|
چکیده
|
Worldwide research on climate change reveals that a slight temperature change can trigger extreme phenomena such as drought, heavy rain, and storm. This study aims to investigate and compare extreme value theory (EVT) methods in the nonstationary analysis of maximum monthly temperature in Arak plain of Iran over a 116-year period (1901–2016). To this end, the maximum temperature data were collected from the CRU(Climate Research Unit) gridded dataset for stations located in the study area including Sarugh, Shanagh, Arak, Ashtian, and Gavar. Then, using EVT methods including two methods of block maxima (BM) and peaks-over-threshold (POT), time series of maximum temperature were extracted. The MK (the Mann–Kendall), ADF (augmented Dickey-Fuller), KPSS (Kwiatkowski-Phillips-Schmidt-Shin), and PPT (Pettitt’s test) tests were used to examine the trends, stationarity, and homogeneity for the selected stations. Then, distribution parameters were computed based on the DE-MC (diferential evolution Markov chain) method. The results showed that all the mentioned stations have a non-stationary trend and the highest diference in the maximum annual temperature was about 0.8 °C over a 2-year period. Moreover, the BM method showed the lowest temperature uncertainty in diferent return periods, which is recommended as the best method in this research
|