A Surrogate Weather Generator for Estimating Natural Gas Design Day Conditions

Natural gas customers rely upon utilities to provide for heating in the coldest parts of winter. Heating capacity is expensive, so and end users (represented by commissions) must agree on day which a utility expected meet demand. The return period such long relative amount weather data that are typically available. This paper develops resampling method called Surrogate Weather Resampler, creates large dataset support analysis extremely infrequent events. While most current methods generating based simulation, this resamples deviations from typical weather. also shows how extreme temperatures strongly correlated demand natural gas. Resampler was compared in-sample out-of-sample WeaGETS generator using both Kolmogorov–Smirnov test an exceedance-based cold generation. A naïve benchmark examined. These studied National Oceanic Atmospheric Administration AccuWeather. were collected 33 stations across North America, with 69 years each station. We show can reproduce tail distribution better than WeaGETS.