Thanks for the link -- I hadn't seen that. My issue is that water lost faster than delivered by comets is bad for tectonics, which is bad for atmospheric CO2. They'd have to show no net loss of water, or else the system is not stable over Gyr timescales.