Evaluation of current and future hourly weather data intended for building designs: A Philadelphia case study

Abstract

Global warming presents a critical grand challenge for 21st-century building design and operation. Today's buildings are designed using reference year databases that represent past weather observations and do not account for future weather scenarios predicated on global and regional climate models. But these models lack the required spatial and temporal resolution to be utilized in dynamic building simulation tools. The present article reviews methods for generating current, as well as future weather files used in dynamic building simulation tools for the City of Philadelphia. Our findings suggest that current TMY files are not suitable to capture predicted trends of global warming. Running a statistical t-test revealed a tendency to underestimate key parameters such as dry bulb temperature of the weather files. A comparison between outputs of four different weather tools in generating future weather data indicates significant variance in predicting the dry bulb temperature range, and an overall increase in extreme and mean figures. The variations among the generators in predicting global horizontal irradiation and wind speed are small. However, the diffuse normal irradiation increases in all predictions through future time slices. A correlation analysis is conducted on the outputs of each weather generator in order to investigate bivariate correlations for both current and future weather data. The discrepancies and uncertainties among the weather files are identified and a process for addressing uncertainties is outlined. This investigation is intended to guide the development of weather data sets for building design, also address challenges regarding the generation and use of current and future weather data sets.