“Every site is different” is a mantra for design that encompasses the diversity of issues that we all face on projects. Topography, demographics, infrastructure, and other site-specific information is typically accessible to a design team and plays a vital role in the process. But what about climate data? Traditionally, climate data are collected from the ‘nearest’ weather station or from data tables published by ASHRAE or government agencies. In many cases these data may be sufficient, but what if the station and site are each located in two different local climates? The applicability of the ‘nearest station’ may be questionable. Since these data form the basis for decisions related to energy consumption, building massing and design of the envelope, could there be some benefit to having site-specific climate modeling?
We recently conducted a comparative analysis that involved the design of a new residential campus in Southern California. The analysis involved two sites (“Old Site and “New Site”), located as shown in Figure 1. Anecdotal evidence suggested that natural ventilation systems were performing well at the Old Site. Given the significance of the decision on the mechanical design and impact on project cost, there was interest in investigating the feasibility of natural ventilation for the New Site.
Outdoor temperature, humidity and wind velocity are key factors in the determination of the successful application of natural ventilation techniques. During the daytime it is desirable for incoming air to be at a lower temperature than the indoor air. At night, effective natural ventilation systems harness the cooler outside air temperatures to reduce cooling loads. Humidity is one of the most important factors since high levels of humidity negatively influence thermal comfort. If humidity levels are too high, mechanical air conditioning systems are required for dehumidification, countering the viability of natural ventilation systems.
Figure 1: Study sites and meteorological data stations included in the climate model.
Both sites are located within 2 miles (3.5 km) of the coast and within 5 miles (8.5 km) of the nearby Marine Corps Air Station Miramar (Miramar), where 30+ years of meteorological data are available. Based on proximity, one might assume that the climate conditions at all three locations are similar. However, the topography, variations in land use, proximity to a large body of water and nearby desert could produce varying weather conditions in this region marked by steep climatic gradients. Anecdotal evidence suggested the new site is both hotter and drier. But how much hotter and how much drier?
Using methods developed and validated during our research study for ASHRAE, high-resolution climate downscaling was conducted to regenerate 30 years of weather data across a domain that included both sites and the airport. The data were regenerated on a rectangular grid with 150 ft (50 m) node spacing, producing a full 30 years of meteorological data at each point on the domain. This resulted in hundreds of observation points having data equivalent to a weather station.
Inputs to the process included: high-resolution terrain data from USGS, eight years of Mesonet weather station data (see Figure 1), and eight years of upper air (weather balloon) data from nearby airports. This massive volume of information was scrutinized and then processed for input to a four-dimensional atmospheric physics model. Outputs from the procedure included micro-meteorological variables such as temperature, wind speed, wind direction, solar radiation etc. in both space (on a gridded domain), and time. After 8-years of detailed gridded climate data was generated a statistical correction was made with nearby airport weather stations, and 30 years of data sampled hourly were generated at the high-density grid points in the study area. This is no simple interpolation of data! In fact, it requires specialized meteorological knowledge and super-computing capabilities to be done efficiently, economically and correctly.
Sample results from the regenerated weather data at the sites are shown in Figure 2 through Figure 4. The full 30-year regenerated data sets, were also processed to produce the data formats required for energy modelling software (TMY3, .epw etc.), for use by the sustainable design team.
What these example results indicate to a meteorologist:
- Higher wind speeds occur more frequently at the Old Site (thicker green bars on the wind rose for the Old Site versus the New Site), and more easterly winds occur at the New Site (bringing more dry air) than the Old Site
- The Old Site is warmer during the winter months (10th centile dry bulb), and the airport is three-to-six degrees cooler than both sites during the winter months.
- The New Site is hotter than the Old Site during the summer months (90th centile dry bulb), and the airport is four-to-five degrees warmer than both sites during the summer months.
- Conditions are more humid at the Old Site during the winter months compared to the New Site and Airport.
- The humidity conditions are relatively similar at all sites during the summer months, with little variation in wet bulb temperature.
- The design conditions as taken from the airport station differ from the regenerated weather data at both sites (i.e., the 10th and 90th centile values).
Are these differences significant in terms of the decision on natural ventilation? I leave those conclusions to sustainable design experts. Suffices to say that there is in fact a difference in conditions at these sites, and that the use of the ‘nearest airport’ data for design is questionable.
This is just one example of how site specific climate modelling can be used to inform the design process. These types of data are showing to be effective for informing decisions related to sizing of heating and cooling systems, solar radiation, the water-energy nexus and many others. If your project is in a coastal region or a region marked by complex topographical and meteorological conditions, you may need to reconsider the data you use for your design.
A parting word of caution: be careful when selecting your weather station and when using interpolated weather data from multiple stations. The data may not reflect actual site-specific conditions, and the results from an interpolation will not account for the real physical process that occurred between those points – particularly when there are variations in topography, land use and hydrological conditions in between.
Do you think today’s projects could benefit from site specific climate modelling? We would love to hear your thoughts.