Sensitivity and uncertainty analysis of England's housing energy model

SAP/BREDEM-based models are commonly used for estimating national domestic energy consumption and for assessing the energy and environmental impacts of changes in the housing stock. However, policy-makers should recognize that such models are subject to multiple potential sources of uncertainty. A one-at-a-time sensitivity analysis is undertaken on the Cambridge Housing Model, a SAP 2009 model for estimating total domestic energy consumption in England and the UK. The analysis identifies several parameters to which the model is particularly sensitive, determines that the response to variations in individual parameters is typically non-linear, and concludes that changes to multiple parameters are not additive. A Monte Carlo uncertainty analysis is then undertaken to provide an indication of the impact of multiple uncertainties on model outputs. A concept map is developed to outline a number of potential sources of uncertainty, and in the absence of reliable information expert judgement is used to describe these uncertainties. Outputs suggest that any single estimate of total consumption may be subject to considerable inaccuracy compared with actual use. Similar levels of uncertainty and inaccuracy are likely for any SAP/BREDEM-based national domestic energy model, such as the models underpinning the Green Deal.