STUDY ON THE RECALIBRATION OF STEVE AIR TEMPERATURE PREDICTION MODELS

Abstract

As part of Singapore’s economic growth, urban development has become an inevitable process for its national development. The most prominent problem that is related to urbanisation would be the Urban Heat Island (UHI) effect. As such, the availability of a climate impact assessment tool will allow urban planners to evaluate their proposed urban development. In this study, recalibration of the air temperature prediction models was performed to establish the association between air temperature and the climate and urban morphology predictors by inputting wind speed and water body as additional predictors. High R2 value of 0.86 is achieved for the Tavg model, follow by fairly high values of 0.76 and 0.63 for the Tmin and Tmax models respectively. However, the WIND and WATER variables were not significant. Subsequently, the development of models for average daytime (Tavg(daytime)) and average night time (Tavg(nighttime)) showed that wind speed is significant in reducing average daytime and average night time temperatures. Simple linear regression was preformed for the Tavg(daytime) and Tavg(nigttime) models and thereafter, water body was found to be effective in reducing average daytime temperature. The models validation conducted showed that the Tavg and Tavg(daytime) models can better predict the air temperature, while the Tmin, Tmax and (Tavg(nighttime)) models tend to give an overestimate of the measured air temperatures. Results of the sensitivity analyses provide evidence that greenery can reduce air temperatures, increasing building heights can reduce daytime air temperature while widening of canyon increases daytime air temperature. The main limitation of this study is the lack of sufficient measurement data for the WATER variable. Finally, recommendations for future study are made to enhance the effectiveness of the air temperature prediction models.