الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 160 |  |  |
| | | from | 160 | | |
Abstract
water scarcity in Egypt requires better management of water resources. The most important parameters needed for lake water management are the lake surface area and volume at any specific time. This study focused on developing an innovative technique that enables the accurate calculation of the lake water surface area and volume at any specific time using satellite images. Short-term and long-term times for water body observation are essential for powerful administration and preservation of High Aswan Dam Reservoir (HADR) as it is considered to be the main water assets in Egypt, which is hugely benefited from the approach of remote-sensing technology. Nonetheless, an efficient just as powerful technique to perform water discovery from satellite images stays challenging because of the different commotion sources from heterogeneous backgrounds. Among the researches that investigated lake water surface is a research that planned a philosophy that detached free-surface from multi-temporal so as multi-spectral images from Landsat-8 OLI and sentinel-2 satellites.
To achieve the objectives of the study, many water Indices of image-processing techniques are investigated to delineate the water body boundary.
The study shows that the MNDWI is most useful technique to delineate water surface from other land features. The processing of Landsat-8 and setienal-2 satellite images for the purposes of this research, shows very high degree of consistency for the scale 1:50,000. The calculation of water body is applied for all satellite data coverages representing the lake surface area in different dates in the years 2015, 2016 and 2021 using 41
119
scenes from both satellites (Lansat-8 and sentinel-2). The 41 scenes are chosen from more than 150 scenes downloaded to select the proper scenes representing nine complete lake surface coverages. Both types of data (Landsat-8 and setienal-2) show very poor results if used with scale 1:25,000 and of course, the higher scales. The selected nine dates images were representing lake different levels varies from 173.49 to 181.24 meters which corresponds to the surveyed data by HYDROWEB project. Analysing the water levels as a variable and the corresponding lake surface area, lead to the following mathematical power equation. The equations found to fit a curve that represents the relation between the water level (WL) and surface area with a very good -proportional of variance statistical measure- R2 value of 0.997.
Lake Surface area = 8.12548* 10-12 * (wl) 6.589258217Km2
The integration of the lake surface area equation is used to obtain the lake volume equation in WL variable as follows:
Lake volume (wl)
= ∫
= 1.07065*10-15 * (wl) 7.58926 - 1.614989474 BCM
For validation purposes four test images obtained in dates different from those used to develop the surface mathematical equation and Digital Elevation Model (DEM). Images acquired in 24/7/2020 and 25/1/202 by sentinel-2 satellite. The investigated water levels, corresponding to the test images acquisition dates are compared to water level values obtained from HYDROWEB data. The HYDROWEB water level value at
120
27/7/2020 was 177.89 m, and at 25/1/2021 was 180.56. It is concluded that the deviation RMS of the investigated water level values from HYDROWEB as a reference is 13 cm.
For the purposes of accuracy assessment of the lake area calculation and boundary of the processed images from Landsat-8 and sentinel-2, the google earth images with sub-meter resolution are used for the corresponding lake levels. 7 sites around the lake boundary are randomly selected. The selected images from both sources were representing the exact area boundary and location to ensure the comparison accuracy. The water surface area is delineated for each pair of equivalent images and calculated for each of them. It is found that the difference in area percentage varies from 1.34% to 3.44% with an average difference of 2.22%. It is also noticed that the difference is always positive in reference to Google images. It means that, the water body area calculated from the low resolution images (Lansat-8 and sentinel-2) is slightly greater than the corresponding Google Earth images by 2.22%.
The validation process was proven to be a very successful monitoring system for lake level estimation and consequently, the surface area and volume in any date by acquiring a single satellite image such that it covers any small part of the generated monitoring sites.
The thesis is consisting of six chapters including introduction, literature review from more than 126 papers and website. Moreover, Data Acquisition and Manipulation chapter discussing the data acquisition and organization of more than 150 scenes from Landsat-8 and sentinel-2 satellites. Additionally, HADR water level data recoded in the period from 2015 to 2021 is collected from HYDROWEB project. Materials and
121
Methods chapter describing the used techniques and image processing functions using ERDAS IMAGING software. Results and Discussion chapter demonstrating the study results. Finally, conclusion and recommendation chapter is included to summarize the study conclusion.
Future Research
During the thesis work, some important research points have been found. Those points can be summarized as follows.
1- Delineating water boundary in shallow water areas from medium resolution Satellite images.
2- Delineating water boundary in shallow water areas from high resolution Satellite images.
3- Delineating water boundary in sea (salty water) shoreline areas from Satellite images.
4- The use of SAR data in identifying the water surfaces.
5- Satellite images threshold value determination using artificial intelligence techniques.
6- Khore area determination of High Aswan Dam Reservoir using Satellite images.
7- HADR sediments identification and its amount value determination.