The fact that you say this slightly higher reading was at night leads me to believe this temp reading may be caused by something either below or nearby the sensor radiating this difference. In the daytime with the sun shining directly on the shield I would expect the added warmth of the shield bowls to be included in the measurements, but at night, even with no breeze .... Nearby pavement, cement, bricks, stones, decking, etc. that were warmed from the daytime sun sounds more like it.
With regard to the solar fan you mention, if it's possible to mount it directly above the sensor (preferred) drawing up or below the sensor blowing up as well, you'd be better off then outside the shield blowing inside. Plus you mentioned the slightly higher readings were at night so the solar fan wouldn't help you there unless you rig it with rechargeable batteries for nighttime operation.