Nghiên cứu sử dụng dữ liệu viễn thám trong mô phỏng dòng chảy mặt phục vụ quy hoạch thủy lợi và phòng chống thiên tai tt tieng anh

MINISTRY OF EDUCATION AND TRAINING MINISTRY OF AGRICULTURE AND RURAL DEVELOPMENT THUY LOI UNIVERSITY BUI TUAN HAI RESEARCH ON USING REMOTE SENSING DATA IN SURFACE FLOW SIMULATION FOR IRRIGATION PLANNING AND NATURAL DISASTER PREVENTION - APPLYING FOR CA RIVER BASIN Major: Water Resources Engineering Major Code: 9580212 SUMMARY OF TECHNICAL PHD THESIS HA NOI, 2020 This Thesis was completed in Thuyloi University Supervisor 01: Assoc.Prof.Dr Le Quang Vinh Supervisor 02: Assoc.Prof.Dr Pham Quang Vinh Reviewer 01: Assoc.Prof.Dr Nguyen Mai Dang Reviewer 02: Assoc.Prof.Dr Uong Dinh Khanh Reviewer 03: Prof.Dr Nguyen Ngoc Thach The thesis will be defended before the Thesis Assessment Council at: Room -K1, Thuyloi University, No 175 Tay Son, Dong Da, Hanoi, Vietnam At 08:30 am on December 10th, 2020 Thesis can be found at the library: - National Library - Library of Thuyloi University INTRODUCTION The rationale of research Ca river basin is one of the regions severely affected by floods, droughts and saline intrusion Research on the process of flow formation in Ca river basin, especially upstream area, is extremely important in irrigation planning and disaster prevention for downstream areas However, since most of the upstream area with 34.8% of the catchment area where the runoff is formed is located in Laos, there is no data available for the study Research on using remote sensing data combined with appropriate technologies in meteorological forecasting, hydrology, flow and irrigation planning, natural disaster prevention in river basins is a solution to overcome the lack of real measurement documents are of interest to scientists For the above reasons, the Ph.D thesis: " Research on using remote sensing data in surface flow simulation for irrigation planning and natural disaster prevention - applying for Ca river basin " is essential Research objectives - Researching, exploiting, analyzing and selecting suitable satellite rain data to add hypothetical rain stations to areas in the basin where there is a lack of measuring stations and lack of data on actual rain measurements, adding monthly rainfall data for intermittent rain measurement stations in order to improve reliability in calculating flow for irrigation planning and disaster prevention - Study using remote sensing data (satellite rain data and Digital Elevation Model (DEM) data) in surface flow simulation for Ca river basin, especially transboundary flow simulation for the upper part where most of the area is located in Lao PDR with no actual rain records Object and scope of the study a) Research object: Surface runoff serving irrigation planning and disaster prevention in river basins in general and Ca river basin in particular b) Scope of research: - In terms of space: The study area is the Ca river basin, in which, focusing on studying two transboundary river branches located upstream of Nam Mo and Nam Non - In terms of time: Study on simulation of flow process in Ca river basin from 1982 to 2019, in which simulating flood flow from 2011 to 2019 Analysis and selection of concentrated satellite rain data for three year from 2015 to 2017 Approach and research method a) Approach to research: The thesis topic chooses the following two scientific approaches: (1) System approach; (2) Approach combines experimental research and theoretical research b) Research method: Using the following research methods: (1) Inheritance method (2) Methods of investigation, data collection, documents; (3) Statistical analysis method; (4) Mathematical modeling method; (5) Method of seminar and consultation with experts The scientific and practical significance of the study a) Scientific significance: Supplement the method of application of remote sensing technology and use of remote sensing data in combination with mathematical models in surface flow simulation on Ca river basin serving irrigation planning and natural disasters prevention b) Practical significance: Proposing methods of handling and using data from remote sensing for areas with no data or insufficient data in flow simulation for irrigation planning and natural disaster prevention river basin in general and Ca river basin in particular, meeting the requirements of socio-economic development New contributions of the Thesis a) Identify suitable remote sensing data among high resolution satellite rain data CHIRPS, GSMAP, GPM, CMORPH and digital elevation model data ALOS, ASTER, SRTM for adding assuming rain gauge stations, adding monthly rainfall data for areas where there is a lack of rain measurement stations, or lack of actual rain measurement documents to improve reliability in calculating and simulating flow for irrigation planning and natural disaster prevention in Ca river basin b) Clarifying the method of using satellite rain data and digital elevation model (DEM) as input data for the deterministic, lumped, conceptual model MIKE NAM and the distributed hydrological model IFAS to increase the accuracy in the flow simulation for the Ca river basin, especially the cross-border flow simulation of the two branches of Nam Mo and Nam Non rivers with most of the catchment area located in Laos The layout of the thesis In addition to the introduction, conclusion and recommendations, the thesis includes the following three main chapters: Chapter 1: Overview of research and use of remote sensing data in irrigation planning and natural disaster prevention Chapter 2: Research method and data used in the study Chapter 3: Research results using remote sensing data in calculating the flow in Ca river basin CHAPTER OVERVIEW OF RESEARCH AND USE OF REMOTE SENSING DATA IN IRRIGATION PLANNING AND NATURAL DISASTER PREVENTION 1.1 Overview of remote sensing technology Remote sensing is a research science that collects information about things and phenomena on the earth's surface from a long distance through special measuring technology equipment installed on satellites observing the earth's surface Over the past 60 years, remote sensing technology has constantly improved, the number of remote sensing satellites has been increasing in number, diverse in types, diverse in size, diverse in orbit and trending developing into satellite beams Application scope of remote sensing technology is also constantly expanding Currently, remote sensing technology is widely applied in the following main areas: (1) Weather forecasting and forecasting natural disasters (2) Water resources management and water quality (3) Management land (4) Building maps to meet socio-economic development requirements 1.2 Overview of researched scientific works in the world on issues related to the thesis topic 1.2.1 Research and apply remote sensing technology in planning, water resources management and natural disaster prevention in the world The research works have shown that remote sensing technology is being applied a lot in calculating rainfall, evaporation, simulating the process of forming runoff and flooding in river basins The hydrological and geological related data on the earth's surface observed by remote sensing when used together with the on-site measurement data, create important data stores for the study of surface water sources and groundwater, providing input to the mathematical models Remote sensing data processing is often done in GIS software Remote sensing contributes to the consensus in water resource management for international rivers Bad weather conditions during the rainy season are often associated with flooding, inundation and landslides making it extremely difficult to access and assess flooded areas, and remote sensing will help overcome the limitations this Through the selection of appropriate sensors and platforms, remote sensing can provide accurate and timely information on flooded or at-risk areas, assessing the extent of damage caused by floods Floods and landslides in places where it is difficult for people to have direct access to appropriate response measures The thesis also summarizes some research results applying remote sensing technology in planning, water resources management and disaster prevention in Africa under the TIGER initiative, in India and International Mekong River Commission The thesis also introduces the research results of applying remote sensing and GIS technology in flood monitoring and flood damage assessment in Pakistan and Nepal 1.2.2 Research using remote sensing data and mathematical model in flow simulation Overview of research results in this field in the world, the thesis is divided into the following main groups: a) The study using satellite rain data: Typically the studies using satellite rain data TRMM-3B42 v6 and v7, RFE 2.0, PERSIANN-CDR, CMORPH1.0 version 0.x to model simulation of flow regime for the Makhazine river basin in Morocco; the study used TRMM and GPM satellite rain data to simulate halfday floods for basins without measuring stations in Myanmar; using satellite rain data TMPA 3B32RT and rainfall data measuring station to evaluate the error model in determining daily rainfall from satellite rain data b) Research using DEM topographic data: Typically the studies compare DEM data with each other and compare with Australia's national topographic elevation data on the entire continent Ocean; The study uses DEM ALOS data to update hydrological maps, providing information on the impact of hydrodynamic dynamics in arid and semi-arid environments c) Research combining remote sensing data and mathematical model in rainflow relationship study: Remote sensing provides input data for mathematical models such as meteorological - hydrological characteristics according to space and time, soil moisture, surface characteristics and land use, soil cover… Scientists have demonstrated the good combination of GIS with simulation models such as HEC, MODFLOW, SHE, SWAT, MIKE BASIN, WEAP to solve problems related to the hydrological regime, flow of a number of river basins in the world Typically the study used remote sensing data combined with actual rainfall data and IFAS model to calculate flood flows for the Kelantan and Dungun river basins in Malaysia Many scientists have studied the combination of remote sensing data, GIS and SWAT model in assessing river basin water flow and quality under changes in land use structure and climate change They have established detailed models of surface and underground runoff, water quality enhancement model, or a combination of quality modeling with GIS tools, the effects of spatial variability on the flow model field 1.2.3 General assessment of research results in the world Overview of research results of scientists around the world shows that the combination of mathematical modeling, remote sensing technology and GIS is a new and very effective approach in calculating and simulating river basin flows, especially especially in irrigation planning, natural disaster prevention The research results also show that the flow calculation method combining remote sensing and GIS still has many gaps that the thesis needs to continue researching and deploying This thesis will inherit the advantages of methods: mathematical modeling, remote sensing and GIS 1.3 Overview of researched scientific works in Vietnam on issues related to the thesis topic 1.3.1 Research on water resource management, irrigation planning and natural disaster prevention a) Regarding scientific research: The research works to evaluate the fluctuation of dry flow and the impact of dry flow on agriculture and fisheries in the downstream Ca and Ma river basins, hydraulic model MIKE-NAM, hydraulic model MIKE11 and MIKE21 to evaluate the impact of irrigation systems, hydropower, transport and infrastructure on flood drainage in the Central region Based on the research results, the authors have proposed structural and nonstructural solutions to limit adverse impacts, efficiently use water resources in accordance with the specific conditions of each river basin Research and application of hydrological and hydraulic models in medium-term rain and flood forecasting for coordinated operation of water reservoirs in river basins to ensure flood prevention for downstream, and safe operation of lakes contain and limit flood damage in Ca river basin b) Regarding irrigation planning and natural disaster prevention and control for river basins: Typically projects on irrigation planning in Ca river basin to 2020 and orientation to 2030 and irrigation planning in Central region period 20122020 and orientation to 2050 in the conditions of climate change and sea level rise The mentioned irrigation plans are based on hydraulic model MIKE11, the NAM hydrological model and only focus on mainstream research based on data for the catchment area located in the territory of Vietnam, not to mention the downstream impact of upstream catchments located on foreign territories in Vietnam 1.3.2 Research using remote sensing data and mathematical model in flow simulation a) Research and use of remote sensing data to provide rainfall and topographic data in the basin The thesis has introduced an overview of a number of typical scientific works such as: 1) The study used satellite rain data from the GPM global rain measurement program, compared evaluation with ground rain monitoring data to warning of rain and flood in the city Ho Chi Minh; 2) Study to select a satellite rain data source suitable for Vietnam as a series of historical data, supplementing to replace the data at the station Research results from rain data sources including global and regional rain in NETCDF format, rain data at 58 stations in Vietnam distributed by climatic regions and radio response data from radar in Our country shows that the APHRODITE data source of Japan is the most suitable for Vietnam; 3) Research results, comparing data of GSMAP satellite rain with observed data at 10 rain gauging stations in the Central region in the period 2000-2010 show that there is a match for the month when rainfall is over 100 mm and the month with the highest rainfall at most stations, although there is a certain difference in the duration of those rainy months; 4) Study to develop a method that combines remote sensing data DEM ASTER, GPS technology, GIS data and VRSAP model to automatically generate flood maps, to analyze, monitor and warn natural disasters for the study area is the Kon-Ha Thanh river basin; 5) Study on application of image of ALOS Prism to establish DEM in areas with complex terrain such as coast, sand dunes, lagoons, accumulation plains, mountains and hills with catastrophic events (movement of dunes, open -fill the lagoon etc); and 6) Studies using LIDAR data to study features on topographic surface such as EM, K-Means, kNN, MCC and combined with high resolution DEM data for modeling flood b) Research on combining the use of remote sensing data and mathematical models in flow simulation in the basin There are not many scientific researches in this area but also achieved some good results, typically the works: 1) Applied research combining remote sensing data with IFAS model in calculating flow flow Bang Giang river basin in Cao Bang; 2) Research on applying remote sensing and GIS in monitoring environment and natural resources; 3) Research to exploit and use satellite rain data in the flood forecasting model in the Mekong river basin (from Chiang Saen to Strung Stren); 4) Apply SWAT rain-flow model in water resource management: use remote sensing data as input to calculate river flows, assess water quality under the impact of land use scenarios flow up, sedimentation and water quality assessment Most of the new studies are just using mathematical models for calculation, but have not studied and applied new technologies in this field Currently in our country there is no research that combines both mathematical model and remote sensing technology in water resource management, especially choosing remote sensing data to provide input data for the mathematical model The advantage of the method of combining remote sensing data with the mathematical model is that it is possible to determine hydrological parameters for areas where there is no document or insufficient data for calculation due to the wide coverage of remote sensing images, time is relatively continuous This is a good application in transboundary water resources management, especially for the Ca river basin where more than 35% of the area is located in the territory of Laos 1.3.3 Evaluation of research in Vietnam Most of the scientific works related to the topic use suitable mathematical models for research and calculation Research on using remote sensing data in irrigation planning and disaster prevention in river basins is not much Most of the new research deals with some specific tasks for a region or a part of the basin within the territory of Vietnam, while the part of river basins in the territory of other countries has not been mentioned and studied due to the The input data required for the mathematical models entered the nodal points in the basins outside the border are largely absent or unreliable In Vietnam, there are many transboundary river basins with existing problems in water resource management as well as irrigation planning and disaster prevention This is the gap in the research that this thesis topic needs to continue to research and choose 1.4 Conclusion of chapter The developed countries in the world have successfully applied remote sensing technology with mathematical models to simulate the flow and forecast natural disasters in the areas and river basins However, in our country this technology is still new Although the number of research projects using remote sensing and GIS data in river basin disaster prevention and planning has been limited, certain results have been achieved, creating a scientific basis for further studies Research using remote sensing data combined with appropriate technologies in meteorological forecasting, hydrology, flow and water resource management in river basins is a useful solution to overcome the lack of documents measurements, especially for transboundary river basins Ca River has a basin area of 27,200 km2 distributed in the territory of Vietnam and Laos Research on the flow in this basin is very important in warning and mitigating natural disasters downstream, especially flood flows coming from upstream However, the upstream part with 34.8% of the basin is located in Laos with no data or very little data for research Using remote sensing data combined with mathematical models in irrigation planning and disaster prevention is a new research direction, it is necessary to conduct a case study for the Ca river basin This is a very important scientific and practical basis to build this thesis topic CHAPTER RESEARCH METHOD AND DATA USED IN THE STUDY 2.1 Interpretation of research methods and tools 2.1.1 Scientific approach The thesis chooses scientific approaches, including: i) System approach; and ii) Approach combining experimental research and theoretical research Each of the above mentioned approaches has been analyzed and clarified its scientific bases The system approach is a basic and cross-cutting approach in the process of researching and implementing the thesis topic 2.1.2 Scientific research method The thesis uses research methods including: i) Inheritance method; ii) Methods of investigation, data collection, documents; iii) Statistical analysis method; iv) Mathematical modeling method; and v) Method of workshop and consultation with experts Each of the above research methods are analyzed and clarified its scientific basis 2.1.3 Research tools The dissertation's research tools are mathematical modeling software related to hydrological and hydraulic calculations and remote sensing data types that provide input data for mathematical models 2.2 Research site The place of research and application in practice of the thesis is Ca river basin Ca River has a basin area of 27,200 km2 distributed in the territory of Vietnam and Laos 65.2% of the catchment area is located within the administrative introduced the structural diagram and structural parameters of the PWRI-DHM model 2.5 Research remote sensing data analysis and processing tools For the research of the thesis, the focus on processing data from remote sensing to provide input data for the mathematical models is extremely important Therefore, it is necessary to select suitable GIS software for data processing from remote sensing data sources for analysis, evaluation and selection of appropriate data to include in the models calculate The thesis has studied and introduced open source GIS softwares including GRASS GIS; QGIS; MapWinGIS and ILWIS to choose an appropriate database building software The evaluation criteria include: 1) Sustainability: strong community of programmers, few bugs or bugs corrected quickly (upgrades regularly) 2) Popularity: used by many individuals and organizations 3) Meeting the needs of building and managing thematic data, including i) Spatial analysis (map overlapping, distance analysis, spatial interpolation); ii) Draw thematic maps according to the features of the objects; iii) Managing and linking databases; and iv) Ease of use The results of analyzing and evaluating software for building and managing GIS data, the thesis selected Quantum GIS software (QGIS) because this software has fully met the above criteria 2.6 Study on remote sensing data selection 2.6.1 Study on selecting high resolution satellite rain data a) General overview Rain is the most important input factor for hydrodynamics, hydraulics, water demand, and water balance models In the Ca river basin, the basin part of our country with 17,730 km2 has only 23 rain-gauging stations with an average density of 778 km2 / station, the remaining 9,470 km2 is located in Laos, with no rain gauge station According to the WMO standard requires a minimum of 575 km2 / rain gauging station The rain monitoring stations already exist in our territory, mainly concentrated in the plains, towns and townships located in the river valley There are few rain gauge stations in the high mountains Recorded rain records of stations out of 23 rain gauge stations in the basin are missing and interrupted, as shown in Table 2.6 11 Table 2.6 Observed rainfall data gap at the meteorological station in the Ca river basin Name of station Quy Chau Tay Hieu Observed 1961-2015 1960-2015 Quynh Luu 1960-2015 Con Cuong Quy Hop Ky Anh Ha Tinh Kim Cuong Cua Rao 1960-2015 1996-2015 1961-2015 1961-2015 1962-2015 1960-2015 No List of documents Data gap (1986- 2015) 1995, 2011 Apr, Oct, Nov, Dec/1987, Jul/1988 Jun/1987, Oct,Nov/1988, Aug,Sep/1995, from 2011 to 2013 Jun, Dec, 1995 From 1986 to 1995 From 1986 to 2000 From 1986 to 1990, 1996 and 1997 From 1986 to 2000 From 1986 to 1990, From 1996 ro 1999 The document gaps in both space and time greatly affect the calculation results of the runoff in the Ca river basin To overcome this situation, the thesis has collected high resolution satellite rain data including GPM, CHIRPS, GSMAP, CMORPH These satellite rain data are analyzed and evaluated with rainfall data measured directly at 12 level meteorological stations according to the following method: 1) Compare 1,069 days data series of years from 2015-2017 between satellite rain and data at 12 measuring stations (because GPM data only has rain data from February 2014) 2) Daily rainfall assessment (rainfall detection ability, rainfall correlation) 3) Comparing total rainfall and annual rainfall distribution between data series 4) Evaluate monthly rainfall (correlate total rainfall and calculate Pearson's linear correlation coefficients r, R2 and errors RMSE, MAE) b) Results of analysis and evaluation of satellite rain data: - For daily rainfall: Satellite's ability to identify rainy and no-rain days accurately reaches about 70% of the days Pearson correlation between measured rainfall and satellite rain reached the average r = 0.40 ÷ 0.48 is acceptable - For monthly rainfall: The correlation of total monthly rainfall between satellite rain and measured actual rain reaches 0.63 ÷ 0.76, particularly at Vinh station between measured rainfall and CHIRPS reaching R2 = 0.93 Correlation coefficient r and R2 of CHIRPS rain is best with R2 = 0.76; followed by rain of GPM (R2 = 0.72); the worst rainfall GSMAP and CMORPH (R2 = 0.65 and 0.63) RMSE and MAE error of CHIRPS rain give the best results, while GSMAP and CMORPH rain is less accurate Identifying trend of total monthly rainfall of satellite rain is quite good in the rainy season from May to October - For the total annual rainfall: CHIRPS and GSMAP rain show well the rainfall distribution in the delta and coastal areas and well show the rainfall distribution in the western and southwestern regions of the mountainous areas The CMORPH rain 12 did not show a good annual rainfall distribution compared with other satellite rain data c) Conclusion of satellite rain data selection: - Using CHIRPS rain data to create more hypothetical rain stations, combined with actual rain measured to calculate runoff from rain in Ca river basin - Using GSMAP rain data in combination with digitalized altitude modeling data (DEM) to calculate and simulate flood processes over time by IFAS model 2.6.2 Analyzing and selecting topographic data from the digitalized altitude modeling data source (DEM) a) General: Common applications of remote sensing in the study of hydrological, rain-runoff models are to determine the spatial flow-hydrological parameters for models, such as division of basins, sub-basins, and River and stream network based on topographical data, altitude distribution of the entire basin To select integrated topographic and elevation data and suit the structure of the rain-runoff hydrological model, input data on spatial resolution, time and accuracy of numerical model data altitude (DEM) must be compared DEM data sources are diverse and abundant High-resolution DEM maps for in-depth and detailed research are often very expensive, while for general studies a free source of DEM maps with a resolution of 30m-90m is relatively suitable The thesis selected DEM data with 30m-90m resolution: SRTM, ASTER, ALOS and analyzed and evaluated the differences as well as the similarities between DEM data and topographic data in the basin map Ca river at the rate of / 50,000 to choose an appropriate DEM data according to the following method: - Compare the differences in the altitude data of the three DEM datasets mentioned above - Compare each DEM data set and data from Ca river map at / 50,000 scale (altitude correlation and Pearson's correlation coefficient r, coefficient R2 and the mean square error RMSE, wrong average absolute number MAE) b) Results of analysis and evaluation of DEM data together - The DEM data (SRTM, ALOS, ASTER) are quite similar in related parameters such as maximum elevation value, minimum elevation value and especially average elevation value of Ca river basin - The DEM data for the delta area is quite similar, the difference in elevation value is not too large However, for the mountainous area, the elevation value is quite different between the DEM data - The difference of the height value between DEM data of ALOS and SRTM is the lowest: the average value of the difference is -0.29m and the lowest SD standard deviation is 12.95m - The difference of the height value between the ASTER data and ALOS, the ASTER data and the SRTM are quite similar with the mean high level, respectively -1.98m 13 and -1.75m, and the corresponding standard deviation respectively 16.53m and 17.21m c) Results of comparing DEM data with Ca river basin map Map of Ca river basin at the scale of / 50.000 coordinate system and altitude VN2000 was converted to the same international coordinate system and altitude WGS84 for comparison of these elevation points A total of 18,924 elevation points were extracted from each DEM data and divided into different regions in the Ca river basin The comparison results show: - Plain area (where the altitude is less than 50m): 2,638 altitude points have been compared, showing that the ALOS data shows the highest accuracy with RMSE = 4,724 and MAE = 3,152 With the correlation coefficient, ALOS data reaches R = 0.943, R2 = 0.93 SRTM data achieved lower accuracy with RMSE = 5,313m and MAE = 3,527m; ASTER data has the lowest accuracy - Midland area (where the altitude is from 50m to 500m): 8,524 altitude points have been compared, showing that the ALOS data has the highest accuracy, but the errors have increased significantly: RMSE is 13,826 m, MAE is 7,295m; SRTM data has lower accuracy than ALOS, ASTER data has the lowest accuracy of the data - Mountain area (where the altitude is over 500m): 7,762 altitude points have been compared, showing that ALOS data still shows the most accurately, followed by SRTM and ASTER data with the lowest accuracy with error RMSE up to 32,111m, quite high when compared to the other data d) General comments on the research results evaluating the numerical elevation model data (DEM) on Ca river basin: 1) The SRTM, ALOS, ASTER numerical model data are quite similar in terms of related parameters such as maximum height value, minimum height value and especially mean height value and The difference in average altitude in the whole Ca river basin 2) The SRTM, ALOS, ASTER data for mountainous areas have quite similar altitude values, the difference in elevation value is not too large, but for the plain area, there are differences quite large among the DEM data 3) The results of calculating the error and the correlation coefficients of altitude values between the DEM data and the data from the / 50.000 topographic map indicated, the ALOS data has the highest accuracy, the error is lowest when assessed with all types of terrain is plain, midland and mountainous; SRTM data has lower accuracy, ASTER data has the lowest accuracy 2.6.3 Conclusion on the analysis and selection of remote sensing data for Ca river basin From the above research results, in order to match the characteristics of the research basin, the thesis has chosen to use the following remote sensing data: - CHIRPS satellite rain data due to the correlation coefficient with actual rainfall data is the best among the satellite rain data that will be used to create more hypothetical 14 rain stations to increase the density of rain stations in accordance with regulations of and used as input data in calculating flow from the NAM model - GSMAP satellite rain data due to its ability to identify rain and no rain based on actual rainfall data series is the best among other satellite rain data that will be used in combination with numerical modeling data DEM ALOS high resolution of 30m to establish rain data for IFAS model in calculation and simulation of flood process over time 2.7 Conclusion of chapter In the content of this chapter, the thesis topic has analyzed the scientific basis of the approaches and methods of scientific research, the order of research steps The dissertation has synthesized, analyzed and selected the research tool of the subject which is to use suitable hydrographic and hydraulic models From the model properties and the ability to process data from remote sensing, MIKE NAM and IFAS models have been selected to apply for this study The thesis has analyzed and evaluated satellite rain data that are commonly used in the country and the world to provide input data for the mathematical model including GSMAP, GPM, CHIRPS, CMORPH The study results show that the satellite rain data mentioned above have a good correlation with the actual rain data measured at the station, in which: - CHIRPS rainfall with the best results will be used to create more hypothetical rain stations to increase the density of rain stations in accordance with regulations on distance and density of rain gauge stations in the basin and used as data input in flow calculation from the NAM model - GSMAP rain data which is capable of detecting rain and no rain based on actual rainfall data series is the best among other satellite rain data that will be used to establish rain data for IFAS model in calculation simulation of flood flow processes in Ca river basin in next study step Topographic data from the digital elevation model (DEM) was also analyzed and selected to provide input data for the IFAS distribution hydrological models The study results show that the topographic data from SRTM, ASTER and ALOS have very good correlation with topographic map data 1/50,000 VN2000 coordinate system, in which ALOS data has good evaluation results Most dissertation will be used to perform the research in chapter QGIS software has also been selected among the popular GIS software packages to serve remote sensing data processing with the advantages of many features, ease of use and open source The remote sensing tools and data selected here will be established and applied to the computational content in the next chapter The research results show that using remote sensing data in providing or supplementing input data in research on flow simulation for irrigation planning, disaster prevention in river basins, especially the River is very feasible and can be applied in practice 15 CHAPTER RESEARCH RESULTS USING REMOTE SENSING DATA IN CALCULATING THE FLOW IN CA RIVER BASIN 3.1 Study on combining satellite rain and measured rain in calculation of rain - runoff from NAM model 3.1.1 The need to combine satellite rain with measured rainfall With the existing measured rain data, the calculation results of rain - runoff from NAM in the study on irrigation planning of Ca river show that the driest flow is smaller than measured, the flood tibial flow is higher or lower compared with real measurements The reason is that the network of rain gauge stations in some areas such as Cua Rao, Nghia Khanh, Hoa Duyet is too thin, not enough to represent the characteristics of rain forming the basin flow Therefore, it is necessary to use CHIRPS satellite rain data to create more hypothetical rain stations to increase the density of rain stations, combined with real rain stations to improve the results of calculating runoff from rain on Ca river basin 3.1.2 Establish a NAM model with existing rainfall stations Based on topographic map of Ca river basin, available hydrological stations and actual measured data source, the thesis has built a set of NAM models for 11 tributaries Data series for modeling and validation the model are shown in Table 3.1 Table 3.1 Calculation period for the simulation and validation of the NAM model No River Hydrology F (km2) Calculation period stations Simulation Validation Khe Choang Coc Na 417 1963-1969 1972- 1976 Giang Thac Muoi 785 1979-1983 1972-1977 Hieu Quy Chau 1,500 1996-2008 1983-1995 Hieu Nghia Khanh 3,892 1996-2008 1983-1995 Khe La Khe La 27.8 1970-1977 1980-1985 Ngan Truoi Huong Dai 408 1965-1976 1971-1976 Song Trai Hoa Quan 150.7 1980-1985 1975-1979 Tiem Trai Tru 96.2 1965-1976 1965-1970 Rao Cai Ke Go 229 1961-1967 1968-1974 10 Nam Mo Muong Xen 2,620 1960-1964 1965-1976 11 Ca Cua Rao 12,800 1960-1964 1965-1976 Simulation and validation of the model must meet the following requirements: i) There is no difference in water balance; ii) Flow simulation in the appropriate dry season; iii) Flood peak of runoff during antiaircrafting plays an important role The NAM model parameters need to be adjusted to select the standard parameter before studying hydraulic calculation NAM correction is done 16 automatically The evaluation results of the correlation coefficient R2 of the simulation and validation are shown in Table 3.2 Table 3.2 The assessment results of the correlation coefficient R2 of the simulation and validation of the NAM model at the hydrological stations R2 R2 Hydrology Hydrology No No stations stations Sim Val Sim Val Coc Na 0.71 0.56 Hoa Quan 0.69 0.70 Thac Muoi 0.78 0.60 Trai Tru 0.77 0.65 Quy Chau 0.83 0.78 Ke Go 0.70 0.64 Nghia Khanh 0.85 0.80 10 Muong Xen 0.70 0.61 Khe La 0.71 0.62 11 Cua Rao 0.62 0.55 Huong Dai 0.77 0.72 3.1.3 Establish of assumed rain station Principle of adding assumed rain station to areas in shortage: 1) According to Circular 30/2018 / TT-BTNMT dated December 26, 2018 of the Ministry of Natural Resources and Environment, the distance between stations in mountainous areas is from 10 km ÷ 15 km 2) In accordance with the resolution of the CHIRPS satellite rain data used in the study with the spatial resolution of 0.05 ° (equivalent to about 5.55 km) 3) The stations are relatively evenly distributed across the basin and represent the flow in the sub-basin are included in the mathematical model Figure 3.2 Location of assumed rain station According to the above principles, the thesis has studied and arranged 43 additional rain stations in the basin, of which Cua Rao added stations upstream 17 in Vietnam and stations in Laos; Muong Ren added stations (2 stations in Vietnam and stations in Laos); Upstream Quy Chau added stations; The area between Nghia Khanh - Quy Chau added stations; Hoa Duyet added more stations (See figure 3.2) 3.1.4 Research results of NAM model a) Research results for the downstream Ca river in Vietnam: Combining CHIRPS satellite rain with actual rainfall data series measured from 1982 to 2018 to calculate the average discharge (LLTB) from the NAM model at locations in Quy Chau, Nghia Khanh The results of calculating the correlation coefficients in Table 3.4 and 3.5 show that this combination has significantly improved the accuracy of the flow from the NAM Table 3.4 and 3.5 Results of the correlation calculation at Quy Chau and Nghia Khanh stations Quy Chau Nghia Khanh Value Used raindata PBIAS PBIAS (m /s) R2 NSE (%) R2 NSE (%) 0.70 0.48 -13.8 0.71 0.47 4.0 Observed Combination 0.71 0.51 3.60 0.75 0.53 2.7 Observed and RS 0.83 0.65 -13.8 0.85 0.70 4.0 Observed LLTB Combination monthly 0.84 0.70 3.50 0.88 0.75 2.8 Observed and RS 0.52 -0.15 -13.8 0.73 0.45 4.0 Observed LLTB annually Combination 0.55 0.27 3.60 0.72 0.47 2.8 Observed and RS b) Research results for transboundary basins located in Laos and Vietnam Upstream Ca river has tributaries originating from Laos, namely Nam Non and Nam Mo The thesis studies the Nam Mo branch which flows into Vietnam to Muong Ren, then merges with the Nam Non branch at Cua Rao The area of Muong Xen basin is 261,350 (the part in Vietnam accounts for 14.1%, the part in Laos is 85.9%) There is only one rain gauge here The thesis establishes hypothetical rain stations in the territory of Laos and 02 stations in Vietnam (Figure 3.5) to calculate the NAM model, the data at these stations are extracted from the daily rain data of CHIRPS satellite rain LLTB daily The results of flow simulation by NAM model between two cases with and without using satellite rain data show that using satellite rain data through assumed rain stations has increased the accuracy and correlation coefficient 18 between the simulated and real measured values in all cases of calculation for mean discharge value (LLTB) daily, monthly and annually, are summarized in Table 3.6 and Table 3.7 Figure 3.5 Location of assumed rain station of Muong Xen basin Table 3.6 The calculated results of the correlation coefficient R2 and NSE coefficient at Muong Xen station Value Coefficient LLTB daily LLTB monthly LLTB annually PBIAS NSE R R2 NSE PBIAS R2 NSE PBIAS (%) Observed Combination RS (%) (%) 0.65 0.41 3.25 0.70 0.46 3.54 0.39 -0.62 3.27 0.73 0.54 0.23 0.85 0.73 0.25 0.45 0.20 0.22 Table 3.7 Calculation results LLTB monthly Muong Xen station (m3/s) Month Observed value 10 11 12 27.4 22.5 20.9 21.6 38.6 74.5 126.1 168.0 141.2 90.1 49.0 34.3 Simulated values from observed rain 48.7 40.4 33.8 28.7 39.2 64.9 73.4 119.2 143.4 116.2 76.4 58.4 Simulated values with satellite rain 28.0 21.2 16.8 17.2 45.0 67.5 116.9 164.4 158.9 86.9 54.0 39.3 Thus, combining additional satellite rain from assumed rain stations to increase station density in missing locations in Ca river basin has improved calculated 19 flow from NAM, improved accuracy and safety of the maximum value of the driest and maximum flood flows In particular, the quality of flow simulation has been improved for the basin where most of the area has no rain gauge 3.2 Research on combining remote sensing data and IFAS math model in flood flow simulation in Ca river basin 3.2.1 Research scope The thesis focuses on flood flow of Nam Non river branch to Ban Ve hydropower plant The total area of Nam Non basin to Cua Rao junction is about 8,700 km 2, of which about 20% of the catchment area is in Vietnam, the part located in Laos has 6,970 km2, accounting for about 80% of the basin area Figure 3.8 Topographic map of Nam Non river basin, upstream of Ca river 3.2.2 Establishing the IFAS model a) Set up general information about the basin for the IFAS model In the area of Nam Non basin located in Laos, there is no topographic data The thesis uses DEM ALOS data at 30m resolution and GIS tool to divide the basin and rivers and streams in the basin The results of the Nam Non basin division are included in the IFAS model through the Basin Data Manager module, see Figure 3.9 (a) (b) Figure 3.9 Distribution of (a) catchment and (b) Nam Non sub-basin 20 For the area without surface cover and geological data, using USGS Global Surface Cover (GLCC) data and FAO's Worldwide Soil Data Data (DSMW) Map, see Figure 3.12 (a) (b) Figure 3.12 Data (a) geology and (b) surface cover in the IFAS model b) Set up IFAS model parameters for Nam Non basin (Parameter Manager): For simulating the flood flow with a relatively fast time, taking place in hours or days, the thesis has selected a two-layer tank model to simplify the model as well as speed up the running Figure Parameters of surface runoff tanks and rivers are set up in the IFAS model as shown in Figure 3.13 (a) (b) Figure 3.13 Parameter (a) surface flow and (b) river course c) Rainfall Data Manager (IFAS) model: To set rain data for the IFAS model, use the Rainfall Data Manager module The thesis selects GSMAP satellite rain based on the characteristics and advantages of this data compared to other satellite rain data 3.2.3 Simulation, calibration and validation of IFAS model The scope of the IFAS model study is the flood flow on the Nam Non river to Ban Ve hydropower In 2010, Ban Ve hydropower plant was put into operation From 2011 to 2019, there appeared four major floods on the Nam Non stream to Ban Ve in June 2011, August 2016, August-September, 2018 and August 2019 The thesis will simulate and calibrate IFAS for the Nam Non flood flow in June 2011, validate the flood model in August 2019 and 8/2018 a) Rain adjustment method and IFAS model parameter correction 21 - To calibrate the GSMAP rainfall in the IFAS model, it is necessary to adjust the coefficient extracted from the relationship between the traveling speed of the rainy area and the rainfall error coefficient calculated from the Robs and Rsat indices in which Robs is precipitation at the ground measuring station (mm/3h), Rsat is satellite rainfall (GSMAP) (mm/3h) - The model parameters including the surface layer and the aquifer layer have been calibrated to ensure the most accurate simulation of flood peak occurrence time and maximum flood discharge b) Simulation and calibration of IFAS model for the June 2011 flood: Comparison of maximum discharge value between actual measured and simulated flood in June, 2011 of Nam Non basin at Ban Ve is shown in Figure 3.17 and 3.18 The study results show that after adjusting the model parameters, the flood peak water level has been adjusted closer to the actual measured value Calculation results show that the correlation coefficient between real measured and simulated reaches R2 = 0.90 and the coefficient NSE = 0.95 2500 2500 Observed Simulated Simulated Q (m3/s) Discharge Q (m3/s) 2000 1500 1000 y = 1.2578x - 453.96 R² = 0.9016 2000 1500 1000 500 500 0 500 1000 1500 Observed Q (m3/s) 28/6/2011 25/6/2011 22/6/2011 19/6/2011 16/6/2011 13/6/2011 10/6/2011 7/6/2011 4/6/2011 1/6/2011 2000 2500 Figure 3.17 and 3.18 Comparison of results between simulated and observed flood in 06/2011 c) Validation of the IFAS model with the flood in August 2019: After adjusting the rainfall in the IFAS model corresponding to the flood in August 2019, continue to study the adjustment of the flow value using parameters in the IFAS model with the flood from August 1st to August 10th, 2019 The results of adjusting the model parameters with this flood are shown in Figure 3.22 3.23 Correlation coefficient between simulation and real measurements: R2 = 0.76 and NSE coefficient = 0.54 3500 4000 Simulated Q (m3/s) Discharge Q (m3/s) 3000 Observed 2500 Simulated 2000 1500 1000 500 3000 2000 y = 1.1059x + 101.3 R² = 0.756 1000 0 10/8/2019 9/8/2019 8/8/2019 7/8/2019 6/8/2019 5/8/2019 4/8/2019 3/8/2019 2/8/2019 1/8/2019 500 1000 1500 2000 2500 3000 3500 Observed Q (m3/s) Figure 3.22 and 3.23 Comparison of results between simulated and observed flood in August 2019 22 d) Validation of the IFAS model with the flood in August-September 2018 The results of the model test with the flood from Aug 10th to Sep 10 th, 2018 are shown in Figure 3.25 and 3.26, showing that the correlation coefficient between simulation and real measurements is R2=0.83 and the coefficient NSE=0.81 5000 4500 Observed 3500 Simulated 3000 2500 2000 1500 1000 500 7/9/2018 3/9/2018 30/8/2018 26/8/2018 22/8/2018 18/8/2018 14/8/2018 10/8/2018 Simulated Q (m3/s Lưu lượng (m3/s 4000 5000 4500 4000 3500 y = 1.0699x - 293.45 R² = 0.8982 3000 2500 2000 1500 1000 500 0 1000 2000 3000 Observed Q (m3/s) 4000 5000 Figure 3.25 and 3.26 Comparison between simulated and observed flood in the August -September 2018 3.2.4 Conclusion on IFAS model research results The thesis has studied using IFAS model with input data from remote sensing data including GSMAP satellite rain, ALOS topography, global cover (GLCC) and worldwide land (DSMW) to simulate flood flow for the Nam Non basin in the upstream of Ca river to Ban Ve hydropower plant Results of simulation and calibration of satellite rain data and parameters of IFAS model for the 6/2011 flood, validation with the flood in 8/2019, August-September 2018 for the correlation coefficients R2 and NSE between acceptable measurement and simulation Thus, using remote sensing data combined with a mathematical model capable of simulating flood flows for areas with little or no data, support the operation of Ban Ve hydropower and actively respond to other problems arise of water resources in Ca river basin 3.3 Conclusion of chapter Results of studies that incorporate remote sensing data and mathematical modeling into flow simulation for Ca river basin show: - Using CHIRPS satellite rain data to create more assumed rain stations to increase station density for the missing areas and as input data of the NAM model improved the simulation results of runoff on Nam Mo River both flood season and dry season with actual data measured at Muong Xen hydrological station - Using IFAS model combined with input parameters from remote sensing data to simulate floods occurring in Nam Non river basin to ensure accuracy and suitability for the formation of floods occurred in reality The research results have practical significance, meeting the research objectives of the topic, demonstrating the effectiveness of supplementing data from satellite sources, remote sensing in setting up flow calculation models and hydrographic modes resources for irrigation planning, natural disaster prevention and control in Ca river basin as well as transboundary water management The study results can be applied to river basins with little or no data to measure meteorological and hydrological factors 23 CONCLUSIONS AND RECOMMENDATIONS Conclusions 1) The results of an overview of domestic and foreign scientific and technological works related to the thesis topic show that up to now, there has not been any research in our country that combines both mathematical models and Remote sensing technology in flow calculation for irrigation planning, natural disaster prevention, especially research, analysis and data processing to provide input data for this mathematical model 2) The thesis has selected appropriate research approaches, methods and tools These are: i) MIKE NAM and IFAS models to simulate hydrological and hydraulic modes; ii) CHIRPS satellite rain to create more assumed rain stations and as input data for simulation models of runoff in Ca river basin; iii) GSMAP satellite rain and DEM ALOS in rainfall calculation and flood simulation over time; and iv) QGIS software for building, processing, analyzing and managing GIS and remote sensing data for the research topic 3) The thesis has successfully researched combining remote sensing data and mathematical modeling in flood flow simulation for Nam Non basin in the upstream of Ca river outside our country The study combining the measured rainfall and the CHIRPS satellite rain simulating rainfall-runoff from the NAM improved the accuracy of the flow simulation in Nam Mo river basin IFAS used GSMAP satellite rain data, well simulated the flood flow in the river basin across Nam Ne border 4) Using remote sensing data to improve the quality and accuracy of mathematical models that simulate flood flows, dry flows, spatial and temporal rain distribution for areas with little or no have real measured data The research results contribute to create a good supporting tool for water resource management and proactively cope with problems arising in the river basin with little or no document to measure the factors meteorology and hydrology Recommendations The research results are only the first step in the selection of using satellite rain data, the appropriate altitude numerical model to calculate and simulate transboundary flow in Ca river basin serving irrigation planning, anti-disaster There are many factors affecting the runoff, which need further research using remote sensing data to supplement data on land cover, soil type, evaporation and underground runoff for Ca river basin Especially, the adjustment of satellite rain data needs to continue to study in the next stage 24 LIST OF PUBLICATIONS Bui Tuan Hai and Le Viet Son, "Research and application of IFAS model and remote sensing data in simulating cross-border flood flow in Thao river basin", Journal of Meteorology - Hydrology, No 713, 05/2020, pages 24-36, ISSN 2525 - 2208 Le Viet Son, Luong Ngoc Chung, Bui Tuan Hai, Sai Hong Anh and Nguyen Duy Quang, "Assessing Satellite-Based Precipitation Products to Create Flood Forecasting in the Da River Basin, Vietnam," Journal of Geoscience and Environment Protection, vol 7, no.11, pp 113-123, 2019 Bui Tuan Hai and Le Quang Vinh, "Research and application of combining remote sensing data and IFAS model in flood flow simulation in Nam Non river basin in Ca river system", Journal of Agriculture & Rural Development, No 369, 18/2019, pages 96-101, ISSN 1859-4581 Bui Tuan Hai, Vuong Tan Cong and Pham Quang Vinh, "Comparing and evaluating the digital elevation model (DEM) data on Ca river basin", Proceedings of the National Geological Science Conference XI 2019, Hue University, City Hue, Thua Thien Hue province, April 2019, Book 2, pages 881890 Bui Tuan Hai and Nguyen Van Tuan, "Research, evaluation and comparison of high resolution satellite rain data in Ca river basin," Journal of Meteorology Hydrology, No 695, 11/2018, page 17-28, ISSN 2525 - 2208 ... meteorological station in the Ca river basin Name of station Quy Chau Tay Hieu Observed 1961-2015 1960-2015 Quynh Luu 1960-2015 Con Cuong Quy Hop Ky Anh Ha Tinh Kim Cuong Cua Rao 1960-2015 1996-2015 1961-2015... from the NAM Table 3.4 and 3.5 Results of the correlation calculation at Quy Chau and Nghia Khanh stations Quy Chau Nghia Khanh Value Used raindata PBIAS PBIAS (m /s) R2 NSE (%) R2 NSE (%) 0.70... added stations (2 stations in Vietnam and stations in Laos); Upstream Quy Chau added stations; The area between Nghia Khanh - Quy Chau added stations; Hoa Duyet added more stations (See figure 3.2)