Estimating energy savings for the residential building stock of an entire city: A GIS-based statistical downscaling approach appliedto Rotterdam

Energy retrofit of buildings represents an important sector for mobilizing investments to address carbonmitigation of cities. The identification of the actual energy consumption profile of large building stocksis a necessary step to evaluate the impact of retrofit measures, e.g. energy savings, at city scale.The present study introduces a bottom-up statistical methodology based on a Geographical InformationSystem (GIS) to estimate the energy consumption of residential stocks across an entire city.The adoption of a multiple linear regression model allows the downscaling of measured natural gasand electricity consumption from the aggregated post-code level to single dwellings, based on severaldescriptors, such as dwelling type, period of construction, floor surface and number of occupants. Theenergy consumption is apportioned to different end-uses and corrected for weather, then the energysavings potential is estimated by accounting for the implementation of typical refurbishment measures.Results are finally aggregated across the whole city for evidence-based decision support in sustainableurban planning.The study provided relevant results to prioritize the implementation of energy retrofit measures forthe residential stock of Rotterdam city, consisting of about 300,000 dwellings. The methodology can befurther applied to other contexts due to its generic nature.