1 災(zāi)害天氣監(jiān)測(cè)

1.1 華南云降水微物理結(jié)構(gòu)觀測(cè)和不同波段不同觀測(cè)體制雷達(dá)的數(shù)據(jù)融合方法

在2017年華南暴雨外場(chǎng)試驗(yàn)期間，利用Ka波段云雷達(dá)、C波段連續(xù)波雷達(dá)和激光云高儀等設(shè)備，獲取了華南云降水垂直結(jié)構(gòu)綜合數(shù)據(jù)；利用X波段相控陣天氣雷達(dá)，獲取了強(qiáng)對(duì)流三維結(jié)構(gòu)數(shù)據(jù)，這些數(shù)據(jù)在云降水機(jī)理研究方面得到應(yīng)用，實(shí)現(xiàn)了外場(chǎng)試驗(yàn)數(shù)據(jù)的遠(yuǎn)程監(jiān)控。對(duì)比了Ka波段云雷達(dá)與C波段連續(xù)波雷達(dá)不同波長(zhǎng)和不同探測(cè)體制的雷達(dá)的回波強(qiáng)度、徑向速度的一致性，提出了云雷達(dá)、C波段連續(xù)波雷達(dá)和激光云高儀的數(shù)據(jù)融合方法。云雷達(dá)在高云觀測(cè)中起重要作用，同時(shí)C波段連續(xù)波雷達(dá)在中云、低云和強(qiáng)降水中觀測(cè)起重要作用，激光云高儀對(duì)識(shí)別云降水回波中的晴空回波有一定幫助。3種數(shù)據(jù)融合填補(bǔ)了云雷達(dá)觀測(cè)強(qiáng)降水的弱點(diǎn)，減小了晴空回波和霧對(duì)邊界層內(nèi)云降水觀測(cè)的影響（圖1）。（劉黎平，阮征，崔哲虎）

1.2 X波段相控陣天氣雷達(dá)和雙線偏振雷達(dá)聯(lián)合分析的一次華南超級(jí)單體微物理和動(dòng)力精細(xì)結(jié)構(gòu)

利用2016年X波段相控陣天氣雷達(dá)和雙線偏振雷達(dá)聯(lián)合觀測(cè)的5月9日發(fā)生在廣東佛山一次超級(jí)單體數(shù)據(jù)，以及發(fā)展的雙線偏振雷達(dá)降水粒子相態(tài)識(shí)別方法進(jìn)行對(duì)流系統(tǒng)微物理參數(shù)反演和分析。結(jié)果顯示，相控陣?yán)走_(dá)對(duì)強(qiáng)對(duì)流精細(xì)結(jié)構(gòu)及其演變的觀測(cè)能力優(yōu)于常規(guī)的機(jī)械掃描雷達(dá)。對(duì)流過程的觸發(fā)和新單體的合并常常發(fā)生在強(qiáng)的入流區(qū)，懸垂回波的形成與消亡與入流和出流有一定關(guān)系。通過這一過程分析，進(jìn)一步檢驗(yàn)和驗(yàn)證了我國(guó)生產(chǎn)的相控陣天氣雷達(dá)和雙線偏振雷達(dá)是對(duì)流過程分析研究和短臨預(yù)報(bào)的重要工具。（劉黎平，吳翀）

1.3 基于雙線偏振雷達(dá)觀測(cè)數(shù)據(jù)的那曲對(duì)流云結(jié)構(gòu)

利用可移式C波段雙線偏振雷達(dá)（C-POL）以及那曲新一代天氣雷達(dá)（CINRAD/CD）2014年7—8月在西藏那曲地區(qū)的觀測(cè)資料，并通過雙多普勒雷達(dá)風(fēng)場(chǎng)反演和偏振雷達(dá)相態(tài)識(shí)別，清晰展示了高原冰雹云發(fā)生發(fā)展的動(dòng)力、微物理、熱力結(jié)構(gòu)特征。RHI掃描的ZH、ZDR和反演的相態(tài)（Class）分布圖上顯示，粒子跟隨“零線”抬高不斷增長(zhǎng)，回波強(qiáng)度也越來越大，并最終超過主上升氣流從另一側(cè)降落，形成冰雹墻的整個(gè)動(dòng)力與微物理過程。從連續(xù)時(shí)次的RHI掃描圖上，還觀測(cè)到一次對(duì)流單體發(fā)生發(fā)展過程中相態(tài)從濕雪到冰雹的變化。單體剛剛觸發(fā)時(shí)，回波高度不高，強(qiáng)度還很弱，但是卻出現(xiàn)成片的濕雪區(qū)域，說明上升氣流非常旺盛，將本來落到0 ℃以下的未完全融化的濕雪重新帶到0 ℃層以上，通過凝華、淞附、攀附等物理過程，僅僅10 min，這些濕雪就能夠迅速增長(zhǎng)成為冰雹。這些濕雪重新凝結(jié)過程中潛熱釋放，進(jìn)一步促進(jìn)了不穩(wěn)定結(jié)構(gòu)，加強(qiáng)了上升氣流和下沉氣流。因此，如果在某個(gè)剛剛生成的弱回波區(qū)域內(nèi)，融化層以上出現(xiàn)大量的濕雪，往往預(yù)示著該區(qū)域上升氣流強(qiáng)勁，會(huì)迅速發(fā)展成強(qiáng)回波單體。通過三維風(fēng)場(chǎng)反演得到的不同高度層的風(fēng)場(chǎng)配置可以看出，風(fēng)切變的存在，不僅提供了水平旋轉(zhuǎn)，并且加速了垂直環(huán)流，有利于不穩(wěn)定能量向動(dòng)能的轉(zhuǎn)換，單體從而能夠迅速發(fā)展起來（圖2）。（胡志群）

1.4 我國(guó)新一代天氣雷達(dá)網(wǎng)波束傳播條件的評(píng)估

大氣折射率垂直梯度（VRG）對(duì)天氣雷達(dá)波束傳播至關(guān)重要。雷達(dá)波束路徑計(jì)算中最常用的是標(biāo)準(zhǔn)折射條件，即4/3地球半徑模式。應(yīng)用探空數(shù)據(jù)分析了我國(guó)天氣雷達(dá)網(wǎng)分布區(qū)域地表上1 km內(nèi)VRG的時(shí)空間變化特征。指出，在區(qū)域氣候條件與地理環(huán)境的影響下，大部分天氣雷達(dá)VRG較標(biāo)準(zhǔn)條件下偏小，極少數(shù)雷達(dá)稍有偏大。VRG總體上以偏小為主，尤其是雷達(dá)觀測(cè)特別重要的雨季，同時(shí)雷達(dá)波束較標(biāo)準(zhǔn)條件下偏低。不過考慮到1°的雷達(dá)波束寬度導(dǎo)致的回波高度不確定性，大氣折射引起的波束高度誤差是可以接受的。但在沿海地區(qū)和長(zhǎng)江中下游地區(qū)，雷達(dá)波束受遮擋和地物污染的風(fēng)險(xiǎn)更高，在雷達(dá)數(shù)據(jù)應(yīng)用過程中有必要更加謹(jǐn)慎。評(píng)估結(jié)果更準(zhǔn)確地給出了我國(guó)新一代天氣雷達(dá)電磁波傳播的大氣環(huán)境，有助于了解雷達(dá)觀測(cè)數(shù)據(jù)受地物雜波污染風(fēng)險(xiǎn)，以便更好地應(yīng)用雷達(dá)觀測(cè)數(shù)據(jù)。（王紅艷）

2 青藏高原研究

2.1 夏季青藏高原對(duì)流系統(tǒng)（TCS）的地區(qū)差異

夏季青藏高原對(duì)流系統(tǒng)（TCS）存在高原中部（97°～102°E，28°～33°N）和高原東部（89°～94°E，28°～33°N）2個(gè)高發(fā)中心，且這2個(gè)高發(fā)中心的TCS存在明顯的地區(qū)差異。東部中心生成的TCS更多，但這些TCS只有20%～30%生成于青藏高原切變線、低渦、低壓等天氣系統(tǒng)，面積較小，水汽相對(duì)豐富，更容易產(chǎn)生降水，降水最大值出現(xiàn)在00∶00—03∶00，其高空西風(fēng)氣流更強(qiáng)，更容易（60%～90%）移出高原；而高原中部中心生成的TCS大多（50%～70%）伴有切變線/低渦/低壓天氣系統(tǒng)，面積更大，但更干燥，平均降水率小，最大降水率出現(xiàn)在18∶00—21∶00，由于其高空西風(fēng)氣流較弱，且離高原東部邊界較遠(yuǎn)，大多（60%～90%）無法移出高原（圖3）。（胡亮，徐祥德，趙平)

2.2 利用大渦模擬研究青藏高原淡積云

通過大渦模擬由地表感熱通量和潛熱通量強(qiáng)迫生成的淡積云，得到淡積云的宏觀特征量(主要水平尺度譜、云底高度、云的覆蓋度、云中垂直速度等參量)，重點(diǎn)討論高原低空氣密度條件下，邊界層內(nèi)強(qiáng)湍流對(duì)淡積云影響的物理機(jī)制。（王寅鈞， 徐祥德）

2.3 青藏高原與中國(guó)其他地區(qū)邊界層高度的統(tǒng)計(jì)特征

使用中國(guó)地區(qū)128個(gè)探空站數(shù)據(jù)計(jì)算青藏高原與其他地區(qū)的邊界層高度，用Gamma函數(shù)擬合不同地區(qū)在不同穩(wěn)定度條件下(對(duì)流邊界層CBL，中性殘余層NRL，穩(wěn)定邊界層SBL)邊界層高度頻率的分布，得到了形狀參數(shù)k和尺度參數(shù)s，指出青藏高原與其他地區(qū)邊界層高度的異同。（王寅鈞， 徐祥德等）

2.4 青藏高原對(duì)我國(guó)霧霾天氣的影響及其對(duì)降水的影響

中國(guó)降水極端事件變異與大地形東部霾污染區(qū)域氣候“脅迫”相關(guān)，青藏高原和黃土高原大地形結(jié)構(gòu)“背風(fēng)坡”“約束”下構(gòu)成了大地形東側(cè)人為氣溶膠持續(xù)高值區(qū)，且氣溶膠濃度高值區(qū)（對(duì)應(yīng)霾天氣高頻區(qū)）與中國(guó)西部氣溶膠相對(duì)“清潔區(qū)域”的不同降水強(qiáng)度極端事件年代際變化趨勢(shì)顯著性特征均存在明顯差異。從華北地區(qū)飛機(jī)觀測(cè)與地面氣溶膠、氣象站點(diǎn)多源信息分析中發(fā)現(xiàn)華北區(qū)域氣溶膠（能見度）與云結(jié)構(gòu)特征相關(guān)，且中國(guó)高原大地形東部降水極端事件與能見度存在年際同步變化趨勢(shì)，即隨著能見度降低（氣溶膠增多），小雨事件呈減少趨勢(shì)，暴雨、特大暴雨事件呈顯著增多趨勢(shì)，研究結(jié)果揭示出氣溶膠與暴雨極端事件的關(guān)聯(lián)性，從實(shí)際天氣事件年際變化視角，清晰地描述出氣溶膠變異的氣候效應(yīng)及其對(duì)極端天氣災(zāi)害的影響特征，提出了氣溶膠變異的氣候效應(yīng)的新事實(shí)。（徐祥德，郭學(xué)良等)

2.5 低緯海洋水汽對(duì)中國(guó)夏季大陸降水的調(diào)制

基于OME-based降水診斷技術(shù)，揭示并量化了3個(gè)低緯海洋子區(qū)域（阿拉伯海、孟加拉灣和南中國(guó)海）水汽輸出對(duì)中國(guó)大陸夏季降水的貢獻(xiàn)；同時(shí)，還探討了3個(gè)子區(qū)域海洋水汽輸出異常與年際尺度上觀測(cè)到的降水變化的相關(guān)性。研究發(fā)現(xiàn)，從多年氣候態(tài)來看，來自阿拉伯海、孟加拉灣和南海的海洋水汽輸出對(duì)整個(gè)中國(guó)大陸降水貢獻(xiàn)相當(dāng)，但在中國(guó)次季節(jié)尺度上，3個(gè)不同地位海洋區(qū)域的貢獻(xiàn)存在顯著反位相季節(jié)內(nèi)變化。分析結(jié)果還顯示，年際變化尺度上，中國(guó)南?！霞永瓰车貐^(qū)與阿拉伯海地區(qū)導(dǎo)致的OME降水呈顯著負(fù)相關(guān)，并影響著中國(guó)夏季降水變化。（陳斌，徐祥德)

2.6 青藏高原通過亞洲—太平洋濤動(dòng)（APO）影響全球氣候的新途徑

通過改變高原地形高度和改變高原植被類型兩類數(shù)值試驗(yàn)，分別模擬出青藏高原抬升加熱的動(dòng)力、熱力混合效應(yīng)以及青藏高原單獨(dú)的熱力效應(yīng)。研究顯示，上述兩類試驗(yàn)均再現(xiàn)了青藏高原激發(fā)亞洲與北太平洋熱帶外異常垂直環(huán)流的物理過程，成功強(qiáng)迫出了典型的夏季亞洲—太平洋濤動(dòng)（APO）異常結(jié)構(gòu)，從而證實(shí)了青藏高原熱力作用在夏季占主導(dǎo)地位，是影響夏季APO的一個(gè)重要因素。這一工作指出了青藏高原熱力異常通過對(duì)APO的調(diào)節(jié)作用進(jìn)而影響全球氣候的新途徑。（劉舸， 趙平， 陳軍明）

3 暴雨和強(qiáng)對(duì)流研究

3.1 全球五大業(yè)務(wù)中心的集合預(yù)報(bào)系統(tǒng)對(duì)華南前汛期定量降水預(yù)報(bào)的評(píng)估

利用THORPEX TIGGE的數(shù)據(jù)集，評(píng)估了歐洲中期天氣預(yù)報(bào)中心（ECMWF）、美國(guó)國(guó)家環(huán)境預(yù)報(bào)中心（NCEP）、日本氣象廳（JMA）、韓國(guó)氣象廳（KMA）和中國(guó)氣象局（CMA）共5個(gè)全球集合預(yù)報(bào)系統(tǒng)（EPS）對(duì)華南前汛期（4—6月）降水的預(yù)報(bào)能力。3個(gè)季度（2013—2015年）的0～5天預(yù)報(bào)評(píng)估表明，與簡(jiǎn)單的數(shù)值平均預(yù)報(bào)相比，概率匹配方法處理的集合預(yù)報(bào)結(jié)果對(duì)降水強(qiáng)度的預(yù)報(bào)能力更高，確定性預(yù)報(bào)的能力與概率匹配方法較為接近。所有集合預(yù)報(bào)均高估弱降雨量（＜30 mm/12 h），低估強(qiáng)降雨量（＞30 mm/12 h），其中JMA低估最嚴(yán)重。對(duì)于4個(gè)代表性暴雨事件，對(duì)比最好的集合預(yù)報(bào)（ECMWF）和較差的集合預(yù)報(bào)（JMA和CMA）的天氣形勢(shì)，并開展集合敏感性分析。結(jié)果表明，將暖濕空氣運(yùn)送到華南地區(qū)的暴雨區(qū)上游的低層西南氣流是控制定量降水預(yù)報(bào)的關(guān)鍵天氣因素；預(yù)報(bào)局地的暴雨比預(yù)報(bào)更廣泛分布的暴雨更具挑戰(zhàn)性；通過縮短30～36 h至18～24 h至6～12 h的起報(bào)時(shí)間，可以稍微改善受大尺度動(dòng)力強(qiáng)迫影響的降水，但未改善局地降水（圖4）。（黃齡，羅亞麗）

3.2 臺(tái)灣2003—2012年極端小時(shí)降水天氣背景分析

基于2003—2012年臺(tái)灣密集自動(dòng)觀測(cè)站的小時(shí)降水?dāng)?shù)據(jù)，對(duì)臺(tái)灣5年、10年、20年一遇以及超過100 mm/h 4種閾值的極端小時(shí)降水進(jìn)行統(tǒng)計(jì)分析。極端小時(shí)降水記錄根據(jù)其發(fā)生時(shí)的天氣背景共分為4類：熱帶氣旋型，鋒面型，弱天氣尺度強(qiáng)迫型以及低渦／切變線型。在4種閾值的分析結(jié)果中，熱帶氣旋型均超過總記錄數(shù)的75%，鋒面型和弱天氣尺度強(qiáng)迫型所占比例相當(dāng)（9%～13%）。極端小時(shí)降水在5月到6月中旬主要由梅雨鋒造成，在7—10月則主要由熱帶氣旋造成。熱帶氣旋型極端降水事件維持時(shí)間傾向較長(zhǎng)（＞12 h），且在極端小時(shí)降水發(fā)生前后的降水強(qiáng)度演變具有對(duì)稱結(jié)構(gòu)；鋒面型和弱天氣尺度強(qiáng)迫型降水事件發(fā)生時(shí)間較短（＜6 h）且降水演變具有不對(duì)稱性結(jié)構(gòu)。熱帶氣旋型根據(jù)熱帶氣旋中心相對(duì)臺(tái)灣島的位置被分為7個(gè)子類型。子類型Ⅰ—Ⅳ發(fā)生在熱帶氣旋中心登陸臺(tái)灣島或在距離最近海岸線100 km范圍內(nèi)時(shí)，極端小時(shí)降水的空間分布主要由當(dāng)熱帶氣旋中心位于臺(tái)灣西北、南部、東部及東北方位角時(shí)熱帶氣旋環(huán)流與臺(tái)灣地形的相互作用關(guān)系所決定；子類型Ⅴ—Ⅶ發(fā)生在熱帶氣旋中心遠(yuǎn)離臺(tái)灣島（＞100 km）時(shí)，西南/東北季風(fēng)的強(qiáng)度以及熱帶氣旋外部環(huán)流與臺(tái)灣中央山脈在山脈東側(cè)的相互作用也是決定極端小時(shí)降水空間分布的重要因子。（羅亞麗，吳夢(mèng)雯）

3.3 華南沿海一次極端降水過程中多雨帶和多單體排列的作用

2014年5月11日，由多條30～80 km長(zhǎng)的雨帶組成的2個(gè)中尺度對(duì)流系統(tǒng)（MCS），在廣東沿海中部產(chǎn)生的24 h累積降水高達(dá)452 mm。雙多普勒雷達(dá)風(fēng)場(chǎng)反演表明，沿著海岸線有一條準(zhǔn)靜止中尺度出流邊界，海上來的暖濕氣流被邊界抬升觸發(fā)對(duì)流，使得2個(gè)中尺度對(duì)流系統(tǒng)得以長(zhǎng)時(shí)間維持。在對(duì)流層中低層環(huán)境西南氣流的影響下，出流邊界上新生的對(duì)流單體向東北方向移動(dòng)組成單體波列，形成多條中β尺度雨帶。這些雨帶幾乎相互平行，沿著海岸線向下游移動(dòng)并產(chǎn)生極端降水，其中下午的第2個(gè)MCS的降水貢獻(xiàn)更大。S波段雙線偏振雷達(dá)資料表明，與上午第1個(gè)MCS相比，第2個(gè)MCS有數(shù)目更多的中等大小的降水粒子以及更大的大雨滴，其對(duì)流上升運(yùn)動(dòng)更強(qiáng)，冰相過程也更活躍，這可能與下午更大的環(huán)境對(duì)流不穩(wěn)定能量有關(guān)。進(jìn)一步分析第2個(gè)MCS多雨帶結(jié)構(gòu)的形成發(fā)現(xiàn)，其最前方對(duì)流雨帶的后部形成強(qiáng)的西風(fēng)氣流，使得雨帶中部向前突起成為弓形，并分裂成南北2段，中尺度出流邊界的對(duì)流觸發(fā)作用，使得分裂的兩部分各自發(fā)展為完整的雨帶，這種弓形雨帶的斷裂—再生的現(xiàn)象使得強(qiáng)降水雨帶的數(shù)目迅速增加。建立了此次過程多雨帶結(jié)構(gòu)極端降水MCS的概念模型。（羅亞麗，劉希，張大林）

3.4 華南前汛期極端小時(shí)降水統(tǒng)計(jì)特征分析

利用基于稠密雨量計(jì)觀測(cè)資料生成的8 km格距的小時(shí)降水?dāng)?shù)據(jù)、天氣圖、再分析資料和逐6 min雷達(dá)組合反射率拼圖數(shù)據(jù)，分析了2011—2017年華南前汛期極端小時(shí)降水（＞60 mm/h）的基本特征。分析時(shí)間段內(nèi)，華南（主要是廣西、廣東兩省）共錄得2659個(gè)極端小時(shí)降水記錄，4、5、6月分別錄得262、1405和992個(gè)。有3個(gè)極端小時(shí)降水發(fā)生的高頻中心，分別位于廣西北部，廣西東南部，以及廣東中部和沿海地區(qū)。3個(gè)中心極端小時(shí)降水地日變化特征各不相同，廣西北部是單峰結(jié)構(gòu)，峰值出現(xiàn)在午夜，廣東中部和沿海是雙峰結(jié)構(gòu)，主峰在清晨，次峰在下午，廣西東南部中心沒有明顯的日變化。根據(jù)發(fā)生時(shí)的天氣背景，把這2659個(gè)極端降水分為鋒面型、低渦型、切變線型、弱天氣尺度強(qiáng)迫型，和熱帶氣旋型5種類型，它們分別占總頻次的35.5%、8.9%、16.4%、38.1%和1%。相比于廣東，廣西的極端小時(shí)降水受到天氣尺度系統(tǒng)產(chǎn)生的動(dòng)力抬升作用更明顯。對(duì)發(fā)生在2013—2017年5月1日至6月15日的1716個(gè)極端小時(shí)降水記錄，利用逐6 min的雷達(dá)組合反射率拼圖數(shù)據(jù)分析降水系統(tǒng)的組織模態(tài)特征。結(jié)果表明：由線狀中尺度對(duì)流系統(tǒng)MCS（MCS中只有1條超過100 km長(zhǎng)的主對(duì)流帶）、多雨帶型MCS（MRB，MCS包含多條對(duì)流帶）和更小尺度的降水系統(tǒng)產(chǎn)生的極端小時(shí)降水分別占總頻次的51.4%、7.7% 和41%；線狀MCS造成的極端降水廣泛分布在華南地區(qū)，在廣西北部有一個(gè)高頻中心，在廣東中部沿海地區(qū)有一個(gè)次高頻中心；由MRB型MCS產(chǎn)生的極端降水主要影響廣東中部?jī)?nèi)陸和沿海地區(qū)；更小尺度的降水系統(tǒng)造成的極端降水分布在整個(gè)華南地區(qū)，且在廣西和廣東中部有兩個(gè)高頻中心。極端小時(shí)降水事件的時(shí)長(zhǎng)和發(fā)展類型也具有明顯的區(qū)域差異。（羅亞麗，陳楊瑞雪，李爭(zhēng)輝）

3.5 同化廣東沿海業(yè)務(wù)雷達(dá)徑向風(fēng)觀測(cè)資料對(duì)模擬強(qiáng)降水對(duì)流活動(dòng)的影響

針對(duì)2013年前汛期發(fā)生在華南沿海的一次強(qiáng)降水過程，研究WRF-EnKF系統(tǒng)同化多普勒雷達(dá)徑向風(fēng)速度觀測(cè)資料對(duì)于中尺度對(duì)流系統(tǒng)（MCS）預(yù)報(bào)技巧的影響。同化試驗(yàn)首先通過擾動(dòng)生成60個(gè)集合成員，然后分別以6、12、24、30 min的時(shí)間間隔、不同的同化時(shí)長(zhǎng)（0～2 h）同化了雷達(dá)資料，運(yùn)行40個(gè)確定性預(yù)報(bào)試驗(yàn)（DADF），運(yùn)行了1組沒有進(jìn)行資料同化的試驗(yàn)（NODA）以作參照。結(jié)果表明：同化試驗(yàn)?zāi)軌蝻@著改進(jìn)對(duì)流預(yù)報(bào)，雖然與觀測(cè)對(duì)流在位置上有些差別，40 個(gè)DADF試驗(yàn)?zāi)軌虼蟾怕实仡A(yù)報(bào)出廣東沿海陸地上和海上的中尺度對(duì)流系統(tǒng)；與更長(zhǎng)時(shí)間間隔的試驗(yàn)成員相比，最短時(shí)間間隔（6 min）的成員具有更高的預(yù)報(bào)技巧，其中最長(zhǎng)同化時(shí)長(zhǎng)和最短同化間隔的成員具有最優(yōu)的預(yù)報(bào)效果，然而縮短時(shí)間間隔（例如12 min）或者延長(zhǎng)同化時(shí)長(zhǎng)并不總是對(duì)預(yù)報(bào)技巧有正的影響；在最優(yōu)成員中，改進(jìn)的初始場(chǎng)提供的動(dòng)力和熱動(dòng)力條件更有利于陸地上的MCS更早觸發(fā)、海洋上的MCS能夠維持更長(zhǎng)時(shí)間。（寶興華，羅亞麗）

3.6 “5.7”廣州破紀(jì)錄降水的對(duì)流系統(tǒng)的觀測(cè)分析

利用華南季風(fēng)降水試驗(yàn)（SCMREX）加密觀測(cè)期間多種儀器的綜合觀測(cè)資料，以及四維變分多普勒雷達(dá)分析系統(tǒng)（4D VDRAS）生成的高分辨率格點(diǎn)化數(shù)據(jù)，對(duì)造成2017年5月7日廣州極端降水的中β尺度對(duì)流系統(tǒng)進(jìn)行了研究。結(jié)果表明，包裹在弱環(huán)境風(fēng)中并攜帶著充分水汽的上岸急流在經(jīng)過城區(qū)時(shí)不斷減速形成中β尺度輻合區(qū)域，最終在城區(qū)北側(cè)的一座小山前緣導(dǎo)致對(duì)流觸發(fā)，城市下墊面增強(qiáng)的摩擦作用和地形阻擋作用是該β尺度輻合區(qū)形成的原因，而城市熱島效應(yīng)有利于夜間在城市上方及其下風(fēng)方（北側(cè)）維持不穩(wěn)定大氣，并形成相對(duì)較深厚的邊界層。上岸氣流導(dǎo)致了城市上方深厚濕空氣柱的形成，深厚的濕空氣柱、弱的環(huán)境風(fēng)場(chǎng)以及中等高度的中性浮力層（LNB，大約10 km ASL）抑制了強(qiáng)上升氣流的出現(xiàn)和高層水凝物的卷出，從而有利于對(duì)流系統(tǒng)維持準(zhǔn)靜止，并產(chǎn)生較高的降水效率。利用VDRAS輸出結(jié)果發(fā)現(xiàn)，極端降水系統(tǒng)內(nèi)部存在一個(gè)中β尺度渦旋，該渦旋內(nèi)存在一個(gè)中等厚度、中γ尺度、圓形的對(duì)流核（C3）。渦度診斷表明，對(duì)流環(huán)境水平渦度的傾側(cè)作用對(duì)渦旋的早期生成起重要作用，而與低層輻合相關(guān)的拉伸作用對(duì)垂直正渦度的維持起重要作用。潛熱釋放造成的熱浮力是維持C3內(nèi)上升運(yùn)動(dòng)的主要原因，動(dòng)力垂直擾動(dòng)氣壓梯度力的作用次之?；钴S的暖雨過程生成了大量中等到大粒徑的雨滴，這可能與偏南入流攜帶的大量海鹽氣溶膠充當(dāng)云凝結(jié)核有關(guān)。（羅亞麗，馬若赟，張大林，陳明軒）

3.7 大尺度環(huán)境場(chǎng)對(duì)華北中尺度對(duì)流系統(tǒng)對(duì)流性降水率和地閃頻次的影響

對(duì)有利于華北中尺度對(duì)流系統(tǒng)（MCS）產(chǎn)生不同強(qiáng)度降水和不同頻次地閃的環(huán)境場(chǎng)進(jìn)行了合成研究。根據(jù)MCS產(chǎn)生的對(duì)流性降水率的高低（HR/LR）和地閃頻次的多少（HL/LL），將2008—2013年6—8月期間的60個(gè)MCS分為強(qiáng)降水率多地閃（HRHL）、強(qiáng)降水率少地閃（HRLL）、弱降水率多地閃（LRHL）和弱降水率少地閃（LRLL）等4類MCS。結(jié)果表明：伴有HR（HL）類MCS多發(fā)生在7月（8月），而伴隨LR或LL的MCS多發(fā)生在6月，與季節(jié)變化一致。MCS多在午后形成，其中HRLL MCS也會(huì)在夜間形成，而HRHL MCS在一天中任何階段都有可能形成。環(huán)流合成分析表明，除了發(fā)生頻次非常低的LRHL MCS外，其他3類MCS的環(huán)流場(chǎng)存在顯著差異。HRHL和HRLL MCS發(fā)生在高空異常輻散，中層槽和低層西南氣流輸送水汽的環(huán)流場(chǎng)中，而LRLL MCS多緣于山區(qū)日間加熱。HRHL、HRLL、LRHL和LRLL MCS環(huán)流場(chǎng)表現(xiàn)出從高到低的對(duì)流有效位能和可降水量分布，和從低到高的對(duì)流抑制和抬升指數(shù)分布。大尺度環(huán)境條件可在一定程度上決定MCS中地閃的頻次和降水率的強(qiáng)弱，但還受到其他過程的影響，如云微物理過程對(duì)地閃頻次的影響。（夏茹娣，張大林）

3.8 北大西洋濤動(dòng)事件的發(fā)生發(fā)展機(jī)制研究

基于ERA-Interm再分析資料，揭示了北大西洋濤動(dòng)（NAO）事件與熱帶季節(jié)內(nèi)振蕩（MJO）的可能關(guān)系，指出不同類型NAO事件提前2周左右的前期征兆信號(hào)為NAO事件在延伸期范圍內(nèi)的預(yù)報(bào)提供了指導(dǎo)。研究表明，受MJO影響的NAO事件可能是一種北半球環(huán)狀模（NAM），生命周期較長(zhǎng)，大概30天左右；而不受MJO影響的NAO事件可能為經(jīng)典的局地NAO事件，生命周期較短，大概10天左右。（姜智娜）

3.9 海風(fēng)環(huán)流對(duì)海南島對(duì)流系統(tǒng)觸發(fā)和傳播作用的數(shù)值模擬

基于1次高分辨率數(shù)值模擬，研究了海風(fēng)環(huán)流對(duì)海南島對(duì)流系統(tǒng)的觸發(fā)和傳播的作用機(jī)制。結(jié)果發(fā)現(xiàn)，山地對(duì)環(huán)境風(fēng)的阻擋效應(yīng)抑制了迎風(fēng)坡的谷風(fēng)環(huán)流和海風(fēng)環(huán)流的發(fā)展，而在海南島低地形條件下，較弱環(huán)境風(fēng)速則有利于背風(fēng)坡的谷風(fēng)環(huán)流和海風(fēng)環(huán)流的發(fā)展和合并。背風(fēng)坡合并的局地環(huán)流與較弱環(huán)境風(fēng)在山頂附近產(chǎn)生穩(wěn)定而強(qiáng)的氣流輻合，該氣流輻合造成了大量水汽在山地附近堆積，導(dǎo)致比周邊地區(qū)更大有效位能的產(chǎn)生，從而使得對(duì)流系統(tǒng)容易在山頂附近觸發(fā)。對(duì)流系統(tǒng)產(chǎn)生后在環(huán)境風(fēng)的引導(dǎo)下往山地背風(fēng)坡下游移動(dòng)，陸續(xù)與下游的海風(fēng)鋒及其背后一系列開爾文—赫姆霍茲波相互作用。由于海風(fēng)鋒和這些開爾文—赫姆霍茲波的前沿均為強(qiáng)對(duì)流運(yùn)動(dòng)區(qū)以及豐富水汽堆積區(qū)和高有效位能區(qū)，造成了對(duì)流系統(tǒng)往下游移動(dòng)時(shí)出現(xiàn)“波浪式”傳播發(fā)展的現(xiàn)象。（梁釗明，王東海，劉英）

4 臺(tái)風(fēng)研究

4.1 登陸我國(guó)臺(tái)風(fēng)降水的日變化信號(hào)

基于常規(guī)氣象站的小時(shí)雨量觀測(cè)資料，通過諧波分析研究發(fā)現(xiàn)登陸我國(guó)熱帶氣旋降水存在明顯日變化信號(hào)，波峰出現(xiàn)在凌晨，波谷出現(xiàn)在午后，且熱帶氣旋外圍的降水日變化信號(hào)比主體環(huán)流區(qū)域的日變化信號(hào)更加明顯。在這種日變化信號(hào)的影響下，夜間登陸的熱帶氣旋登陸后12 h累計(jì)降水量比中午或下午登陸的熱帶氣旋大30%左右（圖5）。(胡皓，端義宏，王玉清)

4.2 熱帶氣旋（TC）結(jié)構(gòu)與強(qiáng)度變化的關(guān)系

基于高分辨率數(shù)值理想試驗(yàn)的分析表明，在渦旋初始強(qiáng)度相同的情況下，尺度較小渦旋快速加強(qiáng)的速率明顯快于尺度較大和初始渦旋的速度廓線沿最大速度向外（眼區(qū)方向）衰減慢的初始渦旋。TC的快速增強(qiáng)速率（RIR）主要受到慣性穩(wěn)定度的影響，而渦旋的初始結(jié)構(gòu)決定了慣性穩(wěn)定度的分布，當(dāng)初始渦旋尺度較大或者初始速度沿徑向衰減慢時(shí)，最大風(fēng)半徑外側(cè)的較大的切向速度有利于渦旋獲得更多的海表熱通量，進(jìn)而有利于組織形成螺旋云帶。大范圍的對(duì)流云帶中的非絕熱加熱，一方面將邊界層中的徑向入流沿徑向擴(kuò)展的范圍更大，從而減弱邊界層中向眼墻的入流；另一方面，眼墻外側(cè)切向速度的增加也增加了眼墻外側(cè)的慣性穩(wěn)定度，這兩方面的綜合作用減弱了向眼墻的絕對(duì)角動(dòng)量輸入，從而減弱了渦旋加強(qiáng)的速度同時(shí)也增加了風(fēng)暴的內(nèi)核尺度（圖6）。（徐晶，王玉清）

4.3 我國(guó)沿海復(fù)雜地形上臺(tái)風(fēng)風(fēng)場(chǎng)變化特征

利用1979—2015年ERA-Interim再分析數(shù)據(jù)、CMA與JTWC最佳路徑資料，統(tǒng)計(jì)分析我國(guó)沿海臺(tái)灣島和海南島附近區(qū)域臺(tái)風(fēng)外圍與內(nèi)核大風(fēng)的分布特征。結(jié)果表明：（1）當(dāng)熱帶氣旋（TC）位于在島嶼地形附近時(shí)，TC中心與島嶼連線上有正、負(fù)渦度帶相間分布。一般TC中心附近和島嶼的另一端為正渦度帶，TC與島之間為弱的負(fù)渦度區(qū)。（2）臺(tái)灣島附近區(qū)域TC的大風(fēng)主要出現(xiàn)在臺(tái)灣島東側(cè)、東北側(cè)以及臺(tái)灣海峽內(nèi)。海南島附近區(qū)域TC大風(fēng)主要位于島東北側(cè)附近海域及西北側(cè)北部灣內(nèi)，島上風(fēng)速相對(duì)較小。（3）臺(tái)灣島附近區(qū)域TC的最大風(fēng)速半徑（RMW）均值為51.3 km，小于海南島附近區(qū)域的58.3 km。平均而言，位于臺(tái)灣島西側(cè)TC的RMW最大，東北側(cè)最?。欢Ｄ蠉u附近則為島上TC的RMW最大，西北側(cè)最小。（4）島嶼附近區(qū)域TC的34節(jié)指定風(fēng)圈均有一定變形，最長(zhǎng)半徑均位于臺(tái)風(fēng)東部象限。平均而言，臺(tái)灣島西北側(cè)臺(tái)風(fēng)風(fēng)圈非對(duì)稱性最強(qiáng)，海南島西南側(cè)區(qū)域TC風(fēng)圈形變最大。(李英，薛霖)

4.4 西北太平洋雙臺(tái)風(fēng)活動(dòng)氣候特征

利用臺(tái)風(fēng)最佳路徑資料以及NCEP全球再分析資料，采用客觀判定標(biāo)準(zhǔn)對(duì)1951—2014年西北太平洋兩類雙臺(tái)風(fēng)活動(dòng)進(jìn)行識(shí)別。得到如下氣候特征：1951—2014年西北太平洋一共出現(xiàn)699對(duì)雙臺(tái)風(fēng)活動(dòng)，其中典型類雙臺(tái)風(fēng)和非典型類雙臺(tái)風(fēng)分別為446對(duì)和253對(duì)，分別占到63.8%和36.2%；典型類雙臺(tái)風(fēng)的比例隨過程最短距離減小而增大，非典型類雙臺(tái)風(fēng)的比例隨過程最短距離減小而減小。雙臺(tái)風(fēng)活動(dòng)主要出現(xiàn)在110°～150°E、10°～30°N的洋面上；頻發(fā)區(qū)集中在南海北部和菲律賓以東洋面，臺(tái)灣島亦包含其中；雙臺(tái)風(fēng)活動(dòng)盛期一般出現(xiàn)在8—9月。典型類雙臺(tái)風(fēng)過程相互靠近速度最大時(shí)刻，雙臺(tái)風(fēng)對(duì)主要表現(xiàn)為東—西向分布，正好受副高南側(cè)東風(fēng)氣流控制，有利于東臺(tái)風(fēng)快速靠近西臺(tái)風(fēng)；典型類雙臺(tái)風(fēng)過程逆時(shí)針互旋速度最大時(shí)刻，雙臺(tái)風(fēng)對(duì)主要呈東北—西南或東北東—西南西向分布，并位于副高西側(cè)和西南方主要受東南風(fēng)控制，有利于雙臺(tái)風(fēng)的逆時(shí)針互旋。（任福民）

4.5 東風(fēng)切變和西風(fēng)切變對(duì)西北太平洋熱帶氣旋強(qiáng)度變化影響對(duì)比研究

統(tǒng)計(jì)分析1982—2015年西北太平洋熱帶氣旋（TC）強(qiáng)度變化與不同方向風(fēng)速垂直切變（VWS）的關(guān)系。結(jié)果表明，西風(fēng)切變與TC強(qiáng)度變化的相關(guān)系數(shù)（?0.36）要比東風(fēng)切變（-0.07）大，尤其是與西南風(fēng)切變相關(guān)系數(shù)達(dá)到?0.43。分析表明，這種相關(guān)性的差異與海表面溫度（SST）有關(guān)，主要是由于SST和緯向VWS存在顯著的相關(guān)關(guān)系（-0.48）。SST隨著東風(fēng)切變?cè)龃蠖龃?，而高SST對(duì)強(qiáng)度增強(qiáng)有利，可以抵消部分強(qiáng)切變帶來的負(fù)作用，從而導(dǎo)致東風(fēng)切變與強(qiáng)度變化相關(guān)關(guān)系減弱。相反，西風(fēng)切變?cè)鰪?qiáng)時(shí)SST降低，使得西風(fēng)切變與強(qiáng)度變化相關(guān)關(guān)系較強(qiáng)，而最強(qiáng)的相關(guān)關(guān)系出現(xiàn)在SST為301 K時(shí)。所以，建議在目前的強(qiáng)度統(tǒng)計(jì)預(yù)報(bào)模型中考慮VWS的方向（圖7）。（魏娜，李英）

5 雷電研究

5.1 特種觀測(cè)資料同化及雷電災(zāi)害天氣系統(tǒng)的監(jiān)測(cè)預(yù)警方法

利用云分辨率雷暴數(shù)值模式開展了閃電激發(fā)位置動(dòng)力微物理?xiàng)l件的分析研究。通過對(duì)模擬閃電始發(fā)點(diǎn)附近垂直運(yùn)動(dòng)速度 (Wini) 和霰粒子質(zhì)量混合比 (qg-ini) 的分析發(fā)現(xiàn)：（1）Wini與單體中最大上升運(yùn)動(dòng)速度之間存在3次方的多項(xiàng)式相關(guān)，相關(guān)系數(shù)達(dá)到約0.97。（2）多數(shù)閃電始發(fā)于霰粒子分布區(qū)的上部，qg-ini與單體內(nèi)霰粒子比質(zhì)量濃度中心值以及閃電的始發(fā)高度均相關(guān)，相關(guān)系數(shù)分別達(dá)到0.86和0.85。特別是在閃電活動(dòng)的中期和后期，線性相關(guān)更加明顯。（3）單體內(nèi)主正、負(fù)電荷區(qū)之間的弱凈電荷密度區(qū)，對(duì)于閃電的激發(fā)非常重要，通常在霰粒子濃度中心之上到霰粒子分布區(qū)上邊界之間變化。在這個(gè)弱凈電荷密度區(qū)內(nèi)多數(shù)出現(xiàn)了較強(qiáng)電場(chǎng)強(qiáng)度的位置上，閃電激發(fā)位置上的冰晶濃度與霰粒子濃度之比與霰粒子比質(zhì)量濃度之間存在指數(shù)關(guān)系。(王飛)

利用中尺度起電放電模式以及衛(wèi)星和閃電定位等觀測(cè)資料，對(duì)比分析了臺(tái)風(fēng)莫拉菲 (2009) 登陸前后以及衰亡階段的電荷形成和結(jié)構(gòu)。結(jié)果表明：莫拉菲在登陸前存在近海加強(qiáng)過程，加強(qiáng)中逐漸形成清晰的臺(tái)風(fēng)眼并伴隨眼壁區(qū)閃電活動(dòng)的多發(fā)。眼壁區(qū)對(duì)流在近海加強(qiáng)階段呈現(xiàn)正的三極性電荷結(jié)構(gòu)，主負(fù)電荷區(qū)位于?25～?10 ℃層之間，其上下各有一個(gè)正電荷區(qū)。在臺(tái)風(fēng)達(dá)到最大強(qiáng)度后呈現(xiàn)負(fù)的偶極性電荷結(jié)構(gòu)，且僅存在云中部的負(fù)電荷區(qū)和下部的正電荷區(qū)。眼壁區(qū)電荷結(jié)構(gòu)同臺(tái)風(fēng)強(qiáng)度變化密切相關(guān)，不受登陸直接影響。在臺(tái)風(fēng)發(fā)展的不同階段，外螺旋雨帶對(duì)流主要表現(xiàn)為正的三極性或偶極性電荷結(jié)構(gòu)，而之前的研究一般認(rèn)為外雨帶對(duì)流只呈現(xiàn)正的偶極性電結(jié)構(gòu)。外雨帶三極性電結(jié)構(gòu)的形成類似于眼壁區(qū)三極性結(jié)構(gòu)的形成，但同時(shí)也存在其他形成機(jī)制，即在霰粒子與冰晶組成的正偶極性電荷結(jié)構(gòu)下存在一個(gè)由雹粒子組成的正電荷區(qū)，從而形成正的三極性電荷結(jié)構(gòu)。臺(tái)風(fēng)衰亡階段對(duì)流活動(dòng)較弱，主要表現(xiàn)為負(fù)的偶極性電荷結(jié)構(gòu)，類似于陸地雷暴消散階段的特性。(徐良韜，張文娟)

5.2 廣州塔對(duì)附近區(qū)域地閃活動(dòng)特征的影響

基于1999—2015年廣東電網(wǎng)地閃定位資料，對(duì)廣州塔（高600 m）建成前后其附近區(qū)域地閃活動(dòng)的分布特性進(jìn)行了詳細(xì)分析。分析發(fā)現(xiàn)：廣州塔建成后，與其周圍10 km范圍內(nèi)回?fù)裘芏群烷W電密度的平均值相比，廣州塔周圍1 km范圍內(nèi)的回?fù)裘芏群烷W電密度明顯大很多，而周圍1～4 km范圍內(nèi)的回?fù)裘芏群烷W電密度均比周圍10 km范圍的平均值要低。閃電定位系統(tǒng)得到的廣州塔周圍1 km范圍內(nèi)回?fù)綦娏鞣档乃銛?shù)平均值明顯高于10 km范圍的平均值，而1 km以外區(qū)域的值與大范圍的平均值相比沒有明顯差異。因此，可以推測(cè)廣州塔不僅會(huì)吸引附近一定范圍內(nèi)的下行閃電擊中自己，還會(huì)觸發(fā)大量上行閃電，使得擊中高建筑物的雷電數(shù)量和電磁輻射強(qiáng)度大幅增加，而附近區(qū)域的地閃數(shù)量有所減少。（呂偉濤，齊奇，馬穎，姚雯）

5.3 觸發(fā)閃電始發(fā)階段特征

上行正先導(dǎo)（UPL）是上行負(fù)地閃始發(fā)及下行負(fù)地閃連接中的重要放電過程，但由于觀測(cè)困難其始發(fā)機(jī)制仍不清楚。為此，基于2014—2015 年廣東野外雷電綜合觀測(cè)試驗(yàn)中獲取的人工觸發(fā)閃電通道底部電流和近距離電場(chǎng)等同步觀測(cè)資料，分析了觸發(fā)閃電上行正先導(dǎo)（UPL）整個(gè)起始階段的放電特征。結(jié)果表明：UPL 起始階段通道底部電流波形表現(xiàn)為不連續(xù)的脈沖形式，以單個(gè)或者成組的形式出現(xiàn)，其中UPL連續(xù)傳輸之前的電流脈沖峰值、10%～90%上升時(shí)間、波形持續(xù)時(shí)間、半峰值寬度、轉(zhuǎn)移電荷量、脈沖時(shí)間間隔的幾何平均值分別為28 A、0.33 μs、2.3 μs、0.73 μs、27 μC、25 μs，而連續(xù)傳輸開始時(shí)刻的電流脈沖相應(yīng)的波形參數(shù)幾何平均值分別為32 A、0.54 μs、3.5 μs、0.93 μs、46 μC、24 μs，兩者相比連續(xù)傳輸開始時(shí)電流脈沖具有更大的轉(zhuǎn)移電荷量。同時(shí)，UPL 對(duì)應(yīng)的近距離電場(chǎng)波形呈現(xiàn)出明顯的梯級(jí)變化，每個(gè)電場(chǎng)脈沖與通道底部電流脈沖具有良好的時(shí)間對(duì)應(yīng)關(guān)系，這表明觸發(fā)閃電UPL在起始階段表現(xiàn)出了梯級(jí)的發(fā)展特征。（張陽，張義軍，呂偉濤，鄭棟，徐良韜）

5.4 閃電始發(fā)階段的精細(xì)放電過程

從2016年開始發(fā)展了具有亞微秒、米尺度分辨率的閃電連續(xù)干涉儀，用以研究閃電擊穿放電活動(dòng)。2016年6月9日在廣東閃電綜合觀測(cè)試驗(yàn)中觀測(cè)到一例觸發(fā)閃電，在火箭上升的883 ms時(shí)間內(nèi)，共產(chǎn)生了54次明顯的閃電始發(fā)先驅(qū)放電。連續(xù)干涉儀觀測(cè)結(jié)果表明：這些先驅(qū)放電是一種發(fā)生在上升火箭頭部的擊穿過程產(chǎn)生。每次先驅(qū)放電事件始于向上發(fā)展的正擊穿，發(fā)展尺度為幾米，速度約為5×106m/s，跟隨著更強(qiáng)、更快的向下的負(fù)擊穿，發(fā)展尺度達(dá)到幾十米，平均速度約為3×107m/s。之后伴隨14個(gè)密集的、爆發(fā)式先驅(qū)脈沖簇的出現(xiàn)，穩(wěn)定的上行正先導(dǎo)開始，且甚高頻輻射迅速變得連續(xù)。值得注意的是，穩(wěn)定發(fā)展的上行正先導(dǎo)的擊穿特性展現(xiàn)了與前面先驅(qū)脈沖相似的形式。該研究為觸發(fā)閃電機(jī)制和正擊穿過程研究提供了非常有價(jià)值的發(fā)現(xiàn)（圖8）。（張陽，張義軍，呂偉濤，鄭棟）

5.5 華南人工觸發(fā)閃電初始階段和回?fù)綦娏魈卣?/h3>

開展了華南地區(qū)50例人工觸發(fā)閃電的初始階段和回?fù)暨^程的電流特征的分析研究。結(jié)果發(fā)現(xiàn)，相比于其他地方的人工觸發(fā)閃電，華南負(fù)極性觸發(fā)閃電的初始階段具有更長(zhǎng)的持續(xù)時(shí)間，更大的平均電流、電荷傳輸和比能量，其幾何平均值分別為347.9 ms、132.5 A、45.1 C 和10.0×103A2s。2次沒有回?fù)暨^程的正極性觸發(fā)閃電相比負(fù)極性觸發(fā)閃電產(chǎn)生了較大的平均電流、電荷傳輸和比能量。華南觸發(fā)閃電回?fù)暨^程的峰值電流、回?fù)? ms內(nèi)電荷傳輸和回?fù)? ms內(nèi)電荷比能量的幾何平均值分別為17.2 kA、1.3 C和5.8×103A2s，大于其他地方人工觸發(fā)閃電的對(duì)應(yīng)值。華南人工觸發(fā)閃電回?fù)綦娏?0%～90%上升時(shí)間幾何平均值為0.4 μs，小于其他地區(qū)。上述情況意味著華南人工觸發(fā)閃電具有更強(qiáng)的放電特征。研究表明，回?fù)舴逯惦娏髋c電流上升陡度、回?fù)? ms內(nèi)電荷傳輸和回?fù)? ms內(nèi)比能量之間具有顯著的相關(guān)關(guān)系。此外，當(dāng)回?fù)艏捌潆S后的延續(xù)電流持續(xù)時(shí)間超過40 ms時(shí)，回?fù)舴逯惦娏鳌? ms內(nèi)電荷傳輸和比能量很少出現(xiàn)大于25 kA、2.6 C和1.5×103A2s的情況。研究指出，包含回?fù)舻挠|發(fā)閃電相比于沒有回?fù)舻挠|發(fā)閃電傾向具有更長(zhǎng)持續(xù)時(shí)間但較小電荷傳輸?shù)某跏茧A段?；?fù)舻姆逯惦娏髋c發(fā)生該回?fù)糁巴ǖ纼?nèi)部沒有電流階段的持續(xù)時(shí)間呈現(xiàn)一定的相關(guān)性。（鄭棟，張義軍，張陽，呂偉濤）

5.6 一次超級(jí)單體的閃電尺度時(shí)空特征

對(duì)一次超級(jí)單體過程中的閃電尺度特征及其與雷暴結(jié)構(gòu)關(guān)系研究發(fā)現(xiàn)，在超級(jí)單體演變過程中平均閃電尺度與閃電頻次和雷暴發(fā)展強(qiáng)度之間呈現(xiàn)反相關(guān)，其中平均閃電尺度與閃電頻次可以用一元冪函數(shù)關(guān)系擬合，相關(guān)系數(shù)可達(dá)到?0.87。此外，5 km高度大于30 dBz的反射率面積除以閃電頻次后，與平均閃電尺度具有顯著線性相關(guān)，相關(guān)系數(shù)為0.88。從水平分布看，沿超級(jí)單體移動(dòng)方向，從對(duì)流區(qū)附近到前側(cè)云砧區(qū)，平均閃電尺度由小到大變化。平均閃電尺度最小值所在區(qū)域通常與閃電起始密度和閃電擴(kuò)展密度的大值中心區(qū)相對(duì)應(yīng)。平均閃電尺度和閃電擴(kuò)展密度的空間分布具有冪函數(shù)的反相關(guān)關(guān)系，在超級(jí)單體一半的雷達(dá)體掃時(shí)次中相關(guān)系數(shù)絕對(duì)值大于0.5，且相關(guān)性與超級(jí)單體發(fā)展強(qiáng)度具有正向?qū)?yīng)關(guān)系。在垂直方向上，閃電起始和閃電擴(kuò)展密度高值區(qū)較為接近，但對(duì)應(yīng)高度的閃電尺度通常較小。在對(duì)流區(qū)及其附近，平均閃電尺度在上述高度以下呈現(xiàn)明顯增大趨勢(shì)，在低層4～5 km達(dá)到峰值，在對(duì)應(yīng)高度以上，維持小值或稍有增大。閃電尺度與閃電活動(dòng)強(qiáng)度和超級(jí)單體發(fā)展強(qiáng)度的時(shí)空反相關(guān)關(guān)系可能是由于雷暴強(qiáng)對(duì)流階段或區(qū)域具有較小但密集的電荷區(qū)所致。（鄭棟，張義軍）

5.7 雷暴閃電活動(dòng)與垂直氣流的關(guān)系

利用閃電探測(cè)、雙雷達(dá)三維風(fēng)場(chǎng)反演數(shù)據(jù)和雷達(dá)反射率數(shù)據(jù)，發(fā)展了一種新的基于閃電數(shù)據(jù)的雷暴單體判別方法，針對(duì)該方法所識(shí)別出的多個(gè)對(duì)流單體，研究了雷暴垂直氣流特征與地閃活動(dòng)的空間對(duì)應(yīng)關(guān)系、量化關(guān)系以及在地閃起始和結(jié)束時(shí)的差異。發(fā)現(xiàn)雷暴內(nèi)大多數(shù)（79.1%）地閃出現(xiàn)在0 oC高度垂直速度在-5～5 m/s的弱上升和弱下沉氣流區(qū)，更偏向于出現(xiàn)在弱上升氣流區(qū)，特別是對(duì)于負(fù)地閃以及雷暴發(fā)展階段的地閃。地閃頻次與一定閾值條件下的上升氣流體積之間存在相關(guān)關(guān)系，成熟階段最強(qiáng)（r=0.70），其次是消亡階段（r=0.66）和初始階段（r=0.60）。高度為7～11 km的冰相物垂直通量絕對(duì)值之和與地閃頻次關(guān)系更為密切，對(duì)應(yīng)上述雷暴3個(gè)階段的相關(guān)系數(shù)分別為0.73、0.71和0.74。地閃起始時(shí)，上升氣流體積所占比例占據(jù)主導(dǎo)（79%）；地閃結(jié)束時(shí)，下沉氣流體積所占比例占主導(dǎo)（82%），但雷暴內(nèi)最大上升速度、10 m/s和20 m/s上升速度所能達(dá)到的最大高度在首次地閃和最后一次地閃發(fā)生時(shí)差異較小，說明閃電活動(dòng)對(duì)上升氣流強(qiáng)度的依賴性。研究推測(cè)，具有較為明顯分層特征的大范圍電荷區(qū)分布相比小電荷區(qū)交錯(cuò)分布的形態(tài)更有利于地閃的發(fā)生。（鄭棟，張義軍，劉黎平）

5.8 低頻電場(chǎng)變化探測(cè)陣列（LFEDA）的建設(shè)和初步結(jié)果

利用低頻和甚低頻頻段探測(cè)全閃活動(dòng)是近年來閃電探測(cè)技術(shù)發(fā)展的一個(gè)重要方向。中國(guó)氣象科學(xué)研究院于2014—2015年在廣州地區(qū)建設(shè)了由9個(gè)快電場(chǎng)變化測(cè)量?jī)x構(gòu)成的低頻閃電電場(chǎng)探測(cè)陣列（LFEDA）。基于蒙特卡羅法和人工觸發(fā)閃電試驗(yàn)對(duì)LFEDA的定位精度和探測(cè)效率進(jìn)行了理論模擬和客觀評(píng)估。前者表明站網(wǎng)中具有較長(zhǎng)基線的定位站有助于改善垂直于該基線方向的水平定位精度和沿著該基線方向的高度定位精度；后者表明LFEDA對(duì)觸發(fā)閃電事件和回?fù)舻奶綔y(cè)效率分別為100%和95%，回?fù)羝矫娑ㄎ徽`差平均值為102 m。定位結(jié)果與雷暴結(jié)構(gòu)的對(duì)比分析表明，閃電脈沖放電事件與反射率柱所表征的對(duì)流特征較強(qiáng)區(qū)域有較好對(duì)應(yīng)，脈沖放電事件在垂直方向上呈現(xiàn)合理的分布。單次云閃、地閃的定位結(jié)果發(fā)現(xiàn)，LFEDA能夠描繪閃電發(fā)展形態(tài)，基于三維定位結(jié)果獲得的云閃初始階段垂直發(fā)展速度以及地閃先導(dǎo)垂直發(fā)展速度變化趨勢(shì)與之前觀測(cè)結(jié)果類似。LFEDA具備了全閃電三維定位能力，為閃電發(fā)展特征研究以及雷暴電學(xué)研究提供了新的技術(shù)手段。（鄭棟，張陽，張義軍，呂偉濤）

5.9 閃電初始階段和尺度判別方法及其特征

基于LMA三維閃電定位數(shù)據(jù)，對(duì)2004年10月5日發(fā)生于美國(guó)新墨西哥州的1s次超級(jí)單體過程的閃電初始及其尺度特征進(jìn)行研究，提出閃電初始階段自動(dòng)判別及其特征參量提取方法，并給出參量分布特征。結(jié)果顯示：閃電初始階段上行負(fù)先導(dǎo)（下行負(fù)先導(dǎo)）的持續(xù)時(shí)間中值為13.5 ms（7.5 ms），三維位移中值為1.4 km（1.0 km），三維平均位移速度中值為9.2×104m/s（1.2×105m/s），上行負(fù)先導(dǎo)速度隨時(shí)間遞減，下行反之，二者與垂直方向夾角的中值分別為40°和54°。表征閃電尺度的閃電凸殼面積和閃電總長(zhǎng)度的概率密度呈負(fù)冪函數(shù)分布，在小值方向分布更為集中。閃電水平延展距離中值為6.1 km，垂直延展距離中值為4.3 km，約83%的閃電其水平延展距離大于垂直延展距離；閃電的持續(xù)時(shí)間中值為271.0 ms。分析發(fā)現(xiàn)，以水平延展為主的閃電起始高度分布峰值位于8.5 km，以垂直延展為主的閃電起始高度分布峰值位于11 km。閃電初始階段位移方向越接近水平對(duì)應(yīng)閃電垂直延展越小，說明閃電初始段的傳播方向?qū)τ陂W電垂直延展具有重要影響。（鄭棟，張義軍）

5.10 雷電野外綜合觀測(cè)試驗(yàn)

開展了第12年度的廣東閃電綜合觀測(cè)試驗(yàn)和第1年度高原雷暴閃電活動(dòng)觀測(cè)試驗(yàn)，在人工觸發(fā)閃電、放電過程精細(xì)化觀測(cè)、雷暴的全閃探測(cè)以及雷電物理過程研究等方面獲得了顯著進(jìn)展。（1）成功人工觸發(fā)閃電16次，觸發(fā)閃電成功率約80%，觸發(fā)次數(shù)、成功率和觸發(fā)效率，都保持了較高的水平。（2）完善了連續(xù)干涉儀精細(xì)化定位技術(shù)，實(shí)現(xiàn)了連續(xù)干涉儀的長(zhǎng)預(yù)觸發(fā)采集存儲(chǔ)，獲得了觸發(fā)閃電和自然閃電全放電過程通道發(fā)展的高時(shí)間分辨率、精細(xì)化描繪。（3）進(jìn)一步完善了閃電低頻電場(chǎng)探測(cè)陣列，整個(gè)系統(tǒng)在全年開展了雷暴過程的持續(xù)性觀測(cè)。進(jìn)一步開展了三維定位算法的優(yōu)化研究，基于信號(hào)處理技術(shù)，顯著提高了低頻電場(chǎng)變化探測(cè)陣列（LFEDA）定位數(shù)據(jù)的精細(xì)化和豐富程度。（4）首次實(shí)現(xiàn)了對(duì)閃電始發(fā)不穩(wěn)定放電-先驅(qū)放電輻射過程的超精細(xì)化解析，獲得了對(duì)始發(fā)放電過程和輻射機(jī)理的新認(rèn)識(shí)。（5）調(diào)整和優(yōu)化了廣州高建筑物雷電觀測(cè)方案，進(jìn)一步提高了雷電綜合觀測(cè)能力?；趶V東電網(wǎng)地閃定位資料，對(duì)高建筑物建成前后其附近區(qū)域地閃活動(dòng)的分布特性及不同高度建筑周圍地閃活動(dòng)分布特征的差異進(jìn)行了詳細(xì)分析。（6）首次在高原建設(shè)了包括8個(gè)子站的三維全閃定位陣列（閃電低頻電場(chǎng)探測(cè)陣列），實(shí)現(xiàn)了對(duì)高原雷暴閃電活動(dòng)的觀測(cè)和定位，初步定位結(jié)果顯示了對(duì)高原閃電活動(dòng)三維定位的可靠性。（張陽，呂偉濤，鄭棟，張義軍，姚雯，馬穎，徐良韜，齊奇，王飛，范祥鵬，樊艷峰）

6 模式關(guān)鍵技術(shù)和再分析資料研究

6.1 天氣氣候一體化模式關(guān)鍵技術(shù)研究

在模式動(dòng)力框架方面，針對(duì)非結(jié)構(gòu)網(wǎng)格的技術(shù)特點(diǎn)，對(duì)新動(dòng)力框架進(jìn)行了整體設(shè)計(jì)，并建立了球面非結(jié)構(gòu)網(wǎng)格計(jì)算模擬系統(tǒng)。該系統(tǒng)采用易于并行化擴(kuò)展的間接索引型數(shù)據(jù)結(jié)構(gòu)，包含模式運(yùn)行所需的網(wǎng)格文件，球面算子庫，方程求解器和數(shù)據(jù)讀寫工具等，已初步形成一個(gè)數(shù)值模式動(dòng)力框架雛形，可集成多種球面原型物理問題的全球準(zhǔn)均勻和變分辨率模擬?；谟?jì)算模擬框架，發(fā)展了球面非結(jié)構(gòu)網(wǎng)格高精度算法，并將其應(yīng)用于兩類物理問題的模擬：標(biāo)量平流問題和位渦傳輸問題。結(jié)果顯示，在標(biāo)量物質(zhì)傳輸問題中，采用高精度算法可有效減小低精度算法的計(jì)算誤差。在位渦傳輸問題中，當(dāng)模式采用迎風(fēng)型高階位渦傳輸后，有效抑制了位渦擬能的增長(zhǎng)，消除了渦度場(chǎng)的噪音問題，并減少了計(jì)算誤差。在模式物理過程方面，引入了氣溶膠輻射方案，形成可用于參加CMIP6試驗(yàn)的大氣模式版本。耦合模式方面，完成參加CMIP6的氣候系統(tǒng)模式CAMS-CSM定版；完成CMIP6強(qiáng)迫數(shù)據(jù)下載與制作，并為模式編寫相應(yīng)讀取模塊；完成氣溶膠模塊開發(fā)；利用CAMS-CSM模式初步開展了CMIP6計(jì)劃DECK系列試驗(yàn)，評(píng)估結(jié)果顯示該版本能成功抓住全球氣候大尺度環(huán)流和主要變率模態(tài)的主要特征，為最終開展CMIP6試驗(yàn)奠定了堅(jiān)實(shí)基礎(chǔ)（圖9）。（張祎，陳昊明，容新堯）

6.2 東亞區(qū)域大氣再分析研究進(jìn)展

進(jìn)一步改進(jìn)和優(yōu)化了東亞區(qū)域再分析系統(tǒng)，使該系統(tǒng)能夠長(zhǎng)時(shí)間穩(wěn)定連續(xù)運(yùn)行。改進(jìn)了雷達(dá)徑向風(fēng)同化算子和雷達(dá)反射率組網(wǎng)技術(shù)，在雷達(dá)數(shù)據(jù)同化應(yīng)用方面取得明顯進(jìn)展。在探空資料方面，對(duì)比分析了東亞區(qū)域大氣再分析應(yīng)用的探空資料與GDAS探空資料數(shù)量和垂直層次，并針對(duì)2種資料開展了同化試驗(yàn)和對(duì)比分析。利用優(yōu)化后的東亞區(qū)域大氣再分析系統(tǒng)開展了1年的再分析數(shù)據(jù)的研制，并對(duì)結(jié)果進(jìn)行評(píng)估。結(jié)果表明，研制的再分析數(shù)據(jù)在中國(guó)區(qū)域內(nèi)整體性能優(yōu)于ERA-Interim數(shù)據(jù)。（梁旭東，尹金方，陳鋒，謝衍新，劉英，何會(huì)中，鄒海波）

6.3 中國(guó)東南地區(qū)降水的2種集合預(yù)報(bào)綜合偏差訂正的對(duì)比分析

利用2015年6月1日至7月31日3個(gè)全球數(shù)值預(yù)報(bào)業(yè)務(wù)中心(CMA、ECMWF和NCEP)的24 h降水集合預(yù)報(bào)資料和我國(guó)東南地區(qū)降水觀測(cè)資料，采用貝葉斯模型平均方法（A方案）和基于貝葉斯模型平均方法的統(tǒng)計(jì)降尺度模型二次訂正方法（B方案）對(duì)上述3個(gè)中心和多模式超級(jí)集合降水預(yù)報(bào)進(jìn)行訂正，并對(duì)比2種方案的訂正效果；最后用2015年8月1—31日資料進(jìn)行獨(dú)立樣本檢驗(yàn)，分析訂正前后的降水預(yù)報(bào)效果。結(jié)果表明：以第50百分位的降水預(yù)報(bào)為例，經(jīng)A方案訂正后各中心和多模式的集合平均消除了大量的小雨空?qǐng)?bào)，其對(duì)小雨、中雨的訂正效果很明顯，對(duì)大雨以上的降水量級(jí)訂正效果不明顯。隨著降水閾值增加，A方案的訂正效果隨之減弱。此方案對(duì)雨帶走向的訂正不明顯，會(huì)使降水大值區(qū)量級(jí)降低甚至消失。采用B方案訂正后，不僅可降低原始集合預(yù)報(bào)的空?qǐng)?bào)率，還可對(duì)降水量級(jí)和落區(qū)進(jìn)行訂正，使降水預(yù)報(bào)的范圍和量級(jí)與實(shí)況更接近，但對(duì)大量級(jí)降水，如50.0 mm以上的降水量級(jí)訂正效果仍然不顯著（圖10）。（楊瑞雯，趙琳娜）

7 試驗(yàn)與研究平臺(tái)

7.1 大氣科學(xué)試驗(yàn)與研究平臺(tái)（原科研數(shù)據(jù)共享平臺(tái)）進(jìn)展

完成科研數(shù)據(jù)共享平臺(tái)（一期）建設(shè)，已上線運(yùn)行，提供在線數(shù)據(jù)共享服務(wù)。平臺(tái)二期建設(shè)得到了中國(guó)氣象科學(xué)研究院基本科研業(yè)務(wù)費(fèi)的支持，順利立項(xiàng)。計(jì)劃用3年（2017—2019年）時(shí)間完成大氣科學(xué)試驗(yàn)與研究平臺(tái)建設(shè)。建設(shè)內(nèi)容包括：（1）研制大氣科學(xué)試驗(yàn)數(shù)據(jù)標(biāo)準(zhǔn)規(guī)范；（2）大氣科學(xué)外場(chǎng)試驗(yàn)數(shù)據(jù)資源建設(shè)；（3）科學(xué)試驗(yàn)場(chǎng)景可視化展示；（4）完善共享平臺(tái)。（高梅，張文華，李斌，朱孔駒）

圖1 2017年6月4日云雷達(dá)和C波段連續(xù)波雷達(dá)觀測(cè)數(shù)據(jù)融合的回波強(qiáng)度（a）、徑向速度（b）和數(shù)據(jù)的來源（c）（圖c中，黑色、綠色、藍(lán)色和紅色分別表示數(shù)據(jù)來自云雷達(dá)第1觀測(cè)模式、第2觀測(cè)模式、第3觀測(cè)模式和連續(xù)波雷達(dá)）Fig.1 Time-height cross-sections of reflectivity on 4 June 2017: (a) the merged reflectivity; (b) velocity; (c) data sources for reflectivity and velocity.The black dots in (a) and (b) indicate the CEIL-derived cloud-base heights.In (c), data from CR1, CR2,CR3 and CVPR are shaded in black, green, blue and red, respectively

圖2 2014年8月5日相態(tài)識(shí)別的RHI掃描圖：(a)為15∶41，方位角167°；(b)為15∶46，方位角162°；(c)為15∶51，方位角157°（圖中藍(lán)色實(shí)曲線表示水平風(fēng)速為0的“零線”，黑色帶箭頭實(shí)曲線表示流場(chǎng)，虛直線分別表示探空-20 ℃、-15 ℃、-10 ℃、0 ℃溫度線）Fig.2 The RHI of Class identification, wherein (a) at 1541 BST in azimuth 1670, (b) at 1446 BST in azimuth 162° and (c) at 1551 BST in azimuth 157°.The blue curves represent the “zero contour”, the solid curves with arrows represent the flow field,and the dotted straight lines are the sounding isotherms of -20 ℃, -15 ℃, -10 ℃, and 0 ℃, respectively

圖3 ISCCP衛(wèi)星資料觀測(cè)到的夏季青藏高原對(duì)流系統(tǒng)（TCS）數(shù)目分布(a)；高原地區(qū)夏季TCS移出的數(shù)目（等值線）及移出高原概率（陰影）分布(b)； (c) 和（b）一致，但為無法移出高原的分布Fig.3 (a) Spatial distribution of genesis of TCSs based on ISCCP dataset (1998–2004 JJA) on a 1°× 1° grid.(b) The number(contour) and probability (shaded) of TCSs that move out of TP.(c) Same as (b) except for TCSs that did not move out.The blue and red boxes represent central TP and eastern TP, respectively.The dotted contours show the topography at 3000 m and 5000 m

圖4 全球五大業(yè)務(wù)中心的集合預(yù)報(bào)系統(tǒng)對(duì)華南2013—2015年4—6月0～5天逐12 h累計(jì)降水量的預(yù)報(bào)性能（X 軸表示成功率，即1減去虛報(bào)率，Y 軸表示預(yù)報(bào)成功的概率；灰色虛線表示概率偏差，實(shí)線代表TS評(píng)分；藍(lán)色、紅色和黑色符號(hào)分別代表集合平均、概率匹配和確定性預(yù)報(bào)的評(píng)估結(jié)果；正方形（星號(hào)/三角形/圓圈）表示評(píng)估的降水閾值為0.1（5.0/15.0/30.0）mm/12h；每個(gè)集合預(yù)報(bào)系統(tǒng)在每種閾值下有10個(gè)符號(hào)，代表每隔12 h共5天的預(yù)報(bào)）Fig.4 Performance diagram of (a) ECMWF, (b) CMA, (c) KMA, (d) JMA, and (e) NCEP, for the EM (blue), PM (red), and CTL (black) forecasts of the 12 h accumulative precipitation during April?June 2013?2015.The X-axis refers to the success ratio, i.e., one minus the false alarm ratio (1-FAR).The Y-axis is the probability of detection (POD).The dashed lines indicate the frequency bias (FBIAS), while solid curves show the threat score (TS).The squares (asterisks/triangles/circles) indicate the performance at a threshold of 0.1 (5.0/15.0/30.0, resp.) mm/12h.Each EPS, at each threshold, has 10 symbols for the forecast periods from 12 hours to 5 days, at every 12 h interval

圖5 (a) 登陸臺(tái)風(fēng)距離臺(tái)風(fēng)中心不同半徑圓環(huán)內(nèi)（自0～1000 km每100 km一個(gè)圓環(huán)）降水率距平百分比的日變化信號(hào)（黑色實(shí)線，其中紅色實(shí)線和藍(lán)色實(shí)線分別代表諧波分析后得到的一波和二波信號(hào)）；(b) 標(biāo)準(zhǔn)化的臺(tái)風(fēng)登陸后12 h內(nèi)累積降水量隨登陸時(shí)間的變化（其中黑色、紅色和藍(lán)色實(shí)線分別代表距離臺(tái)風(fēng)中心0～900 km、0～400 km以及400～900 km范圍內(nèi)降水量的平均）Fig.5 (a) Smoothed diurnal cycle in rain-rate percentage anomaly averaged in each 100 km annular area as a function of local time (black curves) up to the annulus between 900- and 1000 km radii.Red and blue curves indicate the diurnal and semidiurnal harmonics, respectively; (b) Standardized cumulative precipitation within 12 h after landfall as a function of landfalling local time.Black, red, and blue curves indicate the average annuli between 0- and 900 km, between 0- and 400 km, and between 400-and 900 km radii from the storm center, respectively

圖6 數(shù)值理想模擬臺(tái)風(fēng)強(qiáng)度隨時(shí)間變化：（a）為10 m 最大風(fēng)速；（b）為中心最低氣壓（細(xì)線為集合預(yù)報(bào)試驗(yàn)，粗線為集合平均；藍(lán)線表示初始渦旋最大風(fēng)半徑為100 km；紅線表示初始最大風(fēng)半徑為60 km，徑向風(fēng)廓線雖半徑緩慢變化；綠線為初始最大風(fēng)半徑為60 km，徑向風(fēng)廓線隨半徑緩慢變化）Fig.6 Time evolutions of maximum 10 m wind speed (m s?1) and minimum central sea level pressure (hPa) in experiments R60B10 (blue), R60B05 (red), and R100B10 (green).Shown are results for all individual runs (thin) and their ensemble mean(thick) for each experiment.The two dashed horizontal lines indicate storm intensity between 25 m s?1 and 45 m s?1, helping identify the starting and ending times of the RI period for the three experiments (see text for more details)

圖7 不同SST環(huán)境下熱帶氣旋24 h強(qiáng)度變化和東（灰色線）/西（黑色線）切變的相關(guān)關(guān)系及對(duì)應(yīng)樣本數(shù)（柱狀）(線上實(shí)心代表相關(guān)系數(shù)通過95%顯著性檢驗(yàn)，空心代表未通過檢驗(yàn))Fig.7 Moving correlation between TC 24 h intensity change and zonal VWS with respect to various SST at 0.5 K intervals(line) and the corresponding samples (bar).The solid marks on line indicate the corresponding correlation is significant at 95%conf i dence level, while the hollow ones are not statistically significant

圖8 閃電始發(fā)先驅(qū)放電的輻射源定位和高速攝像記錄Fig.8 Radiation source location and high speed images of precursor discharge during lightning initiation stage

圖9 采用非結(jié)構(gòu)網(wǎng)格計(jì)算模擬框架進(jìn)行的標(biāo)量物質(zhì)傳輸測(cè)試（a～d為4種不同的傳輸算法的計(jì)算結(jié)果，測(cè)試算例為變形流場(chǎng)+高斯山）Fig.9 Scalar transport tests using the unstructured grid modeling system, (a?d) computational results from four different transport methods, the selected test case is deformational flow of two Gaussian hills

Advances in Severe Weather Research

1 Severe weather monitoring technology

1.1 Microphysical structures of cloud and precipitation observation and radar data merging algorithm in South China

To improve our understanding of the cloud and precipitation properties in South China, the field campaigns on heavy rainfall were carried out by Chinese Academy of Meteorological Sciences (CAMS) in 2017.Vertical structures of nonprecipitating and precipitating clouds, 3D structures of strong convective precipitations have been observed with X band phased array radar, a Ka-band solid-state transmitter cloud radar (CR), a C-band frequency modulated continuous wave (CVPR) vertical pointing radar (CVPR) and a laser ceilometer (CEIL) in Guangdong Province, China.We evaluated the ability of the two radars to deliver consistent observational data and developed an algorithm to merge the CR, CVPR and CEIL data.CR plays a key role in high-level cloud observations, whereas CVPR is important for observing low- and mid-level clouds and heavy precipitation.CEIL helps us diminish the effects of “clear-sky” in the planetary boundary layer.The corrected CR reflectivity and velocity data were then merged with the CVPR data and CEIL data to fill the gaps during the heavy precipitation periods and reduce the effects of Bragg scattering and fog on cloud observations in the boundary layer (Fig.1).(Liu Liping, Ruan Zhen, Cui Zhehu)

1.2 Advances in chinese dual-polarization and phased-array weather radars: observational analysis of microphysical and dynamical processed in a supercell in South China

In the summer of 2016, a joint radar observation of one phased-array weather radar and two polarimetric weather radars, which represent the advance of radar technique in China, was deployed in Foshan area for the study of severe storms over South China.The advantages of above dual-polarization and phased-array radars are discussed through the observational analysis of a supercell on 9 May 2016.The polarimetric signatures within the supercell are associated with specific microphysical processes that can reveal different stages of storm evolution.The hydrometeor classification algorithm (HCA) is a more straightforward and reliable method for nowcasting than conventional algorithms, which makes it possible for further recommendation in China.During the mature and dissipating stages of this supercell, observations of phased-array radar show detailed changes on short time scales that cannot be observed by parabolic-antenna radars.The convective trigger and megger of new cells are found in the peak-inflow region, and formation and dissipation of the hook echo are associated with the relative intensity of inflow and outflow.The above results demonstrate that the phased-array and dual-polarization radars recently developed in China are powerful tools for better understanding the storm evolution for nowcasting and scientific research.(Liu Liping, Wu Chong)

1.3 Convective cloud structures with dual polarimetric radar in Naqu, Tibet

Based on the data in July and August, 2014 in Naqu region, Tibet, detected with the mobile C-band dual polarimetric radar (C-POL) and Chinese new generation weather radar in Naqu (CINRAD/CD), the hailstorm cell and its dynamical, microphysical and thermodynamic characteristics are detailed by means of dual Doppler radar wind field retrieval and dual polarization radar hydrometeor class identification technique.In the RHI(range height indicator) images of ZH, ZDR and Class, the whole dynamical and microphysical processes can be obviously seen where the hydrometeor particles raise and grow up following the “zero line” and accompanying the echo intensity increase, then form a hail wall dropping down in the other side of the main updraft after overtaking it.From the successive three RHI scans, the phases of particle change from wet snow to hailstorm during the evolution process are found in one convective cell.The height of the echo is lower and its intensity is very weak when the cell is just triggered.However, when a large amount of wet snows appear in region above the melting level, it always hints that the updraft is so strong that the wet snows are brought back to high levels while they have not completely melted in the air below the melting level.Through physical processes such as condensation, rime, and attachment, the wet snow can rapidly grow into hailstones in just over 10 minutes.During the re-condensation of wet snows, the unstable structures are promoted and the updraft and downdraft are strengthened further because of latent heat release.Therefore, if numerous wet snows appear in a newly generated weak echo region above the melting level, it usually indicates that the strong updraft occurs in the region and a strong cell will develop rapidly.The configuration of wind field in different heights obtained by 3D wind field inversion shows that the existence of wind shear not only provides horizontal rotation, but also accelerates the vertical circulation.This is conducive to the conversion of unstable potential energy to kinetic energy, and cells can develop rapidly (Fig.2).(Hu Zhiqun)

1.4 Climatological beam propagation conditions for China

The vertical refractivity gradient (VRG) is critical to weather radar beam propagation.The most common method to calculate beam paths uses the 4/3 Earth radius model, which corresponds to standard refraction conditions.VRG spatial and temporal variations in the first kilometers above the surface are explored using 6 years sounding observations.Under the effects of both regional climatic and topographic conditions, VRG values for most of the radars are generally smaller than those of the standard conditions for much of the year.There are similar or slightly larger values at only a few radar sites.Smaller VRG values are more frequent and widespread, especially during rainy seasons when weather radar observations are important.In such conditions, beam heights estimated using standard atmospheric refraction are overestimated relative to actual heights for most of the radars.Underestimates are much less common and of much shorter duration.However,height deviations are acceptable for being well within the uncertainty of radar echo height owing to the 1°beamwidth.In coastal areas and the middle and lower reaches of the Yangtze River, radar observations should be applied with much more caution because of the greater risk of beam blockage and clutter contamination.This work better documents propagation conditions for radar electromagnetic waves and helps fully benef i t from observations taken from the new-generation weather radar network in China.(Wang Hongyan)

2 Research on weather over the Tibetan Plateau

2.1 The regional difference of TCSs in boreal summer

The Tibetan Convective Systems (TCSs) observed over the southeastern Tibetan Plateau (TP) in boreal summer show two local maximum centers in the central and eastern TP.Results indicate that there are more TCSs in the eastern TP than in central TP.20%?30% of TCSs over the eastern TP are associated with shear line, vortex, and low pressure systems, while 70%?80% of TCSs are formed locally.In contrast, 50%?70%of TCSs over the central TP are linked to synoptic weather systems.TCSs over the eastern TP tend to have smaller size, higher humidity, and more rainfall than those over the central TP, and also more likely to move out of the TP.TCSs are mostly generated in mid-afternoon with more rainfall in evening over the central TP and in midnight over the eastern TP (Fig.3).(Hu Liang, Xu Xiangde, Zhao Ping)

2.2 Large-eddy simulation (LES) of shallow cumulus over the Tibetan Plateau (TP)

Shallow cumulus (ShCu) forced by the surface sensible heat flux (H) and latent heat flux (LE) over the TP are simulated by a LES model.The macroscopic variables of ShCu (horizontal scale spectrum, cloud bottom height, cloud friction, vertical velocity in ShCu, etc) are calculated.The focus of the study is on the influence of strong turbulence on ShCu over the TP in low air density conditions.(Wang Yinjun, Xu Xiangde)

2.3 The statistical characteristics of planetary boundary layer height (PBLH) in the Tibetan Plateau(TP) and other regions in China

The PBLH is calculated by using the radiosonde sounding data including 128 sites in China.The frequency distributions of PBLH variations under different regimes (CBL the convective boundary layer, NRL the neutral residual layer, and SBL the stable boundary layer) can be well fi tted by a Gamma distribution at different regions of China, and the shape parameterskand scale parameterssvalues have been obtained, and the difference between them are pointed out.(Wang Yinjun, Xu Xiangde, et al.)

2.4 The influence of the Tibetan Plateau on air pollution and precipitation in eastern China

In eastern China (EC), the strong anthropogenic emissions deteriorate the atmospheric environment,building a south–north distribution of high aerosols harbored by the upstream Tibetan and Loess plateaus in China.The extreme precipitation events presented an interannual variability pattern similar to that of the frequent haze events over EC.Accompanied by the frequent haze events in EC, light rain frequency significantly decreased and extremely heavy precipitation events have occurred more frequently.There were significant differences in the interdecadal variation trends in light rain and rainstorm events between the highly aerosol polluted area in EC and the relatively clean area on the western plateaus of China.The aircraft measurements over EC confirmed that the diameters of cloud droplets decreased under high aerosol concentration conditions, thereby inhibiting weak precipitation process.(Xu Xiangde, Guo Xueliang, et al.)

2.5 Quantifying oceanic moisture exports to mainland China in association with summer precipitation

Oceanic moisture exports (OMEs) are considered major moisture sources for precipitation over Mainland China during the boreal summer season.The spatiotemporal structure of OME-based precipitation over Mainland China was explored.The main research conclusions are summarized as follows.(1) Climatologically,the OMEs were originated from the Arabian Sea (AS), the Bay of Bengal (BOB), and the South China Sea(SCS) made roughly equivalent contributions to the entire areal-averaged precipitation over Mainland China on a seasonal scale, but the preferred regions influenced by the three oceanic sources differed strongly from each other.(2) The relative contributions of OME from the three specific subsections to precipitation varied significantly on the sub-seasonal scale.During the onset of summer monsoon, the AS region ranked the first as an important oceanic source, followed by the BOB and the SCS, whereas during the withdrawal of summer monsoon, this order was reversed.(3) The interannual anomalies of OME-based precipitation from the SCS and the BOB regions were negatively correlated with those outside the AS region.(Chen Bin, Xu Xiangde)

2.6 A new path for the effect of the Tibetan Plateau heating on the global climate trough modulating the Asian-Pacific Oscillation (APO)

Two kinds of numerical experiments are carried out to simulate the synergic effect of the Tibetan Plateau(TP) dynamical and thermal anomalies and the individual effect of the TP thermal anomaly, through adjusting the TP topographic height and modulating the surface vegetation type of the TP, respectively.The results show that the two kinds of numerical experiments reproduce the physical processes that the TP triggers anomalous vertical circulations between Asia and the North Pacific and therefore successfully force the typical structure of the summer Asia-Pacific Oscillation (APO).The above results confirm that the TP thermal anomaly dominates during summer, and it should be considered as an important contributor to the variation of the summer APO.This work reveals a new path that the TP thermal anomaly can affect the global climate through regulating the APO.(Liu Ge, Zhao Ping, Chen Junming)

3 Heavy rainfall and strong convection

3.1 Evaluation of quantitative precipitation forecasts by TIGGE ensembles for South China during the presummer rainy season

Based on The Observing System Research and Predictability Experiment (THORPEX) Interactive Grand Global Ensemble (TIGGE) dataset, this study evaluates the performance of five global ensemble prediction systems (EPSs) from the European Centre for Medium-Range Weather Forecasts (ECMWF), US National Centers for Environmental Prediction (NCEP), Japan Meteorological Agency (JMA), Korean Meteorological Administration (KMA), and China Meteorological Administration (CMA), respectively, on predicting the presummer rainy season (April?June) precipitation in South China.Evaluation for the 5 day forecasts in three seasons (2013?2015) indicates higher skill of the probability matching forecast than either the simple ensemble mean or the deterministic forecasts.The EPSs overestimate the light-to-heavy rainfall (0.1 to 30 mm/12h),while underestimate the heavier rainfall (＞30 mm/12h), with JMA being the worst.By analyzing the synoptic situations predicted by the identified skillful (ECMWF and CMA) and less skillful (JMA) EPSs, as well as the ensemble sensitivity for four representative torrential rainfall cases, the low-level southwesterly flow upstream of the torrential rainfall regions that transports warm, moist air to South China is found to be a key synoptic factor that controls the QPF.The results also suggest that prediction of the locally produced torrential rainfall is more challenging than the more extensively distributed torrential rainfall.A slight improvement in the performance by shortening the forecast lead-time from 30?36 h to 18?24 h and to 6?12 h is noticed in the large-scale forcing case, but not in the locally produced cases (Fig.4).(Huang Ling, Luo Yali)

3.2 Synoptic analysis of extreme hourly precipitation in Taiwan during 2003?2012

This study investigates the statistical characteristics of extreme hourly precipitation over Taiwan during 2003?2012 that exceeds the 5, 10, and 20 years return values and 100 mm h?1.All the extreme precipitation records are classified into four types according to the synoptic situations under which they occur: tropical cyclones (TCs), fronts, weak-synoptic forcing, and vortex/shear line types.The TC type accounts for over three-quarters of the total records, while the front type and weak-synoptic forcing type are comparable(9%–13%).Extreme hourly precipitation is mostly caused by meiyu fronts during May–mid–June and by TCs during July–October.The TC type tends to have a long duration time (＞12 h) with a symmetrical evolution of hourly rainfall intensity, while the front type and weak-synoptic forcing type mainly occur over a short period(＜6 h) with a slightly asymmetrical evolution pattern.The TC type is further divided into seven subtypes according to the location of the TC center relative to the island.When the TC center is over the island or near the coastline (distance＜100 km), the spatial distribution of subtypesⅠ?Ⅳ is largely determined by the interaction between the TC circulation and topography when a TC center is over the northwest, south, east, or northeast portion of Taiwan, respectively.When the TC center is far away (distance＞100 km) from the island,the strength of the environmental southwesterly or northeasterly winds and the impingement of TC circulation on the eastern side of the Central Mountain Range are also key factors determining the spatial distribution of subtypesⅤ?Ⅶ.(Luo Yali, Wu Mengwen)

3.3 An extreme rainfall event in coastal South China during SCMREX-2014: Roles of rainband and echo trainings

This study investigates an extreme rainfall event that occurred in coastal South China on 11 May 2014 using integrated observations provided by the South China Monsoon Rainfall Experiment (SCMREX).The extreme rainfall is mostly produced by two consecutive mesoscale convective systems (MCSs) that consist of multiple meso-β-scale rainbands of 30?80 km length.A dual-Doppler wind synthesis shows that the two MCSs are maintained by lifting southerly oceanic flows over a quasi-stationary mesoscale outflow boundary (MOB),enhanced by convectively generated weak cold pool, along the coastline.The subsequent northeastward “echo training” of convective cells, under the influence of environmental southwesterly flows, leads to the formation of various meso-β-scale rainbands.The subsequent propagation of the rainbands, together with the “echo training”, along the coastline accounts for the production and distribution of extreme rainfall.An analysis of the S-band dual-polarization radar data reveals the generation of more moderate-sized raindrops and slightly larger maximal size of raindrops with stronger convective updrafts and more active ice-phase microphysical processes in the second MCS with heavier-rain-production than in the first MCS, likely due to the presence of larger convective available potential energy in the afternoon.The generation of the multiple rainbands in the second MCS is examined.The results show that its leading rainband develops a bow structure and later breaks into two parts due to the development of strong westerly rear inflows.Shortly after they strengthen into two well-organized rainbands as a result of convective initiation near the MOB, leading to a growing member of heavy-rain-producing bands.A conceptual model of the extreme-rain-producing MCS consisting of multiple rainbands along the coasts of South China is established.(Luo Yali, Liu Xi, Zhang Dalin)

3.4 Features of extreme hourly precipitation over South China during the presummer rainy season

The general features of extreme hourly rainfall (＞60 mm h?1) over South China (SC) during the presummer rainy season (April–June) of 2011–2017 are studied using an 8 km-resolution gridded hourly rainfall analysis based on rain gauge measurements, weather maps, reanalysis data, and every 6 min mosaics of composite radar reflectivity (1 May–15 June 2013–2017).A total of 2659 extreme hourly rainfall records are found over the analysis region (covering mainly Guangxi and Guangdong provinces), with 262, 1405, 992 in April, May, and June, respectively.Three areas with high frequency of occurrence are located in the north Guangxi (NGX), southeast Guangxi (SEGX), and central coastal area of Guangdong (MCGD), respectively.Their diurnal variations in the three areas, respectively, show a single peak around the middle night, a major peak in the early morning, and a minor peak in the late afternoon, and a lack of distinct peak.The hourly extreme rainfall records are classified into five types according to the synoptic situations under which they occur: the surface front, low-level vortex, low-level shear line, weak synoptic forcing, and tropical cyclone.They contribute 35.5%, 8.9%, 16.4%, 38.1%, and 1%, respectively, to the total occurrence and present distinctive characteristics in regional distribution over the analysis region.The extreme rainfall in Guangxi occurs generally with favorable dynamical lifting by the synoptic weather systems, while that occurs in Guangdong often without.The mosaics of composite radar reflectivity patterns are analyzed for the 1716 extreme rainfall records during 1 May–15 June of 2013–2017.Among them, 51.4% are produced by MCSs with a prominent convective line (linear MCSs), 7.7% by MCSs consisting of multiple rainbands (MRBs), and 41% by smaller-scale storms.The extreme rainfall produced by the linear MCSs is distributed quite extensively over South China, with a major high frequency area in north Guangxi and a secondary high frequency area in central costal Guangdong.The extreme rainfall produced by the MRB-type MCSs mostly influences central inland and coastal Guangdong, and that by smaller-scale storms distributes extensively over South China but have higher frequency in Guangxi than in Guangdong except for central Guangdong.Further analysis of the duration and the development types (abrupt, growing, continuous) of the hourly extreme rainfall events also demonstrates distinct regional features.(Luo Yali, Chen Yangruixue, Li Zhenghui)

3.5 Assimilating Doppler radar observations with an ensemble Kalman filter for convectionpermitting prediction of convective development in a heavy rainfall event during the pre-summer rainy season of South China

This study examines the effectiveness of an ensemble Kalman filter based on WRF model to assimilate Doppler-radar radial-velocity observations for convection-permitting prediction of convection evolution in a high-impact heavy-rainfall event over coastal areas of South China during the pre-summer rainy season.An ensemble of 40 deterministic forecast experiments (40 DADF) with data assimilation (DA) are conducted,in which the DA starts at the same time but lasts for different time spans (up to 2 h) and with different time intervals of 6, 12, 24, and 30 min.The reference experiment is conducted without the DA (NODA).It is found that the DA experiments generally improve the convection prediction.The 40 DADF experiments can forecast a heavy-rain-producing MCS over land and an MCS over the ocean with high probability, despite slight displacement errors.Compared with the experiments using the longer DA time intervals, assimilating the radial-velocity observations at 6-min intervals tends to produce better forecasts.The experiment with the longest DA time span and shortest time interval shows the best performance.However, a shorter DA time interval (e.g., 12 min) or a longer DA time span does not always help.In the optimal DADF experiment, an improved representation of the initial state leads to dynamical and thermodynamic conditions that are more conducive to earlier initiation of the inland MCS and longer maintenance of the offshore MCS.(Bao Xinghua,Luo Yali)

3.6 Observational analysis of a record-breaking rain-producing convective storm influencing Guangzhou during SCMREX-2017

A meso-β-scale storm that produced extreme rainfall over Guangzhou megacity on 7 May 2017 is studied using integrated observations by various instruments deployed during the IOP of the South China Monsoon Rainfall Experiment (SCMREX), and the high-resolution gridded datasets generated using a four-dimensional variational Doppler radar analysis system (4D VDRAS).Results show the initiation and subsequent organization of convective cells in a meso-β-scale convergence region on the northern edge of Guangzhou,where an onshore low-level jet stream with ample moisture embedded in a weak ambient flow decelerated as it moved across the urban areas, and then converged as it hit the mountainous region to the north.The enhanced urban frictional dissipation and the topographical blocking effects accounted for the formation of such a convergence region, while the urban heat island effect helped maintain conditionally unstable columns during this night-time period with a relatively deeper boundary layer over the Megacity.The onshore flows gave rise to the development of deep moist columns over the Megacity.The deep moist columns, the weak environmental winds, and the moderate height of LNB (about 10 km ASL) favored the storm’s stagnation and high precipitation efficiency by inhibiting intense updrafts and detrainment of hydrometeors at higher levels.

Examination of the VDRAS outputs revealed the presence of a moderately deep, meso-γ-scale, circular convective core (C3) within a meso-β-scale vortex.Vorticity diagnosis indicated that tilting of environmental horizontal vorticity played a key role in the early development of the vortex, while stretching effect associated with low-level convergence was responsible for the vorticity maintenance.The thermal buoyancy resulting from latent heat release was the prime forcing and the dynamical vertical perturbation pressure gradient force was secondary in supporting the upward motion in the C3.A large number of raindrops with moderate-to-large sizes were produced through active warm rain processes, which might be attributed to the presence of large sea-salt aerosols in the southerly inflow acting as cloud condensation nuclei.(Luo Yali, Ma Ruoyun, Zhang Dalin, Chen Mingxuan)

3.7 Synoptic control of convective rainfall rates and cloud to ground lightning frequencies in warmseason mesoscale convective systems over North China

This study examines whether environmental conditions can control convective rainfall rates and cloudto-ground (CG) lightning frequencies in mesoscale convective systems (MCSs) over North China (NC).Sixty identified MCSs over NC during June–August of 2008–2013 were classified into four categories based on their high/low convective rainfall rates (HR/LR) and high/low lightning frequencies (HL/LL), i.e., HRHL,HRLL, LRHL and LRLL MCSs.MCSs with HR (HL) occurred most frequently in July (August), while those with LR or LL occurred most frequently in June; they followed closely seasonal changes.All MCSs were apt to form during afternoon hours.HRLL MCSs also formed during evening hours while HRHL MCSs could occur at any time of a day.A composite analysis of environmental conditions shows obvious differences and similarities among the HRHL, HRLL and LRLL categories, while the LRHL MCSs exhibited little differences from the climatological mean due to its small sample size.Both the HRHL and HRLL MCSs occurred in the presence of upper-level anomalous divergence, a midlevel trough, and the lower-tropospheric southwesterly transport of tropical moist air.In contrast, LRLL MCSs took place due to daytime heating over mountainous regions, with little midlevel forcing over NC.The HRHL, HRLL, LRHL and LRLL categories exhibited orders of the highest to smallest convective available potential energy and precipitable water but the smallest to largest convective inhibition and lifted indices.It is concluded that the environmental conditions determine to some extent convective rainfall rates and CG lightning activity, although some other processes (e.g., cloud microphysics) also play certain roles, especially in CG lightning production.(Xia Rudi, Zhang Dalin)

3.8 On the mechanism of the North Atlantic Oscillation

In this study, the relationship between the Madden-Julian Oscillation (MJO) and the North Atlantic Oscillation (NAO) is investigated based on the European Centre for Medium-Range Weather Forecasts Re-Analysis (ERA-Interim) dataset, which provides a guideline for the extended-range prediction.When the MJO is active, negative (positive) NAO events are preceded by enhanced MJO closely resemble the northern annular mode (NAM).The MJO-related NAO events are also long-lived, persisting for approximately 30 days.In contrast, when the MJO is inactive, there is little change to the zonally symmetric component of the lower stratospheric zonal wind, these events are relatively short-lived, lasting for about 10 days, and the NAO corresponds to a dipole that is confined to the North Atlantic.(Jiang Zhina)

3.9 A numerical study of the convection triggering and propagation associated with sea breeze circulation over Hainan Island

Based on high-resolution numerical simulations, the effects of sea breeze circulation (SBC) on associated convective initiation and propagation over Hainan Island are investigated.The results show that the blocking of synoptic wind on the windward side of the mountains inhibits the occurrence of windward-side valley breeze and sea breeze.In the context of low mountains, weak synoptic wind not only favors the combination of the SBC with the valley breeze circulation on the leeward side of the mountains but also makes the convergence zone between the synoptic wind and the leeward-side combined circulation stay near the tops of the mountains,leading to the accumulation of large quantities of water vapor taking place over the mountaintops.The combined effect of enhanced water vapor and high temperatures over the mountaintops can result in the generation of much larger convective available potential energy (CAPE) than the surrounding areas.The coupling of the large CAPE with the considerable converging ascending motion and abundant water vapor near the tops of the mountains promotes the occurrence of a strong convection band (CB) over the mountaintops.During the CB propagation under the steering of synoptic wind, the flow across the CB continuously and strongly opposes the SBC front and the fronts of the Kelvin-Helmholtz billows behind, resulting in successive and strong lifting motion as well as large quantities of water vapor and high CAPE near these fronts.These lead to an apparent downstream wave-like propagation of the CB.(Liang Zhaoming, Wang Donghai, Liu Ying)

4 Typhoon research

4.1 Diurnal cycle of rainfall associated with landfalling tropical cyclones in China from rain gauge observations.

The diurnal variation of rainfall over China associated with landfalling tropical cyclones (TCs) is investigated using hourly rain gauge observations obtained from conventional meteorological stations in China.The harmonic analysis shows an obvious diurnal signal in TC rainfall with a rain-rate peak in the early morning and a minimum in the afternoon.The diurnal cycle in the outer region (between 400 and 900 km radii from the storm center) is found to be larger than in the core region (within 400 km of the storm center.As the result of this diurnal cycle, TCs making landfall at night tend to have cumulative precipitation, defined as the precipitation cumulated from the time at landfall to 12 h after landfall, about 30% larger than those making landfall around noon or afternoon (Fig.5).(Hu Hao,Duan Yihong, Wang Yuqing)

4.2 Effect of the initial vortex structure on intensification of a numerically simulated tropical cyclone

The dependence of intensification rate (IR) of a tropical cyclone (TC) on its initial structure, including the radius of maximum wind (RMW) and the radial decay rate of tangential wind outside the RMW, is examined based on ensemble of simulations using a nonhydrostatic axisymmetric cloud-resolving model.It is shown that the initial spinup period is shorter and the subsequent IR is larger for the storm with the initially smaller RMW or with the initially more rapid radial decay of tangential wind outside the RMW.The results show that the longevity of the initial spinup period is determined by how quickly the inner-core region becomes nearly saturated in the middle and lower troposphere and thus deep convection near the RMW is initiated and organized.Because of the larger volume and weaker Ekman pumping, the inner-core of the initially larger vortex takes longer time to become saturated and thus experiences a longer initial spinup period.The vortex initially with the larger RMW (with the slower radial decay of tangential wind outside the RMW) has lower inertial stability inside the RMW (higher inertial stability outside the RMW) develops more active convection in the outer-core region and weaker boundary-layer inflow in the inner-core region and thus experiences lower IR during the primary intensification stage (Fig.6).(Xu Jing, Wang Yuqing)

4.3 Characteristics of tropical cyclone-induced winds over complex terrains of China coast

Based on the ERA-Interim reanalysis data (0.5°× 0.5°, 6-hourly) from 1979 to 2015, the best track dataset from the CMA and the JTWC, characteristics of tropical cyclone (TC) low-level wind fields around the Taiwan and the Hainan Island are statistically analyzed, respectively.Main results are summarized here.(1) When TC centers locate around the island, positive and negative vorticity belts are found to distribute alternately in TC’s periphery between the islands.(2) The TC-induced high winds mainly appear on the east and northeast of Taiwan Island and the Taiwan Strait.For Hainan Island, high winds tend to occur on the northeast of the island and the Beibu Gulf.Winds on the island are much weaker.(3) The averaged radius of maximum wind (RMW)for TCs around Taiwan Island is 51.3 km, which is smaller than 58.3 km around Hainan Island.On average,RMWs of TCs on the west (northeast) of Taiwan Island are the largest (smallest).For Hainan Island, RMWs of TCs on the island (the northwest of the island) are the largest (smallest).(4) The 34-kts wind circle transforms to a certain degree when a TC approaches the islands.The longest wind radius is located on the eastern part of the TC circulation.On average, the wind circle may have the largest deformation when a TC occurs on the northwestern (southwestern) side of Taiwan (Hainan) Island.(Li Ying, Xue Lin)

4.4 Climatic characteristics of binary tropical cyclones in the Northwest Pacific Ocean

Based on typhoon best track data of China Meteorological Administration and NCEP global reanalysis data, this study analyzes characteristics of binary tropical cyclones in the Northwest Pacific during 1951–2014 by using the objective determination standard.Results show that there were 699 pairs of binary typhoons in the Northwest Pacific during 1951–2014.Among these cases, there were 446 pairs of typical binary typhoons and 253 pairs of typical cases, respectively, or 63.8% and 36.2% of the total.The proportion of typical cases increased with the separation distance decreased, while the proportion of atypical cases decreased with the separation distance decreased.Binary typhoons mainly appeared between 110° to 150°E and 10° to 30°N.Most cases appeared in the northern South China Sea and east of Philippines, with Taiwan Island being included.Binary typhoons most frequently occurred in August and September.When the speed of typical binary typhoons moving towards each other reached the peak, binary typhoons mainly orientated in the east to west direction.At this time typhoons were controlled by the easterlies south of the subtropical high.In this situation,the east typhoon moved toward the west typhoon quickly.When the anticlockwise angular velocity of typical binary typhoons reached the peak, the binary cases distributed northeast to southwest or east-northeast to west-southwest, appearing in the west and southwest edge of the subtropical high and mainly controlled by southeasterly flow, thus benef i ted the anticlockwise rotation between the typical binary typhoons.(Ren Fumin)

4.5 Comparison of the effect of easterly and westerly vertical wind shear on tropical cyclone intensity change over the western North Pacific

The effects of vertical wind shear (VWS) with different directions on tropical cyclone (TC) intensity change are compared in this statistical study based on TCs occurring between 1982 and 2015 over the western North Pacific (WNP).Results show that a westerly VWS has a much higher correlation (?0.36) with change in TC intensity than an easterly VWS (?0.07) over the WNP, especially a southwesterly VWS (?0.43).Sea surface temperature (SST) is found to modulate the effect of VWS on TC intensity change as it has a close relationship with zonal VWS (?0.48).The favorable effect of SST, which increases with the increase in easterly VWS, could offset the detrimental effect of VWS, leading to a relatively low correlation coefficient between easterly VWS and TC intensity change.By contrast, westerly VWS increases with decreasing SST, and the largest correlation coefficient appears when SST is around 301 K.Therefore, it is suggested that the direction of VWS as well as its value should be taken into consideration in statistical models used to predict TC intensity(Fig.7).(Wei Na, Li Ying)

5 Lightning research

5.1 Assimilation of special observation data and research on detecting and forecasting methods for severe thunderstorms

A three-dimensional charge-discharge numerical model is used, in a semi-idealized mode, to simulate a thunderstorm cell.Characteristics of the graupel microphysics and vertical air motion associated with lightning initiation are revealed, which could be useful in retrieving charge strength during lightning when no chargedischarge model is available.The results show that the vertical air motion at the lightning initiation sites (Wini)has a cubic polynomial correlation with the maximum updraft of the storm cell (Wcell-max), with the adjusted regression coefficient of approximately 0.97.Meanwhile, the graupel mixing ratio at the lightning initiation sites (qg-ini) has a linear correlation with the maximum graupel mixing ratio of the storm cell (qg-cell-max)and the initiation height (zini), with the coefficients being 0.86 and 0.85, respectively.These linear correlations are more significant during the middle and late stages of lightning activity.A zero-charge zone, namely, the area with very low net charge density between the main positive and negative charge layers, appears above the area of qg-cell-max and below the upper edge of the graupel region, and is found to be an important area for lightning initiation.Inside the zero-charge zone, large electric intensity forms, and the ratio of qice (ice crystal mixing ratio) to qg (graupel mixing ratio) illustrates an exponential relationship to qg-ini.These relationships provide valuable clues to more accurately locating the high-risk area of lightning initiation in thunderstorms when only dual-polarization radar data or outputs from numerical models without charging/discharging schemes are available.The results can also help understand the environmental conditions at lightning initiation sites.(Wang Fei)

The charge structure and formation in Typhoon Molave (2009) before and after its landfall and during the decaying stage are investigated using satellite, lightning detection data and a mesoscale simulation.Results show that Molave was intensifying prior to landfall, with a well-defined eye and relatively high-frequency lightning activity in the eyewall.Convection near the eyewall exhibited positive tripole charge structure, with negative charge located between the levels of ?25 ℃ and ?10 ℃ sandwiched by two positive charge regions.However, the charge structure of convection becomes negative bipole, along with negative charge in the middle and positive charge at the bottom of convective clouds after Molave reaches its maximum intensity.The charge structure of eyewall convection is closely associated with typhoon intensity, but not in a direct correlation with landfall.The outer spiral rainband cells display a positive tripole charge or a positive bipole in different stages of the typhoon.Previous studies suggested that outer rainband only features a positive bipole charge structure.The positive tripole charge structure forms with different mechanisms: one resembles that in the eyewall, and the other has a positive charge region composed by hail and a positive bipole region composed by graupel and ice crystals in the upper levels.During the decaying stage of the storm, weak convection is mainly featured by a negative bipole, similar to the terrestrial thunderstorms in the dissipative stage.(Xu Liangtao, Zhang Wenjuan)

5.2 Influence of the Canton Tower on the cloud-to-ground lightning in its vicinity

The influence of the Canton Tower on cloud-to-ground lightning in its vicinity is analyzed based on the data obtained by Lightning Location System (LLS) of the Guangdong Power Grid Corporation from 1999 to 2015.The LLS data obtained before (1999–2005) and after (2010–2015) the erection of the Canton Tower were compared.The flash/stroke density showed a significant increase in the immediate vicinity around the Canton Tower within a radius of 1 km, and a clear decrease in the annular vicinity with a radius from 1 km to 4 km.The arithmetic mean and median LLS-inferred peak current of strokes occurred in 1 km radius area around the Canton Tower exhibited a significant increase, while that of the strokes occurred beyond 1 km did not show a clear change.It is speculated that the Canton Tower induced lots of upward flashes and attracted some downward flashes around it within several kilometer radius to itself.(Lyu Weitao, Qi Qi, Ma Ying, Yao Wen)

5.3 Characteristics of initial stage for triggered lightning

Upward positive leader (UPL) is the key discharge during the initiation of upward negative lightning and the connection of downward negative lightning.However, it is not enough to understand the physical mechanism during the inception stage of UPL.Based on the synchronous data of channel-base current and close-distance electric field generated by triggered lightning detected in Guangdong comprehensive observation experiment for lightning discharge, the discharge characteristics of UPL during the initial stage of triggered lightning are analyzed.The waveform of channel-base current for upward positive leader exhibits discontinuous pulses, which occur separately or in groups.Statistical analysis shows that the geometric mean(GM) values of current peak, 10%–90% rise time, duration, half peak width, transferred charge, and pulse interval for the pulses generated before the appearance of continuously upward leader are 28 A, 0.33 μs, 2.3 μs, 0.73 μs, 27 μC, and 25 μs, respectively.The GM values of the corresponding parameters for the pulses generated during the initiation of continuously upward leader are 32 A, 0.54 μs, 3.5 μs, 0.93 μs, 46 μC and 24 μs, respectively.Compared with the discharge process before continuous propagation of UPL, the discharge process during the inception of continuously upward leader transfers more charge.Also, the waveforms of close-distance electric field show obviously a step-change, and each pulse of electric field corresponds to the channel-base current pulses in time sequence, indicating that the cascade development has characteristics of upward positive leader during the initial stage of artificially triggered lightning.(Zhang Yang, Zhang Yijun,Lyu Weitao, Zheng Dong, Xu Liangtao)

5.4 Detailed discharge process of lightning initial stage

During 2016, we developed continuous interferometer which can accurately locate and detect the breakdown activity with sub microsecond and meter-scale resolutions.We observed a triggered lightning on 9 June 2016 during the Guangdong Comprehensive Observation Experiment on Lightning Discharge.The discharge produced 54 precursor current pulses (PCPs) over 883 ms during the rocket’s ascent.The interferometer observations show that the PCPs were produced by breakdown at the ascending tip of the rocket,and that individual PCPs were produced by weak upward positive breakdown over meter-scale distances,followed by more energetic, fast downward negative breakdown over several tens of meters distance.The average propagation speeds were 5×106m s?1and 3×107m s?1, respectively.The sustained upward positive leader (UPL) was initiated by a rapid, repetitive burst of 14 precursor pulses.Upon initiation, the VHF radiation abruptly became continuous with time.Significantly, breakdown during the UPL appeared to extend the discharge in a similar manner to that of the precursor pulses.Our research provides valuable insights into the triggering mechanism and positive breakdown processes (Fig.8).(Zhang Yang, Zhang Yijun, Lyu Weitao,Zheng Dong)

5.5 Characteristics of the initial stage and return stroke currents of rocket-triggered lightning flashes in South China

This study investigates the initial stage (IS) and return stroke (RS) currents of 50 triggered lightning flashes (TLFs) that were conducted in South China.The IS of the negative TLFs has a longer duration and larger average current, charge transfer, and action integral than those reported elsewhere, with geometric means (GMs) of 347.9 ms, 132.5 A, 45.1 C, and 10.0× 103A2s, respectively.Two positive TLFs containing no RS have much greater average currents, charge transfers, and action integrals in the IS when compared with the negative TLFs.The RS has a greater peak current (17.2 kA; GM, same as below), charge transfer within 1 ms (1.3 C), and action integral within 1 ms (5.8×103A2s), and shorter 10% to 90% rise time (0.4 μs)than elsewhere.The peak current is prominently correlated with the rate of rise, charge transfer within 1 ms,and action integral within 1 ms.Furthermore, when the total duration of the RS and any following continuing currents is longer than 40 ms, the peak current, charge transfer within 1 ms, and action integral within 1 ms of the RS are seldom greater than 25 kA, 2.6 C, and 15×103A2s, respectively.It is indicated that TLFs containing RSs tend to have a longer duration but a smaller charge transferred during the IS than those without RS.The peak current of the RS is weakly correlated with its preceding silence period when there was no channel base current.(Zheng Dong, Zhang Yijun, Zhang Yang, Lyu Weitao)

5.6 Spatial-temporal characteristics of lightning flash size in a supercell storm

The flash sizes of a supercell storm, in New Mexico on October 5, 2004, are studied using the observations from the New Mexico Lightning Mapping Array and the Albuquerque, New Mexico, Doppler radar (KABX).First, during the temporal evolution of the supercell, the mean flash size is anti-correlated with the flash rate, following a unary power function, with a correlation coefficient of ?0.87.In addition, the mean flash size is linearly correlated with the area of reflectivity ＞30 dBz at 5 km normalized by the flash rate,with a correlation coefficient of 0.88.Second, in the horizontal, flash size increases along the direction from the region near the convection zone to the adjacent forward anvil.The region of minimum flash size usually corresponds to the region of maximum flash initiation and extent density.The horizontal correspondence between the mean flash size and the flash extent density can also be fi tted by a unary power function, and the correlation coefficient is ＞0.5 in 50% of the radar volume scans.Furthermore, the quality of fi t is positively correlated with the convective intensity.Third, in the vertical direction, the height of the maximum flash initiation density is close to the height of maximum flash extent density but corresponds to the height where the mean flash size is relatively small.In the discussion, the distribution of the small and dense charge regions when and where convection is vigorous in the storm, is deduced to be responsible for the relationship that flash size is temporally and spatially anti-correlated with flash rate and density, and the convective intensity.(Zheng Dong, Zhang Yijun)

5.7 Relationship between lightning activity and vertical airflow characteristics in thunderstorms

Cloud-to-ground (CG) lightning data, wind field data derived from dual-Doppler radars, and radar reflectivity data are combined to investigate the relationships between CG lightning and the parameters associated with vertical airflow.A new method for the identification of thunderstorm cells based on the aggregation of flashes is developed.It is found that approximately 79.1% of CG flashes are located in the region featuring weak vertical velocity at the 0 ℃ level, ranging from ?5 to 5 m s?1, with the majority in the weak updraft region, especially for negative CG lightning and the CG lightning in the initial stage of thunderstorms.The CG lightning rate is correlated with the volume of updraft for vertical velocities within certain limits.The sum of absolute precipitation ice mass flux in the region from 7 to 11 km is more significantly correlated with the CG flash rate, with correlation coefficients of 0.73, 0.71, and 0.74 for the initial, mature, and dissipating stages of thunderstorms, respectively.On average, the updraft in the thunderstorm at the stage when the last CG flash occurs accounts for a much smaller ratio to the whole volume of the thunderstorm than that corresponding to the first CG flash.The maximum updraft and maximum height of the 10 and 20 m s?1updraft speeds are close for the first and last CG flashes, indicating the dependence of the lightning on strength of updraft.It is deduced that layered large-range charges may be more conducive to the generation of CG flashes than charge pockets in the thunderstorm.(Zheng Dong, Zhang Yijun, Liu Liping)

5.8 Low-frequency e-field detection array (LFEDA): Construction and preliminary results

In recent years, locating total lightning at the VLF/LF band has become one of the most important directions in lightning detection.The low-frequency e-field detection array (LFEDA) consisting of nine fast antennas was developed by the Chinese Academy of Meteorological Sciences in Guangzhou between 2014 and 2015.This study documents the composition of the LFEDA and a lightning-locating algorithm that applies to the low-frequency electric field radiated by lightning pulse discharge events (LPDEs).Theoretical simulation and objective assessment of the accuracy and detection efficiency of LFEDA have been done using Monte Carlo simulation and artificial triggered lightning experiment, respectively.The former results show that having a station in the network with a comparatively long baseline improves both the horizontal location accuracy in the direction perpendicular to the baseline and the vertical location accuracy along the baseline.The latter results show that detection efficiencies for triggered lightning flashes and return strokes are 100% and 95%,respectively.The average planar location error for return strokes of triggered lightning flashes is 102 m.By locating LPDEs in thunderstorms, we find that LPDEs are consistent with convective regions as indicated by strong reflectivity columns and present a reasonable distribution in the vertical direction.In addition, the LFEDA can reveal an image of lightning development through mapping the channels of lightning.Based on three-dimensional locations, the vertical propagation speed of the preliminary breakdown and the changing trend of the leader’s speed in an intra-cloud and a cloud-to-ground flash are investigated.Results show that the LFEDA has the capability for three-dimensional location of lightning, which provides a new technique for the study of lightning development characteristics and thunderstorm electricity.(Zheng Dong, Zhang Yang, Zhang Yijun, Lyu Weitao)

5.9 Identification method and analysis on lightning flash initiation phase and size

Based on the observation of Lightning Mapping Array, the statistical distribution of the characteristic parameters describing the lightning flash initiation and size in a supercell storm occurring on October 5, 2004 in New Mexico, United States is studied.A method automatically identifying the start and end of the negative leaders at initial stage of lightning is developed, and the flash convex hull, total channel length, horizontal and vertical extent are used to represent the flash size.The distributions and characteristics of flash initiation and size in this storm are summarized here.(1) It is found that the median values of the duration, three-dimensional displacement, vertical displacement and the average displacement velocity for the upward (downward)negative leaders at IS are 13.5 ms (7.5 ms), 1.4 km (1.0 km), 0.9 km (0.5 km), and 9.2×104m s?1(1.2×105m s?1), respectively.In addition, the average flash initiation velocity decreases with height from 6 km to 11 km.With time going by, the speed of upward negative leader at initial stage decreases before 24 ms (to ensure the samples is larger than 100), while that of the downward negative leader increases before 12 ms (to ensure the samples is larger than 50).The negative leaders are dominantly tilted at initial stage, considering that the median angles between the 3-D displacement direction and the vertical direction are 40° for upward leader and 54° for downward leader, respectively.(2) The probability density distribution of flash size described by flash convex hull and total channel length can be well fi tted by negative power function, and both of them bias to the small value, which means that the distribution and evolution of flash convex hull is consist with that of total channel length.In the vertical direction, the average flash sizes at middle and upper levels are smaller than at the lower levels.The median of flash duration is 271.0 ms, and the mean of that is 329.1 ms.The flash duration time and size are not significantly correlated.The flash with long duration time is not necessarily large.Moreover, the median of flash horizontal extent is 6.1 km while the vertical extent is 4.3 km, and there are 83%of flashes whose horizontal extent is greater than vertical extent.The flashes with horizontal extent greater than vertical extent are mainly initiated at 8.5 km, and those with vertical extent greater than horizontal extent are mainly initiated at 11 km.It is shown that the more horizontal the leader displacement direction at Initial stage is, the less the flash vertical extent is, which indicates that the leader displacement direction at the initial stage has an important influence on flash vertical extent.(Zheng Dong, Zhang Yijun)

5.10 Development of lightning detection technology and outfield experiment

During the 12th Guangdong comprehensive observation experiment on lightning discharge and the first observation on electric activity in Tibet thunderstorms, detailed observations of lightning discharge process and total flash in thunderstorm have been conducted, and significant progress in triggered lightning and research on lightning physical mechanism have been gained.The specific progresses are summarized as follows.(1)The numbers and success rate of triggered lightning were 16% and 80%, respectively, both keeping relatively on a high level.(2) Detailed positioning of lightning channel by continuous interferometer was improved.The long pre-trigger technology was developed, which helped describe the process of lightning discharge in a detailed and on a high time-resolution.(3) A further improvement of low frequency electric-field detection array (LFEDA) was made.Using the LFEDA, lightning activity in thunderstorms during 2017 was observed.The location was finer by optimizing the 3D location method.(4) A new sight about lightning initial discharge and radiation mechanism was obtained for the first time, based on the detailed observation on the instability discharge stage (precursor discharge).(5) The observation ability for high building lightning was significantly improved by optimizing the observation scheme.Based on could-to-ground (CG) lighting location data of Guangdong, the distribution differences of CG lightning flash before and after the construction of high building were revealed and the effects on CG flash distribution for different height buildings were analyzed.(6) The 3D total flash location array, LFEDA, including 8 substations were constructed on Tibet for the first time.The preliminary observation on Tibet thunderstorm activity indicated that the 3D locations of Tibet electric activity were reliable.(Zhang Yang, Lyu Weitao, Zheng Dong, Zhang Yijun, Yao Wen, Ma Ying, Xu Liangtao, Qi Qi,Wang Fei, Fan Xiangpeng, Fan Yanfeng)

6 Model and reanalysis data

6.1 Studies of key techniques associated with unified modeling of weather and climate

On the model dynamical framework, based on the technical features of unstructured grids, we presented a general formulation for the dynamical core and developed an unstructured grid modeling system.This system adopts the indirect addressing data structure, which is well suited for massive parallel computing.The system is equipped with separate mesh files, spherical operators, numerical solvers and I/O capability, which make it a new prototype of a dynamical framework.The model may be combined with both quasi-uniform and variable-resolution meshes and can be used for resolving various physical problems.Based on this system, we developed high-order numerical methods and applied them to two-type physical problems: a scalar transport problem and a potential vorticity transport problem.Results suggest that in the scalar transport problem, using high-order methods can effectively reduce numerical errors produced by low-order schemes; in the potential vorticity transport problem, using high-order upwind methods well controls the increase of potential entrophy,damps the numerical noise and reduces the numerical errors.On the model physics, a new aerosol scheme is incorporated to the CAMS-CSM model, which will be used for participating the CMIP6 experiments.Regarding the coupled model, we have completed the final version of CAMS-CSM model for participation in CMIP6, downloading and generating of CMIP6 forcing data for CAMS-CSM model, and developing corresponding reading modules for the model, the development of aerosol modules for CAMS-CSM model.Tentative CMIP6 DECK experiments were carried out using CAMS-CSM model.The assessment of the results shows that this version of CAMS-CSM can successfully capture the major features of large-scale circulation and major modes of the global climate, laying a solid foundation for conducting the final CMIP6 experiments(Fig.9).(Zhang Yi, Chen Haoming, Rong Xinyao)

6.2 Progress in East Asia regional atmospheric reanalysis data

The East Asian regional reanalysis system has been further improved and optimized so that the system can operate continuously and steadily in a long period integration.Besides, assimilation operator of radar radial wind and radar mosaic technology have been improved, and significant progress has been achieved in the application of radar data assimilation.As for sounding data, a comparison has been made between the number of the sounding data and its vertical levels used in the East Asia region reanalysis data and those included in GDAS, and comparative analysis was conducted by assimilating the two types of data.Furthermore, one year reanalysis dataset was produced using the optimized East Asian regional reanalysis system and the data were evaluated based on both surface and sounding observations.The results show that the developed reanalysis data performs a little better than ERA-Interim data within China.(Liang Xudong, Yin Jinfang, Chen Feng, Xie Yanxin, Liu Ying, He Huizhong, Zou Haibo)

6.3 Comparative analysis of integrated bias correction to ensemble forecast of precipitation in Southeast China

Based on the 24 h precipitation ensemble forecasts from three global numerical forecast centers (CMA,ECMWF and NCEP) and rain gauge data in the southeast region from 1 June to 31 July 2015 and using the Bayesian Model Averaging (scheme A) and combined Bayesian Model Averaging and statistically downscale(scheme B), we have corrected the three single-center and multi-center grand ensemble precipitation predictions, compared the adjustment effects of the two schemes, and then selected the precipitation forecasts from 1 to 31 August in 2015 to perform an independent sample test, and analyzed the skills of precipitation prediction before and after the correction.Taking 50th percentile precipitation forecast of three single center and grand ensemble as example, the results indicate that scheme A eliminates a large number of the false alarm of light rain and corrects the bias for light and moderate rains remarkably.But the correction of the precipitation intensity for those exceeding heavy rain is not evident.With the precipitation thresholds increasing, the correction of scheme A becomes weak.The correction to the orientation of rain belt is not clear and reduces the magnitude of precipitation area of heavy rain, or even makes the area disappear.After adopting scheme B adjustment, it not only reduces the false alarm of raw ensemble forecasts, but also corrects the precipitation intensity and the rainfall area, so that the range and magnitude of the precipitation forecasts are closer to observations.But its effect is still not significant to correct high-graded precipitation such as those exceeding 50.0 mm.(Yang Ruiwen, Zhao Linna)

7 Experiment and research platform

7.1 Atmospheric Science Experiment and Research Platform (formerly Scientific Research Data Sharing Platform) progress

We have completed construction of phaseⅠof Scientific Research Data Sharing Platform, which is fully operational and provides online data sharing service.We have received basic research funding support from department for construction of phaseⅡof the platform for a new project.We plan to complete the construction of Atmospheric Science Experiment and Research Platform in three years from 2017 to 2019.The project includes (1) development of atmospheric science experiment data management standards; (2) establishment of atmospheric science experiment data resources; (3) visualization of scientific experiment scene; (4)optimization of data sharing platform.(Gao Mei, Zhang Wenhua, Li Bin, Zhu Kongju)