Extreme climatic conditions are those falling in the upper or lower tenth percentile of the local period of record. North American hygrothermal datasets rely on such extremes in simulating climatic effects on building performance. Prior to 2012, WUFI employed customized 30-year hourly data provided by the U.S. Department of Energy for the 1961-1990 period of record. This dataset was supplemented with hourly rainfall data, providing hygrothermal years selected from the 10th percentile of the warmest and coolest years. The DOE-WUFI climate files represent composite Hygrothermal Reference Years, not contiguous weather years.
In 2011, ASHRAE 1325-RP developed Moisture Design Reference Years (MDRY) from hourly climate records for 100 locations in the United States and 7 locations in Canada. The current data include a collection of three ‘worst’ years for each location. ASHRAE 1325 recommends the third year, which corresponds to the 10th percentile for a 30-year dataset and a severity event that is expected to occur once in every 10 years. These MDRY’s represent Actual Meteorological Years (AMY) comprised from contiguous data for a given location.
Case Study: Miami, Florida; 1992
This study assessed climatic effects on a 2×6 framed stucco wall assembly configured for typical south Florida construction. The OSB moisture loading component faces northeast.
The chosen weather year was 1992, a year marked by Hurricane Andrew (August 24, 1992), one of the most devastating hurricanes in U.S. history. Climatic patterns for the south Florida region were atypical, having prolonged periods of drought and punctuated periods of heavy rainfall. Although unusual in terms of rainfall periodicity, 1992 was actually fairly ‘normal’ in terms of temperature and total rainfall (see 1992 rainfall & temperature graph).
Hygrothermal simulations were performed using hourly measured meteorological data for 1992, WUFI Cold Year, WUFI Warm Year, ASHRAE Year 1 (most severe year according to ASHRAE’s dataset selection), ASHRAE Year 2 (second most severe year according to ASHRAE’s dataset selection), and ASHRAE Year 3 (third most severe year according to ASHRAE’s dataset selection).
As expected, predicted moisture loading within OSB sheathing corresponded with periods of heavy rainfall in June, August/September, and November. For this particular location and assembly type, most of the climate datasets predicted similar peak relative humidity, however, the time period and duration of these peaks varied considerably. The same was generally true for the OSB water content. WUFI’s cold years are generally more severe than warm years for most assembly types and locations. Here, WUFI’s cold year provided peak RHs and moisture contents that were very similar to ASHRAE’s severe years.
Interestingly, ASHRAE Year 1 (i.e. most severe year according to ASHRAE’s dataset selection) was actually less severe than Year 3. This is probably due to the use of a northeast wall rather than a north wall assumed by the ASHRAE selection method. Still, it is easy to see that the rank order of ASHRAE’s three severe years will vary based on assembly type, orientation, and assumed failure criteria. ASHRAE Year 2 (not shown) yielded results that were lower than Year 3 but higher than Year 1.
The 1992 weather year was fairly typical in terms of annual temperature and overall rainfall totals. Still, climatic patterns in 1992 yielded hygrothermal conditions that were more in keeping with extreme years (ASHRAE Years 1 and 3). Obviously, Hurricane Andrew was a very significant event as it increased moisture loads prior to even greater rainfall events in November, which were 10 inches greater from than its 30-year November average. WUFI’s Cold Year as well as ASHRAE’s Year 3, which were both derived from 1961-1990 data, provided a good predictor of this November 1992 extreme.
Although the modeled wall assembly performed similarly under all datasets, the results underscore the need for comparing multiple climate conditions when performing robust analyses. Furthermore, one should not rely on the rank order of the ASHRAE weather years, especially when deviating from the assumed north orientation. Obviously, annual extremes in parameters such as temperature, rainfall, RH, wind, and cloud cover/solar radiation are important when assuming ‘representative’ extreme years. But equally important are extremes in weather patterns for a given contiguous year, including frequency, periodicity, and duration of critical climatic events.
Unlike energy modeling, where simulations focus on long-term performance under typical temperature patterns, hygrothermal datasets emphasize atypical or extreme conditions. This is the preferred approach for moisture-control design as the purpose of hygrothermal modeling is to evaluate failure tendencies. Nonetheless, the approach is at odds with the assumptions made by ASHRAE 160-2009, which states, “An international consensus has emerged that the analysis should be predicated on loads that will not be exceeded 90% of the time. This standard adopts this approach.” Some hygrothermal modelers are critical of extreme datasets, especially when used in conjunction with ASHRAE 160’s evaluation criteria and the 1% wind-driven rain fraction. They claim unrealistic conditions, if not insurmountable odds. But decades of failed buildings tells us otherwise. The use of extreme climate datasets, including AMY data, provides an important tool in robust resilient design. Failure to embrace these ‘conservative’ measures can yield surprisingly negative results.