I saved this old comment from Mike because it was an eye opener for me and addresses something most people ignore, that being chronically warm temperatures:
Allow me to quote Mike:
"Fish do not respond to thermal averages; they respond to thermal maxima and minima. They also respond to what are, in effect degree-days or other time-based thermal units. Warm temperature caused mortality can be chronic or acute and the chronic occurs at temperatures that are frequently viewed as being sub-lethal by laymen.
With respect to lethal water temperatures it is important to not focus on just one form of lethality, that being thermal maxima. There is a whole other form of lethal temperatures:
Chronically warm temperatures:
Ignoring Brook Trout, that problem begins around 68° F; 68.1 F° for hatchery Rainbow Trout (RT) and I expect probably a very similar number for Brown Trout. At that point, the percentage of the RT that die throughout a late spring and summer due to temperature stress alone is directly related to the number of 15-minute periods per month that the water temperature exceeds 68° F.
It does not matter if the temperature drops below 68° each night because the tally continues the next day once the temperature rises above 68 again. The 15 min periods above 68° accumulate daily through the end of the month.
In addition, when fish are in thermally stressful conditions or in thermal refuge, they are more vulnerable to predation such as by great blue herons, which are effective predators in daylight and at night."