Adapting the Iowa Flood Center’s Hillslope Link Model for compatibility with the NextGen Framework

Objective:

The goal of the project is to adapt the Iowa flood center Hillslope link model (HLM) to make it compatible with the Nextgen Framework

Approach:

The hillslope link model (HLM) is a conceptual distributed hydrologic model developed by the Iowa Flood Center, designed with focusing on prediction of high flow events. The model was originally intended to simulate the flow conditions for the typical agricultural Midwest; however, the model has proven its skills that simulate water quantity in various climate conditions because of its flexible modular structure. The modules simulate snow accumulation and melting, infiltration, percolation, overland flow, interflow, groundwater flow and open channel flow. The flexible structure allows simulating processes with different levels of complexity, e.g., from simple constant infiltration schemes to more complex modeling. The HLM has been used for real-time streamflow prediction in Iowa and has proved its operational capability and reliability . In an operational mode, the model simulates and updates the most current state of water stored in the basins and channels, as new forcing data (i.e., precipitation, evapotranspiration and temperature) become available. The most current state is used as initial condition to produce streamflow forecasts for the next 10 days using quantitative precipitation forecasts. The model has shown its skill predicting high flows due to snowmelt events during winter, intense rainfall events during the summer; and fast hydrologic response due to intensive agricultural tile drainage.

The goal of the project is to enable the HLM to be compatible with the NextGen modeling framework. The HLM codes will be refactored to adopt the BMI standards that will allow communication with NextGen Framework. For this purpose, the legacy HLM codes written in C will be migrated to the modules in Python. The HLM equations will be rewritten in generic modules. The modules will focus only on rainfall-to-runoff conversion processes. All BMI functions will be implemented to create an interface that allows forcing data inputs into the model. The interface will also provide a bridge for modifying model parameters and states, offering flexibility for model calibration and data assimilation, by replacing states variables (e.g., discharge, soil moisture). Unit tests will be developed to evaluate the implementation of each BMI interface method. To show the operational benefit for River Forecast Centers we will conduct tests the new development using each one medium size basin in Iowa and Missouri, respectively. We will explore how changes in the landscape and climate conditions, reflect in the output of the new model, and compare to simulations using legacy codes.

Impact:

Abstract:

The goal of the project is to adapt the Iowa flood center Hillslope link model (HLM) to make it compatible with the Nextgen Framework. HLM is a distributed hydrologic model where hillslopes and channels are the spatial units where runoff generation and transport processes take place. Making HLM available in Nextgen suite will ease the struggle of software installation, and preparing customized inputs through the use of Basic Model Interface in the Framework. HLM has been used extensively to provide operational forecasts in the state of Iowa and other areas of the US Midwest. Bringing HLM into NextGen will facilitate that the model can be used by more members of the hydrologic community, and to continue using it as a tool to address questions regarding scaling of peak flows, flood frequency analysis, quality of precipitation estimates, among others. This project has a direct impacts in the operations of North Central and Missouri River Forecasting Centers, who are in charge of providing forecasts for the rivers in Iowa, encompassed between the Missouri and Mississippi rivers. The availability of HLM in Nextgen allows to operational forecasters to use interchangeable modules of runoff routing and transport, and find suitable configurations to produce reliable forecasts for these basins. This idea can be extended not only to basins in Iowa, but over any river basin in the US where the processes represented in HLM are relevant. The project will address challenges in the model implementation and integration, and also in the transition from using hillslopes and channels to hydrofabric units.