Evaluation of the Morphometric Effect of Landforms on Hydrological Processes of the Basin (Case Study: Northern Slopes of Sabalan)

Geomorphometry is the science of quantitative land-surface analysis (Pike, 1995, 2000a; Raseman et al., 2004). It is an interdisciplinary field that has evolved from mathematics, the Earth sciences, and most recently computer sciences (Pike et al, 2008, 3). It is well to keep in mind the two overarching modes of geomorphometric analysis first distinguished by Evans (1972) as specific, addressing discrete surface features (i.e. Landforms), and general, treating the continuous land surface. The morphometry of landforms per se, with or without the use of digital data, is more correctly considered part of the quantitative geomorphology (Thorn, 1988; Scheidegger, 1991; Leopold et al., 1995; Rhoads &Thorn, 1996). The shape of terrain, i.e. landforms, influences flow of surface water, transport of sediments, and soil production, and determines climate on local and regional scales. Furthermore, natural phenomena like vegetation are directly influenced by landform patterns and their relative position across the landscape (Blaszczynski 1997; Blaschke & Strobl, 2003).
The Earth’s surface is structured into landforms as a result of the cumulative influence of geomorphic, geological, hydrological, ecological, and soil forming processes that have acted on over time. Landforms define boundary conditions for processes operative in the fields of geomorphology, hydrology, ecology, pedology and others (Dikau, 1989; Dikau et al., 1995; Pike, 1995, 2000a; Dehn et al., 2001). In this study, using MRS algorithms and Ecognition software, landforms in the northern slopes of Mount Sabalan have been extracted and the effects of Landform morphometry on its hydrology have been investigated The semi-automated methods refer to the automatic procedures of extracting a landform based-process. This is mainly relying on unsupervised isodata classification, pixel-based classification (supervised /subpixel classifier based on training material), the analysis of digital elevation models (DEM), algorithms, hydrological modelling, and object oriented analysis (Nabil and Moawad, 2014:42).
In this study object-oriented methods and Ecognition software were used for the classification and the extraction of landforms. The object-oriented classification was used as an alternative to traditional pixel-based classifications, to cluster grid cells into homogeneous objects, which can be classified as geomorphological features (Seijmonsbergen, 2012). In addition, the DEM and its derivation (Slope, Profile and plan curvatures, maximum and minimum curvatures), were used in order to extract landforms. Then, using fuzzy logic method, the landform, land use, NDVI index , precipitation, density of river, and lithology layers were Overlaid and the potential flooding area was obtained. In the object-oriented method, determining the scale parameter is a very important factor in the separation of different objects in an image. Scale parameter is a crucial threshold that determines the maximum allowed heterogeneity for segmentation and has a direct influence on the size of the objects to be obtained. The scale parameter, after a trial and error process, is recognized to be within a particular range (Gerçek, 2010:115). A novel method that was introduced by Dragut et al. (2010) and the ‘Estimation of Scale Parameter (ESP) that built on the idea of ‘Local Variance’ (LV) were employed to obtain the optimum scale out of a range of scales. By interpreting thresholds and prominent peaks in the ROC-LV graph, characteristic scales relative to data properties at the scene level could be found. This curve in 100 scale level was produced for the study area by using the ESP software and with respect to curve, the scale of 25 was selected for the segmentation. After segmentation, using the morphometric differences between the landforms, the landforms were extracted. After this stage, the landforms along with three layers of NDVI index, land use, and lithology was fuzzy. Finally, using gamma 0.8, they were combined and the zoning map of the potential flooding was estimated. Flood zoning map was classified into 5 classes and the percentage of each zone risk was calculated in each landform. In this research, using an object-oriented model, landforms were extracted as plain, peak, pit, ridge, channel, nose, shoulder slope, hollow shoulder, spur, planar slope, hollow, spur foot slope, and hollow foot. An assessment of the effect of landforms on the hydrology of the area revealed that three landforms of hollow, shoulder and planar slope which were respectively 67.3%, 62.9%, and 53.2% had the greatest impact on flooding and their area were zoned as high and very high flooding. On the other hand, plain and pit landforms were zoned in the form of low and very low flooding areas.