Knowledge-guided graph machine learning for spatially distributed prediction of daily discharge and nitrogen export dynamics

Water Research · 2026-02-24

Knowledge‐Guided Machine Learning for Operational Flood Forecasting

Water Resources Research · 2025-11-01 · 3 citations

Abstract We present a knowledge‐guided machine learning framework for operational hydrologic forecasting at the catchment scale. Our approach, a Factorized Hierarchical Neural Network (FHNN), has two main components: inverse and forward models. The inverse model uses observed precipitation, temperature, and streamflow data to generate a representation of the current underlying catchment state. The forward model predicts streamflow using the learned catchment state. The FHNN architecture is designed to model multi‐scale processes and capture their interactions, a critical ability for flood modeling. FHNN also improves forecasts based on real‐time data through an inference‐based data integration approach using inverse modeling. FHNN's data integration approach improves forecasts in response to observed data more efficiently than data assimilation methods that require computationally intensive optimization. We compare the FHNN to a leading deep learning alternative (autoregressive LSTM) on the large‐sample CAMELS‐US data set, and operational flood forecast data from the US National Weather Service (NWS). Official NWS flood forecasts are generated by expert human forecasters using a physics‐based model, in a human‐in‐the‐loop process. Thus, we assess the flood forecast ability of FHNN by directly comparing its performance against these NWS expert‐derived forecasts. The human forecaster creates a more accurate forecast within the first 12–18 hr of a forecast's issuance, but FHNN has significantly better predictions thereafter. This research lays the groundwork for leveraging the predictive performance of AI‐based models with the expertise in forecasting agencies to produce better river forecasts.

Knowledge-Guided Machine Learning for Operational Flood Forecasting

HydroShare Resources · 2025-05-07 · 1 citations

Biochemical Processes within a Two-Stage Agricultural Drainage Ditch in Mower County, Mn: Methods for Estimating Nitrogen Removal Rates and Efficiencies

SSRN Electronic Journal · 2025-01-01

Hierarchically Disentangled Recurrent Network for Factorizing System Dynamics of Multi-scale Systems: An application on Hydrological Systems

2025-11-12

We present a framework for modeling multi-scale processes, and study its performance in the context of stream-flow forecasting in hydrology. Specifically, we propose a novel hierarchical recurrent neural architecture that factorizes the system dynamics at multiple temporal scales and captures their interactions. This framework consists of an inverse and a forward model. The inverse model is used to empirically resolve the system's temporal modes from data (physical model simulations, observed data, or a combination of them from the past), and these states are then used in the forward model to predict streamflow. Experiments on several catchments from the National Weather Service North Central River Forecast Center show that FHNN outperforms standard baselines, including physics-based models and transformer-based approaches. The model demonstrates particular effectiveness in catchments with low runoff ratios and colder climates. We further validate FHNN on the CAMELS (Catchment Attributes and MEteorology for Large-sample Studies), which is a widely used continental-scale hydrology benchmark dataset, confirming consistent performance improvements for 1–7 day streamflow forecasts across diverse hydrological conditions. Additionally, we show that FHNN can maintain accuracy even with limited training data through effective pre-training strategies and training global models.

Biochemical processes within a two-stage agricultural drainage ditch in Mower County, MN: Methods for estimating nitrogen removal rates and efficiencies

Agricultural Water Management · 2025-07-28 · 1 citations

Drainage ditch design has historically focused on providing adequate water conveyance. More recently, greater attention has been placed on alternative designs that assimilate excess nutrients. However, due to the hydrologic complexities, accurately calculating nutrient removal in these systems presents a challenge. In 2009, 1.89 km of a conventional drainage ditch in Mower County, MN, was converted to a two-stage design. The objective was to evaluate three different methods for calculating nitrogen removal in this system. Continuous data from the growing season of 2010 was used to produce average monthly removal efficiencies by comparing influent and effluent concentrations, which ranged from 19.5 % in May to 12.9 % in September. Three dates were used in 2013, 2011, and 2010 to produce mass-balance relationships for in-channel denitrification using isotopic tracers. Removal efficiencies were estimated at 21 % (2013), 32 % (2011), and ∼20 % (2010) using the mass-balance approach. Nitrous oxide production was measured from soil samples taken for one date in 2013 to estimate potential soil denitrification using the acetylene inhabitation assay, which varied greatly among habitat zones from 0.08 to 1.85 µg N 2 O-N g DW −1 g h −1 . Potential habitat-weighted soil denitrification ranged from 19 % to 42 % compared to 1–3 % estimated for a hypothetical conventional drainage ditch. Although denitrification rates and removal efficiencies are difficult to quantify, comparing the results from multiple dates using an array of methods can validate and add robustness to studies of two-stage ditch denitrification, providing further support for alternative drainage ditch designs as an effective method for addressing nutrient pollution to our natural waterways. • A conventional drainage ditch was converted to a two-stage design in MN, USA. • Three different methods were used to calculate nitrogen removal efficiencies. • Nitrogen removal efficiencies ranged between ∼13 and 42 %. • Removal efficiency for a conventional drainage ditch was estimated at 1 – 3 %. • All removal methods produced results similar to one another and the literature.

Knowledge-Guided Machine Learning for Operational Flood Forecasting

2025-05-21

We present a knowledge-guided machine learning framework for operational hydrologic forecasting at the catchment scale. Our approach, a Factorized Hierarchical Neural Network (FHNN), has two main components: inverse and forward models. The inverse model uses observed precipitation, temperature, and streamflow data to generate a representation of the current underlying catchment state. The forward model predicts streamflow using the learned catchment state. The FHNN architecture is designed to model multi-scale processes and capture their interactions, a critical ability for flood modeling. FHNN also improves forecasts based on real-time data through an inference-based data integration approach using inverse modeling. FHNN’s data integration approach improves forecasts in response to observed data more efficiently than data assimilation methods that require computationally intensive optimization. To show the performance of FHNN, we compare the FHNN to a leading deep learning alternative (autoregressive LSTM) on the large-sample CAMELS-US dataset, and operational flood forecast data from the US National Weather Service (NWS). Official NWS flood forecasts are generated by expert human forecasters using a physics-based model, in a human-in-the-loop process. Thus, we assess the flood forecast ability of FHNN by directly comparing its performance against these NWS expert-derived forecasts. The human forecaster creates a more accurate forecast within the first 18 hours of a forecast’s issuance, but FHNN has significantly better predictions thereafter. This research lays the groundwork for leveraging the predictive performance of AI-based models with the expertise in forecasting agencies to produce better river forecasts.

Biochemical Processes within a Two-Stage Agricultural Drainage Ditch in Mower County, Mn: Methods for Estimating Nitrogen Removal Rates and Efficiencies

SSRN Electronic Journal · 2025-01-01

Machine learning applications in vadose zone hydrology: A review

Vadose Zone Journal · 2024-06-14 · 13 citations

Abstract Machine learning (ML) has been broadly applied for vadose zone applications in recent years. This article provides a comprehensive review of such developments. ML applications for variables corresponding to different complex vadose zone processes are summarized mostly in a prediction context. By analyzing and assessing these applications, we discovered extensive usages of classic ML models with relatively limited applications of deep learning (DL) approaches in general. We also recognized a lack of benchmark datasets for soil property research as well as limited integration of physics‐based vadose zone principles into the ML approaches. To facilitate this interdisciplinary research of ML in vadose zone characterization and processes, a paradigm of knowledge‐guided machine learning is suggested along with other data‐driven and ML model‐based research suggestions to advance future research.

Biochemical Processes within a Two-Stage Agricultural Drainage Ditch in Mower County, Mn: Methods for Estimating Nitrogen Removal Rates and Efficiencies

SSRN Electronic Journal · 2024-01-01