The first problem is that climate models are not designed to function properly in conditions far removed from current earth conditions. They use thousands of empirically derived parameters, and by reducing CO2 to zero they probably made the output of the model even more meaningless than the usual output.
The next problem can be highlighted by performing a related thought experiment. Instead of zeroing out CO2, let’s “zero” out temperature. We are going to make the Earth cold – with temperatures at the equator being 0°C. At that temperature, the equilibrium vapour pressure of water is 6 millibars, which is 25% of it’s value at 20C.
This means that the amount of water vapour in the air is about one fourth of current tropical conditions. Let’s use 0.5% of the atmosphere as a conservative value. That is 50,000 ppm – which is more than 100 times greater than the amount of CO2 in the atmosphere.
Even on a frozen planet, H2O would be much more abundant than CO2, and H2O also absorbs a much wider spectrum of LW radiation.
Conclusion – on NASA’s hypothetical frozen planet, H2O would still be (far and away) the dominant greenhouse gas. Claims that CO2 is the Earth’s thermostat don’t work empirically or theoretically. Their assumption that H2O is completely condensable, is bogus. It’s ability to condense tails off logarithmically at lower temperatures, where H2O behaves more and more like “non-condensable” CO2.