Accuracy of Skin Temperature Sensors Under Cold and Windy Conditions: Effects of Sensor Type and Insulation

Accurate skin temperature measurement is essential for assessing thermal strain in cold environments. This study evaluated the accuracy of four contact sensors under cold and windy conditions and examined how insulation affected measurement accuracy. Sensors were first calibrated in a temperature bath (5-40°C) against a reference thermometer. Nineteen participants then completed two climate-chamber trials: a windy-chill condition (-7°C, 2.5 m/s wind) and cold-chill condition (-15°C, 0.2 m/s wind). Each sensor (iButtons®, BodyCAP, MSR, ADInstruments) was tested on the forearm with and without insulation. Infrared thermography (IRT) served as reference. In the water-bath comparison, all sensors performed within acceptable accuracy ranges, with biases from -0.13 to 0.29°C and 95% limits of agreement ≤ 0.63°C. During human trials, all sensors showed systemic overestimation, most pronounced for BodyCAP. Uninsulated, iButtons® overestimated temperatures relative to IRT in cold-chill and underestimated them in windy-chill conditions, while BodyCAP, MSR, and ADInstruments were unaffected by condition. Insulation consistently led to an overestimation of temperatures relative to IRT (bare bias: -1.24-2.84 °C, insulated bias: 1.41-7.5°C, P≤0.004). All sensors demonstrated temperature-dependent bias in the windy-chill condition, with larger deviations at higher skin temperatures. Error distributions were not significantly different over insulation or wind condition. Although all sensors showed acceptable accuracy in controlled water-bath calibration, systematic biases emerged during human trials. Insulation consistently caused overestimation, and iButtons® were sensitive to wind, whereas other sensors were not. Bias increased at higher skin temperatures under windy conditions across all sensor types.