Some of the bright x-ray sources in the sky are neutron stars that are accreting gas from much larger companions. The gas, consisting of Hydrogen and Helium, accumulates in an ocean that covers the surface. As accretion proceeds, the ocean becomes more massive and the Hydrogen and Helium at the base of the ocean become increasingly compressed. Once the density at the base reaches about 1 million grams per cubic centimeter (roughly the same as the density at the center of a white dwarf), the ocean ignites and burns up. This produces a brilliant flare in x-rays and other bands, known as a "Type-I x-ray burst". Type-I x-ray bursts are routinely observed.
The rises of the bursts are typically quite short, less than a second. This means that the nuclear flames, once ignited at some spot on a neutron star surface, propagate around the neutron star on a timescale less than a second. The flame speeds should be impressively high, between 10 and 100 km/sec! Together with Anatoly Spitkovsky and Greg Ushomirsky, I made a crack at understanding such high flame speeds and their geometry in 2001-2. Anatoly was a PhD student at Berkeley, working on PIC simulations that would one day make him famous in the Astrophysics world. He nonetheless found the time to take a lead on this project. We worked on it closely for about a year. None of us knew any geophysical fluid dynamics, but we had to learn it rather well to address this problem. The strongly accreting neutron stars are incredibly fast rotators, and the Coriolis force plays a huge role. The flames are essentially nuclear hurricanes!
The SLU paper only used two vertical zones, and left many unanswered questions about the details. Many of these were addressed in Yuri Cavecchi's PhD thesis which was co-supervised by Anna Watts and myself. The real engine for this research was the magneto-hydrodynamical code that was developed by Jonathan (Jon) Braithwaite. This code enforced hydrostatic equilibrium in the vertical direction, thus avoiding unnecessary and costly tracking of vertical sound waves. This allowed Yuri to carry out vertical multizone simulations of propagating flames that clarified many aspects of this process.