It's possible to use "dry cooling" for a nuclear reactor. [1][2] The thermodynamic efficiency would be much lower, but it is possible to use alternative thermodynamic cycles/methods, for example reactors sent into space have used radiative cooling since there is no realistic way to use fluids to finally reject the heat.
So basically radiators. They would have to be huge, and they would have to move huge amounts of air through them. As a backup that might work, probably using a hybrid where a water body is used for cooling and air is the fallback.