Evaluation of downscaled CMIP5 model skill in simulating daily maximum temperature over the southeastern United States
Evaluation of downscaled CMIP5 model skill in simulating daily maximum temperature over the southeastern United States
复制标题
DOI：
10.1002/joc.4612
复制
发表时间：
2016-10-01
期刊：
影响因子：
--
通讯作者：
Keellings, David
中科院分区：
文献类型：
--
作者：
Keellings, David
在美国东南部地区，对来自Downscaled CMIP3和CMIP5气候和水文预测档案的Downscaled CMIP5（耦合模式相互比较项目第5阶段）最高日温度气候预测进行了区域性研究。模型技能的三个措施（平均值为基础的，分布为基础的，极端为基础的）被用来评估15个降尺度模型模拟每日最高温度观测的能力。提出了一种新的测试来确定基于概率密度函数的技能测量的统计显著性。技能得分被认为是普遍较高的所有三个措施在整个研究区域，但较低的分数是存在于沿海和山区。显著性检验的应用表明，虽然技能分数可能很高，但在某些领域，它们并不显著高于随机预期。在佛罗里达和阿巴拉契亚山脉的大部分地区，基于分布的技能得分并不显著。基于极端的技能分数是不显着的，在超过90%的区域的所有模型的调查。研究结果表明，虽然缩小的模型很好地模拟了观测的平均值，并且与整个观测分布很好地匹配，但它们并没有模拟极端（高于第90百分位）日最高温度的发生。
Downscaled CMIP5 (Coupled Model Intercomparison Project Phase 5) climate projections of maximum daily temperature from the Downscaled CMIP3 and CMIP5 Climate and Hydrology Projections archive are examined regionally over the southeastern United States. Three measures of model skill (means-based, distribution-based, extreme-based) are utilized to assess the ability of 15 downscaled models to simulate daily maximum temperature observations. A new test is proposed to determine statistical significance of the probability density function-based skill measures. Skill scores are found to be generally high for all three measures throughout the study region, but lower scores are present in coastal and mountainous areas. Application of the significance test shows that while the skill scores may be high, they are not significantly higher than could be expected at random in some areas. The distribution-based skill scores are not significant in much of Florida and the Appalachians. The extreme-based skill scores are not significant in more than 90% of the region for all models investigated. The findings suggest that although the downscaled models have simulated observed means well and are a good match to the entire distribution of observations, they are not simulating the occurrence of extreme (above 90th percentile) maximum daily temperatures.