How are total solar eclipses used by scientists?

Of course the most spectacular use has been to study the faint corona of the sun, which can be observed by spacecraft such as the Solar and Heliospheric Observatory (SOHO) by making artificial eclipses, but ground-based telescopes and photography have also made many historical contributions to understanding the shape, structure and extent of the corona. Also, total solar eclipses have been invaluable in improving our understanding of the lunar orbit. Whether a total solar eclipse occurs at a specific location and time on the surface of Earth depends on the lunar orbit, the motion of the moon along the orbit, the earth-moon distance and other factors. Sophisticated physics-based computer models have been used for over a century to make accurate predictions of each eclipse to the second, and to the nearest mile on Earth. The best way to do this is to look at historical sightings of total solar eclipses from centuries or even millennia in the past.  These sightings are often made by observers at specific geographic locations and who indicate the time of the eclipse from that location. These distant-in-time observations can be calculated by the modern eclipse models and compared with the historical sighting, then the models can be adjusted by improving the parameters of the physics calculation until agreement is reached. This process sometimes results in new ‘science’ related to the shape of the moon, or gravitational perturbations in the lunar orbit that can take centuries to build up to measurable effects. For example, in 1989 an astronomical event recorded on a clay tablet found in 1948 among the ruins of the ancient city of Ugarit, Syria, was identified as a description of a total solar eclipse that occurred on 3 May 1375 BCE. The information was used to  provide a reference point to establish the long-term evolution of angular momentum in the Earth-Moon system