There's lots of speculation, but it sounds like the mppt controller has issues. You need to measure solar current and voltage as well as battery charging current and voltage. How many watts are they apart?
A good mppt controller has better than 95% efficiency at max current.
Regarding the temperature effect: Even in spring time the panel should still get quite hot. Look at the numbers, solar is only 17& efficient. Most of the remaining 83% goes to heating the panel. If you have a meter squared sized panel with 1000W/m^2 irradiance, then 0.83 * 1000W = 830 W goes to heat. Solar is a huge heat collector.
If you have 1000W irradiance then the panel temperature will be greater than 25C and the panel wattage will be lower. Isc will remain about the same.
You really need to look at the NOTC spec, not the STC spec. For example the 140W Kyrocera panel has 140W (STC) and 101W (NOTC). Granted NOTC is at 800W/m^2, but panel temperature is allowed to go up.