Advance Land Surface Modeling Coupled with the NexGen NWM and Integrate with NOAA UFS to Improve Hydrologic Forecasting Capability

The operational weather/climate model, i.e., the Unified Forecasting System (UFS), and National Water Model (NWM) have a low predictability of precipitation and floods over the large areas of the Mississippi River due to model deficiencies in representing soil-mediated hydrological processes. This project aims to improve the predictability of streamflow and precipitation by advancing the hydrological schemes in the operational water and weather/climate models in the contiguous US (CONUS). More specifically, we aim to

1. Improve model representations of soil hydrology, surface ponding, groundwater recharge/discharge, and plant hydraulics (in the Noah-MP land surface model, which is used in both NWM and UFS).
2. implement alternative datasets of remotely-sensed of vegetation traits and soil hydraulics parameters (e.g., machine-learning based), bedrock depth, organic matter, and coarse fragments.
3. Test the coupled water model (NWM/Noah-MP) over selected river basins in CONUS and the coupled weather/climate model (UFS/Noah-MP) over global domain with a focus on CONUS.

In the first year, we have

1. Investigated the impacts of frozen ground on infiltration and runoff. We found that higher permeability of frozen soil improves the predictability of streamflow at basin scales due mainly to presence of soil macropore networks that are formed during the freezing-thawing cycles.
2. Developed an explicit preferential flow model (dual-permeability model, DPM) with explicit surface ponding and infiltration-excess runoff to represent rapid flow through the macropore networks. Through extensive testing over CONUS, we found that
   • Explicitly representing surface ponding substantially improve peak flow predictions over the major tributaries of the Mississippi River
   • The van Genuchten soil hydraulics (vs. Brooks-Corey) improves soil moisture predictions.
   • The preferential flow model improves streamflow predictions in most rivers in CONUS.
3. Demonstrated that implementing vegetation dynamics improves soil moisture and streamflow simulations. We have been testing this effect in the CFE+Noah-OWP-Modular configuration of the NextGen Framework.

This is the first time to implement an explicit preferential flow model with explicit surface ponding and infiltration-excess schemes into a land surface model (Noah-MP) for operational weather and water predictions. Coupling the improved hydrological representations with the NextGen NWM and NCEP/UFS and testing the two systems over global domain and selected river basins will impact operational weather, subseasonal-to-seasonal (S2S) climate, and water predictions over CONUS.