As "thermal" power plants (Rankine cycle) they would have to obey the laws of thermodynamics which require that they reject heat while producing power. Thermodynamics sets an absolute upper limit to their thermal efficiency of about 47% (the actual number being the 1 minus the ratio of condenser to boiler temperatures measured in absolute units). That is, they must dissipate about half of the heat they collect to the environment in order to produce any output power. The conventional methods used to meet this requirement are flow through systems using lake or river water or cooling towers to condense the steam exiting the turbines. A demand for cooling water will certainly impact the site planning for these power plants since deserts are deserts because water is not widely available.
If the designers try an end run on thermodynamics and specify an air cooled radiator it would have to be very large to dissipate half of the heat flux passing through the plant. Air cooling would also drop the thermal efficiency of the plant because it would unavoidably raise the condenser temperature to near the boiling point of water.
Regards,
GtG
There is also the fairly obvious fact that air-cooling doesn’t work nearly as well where the temperature is really high.
OTOH, humidity is very low and temp high, so evaporative cooling is more efficient. But that gets you back to needing lots of water.