Generating Future Weather Files for Resilience

By Robert Dickinson of Argos Analytics, LLC and Benjamin Brannon of Arup

Download PDF


Weather files provide hourly times series of weather variables for a typical year based on historical observations, and are widely used for modeling energy use and human comfort in the built environment. However, since they are based on historical data, they are increasingly inaccurate tools for estimating the future performance of buildings and communities, especially those with lifetimes exceeding 30 years. A “morphing” technique [1] has been used to transform historical time series based on projected changes in the monthly averages of several climatic variables. Since future values of these variables are uncertain, especially at local scales, changes in the monthly mean value of each of the variables, or offsets, have been calculated for an ensemble of climate projections generated by models from the Coupled Model Intercomparison Project Phase 5 (CMIP5). These models were run for the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment (AR5). The offsets were calculated for several future time periods during the 21st century for both the 4.5 and 8.5 Representative Concentration Pathways (RCPs). For each combination of time period and RCP, a cumulative frequency distribution (CFD) of the offset in mean monthly temperature was constructed based on the percentile rankings of the projections in the ensemble. The association between the mean monthly temperature and the other variables for each model is maintained, providing a consistent and physically plausible set of inputs to the morphing algorithm for each point on the CFD. The Weather File Module of the WeatherShift™ tool provides an intuitive user interface that allows the user to select a location, emissions scenario, future time period and point on the CFD to obtain a projected weather file reflecting a plausible future climate.