Experimental model predicts the effect of the 2017 eclipse on weather

Image: Evolution of 2m Temperature from HRRR model.

Downward solar radiation with effects of solar eclipse from HRRR model experiment. Case for 4 August 2017 but simulating eclipse with 21 August solar conditions.

August 15, 2017

On a typical clear August day in the Great Plains, maximum downward shortwave radiation from the sun occurs about noon. But if the sun is blocked partially for a few hours and completely for a few minutes, how will this affect temperature near the surface? Or to low-level winds? What impact will this drop in radiation have on existing weather, such as thunderstorms in progress? Current operational weather forecasting models are not equipped to represent the eclipse and its effect on the weather.

Researchers at the NOAA Global Systems Division (GSD) and the Cooperative Institute for Research in Environmental Sciences (CIRES) have adapted an eclipse algorithm to use with their experimental version of the 3-km High-Resolution Rapid Refresh (HRRR) model. The algorithm, developed by a University of Barcelona team and shared by NCAR with the Weather Research and Forecasting community model, computes the degree of obscuration of the solar disk at each model grid point. Based on this calculation, the model modifies the incoming solar radiation, which impacts the heating of the earth and, subsequently, the weather.

Researchers at GSD and CIRES implemented and tested this algorithm within the HRRR model. They used meteorological conditions from 4 August 2017 but paired those with the anticipated astronomical information from 21 August when the eclipse will occur. The researchers found that the eclipse will primarily affect temperatures: the model predicts a reduction in near-surface temperatures—from 1-4 C cooler in the penumbra region (widespread area over the U.S. that experiences a partial eclipse) to 3-7 C (5-12 degrees F) cooler along the path of the full eclipse maximizing near 1800-1900 UTC (noon-1pm Mountain Time Zone). The model also predicts that incoming solar radiation will be reduced by up to 900 W/m2 during the eclipse (see figure below), but the exact local effect will depend on the local cloud conditions next Monday. NOAA/GSD’s experimental version of the HRRR model (HRRRx) has now been modified to account for the eclipse in its real-time forecasts.

Learn more about the results from GSD’s experiments of the likely effect of the eclipse on August 21.

For real-time HRRRx experimental forecasts, now available and before and during the actual eclipse itself—including the effects of the eclipse starting Saturday night looking ahead 48 hours with the 00 UTC model run—visit https://eclipse2017.noaa.gov. Selected weather fields are available, including downward solar radiation and cloud fields and 2-meter temperature, for HRRRx (with eclipse effect), HRRR-NCEP (without eclipse effects and some other differences) and HRRRx - HRRR-NCEP difference fields. NOAA’s National Weather Service runs the HRRR-NCEP model for operational usage.

The eclipse and state-of-the-art weather models like the HRRR offer a rare opportunity to compare conditions of the atmosphere with and without the eclipse, and see how it interrupts processes in the atmospheric boundary layer from coast-to-coast.


  • NOAA/ESRL/GSD/CIRES Joe Olson, Michael Toy, and Jaymes Kenyon
  • NCAR/MMM - Wei Wang, Jimy Dudhia, Dave Gill Alex Montornes, U. of Barcelona

For more information contact: Susan Cobb 303-497-5093