Stationary and nonstationary extreme value modeling of maximum temperatures in Peninsular Malaysia using different reanalysis datasets
The increasing frequency and intensity of extreme temperature events pose significant challenges to environmental sustainability and public health, particularly in regions like Peninsular Malaysia (PM). This study employs the block maxima (BM) approach with the generalized extreme value (GEV) distri...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | en |
| Published: |
Springer
2025
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/121280/1/121280.pdf http://psasir.upm.edu.my/id/eprint/121280/ https://link.springer.com/article/10.1007/s00704-025-05547-3?error=cookies_not_supported&code=30584535-50b4-4057-9eb0-a17dfe82b3dc |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The increasing frequency and intensity of extreme temperature events pose significant challenges to environmental sustainability and public health, particularly in regions like Peninsular Malaysia (PM). This study employs the block maxima (BM) approach with the generalized extreme value (GEV) distribution to model annual maximum temperatures (Tmax) from 1958 to 2023, using three reanalysis datasets: Japanese 55-year Reanalysis (JRA-55), Japanese Reanalysis for Three Quarters of a Century (JRA-3Q), and Fifth Generation European Centre for Medium-Range Weather Forecasts Reanalysis (ERA5). Both stationary and nonstationary models are applied to account for temporal changes in temperature extremes. The spatial distributions of return levels for 5-, 10-, 20-, and 50-year return periods highlight regional differences, with northern areas such as Kedah and Perlis consistently showing higher Tmax compared to the southern and eastern regions. Notably, JRA-3Q and ERA5 predict higher return levels, with JRA-3Q showing some locations approaching 45 °C for the 50-year return period. In contrast, JRA-55 estimates lower return levels, reflecting differences in spatial resolution among the datasets. ERA5 demonstrates superior accuracy in capturing spatial patterns, with the highest correlation (0.5458), lowest bias (-0.5218) and lowest RMSE (2.2245), emphasizing the critical need for accurate climate modeling in formulating localized risk management strategies. These findings contribute to a deeper understanding of extreme temperature risks in PM, providing valuable insights for climate adaptation and mitigation efforts. |
|---|