Weather files for remote places: Leveraging reanalyses and satellite datasets

Weather files capture the time-varying conditions under which buildings perform and, as such, they constitute one of the fun-damental inputs for building performance simulation. In theory, the creation of weather files only requires collecting data at a certain frequency for a key number of variables during the time of interest. In practice, several problems arise. Direct measurement on a project basis can be a costly operation considering the site accessibility and the number of instruments needed to collect complete weather observations. Sometimes, this is simply impossible if a study requires historical data. These issues are traditionally overcome using the weather data collected at a nearby public weather stations, but this can be equally challenging, or even impossible, depending on how far away the station is and the frequency and completeness of observations.
Arising from the need to simulate the thermal performance of buildings at remote locations, this study presents an approach to generate weather files based on satellite imaging and reanalysis datasets. Given the good agreement with local station’s observations, it is shown how these publicly available datasets can be combined to create weather files suitable for building performance simulation. This is applied to a case study to compare the performance of a building and its systems against traditional weather files. The work quantifies and discusses the discrepancy obtained between the different sources. Overall, results indicate that satellite and reanalysis datasets constitute a suitable resource to create weather files for building performance simulation.