Projecting extreme climate events in China's Loess Plateau: multiple RCMs and emission scenarios corrected by a trend-preserving method.

Projecting extreme climate events under global warming is fundamental but challenging because of uncertainties from multiple sources. This study aims to project the changes in intensity, frequency, and duration of climate extremes for the period of 2021–2050 in China's Loess Plateau. In particular, the uncertainties in climate projection will be considered by employing five regional climate models (RCMs) under two representative concentration pathway scenarios (RCP4.5 and RCP8.5), decreased by a trend-preserving bias correction method, and further partitioned for each uncertainty source. The bias-corrected RCMs kept the change trends but satisfactorily corrected the climatic means and extremes. Compared with 1980–2005, the intensity of precipitation extremes would decrease while the frequency and duration of precipitation extremes would increase, and the warm/cold extreme events would increase/decrease for 2021–2050. The precipitation extremes had larger changes in the southeast, while the temperature extremes had larger changes in the west. For the projection of climate extremes, the uncertainties from RCMs (29% ± 27%) and RCPs (8% ± 11%) were far smaller than those from internal climate variability (62% ± 27%), implying that the projected climate extremes are reliable. The techniques relating to bias correction and uncertainty analysis can be referred by other studies, and the projected changes in climate extremes are helpful for soil conservation and water resource management in the study area. [ABSTRACT FROM AUTHOR]