I'm not following the argument for being able to estimate deaths per [T]Wh for hydro, gas, etc. but not nuclear. I think hydroelectric is especially analogous