How much is COVID-19 impacting forecast model accuracy?
Fewer aircraft observations may be impacting forecast model accuracy
Go Deeper.
Create an account or log in to save stories.
Like this?
Thanks for liking this story! We have added it to a list of your favorite stories.
Chalk up one more impact of COVID-19. Less air travel means less weather data feeding into weather forecast models.
The European Centre for Medium-Range Weather Forecasts (ECMWF) first flagged the reduction in aircraft weather data in March.
Between March 3 and March 23 March there was a reduction of 65 percent in weather reports received over Europe. Globally the reduction was about 42 percent.
Garbage in, garbage out?
Less complete weather data into a weather forecast model will impact accuracy as you might expect. The ECMWF estimates a 15 percent reduction in model forecast accuracy in wind and temperature forecasts at flight levels if all aircraft data is removed.
Turn Up Your Support
MPR News helps you turn down the noise and build shared understanding. Turn up your support for this public resource and keep trusted journalism accessible to all.
Sensitivity studies at ECMWF have shown that removing all aircraft data degrades the short-range wind and temperature forecasts at those levels by up to 15%, with significant degradations at all forecast ranges up to seven days. There is a smaller, but still statistically significant, impact on near-surface fields, up to 3% on surface pressure.
Surface weather parameters impacted?
A July study in the American Geophysical Union journal Geophysical Research Letters suggests a measurable reduction in forecast accuracy due to flight reductions during COVID-19.
In this study, we verify global weather forecasts in March to May 2020 against high‐resolution global reanalysis data set and an observation‐based global precipitation data set, which are the best estimate of the atmospheric state. To investigate the forecast deterioration during the pandemic, the forecast accuracy during March to May 2020 is further compared with the average accuracy during March to May 2017–2019.
We report a significant deterioration in the forecasts of surface temperature, RH, wind speed, and pressure, but no significant deterioration in precipitation forecast is observed. A similar analysis for February 2020 suggests that the forecast accuracy of surface meteorology could have been expected to improve in 2020 compared with 2017–2019, if aircraft observations were carried out as usual.
It should be noted that the study methodology has been questioned by several meteorologists. There are other variables that may be at play in this complex situation.
One area to watch for model forecast differences is in the tropics. The lack of weather data could be causing some big forecast model differences there. There have been some interesting inconsistencies between models lately in the tropical Atlantic.
It is likely that the reduction in flight-level weather data feeding into forecast models is reducing accuracy somewhat. It’s just difficult to pinpoint the magnitude and spatial distribution of the effects on forecast accuracy.
Stay tuned.