Time-varying Distribution Analysis for Rainfall and Air Temperature Data in Jakarta in Response to Future Climate Change

AbstractIndonesia is vulnerable to climate change (rainfall and air temperature), which can increase the chances of climatic disasters. An organized risk analysis is a strategic plan to minimize the impact. The purpose of this research is to estimate time-varying distribution parameters for normal, generalized extreme value (GEV), and lognormal distributions using fminsearch and MLE algorithms on rainfall and air temperature data in Jakarta, as well as visualize and analyze the best time-varying distribution. The maximum likelihood estimation (MLE) method is used for stationary distribution parameter estimation. The fminsearch algorithm is used for stationary and nonstationary distribution parameter estimation. The highest difference value of stationary distribution parameter results from both methods is 5.3768 mm for rainfall data and 0.2670°C for air temperature data. The results of the best distribution based on the AIC value are the 3-parameter lognormal distribution for rainfall data and the 4-parameter GEV distribution for air temperature data. Over time, the variance of rainfall increases, and the average air temperature increases with a fixed variance.