>Those periods of near zero production can last non-trivial amounts of time no realistic battery could provide for
There are many options (electrochemical, gravity etc) for grid-scale energy storage of various durations that are commercially-viable today. 0-8 hours tends to be the most needed duration in the US but there are solutions that can store power for days or weeks without losses. Even in a 100% nuclear or gas-fired world storage is critical. During the Texas blackout both nuclear and gas assets failed due to the cold temperatures.
Besides the fact that electricity generation is merely one aspect of GHG emissions (hello agriculture, land use, transportation, etc). The ultimate argument against the "let's not radically change our ways and only retool the engine" view is the following: we have ~120 years of known reserves for the current fission reactors, with a potential 10x multiplier with fast neutron reactors should anybody (other than the russians and the chinese) bother to invest into researching the field. How many more doublings in our economy (and hence energy use) does it take to consume it all ? At a measly (by economists and politicians standards) rate of 2%/year, you get a doubling every 35 years. in 350 years (yeah I know this sounds very far in the light of the current news) we'd have increased our demand by 3 orders of magnitude...
There are many options (electrochemical, gravity etc) for grid-scale energy storage of various durations that are commercially-viable today. 0-8 hours tends to be the most needed duration in the US but there are solutions that can store power for days or weeks without losses. Even in a 100% nuclear or gas-fired world storage is critical. During the Texas blackout both nuclear and gas assets failed due to the cold temperatures.