Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

Home » Recent Research Supports NOAA Artificial Intelligence Strategy and CPO Water Resources Climate Risk Area

Recent Research Supports NOAA Artificial Intelligence Strategy and CPO Water Resources Climate Risk Area


Emerging machine learning technologies are key to improving the quality of predictions in complex and rapidly changing Earth systems, as outlined in the NOAA Artificial Intelligence Strategy. Two recent studies, partially funded by the Climate Program Office’s (CPO) Climate Observations and Monitoring program, use machine learning techniques to support research on water in the western U.S as part of the Climate Observation and Monitoring program’s larger initiative focused on datasets and analyses for improved representation of surface-atmosphere processes in models. These studies contribute to the growing pool of research aligned to CPO’s water resources climate risk area.

One paper, published in the Journal of Geophysical Research: Atmospheres, used water, carbon, and energy flux data that was up-scaled by machine learning methods to represent a larger portion of the western US that frequently experiences drought. The University of Arizona researchers found that this method was helpful to determine that annual carbon uptake by plants is highly dependent on water availability and regional drought conditions, highlighting important interactions between ecosystems and hydrology. Another paper, published in Water Resources Research, used a hybrid approach to characterize streamflow in a watershed in northern Utah in which a high resolution model represents snowmelt, while deep learning methods represent the surface and subsurface properties. This region has trouble predicting streamflow because of variable snow melt and heterogeneous pores and conduits in the underlying geology. The researchers, from Arizona State University, Utah State University, and the Jordan Valley Water Conservancy District, produced model results with a higher accuracy than previous studies. 

The results from these studies provide evidence that combining traditional observations and models with machine learning techniques is a valuable way to realistically characterize water conditions in the western U.S. This research can help to improve the quality and timeliness of predictions used for water resource management in the west.


Read the article in the Journal of Geophysical Research: Atmospheres »

Read the article in Water Resources Research »

For more information, contact Clara Deck.

More News

Scroll to Top