灾害天气
Severe Weather

灾害天气研究进展

1 灾害天气监测与资料应用

1.1 雨滴谱数据在双线偏振雷达雨滴谱反演及降水估测中的应用

利用双线偏振雷达在广东鹤山（2014年5月1日至6月15日）参加华南前汛期暴雨外场观测试验，获得大量的暴雨、大暴雨野外观测资料，对资料进行了质量控制。利用雨滴谱观测数据分析了华南前汛期滴谱结构，重新拟合偏振参量进行降水估测，提高了偏振雷达定量降水估测精度；利用双线偏振雷达在西藏那曲（7月1日至8月31日）开展的第3次青藏高原科学考察野外观测试验数据，研究了高原对流云的宏、微观特征。（胡志群）

1.2 C波段调频连续波雷达降水云结构研究及降水参数反演

C-FMCW雷达可以描述降水云的精细垂直结构和快速演变特征。利用C-FMCW雷达参加江淮梅雨锋降水观测试验、华南季风强降水外场试验和青藏高原那曲局地对流云的高原观测试验所获得的资料，开展了降水云垂直结构分析和降水参数反演研究。回波谱参数和谱分布的协同分析，有助于深入认识云中的水汽与动力过程，上升和下降运动的强烈变化在回波强度谱分布中的表现更加清晰，在弱回波强度区中发现了高空强的大气下沉速度；反演的广东阳江降水云中雨滴谱分布与地面雨滴谱仪观测数据一致性较好（图1～2）。（阮征）

1.3 联合多种观测设备在多个科研项目上开展外场试验

完成行业专项“第3次青藏高原科学试验——边界层和对流层观测”中的青藏高原云降水预观测试验、华南季风强降水外场试验和973项目“高原东移降水系统观测”等外场试验，获取了云降水微物理结构和动力结构数据。主要创新点：在国内首次利用云雷达、微降水雷达和水汽云探测激光雷达等云观测设备，以及X波段相控阵天气雷达和C波段双线偏振雷达的强对流观测设备，在青藏高原和高原东坡实现了云垂直结构和对流系统中尺度结构的观测。以此为基础，初步形成了多种遥感手段反演高原对流系统空气上升速度和云水/云冰/雨水/雪水的垂直廓线的反演方法，验证了国产新型探测设备在高原工作的能力，为青藏高原云和降水的研究提供了数据（图3）。（刘黎平）

1.4 我国持续性强降水的动力预报理论与方法和可预报性研究

构建了基于集合预报、全球海陆气耦合模式和高分辨率区域模式嵌套的持续性强降水动力预报原型系统；开展了全球海陆气耦合模式和同化系统研究，形成了全球海陆气耦合预报与同化一体化系统，并进行了初步的数值试验。结果显示，对于20天内预报，相对于使用FNL直接插值作为初值的确定性预报，使用CESM-DART预报同化系统的输出作为初值的集合预报在北半球5天后的预报效果较好。开展了现有业务模式对持续性强降水预报能力的评估；研究了高分辨率区域模式的持续性强降水预报方法与技术，表明谱逼近方法对降水的量级、落区以及连续性都有较好的预报，2波及其以上波段的逼近效果更好，其中对降水TS评分的提高尤其体现在中雨及其以上量级的预报。（王东海）

1.5 东亚云水特征及其在参数化中的应用

利用1980—1982年北方5省（新疆，内蒙古，陕西，宁夏，吉林）外场飞机观测资料，探讨了云中云水含量与云滴浓度的关系，并将该统计关系用于云微物理参数化方案中云滴数浓度的约束。利用2006—2011年CloudSat卫星观测数据统计分析了东亚地区云垂直结构特征，并基于统计结果改进了Kessler型云雨自动转换过程参数化的表达。利用CALIPSO-GOCCP数据统计分析了东亚地区3D云垂直结构特征。（尹金方，王东海）

2 灾害天气结构与形成机理

2.1 第3次青藏高原大气科学试验

全面开展了2014年度观测任务，进行了加密探空、边界层和雷达观测试验数据的质量控制。初步完成了雷达径向风和反射率、卫星产品、风廓线雷达资料、GPS/MET水汽资料的同化。研究了高原地区不同下垫面地气能量交换特征、辐射收支和水热交换变化特征，以及感热、潜热通量变化特征。研究了青藏高原地区云的时空变化特征，完善了混合相态的云微物理过程，改进了目前双参数云物理方案中的气溶胶活化方案；研究了青藏高原及周边地区对流系统的变化特征，对比了青藏高原低涡的东移机制。分析了高原影响下中尺度气旋在向淮河流域移动过程中引发暴雨过程的机制；揭示了青藏高原大气加热场异常变化对我国旱涝的影响，分析了年际尺度高原异常与我国西南地区5、6月降水的相关关系。（徐祥德）

2.2 青藏高原对梅雨区水分循环及降水变异的影响

进行了梅雨系统上游关键区青藏高原及东缘数据再分析与同化试验研究；开展了青藏高原热力特征变化及其对季风梅雨区环流系统影响的诊断分析、青藏高原热力特征变化对中低纬水汽交换及梅雨区大气水分循环过程的影响研究、青藏高原热力特征变化对梅雨带异常及其旱涝灾害分布影响机理研究，以及青藏高原热力特征变化对梅雨区大气水分循环及降水异常影响的模拟试验研究。（徐祥德）

2.3 不同天气系统特征和变化规律及其对气候变化的响应

基于重污染天气（重点霾天气）时空变化特征，采用历史相关各类气象资料，剖析我国重污染天气大范围时空变化及其对气候变化的年代际响应，探讨大地形对我国重污染天气季节特征及其时空分布变化规律的影响，归纳出重污染天气过程预警业务应用的天气背景各类模型。（徐祥德）

2.4 WMO/WWRP RDP“华南季风降水试验”（SCMREX）

2014年4月3—6日，进行了WMO/WWRP RDP“华南季风降水试验”（SCMREX）外场试验仪器设备的选址工作。5月1日至6月15日，在业务观测网基础上开展了加密观测，暴雨发生期间对移动观测设备进行24 h监控。探空加密观测从2014年5月1日08：00开始，2014年6月15日08：00结束，华南3省的7个高空观测站每站每日增加了14：00和02：00的观测。观测试验之后，收集了移动设备观测数据和华南地区业务数据，并进行了质量控制，完成了项目网站与数据库建设与维护，观测资料已经初步整理并在项目网站上进行共享。针对SCMREX 2014和2013年观测试验期间的暴雨过程，开展了多尺度机理研究以及资料同化影响、模式物理过程参数化方案等相关数值天气预报试验和研究，取得了初步的成果。2014年2次向WMO汇报SCMREX项目进展，得到肯定。（罗亚丽）

2.5 华南暖区暴雨发生发展机制及关键预报技术研究

把具有一定区域意义的暖区暴雨的天气形势分为锋面类、暖湿急流北上类和低涡东移动类3种类型。给出3种类型暖区暴雨的统计特征及其差异。基于综合分析几个华南暖区暴雨个例，给出暖区对流触发与组织过程与冷池传播、地形等的可能关系。推导出适合于华南暖区暴雨的不稳定判据，修正了Richardson数和Brunt-Vaisala频率。个例研究表明，修正后的Richardson数和Brunt-Vaisala频率表征的不稳定区较之以前的不稳定参数与未来雨区有较好的对应。（姜智娜）

2.6 基于EnKF资料同化方法的华南前汛期暖区暴雨集合数值模拟研究

华南前汛期暖区暴雨的定量降水预报（QPF）水平很低，应用集合卡尔曼滤波（EnKF）资料同化技术可以改善模式初始条件，从而提高定量降水预报水平。利用WRF-EnKF系统开展集合预报的模拟试验，研究同化地面资料、探空资料和多普勒雷达径向风资料对模式初值和模拟结果的影响；将集合模拟与实际观测分析相结合，研究对流系统的结构和生消演变机制，重点关注对流触发和维持机制、对流系统的组织模态、热力和动力特征及其与中尺度环境大气的相互作用。个例研究表明，利用WRFEnKF系统同化常规探空资料，显著改善了数值预报的初始场，减小了各物理量的预报偏差和预报均方根误差，进而提高了暴雨过程的降水落区和强度的预报准确率。（宝兴华）

2.7 高污染区大气冰核活化参数化方法及对强降水的影响

利用华北（北京及周边）地区1960—2013年的外场冰核观测资料，分析冰核浓度和活化谱特征，结合当前已有的多种冰核活化参数化方法，提出高浓度冰核背景条件下冰核的活化新方法。将新方法耦合到WRF-SBM分档模式中，利用改进后的模式，以华北地区常规观测、地基雷达观测、A-Train系列卫星探测、风云（FY）卫星观测等资料的实例作为研究对象，进行精细模拟，定量分析强降水过程中不同相态水物质之间相互转换，揭示冰核对强降水粒子群的影响途径和量级，以便改进数值模式中冰核活化的参数化描述方法，提高数值模式对高污染区强降水的预报能力。（尹金方，王东海）

2.8 我国陆面水体对台风残涡维持及其降水的影响

内陆水体与陆面其他种类特征差异显著，其热量通量分布不均性更加明显，导致台风中尺度对流系统活动以及降水强度和落区差异。富含水汽的台风下垫面（鄱阳湖及其周围湿地）具有较高的感热和潜热通量，可为台风维持提供水汽供应，从而也有利于降水。敏感性试验发现，陆面水体移除（扩大）时可产生正（负）气压扰动，升高（降低）台风海平面气压，即陆面水体对陆上台风残涡维持有积极作用。宽广水面表层摩擦较小，低层风加大，从而改变低层风场辐合的强度和位置，影响降水强度及其落区。水体在夜间作为一个热源，比陆地土壤释放更多的表层水热通量，从而加热加湿低层大气，是降水的一个能量源。（李英）

2.9 孟加拉湾热带气旋活动及其对我国天气的影响

对中国降水与孟加拉湾风暴活动的相关性进行了分析，并对风暴活动期间的降水进行EOF分解。发现二者有较好的相关性，其通过0.01的相关显著性检验的高相关系数区位于中国西南地区、西藏东南部、西北地区东南部和山西南部，长江下游地区和台湾部分地区也有较小范围的高相关系数区。降水量多少与风暴活动日数有较好的对应。风暴活动主要并同时影响中国西南地区和长江下游地区降水，且对秋季降水的影响比春季更显著。（李英）

2.10 西北太平洋高空冷涡对台风路径突变影响的机理研究

统计分析发现，当台风与冷涡中心距离小于5经（纬）距时，其未来12 h移向转角明显大于气候平均值，而且台风在冷涡影响下平均移速有所减缓，尤其是移向突变的台风。台风“莫兰蒂”（2010年）即为在冷涡影响下发生突然北折现象的典型个例。诊断分析表明，高空冷涡自台风中心东侧向西北方向移动过程中改变了“莫兰蒂”的高空引导气流的方向，是其路径突变的一个重要原因（图4）。（李英）

2.11 超强台风“丹娜丝”对2013年第23号强台风“菲特”极端降水的作用

利用地面观测资料、台风定位资料、ECMWF全球再分析资料等，采用热带气旋（TC）降水天气图客观识别法（OSAT）、TC路径相似面积指数（TSAI）和气流轨迹模式HYSPLIT等方法，从2013年第23号强台风“菲特”在我国东南沿海引发台风暴雨的极端性分析及其成因诊断入手，揭示了双台风作用对极端暴雨的增强作用。结果表明：首先，强台风“菲特”给浙江带来了1958年以来单站日降水排名第2的极端降水，余姚和奉化日降水量均为395.6 mm；“菲特”降水过程有2个明显的强降水阶段；其次，秋季强台风“菲特”登陆后之所以出现如此强度且持续的台风暴雨，与超强台风“丹娜丝”的存在密不可分。在强降水第1阶段，双台风作用增强了降水的极端性，台风“丹娜丝”向降水区域输送了约79%的水汽，对杭州湾南侧的强降水过程有重要贡献；在台风“菲特”强降水第2阶段，“菲特”的环流已经基本消散，超强台风“丹娜丝”与冷空气的共同作用主导了这一阶段强降水的发生。（任福民）

2.12 孟加拉湾和阿拉伯海热带气旋活动双峰型差异及可能成因

针对北印度洋热带气旋（TC）研究中存在的TC资料对比不足和TC活动双峰型特征的区域性差异分析不足的现状，采用联合台风警报中心（JTWC）TC资料和印度气象局（IMD）TC资料开展了针对性的研究。结果如下：IMD资料的时段（1990—2012年）较短且资料记录时刻规律性差；而JTWC资料自1977年突变之后时段仍明显长于IMD资料且资料记录时刻稳定，故采用1977—2012年JTWC资料进行TC气候特征分析。1977—2012年，孟加拉湾TC频数呈下降趋势，而阿拉伯海TC频数呈显著增多趋势；二者多年平均TC频数分别为3.6个和1.5个；两海区TC频数的季节变化都表现为双峰型，但在双峰型的峰值时间、强度以及双峰的强弱配置上存在较大差异。研究表明，纬向风垂直切变和相对涡度的季节变化可能分别是影响孟加拉湾和阿拉伯海TC活动双峰型的关键因子；就整个北印度洋而言，南亚夏季风开始前和结束后，风垂直切变维持在10 m/s及以下、正的相对涡度、较高的海表温度、较大的相对湿度都是TC生成的有利条件。（任福民）

2.13 热带气旋初始结构对快速加强速率的影响

利用多层嵌套、可压缩、非静力热带气旋数值模式研究了不同的涡旋初始结构对热带气旋的快速加强过程的影响，涡旋的初始结构主要包括通过设置不同大风半径来描述的涡旋的大小以及用不同初始最大风速径向衰减曲率来表示的涡旋的不同的眼墙宽度。试验结果表明，当初始涡旋切向风自最大风半径沿径向迅速衰减，同时向台风中心平缓衰减时，热带气旋快速加强明显。反之，当初始涡旋切向风自最大风半径沿径向衰减缓慢时，同时向台风中心衰减迅速时，热带气旋快速加强受到抑制。另外，初始涡旋尺度的敏感性试验表明，初始涡旋尺度大不利于热带气旋快速加强，而小的初始涡旋有利于热带气旋的快速加强。（徐晶）

2.14 海南岛海陆风演变特征及其引发的中尺度对流天气预报技术研究

统计分析了2013年海南岛海陆风演变规律的年变化、月变化和日变化特征及其与温度场和降水场的关系。分析显示海南岛各个月份的海陆风特征明显，海风一般发生在上午11：00至晚上20：00之间，11：00附近和20：00附近为海风与陆风的转换阶段，20：00以后至次日11：00前为陆风明显的阶段。海陆风演变造成的辐散关系与降水分布关系密切，海陆温差分布特征是造成海陆风演变的主要原因。海陆风的辐合为降水提供了很好的抬升机制，但需要水汽配合方能发生明显的降水过程，地面相对湿度很好地反映了水汽条件的状况。（梁钊明，王东海）

2.15 我国东部小时降水的特征分析

利用中国区域1951—2013年6月1日至8月31日2420个站点逐小时降水资料，经过质量控制、无资料剔除等处理，建立了我国105°E以东区域内1980—2012年6—8月无缺测记录的1489个站点的逐小时降水资料序列。分析了夏季年平均小时降水频次、雨强和小时极端强降水频次、雨强的分布特征；夏季降水频次的变化趋势；夏季和夏季分月雨强的趋势变化；夏季和各月极端强降水频次的趋势变化；夏季和各月极端强降水雨强的趋势变化。结果表明，我国东部夏季小时降水频次整体以减少趋势为主，小时降水平均雨强以增加趋势为主，小时极端强降水频次整体呈减少趋势，极端强降水雨强整体呈增加趋势（图5）。（赵琳娜）

3 数值天气预报和模式关键技术

3.1 高分辨模式物理过程研发

分析检验了CAMS云微物理方案在青藏高原地区雨滴谱截距偏小导致雨滴偏大和雷达反射率偏大的问题；利用飞机观测资料总结了云水物质含量与云滴浓度的关系，并采用卫星遥感资料分析了东亚地区云垂直结构，改进了云雨自动转化参数化过程，从而有效改进了CAMS云微物理方案对我国高原和华南地区降水的模拟效果。利用观测资料，对半干旱区地表交换系数检验修正，改善了交换系数和地表冠层阻尼参数化，改善了地表热通量计算（图6）。（高文华，尹金方，张果）

3.2 GRAPES模式准均匀网格技术研究

采用Schwarz方法在阴阳网格第1类边界条件强迫条件下首次成功实施了对阴阳网格非静力大气动力模式的半隐式半拉格朗日求解，实现了GRAPES阴阳网格模式动力框架的稳定积分；采用ILU预条件处理提高了阴阳网格GCR法求解Helmholtz方程的效率，加速了数值解的收敛速度。成功确认了准均匀网格GRAPES模式的二阶精度算法改进和长期积分试验，3D动力框架在2个月时间积分中保持很好的稳定性和计算精度，为下一代高分辨率GRAPES模式发展提供了一个解决方案（图7）。（彭新东）

3.3 降水短期预报的客观订正

图1 2013年8月24日安徽蚌埠SA雷达和定远C-FMCW雷达探测产品(a：SA雷达回波强度；b：C-FMCW雷达谱参数时序图；c：C-FMCW回波强度谱密度分布)Fig. 1 Images of SA radar in Bengbu and C-FMCW radar in Dingyuan on August 24, 2013 (a: ref ectivity of SA Radar; b: timeheight section of C-FMCW radar; c: spectral density of ref ectivity of C-FMCW radar)

利用高分辨率降水分析资料和距平积分数值订正（ANO）方法，对2013年7月四川暴雨中尺度模式环流预报场和高分辨降水模拟进行了订正试验。结果表明，订正对高分辨大气环流场有明显的改进，明显改善了近地层大气的日周期环流特征预报，位势高度、温度和比湿分别在对流层中层550 hPa和低层750 hPa改进尤为显著，同时确认了高分辨WRF中尺度模式的短期降水预报的改善（图8）。（彭新东）

图2 广东阳江C-FMCW雷达反演(120 m)雨滴谱与地面雨滴谱分布比对(左图：1 min平均；右图：10 min平均)Fig. 2 Raindrop size distributions at 120 m retrieved from C-FMCW radar in comparison with Parsivel disdrometer on surface (Yangjiang, Guangdong Province, May 8, 2014. left: average_ 1 min; right: average_10 min )

图3 青藏高原云降水外场试验观测主要设备Fig. 3 The main equipment used in the experiment in the Tibetan Plateau

图4 台风移向12 h转折角度(曲线)及台风频数(柱状)随台风与冷涡中心相对距离的变化(a)、台风“莫兰蒂”地面中心(空心圆)和高空冷涡(实心圆)的路径(b)Fig. 4 (a) Variation of average 12 h directional change of TC movement (red solid line) and accumulative frequency (blue bar) with the relative distance between TC and UTCL (The red dash line indicates the western North Pacif c climatology). (b) The tracks of typhoon Meranti (empty circle) and UTCL (dot)

图5 我国东部夏季及夏季各月极端强降水雨强的变化趋势空间分布(单位：%/10a):(a)6月；(b)7月；(c)8月Fig. 5 The spatial distribution features of the extreme precipitation intensity in summer over the east of China (Unit: %/10a): (a) June; (b) July; (c) August

图6 2013年5月8日TRMM/TMI观测降水(中)和CAMS云微物理方案改进前(左)后(右)模拟降水Fig. 6 The observed (middle) and WRF-CAMS microphysics simulated (left, original; right, improved) 24 h accumulated precipitation during 00:00 to 24:00 (UTC) on 8 May 2013

图7 Haurwitz-Rossby波传播第14天位势高度(a)、地面气压(b)、u风分量(c)和v风分量(d)Fig. 7 The Haurwitz-Rossby wave propagation on day 14: (a) geopotential height; (b) surface pressure; (c) u component; (d) v component

图8 WRF模式预报(ORI)和ANO订正后的暴雨(a)和大雨(b)预报的ETS评分Fig. 8 Torrential (a) and heavy (b) rainfall forecasting of the original WRF model prediction (ORI) and ANO correction

Advances in Research on Severe Weather

1 Severe weather monitoring technology and data application

1.1 Application of rain drop size distribution (DSD) data to DSD retrieval and precipitation estimation with a dual polarization radar

Heavy rain data were observed with the dual linear polarization radar and disdrometer from May 1 to June 15, 2014 in Heshan, Guangdong Province during pre-rainy season in South China. The dual polarization radar data were controlled in quality before usage. By using the data observed with the disdrometers, the drop size distribution (DSD) of the pre-rainy season in South China was analyzed, and the algorithm of radar quantitative precipitation estimation (QPE) by polarimetric parameters was ref tted to improve the accuracy of QPE. The DSD was also retrieved by a dual linear polarization radar in Naqu Tibet during the third Tibetan Plateau Atmospheric Science Experiment (from July 1 to August 31, 2014), which was compared with distrometers. (Hu Zhiqun)

1.2 Micro-physics parameters retrieving vertical structure in the precipitation cloud using a C-FMCW radar

The C-FMCW radar data were used in such experiments as the Meiyu seasonal precipitation in Changjiang-Huaihe River valley, the South China sea monsoon onset and the precipitation cloud structure at the main region of Tibetan Plateau. Data from other active and passive remote sensing systems were used to research the microphysical parameters. The rain parameters were used to retrieve detailed vertical structures and raindrop size distributions in precipitation cloud (Fig.1–2). (Ruan Zheng)

1.3 Various kinds of equipment were employed in f eld observation experiments for many scientif c projects

The Third Tibetan Plateau Atmospheric Science Experiment, heavy rainfall experiment in Huanan for RDP project and precipitation system observation from Tibetan Plateau in 973 Project were carried out and comprehensive measurements of water vapor, clouds, and precipitation were conducted. The most advanced radars in China, such as Ka-Band millimeter-wave cloud, Ku-Band micro-rain, C-Band continuous-wave and lidar, and microwave radiometer and disdrometer were deployed to observe high spatial-temporal resolution vertical structures of clouds and precipitation. Based on the radar measurements in this experiment, the air vertical draft and content prof les for water cloud, ice cloud, rain and snow were retrieved. The observation capabilities of equipment made in China were verif ed. The above measurements and preliminary analyses provide a basis for further in-depth study of cloud physics and precipitation processes in the Tibetan Plateau (Fig.3). (Liu Liping)

1.4 Forecasting theory and method and predictability of persistent extreme precipitation events in China

A dynamic prototype forecast system for persistent extreme precipitation is constructed based on the ensemble forecast technique, globally coupled oceanic-land-atmospheric model and nested high-resolution regional model. An integrated system composed of globally coupled forecast and global assimilation is developed and used to perform preliminary numerical experiments. The results show that for 20-day forecast, compared with deterministic forecasts, which use the interpolated FNL analysis as initial conditions, the ensemble forecasts, which use the output of the CESM-DART forecast-assimilation system as initial conditions, exhibit better forcast skills after 5 days for Northern Hemisphere. The prediction ability of the operational model for persistent extreme precipitation is verified. The forecasting theory and method of persistent extreme precipitation events by using the high-resolution regional model are studied. The results show that the numerical simulations by using the spectral nudging and filtering methods have improved obviously precipitation forecasting for magnitude, rain band and continuity, the better nudging effect ref ects above 2-wave bands, and, the TS scores also display an improvement in the precipitation rate categories above light rain. (Wang Donghai)

1.5 The cloud-precipitation characteristics and their application in cloud microphysical parameterization

A statistical analysis of cloud microphysical properties was performed based on the in-situ observations from the North China Cloud Physics Detection Project (NCCPDP) during the period from 1980 to 1982. From the statistical results, the relationship between liquid water content (LWC) and cloud number concentration (NC) was investigated. The vertical distributions of LWC in precipitating and non-precipitating clouds were analyzed based on a large number of CloudSat observations over Asian land areas between June 2006 and April 2011. The results of this analysis were used to propose a new def nition of the threshold value for Kesslertype parameterization of warm cloud autoconversion. The spatial distribution of clouds and their seasonal variations, and the three-dimensional (3D) cloud structures over East Asia were analyzed with the CALIPSOGOCCP data during the period from 2007 to 2012. (Yin Jinfang, Wang Donghai)

2 Structures and mechanisms of severe weather

2.1 The 3rd Tibetan Plateau Atmospheric Science Experiment (TIPEX III)

During the period of TIPEX III in 2014, all the following actions were conducted: (1) the observation was executed comprehensively, in which the quality control for intensive sounding, boundary layer and radar observation data was carried out, and the assimilation of observations from radar radial velocity and ref ectivity, satellite products, wind prof ler radar data, GPS/MET water vapor was completed initially; (2) the exchange characteristics between the ground and the atmosphere, the change in radiation budget and waterheat exchange, and f ux change in sensible and latent heats underlying surface on the plateau were researched; (3) the spatial and temporal variation of the cloud in the Tibetan Plateau was investigated, in which the cloud’s microphysical processes in mixed phases were refined, and the aerosol activation scheme in existing dualparameter cloud’s physical processes was improved; (4) the change characteristics of a convective system over the Tibetan Plateau and its neighboring areas were studied, in which the mechanism of eastward movement of low vortex in Tibetan Plateau was detected, and the mechanism of storms triggered by the mesoscale cyclones when moving eastward to Huaihe River Basin was analyzed; (5) the inf uence of abnormal atmospheric heating f eld in Tibetan Plateau on f ood and drought in China was revealed, in which the relationship between the interannual-scale anomalies in plateau and precipitation in May and June in Southwest China was analyzed. (Xu Xiangde)

2.2 The effect of Qinghai-Xizang Plateau on water cycle and precipitation variation in Plum Rain region

We have made a reanalysis and assimilation experiment for the data of the Qinghai-Xizang Plateau and its eastern margin, the key upstream areas for the Plum Rain system, a diagnostic analysis of the thermodynamic variation characteristics in the Qinghai-Xizang Plateau region and a study of its impact on the circulation system in the monsoon and Plum Rain region, a study of the influence of the thermodynamic variation in the Qinghai-Xizang Plateau on the water vapor exchange between low and middle latitude areas, as well as atmospheric water circulation processes in the Plum Rain area, a study of the mechanism on the influence of the thermodynamic variation in the Qinghai-Xizang Plateau on the anomalies of the Plum Rain zone and the distribution of flood and drought disasters, and a simulated experimental study of the influence of the thermodynamic variation in the Qinghai-Xizang Plateau on the anomalies of atmospheric water cycle and precipitation in the Plum Rain area. (Xu Xiangde)

2.3 Characteristics and evolution of different weather systems and their response to climate change

Building on the spatial and temporal characteristics of the seriously-polluted weather (particularly the haze) derived from all kinds of historical meteorological data associated, we described the large-scale temporal and spatial variation of seriously-polluted weather and its decadal response to the climate change in China, and explored the effects of large scale topography on the seasonal evolution of seriously-polluted weather and its spatial and temporal distribution. At last, we established diverse synoptic background models for warning application during the formation of seriously-polluted events. (Xu Xiangde)

2.4 WMO/WWRP RDP South China Monsoon Rainfall Experiment (SCMREX)

During 3–6 April 2014, the site selection of the field experiment instruments and equipment was completed. Based on the conventional observation network, intensive observations were made during May 1–June 15. The mobile observation equipment was under the 24-hour monitoring during periods of heavy rainfall. Intensive sounding observations started at 8 BST on May 1, 2014 and ended at 8 BST on June 15, 2014. 7 upper-air stations in the three South China provinces made additional observations at 14 BST and 02 BST per day. After the observation experiment, not only the conventional observation data but also the mobile equipment observation data were collected and controlled in quality. Also, the development and maintenance of the project website and database were achieved. After being primarily processed, the observation data were shared on the project website. Focusing on these heavy rainfall cases during 2013–2014 SCMREX experiment, we studied the multiple scales mechanism, the impact of data assimilation, and the difference of various model physics Parameterization Schemes, with initial results achieved. We reported SCMREX progress twice in 2014, which was recognized by WMO. (Luo Yali)

2.5 Mechanism and forecast skill of warm sector heavy rainfall in South China

According to weather regimes, warm sector rainfall is divided into three types: (1) Frontal rainfall, (2) Northward motion of warm and moisture jet, (3) Eastward motion of low vortex. The above three types of warm-sector rainfall are analyzed and compared. A case study pointed out that the triggered convection in a warm sector may be closely related to its organization process, the cold cell spreading and the terrain. Besides, the instability criterion f t for the warm-sector rainfall is derived, e.g. the modif ed Richardson number and Brunt-Vaisala frequency. The study shows that the instable region revealed by the modif ed Richardson number and Brunt-Vaisala frequency can well represent the future rainfall region. (Jiang Zhina)

2.6 On the warm-sector heavy rainfall during the early summer rainy season over South China: An EnKF data assimilation based simulation study

The Quantitative precipitation forecast (QPF) skill for the warm-sector heavy rainfall over South China during the early summer rainy season is very low. The application of the Ensemble Kalman Filter (EnKF) data assimilation technique can potentially improve the initial conditions of simulation and thus improve the QPF skill. On the one hand, ensemble simulation experiments will be conducted using the WRF-EnKF system. Impacts of a variety of observational data, including not only the surface and sounding observations but also the Doppler radar velocity, on the initial conditions and the simulation results will be investigated. On the other hand, combining the ensemble simulation experiments with the observational analysis, we will study internal structures and organizational modes of the rainy storm, and inf uencing factors for the convective interactions between the rainy storm and its mesoscale environments. One case study results show that, compared to no data assimilation experiment (NODA), not only the initial conditions of EnKF data assimilation experiment (DA) are much closer to the observed f elds, but also the DA predicted physical parameters are improved in terms of both biases and root-mean-square errors, which leads to a more accurate prediction of the location and magnitude of precipitation from DA. (Bao Xinghua)

2.7 Study of the ice nuclei activation parameterization and its influence on severe precipitation under the conditions of high pollution

An attempt was made to delineate the characteristics of ice nuclei (IN) over the eastern central China using the ground-based measurement data over 1960–2013. Based on the statistical results, a new parameterization was proposed for ice nuclei activation within high ice nuclei conditions. Based on the ice nuclei field observational data over the last five decades and the widely used ice nuclei activation parameterizations in numerical models, the new scheme will be coupled into the WRF model with Spectral Bin Microphysics (WRF-SBM). Based on the improved model, we will focus on the conversion among the hydrometeors (vapor, cloud water, rain water, cloud ice, snow, graupel, and hail) in a severe precipitation event which has been captured by ground-based radars, and satellites of A-train (mainly including CloudSat and CALIPSO) and Fengyun (FY). The results will be used to identify the effects of high ice nuclei concentration on the development and formation of a severe precipitation. Based on the results, the parameterization for ice nuclei activation will be improved to improve severe precipitation forecasting within polluted conditions by a numerical model. (Yin Jinfang, Wang Donghai)

2.8 The impact of land surface water on the inland behavior of landfalling typhoon

Inland water surface has remarkable differences in heat property and roughness from other land surface covers. It could lead to the uneven distribution of surface f ux, which inf uences the mesoscale systems and rainfall of landfalling typhoon significantly. That is, inland water surface (such as Poyang Lake and wet land around it) could release heat f uxes to the boundary layer of typhoon, providing favorable condition for typhoon maintenance and rainfall. Sensitivity experiments on Poyang Lake demonstrate that a mesoscale vortex would be weakened in the absence of the lake to lead to the increase of sea level pressure in typhoon center. In addition, broader water bodies have less friction, which increases wind speed near the underground to enhance convergence over heavy rainfall area on the one hand. On the other hand, water bodies may release more surface heat f uxes to make air warmer and wetter in lower atmosphere, which is benef cial to typhoon heavy rainfall. (Li Ying)

2.9 The impact of cyclonic storms over the Bay of Bengal on the weather of China

A correlation analysis is used to investigate the relationship between rainfall of China and activities of cyclonic storms over the Bay of Bengal. Results show that the high correlation coeff cient areas are mainly found in the Southwest China, southeast of Tibet plateau, southeast of Northwest China, the lower reaches of Yangtze River and Taiwan Island. Storm activities can inf uence rainfall of the southwest area and the lower Yangtze region concurrently，and the impact is more signif cant in autumn than in summer. (Li Ying)

2.10 The impact of upper tropospheric cold low on abrupt turning of tropical cyclone movement

Statistical study indicates that when the relative distance between a tropical cyclone (TC) and an upper tropospheric cold low (UTCL) is less than 5 latitudes/longitudes, the average 12 h directional change of TC movement is obviously larger than the average western North Pacif c climatology. In addition, the TC average moving speed slows down under the impact of UTCL, especially that of the TCs with abrupt turning. Meranti (2010) is a typical typhoon which turned abruptly under the inf uence of an UTCL. A diagnostic study indicates that the UTCL changes the steering f ow when it is moving from the east of the central TC to the northwest, which is an important reason for abrupt turning of Meranti (Fig. 4). (Li Ying)

2.11 The role of super typhoon Danas in extreme precipitation associated with severe typhoon Fitow

Based on surface observational data, typhoon track data and ECMWF global reanalysis data, using the objective synoptic analysis technique (OSAT), tropical cyclone (TC) track similarity area index (TSAI), and the airf ow trajectory model (HYSPLIT4.9) and analyzing the characteristics and causes of the extreme precipitation over coastal Southeast China associated with the sever typhoon Fitow, this study reveals the intensification by a binary typhoon of an extreme precipitation. First, Fitow causes the maximum daily precipitation of 395.6 mm at both Yuyao and Fenghua, which rank the second most extreme daily TC precipitation in Zhejiang Province in record. The precipitation process has two distinct intense precipitation stages. Secondly, such a heavy and continuous rainfall during and after Fitow’s landfall is mainly due to the existence of super typhoon Danas. In the f rst stage, because of the binary tropical cyclone interactions, Fitow moves much faster than before. Moreover, Danas transports about 79% moisture to the raining region, which is an important contribution to the extreme precipitation over the southern coast of Hangzhou Bay. In the second stage, as the circulation of typhoon Fitow almost dissipates, the combined interaction of super typhoon Danas and cold air mainly causes the extreme precipitation. (Ren Fumin)

2.12 Differences of bimode patterns in TC activity and their possible causes in the Bay of Bengal and the Arabian Sea

To address the inadequate tropical cyclone datasets for comparison research purpose and to analyze the regional differences in TC bimodal patterns in the North Indian Ocean, a targeted study on these issues has been carried out using the Joint Typhoon Warning Center (JTWC) TC dataset and the India Meteorological Department (IMD) TC dataset. Results are as follows: time period (1990–2012) of IMD dataset is much shorter and data recording time is irregular, while time period (1977–2012) of JTWC dataset after the abrupt change in 1977 is much longer and data recording time is regular. Thus, JTWC dataset of 1977–2012 is adopted in TC climatic characteristics analysis. During 1977–2012, the Bay of Bengal (BoB) TC frequency decreases, while the Arabian Sea (AS) TC frequency increases signif cantly. Average TC frequencies are 3.6 and 1.5 in BoB and AS, respectively. Meanwhile, seasonal variations in TC frequency show a bimodal pattern with two peaks in both BoB and AS but with obvious differences in bimodal peak time, intensity and strength. It is revealed that vertical zonal wind shear and seasonal changes in relative vorticity might be the key factors affecting the bimodal patterns of TC activity in BoB and AS, respectively. Meanwhile, for the whole North Indian Ocean, before the onset and after the f nish of South Asian Summer Monsoon, below 10 m/s vertical zonal wind shear, positive relative vorticity, high sea surface temperature and high humidity are all favorable conditions for TC genesis. (Ren Fumin)

2.13 Sensitivity of tropical cyclone rapid intensif cation to the initial vortex

The multiply nested, fully compressible, non-hydrostatic tropical cyclone model is used to examine and understand the sensitivity of the simulated tropical cyclone (TC) rapid intensif cation (RI) to its initial vortex structure, including the radial prof le of tangential wind and the size of the vortex. The results show that when the initial vortex is with a broader prof le outside but a narrower prof le inside the eyewall, the slower (faster) RI is in the simulated storms. The results also show that the larger and broader (smaller and narrower) the initial vortex is, the slower (faster) the RI is but the larger (smaller) inner-core size is in the simulated storms. (Xu Jing)

2.14 Technical study of how to forecast a land-sea-breeze and associated mesoscale convective weather in Hainan island

The annual, monthly and daily changes of land-sea-breezes in Hainan Island and their relations with the temperature and rainfall fields are investigated by using the observational data of year 2013. Results show that a land-sea-breeze obviously happens throughout the year with a sea breeze occurring between 11:00 and 20:00 LST (local standard time) and a land breeze occurring between 20:00 and 11:00 (the next day) LST. A close relation is shown between the rainfall and the convergence caused by the land-sea-breeze, in which the thermal heating difference determines the evolution of the land-sea-breeze. The convergence caused by the land-sea-breeze provides a good condition for the lifting of air, which can lead to significant rainfall when accompanied with suff cient water vapor. The relative humidity near the surface is a good indicator of water vapor conditions. (Liang Zhaoming，Wang Donghai)

2.15 The characteristic analysis of hourly precipitation over eastern China

Based on the hourly rain gauge data of 2420 stations across China from June 1st to August 31st during 1951 to 2013, the hourly precipitation sequence of 1489 stations for areas in China to the east of 105°E from June 1st to August 31st during 1980 to 2012 is established with the quality controlled and no-data records rejected. The annual average precipitation frequency, rainfall intensity, hourly extreme precipitation frequency, including the distribution of the summer rainfall intensity, are analyzed. The trend of summer precipitation frequency, that of the monthly summer rainfall intensity change, and the frequency of the extreme strong precipitation events is also investigated. The results indicate that the frequency of hourly precipitation in summer mainly declines while its intensity increases on the whole over east China. The frequency of hourly extreme precipitation decreases while its intensity increases as a whole (Fig. 5). (Zhao Linna)

3 Numerical weather prediction and key techniques in numerical modeling

3.1 Study of the high-resolution physical processes

The reasons of large raindrop and high radar ref ectivity simulated by CAMS cloud microphysics over the Tibetan Plateau are investigated at 1 km horizontal resolution. The relationship between cloud water content and cloud droplet concentration is analyzed using aircraft observations, and the cloud vertical structure over East Asia is explored with multi-satellite data. As a result, the performances of CAMS cloud microphysics are considerably improved over the plateau and South China. In addition, the surface exchange coeff cient in a semi-arid region is assessed and modif ed using the observations, while the calculations about surface exchange coeff cient, canopy damping parameter and surface heat f ux are also improved (Fig. 6). (Gao Wenhua, Yin Jinfang, Zhang Guo)

3.2 Quasi-uniform grid GRAPES model development

Using Schwarz method, the non-hydrostatic GRAPES model is successfully established on the Yin-Yang quasi-uniform grid with a first-type boundary condition and semi-implicit semi-Lagrangian solver. A longterm stable integration of the model dynamic core is conf rmed. The application of the ILU pre-conditioner accelerates the GCR solver for Helmholtz equations, and consequently the convergence rate is speeded up signif cantly in the temporal integration of the model on Yin-Yang grids. An overall second-order accuracy and excellent stability of the model frame are verif ed in the three-dimensional idealized test cases. The frame provides a possible solution for the next-generation high-resolution GRAPES model development (Fig. 7). (Peng Xindong)

3.3 Objective correction of short-term rainfall prediction

The Anomaly Numerical-correction with Observations (ANO) method and a high-resolution precipitation analysis data are used to test the model results of correction in a Sichuan rainstorm case in July 2013, which shows the amelioration of meso-scale circulation and high-resolution rainfall forecasting. Great improvement to the high-resolution circulation, especially near surface diurnal variations, is displayed. The geopotential height, temperature and relative humidity are prominently corrected at 550 hPa and 750 hPa. Notable improvement to the short-term rainfall forecasting with the WRF mesoscale model is demonstrated (Fig. 8). (Peng Xindong)