Among the worst is the Birkeland-Eyde process, hot plasma by electric arcing, like it happens naturally during a thunderstorm. Even microwaving the air is more energy efficient. And amongst the most promising are transient non-equilibirum / cold plasma approaches using ns pulse generators.
As far as I understand, the reasons for using methane as the hydrogen source in industrial processes is because of costs and the difficulty in scaling electrolysis industrially.
So this is pretty far outside of my area of expertise, but https://pmc.ncbi.nlm.nih.gov/articles/PMC8133363/ suggests that there may be an evolution of Birkeland-Eyde that isn’t nearly as bad as the original process was. I haven’t gone through this in detail and even if I had I probably don’t have the background to do it justice.
Plasma to take N2 + O2 -> NO2 and NO gas
The NO and NO2 are dissolved in water to become HNO2 and HNO3, which then become NO2- and NO3- ions in the water.
It’s a low pH solution which stabilizes it and then you dilute right before application so that you don’t burn the crops.