Without protection, temperature sensors absorb solar radiation and read higher than the true air temperature. To reduce this error, temperature sensors on weather stations are always shielded from direct sunlight. But even with shading, temperature sensors read high when the sun is bright and the wind speed is low.
We are evaluating the magnitude of temperature measurement errors on the USU Environmental Observatory. One sensor is shaded with a multi-plate shield (sometimes called a Gill shield); the other sensor is inside a new type of fan-aspirated shield that can be operated using solar power. The Gill shield is 'naturally ventilated,' and uses multiple stacked white plates to block both direct and reflected solar radiation, yet allow the wind to blow through it. In the aspirated shield, a solar-powered fan continuously draws air past the internal sensor to produce a more accurate measurement.
We are testing these shields in several locations, including the famous “Peter Sinks” near the top of Logan canyon https://en.wikipedia.org/wiki/Peter_Sinks. At this location we observed the largest difference we have recorded. It occurred on a clear and calm winter day when the low-angle sunlight reflected off the bright white snow went in between the plates on the multi-plate shield and caused it to read 13 C (23.4 F) higher than the fan aspirated shield. These large errors are common in winter and cause our mountain snow melt predictions to have significant errors.