Degree-day based non-domestic building energy analytics and modelling should use building and type specific base temperatures

[Display omitted] •Studied HDD base temperature of 119 buildings using 3pH model on 4-year gas use.•Base temperature range: 11.6–20.5°C; varies because of characteristics and type.•Mean is 1.2°C higher & 1.6°C lower than CIBSE (15.5°C) and ASHRAE (18.3°C).•Highest mean, 17.7°C for clubs and community; lowest, 12.8°C for storage buildings.•Published base temperature must be building type specific and updated regularly. A deeper understanding of building performance is essential to reduce their energy consumption and corresponding greenhouse gas emissions. Heating degree-days (HDD) encapsulates the severity and duration of cold weather, which is routinely used for weather related analysis of fuel consumption, performance benchmarking, and compliance. The accuracy of HDD-based prediction largely depends on the correct base temperature, which varies depending on building thermal characteristics, and their operation and occupancy. We analysed four years’ (2012–2016) half-hourly metered gas consumption from 119 non-domestic buildings representing seven types, to: (a) identify their base temperature using a three-parameter change point (3pH) regression model, and (b) their relationships with intrinsic building parameters. The highest mean base temperature, 17.7°C was found for clubs and community centres, and the lowest, 12.8°C was for storage buildings. The average of all base temperatures is 16.7°C, which is 1.2°C higher and 1.6°C lower than the British (15.5°C) and American (18.3°C) standards respectively. The current practice of a fixed base temperature degree-days for all buildings has been found to be unrealistic. Building type specific base temperatures must be developed, agreed upon and published for increasing accuracy in energy analytics and legislative compliance, as well as for developing effective standards and policies.