Evaluating the long-term changes in temperature over the low- latitude plateau in China using a statistical downscaling method

摘要

A statistical downscaling method (SDM) has been established through multiple stepwise regressions of predictor principal components using the ERA-Interim reanalysis data and the meteorological data collected from 115 stations in the low-latitude plateau in China from 1981 to 2015. The skill of the SDM was checked by comparing the results of the different predictor combinations and the different time lengths used to construct the SDM. In addition, compared to the historical simulation of the coupled Max Planck Institute Earth System Model (MPI-ESM-LR), better performance can be achieved by using the ERA-Interim data to construct the SDM in the low-latitude plateau. The long-term changes in temperature from 1981 to 2015 in the ERA-Interim reanalysis data are calibrated by the SDM over the low-latitude plateau of China. Furthermore, the SDM is projected into the simulation results of the MPI-ESM-LR model to construct a suitable SDM (ERA-SDM), and then the ERA-SDM is implemented to evaluate the future temperature changes in the low-latitude plateau during the period of 2018-2100 using the simulation results of the MPI-ESM-LR model under the RCP2.6, RCP4.5, and RCP8.5 scenarios, respectively. The results showed that an increase in temperature of 0.3 degrees C decade(-1) was found from 1981 to 2015, in which the fastest increase of 0.4 degrees C decade(-1) occurred in winter and the slowest increase of 0.2 degrees C decade(-1) occurred in autumn. Most models in CMIP5 failed to simulate the long-term changes in temperature over the last 30years in the low-latitude plateau region, and the temperature in the low-latitude plateau was underestimated by 2.4 degrees C using the 22 models. The SDM improved the annual and seasonal temperature characteristics and inter-annual and seasonal changes simulated by the MPI-ESM-LR. The future temperature predictions by the ERA-SDM indicated that the fastest temperature increase of 0.271 degrees C decade(-1) was found in spring under the RCP8.5 scenario. A faster rate of temperature increase was found in the northern part of the low-latitude plateau than in the southern part under the RCP8.5 scenario.