Possible/probable (almost certain? unless you corrected for it) confounding: as wind speed will correlate (causatively, but the wind isn't just local) to temperature gradients, you will find correlation to time of year and time of day, both of which correlate highly to light and temperature which correlate to power demand.

Also, as I always note: you should include credit to yourself, date of graph creation, and cited data sources, in text on the image itself, since jerks like to copy reddit images everywhere without backlinking.