To verify the quality of my weather forecast I usually compare my prediction with measurements. In my case this location is in the vicinity of Hannover in Northern Germany. For March 20, 2015 I predicted a maximum temperature of about 12°C. To my horror the measured temperature was only 9°C. My forecast is usually pretty good and we are talking about a forecast for the same day, which is not a very long forecast. Another large online weather service predicted a maximum temperature of even 13°C for the above mentioned region. What went wrong?
On the day, that we are talking about, namely March 20, 2015, we had a partial solar eclipse in Germany. It started at about 09:30 CET and ended up at 12:00. The weather on this day was cloudless with low winds, which means that the air temperature near the ground is highly influenced by the solar radiation. But of course! That's it! The solar eclipse! The numerical weather prediction model did not consider this phenomenon. The moon covered a significant part of the solar disk. In Hannover the part of this coverage increased up to 77.5 %. Thus the solar radiation was weakened and the air near the ground could not warm up to the predicted values.
The corona, that surrounds the sun, is often mentioned in conjunction with solar eclipses. To observe the corona you need a total solar eclipse and not a partial one as in this particular case. The next total solar eclipse is on August 21, 2017 in the United States.