Decomposition analysis of air temperature data in Bangka Island

Decomposition analysis is performed for air temperature data from weather station in Nuclear Power Plant (NPP) candidate site in Bangka Island. Air temperature data at the height of 10 meters is obtained from an hourly record using an Automatic Weather Station (AWS). Data preprocessing is performed by selecting, interpolating and averaging and then analyzed using decomposition method and statistical tests to obtain stationerity, periodicity and residual parameters as well as stasionerity. The result of analysis shows that within the March, 2012 to July, 2016 monitoring period, the air temperature pattern is categorized as stationer. Significant yearly change in average air temperature is not apparent and the manifestation of global warming phenomena cannot be clearly identified from the data. In Bangka Island, there is no significant average air temperature difference within one-year period and relatively high average air temperature can be identified in the months of May and June. Outliers in the air temperature data are mostly below the lowest average temperature. (author)