Assessment of Basin Morphometric Measurements Accuracy Based on Different DEM Data, Chai Dakuk Basin, Northeast Iraq : A Case Study

Abstract


Introduction
Accurate delineation of drainage networks is a prerequisite for many natural resource management issues.The morphometric analysis is essential for evaluating the drainage basins' geomorphic development and hydrologic behavior.The measurement accuracy depends mainly on the data source and the analysis techniques.The availability of new satellite-based topographic datasets has opened new venues for hydrologic and geomorphologic studies, including surface morphology analysis (Reddy et al., 2018).Remote Sensing (RS), Geographical Information System (GIS), and Global Positioning System (GPS) have proved to be efficient tools in the delineation of drainage patterns and the management of water resource and its planning.The watershed morphometry analysis has been performed with manual and computerized delineation, enabling detailed morphometrical measures (Rai et al., 2018).Topography plays a vital role in various natural processes (Das et al., 2016).The topographic and terrain data source is essential in the drainage network's extraction, delineation, management of the watershed, and estimations of contribution area inflow patterns (Ahmed et al., 2010).The automated derivation of topographic watershed data from Digital Elevation Models (DEMs) is faster, less subjective, and provides more reproducible measurements than traditional manual techniques applied to topographic maps (Garbrecht and Martz, 2000).
DEMs are an important form of satellite or remote sensing data used in hydrological, hydraulic, climate change, agricultural management, and water resource development studies.Before introducing DEMs, landforms and drainage networks were only manually identified by surveying data when available through the interpretation of topographic maps and aerial photographs (Ashmawy et al., 2018;Garbrecht and Martz, 2000).The available open-source digital topographic data include the Shuttle Radar Topography Mission (SRTM), Advanced Spaceborne Thermal Emission, and Reflection Radiometer (ASTER) GDEM.The other thematic datasets aided by mathematical tools on different GIS platforms have greatly enhanced the capability of extracting morphometric parameters, which are essential for spatial landscape pattern characterization (Jaiswara et al., 2020).The sources and resolution of DEMs impact the results obtained from hydraulic and hydrologic models, e.g., flood inundation mapping of river channels was affected by DEM sources (Tesema, 2021).
The morphometric parameters of a drainage basin are the quantitative attributes of the landscape derived from the terrain, the elevation surface, and the drainage network (El Bastawesy et al., 2021).It gives an idea about the basin characteristics regarding basin geometry, slope, topography, soil condition, runoff characteristics, and surface water potential (Kiliç et al., 2022).Morphometric analysis is critical in evaluating the water resources and identifying the recharge sites, runoff modeling, soil erosion, flood susceptibility, groundwater mapping, and watershed delineation (Kasi et al., 2020).Morphometric analysis of a drainage basin demonstrates the dynamic equilibrium achieved due to the interaction between matter and energy.This helps to understand the prevailing geohydrological characteristics of the drainage basins (Kolapkar and Thakare, 2018).
The current study reviewed several related studies.Kiliç et al. (2022) used five DEMs data to calculate14 morphometric parameters and concluded the SRTM30m revealed more consistent results with TOPO30m DEM.Kasi et al. (2020) compared four DEMs data and toposheet 1:25000, and the result observed that the CARTOSAT 10 m and 30 m are more accurate than the SRTM DEMs in terms of vertical elevation.However, estimated morphometric parameters were not highly sensitive to the used data sources.Niyazi et al. (2019) showed that the SRTM 30 m is characterized by high accuracy and perfectly matches with Google Earth maps and topographic maps of scale 1:50,000.The study of Das et al. (2016) showed that the DEMs derived from the 1:50000 topographical map and ASTER GDEM datasets are relatively more accurate and consistent than other DEMs of 90 m resolution.Ahmed et al. (2010) referred to the morphometric parameters derived from the SRTM and ASTER DEMs-30m data provide good and satisfying results.Regarding comparative studies, the ArcGISmorphometric toolbox was used in the hypsometric analysis of Al-Adhaim and Tigris rivers subbasins based on SRTM-DEM data, and the observed results satisfied the evolution of geomorphic development of the river basins (Al-Sulttani and Beg, 2020;Beg, 2020).
The current study aims to prepare the quantitative morphometric analysis of the Chai Dakuk River basin using different DEMs data, i.e., ALOS PALSAR (ASF12.5),SRTM 1arc second, ASTER-GDEM 1arc second, SRTM 3arc second, and GMTED-7.5arcsecond, because this basin extends along an area characterized by various geological lithology, structure, soil types and diverse topography, to evaluate their accuracy with the measurements carried out on reference points from topo maps and in-situ GPS measurements.In the next step, the morphometric parameters were calculated using different DEM data and the Morphometric toolbox developed by Beg (2015) in ArcGIS 10.8.1 software to investigate the impact of DEM accuracy on the calculation of morphometric parameters and to overcome the problem of DEM data selection for research purposes.

Description of the Study Area
The Chai Dakuk River is one of the main branches of Al-Adhaim River.The basin extends over an area of about 3481 square kilometers and is bounded between latitudes 34°44'30"-35°35'49" N and longitudes 44°17'39"-45°28'16" E (Fig. 1).The geological setting of the study area (Fig. 2a) belongs mainly to Paleocene-Eocene sequence including Pila Spi Formation the upper part comprises wellbedded, bituminous chalky and crystalline limestone with bands of chalky marl, while the lower part comprises well-bedded bituminous fossiliferous limestone.Gercus Formation consist of Molasse deposited of red and purple shales, mudstones, sandy and marls, pebbly sandstone and conglomerate with gypsum and halite lenses near the top of the formation, while middle Miocene sequence represented by Fatha formation comprises of anhydrite, gypsum and salt interbedded with limestone and marl, whereas, the Late Miocene-Pliocene sequence including Injana Formation comprises finegrained pre-molasse sediments.Mukdadiya Formation characterized by gravelly sandstone and Bai Hassan Formation consists of sandstone beds, and at downstream area consist of flood plain and local deposits (Jassim and Goff, 2006).Intensive lineaments are recognized in the study area (Fig. 2b).Most of the major lineaments are extended in the northwest-southeast direction, which controls the drainage network and the shape of the basin.Regarding soils, many types cover the study area (Fig. 2c), i.e., brown soils and shallow phases over gravel cover most of the basin.In contrast, the upstream area is covered by calcareous rock outcrops roughly broken-rocky land and reddish-brown soils, medium and shallow phases are dominant in the area above nick point of the main channel, gilled land, sandy gravel, reddish-brown soil, which are distinct at the valley of the river in the downstream area of the basin (Buringh, 1960).

Building of Geodatabase
Building a geodatabase is the first step in the spatial analysis project using ArcGIS software.In the case of basin morphometric analysis, the geodatabase is built as follows (Fig. 4): - • Convert all the used DEM data from the GIS to the projected coordinate system (UTMWGS84zone XX) or any suitable projected coordinate system; in the current study, UTM-WGS84-zone-38 was used.• Extract the drainage network and watershed using the ArcGIS-Hydrology Toolbox.
• Identify the main channel route using ArcGIS-Network analyst tools-new route.
• Calculate basin length using ArcGIS-measure icon.
• Add the morphometric toolbox from the folder to ArcGIS tools.
• Execute the morphometric script and follow the required inputs from the drop-down list in the pane, and insert the main channel and basin lengths as given in Table 1.• Execute the Hypsometric script to do the hypsometric analysis.

Collection of Reference Points
Elevation measurements are collected from 133 reference points randomly located along the basin area to assess the accuracy of DEM data.The reference points are collected from benchmark elevations fixed on topo maps of scale 1: 20000 (National Imagery and Mapping Agency, USA, 2003, sheets of Qara Dagh, Kirkuk, Kallar, Dokan, and Sulaymaniyah), and some of the points are collected using differential GPS (Fig. 3f).These data are used to evaluate the elevation accuracy of the DEMs data in a proactive step of calculating morphometric variables from different DEMs data.

Assessment of Relative Errors
Assessment of the relative errors between ground truth measurements and DEMs data was carried out using root mean squared error (RMSE), Mean Absolute Error (MAE), and Relative Absolute Error (RAE), as shown in equations 1,2 and 3 (Hamner et al., 2018). (1) (2) Where Zob.: Represent observed elevation, ZDEM: is the DEM elevation, Z ̅ Ob. is the average observed elevation, and n: is the number of reference points.

Calculation of Morphometric Parameters
Morphometric analysis was carried out according to equations summarized in Table 2 and using an ArcGIS-Morphometric toolbox developed by (Beg, 2015) and published on the ESRI website.

Results and Discussion
The drainage morphometric analysis was done based on five DEMs data to assess the hydrological characteristics of the Chai Dakuk watershed.The DEMs data are tested for their accuracies based on reference points and then used to calculate the basin morphometric parameters as discussed in the following paragraphs: -

Elevation Accuracy of DEMs Data
The comparison results of the DEMs data with reference points are summarized in Table 3. Noticeable variations in the values of standard deviation (SD.), RMSE, MAE, and RAE are observed.The values of SD range from 5.18 for SRTM-1arcsec to 23.14 for GMTED7.5 arcsec.RMSE values varied from 5.65 to 23.35, MAE values varied from 4.02 to 14,72, and RAE values ranged between 0.019 to 0.08.Based on that, the SRTM-1 arc-second data are identified with the least values of standard deviation, RMSE, MAE, and RAE that indicating the SRTM-1 arc-second shows a considerable matching with reference points followed by ALOS and ASTER 1 arcsec.The essential concluded indication is that SRTM-1 arc-second ALOs and ASTER are more favorable for measuring the morphometric parameters of basins, and the achieved result was consistent with previous studies in this field (Ahmed et al., 2010;Niyazi et al., 2019;Rai et al., 2018).

Morphometric Parameters
The morphometric parameters calculated from different DEMs data show clear variation in drainage patterns, densities, and many morphometric parameters.

Drainage network parameters
According to the Strahler stream order, the drainage networks of the basin (Figs.5a-e) show wide variation in stream orders.The basin was classified into nine orders from ALOS DEM, eight from SRTM 1arcsec, seven from ASTER, six from SRTM3-arcsec, and five from GMTED data.The drainage networks analysis results, i.e., numbers, lengths, bifurcation ratios, and main channel length, are given in Table 4 and Figs.6a-d.The result shows distinct variations in their values; regarding stream numbers, the minimum stream number is 257 extracted from GMTED, and the maximum number is 71796 from ALOS data with a standard deviation of 29490.43.Whereas the total minimum length of streams is 1239.8km, and the maximum length is 19038.13km, with a standard deviation of 6936.16 km, which indicates the role of pixel size in the extraction of the basin drainage lines.While Bifurcation ratio (Rb) is not affected by the same degree of number and length of streams which shows minimum Rb is 3.81 and maximum Rb is 4.76 with a standard deviation of 0.38 that is because all the drainage lines are affected by the same ratio when pixel size of DEM data was changed, that means most of DEMs data with resolution range between 30 -90m is considered to be suitable for calculation of bifurcation ratios.

Geometric parameters
The most geometric parameters which depend on DEMs resolution are the basin area, perimeter, main channel length, and compactness coefficient (Table 5, Figs.7a-d).The variation in pixel size will affect the delineation of river or valley watershed and consequently affect perimeter length and basin area and all parameters calculated as a function of them like compactness coefficient and Relative perimeter.

Texture parameters
Several measured drainage texture parameters are exposed to considerable change with DEM resolution variation.The reason for that variation is because all the texture parameters are a function of (Basin area, perimeters, drainage lengths, and drainage numbers) which are already affected by the pixel size of DEMs data.The affected parameters are drainage density, stream frequency, constant channel maintenance, drainage texture and infiltration number, drainage intensity, and the average length of overland flow (Table 6, Figs.8a-g).

Relief parameters
Total basin relief (H) is the difference between basin mouth and summit (Strahler, 1957).The height of the basin outlet and maximum height of basin values are demonstrated in the total basin relief value, which varied from 1556 to 1623 m with an SD value of 25.29.This indicates that variation of pixel size affects the elevation values, and the elevation error will increase directly with pixel size.Regarding the hypsometric and volumetric analysis of the basin, which is the relation of horizontal cross-sectional drainage basin area, to evaluate and determine how the mass is distributed within a basin from base to top (Schumm, 1956;Strahler, 1952;Strahler, 1957).While Surf_Hypsometric Integral is calculated as the ratio of the surface area of the horizontal cross-section to elevation, volumetric analysis means the ratio between the volume of basin mass rocks above the horizontal cross-section to the total volume of mass above the horizontal cross-section passing through basin outlet (Beg, 2015).Hypsometric, surf-hypsometric, volumetric curves and integral values are highly correlated (Fig. 9a-e), and the SD values range from 0.31 to 0.67 (Table 7), which means all the tested DEMs data are suitable for the evaluation of the geomorphic development of a basin using hypsometric analysis.Longitudinal profiles of the main basin channel are important in geomorphometric and tectonic geomorphic studies (Das et al., 2016).In the current study, all the profiles extracted from different

Fig. 2 .
Fig. 2. Physical characteristics of Chai Dakuk River basin, a)Geological setting of the basin, b) Lineaments map and c) Soil types map

Fig. 6 .
Fig. 6.Differences in linear parameters, a) Numbers of stream orders, b) Total lengths of Streams and c)Values of bifurcation ratios

Table 1 .
Lengths of the main channel were measured from different DEM data.

Table 3 .
Calculated error values of elevation differences between DEMs data and references points

Table 4 .
Values of drainage network linear parameters of Chai Dakuk River basin

Table 5 .
Values of geometry parameters of Chai Dakuk river basin

Table 6 .
Drainage texture parameters of Chai Dakuk River basin

Table 7 .
Basin relief parameters of Chai Dakuk river basin