Verification of the WRF-Eclipse Model Using Data From the 14 October 2023 Annular and 8 April 2024 Total Solar Eclipses
Loading...
Authors
Picciano, Genevieve
Date
2025-08
Type
Thesis
Language
en_US
Keywords
Alternative Title
Verification of the WRF Eclipse | 2023 & 2024
Abstract
Solar eclipses are astronomical phenomena that occur when the moon
passes between the Sun and the Earth, casting a shadow on the Earth’s surface
and causing a measurable decrease in solar radiation. In recent years, the
United States experienced two eclipses, on 14 October 2023, when an annular
solar eclipse swept from Oregon to Texas, and again on 8 April 2024, when a
total solar eclipse passed from Texas to Maine. As a part of the NASA-funded
Nationwide Eclipse Ballooning Project, 20 teams were positioned underneath the
paths of annularity and totality to collect surface observations and conduct
radiosonde launches to gather atmospheric profiles starting 24 hours before
maximum obscuration and ending six hours after. The data from six of these sites
(2023: Elko, NV, Moriarty, NM, and Corpus Christi, TX and 2024: Junction, TX,
Bloomington, IN, and Pittsburg, NH) were then compared against pairs of WRFARW
simulations with varying planetary boundary layer (PBL) schemes run with
and without an eclipse-radiation parameterization (WRF-Eclipse). The purpose of
this study was to investigate the accuracy of the WRF-eclipse model at
simulating the effects of these eclipses on 2 m temperature, relative humidity,
and wind speed and direction and on temperatures throughout the height of the
atmosphere.
At the three sites for each eclipse, WRF successfully reproduced the
effects of synoptic-scale patterns and frontal passages but struggled with finescale
processes including cold pool formation in valleys and downslope wind
events. These discrepancies were largely attributed to the smoothed topography
in the model and limitations in resolving small-scale boundary layer variability. At
the surface during the annular eclipse, the WRF-eclipse-on models were unable
to replicate the changes in temperature from first to second and third to fourth
contacts, over cooling these changes by ~2.5°C. The models were found to be
more accurate in regards to the changes in relative humidity and wind speed
during these time periods, with the relative humidity values being within 8% of the
observations and wind speeds being within 1.3 m s-1. During the 2024 total solar
eclipse, the WRF-eclipse-on model runs were able to reproduce the temperature
changes around totality while generally underestimating the changes in relative
humidity and wind speed.
For both eclipses, Mean Absolute Error (MAE) analyses showed that the
greatest variability in temperature between the ensemble members was near the
surface in the surface – 850-hPa layer, where the effects of changes in solar
radiation are most pronounced. The MAE values in the mid- and uppertropospheric
layers remained small and largely unaffected by the use of WRFeclipse.
This study aims to advance our understanding of how WRF, specifically
WRF-Eclipse, simulates the tropospheric response to the decreased solar
radiation during solar eclipses on a fine spatiotemporal scale. This is
accomplished by investigating the accuracy of WRF simulated changes in 2 m
temperature, relative humidity, and wind speed and direction, as well as
temperature profiles up to ~10 hPa.