High mountain areas are poorly represented by official weather observatories. It implies that new instruments must be evaluated over snow-covered and strongly insolated environments (i.e. mid-latitude mountain areas). We analyzed uncertainty sources over snow covered areas including: 1) temperature logger accuracy and bias of two widely used temperature sensors (Tinytag and iButton); 2) radiation shield performance under various radiation, snow, and wind conditions; 3) appropriate measurement height over snow covered ground; and 4) differences in air temperature measured among nearby devices over a horizontal band.

The major results showed the following. 1) Tinytag performance device (mean absolute error: MAE ≈ 0.1–0.2 °C in relation to the reference thermistor) was superior to the iButton (MAE ≈ 0.7 °C), which was subject to operating errors. 2) Multi-plate radiation shield showed the best performance under all conditions (> 90% samples has bias between ±0.5 °C). The tube shield required wind (> 2.5 m s⁻¹) for adequate performance, while the funnel shield required limited radiation (< 400 W m⁻²). Snow cover causes certain overheating. 3) Air temperatures were found to stabilize at 75–100 cm above the snow surface. Air temperature profile was more constant at night, showing a considerable cooling on near surface at midday. 4) Horizontal air temperature differences were larger at midday (0.5 °C). These findings indicate that to minimize errors air temperature measurements over snow surfaces should be carried out using multi-plate radiation shields with high-end thermistors such as Tinytags, and be made at a minimum height above the snow covered ground.