I have taken a while to notice a pattern in the weather station readings, namely that on several days the outside temperature reading is elevated over what one might expect from trend lines before and after. It happens around 6 pm at the moment, but looking back into March I can see the effect around 5 pm, so it seems to be linked to the actual time rather than to clock time (I mean by this that it happens at about the same time, 5 pm, if you measure entirely in Standard Time). It doesn’t happen every day, which has obscured the pattern somewhat, but it is consistent otherwise. It also didn’t happen (that I have been able to track) in early March, and seems to have appeared after the equinox. My supposition is that the temperature sensor is being exposed to sunlight and is heating up during the times when the reading is elevated. If true, then the days when it doesn’t happen will be cloudy days. Since I don’t track cloud cover, this is going to be hard to confirm except on weekends when I am in a position to check the sensor. If I do manage to confirm this as the cause I will have to either re-site the sensor or shield it from the sun (I thought I had shielded it in the current position on a north-facing wall, but I may be mistaken).
Later: I had someone check the sensor when the temperature began to rise, and as anticipated the sensor was in full sun, and when the shadows ultimately passed over the sensor the temperature began to fall. It seems I have some planning to do….