A Hetero-functional Graph State Estimator for Watershed Systems: Application to the Chesapeake Bay
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Robotics Radio. Generated commentary on the latest robotics and control papers.
Rosa: I'm Rosa.
Dev: And I'm Dev, here with Taro.
Rosa: We're looking at the paper: "A Hetero-functional Graph State Estimator for Watershed Systems".
Dev: Here's the gist of the paper. A Hetero-functional Graph State Estimator for Watershed Systems: Application to the Chesapeake Bay introduces a modeling framework that integrates open-source geospatial processing with an estimation methodology grounded in Hetero-functional Graph Theory (HFGT) to infer nutrient flows from uncertain data in watershed systems, which is important because it enables the integration of economic and governance systems to support sustainable watershed management Modeling Framework Foundation The paper introduces a modeling framework for watershed systems that integrates open-source geospatial processing with an estimation methodology grounded in HFGT The framework is demonstrated through instantiation in the Chesapeake Bay Watershed First, a region-independent reference architecture is developed Second, the Weighted Least Squares Error Hetero-functional Graph State Estimator, an extension of Hetero-functional Graph Theory (HFGT), is adapted to estimate nutrient flows from uncertain data By establishing a shared ontology grounded in Systems Modeling Language and HFGT, the approach enables integration of economic and governance systems to support sustainable watershed management System Definition and Architecture The paper develops a reference architecture for watershed connectivity and nutrient application This architecture is built upon the Hetero-functional Graph Theory (HFGT) meta-architecture, which represents engineering systems through resources, processes, and operands The organization of these elements is formalized using a Reference Architecture (RA) that categorizes both system form and function within a unified ontology The generic watershed system is defined by a Block Definition Diagram (BDD) and an Activity Diagram (ACT) that specialize the HFGT meta-architecture Estimation Methodology The core estimation methodology is the Weighted Least Squares Error Hetero-functional Graph State Estimation (WLSEHFGSE), which extends the Hetero-functional Network Minimum Cost Flow (HFNMCF) problem The WLSEHFGSE problem extends the HFNMCF problem by explicitly incorporating uncertainty in exogenous data It transforms the HFNMCF from a purely optimization-based flow model into a generalized estimation framework suitable for reconstructing real-world systems, such as the Chesapeake Bay Watershed, from empirical data The objective function minimizes the weighted sum of squared errors on the exogenously defined flows Application and Results The framework is demonstrated through instantiation in the Chesapeake Bay Watershed The synthesis of the hetero-functional incidence tensor for the watershed engineering system net state transition function begins with constructing a spatially explicit formal representation of the CBW This geospatial workflow generates the GIS-based resource definitions needed for the HFGT toolbox input file and, ultimately, the hetero-functional incidence matrix that encodes which capabilities inject or eject each operand for every CAST-defined buffer The reconstructed transport network visualizes nutrient movement through delineated river segments, revealing the implicit topological structure underlying the aggregated delivery factors of CAST The aggregate mass-balance closure (errors < 0.01% at EoT) demonstrates that these local deviations offset rather than accumulate Future Directions The paper establishes an extensible foundation for integrating additional processes and domains that influence watershed function Future work towards the Chesapeake Bay Watershed case study can focus on coupling this model to the estuary model by incorporating freshwater delivery and nutrient phase partitioning to improve representation of land–estuary connectivity Similarly, assimilating monitoring data to estimate intermediate fluxes and refine spatial variability in nutrient transport would enhance accuracy of constructed flows Finally, because CAST is a time-averaged statistical emulator of the original HSPF-based watershed model, a natural next step is to compare the reconstructed HFGT network against outputs from the fully dynamic HSPF lineage to evaluate consistency in spatial routing and mass conservation The extensibility of the MBSE-HFGT architecture enables integration with models of human systems that shape nutrient dynamics Data Integration Details The measurement of exogenously defined variables is largely dependent on the instantiated system CAST behavioral data is used to derive the watershed engineering system net measurement function The formulation supports the incorporation of data that is coarse or incomplete both temporally and spatially In place of a single selection matrix DU, separate capability and temporal aggregation matrices–DE (Defn.14) and DT (Defn.
Title and authors: Rosa: So, focusing on the title and the authors for "A Hetero-functional Graph State Estimator for Watershed Systems: Application to the Chesapeake Bay", what’s actually going on there? What’s that HFGT thing they are using?
Dev: The authors are Megan Harris, John Little, and Amro Farid. They seem to be linking systems modeling language with this graph theory approach.
Rosa: It sounds like the title is highlighting two big ideas: the graph state estimator for watershed systems and the application to Chesapeake Bay. That’s a lot of complexity packed into one name.
Taro: I wonder if they are trying to move away from those old, opaque routing models that we talked about?
Dev: Yeah, they mention that regional-specific models like the Chesapeake Assessment Scenario Tool rely on structurally opaque watershed routing, so this paper is aiming for something more transparent.
The paper's summary: Rosa: So the summary says they developed a region-independent reference architecture first. That means they didn't just jump into the Bay immediately; they built a general way to model any watershed system that can be applied anywhere.
Dev: And then they adapted this Weighted Least Squares Error Hetero-functional Graph State Estimator, or WLSEHFGSE, to estimate nutrient flows from uncertain data.
Taro: So instead of just giving us one fixed answer for the flow, this method lets them infer what the flow actually is based on what we measure.
Rosa: Right. They established a shared language—a systems modeling language grounded in HFGT—which is key because it lets them integrate those economic and governance systems we mentioned earlier.
Dev: So, the core idea is to use this framework to reconstruct real-world systems like the Chesapeake Bay Watershed from empirical data by minimizing weighted squared errors on the flows they estimate.
The paper's improvements: Rosa: Now, looking at what they suggest as improvements for this approach, what are the main things they want to do next? What’s the path forward?
Dev: They mention a few key directions. First, coupling this model to an estuary model by incorporating freshwater delivery and nutrient phase partitioning to get a better picture of land-estuary connectivity.
Taro: That makes sense for refining the land-estuary connection part. What else are they looking at? Where does the uncertainty come from?
Rosa: They also plan to assimilate monitoring data to estimate intermediate fluxes, which should help refine the spatial variability of nutrient transport in the reconstructed flows.
Dev: And a really important one for consistency: they plan to compare this reconstructed HFGT network against outputs from those fully dynamic models, like the HSPF lineage.
Taro: Because they mention that their tool is a time-averaged statistical emulator of the original HSPF-based model, comparing them should show if the spatial routing and mass conservation are consistent.
Conclusion: Rosa: So to wrap this up on "A Hetero-functional Graph State Estimator for Watershed Systems: Application to the Chesapeake Bay", this paper provides a framework that moves us beyond just calculating nutrient estimates in isolation.
Dev: It gives us a structurally transparent foundation for management decisions because it creates an explicit, mass-conserving network that connects land segments, rivers, and the estuary.
Taro: And by using this approach, they can even prioritize data collection efforts because the model shows where transport-capability residuals are largest where CAST behavioral information is most aggregated and least directly observed.
Rosa: It’s about using this modeling power to see how we can better integrate economic and governance systems into watershed management strategies.
Dev: This work sets up a way to infer unknown or partially observed flows from real-world data while preserving physical feasibility and structural consistency, which is crucial for our control loops.
Taro: And it shows us that when the world misbehaves, we can have a system that not only estimates the state but also tells us where the biggest gaps in our knowledge are.
Rosa: That’s what this paper does with the Chesapeake Bay data, by providing a tool to see how systems are connected across scales.
Dev: It’s an extensible foundation, and I think linking it to human systems modeling is where the real power for sustainable management lies.
Virginia Tech · Stevens Institute of Technology
eess.SY, cs.SY
Submitted: 2026-03-02
Updated: 2026-11-03
Code: https://github.com/LIINES/HFGT-CBW-Watershed-Model
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 72/100
The gist: A Hetero-functional Graph State Estimator for Watershed Systems: Application to the Chesapeake Bay introduces a modeling framework that integrates open-source geospatial processing with an estimation…
Key concepts
- Hetero-functional Graph Theory (HFGT)
- HFGT is a mathematical framework used to model engineering systems by defining them through resources, processes, and operands. It provides the underlying structure necessary to represent complex watershed connectivity and nutrient application within a unified ontology.
- Weighted Least Squares Error Hetero-functional Graph State Estimation (WLSEHFGSE)
- This is the core estimation method that extends existing flow models. It takes uncertain real-world data and uses it to estimate unknown flows by minimizing the weighted sum of squared errors, ensuring the reconstructed nutrient transport network remains physically feasible.
- Systems Modeling Language (SML) Ontology
- A shared vocabulary built on SML and HFGT that standardizes how different components of a watershed system are described. This allows for the integration of diverse systems, such as physical nutrient flows and human economic/governance decisions, into one coherent model.
- CAST (Time-Averaged Statistical Emulator)
- CAST is a tool used to derive the net measurement function of the watershed engineering system. It helps in incorporating coarse or incomplete data by aggregating behavioral information, which is then used to measure how nutrients move through the system.
Terminology
Summary
A Hetero-functional Graph State Estimator for Watershed Systems: Application to the Chesapeake Bay introduces a modeling framework that integrates open-source geospatial processing with an estimation methodology grounded in Hetero-functional Graph Theory (HFGT) to infer nutrient flows from uncertain data in watershed systems, which is important because it enables the integration of economic and governance systems to support sustainable watershed management
Modeling Framework Foundation
The paper introduces a modeling framework for watershed systems that integrates open-source geospatial processing with an estimation methodology grounded in HFGT The framework is demonstrated through instantiation in the Chesapeake Bay Watershed First, a region-independent reference architecture is developed Second, the Weighted Least Squares Error Hetero-functional Graph State Estimator, an extension of Hetero-functional Graph Theory (HFGT), is adapted to estimate nutrient flows from uncertain data By establishing a shared ontology grounded in Systems Modeling Language and HFGT, the approach enables integration of economic and governance systems to support sustainable watershed management
System Definition and Architecture
The paper develops a reference architecture for watershed connectivity and nutrient application This architecture is built upon the Hetero-functional Graph Theory (HFGT) meta-architecture, which represents engineering systems through resources, processes, and operands The organization of these elements is formalized using a Reference Architecture (RA) that categorizes both system form and function within a unified ontology The generic watershed system is defined by a Block Definition Diagram (BDD) and an Activity Diagram (ACT) that specialize the HFGT meta-architecture
Estimation Methodology
The core estimation methodology is the Weighted Least Squares Error Hetero-functional Graph State Estimation (WLSEHFGSE), which extends the Hetero-functional Network Minimum Cost Flow (HFNMCF) problem The WLSEHFGSE problem extends the HFNMCF problem by explicitly incorporating uncertainty in exogenous data It transforms the HFNMCF from a purely optimization-based flow model into a generalized estimation framework suitable for reconstructing real-world systems, such as the Chesapeake Bay Watershed, from empirical data The objective function minimizes the weighted sum of squared errors on the exogenously defined flows
Application and Results
The framework is demonstrated through instantiation in the Chesapeake Bay Watershed The synthesis of the hetero-functional incidence tensor for the watershed engineering system net state transition function begins with constructing a spatially explicit formal representation of the CBW This geospatial workflow generates the GIS-based resource definitions needed for the HFGT toolbox input file and, ultimately, the hetero-functional incidence matrix that encodes which capabilities inject or eject each operand for every CAST-defined buffer The reconstructed transport network visualizes nutrient movement through delineated river segments, revealing the implicit topological structure underlying the aggregated delivery factors of CAST The aggregate mass-balance closure (errors < 0.01% at EoT) demonstrates that these local deviations offset rather than accumulate
Future Directions
The paper establishes an extensible foundation for integrating additional processes and domains that influence watershed function Future work towards the Chesapeake Bay Watershed case study can focus on coupling this model to the estuary model by incorporating freshwater delivery and nutrient phase partitioning to improve representation of land–estuary connectivity Similarly, assimilating monitoring data to estimate intermediate fluxes and refine spatial variability in nutrient transport would enhance accuracy of constructed flows Finally, because CAST is a time-averaged statistical emulator of the original HSPF-based watershed model, a natural next step is to compare the reconstructed HFGT network against outputs from the fully dynamic HSPF lineage to evaluate consistency in spatial routing and mass conservation The extensibility of the MBSE-HFGT architecture enables integration with models of human systems that shape nutrient dynamics
Data Integration Details
The measurement of exogenously defined variables is largely dependent on the instantiated system CAST behavioral data is used to derive the watershed engineering system net measurement function The formulation supports the incorporation of data that is coarse or incomplete both temporally and spatially In place of a single selection matrix DU, separate capability and temporal aggregation matrices–DE (Defn.14) and DT (Defn. p. 3 Introduction
The modern societal challenges reflected in the United Nations Sustainable Development Goals (SDGs) are inherently interconnected, arising from
interactions among multiple systems of the Earth and society known as systems of systems (Defn. 1) [1, 2, 3].Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved AI can perform explicit system of systems modeling by integrating economic and governance systems into nutrient flow inference, as stated in
By establishing a shared ontology grounded in Systems Modeling Language and HFGT, the approach enables integration of economic and governance systems to support sustainable watershed management.
-
Improved AI can infer unknown or partially observed flows from uncertain data by extending the Weighted Least Squares Error Hetero-functional Graph State Estimator (WLSEHFGSE), which
enables the inference of unknown or partially observed flows from real-world data while preserving physical feasibility and structural consistency.
-
Improved AI can reconstruct the implicit flow structure of existing empirical models, transforming them into an
explicit, mass-conserving network that connects land segments, rivers, and the estuary,
thereby providing astructurally transparent
foundation for management decisions. -
Improved AI can prioritize data collection efforts by identifying where uncertainty is dominated by
sparsity or aggregation,
as the model reveals wheretransport-capability residuals are largest where CAST behavioral information is most aggregated and least directly observed.
Abstract
Regional watersheds are systems of systems in which hydrology, land use, and governance interact. Nutrient models such as the Chesapeake Assessment Scenario Tool (CAST) route nutrients from land to the estuary through lumped delivery factors, so the river network that connects sources to the estuary is not represented explicitly. This paper develops a region-independent watershed reference architecture in Hetero-functional Graph Theory (HFGT) and adapts the Weighted Least Squares Error Hetero-functional Graph State Estimator to estimate nutrient flows in that network from uncertain data. The framework is applied to the Chesapeake Bay watershed. A mass-conserving network of land segments, outlet points, and river reaches is built from public geospatial data. CAST applied loads, land-segment edge-of-river loads, and river-to-bay delivery factors enter the estimator as measurements, each with its own error term. CAST-reported loads delivered to the Bay are withheld from estimation and used only to test the estimated delivery. The estimator reproduces the applied loads (R 2=1.00) and the land-segment edge-of-river loads (R 2=0.997) and under-predicts delivery to the Bay by 8.0% for nitrogen and 5.7% for phosphorus. The gap is traced to CAST delivery factors that imply in-river gains, which a mass-conserving network cannot represent, and to land segments whose edge-of-river loads exceed their applied loads. The framework provides a common systems representation for coupling watershed models with infrastructure, economic, and governance models.
Sources
- An Introduction to Model-Based Systems Engineering and Hetero-functional Graph Theory for Hydrological Systems of Systems
- The Hetero-functional Graph Theory Toolbox
Related papers
- Industrial Load Modeling and Optimization for Market-Based Interaction with Power Systems
- Exact and Approximate Convex Reformulation of Linear Stochastic Optimal Control with Chance Constraints
- Certified Set Convergence for Piecewise Affine Systems
- Symbolic Control for Stochastic Systems via Finite Parity Games
- Multiobjective optimization-based design and dispatch of islanded, hybrid microgrids for remote, off-grid communities in sub-Saharan Africa
- Physiology-informed layered sensing for intelligent human-exoskeleton interaction