Climate projection using quantile matching bootstrap: A case of temperature and precipitation in Komodo Island, East Nusa Tenggara

Abstract

Changing the physical conditions of the Earth's atmosphere include temperature and distribution of rainfall, which has a wide impact on various sectors of human life. According to the Indonesia Food Security Monitoring Bulletin, Komodo Meteorological Station in East Nusa Tenggara (hereafter denoted as NTT) Province in Indonesia has reported that Komodo was the one of first priority in the classification of districts experiencing drought impacts from climate change that occurred in Indonesia. The study aimed to climate projection in Komodo using the Quantile Matching Bootstrap (QMB) method, which is the block bootstrap simulation technique combined with a quantile prediction and matching method for simulating future daily climate data. The projected data is influenced by the relationship of predictors derived from the Global Climate Models to the observed data. The study was carried out by examining daily rainfall and daily maximum temperature observed data from Komodo Meteorological Station. In this paper, the performance of the QMB approach is improved by three different quantiles i.e. 10th, 50th, and 90th percentiles. The result for QMB produces the average coverage rate for the prediction intervals of around 80%. By using out of sample validation techniques, the results show that this method has successfully projected daily rainfall and daily maximum temperature during rainy season as well as transition from dry season to rainy season.