be used to create optimization scenarios.

Modeling tools play a critical role in optimizing BMP placement, particularly when multiple practices are implemented together at a single location. When multiple BMPs are stacked, their interactions can be nonlinear, meaning their combined effects may be greater—or, in some cases, less effective—than the sum of their individual impacts. However, without a mechanistic understanding of nutrient cycling, denitrification rates, or sediment transport dynamics, models may produce misleading predictions, leading to suboptimal BMP placement.

A more effective approach to BMP placement must begin with identifying suitable locations through advanced modeling tools. Several tools exist for both rural and urban landscapes, such as the Agricultural Conservation Planning Framework (ACPF) (Tomer et al. 2015; Bravard et al. 2022; Rohith et al. 2024), which provides spatially explicit recommendations for conservation practices on farmland. Beyond identification, prioritization is key. A robust prioritization framework should consider multiple ecosystem services, regulatory constraints, and economic factors such as return on investment. This ensures that conservation dollars are spent on projects that provide the greatest environmental and social benefits. Additionally, the placement of BMPs should align with both local and regional objectives, ensuring that conservation efforts contribute to broader watershed health goals.

 

Equally important is the social dimension of BMP implementation. The most effective conservation strategies balance ecological benefits with economic and social feasibility. Engaging stakeholders—whether individual landowners, municipalities, or regulatory agencies—is essential to ensuring long-term adoption and success. Optimizing BMP placement requires a structured approach that integrates spatial modeling to identify feasible locations with prioritization with meaningful stakeholder involvement while ensuring alignment with local and regional needs. For rural areas, tools like the ACPF identify opportunities for implementation by combining publicly available data including topography, drainage networks, and land management practices. Urban environments require comparable tools that account for impervious surfaces, stormwater management infrastructure, and space constraints. Identifying urban corollaries to ACPF is critical for effectively integrating BMPs in developed areas, where runoff dynamics differ significantly from agricultural settings.

 

A robust prioritization framework is needed to balance multiple factors, including ecosystem services, regulatory constraints, and economic feasibility. Effective BMP placement should maximize benefits such as water quality improvement, flood mitigation, or habitat restoration while also considering constraints like zoning regulations and financial viability. A life-cycle cost analysis that includes forfeited income from land conservation and adaptive management considerations is needed to determine approaches that maximize environmental outcomes while also ensuring long-term success. Project placement must be aligned with local and regional priorities, which requires a transparent decision-making framework that incorporates stakeholder input, including landowners, policymakers, and community members. Engaging stakeholders from the outset fosters support for BMP implementation and ensures that selected sites address both environmental objectives and community concerns. The primary tasks for this objective include:

1. Develop and refine modeling tools that identify areas suitable for BMP placement in both agricultural and urban watersheds
2. Collect data to more accurately represent individual practice functions and constraints
3. Conduct economic data analysis on life-cycle of each practice
4. Develop stakeholder-informed decision support through prioritization frameworks. While much of this work is site-specific, our collective efforts will focus on generating repeatable strategies and templates that are transferrable.

Objective 4: Engage Stakeholders for Participatory Model Development

Integrating stakeholder insights into water resources modeling is crucial for enhancing the effectiveness of watershed management policies. This approach fosters collaboration among scientists, policymakers, and local communities, ensuring that models are both scientifically robust and socially relevant. Models developed with stakeholder participation are often considered more useful for decision makers, provide educational benefits to all involved, and help build consensus among diverse stakeholders.

 

Success stories from across the U.S. and beyond highlight the benefits of collaborative modeling and stakeholder engagement to successfully tackle complex water management challenges. In a review of 180 studies on participatory modeling, researchers provided a guide to help modelers and stakeholders select the most appropriate tools and methods for their projects (Voinov et al. 2018). Effective watershed management involves clearly identifying problems and providing strong leadership for stakeholders (Voinov and Gaddis 2008). They further discussed the importance of engaging stakeholders early and often to build trust and maintain neutrality. Further, it’s important for modelers to involve stakeholders in selecting appropriate modeling tools for addressing the identified problems, gaining approval for methods, and incorporating local knowledge throughout the modeling effort. It’s also important for modelers to discuss uncertainties with stakeholders, work with them to develop feasible scenarios and interpret results in a manner that helps stakeholders create needed policies and implementation plans. Finally, the modeling process must be considered by all involved as an ongoing, adaptive, iterative process and not a final solution.

 

While most watershed management decisions benefit from stakeholder input, some issues may not attract wide interest. If stakeholders do not understand or consider the problem important, engaging them in a participatory process will be difficult. Educating the community about water resource issues and their impact is often a good first step. Citizen science monitoring programs offer an excellent opportunity to educate and involve citizens in water resources management efforts, while also filling critical spatio-temporal water quality data gaps needed for effective modeling and watershed planning. Volunteer monitoring programs such as Oklahoma’s Blue Thumb Program, the Texas Stream Team, Water Watch, and others have engaged thousands of citizens in watershed monitoring and management. Volunteer data acquisition, making up over 20% of the Virginia’s Clean Water Act report, saved the state $3.25 million annually (Wyeth 2023).

 

Stakeholder involvement has the potential to address current limitations, which include the lack of high-resolution data, the lack of communication between watershed modelers and stakeholders, and low community-awareness. There is a need for a holistic approach to solve these limitations. Integrating volunteer monitoring, new sensors and satellite data would largely enhance data availability and greatly help watershed management decisions. This would be beneficial, however, only if there would be tools to streamline processes and tools to enhance communication and feedback loops between watershed modelers and stakeholders, increasing awareness and efficacy.

However, two main limitations to the use of current citizen science approaches must be addressed, including the quality and consistency of the data, and ensuring that key parameters are measured. While citizen scientists' involvement offers valuable educational benefits and foster community engagement, ensuring data validity and reliability remains a challenge for regulatory acceptance (Wyeth 2023). Sparse or unharmonized frequency of acquisition is often problematic (Ramírez et al. 2023). However, utilization of citizen science data alongside other data streams in modeling efforts (calibration, validation) offers a real opportunity for using this data in a manner that alleviates many agency concerns regarding quality and consistency because the modeling is not solely relying on the volunteer data for analysis. Citizen-based acquisition of concentration information on water quality parameters can become useless, unfortunately, if essential data such as flow is not measured at the same time. Thankfully, we believe there are methods to overcome the limitations detailed here. One is to use new technologies to fill the educational gaps. This might include educational videos to help users quickly overcome learning curves for a wide variety of tasks. Another is making watershed model platforms more user-friendly to encourage broader stakeholder involvement.

Through this project, we will promote integrated watershed management strategies involving a wide range of stakeholders (government, communities, industries, NGOs), and emphasize shared responsibility and collaborative tools to address emerging contaminants, climate resilience, and cumulative impacts of BMPs. The primary tasks for this objective include:

1. Engage the community to guide model development by identifying key groups, strategies, and effective collaborative structures.
2. Train the next-generation workforce by developing protocols and by providing training in user-friendly tools
3. Organize field demonstrations to showcase the applicability of models and technology in practical settings
4. Enhance modeler awareness of the utility of volunteer monitoring data for improving watershed modeling efforts
5. Develop and advance new streamlined processes and tools to enhance communication and feedback loops between watershed modelers and stakeholders

Measurement of Progress and Results

Outputs

• next-generation hybrid watershed models Comments: The following are the specific outputs: 1. Hybrid process-based models coupled with AI/ML approaches 2. Biogeochemical process modules to integrate into existing models 3. Processes for incorporating high-frequency data into model calibration and validation
• high-temporal frequency monitoring data Comments: The following are the specific outputs: 1. Data to better constrain pollutant sources and transport dynamics 2. Strategies for sensor implementation 3. Novel approaches to detect emerging contaminants
• Optimial BMP placement for multiple outcomes Comments: The following are the specific outputs: 1. Maps to show opportunities and guide decision-making 2. Decision support playbook based on publicly available data and site-specific goals 3. Strategies for aggregating multiple ecosystem outcomes into watershed management
• stakeholder engagement for participatory model development Comments: The following are the specific outputs: 1. Cooperation between researchers, local stakeholders, agencies, and agricultural producers 2. Processes and strategies for successful volunteer monitoring 3. Next-generation workforce training

Outcomes or Projected Impacts

• Improved prediction and forecast skills of water quality and hydrology The following are the specific outcomes: 1. Enhanced prediction of watershed-scale hydrology and water quality 2. Greater appreciation for novel modeling tools 3. Improved modeling efficiency
• Monitoring and understanding pollutant fate and transport at high temporal resolution The following are the specific outcomes: 1. Enhanced understanding of pollutant fate and transport at small watershed scales 2. Timely and targeted monitoring for watershed management approaches.
• More efficient and cost effective BMP placement The following are the specific outcomes: 1. Improved tools for targeted practice placement and adoption 2. Increased awareness of the value of including multiple outcomes, benefits, and constraints in siting BMPs in urban and rural lands
• Stronger stakeholder involvement with increased trust between scientists and stakeholders The following are the specific outcomes: 1. Increased trust in model outputs by stakeholders 2. Enhanced appreciation of the utility of volunteer monitoring data for watershed management

Milestones

(2026):1. Develop AI-enhanced hybrid watershed models integrating watershed models, including but not limited to SWAT, HSPF, and DRAINMOD. Incorporate IoT sensors, remote sensing, and data fusion techniques. Publish initial findings and develop training materials. 2. Develop new and enhance current sensors beyond their marketed capabilities to detect and measure additional nutrients and contaminants 3. Establish watershed exemplars across our multi-state region as test cases for multi-outcome optimization; identify appropriate models, data gaps, and needed stakeholders 4. Work across Objectives 1 and 3 to identify watershed stakeholders, strategies, and effective collaborative structures

(2027):1. Implement AI-driven calibration techniques. Deploy real-time sensor networks for high-frequency data collection.  Integrate biogeochemical process modules capturing wetting/drying cycles. Optimize BMP placement using AI-based spatial modeling. Expand stakeholder engagement and model refinement. 2. Expand current sensor capabilities to short-range spatial resolution, and use high-frequency sensor data to calibrate and validate hydrological and water quality models. 3. Survey literature and collect additional field data related to management practice functions and performance considering multiple ecosystem services 4. Identify processes and strategies for successful volunteer monitoring and develop communication loops and feedback with modeling community.

(2028):1. Develop a web-based AI-enhanced decision-support tool. Expand AI-driven watershed modeling across diverse landscapes. Enhance flood and drought forecasting using deep learning models. Conduct mid-term evaluation and stakeholder feedback and discussions. 2. Establish methods for new high-frequency sensor installation, maintenance, calibration, rapid identification of malfunctions, and missing data gap-filling. 3. Develop and refine modeling tools to identify and prioritize suitable locations for BMP placement; apply to watershed test cases. 4. Organize field demonstrations to showcase the applicability of models and technology in practical settings, and initial pilot test of stakeholder engagement with model development in concert with Objectives 1 and 3.

(2029):1. Deploy AI-assisted real-time watershed management tools. Validate AI-driven BMP optimization across multiple watersheds. Implement large-scale stakeholder training and workforce development. 2. Use high-frequency data to identify and predict watershed water quality traits through pattern recognition in historical and real-time data, and optimize sensor placement to ensure comprehensive spatial coverage while minimizing redundancy. 3. Continue development of watershed test case optimization models; engage stakeholders and incorporate feedback to refine model utility and outputs (e.g., maps) 4. Incorporate lessons learned from year 3 training and engagement to refine strategies; expand stakeholder engagement with modeling efforts.

(2030):1. Finalize AI-enhanced models for policy integration. Ensure long-term sustainability through partnerships and open-access tools. 2. Use high-frequency sensor data to calibrate and validate hydrological and water quality models, track the fate of pollutants in BMPs and extract biogeochemical process kinetics during hydrological events. 3. Produce guidelines and toolkits for prioritization framework to jointly consider BMP ecosystem service provision and economic cost as transferrable template. 4. Produce guidelines to engage stakeholders in participatory watershed modeling, including streamlined processes and tools for communication and feedback loops between watershed modelers and stakeholders

Projected Participation

View Appendix E: Participation

Outreach Plan

The ultimate beneficiaries will be land users, home owners, watershed planners, municipal stormwater managers and other stakeholders who will be impacted by the BMPs program. All stakeholders will benefit from potential water quality improvements, and landowners and taxpayers will benefit from the development of efficient BMP implementation plans that are more economically feasible. The immediate and direct beneficiaries of the project will be Federal, State, and local agency personnel and consultants. The results of this project will be communicated through factsheets, conference presentations and publications authored by members, and annual meetings and reports. This multi-state proposal includes participants with both research and extension appointments. As such, our members will leverage existing programs and disseminate findings through their state and regional programs. At annual project meetings, time will be devoted to interacting with extension professionals at the host institution to create an exchange of information. The project participants will share insights and advances, and the extension professionals can provide feedback and suggestions on how to improve outputs that are useful and accessible to stakeholders.

Organization/Governance

The project’s membership will elect a project chair annually in September. The project chair will be responsible for organizing the annual project meeting and submitting the project’s annual report. The proposal/technical committee consisting of researchers and extension specialists from various land grant institutions (shown below) will serve as the project’s executive committee and will assist the project chair in developing the agenda for the annual meeting and in tracking progress toward achieving the project objectives. The chair elected in the final year of the project will be responsible for submitting the final report. The project’s past chairs and executive committee will assist the current chair with compiling the final report. Administrative guidance will be provided by an assigned Administrative Advisor and a CSREES Representative. The detailed list of the project participants can be found on the project website hosted by the NIMSS.

Literature Cited

Ahmad MM, Ghumman AR, Ahmad S, Hashmi HN. 2010. Estimation of a Unique Pair of Nash Model Parameters: An Optimization Approach. Water Resour Manage. 24(12):2971–2989. doi:10.1007/s11269-010-9590-3.

Bieroza MZ, Heathwaite AL, Bechmann M, Kyllmar K, Jordan P. 2018. The concentration-discharge slope as a tool for water quality management. Sci Total Environ. 630:738–749. doi:10.1016/j.scitotenv.2018.02.256.

Bravard EE, Zimmerman E, Tyndall JC, James D. 2022. The Agricultural Conservation Planning Framework Financial and Nutrient Reduction Tool: A planning tool for cost effective conservation. Journal of Environmental Quality. 51(4):670–682. doi:10.1002/jeq2.20345.

Bronstert A. 1999. Capabilities and limitations of detailed hillslope hydrological modelling. Hydrological Processes. 13(1):21–48. doi:10.1002/(SICI)1099-1085(199901)13:1<21::AID-HYP702>3.0.CO;2-4.

Casper AF, Dixon B, Earls J, Gore JA. 2011. Linking a spatially explicit watershed model (SWAT) with an in‐stream fish habitat model (PHABSIM): A case study of setting minimum flows and levels in a low gradient, sub‐tropical river. River Research & Apps. 27(3):269–282. doi:10.1002/rra.1355.

Du X, Su J, Li X, Zhang W. 2016. Modeling and Evaluating of Non‐Point Source Pollution in a Semi‐Arid Watershed: Implications for Watershed Management. CLEAN Soil Air Water. 44(3):247–255. doi:10.1002/clen.201400773.

Exner ME, Hirsh AJ, Spalding RF. 2014. Nebraska’s groundwater legacy: Nitrate contamination beneath irrigated cropland. Water Resources Research. 50(5):4474–4489. doi:10.1002/2013WR015073.

Gassman PW, Sadeghi AM, Srinivasan R. 2014. Applications of the SWAT Model Special Section: Overview and Insights. J Environ Qual. 43(1):1–8. doi:10.2134/jeq2013.11.0466.

Gesch K, Kiel A, Sutphin T, Wolf R. 2020. Integrating farmer input and Agricultural Conservation Planning Framework results to develop watershed plans in Iowa. Journal of Soil and Water Conservation. 75(4):101A-104A. doi:10.2489/jswc.2020.0226A.

Hawkins GA, Vivoni ER, Robles-Morua A, Mascaro G, Rivera E, Dominguez F. 2015. A climate change projection for summer hydrologic conditions in a semiarid watershed of central Arizona. Journal of Arid Environments. 118:9–20. doi:10.1016/j.jaridenv.2015.02.022.

Jang T, Vellidis G, Hyman JB, Brooks E, Kurkalova LA, Boll J, Cho J. 2013. Model for Prioritizing Best Management Practice Implementation: Sediment Load Reduction. Environmental Management. 51(1):209–224. doi:10.1007/s00267-012-9977-4.

Jayakrishnan R, Srinivasan R, Santhi C, Arnold JG. 2005. Advances in the application of the SWAT model for water resources management. Hydrological Processes. 19(3):749–762. doi:10.1002/hyp.5624.

Kim Y-W, Lee J-W, Woo S-Y, Lee J-J, Hur J-W, Kim S-J. 2023. Design of Ecological Flow (E-Flow) Considering Watershed Status Using Watershed and Physical Habitat Models. Water. 15(18):3267. doi:10.3390/w15183267.

Lewandowski AM, Cates A. 2023. Connecting soil health and water quality in agricultural landscapes. Journal of Environmental Quality. 52(3):412–421. doi:10.1002/jeq2.20390.

Mangukiya NK, Sharma A, Shen C. 2023. How to enhance hydrological predictions in hydrologically distinct watersheds of the Indian subcontinent? Hydrological Processes. 37(7):e14936. doi:10.1002/hyp.14936.

Maxwell BM, Birgand F, Schipper LA, Christianson LE, Tian S, Helmers MJ, Williams DJ, Chescheir GM, Youssef MA. 2019. Drying–Rewetting Cycles Affect Nitrate Removal Rates in Woodchip Bioreactors. Journal of Environmental Quality. 48(1):93–101. doi:10.2134/jeq2018.05.0199.

McGuire PM, Butkevich N, Saksena AV, Walter MT, Shapleigh JP, Reid MC. 2023. Oxic–anoxic cycling promotes coupling between complex carbon metabolism and denitrification in woodchip bioreactors. Environmental Microbiology. 25(9):1696–1712. doi:10.1111/1462-2920.16387.

Meals DW, Dressing SA, Davenport TE. 2010. Lag Time in Water Quality Response to Best Management Practices: A Review. Journal of Environmental Quality. 39(1):85–96. doi:10.2134/jeq2009.0108.

Messer TL, Burchell MR, Bírgand F. 2017. Comparison of Four Nitrate Removal Kinetic Models in Two Distinct Wetland Restoration Mesocosm Systems. Water. 9(7):517. doi:10.3390/w9070517.

Mohamoud YM, Parmar R, Wolfe K. 2010. Modeling Best Management Practices (BMPs) with HSPF. In: Watershed Management 2010. Madison, Wisconsin, United States: American Society of Civil Engineers. p. 892–898. [accessed 2025 Mar 13]. http://ascelibrary.org/doi/10.1061/41143%28394%2981.

Muharemi F, Logofătu D, Leon F. 2019. Machine learning approaches for anomaly detection of water quality on a real-world data set. Journal of Information and Telecommunication. 3(3):294–307. doi:10.1080/24751839.2019.1565653.

Pinay G, Haycock NE. 2019. Diffuse nitrogen pollution control: Moving from riparian zone to headwater catchment approach—A tribute to the influence of Professor Geoff Petts. River Research and Applications. 35(8):1203–1211. doi:10.1002/rra.3488.

Ramírez SB, van Meerveld I, Seibert J. 2023. Citizen science approaches for water quality measurements. Sci Total Environ. 897:165436. doi:10.1016/j.scitotenv.2023.165436.

Rode M, Wade AJ, Cohen MJ, Hensley RT, Bowes MJ, Kirchner JW, Arhonditsis GB, Jordan P, Kronvang B, Halliday SJ, et al. 2016. Sensors in the Stream: The High-Frequency Wave of the Present. Environ Sci Technol. 50(19):10297–10307. doi:10.1021/acs.est.6b02155.

Rohith AN, Karki R, Veith TL, Preisendanz HE, Duncan JM, Kleinman PJA, Cibin R. 2024. Prioritizing conservation practice locations for effective water quality improvement using the Agricultural Conservation Planning Framework (ACPF) and the Soil and Water Assessment Tool (SWAT). Journal of Environmental Management. 349:119514. doi:10.1016/j.jenvman.2023.119514.

Santhi C, Arnold JG, Williams JR, Dugas WA, Srinivasan R, Hauck LM. 2001. Validation of the swat model on a large rwer basin with point and nonpoint sources. J American Water Resour Assoc. 37(5):1169–1188. doi:10.1111/j.1752-1688.2001.tb03630.x.

Sauers N, Rok A, Birgand F. 2025. Evidence of Nitrate Removal “Hot Moments” During Flow and Nitrate Pulses in a Denitrification “Hot Spot.” Journal of the ASABE. 68(2). doi:doi: 10.13031/ja.15988. [accessed 2025 Mar 13]. https://elibrary.asabe.org/abstract.asp?aid=55063&t=3&dabs=Y&redir=&redirType=.

Tomer MD, Porter SA, Boomer KMB, James DE, Kostel JA, Helmers MJ, Isenhart TM, McLellan E. 2015. Agricultural Conservation Planning Framework: 1. Developing Multipractice Watershed Planning Scenarios and Assessing Nutrient Reduction Potential. Journal of Environmental Quality. 44(3):754–767. doi:10.2134/jeq2014.09.0386.

Van Meter KJ, Basu NB. 2017. Time lags in watershed-scale nutrient transport: an exploration of dominant controls. Environ Res Lett. 12(8):084017. doi:10.1088/1748-9326/aa7bf4.

Vivoni ER, Richards KT. 2005. Integrated use of GIS-based field sampling and modeling for hydrologic and water quality studies. Journal of Hydroinformatics. 7(4):235–250. doi:10.2166/hydro.2005.0021.

Voinov A, Gaddis EJB. 2008. Lessons for successful participatory watershed modeling: a perspective from modeling practitioners. Ecological modelling. 216(2):197–207.

Voinov A, Jenni K, Gray S, Kolagani N, Glynn PD, Bommel P, Prell C, Zellner M, Paolisso M, Jordan R. 2018. Tools and methods in participatory modeling: Selecting the right tool for the job. Environmental Modelling & Software. 109:232–255.

Wyeth G. 2023. Integrating Citizen Science into the Work of United States Environmental Agencies. Citizen Science: Theory and Practice. 8(1). doi:10.5334/cstp.490. [accessed 2025 Mar 13]. https://theoryandpractice.citizenscienceassociation.org/articles/10.5334/cstp.490.

Zhang J. 2025. Ensemble learning-based approach for water quality prediction. In: Yuan H, Leng L, editors. Fourth International Conference on Computer Vision, Application, and Algorithm (CVAA 2024). Chengdu, China: SPIE. p. 88. [accessed 2025 Mar 13].

Attachments

Land Grant Participating States/Institutions

TN

Non Land Grant Participating States/Institutions