Bias Correction Methods for Generating Climate Data as Input to Hygrothermal Simulations

The objective of this research is to investigate the effects of different bias correction (BC) methods applied to climate data used to assess the hygrothermal performance on building envelopes based on simulations. To this end, a univariate and two multivariate distribution based bias correction methods were used to prepare weather files to assess the hygrothermal response of a wood frame wall assembly at the Ottawa International Airport. The effectiveness of the bias correction methods was evaluated by comparing the bias corrected climate data to observational data, as well as by comparing their hygrothermal responses as obtained through simulations. The bias correction improved upon the raw Regional Climate Model (RCM), which significantly overestimated the average annual rainfall, relative humidity (RH), and temperature, whilst underestimating the wind-driven rain (WDR). Despite improvements made to the climate data through bias correction, performance indicators such as the moisture content (MC), mould index (MI), and RH on the exterior of the OSB had a diverse response to the climatic conditions. Generally, multivariate BC methods are seen to perform marginally better than the univariate method in this context. However, BC was not able to produce the same time series of climate variables, which have a significant impact on the wetting and drying conditions, and consequently the risk to mould growth in wood frame wall assemblies.