Bug in weight for geos chem analysis increments

[jeromebarre]

https://github.com/GEOS-ESM/geos-chem/blob/830985334191d1f2dfb9f3fe274f736dc695a097/Interfaces/GEOS/geos_analysis.F90#L707C3-L707C92

This can lead to a division by 0 if the condition is met...

@viral211

[viral211]

Thanks for catching it, @jeromebarre.

[jeromebarre]

No problems. I am thinking on improving this code down the line. Those AnaL1-4 options can conflict with introp and instrat options, as if one define those coefficient too low and the tropopause of GEOS is high enough this would create a "zig zagging" profile of weights...
Also, we just found that keeping the NO/NO2 ratio can really mess up the UTLS chemistry. I would refrain from using this secondary species increment. Chemistry is non-linear, so trying to keep the NOx ratio doesn't work, especially messing with NO at those altitudes. This assumption might be more stable in the boundary layer but not in the UTLS with the very strong O3 gradient that we can have.