Exploring the behavior and changing trends of rainfall and temperature using statistical computing techniques

The present study aimed at quantifying the change in surface air temperature and monthly total rainfall. The changing trend was detected using Mann-Kendall trend test, seasonal Mann-Kendall trend test, and Sen's slope estimator. K-means clustering algorithm was used to identify the rainfall distribution patterns over the years and also their changes with time. A comparative analysis was done among different time series prediction models to find out their suitability for forecasting daily temperature in climatic condition of Bangladesh. The analysis was performed using daily temperature and rainfall data of more than last 40 years (till 2009). The study found an increasing trend in maximum temperature during June to November and in minimum temperature during December to January in Bangladesh. There has been seen no significant change in rainfall over the years. However on the western side of the country, the amount of rain is significantly less than the eastern side. The study found that different prediction models were appropriate for different conditions.