- 日本建築学会環境系論文集 (ISSN:13480685)
- vol.81, no.729, pp.1047-1054, 2016 (Released:2016-11-30)
The outputs of weather and climate models have been used in various application fields. For example, future weather data for the building energy simulation (BES) can be provided based on a climate model prediction. However, as the model output has systematical errors (called the bias), some type of bias correction is necessary in order to use the model output for an application field. For temperature or humidity, we often assume normal distribution and correct bias using statistical parameters, such as the average and the standard deviation. However, for solar radiation, a bias correction method (BCM) that uses only the average and standard deviation is insufficient and can result in negative values after bias correction. Consequently, the solar radiation bias is often corrected using only its average. In general, climate models can accurately predict the daily maximum amount of solar radiation on clear days at a given site because solar radiation depends mainly on its geolocation (latitude, longitude, and elevation) and the season (solar altitude). However, it is difficult to model cloud physics processes accurately to establish the weaker amounts of solar radiation on cloudy days. As a result, when we correct the solar radiation bias using only the average value, the daily maximum value deviates from the observed results instead of correcting the average. In this paper, we present a method called quantile mapping (QM) for the bias correction of solar radiation to provide the bias corrected weather data for the BES. The QM has been developed mainly for the correction of precipitation or temperature biases, although there are few studies that apply QM to the correction of solar radiation. In previous studies, QM was applied to the daily or monthly average. However, for the BES, the daily maximum value is also as important as the daily or monthly average, because the peak energy load depends mainly on the daily maximum. In this study, we also applied QM to obtain the daily maximum amount of solar radiation. In addition, we conducted BESs using the bias corrected weather data and evaluated the efficiency of each BCM. From the simulation results, the average energy consumption did not differ according to the difference in the BCM. However, the simulation that used the weather data corrected by only the monthly average could not predict the maximum cooling load; it was underestimated by 12%. Conversely, the simulation with the data corrected by QM, which used either the daily cumulative or the maximum amount of solar radiation, could predict the maximum cooling loads, which were under estimated by only 6% and 2%, respectively.