DOS- Prairie Pothole Modeling - Improving Short-term and Seasonal Hydrologic Forecasting in the Souris River Basin

The International Joint Commission (IJC) is a binational organization that was established in 1909 via the Boundary Waters Treaty to prevent and resolve disputes over transboundary water issues between the United States and Canada. The IJC accomplishes its work through bi-national transboundary Boards composed of equal number of members from both countries. One of the Boards is the International Souris River Watershed Board. The International Souris River Basin (ISRB) is in the prairie pothole region of Canada and the United States. The area has unique hydrology, with streamflow and reservoir inflow affected by both rainfall and snowmelt. Given that climate change is altering the partitioning between rainfall and snowfall, it is expected that these changes will also markedly impact the hydrologic responses in the SRB. To provide robust forecasts under changing climate, it is crucial that hydrologic models appropriately represent the dominant processes driving the hydrologic response in the SRB. Notably, streamflow response in the basin is heavily shaped by the filling and spilling of prairie potholes, and the associated dynamic changes in the contributing area. The goal of this project is to evaluate candidate - fill and spill representations for use within the Sacramento Soil Moisture Accounting Model for incorporation into the Next Generation Water Resources Modeling Framework. The project would assist the responsible flow forecasting agencies in supporting the ISRB by taking steps to improve their river forecasting tools. The purpose of this exercise is to gain insights into the specific time periods and locations where the inclusion of prairie pothole hydrology enhances or constrains predictive capabilities. The first element of this work is to assess the candidacy of prairie pothole hydrology algorithms to represent the fill and spill phenomenon. The second element involves incorporation of one or more recommended prairie pothole hydrology algorithms into the Sacramento model within NOAA's Next Generation Water Resources Modeling Framework. The third element will test these developments in the NextGen Framework. The four elements are described as tasks in the next section. Project Tasks and deliverables Task 1: Evaluate Candidate Prairie Pothole Hydrology Algorithms Numerous algorithms have been developed over the years to address the effects of modified storage, runoff generation, surface-groundwater interactions, and evapotranspiration resulting from the presence of wetlands in the prairie pothole landscape. In this project, the contactors will consider at least three algorithms: the widely utilized Probability Distribution Model based RunOff generation (PDMROF) method, the Hysteric Gatekeeping Depression Model (HGDM), and a novel approach that learns the dynamics of wetland inundation directly from the remote sensing data, to capture wetland inundation's influence on runoff and evapotranspiration. The contractor shall use Algorithm-1 or the PDMROF method to employ a two-parameter Pareto distribution function to depict the relationship between ponding water and critical storage capacity. By utilizing a probabilistic framework, this algorithm assesses changes in storage and inundation extent over time. The contractor shall use Algorithm-2 or the HGDM method to follow the work of Shook et al. (2021) and consider a model agnostic formulation. The contractor shall use Algorithm-3 to establish a global remote-sensing derived monthly dataset of wetland inundation with a spatial resolution of 30 meters (Pekel et al., 2016, Nature) to establish a relation between modeled catchment water storage (or surface water storage) and inundation area for each functional unit of the model, such as a catchment. Given that past studies (e.g., Shook et al., 2013, HP) have reported a hysteretic relation between the two, the contractor shall establish a relation that will explicitly consider wetting and drying seasonal cycles in the assessment. Initially, the contractor shall apply the algorithm so it will begin with a predefined relation, which will be subsequently optimized iteratively to ensure that the model simulations of hydrologic variables align with observed data. It is important for the contractor to use this methodology since it will explicitly incorporate the spatial variations in the relationship between storage and inundation area across the landscape due to variations in topographic and hydrogeological properties. The contractor shall evaluate the three algorithms and will consider performance, input requirements, and potential for implementation in operations. The results of the contractor's evaluation of the candidate algorithms will be reviewed with IJC staff, ISRB, and local experts in forecasting agencies to identify the recommended algorithm(s) to promote for further testing in Task 3. Task 2: Incorporate Prairie Pothole Hydrology into NOAA's NextGen Framework The one (or more) algorithm(s) recommended in this Task will be integrated into the Sacramento Model and incorporated into the NOAA's Next Generation (Nextgen) Water Resources Modeling Framework. The Sacramento Model is widely used by U.S. regional river forecasting centers (RFCs) to generate streamflow forecasts. One of the model's mechanisms for generating streamflow involves direct runoff from variable impervious areas, including inundated areas. The dynamic wetland inundation extent, obtained through the prairie pothole hydrology algorithm, will directly impact the runoff from each functional unit of the model. Furthermore, the new algorithm will also affect simulations of other dynamic storage components that influence evapotranspiration, groundwater recharge, subsurface storage, and baseflow. To integrate the modified model into the Nextgen Framework, the contractor shall develop appropriate Basic Model Interface (BMI) wrappers. These extensions to the Sacramento Model and the Nextgen Framework will produce a feasible tool for hydrologic predictions and forecasting in basins dominated by prairie pothole hydrology. Task 3: Test the Sacramento Model in NextGen Framework The assessment of the benefits associated with incorporating dynamic wetland inundation and storage, as enabled by the incorporated algorithm, will involve a comparison by the contractor of calibrated model results versus observed streamflow (e.g., stream gages, naturalized streamflow), inundation extent (e.g., satellite imagery), and evapotranspiration (e.g., satellite estimates) within the basin for a defined set of test cases. The contractor shall conduct these evaluations using standard metrics (e.g., Nash-Sutcliffe efficiency (NSE), Kling-Gupta efficiency (KGE)) across various spatial and temporal scales, encompassing different hydroclimatic conditions. In addition to the standard model performance evaluation metrics used in hydrology (NSE, KGE), the contractor may also use new metrics being developed. Additionally, computational benchmarking of the models will be conducted by the contractor to ascertain the additional computational requirements that arise from implementing the modified models.