The size of the dish would suggest it’s 23 GHz. Those links are usually less than one watt.
Links running at 18 GHz and higher typically have slip-mounted radios directly attached to the back of the dish and avoid flexible jumpers from the outdoor unit (ODU) to the antenna due to high losses. This antenna is also fed by waveguide, suggesting operation somewhere in between 6 and 8 GHz. The only time I see operators running waveguide above 8 GHz is in the odd 11 GHz link run by a utility where they want to keep the RF unit indoors for maintenance purposes. Telcos generally don't care about availability as much as they do throughput. Waveguide loss at >11 GHz is significant and is generally avoided by placing the ODU adjacent to the dish and using a 2m jumper.
This antenna looks to be at least 4-feet, possibly 6-feet, high-performance dish, based on the relative size to the man-lift, the side shroud, and the single tie-back strut. As soon as I hit 6 feet, I use dual tie-back struts to stabilize from wind buffeting because you can see the signal levels fluctuate once the antenna hits a resonant mechanical frequency due to the wind. Reflectors for 18+ GHz have a different rear profile and taper at the back to mount the ODU.
TX power is going to be on the order of +19 dBm for 11 GHz, up to +33 for 6 to 8 GHz, into a +40 dBi antenna with somewhere between 3 and 6 dB of waveguide and connector losses. The radiation patterns due to the side shrouds and the general front-to-back isolation ratio performance of these dishes are sufficient to allow re-use of the same channel for about 60 degrees on either back side of the center line. The squirrel is only going to get truly cooked if it is inside the diameter of the shroud or if the shroud and the interfaces to the reflector are leaking.
Everyone else's mileage may vary.
