A METHODOLOGY FOR THE SYNTHESIS OF HOURLY WEATHER DATA

The ability to generate weather data from limited inputs and location independent correlations would allow simulations of environmentally driven systems to be run at locations for which detailed weather records do not exist. A further improvement would be to generate just one year of data which would yield the same simulation results as those of long-term data, thereby minimizing the computation needed to assess long-term system performance. In this paper, a methodology to calculate such a year of data is described. It is a combination and modification of several previously published generation methods. Statistics obtained from long-term (22 years), Typical Meteorological Year, and synthetically generated weather data are compared at three locations (Albuquerque, NM, Madison, WI, and New York, NY).