Model Uncertainty Representation for a Convection-Allowing Ensemble Prediction System Based on CNOP-P

Formulating model uncertainties for a convection-allowing ensemble prediction system(CAEPS) is a much more challenging problem compared to well-utilized approaches in synoptic weather forecasting. A new approach is proposed and tested through assuming that the model uncertainty should reasonably describe the fast nonlinear error growth of the convection-allowing model, due to the fast developing character and strong nonlinearity of convective events. The Conditional Nonlinear Optimal Perturbation related to Parameters(CNOP-P) is applied in this study. Also, an ensemble approach is adopted to solve the CNOP-P problem. By using five locally developed strong convective events that occurred in pre-rainy season of South China, the most sensitive parameters were detected based on CNOP-P, which resulted in the maximum variations in precipitation. A formulation of model uncertainty is designed by adding stochastic perturbations into these sensitive parameters. Through comparison ensemble experiments by using all the 13 heavy rainfall cases that occurred in the flood season of South China in 2017, the advantages of the CNOP-P-based method are examined and verified by comparing with the well-utilized stochastically perturbed physics tendencies(SPPT) scheme. The results indicate that the CNOP-P-based method has potential in improving the under-dispersive problem of the current CAEPS.     

全国网格化多模式集成空气质量预报的初步建立

为降低单个模式预报的不确定性和提高多模式集成空气质量预报系统的精细化程度,利用Cressman插值初步建立了我国0.25°X0.25°网格化污染物实况。结合4套空气质量数值预报模式,通过均值集成、权重集成和多元线性回归集成分别逐格点建立了集成预报。在预报当天各单一模式和集成方法前50 d预报效果评估基础上,建立了最优集成预报。对2018年12月19一22日一次重污染过程中集成预报的PM_(2.5)浓度评估结果显示:在污染较重时刻,最优集成预报与观测之间的归一化平均偏差(NMB)值在重污染地区保持在—20%～40%,对污染程度为良及以上区域的预报范围相较于单个模式更接近观测。整个过程中,最优集成在大部分污染区域与观测之间的NMB值为—20%～20%,均方根误差(RMSE)值为35～75μg·m~(-3),相关系数(R)值大于0.4。相较于所有单一模式和其他集成方法,最优集成在全国最多的格点有着较高的总体评分。在污染最重区域的8个城市,最优集成预报的污染过程平均开始和结束时间分别比观测时间早1.8和6.9 h。未来需融合卫星反演和地表观测来提高网格化污染物实况的精细化程度,利用降尺度、主客观融合和滚动订正等方法进一步提高网格化多模式集成空气质量预报的准确率。     

Assessing the Seasonal Predictability of Summer Precipitation over the Huaihe River Basin with Multiple APCC Models

Seasonal rainfall predictability over the Huaihe River basin is evaluated in this paper on the basis of 23-year(1981-2003) retrospective forecasts by 10 climate models from the Asia-Pacific Economic Cooperation(APEC) Climate Center(APCC) multi-model ensemble(MME) prediction system.It is found that the summer rainfall variance in this basin is largely internal,which leads to lower rainfall predictability for most individual climate models.By dividing the 10 models into three categories according to their sea surface temperature(SST) boundary conditions including observed,predicted,and persistent SSTs,the MME deterministic predictive skill of summer rainfall over Huaihe River basin is investigated.It is shown that the MME is effective for increasing the current seasonal forecast skill.Further analysis shows that the MME averaged over predicted SST models has the highest rainfall prediction skill,which is closely related to model’s capability in reproducing the observed dominant modes of the summer rainfall anomalies in Huaihe River basin.This result can be further ascribed to the fact that the predicted SST MME is the most effective model ensemble for capturing the relationship between the summer rainfall anomalies over Huaihe River basin and the SST anomalies(SSTAs) in equatorial oceans.     

3种再分析资料的高空温度与中国探空温度资料的对比:年平均特征

利用中国105站的探空资料以及NCEP/NCAR、ERA和JRA 3种再分析资料,采用线性趋势、标准差、相关系数、EOF分析等多种统计分析方法,对再分析资料年平均的高空温度在中国区域的可信度进行了分析.研究表明:在数值上,3种再分析资料的高空温度均小于探空资料的高空温度,NCEP/NCAR资料在对流层上层更接近于探空资料,ERA和JRA资料则在对流层中下层与探空资料更为接近;在描述年际变化和长期变化趋势方面,ERA资料在我国北方的对流层上层的再现能力较好,NCEP/NCAR资料在我国南方的对流层上层的再现能力较好,而3种再分析资料在对流层中下层的再现能力相当;在时空变化特征方面,NCEP/NCAR和ERA资料能较好地表现高空温度的年代际变化特征,而ERA和JRA资料则能较好地表现年际变化特征.     

中国气温与降水的时空变化趋势分析

如何对离散分布的气象台站观测数据进行高精度曲面模拟,为生态系统及服务功能时空变化趋势模拟及其综合评估提供高质量、高分辨率的空间气候数据,以满足栅格层次上的生态系统过程模型、生态系统格局模型及生态系统综合评估模型的参数需求,一直是存于生态学界的难点问题.在对全国1964一2007年的752个气象台站长期观测的气温和降水数...     

基于集合预报的四类持续性灾害性天气中期概率预报应用技术分析

针对4类持续性、灾害性天气,利用集合平均、离散度、概率以及盒须图等方法对T213集合预报的数据进行再加工,得到其中期概率预报,包括全国范围的持续性降水、南方春季持续低温（重度低温）、南方春季低温（重度低温）连阴雨、江淮梅雨。其结果表明：（1）对于全国范围内的持续性降水过程,各类图都能很好的反映集合预报中不同量级的降水预报在10天之内的落区变化以及持续情况,并与实况对应较好。（2）对于南方低温和低温连阴雨,集合平均、离散度和概率图都能明显的反映出实际天气过程和范围的变化。（3）对于江淮梅雨这5个指标站的盒须图,能明显看出各集合成员在15个集合成员中的相对位置,以及数据的分布情况。同时还能看出集合成员与集合平均数之间的关系。梅雨日概率图则能直观的反映出10天之内每天达到梅雨日标准的概率,更有利于判断梅雨天气．     

The Uncertainty of Mesoscale Numerical Prediction of Heavy Rain in South China and the Ensemble Simulations?

In the context of non-hydrostatic MM5 version we have explored the impact of convective parameterization schemes on uncertainty in mesoscale numerical prediction of South China heavy rain and mesoscale heavy rainfall short-range ensemble simulation by using two kinds of physics perturbation methods through a heavy rain case occurring on June 8, 1998 in Guangdong and Fujian Provinces. The results show the physical process of impacts of convective schemes on heavy rainfall is that different latent heat of convective condensation produced by different convective schemes can make local temperature perturbation, leading to the difference of local vertical speed by the intrinsic dynamic and thermodynamic processes of atmosphere,and therefore, making difference of the timing, locations and strength of mesh scale and subgrid scale precipitation later. New precipitations become the new source of latent heat and temperature perturbation,which finally make the dynamic and thermodynamic structures different in the simulations. Two kinds of methods are used to construct different model version stochastically. The first one is using different convective parameterization and planetary boundary layer schemes, the second is adjusting different parameters of convective trigger functions in Grell scheme. The results indicate that the first ensemble simulations can provide more uncertainty information of location and strength of heavy rainfall than the second. The single determinate predictions of heavy rain are unstable; physics ensemble predictions can reflect the uncertainty of heavy rain, provide more useful guidance and have higher application value.Physics ensembles suggest that model errors should be taken into consideration in the heavy rainfall ensembles. Although the method of using different parameters in Grell scheme could not produce good results, how to construct the perturbation model or adjust the parameter in one scheme according to the physical meaning of the parameter still needs further investigation. The limitation of the current study is that it is based on a single case and more cases will be addressed in the future researches.     

GJ 876行星系统中潜在类地行星研究

用数值积分方法模拟探讨了GJ876行星系统中存在潜在类地行星的可能性．结果表明分别产生于内外两颗巨行星的长期共振v_1和v_2可激发类地行星(其轨道半长轴分布在0．21AU≤a＜0．50AU之间)的偏心率,并使它们在极短时间内被抛射出系统．然而,从动力学角度而言,类地行星潜在存在区域可分布在0．50 AU≤a≤1．00 AU之间,其能在系统中10~5年时间尺度上稳定地存在．     

复杂背景下提高精度的眼睛定位方法

提出了一种在基于肤色的人脸检测中通过进行相似度计算来实现人眼特征提取和定位的算法.其目的是在较复杂的背景环境下对静态图片中的人脸进行检测,找出脸部区域并对该人脸进行眼睛定位.先进行肤色分割来完成人脸的区域检测,然后在检测出的人脸范围中利用人脸的几何特征进行快速的眼睛定位.在人脸区域检测过程中使用均值滤波提高人脸区域定位的准确度.实验结果证明,这种方法是准确的、可靠的和快速的.     

积分第二中值定理“中值点”渐近性探讨

本文研究了积分第二中值定理“中值点”的渐近性，得出了一般性的结果。