AbstractWind-driven rain (WDR) is one of the most important causes for water damage in buildings. Therefore, the first crucial step to assess the hygrothermal performance of the building envelope, is the appropriate estimation of the amount of rainwater striking the building's façade. ISO 15927 offers the annual average index –mainly to assess the moisture content of absorbent surfaces, and the spell index – more related to the likelihood of water penetration through joints. To calculate these indices, assumptions are made concerning the length of the period of ‘no rain’, called ‘spell definition’. Obviously, the choice of this spell definition will characterise the WDR-amount. In this paper, WDR measurements of a 3-storey building in Vancouver, Canada are used to investigate how this spell definition affects the rain load. Different filter criteria are used to exclude errors due to measurement equipment. By means of the catch ratio as a dimensionless parameter, the results of the analysis for different spell definitions are compared to hourly and 5-min data. It is concluded that longer spell definitions result in lower catch ratios and an underestimation of the WDR load. Hourly data turns out to be a more conservative approach for WDR-assessment of this case study, but is able to represent the spread on the catch ratio most closely to the original 5min-dataset.