针对江苏省气温和降水的高分辨率降尺度模拟及预估

Dynamical downscaling and projection of surface air temperature and precipitation in Jiangsu Province with high resolution

本文利用WRF模式,以25 km分辨率数值模拟结果作为驱动场,对江苏省现代和未来15 a的气候进行5 km高分辨率降尺度模拟及预估。结果显示,高分辨率降尺度模拟对其驱动场具有显著的提高,降水的负偏差和气温冷偏差均有所降低,其模拟的降水与气温概率分布与观测更为接近;对于极端指数,WRF模式能够模拟出其基本分布,除连续湿润日数CWD和极端高温TXx之外,高分辨率模拟对其他指数的模拟均有显著的提升。在RCP8.5排放情景下对未来气候变化的预估表明,江苏降水在夏季以减少为主,在春季则以增加为主,全年平均降水存在减少趋势;未来0~1 mm·d^(-1)的微弱降水发生概率将增加,小雨、中到大雨以及暴雨发生的概率均降低,而暴雨强度的增强导致极端强降水R95显著增加;气温25℃以上高温发生的概率在未来有所增加,而0℃以下的低温发生概率减小,从而导致暖持续日数显著增加,而冷持续日数减小,另外,极端高温和极端低温都有显著的升高。

Driven by the numerical simulation output with 25 km resolution, the WRF model was utilized to perform dynamical downscaling for the present climate and the future climate projection in Jiangsu Province with a high resolution of 5 km, covering 15-years for both periods. Results indicate that, the high resolution downscaling significantly outperforms the driving fields, with the negative bias of precipitation and cold bias of temperature both reduced. It also produces more realistic probability distributions for both precipitation and temperature. As for the climate extremes, the WRF model well produces both precipitation and temperature indices, and the high resolution downscaling shows better skill for all the indices except CWD(Consecutive Wet Days) and TXx(max Tmax). The projection under the RCP8.5 scenario shows that, the precipitation would decrease in summer and increase in spring, with decrease for annual mean. The drizzle rain with intensity of 0-1 mm·dwould occur more frequently in the future, while the probability of moderate to heavy rainfall would decrease. The more intensive storm would result in a significant increase of rainfall in very wet days(R95). In the future, the probability of high temperature above 25 ℃ would be higher while that below 0 ℃ would be lower, resulting in longer Warm Spell Duration Indicator(WSDI) while shorter Cold Spell Duration Indicator(CSDI). In addition, the extreme high and low temperature are both projected to be higher.