Combining GIS-based statistical and engineering urban heat consumption models: Towards a new framework for multi-scale policy support

• We compare a statistical and engineering modelling methods for urban heat consumption calculation.
• Total consumptions calculated by both modelling methods deviates by 18% on the case study Rotterdam Bospolder.
• Deeper analysis shows major differences in the encompassing of building subsequently refurbished and occupants’ behaviour.
• Based on both method complementarities, we combine them in an innovative multi-scale framework.

Diagnosing and modelling precisely the actual energy consumption at the urban scale is the indispensable starting point of any low-carbon urban energy policy. This paper compares two building heat consumption models for urban scale applications: a statistical model based on 2D-GIS and multiple linear regression, and an engineering model making use of 3D city models and monthly energy balance of standard EN ISO 13790.Both methods are combined in a new multi-scale framework for improved prediction of heat demand and energy savings potential of building stock at the several scales within the city.This multi-scale framework was tested for the case study of Bospolder – Rotterdam (Netherlands) including around 1000 buildings. Firstly, the statistical model predicts the energy consumption of buildings at the city scale, then relevant neighbourhoods for retrofitting plans are selected and precisely modelled using the engineering model, finally individualized energy savings potentials are predicted building by building.The prediction provided by the two models demonstrated a good agreement with measured gas consumption data at the neighbourhood level (5–25% deviation), while errors become higher at disaggregated level. Major differences result from the ability of each model to cope with the lack of information concerning subsequently refurbished buildings, occupants’ profile and behaviour, and unoccupied buildings.The study showed the ability and effectiveness of the multi-scale framework to support decision about retrofitting plans at different levels and scales.