Investigating Sediment and Velocity Distribution Profiles for Nubia Lake Using RS/GIS and Field Data

Available online at www.sciencedirect.com ScienceDirect Procedia Engineering 154 (2016 ) 291 298 12th International Conference on Hydroinformatics, HIC 2016 Investigating Sediment and Velocity Distribution Profiles for Nubia Lake Using RS/GIS and Field Data Mohamed Elsahabi a, Abdelazim Negm b * and Kamal Ali a a PhD student, Civil Engineering Dept., Aswan University, Aswan 81542, Egypt b Head of Natural Resources Lab., Department of Enviromental Eng, Egypt-Japan University of Science and Technology E-JUST,New Borg EL- Arab, Alexandria, Egypt and Prof. of Hydraulics at Faculty of Engineering, Zagazig University, Zagazig, Egypt Abstract Monitoring the changes provides the decision maker with sufficient information to manage the storage capacity of the lake. Therefore, the main aim of this paper is to detect the changes in AHDL bed surface profiles in the period 2000-2012 using the remote sensing (RS) and (Geographic Information Systems) GIS techniques via building 3D profile of the lake. On the other hand, the measured values and directions of velocities along with the interpolation function are used to map, analyze and correlate the velocities pattern to the variation of the bed profile (the erosion and sedimentation patterns) and to the sediment component (particles) distributions. 2016 The The Authors. Published by Elsevier by Elsevier Ltd. Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of HIC 2016. Peer-review under responsibility of the organizing committee of HIC 2016 Keywords: Aswan High Dam Lake; Sediment and erosion ; Bed Surface changes; Velocity patterns, RS/GIS Introduction 1. Introduction By the construction of the AHD in 1964, a large reservoir, AHDL, has been formed at the dam s upstream side. This Lake is consisted of two main connected parts. One lies in Egypt and is known as Lake Nasser and the second lies in Sudanese side and is known as Lake Nubia. Sediment accumulations, in this lake, resulted in decreasing the lake storage capacity, navigation problems and other water quality and environmental problem [1]. Whereas, AHDL is * Corresponding author. Tel.: +201005735345; fax: +2034599520. E-mail address: negm@ejust.edu.eg 1877-7058 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of HIC 2016 doi:10.1016/j.proeng.2016.07.477

292 Mohamed Elsahabi et al. / Procedia Engineering 154 ( 2016 ) 291 298 vital to Egypt, as it stores and regulates Nile water, being the main source of fresh water, for about 85% of its population. Consequently, it is of the outmost importance for Egypt to understand properly the changes occur in this lake, especially in its bed surface where the sediments are accumulated. This is essential to monitor the changes in the bed surfaces and hence the current actual capacity of the lake can be assessed. Satellite data have been combined with in-situ measurements, using GIS analyst tools, to estimate changes in lakes bed surfaces and analyze the effect of inflow velocity on sediment and erosion patterns, e.g. [2]. The morphological changes due to erosion / sedimentation processes occurred in San Giuliano Lake, Basilicata Region (Southern Italy), during the period 1984 2004 were analyzed using RS data integrated in GIS technique [2]. Nasr and Tarek [3] presents preliminary results on effect of the water velocity direction on distribution patterns of sediment deposition in some detected cross sections of AHDL based on field data analysis. The main target of this paper is to detect the effect of inflow velocity on sediment and erosion patterns of the active sedimentation portion in AHDL using both of remotely sensed data and in-situ data. 2. Study Area and Data Collection 2.1. Selection of the study area The most active portion of the AHDL is the main focus of this study which is located in the Sudanese part of the lake (Lake Nubia). In addition, from the studies and observations done by the repetitive field survey missions, which successfully carried out through the joined efforts of the Aswan High Dam Authority (AHDA) and the Nile Research Institute (NRI), it is obvious that the cross sections have been enlarged and the water velocity have been decreased in the chosen part of Lake Nubia. So, this part represents the area with most intensive sediment deposition. Fig.1. Location of the study area in AHDL. Moreover, the amount of sediment deposited in the study portion of AHDL is about (50%-70%) of the total amount of sediment in AHDL as indicated in the foregoing estimations by AHDA [4], although this portion represents only about 5.96% of the total area of AHDL [5]. Therefore, this reach is called (the active sedimentation portion).the study area of the present research extends between latitudes 21 44 30 N and 22 00 00 N (upstream AHD). It contains 6 cross sections (28, 27, 26, 25, 24 and 22) from the South to the North respectively as indicated in Fig. 1. 2.2. In-situ data Hydrographic survey data: The hydrographic survey data - which describe the geometry of AHDL (Easting, Northing, and Elevation), were conducted via field trips by using the echo-sounder. These trips were done by (AHDA and NRI). The hydrographic survey data of years 2000, 2008, 2010 and 2012 for the study area were used in this paper [4]. Water levels data: Water levels upstream AHD were recorded by AHDA gauge stations in the different dates of the year [6] and they were collected to help in detecting the water surface levels of th lake at the dates of acquiring the satellite images.

Mohamed Elsahabi et al. / Procedia Engineering 154 ( 2016 ) 291 298 293 The inflow velocity data: These data involve the velocity magnitudes and the inflow directions, were measured by using the Vale port velocity meter device via the above mentioned field trips. These data were obtained at the locations of the cross-sections shown in Fig. 1. In this study, the data about the inflow directions of the year 2012 and the velocity magnitudes data of years 2008, 2010 and 2012 are used [4]. Lake bed soil samples data: The Lake bed material samples were collected via the field trips using the bed material sampler at locations of the cross-sections shown in Fig. 1. Then the grain size distribution of these samples was identified in AHDLA and NRI laboratories to investigate the nature of the deposited particles and classifying them according to their geometrical characteristics. The collected samples of the year 2012 are used in the current study. Though studying of these samples, it is noticed that; sand medium grained size particlesand (silt & clay) fine to very fine grained size particles are the major components of these samples. Meanwhile; these samples do not contain any gravel coarse grained size particles - at all. 2.3. Satellite images (remote sensing data) Three Landsat ETM+ images (Path/Row=175/045) were used in this research. These images were acquired in different dates (September 2000, March 2006 and March 2009) from the GLCF website in GeoTIFF (systematic correction) products [7]. The acquired images were used to extract the lake boundaries. The satellite images were shot in September 2000 where the water level in the lake was (178 m) amsl, March 2006 where the water level was (173 m) and in March 2009 where the water level was (176.60 m). 3. Methodology To achieve the objective of the present paper, the methodology presented in Fig. 2 is used as explained in the next subsections. 3.1. Water surface areas extraction Fig. 2 Flowchart of the procedures adopted in this study to achieve its goals. Many studies for water areas estimation were carried out over world [8, 9, 10, 11, 12]. Considering AHDL, few water surface extraction studies have been carried out on AHDL [13, 1]. Recently, the use of remotely sensed imagery data is increasing with many applications in the field of water resources management. These applications involved the extraction of water information by various techniques [10].

294 Mohamed Elsahabi et al. / Procedia Engineering 154 ( 2016 ) 291 298 The unsupervised classification technique of Landsat images to obtain the water body class of AHDL was performed, in this study because it is considered the best technique for water texture recognition [14]. The shape of the lake was formed by using the extracted lake boundaries from the satellite images. Also, these data were used to form a group of scatter points (x,y,z) using the WGS84, UTM Z36N as a defined projected coordinate system. These points were used, combined together with the hydrographic survey points in the generation of the 3D bed surfaces of the study area for all available years (2000, 2008, 2010, and 2012). 3.2. Prediction of the 3D bed surfaces In order to predict the original lake 3D bed surfaces from the year 2000 to 2012, the available hydrographic survey data combined with the points that represent the water surface areas which derived from Landsat satellite images, were used in the interpolation process. In addition the water surface, that represents the highest surface for all predicted bed surfaces, was for the year 2000 Landsat image. The interpolation process is performed with the Radial Base functions (RBF) method [15]. the Mean Absolute Error (MAE) was used to assess the accuracy the interpolation methods. Smaller the MAE, the accurate is the results. 3.3. The inflow velocity contour maps In order to establish the inflow velocity contour maps for years 2008, 2010 and 2012, the available inflow velocity magnitude data were used in the interpolation process. This process was performed by using ArcGIS software. These contour maps were produced in order to explain the effect of the velocity magnitudes on the erosion and sedimentation patterns. 3.4. Establishing the inflow direction maps The inflow direction map of year 2012 was derived from the interpolation process of the inflow direction data. This map was generated to detect the effect of the inflow direction on the sediment distribution and formation. 3.5. Producing of the sediment particles distribution maps The sediment particles distribution maps generated from the interpolation process of the bed soil samples data of year 2012; represent the distribution of sand and (silt & clay) through the bed of the study area. 4. Results 4.1. Creation of the 3D bed profiles The 3D bed surface of the lake was produced by using the extracted water surfaces generated by using the unsupervised technique from the available Landsat images for the year 2000 to 2012. The Radial Base functions (RBF) method for interpolation was used for gently varying surfaces such as elevations [15]. To assess the accuracy of the used RBF method, MAE was computed which is almost near to zero for all studied years. The MAE for the year 2012 equals 0.02554 indicating high accuracy of interpolation process. The 3D bed surfaces are predicted for the years 2000, 2008, 2010 and 2012. Sample results are presented in Fig. 3 for the year 2000 and year 2012. Fig. 3 Typical results for the predicted bed surfaces for the years 2000 and 2012.

Mohamed Elsahabi et al. / Procedia Engineering 154 ( 2016 ) 291 298 295 4.2. Sedimentation distribution and delta formation According to, sedimentation distribution and delta formation map that was produced from subtracting process between the bed surfaces of years 2000 and 2012 as shown in Figure 4. It is obvious that there was a variation in sedimentation thickness from the year 2000 to 2012 along the study area, the maximum depth of sediments through this period was estimated to be about 28 m. Fig. 4 Sedimentation distribution and delta formation map for the period (2000 to 2012). 4.3. Velocity maps Data S. Then, the inflow direction data of 2012 was interpolated too, by using ArcGIS software. Afterwards, the velocity contour maps and the inflow direction map were produced. Figs. (5a, 5b and 5c) show the inflow velocity contour maps for years; 2008, 2010 and 2012 respectively. Fig. 5 The inflow velocity contour maps: (a) map of year 2008; (b) map of 2010; (c) map of 2012. The inflow direction map of year 2012 is generalized by the interpolation process of the inflow directions field data. Fig. 6 represents the overlying of the inflow directions on the delta formation map of Fig. 4. 4.4. The lake bed soil particles distributions maps Fig. 6 Overlying of the inflow directions on the delta formation map. Figs. (7a and 7b) show the lake bed soil particles distributions maps of year 2012 that represent the sediment components map. These maps represent sand distribution map and (silt & clay) distribution map respectively which were generated from the interpolation process of the collected soil samples data using ArcGIS software. According to these maps, it can be clearly noticed that, silt & clay particles (fine soil particles) are distributed with high amount all over the study area. On contrast, the sand particles are distributed with low amount, which indicate

296 Mohamed Elsahabi et al. / Procedia Engineering 154 ( 2016 ) 291 298 that (silt & clay) are the predominant components of the deposited sediment particles. Fig. 7 The lake bed soil particles distributions maps of year 2012: (a) sand distribution map; (b) silt & clay distribution map According to these maps, it can be clearly noticed that silt & clay particles (fine soil particles) are distributed with high amount all over the study area. On contrast, the sand particles are distributed with low amount, which indicate that (silt & clay) are the predominant components of the deposited sediment particles. 5. Discussions According to the interpolation and analysis of the inflow direction, the inflow velocities values and the lake bed soil samples, the related results were obtained as follows: 5.1. The relation between the inflow direction and the deposited sediments By comparing the inflow directions with the delta formation, it is observed that; the mean inflow direction is almost coincide with the same direction of the deposited sediments to the northern east side and with the same trend of the formed sedimentation delta as shown in Fig. 6. 5.2. Effect of the inflow velocities magnitudes on the lake bed surface relief The inflow velocity magnitudes affect directly on the lake bed surface relief as it is obviated in the following examples. 1.An example of the bed surfaces cross sections at sections (24 and 25) - which are represented above in Fig. 1- and the corresponding inflow velocity profiles are shown in Figs. (8 and 9). It is noticed that; the increase in the velocity rates is associated with decrease in bed surface levels (high amounts of erosion), as it was occurred during 2008 to 2010. On the other hand; the decrement in the velocity rates is accompanied with increment in bed surface levels (high amounts of sedimentation), as it was occurred during 2010 to 2012. Fig. 8. Comparison between bed and velocity profiles at section 24 (347 Km U.S AHD): (a) Bed surfaces cross sections; (b) Velocity distribution Fig. 9. Comparison between bed and velocity profiles at section 25 (352 Km U.S AHD): (a) Bed surfaces cross sections; (b) Velocity distribution

Mohamed Elsahabi et al. / Procedia Engineering 154 ( 2016 ) 291 298 297 2. Another example for indicating the effect of the inflow velocity magnitude on sediment and erosion pattern is shown in Fig. 10, which illustrates the comparison between changes of the bed profiles - (longitudinal sections that pass through the lowest points in the study area bed surface) and the corresponding inflow velocity profiles for the years 2008, 2010 and 2012. Figures 10 confirmed the previously mentioned results about the velocity effect on sedimentation and erosion patterns through the achieved comparisons in Figs. (8 and 9). Fig. 10. Comparison between bed surfaces and velocity profile :( a) Bed surfaces longitudinal sections; (b) velocity distribution 5.3. Effect of the inflow velocities magnitudes on the sediment component (particles) distributions Fig. 11 indicates the comparison between the sediment particles profiles - (longitudinal sections in the lake bed soil particles distributions maps that pass through the lowest points in the study area bed surface) - and the corresponding inflow velocity profiles for the year 2012. Fig. 11. Comparison between the sediment particles profiles and velocity profile of year 2012: (a) longitudinal section in sand map; (b) longitudinal section in silt & Clay map; (c) velocity distribution In general, Fine grained size particles (silt & clay) deposits accumulate in low velocity. On contrast medium grained size particles (sand) deposits accumulate in high flow velocity. Whereas, the inflow in the study area is characterized by low inflow velocity as shown above in Figure 10 and hence, the most deposited particles are fine which confirm the particles distributions in the particle distribution maps. On the other hand, most of sand particles deposited in the southern narrow part of the study area with relatively low quantity (5 to 1.5% of total sediment amount) up to velocity 0.07 m/sec (the interference margin between fine and medium sediment particles), then these particles nearly deplete in the northern part as deduced from Fig. 11.

298 Mohamed Elsahabi et al. / Procedia Engineering 154 ( 2016 ) 291 298 Finally, it can be noticed that there was a great effect of the inflow velocity on the sediment and erosion patterns. Meanwhile, a direct relation between the change in velocity and the difference in sediment and erosion amounts (change in bed surfaces levels) can't be detected because of the lack of data for some years. So, it is recommended to repeat the field trips annually to collect the data continuously. This data will help in the future studies to detect this relation. 6. Conclusions This paper presents and discusses the variations in bed profiles (sediment / erosion) of the active sedimentation Portion of Aswan High Dam Lake, based on the created 3D bed surfaces for this portion, using RS/GIS techniques. Moreover, the measured velocity patterns are mapped, analyzed and correlated to the variations in bed profiles (the erosion and sedimentation patterns). According this correlation, it can be concluded that the increment in the inflow velocity rate is associated with the erosion phenomenon. On the other hand, the decrease in the inflow velocity rate is accompanied with the sedimentation phenomenon. References [1] M. S. El-Sammany, A. M. El-Moustafa, Adaptation of surface water modeling system for sediment transport investigations in Lake Nasser, Journal of Nile Basin Water Science and Engineering, Vol. 4, Issue 1, 2011, pp. 71 85. [2] S.L. Curzio, F. Russ, M. Caporaso, Application of remote sensing and GIS analysis to detect morphological changes in an artificial lake, International Journal of Remote Sensing & Geoscience (IJRSG), Vol. 2, Issue 4, ISSN No: 2319-3484, 2013. [3] H. Nasr, M. Tarek, Effect of transverse water velocity distribution on sedimentation at aswan high dam reservoir, Eighteenth International Water Technology Conference, IWTC18, Sharm ElSheikh, 12-14 March 2015, pp. 105 119. [4] NRI: Nile Research Institute, Annual report of Sedimentation in Lake Nubia Wadi Halfa Field trips- (1973-2012), National water Research Center, Cairo, Egypt, 2012. [5] MWRI: The Ministry of Water Resources and Irrigation, Egypt, Nile Water Sector. Annual report (2011-2012), 2012. [6] MALR: The Ministry of Agriculture and Land Reclamation, Egypt, the General Authority for AHDL Development, AHDL levels (1978 to 2010), 2010. [7]GLCF: The Global Land Cover Facility, provides earth science data and products. Available from: http://glcfapp.glcf.umd.edu/data/landsat/, last accessed: May 09, 2014. [8] G.L. Feyisa, H. Meilby, R. Fensholt, S.R. Proud, Automated water extraction index: A new technique for surface water mapping using Landsat imagery, Remote Sensing of Environment, 140, 2014, pp. 23 35. [9] F. Ling, X. Cai, W. Li, F. Xiao, X. Li, Y, Du, Monitoring river discharge with remotely sensed imagery using river island area as an indicator. Journal of Applied Remote Sensing, 6: 063564, 2012, pp. 1 14. [10] H. Xu, Modification of normalised difference water index (NDWI) to enhance open water features in remotely sensed imagery. International Journal of Remote Sensing, 27(14), 2006, pp. 3025-3033. [11] Z. Du, B. Linghu, F. Ling, W. Li, W. Tian, H. Wang, Y. Gui, B. Sun, X. Zhang, Estimating surface water area changes using time-series Landsat data in the Qingjiang River Basin, China. Journal of Applied Remote Sensing, 6: 063609, 2012, pp. 1 16. [12] M. Ma, X. Wang, F. Veroustraete, L. Dong, Change in area of Ebinur Lake during the 1998 2005 period. International Journal of Remote Sensing, 28, 2007, pp. 5523 5533. [13] E. Muala, Y. Mohamed, Z. Duan, P. Zaag, Estimation of Reservoir Discharges from Lake Nasser and Roseires Reservoir in the Nile Basin Using Satellite Altimetry and Imagery Data. Remote Sens., 6, 2014, pp. 7522-7545. [14] M. A. Elsahabi, A. A. Negm, M.A. El-Tahan, Performances evaluation of surface water areas extraction techniques using landsat ETM+ Data: case study Aswan High Dam Lake (AHDL), Procedia Technology Journal, 22 ( 2016 ) 1205 1212 [15] ESRI: Environmental Systems Research Institute, Help topics of ArcGIS version 9.3 Desktop, Developer center of geographic information systems (GIS) software, Redlands, California, 2008.