Maybe because "We propose burying this device beneath the Kerguelen Plateau in the Southern Ocean, 3-5 km into the basalt-rich seafloor and 6-8 km below the water’s surface." would be prohibitively expensive for hundreds of nukes.
The nukes just laying around in stockpiles are mostly in the 100-800kT range. You could use nukes that would otherwise need refurbishment (exploding old stockpiles and producing new stock instead of refurbishing old stock), or maybe even spin it as a disarmament treaty where the method of disposal are underground explosions for carbon sequestration. Or use it as an opportunity to use old nukes for a good purpose when you want to switch to a newer model (instead of needing facilities to disassemble them)
The point would be experimentation. You do a small one, measure the results, and if it aligns with theoretical predictions, you scale up. It transforms it from high risk and high reward, into low risk high reward.