Authors: Daniel Philippus, Claudia R. Corona, Terri S. Hogue – Colorado School of Mines
Title: Daily Stream Temperature Predictions for Ungaged Watersheds with NextGen
Presentation Type: Lightning Talk
Abstract: Stream water temperature (SWT) affects processes ranging from water treatment to fish spawning, which has motivated SWT modeling development in recent years. Large-domain, high-resolution models support evaluation of stream thermal conditions across a range of scales, particularly for ungaged sites. Recently, 1 km/daily-resolution, ungaged SWT models have been developed for the contiguous United States (CONUS). However, these models do not support forecasting, and lack Basic Model Interface (BMI) implementations for NextGen use.
We present the first NextGen-compatible, CONUS-scale, 1 km/daily-resolution, forecasting-capable SWT model for ungaged watersheds: TempEst-NEXT (temperature estimation: near-term expected temperatures). TempEst-NEXT supports hindcasting and 1-16 day forecasting, and has full BMI support, with integration into a real-time NextGen system in progress. Historical tests for mean daily temperature show a median R2 of 0.93, RMSE of 2.1 C, and bias of 0.88% for ungaged watersheds, with similar performance in preliminary real-time forecasts. TempEst-NEXT’s ungaged forecasting capabilities can support real-time decision making without local calibration for water-resource management, such as reservoir releases, particularly when integrated with operational systems running on NextGen. Broadly, TempEst-NEXT is available to support efficient, large-scale prediction of stream thermal regimes, especially in watersheds where SWT observations are unavailable.