Abstract

Despite ET's critical importance in the hydrological cycle, very little has been done to provide ET forecasts at a national level. This study comprises two research chapters examining the National Water Model's (NWM) evapotranspiration (ET) simulations. In Chapter II, we conducted the first comprehensive NWM ET model performance assessment by comparing simulations to eddy-covariance flux tower measurements across the Continental United States (CONUS) on a high-resolution, 1 KM square grid. We clustered results by National Weather Service (NWS) River Forecast Centers (RFCs), land cover classifications, elevation bands, and Köppen-Geiger Climate Zones. The NWM performs best in the Northeast and Ohio RFC regions, humid and wet regions, forested ecosystems, winter season, and mid-elevation ranges. General positive bias is seen in all RFCs except the California Nevada RFC and regions with significant anthropogenic interactions. Our evaluation of the AORC V1.1 temperature forcings found no significant bias that could impact the ET outputs. In Chapter III, we addressed the biases identified in Chapter II by implementing Machine Learning (ML) techniques to post-process the NWM accumulated evapotranspiration (ACCET) Medium Range Forecasts in California. We applied Random Forest (RF), Extreme Gradient Boosting, and Light Gradient Boosting Machine to post-process 1-day to 7-day forecasts using both NWM forcings and additional features capturing spatial and temporal dynamics. All three ML models effectively corrected bias, with RF as the top performer reducing the initial Normalized Root Mean Squared Error from 2.64–2.46 to 0.26–0.37 across all lead times. The most significant features were wind component and radiation variables from NWM forcings, hours of the year fraction, and lead time. This bias correction approach can be extended across the US, enabling NWM use in diverse hydrological applications beyond its primary streamflow monitoring function. Overall, our findings suggest the need for more critical examination and precise use of meteorological forcings beyond temperature, as well as refining the ET outputs using ML post-processing in the forthcoming NextGen framework-based NWM.

Degree

MS

College and Department

Ira A. Fulton College of Engineering; Civil and Environmental Engineering

Rights

https://lib.byu.edu/about/copyright/

Date Submitted

2025-03-31

Document Type

Thesis

Handle

http://hdl.lib.byu.edu/1877/etd13530

Keywords

Evapotranspiration, National Water Model, Machine Learning, Retrospective, Forecast

Language

english

Download

Included in

Engineering Commons

Share

COinS