定量降雨與實(shí)時(shí)洪水預(yù)報(bào)研究

1、附件 2論文中英文摘要作者姓名： 吳志勇論文題目 ：定量降雨與實(shí)時(shí)洪水預(yù)報(bào)研究作者簡介：吳志勇，男， 1979年 01月出生， 2003年 09月師從于河海大學(xué)陸 桂華教授，于 2008年 06月獲博士學(xué)位。中文摘要基于實(shí)測(cè)降雨的傳統(tǒng)洪水預(yù)報(bào)方法，預(yù)見期較短，難以完全滿足防汛調(diào)度、洪水資源化 應(yīng)用等洪水管理的要求。預(yù)見期短已成為水文預(yù)報(bào)發(fā)揮更大效益的主要障礙。從研究與實(shí)踐 均表明，預(yù)見期的延長必須依賴有足夠精度的定量降雨預(yù)報(bào)。本文以增長洪水預(yù)報(bào)預(yù)見期為 主要目標(biāo)，重點(diǎn)研究了引入降雨預(yù)見期的實(shí)時(shí)洪水預(yù)報(bào)技術(shù)，突出數(shù)值天氣預(yù)報(bào)模式、陸面 過程模型以及基于網(wǎng)格的流域匯流模型和實(shí)時(shí)洪水預(yù)報(bào)系統(tǒng)等方面問

2、題的解決。較長預(yù)見期的降雨預(yù)報(bào)必須基于全球數(shù)值天氣預(yù)報(bào)模式。目前全球數(shù)值天氣預(yù)報(bào)模式模 擬范圍覆蓋全球，由于計(jì)算條件的限制，模式的網(wǎng)格往往較大，水平網(wǎng)格為25100 km。相對(duì)于數(shù)值天氣預(yù)報(bào)模式，流域洪水預(yù)報(bào)要求較小尺度下進(jìn)行，如分布式水文模型采用的網(wǎng)格 為0.011 km。因此，將大氣數(shù)值模式和水文模型耦合進(jìn)行洪水預(yù)報(bào)時(shí)，需解決它們之間的尺度問題。本文采用 MC2 模式( Mesoscale Compressible Community Mode)l 作為動(dòng)力尺度降解 工具，進(jìn)行了具有預(yù)見期長達(dá)96 h,空間分辨率達(dá)20 km和5 km的降雨預(yù)報(bào)研究。分析了預(yù) 報(bào)區(qū)域范圍大小、時(shí)空分辨率和物

3、理過程參數(shù)化方案等關(guān)鍵要素的確定和協(xié)調(diào)問題，提出了 考慮水汽源地影響的區(qū)域數(shù)值天氣預(yù)報(bào)模式構(gòu)建方法， 建立了 MC2 多層自嵌套方案。 選取了 淮河流域 1998 年、2003 年以及 20052007年汛期資料對(duì)降雨預(yù)報(bào)進(jìn)行了后預(yù)報(bào)和實(shí)時(shí)預(yù)報(bào) 檢驗(yàn)，結(jié)果表明， 本文建立的 MC2 模式對(duì)淮河流域暴雨具有很好的預(yù)報(bào)能力， 特別是對(duì)梅雨 期強(qiáng)降雨預(yù)報(bào)，無論其雨帶分布和結(jié)構(gòu)，還是暴雨中心位置及強(qiáng)度，均能獲得較高的預(yù)報(bào)精 度。陸面是陸氣系統(tǒng)的重要組成部分，它控制著地表能量在感熱和潛熱間的分配，以及地表 水分在蒸發(fā)和徑流之間的分配，對(duì)全球和區(qū)域天氣、氣候有重要影響作用。本文引入 CLASS (Cana

4、dian Land Surface Scheme模型作為陸面過程模型與 MC2模式耦合。原 CLASS模型 具有很強(qiáng)的模擬能量通量能力，但缺乏模擬重要的壤中流產(chǎn)流的機(jī)制。為此，本文提出了對(duì) CLASS 的產(chǎn)流模塊的改進(jìn)方案。改進(jìn)方案采用田間持水量閾值來實(shí)現(xiàn)壤中流的產(chǎn)生，引入了 空間分布函數(shù)來表示土壤田間持水量在次網(wǎng)格內(nèi)的變化。選用了 1998和1999年淮河流域能 量與水分循環(huán)試驗(yàn)( HUBEX )的觀測(cè)資料，對(duì)改進(jìn)前后的 CLASS 模型進(jìn)行了點(diǎn)尺度和流域 尺度的檢驗(yàn)。結(jié)果表明，改進(jìn)后的 CLASS 模型，不僅保持原 CLASS 模型具有很好的模擬淮 河流域不同地表和不同季節(jié)的能量通量、土壤

5、溫度、土壤含水量過程以及總徑流量的功能， 且對(duì)模擬洪峰流量、峰現(xiàn)時(shí)間和洪水過程的能力有了較大的提高，使產(chǎn)流過程的模擬大為改 善。壤中流模擬引入 CLASS 模型，也為水文資料短缺地區(qū)建立水文計(jì)算模型提供了很好技術(shù) 支撐。流域匯流是洪水預(yù)報(bào)的重要組成部分，為進(jìn)行流域洪水匯集的分布式模擬，以及有效利 用數(shù)值天氣預(yù)報(bào)模式輸出的分布式降雨預(yù)報(bào)結(jié)果，本文開發(fā)了基于網(wǎng)格的分布式流域匯流模 型。提出了空間分布匯流演算參數(shù)的求解方法和基于新一代交互式數(shù)據(jù)語言（ IDL ）的網(wǎng)格 匯流模型程序設(shè)計(jì)方法，為深入研究流域洪水匯流的時(shí)空過程提供了支持。模型基于數(shù)字水 系，構(gòu)建河網(wǎng)拓?fù)潢P(guān)系；采用造床流量作為河段的參考穩(wěn)

6、定流量，通過構(gòu)建地理因子和水力 學(xué)因子的經(jīng)驗(yàn)關(guān)系，求解空間分布的匯流演算參數(shù)；基于新一代交互式數(shù)據(jù)語言（ IDL ），設(shè) 計(jì)開發(fā)計(jì)算程序。經(jīng)淮河流域?qū)嶋H洪水過程的檢驗(yàn)，表明該模型計(jì)算效率高，能夠較好地模 擬洪水的時(shí)空演進(jìn)過程，成果具有較高的精度，滿足洪水預(yù)報(bào)的要求。將定量降雨預(yù)報(bào)模式與洪水預(yù)報(bào)模型耦合起來進(jìn)行實(shí)時(shí)洪水預(yù)報(bào)，增長洪水預(yù)報(bào)的預(yù)見 期和預(yù)報(bào)精度，是一項(xiàng)復(fù)雜的系統(tǒng)工程，本文基于先進(jìn)的計(jì)算機(jī)及網(wǎng)絡(luò)技術(shù)，采用計(jì)算服務(wù) 器和客戶端的構(gòu)建方式，綜合運(yùn)用多種程序開發(fā)語言，開發(fā)了基于定量降雨預(yù)報(bào)的實(shí)時(shí)洪水 預(yù)報(bào)系統(tǒng)。提出了多任務(wù)并行的協(xié)同工作方式，開發(fā)了數(shù)值天氣預(yù)報(bào)、洪水預(yù)報(bào)和 GIS 的集 成技

7、術(shù)，實(shí)現(xiàn)了降雨預(yù)報(bào)和洪水預(yù)報(bào)的全自動(dòng)運(yùn)行。構(gòu)建了一個(gè)可滾動(dòng)運(yùn)行的實(shí)時(shí)洪水預(yù)報(bào) 系統(tǒng)，預(yù)報(bào)所需的實(shí)時(shí)降雨信息從全國水文站網(wǎng)信息系統(tǒng)數(shù)據(jù)庫獲取， 預(yù)報(bào)降雨由 MC2 模式 提供；實(shí)時(shí)預(yù)報(bào)結(jié)果通過基于 WebGIS 的全國水文站網(wǎng)信息系統(tǒng)平臺(tái)發(fā)布。在經(jīng)過 1998 和 2003年的檢驗(yàn)之后， 系統(tǒng)對(duì) 20052007年 3個(gè)汛期淮河王家壩以上流域的洪水進(jìn)行了實(shí)時(shí)跟 蹤預(yù)報(bào)，獲得了預(yù)見期長達(dá) 35 d 的洪水預(yù)報(bào)，預(yù)見期與精度滿足實(shí)用要求。 2007年 7月， 應(yīng)用本系統(tǒng)比傳統(tǒng)預(yù)報(bào)方法提早 132 小時(shí)，預(yù)報(bào)出 7 月 11日淮河流域王家壩（含地理城）的 洪峰流量8740m3/s，與實(shí)測(cè)8100m3/

8、s僅相差9%,峰現(xiàn)時(shí)間誤差僅為6 h。可見，本文建立的 定量實(shí)時(shí)降雨預(yù)報(bào)與實(shí)時(shí)洪水預(yù)報(bào)相結(jié)合的系統(tǒng)能夠增長洪水預(yù)報(bào)的預(yù)見期、保證洪水預(yù)報(bào) 精度，具有較好的應(yīng)用前景。本文研究形成的技術(shù)成果，如數(shù)字水系構(gòu)建程序、降雨預(yù)報(bào)運(yùn)行和發(fā)布模塊以及洪水預(yù) 報(bào)模塊等，是2007年水利部大禹水利科學(xué)技術(shù)一等獎(jiǎng) “基于WebGIS的全國水文站網(wǎng)信息系 統(tǒng)”的重要組成部分。 以論文的研究成果為主要內(nèi)容， 作者與導(dǎo)師和何海博士合作完成了 水 文循環(huán)過程及定量預(yù)報(bào)專著的撰寫，即將于 2010 年由科學(xué)出版社出版。關(guān)鍵詞：洪水預(yù)報(bào)、預(yù)見期、陸氣耦合、數(shù)值天氣預(yù)報(bào)模式、陸面過程模型、匯 流模型Quantitative Pr

9、ecipitation and Real-time Flood ForecastWu ZhiyongABSTRACTMost traditional flood forecast methods in China use observed precipitation as model input. The flood forecast lead-time is thus limited to the natural flow concentration time of a basin, which cannot meet the growing needs in the operational

10、 practice of flood prevention and disaster reduction. The thesis focuses on real-time precipitation and flood forecasts using coupled hydro-metrological modeling systems, which can help to enhance the flood forecast lead-time. The ultimate goal of the study is to produce accurate and timely determin

11、istic forecasts with long lead times for the authorities in operational flood management decision making.The atmospheric model MC2(Canadian Mesoscale Compressible Community) is used as a dynamical downscaling tool, which is run at different resolution (20 and 5 km) over two regional domains and can

12、gain 96-h lead time. The determination and coordination of vital factors such as the size of forecast area, the spatial-temporal resolution and the parametrical schemeof physical process were discussed. The precipitation data during flood season of 1998, 2003 and 2005 to 2007 in Huaihe River Basin w

13、ere chosen to make hindcast and real-time forecast validation. MC2 is found to be capable of simulating precipitation over the basin both for the position and intensity; which is especially true for heavy precipitation events during the Meiyu season. This is important as Meiyu induced flooding is th

14、e primary concern in the Huaihe River Basin.The original version of CLASS has no interflow and a baseflow that is not appropriate for hydrograph simulation in real-time flood forecasts. The original CLASS is thus improved for interflow generation using a field capacity threshold together with a spat

15、ial probability distribution function to represent sub-grid variability in soil field capacity. A linear reservoir at the bottom the third CLASS soil layer is introduced to model the baseflow. The 1998-99 HUBEX/GAME were chosen to test the original and improved CLASS in point scale and catchment sca

16、le. The result indicates that the two versions of CLASS give good results for the net radiation, sensible and latent heat fluxes, ground heat flux, and soil temperature and moisture in various land surfaces during various seasons. However, The improved CLASS is significantly improved in both peak fl

17、ow and flood timing simulation.Flow routing in catchments is an important part in flood forecast technology. In order to carry on distributed simulation of flood convergence in catchments, a grid-based flow routing model was developed. Based on digital water system, the model constructed topology re

18、lationship of channel networks. Dominant discharge was taken as the stable discharge of channel to consult, and the solutions of distributed parameters of flow routing were given via the construction of the relationship between geographical and hydraulic factors. The calculating programs in design a

19、nd exploitation were based on a new generation of interactive data language-IDL. After the validation using practical flood in Huaihe River Basin, the model were efficient in calculation and could simulate the spatial-temporal development of flood very well. And the results of this model were shown

20、to be a high level in precision and could satisfy the need of the application of flood forecast.A coupled hydro-metrological modeling system is established on the bases of modern science and technology. It involves the use of numerical weather prediction models, flood forecast models, acquiring real

21、-time gauge precipitation, a GIS based platform for spatially displaying the propagation of the forecast floods in real-time, various programming languages, personal and sup-computers, different operating systems, client-server modes, and internet technologies. The system has become fully automated.

22、 After the validation for 1998 and 2003, real-time rolling forecasts in catchments above Wangjiaba were carried on during 2005 and 2007 using the system in Huaihe River Basin. The results demonstrate the applicability and the value of using mesoscale model precipitation for real-time flood forecast, which can provide a long lead time of heavy precipitation and subsequent flooding for authorities in operational flood management decision maki