Swamp coolers get less efficient, they don't stop working altogether.
The metal bench doesn't get continuously cooler; at some point the delta-T between the bench and the air is warming the bench at the same rate that the evaporative cooling is cooling it, and an equilibrium is reached. But it will reach a point cooler than the ambient air, depending on the wind and humidity. Even when it's raining out, wind causes evaporation that causes cooling.
Hang one thermometer outside on a rainy day under the eaves to measure the air temperature. Put another in a bowl of water (that's at ambient temperature) with a fan blowing across it. The water will cool down and the thermometer will register cooler than the first.
Right. That's what I've said, no? Unless you mean to take me to task for "noticeably ineffective" being too close to "stop working altogether".
You're describing the difference between "wet bulb" and "dry bulb" temperature. I'm familiar with it. One measures wet bulb by actually with a sling psychrometer: literally a thermometer in a wet sock that is swung on a string around in the air.
The wet bulb temp will be colder than dry bulb, but it would not continue to get colder. It would reach some differential, and then stay there.
That differential is directly proportional to the relative humidty.
At 50d F, and 74% humidity, the wet bulb temp would be 4d below (thus 46d F).
On a 100% humidity day, that difference would be zero. There would be no
evaporation (unless by means of radiant heat, and even so it would be a feedback loop).
edit: looking into it, game-time humidity is reported at ~74% (give or take, because 74% is convenient in the chart I referenced).
So the wet
balls temp could have been as much 4d F below ambient air.