《基于ArcGIS的水利大数据及应用ppt课件.pptx》由会员分享，可在线阅读，更多相关《基于ArcGIS的水利大数据及应用ppt课件.pptx（56页珍藏版）》请在三一办公上搜索。
1、基于ArcGIS的水利大数据及应用,研究中心及团队简介,水利大数据及其面临的挑战,基于水利大数据的多灾害信息集成与风险预警案例,主要内容,1,2,3,一、研究中心及团队简介,科研平台：清华HydroSky创新团队全球遥感大数据与水科学工程环境资源前沿交叉基于卫星雷达遥感和云计算大数据信息技术的现代水文水资源新理论技术全球海量天空地遥感大数据信息挖掘与多源数据集成同化技术多时空尺度上跨越系统观测、模拟和预报分析及动态可视化技术水文气象地质灾害与极端气候变化监测预警技术海洋遥感信息技术和海洋大数据平台建设遥感金融大数据创新创业研究智慧产业、优化配置、高效利用管理等,资源整合 跨院系合作 多学科交叉
2、 政府、社会,平台建设 天-空-地-海 校地合作 海外合作,论坛培训 学术交流 创新创业 教育教学,跨院系平台：清华大学遥感大数据研究中心2015年10月23日成立，土水学院、水利系、水沙科学国家重点实验室建筑学院、环境学院、地学中心、3S中心、电子系、计算机系,产学研用平台： 物联网遥感大数据联合研究中心Joint Center for Internet of Things and Remote Sensing Big Data2016年5月24日成立(国内外第一家)理论顶天创新 实践立地创业,天地空海遥,感信息采集,能物联网,万物相连智 开放大数据,服务平台,Sensor Technolo
3、gies: All Data/Info.IOT: Connecting/Interacting all thingsBig Data Technologies Washing/MiningAI: Artificial Intelligence/Deep Learning围绕天地空海遥感信息采集、万物相连物联网、人工智能以及开放性大数据服务平台等核心领域，以前沿交叉创新技术研发及产业化应用为主线，形成“理论顶天创新、实践立地创业”，引领推动国内外物联网遥感大数据交叉领域的创新发展及产学研创业孵化。,遥感大数据平台项目,导航卫星大数据,海洋水利大数据,农业遥感大数据,医疗金融大数据,三维智慧城市,
4、水文,气象地质灾害大,数据,商业航天,遥感大数据研究中心产学研项目团队,1. 水文洪涝干旱灾害模型系统1.1 全球分布式水文模型：CREST2.0-Fortran1.2 全球分布式水文模型：CREST2.1-Matlab1.3 城市洪水模型uCREST1.0：高精度Urban CREST 1.01.4 水文洪涝淹没四维模拟系统：CREST_iMap 1.01.5 Global Multi Droughts Indicator System：全球多干旱指标体系1.6 基于GIS可视化平台的：Arc CREST 1.02 滑坡泥石流模型系统2.1 滑坡风险预警模型：RIDL1.02.2 SLIDE
5、1.02.3 TRIGRS 2.03 多灾害耦合系统及开发平台3.1 水文、滑坡耦合模型：iCRESLIDE1.03.2 EF5： Ensemble Framework for Flash Flood Forecasting3.3 NFL：美国国家山洪泥石流系统3.4 HFL_DEWS：台风洪水灾害预警系统3.5 CI-FLOW：海暴潮近岸带防灾预警系统3.6 HyPRO：专业水模型系统工程开发平台4. 遥感反演算法产品系统4.1 PERSIANN，1983-now, global4.2 PERSIANN-CCS, 02-now, 4km global4.3 TRMM/TMPA, 98-now
6、, 25km, global4.4 GPM/iMERG, 4km, global4.5 低空雷达融合方法VPR-IE, 94-now, 250m, CONUS4.6 天地空多源降水系统MRMS， 250米，2.5分钟4.6 M2ET 遥感蒸散ET算法4.7 SatET 全球遥感蒸散ET算法4.8 导航卫星大气及土壤含水量、积雪等反演技术,5. 大数据，移动平台、云计算技术平台5.1 mPING 美国版移动平台技术5.2 mPING 全球多语种移动平台技术5.3 Disaster中国民政多灾害信息搜集移动平台5.4 CyberFlood全球洪水数据库云计算平台技术5.5 CsLID中国滑坡数据库
7、云计算平台技术5.6 基于云计算的WebCREST1.0： mCREST移动终端6. 遥感硬件技术6.1 Roughness Meter for 3-D Surface（ US Invention Model Patent）6.2 XP1000双偏振X-band大气雷达6.3 多普勒天气雷达系统 （SDR-100X）6.4 StreamRadar 水雷达技术7. 临近预报方法及预报评估7.1 A Lagrangian Pixel-Based Approach7.2 An Object-based Short-term QPF approach7.3 Hybrid Nowcasting Appr
8、oach8. 优化及模拟预报算法8.1 An Automatic Seeded Regional Growth Segmentation Algorithm for Satellite Images8.2 SOLO优化模拟预报合成器8.3 SONO优化模拟预报合成器8.4 多源同化ENSRF: Ensemble Square Root Filter8.5 同化方法SPF: Sequential Particle Filter8.6 联合同化HKV: Hybrid of K-Filter and 3/4D Variation Methods8.7 同化GSI: Gridpoint Statist
9、ical Interpolation DA System（NCEP Radar-WRF）,成果,1. SATELLITE PRECIPITATION DATA1.1 TRMM-based Multi-Satellite Precipitation Analysis (1998-present): Quasi-global, 3 Hour 0.25 Degree1.2 PERSIANN (1998-Present): Quasi-global, 3-hour 0.25 Degree1.2 PERSIANN-CDR (1983-Present): Quasi-global, Daily, 0.25
10、 Degree1.3 PERSIANN-CCS CONUS (2002-present): CONUS, 4km, 30-minutes1.4 PERSIANN-CCS Global (2002-present): Global, 4-km, 30-minutes1.5 Hydro-Estimator Data: CONUS 4-km hourly1.6 GPCP/CMAP (1979-present): Global Monthly 2.5 x 2.5 Degree1.7 GPM/iMERG: 4km, 3-hour, Global2. RADAR PRECIPITATION DATA2.1
11、 NOAA/NSSL/MRMS: 1-km 2.5 minute for Contiguous U.S, 2004-present2.2 Multi-Sensor Precipitation Estimation (Radar/Satellite/Gauge/Model)2.3 Stage IV, Stage II, and MPE multi-senosr Precipitation Estimation2.4 S-band KOUN and C-band OU-PRIME Dual Polarization Radar QPE2.5 Phased Array Radar QPE3. GAU
12、GE PRECIPITATION DATA3.1 Africa Lake Victory Nzoia Basin Precipitation and Discharge data , 1985-20063.2 MESONET3.3 GPCC: 1979Present3.4 CPC Daily Gauge3.5 North American Monsoon Rain Gauge Netwrok (NAME NERN)3.6 Micronet Ft Cobbs Basin and Washita Basin3.7 CONUS HADZ Gauge Network3.8 Bagmati Basin
13、Nepal (daily data for more than 50 stations for 1999-2006)4. GLOBAL AND REGIONAL RUNOFF/DISCHARGE DATA4.1 GRDC: Daily Discharge from more than 1600 stations in Central/South America and Africa4.2 Nzoia basin Discharge, 1 station, 1985-20064.3 10+ years TRMM-based Rainfall-Runoff Data4.4 Africa Lake
14、Victoria and Kenya rainfall gauge and discharge4.5 Hydrometeorological Testbed East: TAR-Pimlico and Neuse Basin4.6 USGS Discharge data4.7 Nepal Mountainous Basins (Daily discharge at one station for 1999-2006),5. ET DATA and Soil Moisture5.1 GDAS 1-Degree Daily Global Potential ET5.2 MODIS-based Po
15、tential ET5.3 MESONET Reference ET5.4 Remote Sensing M/M-ET: Oklahoma Actual ET (3-year daily 30m-250m)5.5 Global Monthly Mean PET5.6 SatET: Satellite-based ET products (1-km, weekly, global 1983-present)5.7 GNSS-R Soil Moisture Retrieval, Validation, and Application5.8 AMSR-E , ASCAT, FY-3, SMAP6.
16、GLOBAL LAND SURFACE DATA6.1 SRTM 30m-90m Global Digital Elevation Datab6.2 HydroSHEDS 30m-1000m Global River Channel Network Data6.3 Hydro1k Global 1km Hydrological Network Data6.4 MODIS Global Multi-year Land cover/types/LST/NDVI6.5 LandSat 30m Multi-Band Remote Sensing Data6.6 Global Soil Type Cla
17、ssification Data, 1km7. GLOBAL DISASTER DATABASE7.1 Global Flood Inventory Digital Database (1998-2009)7.2 Global Landslide Inventory data (2003-2009)7.3 Global Landslide Susceptibility data7.4 Global MODIS-based Fire Map8. GLOBAL SOCIOECONOMIC DATABASE8.1 Global Gridded Population/GDP/HDI9.Cyber/Vi
18、rtual Big Data form Mobile Apps and Cloud Technologies9.1 mPING: Meteorological Phenomena Identification Near the Ground9.2 mPING_Glob: mPING Multi-language Global Version：9.3 iDisaster: integrated Disaster Report and Visualization Apps System9.4 CyberFlood: Cloud-based Global Cyber Flood RD Platfor
19、m,成果,清华大学高分卫星数据与应用中心高校第一家服务全国科教产学研,高分立体观测体系,高分数据使用用户培训,清华高分中心一期建设高分技术及产品研发,北斗+ ：点石成金，增值创新,目标： 拓展北斗从传统行业到新细分行业的应用创新！,Satellite InSAR Monitoring All Deformation:1mm卫星合成孔径雷达干涉测量形变监测,高速公路,火山现象,采矿活动,关键构筑物,大坝,下沉现象,铁路,InSAR监测应用领域,管线,关键区域,建筑物,滑坡,油气,13,溪洛渡水电站坝体形变监测 ：Sentinel-1、TerraSAR,TerraSAR监测结果 2015.7.28
20、-2015.8.19 垂直向上形变,Sentinel-1监测结果 2015.5.18-2015.6.11垂直于河道方向（北偏东48.12）向形变,二、水利大数据及其面临的挑战, 水利工作关系到国计民生，尤其是我国水资源分布存在严重的时空分布不均特性，旱灾洪涝易发多发。水利行业在经济、生态、社会等方面都扮演着重要角色，对水利大数据的研究具有重要的现实意义和应用价值。 水利大数据是在大数据的理论指导及技术支撑下的水利科学和工程的重要实践。,水利工作及水利大数据的重要性,水利大数据 水利大数据是指产生于各种水文监测网络、水利设施、用水单位和水利相关经济活动，并通过现代化信息技术高效传输、分布存储于各
21、地存储系统、,但又可以快速读取集中于云端、实现深度数据挖掘并可视化的海量多源数据总和。,Volume海量,Velocity快速,Value价值,Variety多样,Veracity真实,交叉性，由于水利和其它领域具有交叉性，因此水利大数,据和遥感大数据、气象大数据、海洋大数据等交叉；,时空分布性，需要依赖先进大数据技术进行处理分析，包,括分布式大数据存储框架、机器学习等数据挖掘方法；多元循环性，由水的多元循环决定的水利大数据在经济、,社会、生态等领域的价值循环。,水利大数据的外延,挑战一：水利大数据的收集与集成 水利大数据来源广泛，不同的监测平台得到的数据具有不同的数据结构、存储系统，非结构化
22、数据、半结构化数据、结构化数据并存； 由于观测条件的差异，数据可信度层次不齐，对数据清洗和质量的确保提出了很高的要求； 大数据的存储与管理需要新型数据库的支持，水利大数据的信息化还未与新型数据库接轨。,水利大数据面临的挑战,挑战二：水利大数据的时空多维度分析 水利大数据具有明显的时空分布特性，时间、空间双维度下的数据分析具有难度； 水利大数据在其应用领域讲究实时性，比如洪水预报等，这对大数据的处理分析速度提出了高要求； 水利大数据的深度挖掘有赖于引入先进的人工智能算法，两者的有效结合至关重要。,水利大数据面临的挑战,挑战三：水利大数据的共享与安全 众多水利数据掌握在政府机关部门，为非公开数据，
23、形成数据孤岛现象；, 水利数据是国家安全的重要组成部分，水利数据的共享与安全是一个值得探讨的问题。,水利大数据面临的挑战,三、基于水利大数据的多灾害信息集,成与风险预警案例介绍,基于水利大数据的多灾害信息集成与风险预警案例介绍,1、天、地、空、海，多基多源降水数据采集2、移动众包信息收集可视化云平台mPing3、基于水利大数据的全球洪水泥石流灾害预测预报4、基于概率洪水风险预报EF55、城市洪水模型Urban CREST介绍6、全球风暴数据库及CI-FLOW7、中国区域多尺度洪水模拟及预警系统的建立8、基于ArcGIS的FFG介绍9、基于ArcGIS平台开发的ArcCREST介绍,基于水利大数
24、据的多灾害信息集成与风险预警案例介绍,3小时临近预报（250米2.5分钟）36小时模型预报（1公里小时）,1.天、地、空、海多基多源降水数据采集双偏振雷达卫星站点模型,PERSIANN 全球卫星产品(4km, hourly),Hong et al., 2004, JAM;,5颗地球静止卫星（可见光红外）以及4颗极轨卫星（雷达和被动微波）通过,人工神经网络ANN机器学习训练反演 High Quality 卫星降水产品,Merge Satellites, ground (Radar & Gauge), and Model (NWP),TRMM,Aqua,DMSP,NOAA,METEOSAT,(Eu
25、rope),GOESGMS/MTSAT(Japan),2005 加入 NASA：多卫星联合反演共性技术；(1700+引用)全球天地空标准产品系列：TMPA17+ years (98-16) of data; Most requested TRMM product from NASAWith Huffman et al. 2007 : （1700+ 引用）,Instant-aneousSSM/ITRMM,AMSR,AMSU,30-day HQ coefficients,3-hourly merged HQ,HourlyIR Tb,Hourly HQ-calib IRprecip,3-hourly
26、 multi-satellite (MS),Monthlygauges,Monthly SGRescale 3-hourly MSto monthly SGRescaled 3-hourly MSTMPAuses 4 Polar-orbital microwave satellites (NOAA, DoD, NASA) and 5 Geo-IR satellites(GOES8-10, GMS, MYSAT, MeteoSAT); all calibrated by TRMM Preci Radar,Calibrate High-Quality(HQ) Estimates to“Best”
27、Space Radar,Merge HQ Estimates,Match IR and HQ,generate coeffs,Apply IR coefficients,Merge IR, merged HQestimates,Compute monthlysatellite-gaugecombination (SG),30-day IR coefficients,26,深度学习方法研制全球卫星产品研制,在深度学习中，我们可以将不同频段的可见光、红外、微波影像同时作为训练数据输入模型，且不需要事先设定Feature，海量的遥感影像下，让模型自己去寻找Feature。,青藏西南部IR云图,相应时
28、段降水情况,5-minute250m,Rainfall Dataover USA,2. mPING 美国版灾害Crowdsourcing移动平台技术,2.移动众包信息收集可视化云平台,mPING Crowd Sourcing Tool and Data,750,000+ App Downloads Since Dec 2013,硅谷SF IoT/BigData Weather 2.0 Service Inc.,Ensemble Coupled Hydro-Landslide Modeling SystemWater Balance ComponentCREST (Variable Infilt
29、ration,Curve)SAC-SMA,Cell-by-cell linear reservoir,Landslide Model Ensemble,TRIGRSSLIDE,+Runoff Routing,Surface Flow andInundation,Soil Water Content,Other variables,Occurrence andLocations of landslides,Remote Sensing basedPrecipitation Estimates,Topography,Land cover/Land Use,3.基于水利大数据的全球水洪泥石流灾害预测
30、预报National Flash Landslide System,LANDSLIDE:,SLope-Infiltration-Distributed Equilibrium Model,3. 基于水利大数据的全球水洪泥石流灾害预测预报美国暴雨山洪泥石流灾害链业务化系统NFL:NMQ： National Mosaic and Multi-Sensor QPE (NMQ)FLASH： Flooded Locations And Simulated Hydrographs,NMQ Radar Precipitation,Observations 250 m/2.5 min,Hydrologic M
31、odels,10-11 June 2010, Albert Pike RecArea, Arkansas,250 mm,150200,Simulated surface water flow,20fatalities,FLASH Distributed CREST LANDSLIDE,Landslide Hotspot Models,Red: ObservationsPink: Predictions,Landslide prediction,model,Integrated Hydrologic-Landslide ModeliCRESLIDE = CREST + SLIDE,Coupled
32、 Routing and Excess STorage (CREST),Jointly developed by,OU/NASARun operationally overglobeDistributed, fullycoupled runoff,generation and routingWang and Hong et al. 2011 HSJ,Integrated Hydrologic-Landslide Model:iCRESLIDE,Development andApplication,- CREST has been set up at both national and basi
33、nscales in China;- iCRESLIDE shows great capability in forecastingshallow landslides around the world;,- More flood and landslide event data is needed.,NFL: Real-time, direct prediction of flash floods a reality,Photo source: National Geographic250m/5-min resolution of Q2 precipitation forcing and m
34、odel outputsAddresses service needs in NWS; flash flooding is #1 weather-related killer6/11 12:30am-4am 20 deaths: Little Missouri River Crested from 3 ft to 23.5 ft within 2 hoursInclude data assimilation and probabilistic productsReadily incorporate dual-pol radar products (Q3) and stormscale ense
35、mble forecasts,20,40,60,80,100,120,00,美国暴雨山洪泥石流灾害链耦合系统核心模型Physically-coupled iCRESTSLIDE (SLope Infiltration- Distributed Equilibrium)10.80.60.40.2,Validation with inventory data,Red: ObservationsPink: Predictions,美国北卡州 梅肯县Within 18-m 120-meter buffer zonePOD 0.5 0.9CSI 0.1 0.8FAR 0.9 0.2,(Liao et a
36、l., 2011, Nat. Hazards ),16th hr,Radius (m)FS Map vs. Time18th hr,21st hr,State-Param Estimation DREAM (2010)Observed Streamflow,Routing Kinematic wave (2014) Linear reservoir (2010)Forecast Streamflow (2010) Recurrence Interval(2010) Inundation (2015),4.基于概率洪水风险预报 EF5Ensemble Framework For Flash Fl
37、ood ForecastingBest distributed hydrologic System yet,PrecipForcing1. MRMS2. TMPA RT3.WRR/HRRR QPF,Evapotranspiration1. FEWS NET PET2. HRRR temp3. VIIRS?,Surface Runoff CREST (2010) SAC-SMA (2013) Hydrophobic (2015)Groundwater MODFLOW,Snowmelt SNOW-17(2015)- 2m Temp,CurrentVersionFutureAddition,EF5:
38、 Probability of Flash Flood Forecast (PFFF)基于概率洪水风险预报,PFFF( RP = 5 yr )100%50%0%,The New Features of uCREST Model 1-10 Meter DEM and Urban Drainage System, Urban Canopy and High Rise Building Impact on the Rainfall,Interception, Enhanced Impervious (pavement, roof etc.) and Non-impervious,surface in
39、filtration and Surface Processes (runoff, ET etc), Urban Sewer/Pipeline Module included as a special Interflow,Process/reservoir, Has been tested and implemented in Oklahoma City and Dallas,Metropolitan at spatial resolution,5.城市洪水模型Urban CREST介绍,AHigh-Resolution UrbanCREST Flood Modeling and Mappin
40、g System,For Urban and Built-up Environments,101 km,2010 June 14, OKC Flash Flood,Return Period (years),1,2,10,200+,NoFlooding,Flooding,SevereFlooding,Urban-CREST Flood Model Implemented at,Oklahoma City &Dallas Metropolitan,137 km,6.全球风暴数据库及CI-FLOWGlobal Storms (2000-2010),*Sellars et al. (2013), C
41、omputationalEarth Science: Big Data TransformedInto Insight, EOS Trans. AGU, 94(32),277,Nov 2011 BAMSThe CI-FLOW Project:A System for Total Water Level PredictionFrom The Summit To The SeaCI-FLOW summary paper with HurricaneIsabel, Hurricane Earl, & TropicalStorm Nicole results,Volume # Number # Nov
42、ember 2011BAMSAmerican Meteorological Society,Suzanne Van Cooten, , Yang Hong, et al., 2011: Theci-flow project: a system for total water levelprediction from the summit to the sea. Bull. Amer.Meteor. Soc., 92, 14271442.已应用到美国北卡罗来纳州、墨西哥湾等易受飓风和风暴潮影响的海岸带地区,海洋风暴潮与内陆洪水监测预警系统(CI-FLOW),Tracking the raindr
43、opsand disasters from theSKY and the SUMMIT to,the sea,CI-FLOW,Coastal and Inland FloodingObservation and Warning,CI-FLOW: HL-RDHM/SWAN/ADCIRC Coupled Model,Precipitation,Total Water Levels,Hydrodynamic Model (ADCIRC),HydrologicModelRiver BCsDischarge,AtmosphericModelSurface BCsPressureWind Forcing,
44、Wave ModelSurface BCsWave Forcing,Precipitation Source: QPE/QPFAtmospheric Model: NAM or NHC trackHydrologic Model: HL-RDHM, Vflo or CRESTWave Model: unstructured SWAN,7.中国区域多尺度洪水模拟及预警系统的建立, 与气象局以及国家气象中心合作开发中国的山洪预警系统 多源降水产品和地面台站数据进行雨量融合，驱动CREST模型，模拟径流分布 地貌水动力学模型模拟洪水淹没情景的时空演进，实时动态提取洪水淹没范围、水深分布和淹没时间分布
45、，实现对洪水的模拟,Date,3/5/1997,5/8/1997,7/11/1997,9/13/1997,11/16/1997,1/19/1998,3/24/1998,5/27/1998,7/30/1998,10/2/1998,12/5/1998,2/7/1999,4/12/1999,6/15/1999,8/18/1999,10/21/1999,12/24/1999,2/26/2000,4/30/2000,7/3/2000,9/5/2000,11/8/2000,1/11/2001,3/16/2001,5/19/2001,7/22/2001,9/24/2001,11/27/2001,1/30/2
46、002,4/4/2002,6/7/2002,8/10/2002,10/13/2002,12/16/2002,2/18/2003,4/23/2003,6/26/2003,8/29/2003,11/1/2003,1/4/2004,3/8/2004,5/11/2004,7/14/2004,9/16/2004,11/19/2004,1/22/2005,3/27/2005,5/30/2005,8/2/2005,10/5/2005,12/8/2005,洪水模拟的时间：19980628,050,100,150,200250300,1000050000,15000,2500020000,R_Obs in (m
47、3/s)R(v2.1) in (m3/s),rain,率定期,验证期,NSCE=0.897CC=0.947,Bias=-1.57%,20 年、,10 年、,5年、,2年、,1年 一遇洪水,外州站CREST模型率定/模拟效果：气象台站数据驱动,7.中国区域多尺度洪水模拟及预警系统的建立,114,114.5,115,115.5,116,116.5,117,2928.52827.52726.52625.525,7.中国区域多尺度洪水模拟及预警系统的建立,iMAP 在嘉陵江流域的应用结果,7.中国区域多尺度洪水模拟及预警系统的建立,9.基于ArcGIS平台开发的ArcCREST介绍ArcCREST U
48、I,Precip Thiessen,Evap Thiessen,Geo Data,Used for rainfall sites (Cell-based data need some effort),Parameters distribution need more advanced methodBugs in code, the results are not correct,Geo and Hydro data management and operation ,Parameters distribution settingModel running and results show,Ar
49、cCREST,Discharge(m3),1,13,25,37,49,61,73,85,97,109,121,133,145,157,169,181,193,205,217,229,241,253,265,277,289,301,313,325,337,349,361,ArcCREST运行结果分析,4003002001000,500,Calib,Time(24h)UnCalib,Actual,0,30025020015010050,0,50,100,150,200,250,300,350,Gage,Discharge:ArcCREST vs Gage, R2= 0.7025 ArcCREST
50、tends to overestimatedischarge Uncalibrated results indicate no modelsensitivity and unreliable estimationsDischarge:ArcCREST vs Gage, Flash Flood Guidance: FFG is theamount of rainfall required in agiven period of time to producebank full conditions on smallbasins from Flash Flood Guidance 1970toHy
