A Study using Future Weather Data in Consideration of Climate Change and Local Climate Phenomena using Dynamical Downscaling for Building Energy Simulations

1. INTRODUCTION

  This study uses a dynamical downscaling method with a global climate model (GCM) and a regional climate model (RCM) to develop future standard weather data for use in the building energy simulations. Climate change due to global warming causes serious problems. In the architectural design process, designers need to take the climate into consideration for proper energy strategies. Energy simulations are often used to evaluate the indoor thermal environment and energy consumption of buildings, and in these simulations, it is common to use regional climate weather data based usually on current, or past, observed climate information known as "standard weather data." However, most buildings are designed for use over several decades, and it is questionable how they will adapt in a gradually changing climate. It is therefore important to develop future weather data so as to consider a gradually changing climate when developing energy simulation programs. Although GCMs can predict long-term global warming, they cannot describe the details of local phenomena due to the coarse grid resolution (~100 km), and standard weather data requires more detailed climate information. Therefore, a dynamical downscaling method was employed using a high resolution (~1 km) RCM, to obtain detailed spatial and temporal weather data, which is expected to reproduce the locality. The objective of this paper is to confirm that the weather data predicted by this method reproduces regional characteristics and local climate change.

2. METHOD

  In this study, we employed the model for interdisciplinary research on climate version4 (MIROC4h) as the GCM and the weather research and forecasting (WRF) model as the RCM. MIROC was developed mainly by the center for climate system research, and among GCM models, it has a relatively high grid resolution and a horizontal scale of about 60 km. The target area of this study in the Kanto region, or Tokyo and its surrounding area, and this locality is used in the simulation using the WRF model. In the dynamical downscaling method, we used GCM data to RCM as initial and boundary conditions and downscaled the GCM data physically. We used four levels of nested regional climate modeling: where the first and fourth levels have horizontal spatial resolutions of 54 km and 1 km, respectively. To confirm the regional characteristics reproduced by downscaling, we conducted the simulation using current data (2006 - 2010) and compared the weather data (temperature, humidity, wind velocity and direction) in three cities: Otemachi (lat35.9, lon139.76), Tsukuba (lat36.65, lon140.12), and Kumagaya (lat36.15, lon139.38). We then conducted the simulation using future data (2031 - 2035), and compared the future weather data with the current weather data to assess the differences in climate change among the three cities.

3. RESULTS

  Figs. 1-4 represent the weather component mean average daily changes and the mean wind direction frequency from August 2006 - August  2010 for each of the three cities. Fig.1 represents the daily temperature, and shows that the night-time temperature at Otemachi is higher than that in Tsukuba and Kumagaya, due to urban heat storage. It is noticeable that the maximum daytime temperature in Kumagaya is attributed to heat generated in urban areas south of Kumagaya, which is transported by the south sea breeze from Tokyo Bay. Fig. 3 represents the daily wind velocity changes, and it is evident that there is a weak wind in Kumagaya, due to its inland location. After obtaining the regional climate information, we examined local changes using present and future data. Table.1 shows the average temperature and humidity of the five-yr mean for present and future, as reproduced by MIROC and WRF. It also shows temperature and humidity changes reproduced by WRF and MIROC. From the WRF result, it is evident that the maximum temperature change is at Tsukuba, (1.27°C), and the minimum temperature change is at Kumagaya, and that the maximum change in specific humidity is at Tsukuba, (0.00173[kg/kg]), and the minimum change at Kumagaya, (0.00111[kg/kg]). In the WRF result, the maximum changes of both temperature and humidity are found at Tsukuba, and the minimum changes are at Kumagaya. However, in the MIROC result, although the maximum change of both temperature and humidity is found at Tsukuba and the minimum change at Kumagaya, the predicted temperature and humidity changes at Otemachi and Tsukuba are smaller than those predicted by WRF. In contrast, the changes predicted by WRF at Kumagaya are smaller. It is evident that local climate change can be reproduced using a dynamical downscaling method.

4. CONCLUSIONS

  Using a dynamical downscaling method with the WRF model, we can obtain regional climate information and the local climate changes. In other existing methods climate change is predicted by GCM and added to current weather data in order to predict future local weather data, the effect of local climate change is not considered.  To create future standard weather data, which requires high-resolution climate information in the order of a few km, the deployment of this dynamical downscaling method is a useful way to consider regional characteristics and regional climate change. Future weather data derived from dynamical downscaling is expected to reproduce both global climate change and local climate phenomena, and by using this method, designers can take the future local climate into consideration.

Table.1Average temperature and, humidity of present and future 5-yr mean reproduced by MIROC and WRF, and temperature and humidity changes reproduced by MIROC and WRF
City and Period	Average Temperature [°C]		Average Specific Humidity [kg/kg]		Temperature Change [°C]		Specific humidity Change [kg/kg]
	MIROC	WRF	MIROC	WRF	MIROC	WRF	MIROC	WRF
Otemachi  2006-2010	25.72	26.24	0.0187	0.0147	0.86	1.12	0.00144	0.0016
Otemachi  2031-2035	26.59	27.36	0.0198	0.0163
Tsukuba     2006-2010	26.27	25.06	0.0187	0.0149	0.92	1.27	0.0015	0.00173
Tsukuba     2031-2035	27.19	26.33	0.0202	0.0166
Kumagaya 2006-2010	24.89	26.67	0.0181	0.0150	0.81	0.50	0.00144	0.00111
Kumagaya 2031-2035	25.70	27.17	0.0196	0.0161