Accuracy Evaluation of Satellite-Based Bathymetry Data Extracted from Digital Elevation Model against Echo Sounder Measured Data in Gurara Reservoir, Nigeria

data was collected from bathymetry survey conducted using Echo Sounder. Materials and Method Case Study The research was conducted in the Gurara reservoir which lies between latitude 8°15' and 10°05' N and longitude 6°30' and 8°30' E on the Gurara river in the semi-arid region of North west Nigeria to obtain the measured near-reliable elevation data for all water uses in the reservoir and useful information for current and future water uses are provided with the development of this important tool as the water body is relatively calm allowing sunlight to penetrate to a maximum depth that can captured by satellites. The reservoir effectively controls 86% of the sediment inflow and 80% of the flood inflow of upstream reaches and meets the irrigation and municipal IJEMT DOI: 10.56201/ijemt..pg198.203 water supply to Abuja (FMWR, 2002). Therefore, the safety of the reservoir directly relates to the safety and property of people of the Abuja and the downstream riparian population. The extracted elevation data has a scale of scale 1:50,000 obtained from the USGS website (https://earthexplorer.usgs.gov) and shown in Figure 1. Figure 1: Case study map showing Gurara Reservoir, the Catchment and Country boundary Location Digital Elevation Model The digital elevation model as the satellite-derived imagery has a resolution of 90 m from the Shuttle Radar Topography Mission. The digital elevation model was used for the extraction of the relevant quality hydrographic data to aid the satellite-derived technique for this study. Also, the digital elevation model was used to analyze the drainage patterns of the terrain, slope and slope length of the terrain which are parameters of the catchment as well as the stream network characteristics: channels slope, length, and width were obtained from the digital elevation model. A sink and fill operation were also performed on the digital elevation model. IJEMT DOI: 10.56201/ijemt..pg198.203 Extraction of Digital Data The extraction of the digital data involves the application of aerial multispectral data and radiometric technique to estimate depth of water. According to Terrain Zonum, (2015), the method shows electromagnetic spectrum to having positive potential in its ability to penetrate water columns, which presents the background knowledge to enable water depth extraction in principle. In its physical application, it involves passing light rays through columns of water which makes it attenuated, with shallow water appearing bright and darks indicate deep water areas. This is because solar radiation generally when reflected from the earth undergoes significant interaction with the atmosphere and in water column. The methodical flow chart to extract digital data especially water depth to achieve the objectives of this study is shown in Figure 2. Broadly, two types of data were used in this study, the Echo Sounder Bathymetry elevation data obtained from National Water Resources Institute and topographic data extracted from DEM using Geographical Information Program While the Echo Sounder Bathymetry elevation data obtained from NWRI involves the use of a boat, handheld GPS, the odelite, stadia rod, tripod, leveling staff and measuring tape. The extracted elevation data using Geographical Information Program in the form of ‘x,y,z’ were processed from the Geographic Information System laboratory in the National Water Resources Institute IJEMT DOI: 10.56201/ijemt..pg198.203 Geographical Information Program The geographical information program utilizes the shoreline mapped to mask the reservoir area from the digital elevation model in order to compliment the dataset generated for the bathymetry survey. The program is based on ratio imaging which enhances the presence of open water surface using a Normalized Difference Water Index and takes advantages of the difference in the reflectance of different wavelength of light objects captured by spatial imagery using Single Output Map Algebra (Oseke et al., 2020) In the use of NDVI, the water surface mask can be generated from the DEM and converted into a shoreline via a slicing process followed by reclassification. The reservoir shoreline is subsequently obtained from the classification map by employing the raster to polygon feature, the raster to polyline, and the feature to point operations. Lastly, this point feature class is merged with the fathometer feature class The accuracy of the shoreline obtained is influenced by the digital elevation models spatial resolution, so utilizing high-resolution spatial data will provide a more accurate outcome. Ideally, the spatial data should be taken on a day when the reservoir is at its highest water level. To convert the shoreline into a topographic contour line, a reference water level at the time of Figure 2: Methodology Flow Chart (Modified: form Khalid, 2016) Collection of Echo Sounder Bathymetry elevation data for the determination of ground coordinates as reference elevation data points Exporting the extracted digital elevation data in the form of WGS84 latitude and longitude and elevation (x, y and z) Transformation of GPS latitude and longitude to UTM East and North coordinates for digital elevation model data Generating the elevations for the region area using UTM East and North and elevation coordinates Comparing the reference and extracted elevation data points respectively Statistical accuracy assessment by computing Root Mean Square Error-observation standard deviation ratio, coefficient of variation and coefficient of determination on the actual and generated elevation data points IJEMT DOI: 10.56201/ijemt..pg198.203 data capture is necessary. The shoreline that represents the water surface is extracted from the digital elevation data by using the mask extraction operation, which is a geographical Information program based spatial analysis tool. Through this spatial analysis, a fact can be ascertained between the reservoir elevation and area by utilizing the area and elevation statistics established. This is subsequently followed by a statistical evaluation for the generated spatial data accuracy assessment. Development of Longitudinal and Cross-Sectional Profiles The digital elevation model from the bathymetric data is to be extracted involves the creation of a Triangular Irregular Network . The Triangular Irregular Network is a digital representation of the underwater surface topography of the reservoir and is used to create the longitudinal and cross-sectional profiles of the reservoir bed as captured for further analysis to establish the bathymetry datasets. The creation of the cross-section profile lines usually extends between both banks of the reservoirs at 20m interval for the first 100m; 50m interval for the next 500m; 100m interval for the next 1000m and 250m interval for the next 14.5km. The Spatial Technique The technique spatial technique for bathymetric data extraction involves the application of aerial multispectral technique to estimate depth of water (Terrain Zonum, 2015). The technique demonstrates that electromagnetic radiation has a beneficial capacity to penetrate through columns of water, a unique advancement in bathymetric data acquisition which provides the foundational knowledge necessary for extracting water depth conceptually (Oseke et al., 2020). In practical application, the spatial technique entails directing light beams through water columns, which results in attenuation (Oseke et al., 2020). In reservoir regions that are submerged, bathymetric features appear as bright and dark areas, representing shallower and deeper water respectively captured by satellites (Nicolas et al., 2017). This occurs because solar radiation, when reflected off the earth, interacts significantly with both the atmosphere and the water column before reaching the satellite sensor (Oseke et al., 2020). While the primary objective is to automatically extract information, particularly water depth from bathymetric mapping, this approach has been progressively developed by numerous researchers, including Nicolas et al. (2017) to accomplish those objectives. Accuracy Assessment There are numerous methods to assess the accuracy of dataset for research such as bathymetry. One approach is to quantify the differences between the surface of the previous dataset and that of the newly measured dataset (Erdogan, 2009). This in most cases often not possible due to the time and monetary resources needed to collect the new dataset. It is in this regard that researchers propagated the need to develop a digital elevation data using interpolation technique as digital elevation data is readily available. Statistical Evaluation The statistical evaluation involves the statistical comparing of the reference dataset and the extracted digital dataset. The statistical evaluation is performed by three statistical indicators namely: statistically in terms of coefficient of variation , Root Mean Square Error- observation standard deviation ratio and coefficient determination (R2) using the following formulas. IJEMT DOI: 10.56201/ijemt..pg198.203 100 ) ( ) ( 1 1 − = = = n i obs i n i sim i obs i Q Q Q COV ( ) − − = = = = n i obs mean obs i n i sim i obs i obs Q Q Q Q STDEV RMSE RSR 1 2 1 2 ) ( ( ) ( ) ( ) ( ) − − − − = = = = n i sim mean sim i n i obs mean obs i n i sim mean sim i obs mean obs i Q Q Q Q Q Q Q Q R 1 1 1 2 Where Qi obs and Qmean obs are the observed values and mean of the observed values respectively, while Qi sim and Qmean sim are the simulated values and mean of the simulated values. The coefficient of variation is used to determine how well the model predicts the average magnitudes for the output response of interest for continuous long-term modeling. The coefficient determination is used to estimate the performance factors (fit-to-observations, equation). The RSR is computed as the ratio of the root mean square error and standard deviation of measured data, and ranges between 0 and 1 a large positive number. The lower the RSR value is, the better the extracted simulates the performance. The statistical range of the computed accuracy assessment parameters based on Ghoraba, (2015) and Heyman et al, (2007) is presented in Table 1. Table 1: Statistical rating indicators recommended for model analysis Accuracy Rating COV (%) RSR R2 Very good 75<COV <100 0.0 < RSR < 0.50 R2> 0.70 Good 65<COV <75 0.5 < RSR < 0.60 0.60< R2< 0.70 Satisfactory 50<COV <0.65 0.6 < RSR < 0.70 0.50< R2<0.60 Unsatisfactory COV<0.50 RSR > 0.70 0.00< R2<0.50 Results and Discussions Preliminary Dataset and Performance Indicators The preliminary data extracted using the digital model produced a maximum elevation of 626 m - amsl, while the minimum elevation was 574m - amsl (see Figure 3), indicating an average elevation of 600m-amsl above mean sea level as shown in Figure 3. (1) (2) (3) IJEMT DOI: 10.56201/ijemt..pg198.203 Figure 3: The elevation dataset extracted from Digital Elevation Model and selected Transverse Lines for Cross Sectional Profiling Accordingly, the preliminary findings reveals that the digital dataset extracted from the digital model can be used as a quality control assessment tool for assessing the deviation of measured dataset. The summarized statistics from the preliminary results of the Echo sounder measured elevations dataset and the reference dataset (NWRI 2017) with their corresponding volumes datasets for this study is shown in Table 2. Table 2: Reference data points elevation accuracy for the case study area Measured Elevation Dataset Performance Indicators Digital Elevation Dataset Performance Indicators Minimum Dataset Maximum Dataset Minimum Dataset Maximum Dataset Elevation Dataset (m- amsl) 581 630 574 626 Note: m - amsl – meters - above mean sea level The preliminary findings established that minimum dataset between the Echo Sounder and the digital elevation dataset shows an accuracy of 98.7% representing an error of 1.30%. This reveals there is a significant correlation between the Echo Sounder measured elevation data Line 15 Line 29 Line 32 IJEMT DOI: 10.56201/ijemt..pg198.203 obtained from the National Water Resources Institute (NWRI, 20217) and the extracted digital elevation dataset from the same year. Topographic and Cross-Sectional Profiles The topographic profile showing the cross-sectional of the measured dataset and the digital dataset is shown in Figure 4. The profile is an indication of a temporal sequence of activities as to how much the reservoir's characteristics have changed over time (Syvitski et al., 2005), and enables the determination of the terrain area at certain elevations. Further revealed in Figure 4 on the cross-sectional profile is that the mid-section on both flange of the datasets, is having a higher elevation than the right and the right flanges. Though the digital cross-sectional profile reveals no major deviations from the Echo Sounder measured dataset, these results support the fact digital elevation models can be used to extracted elevations datasets for bathymetric survey (Heyman et al., 2007). 630 620 610 600 590 580 570 560 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00 Distance (m) from right Measured Bathymetric Dataset DEM Extracted Bathymetric Dataset Figure 4: S e l e c t e d cross sectional profile comparison of of dataset at transverse Line 15 IJEMT DOI: 10.56201/ijemt..pg198.203 630 620 610 600 590 580 570 560 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00 Distance (m) from right bank Measured Bathymetric Dataset DEM Extracted Bathymetric Dataset Figure 5: Selected cross sectional profile comparison of dataset at transverse Line 29 630 620 610 600 590 580 570 560 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00 Distance (m) from right bank Measured Bathymetric Dataset DEM Extracted Bathymetric Dataset Figure 6: s e l e c t e d cross sectional profile comparison of dataset at transverse Line 32 IJEMT DOI: 10.56201/ijemt..pg198.203 Performance Indicators for Accuracy Assessment The obtained indicators for assessing the performance accuracy for the overall surveying operations presented in Table 3, as the accuracy is highly dependent on the surveying application, data, technique and the expected resulted production. Table 3: Statistical variables for accuracy assessment for the case study regions Variables Digital Elevation / Echo Sounder Dataset Number of Considered Points 11 Maximum Elevation Difference (m- amsl) 4 Minimum Elevation Difference (m-amsl) 7 Standard Deviation 440.74 Coefficient of Variation (CV %) 8.57 Coefficient of Determination (R2) 0.77 Root Mean Square Error-observation Standard Deviation Ratio 0.50 From results of Table 3, it can be note that: 1. The Echo Sounder dataset and the digital dataset produced a Standard Deviation of 440.74 as the result for estimating Coefficient of Variation (CV %) of 8.57 for elevation. 2. Accordingly, the output from the performance indicators for accuracy assessment shows the measured dataset for elevation captures the digital elevation dataset reasonably. With R2 being more than 0.77, it shows a very good correlation between the measured elevation and the digital elevation and least error variance between the two datasets. The dependability of the model digitally generated suitability for bathymetric survey application was additionally established by RSR values less than 0.53. Conclusion Digital elevation model is an easy tool that enables the earth’s surface mapping using remote sensed imagery. The digital dataset extracted from the digital elevation model helped in the determination reservoir elevation and topographic datasets. The proposed method of this paper is simple for generating digital bathymetric datasets from Digital elevation model. The obtained accuracies for the extracted digital dataset from the digital elevation model is suitable for some engineering application but inadequate to meet the standard required for actual bathymetry requiring accurate engineering precision engineering. However, digital elevation model dataset can be used for investigation and preliminary studies with low cost. It is strongly concluded that the users of digital elevation model have to test the accuracy of elevation data by comparing with reference data before using it. IJEMT DOI: 10.56201/ijemt..pg198.203 Reference Oseke I.F, Anornu, G.K, Adjei. A.K and Eduvie O.M, (2020): Development of water surface area–storage capacity relationship using empirical model for Gurara reservoir, Nigeria, Modeling Earth Systems and Environment, Vol: 7, pp 2047-2058 Miller, F. P., Vandome, A. F., McBrewster, J. (2010): Bathymetry. VDM Publishing House Ltd., 2010 –68 p. ISBN: 6130704542. Obregon, O., Chilton, R.E., Williams, G.P., Nelson, E.J., and Miller, J.B. (2011): Assessing Climate Change Effects in Tropical and Temperate Reservoirs by Modeling Water Quality Scenarios, Proceedings of the 2011 World Environmental and Water Resources Congress, paper 407, Palm Springs, USA, ISBN: 978‐0‐7844‐1173‐5. Nicolas A, Amaury T, Marc F.M, and Hua Z, (2017): Monitoring small reservoirs storage from satellite remote sensing in inaccessible areas, Hydrol. Earth Syst. Sci. https://doi.org/10.5194/hess-2017-373 Federal Ministry of Water Resources (FMWR, 2002): Gurara Power Plant Project: Feasibility study. Abuja: COB/DEC JV. 2002 Terrain Zonum (2015): Retrieved on October 12, 2015, Available at http://www.zonums.com/gmaps/terrain.php (Retrieved on 15th June 2025) Khalid L.A. and El-Ashmawy (2016): Investigation of the Accuracy of Google Earth Elevation Data, Artificial Satellites, Vol. 51, No. 3. Ghoraba, S. M, (2015): Hydrological Modelling of the Simly Dam watershed (Pakistan) using GIS and SWAT model, Alexandria Engineering Journal. Vol: 54(3): pp 583-594. Syvitski J.P., Vörösmarty A.J., Kettner P. (2005): Impact of humans on the flux of terrestrial sediment to the global coastal ocean. Science, Vol 308(5720): pp 376–380 Heyman, W.D., J.-L.B. Ecochard, and F.B. Biasi, (2007): Low-Cost Bathymetric Mapping for Tropical Marine Conservation—A Focus on Reef Fish Spawning Aggregation Sites. Marine Geodesy, Vol 30(1-2): pp. 37-50