Hygrothermal numerical simulations are essential for assessing the long-term performance of building structures under varying climatic conditions. However, discrepancies between simulated and measured relative humidity (RH) profiles, particularly at subzero temperatures, negatively impact the reliability and introduce uncertainty for practitioners in hygrothermal assessments. While both simulations and measurement sensors calculate RH similarly above freezing, referencing liquid water, differences emerge below freezing due to variations in calculation methods. Simulations typically account for ice formation at subzero temperatures, whereas measurements reference supercooled liquid water, causing RH values to diverge as temperatures decrease. This novel study investigated how RH calculation methods influence simulation accuracy in cold climates. Converting subfreezing RH values from the climate cabin to align with simulations improves agreement and validation accuracy. The results showed that the RH calculation method at subzero temperatures have a minimal impact on simulation end results or mold growth risk, allowing direct use of sensor data referenced to water. As temperatures decrease, the difference between RH values with respect to water and ice increases, making it advisable to use ice-referenced values in extremely cold conditions (<−30 ℃) for better accuracy.