| Summary: | This report states the main purpose of the final year report on statistical downscale of rainfall and temperature which is to make use of R programming language and its packages, such as CaDENCE, nnet and MASS to conduct a rainfall and temperature study by statistical approach. The statistical models proposed to be used are multiple linear regression model and artificial neural network model. Even though Singapore is a country located at the tropical region and experience tropical rainforest climate with plenty of rainfall throughout the year, it is still possible for Singapore to face water shortage issue. Hence, carefully monitor the rainfall and conduct study to predict precise future climate data is very important. Besides rainfall, temperature is another study aspect. Singapore has relatively high temperatures throughout the whole year, when the temperature is pretty high and the rainfall is abnormal little, Singapore will suffer drought and this will affect people’s daily life greatly. Hence, if the rainfall and temperature can be predicted in a more precise way, it is possible for government and population of Singapore to come out with effective measures to deal with the drought and the rise of temperature will also cause the rise of sea level and affects island countries like Singapore. In this project, a set of historical rainfall data from 1967 to 2010 will be used to build up the correlation with the rainfall data generated by the GCM (Global Climate Model) from 1967 to 2010. This relationship can be used to predict more precise future climate data with the future climate data generated by GCM. Two set of future climate data from 2041 to 2060 and 2081 to 2100 from GCM (model HADCM3 and model CGCM3T47) will be selected to predict the future climate data based on statistical model constructed.
|
|---|