Quantitative Analysis of Slope Instability in Under-Construction Highway of Tehran-North (Tehran-Soleghan) by the Use of Logistic Regression

Slope instabilities are one of the most distinctive types of geomorphic hazards that are exacerbated by human interference and threaten most of the human installations, especially mountainous highways and impose heavy costs on the government and local residents. Each year, slope instabilities cause enormous economic damages to highway, railways, power transmission and communication lines, irrigation and watering canals, ore extraction, as well as oil and gas refining installations, infrastructures in cities, factories and industrial centers, dams, artificial and natural lakes, forests, pastures and natural resources, farms, residential areas and villages or threaten them. Nowadays, many instabilities are resulted by human intervention and manipulations. One of the effective human factors in instability occurrence is the construction of highway. Highway construction, especially in mountainous areas, increases the probability of occurrence of various types of instabilities, as it changes the natural balance of the slopes and causes deformations in the land. Each year, lots of casualties and financial losses are imposed by the occurrence of various types of instabilities in the slopes overlooking the highways, which also cause the destruction of many natural resources in the country. However, the construction of roads, highways and freeways is necessary and unavoidable in today’s life.The Tehran-North highway will be the route that connect the Iran’s capital Tehran with the southern shores of the Caspian Sea.

This contribution aims to study slope instabilities along this highway using logistic regression method. In this regard, layers of 14 effective factors were identified, comprised of elevation classes, slope, aspect, geology, land use, precipitation, distance from fault, river and highway, normalized difference vegetation index (NDVI), climate, slope length (LS), stream power index (SPI) and topographic wetness index (TWI). Consequently, maps of the factors responsible for instabilities were prepared as separate layers in the GIS environment and transferred into the Idrisi software. The whole procedure included: (1) preparation of digital elevation model (DEM), river and fault layers based on the 1:25,000 topographic map of the area, as well as distance maps from rivers and faults, (2) creating slope and aspect maps from DEM, (3) preparation of land use and NDVI maps of the region based on unmatched classification of Landsat 8 image of OLI sensor, (4) preparation of geological map, (5) preparation of precipitation and climate layers based on the information obtained from the meteorological organization, (6) creating LS, SPI and TWI layers based on the DEM, (7) conversion of the distribution data of the regional instabilities using Landsat satellite and Google Earth images, (8) correlating the information layers with the regional instability map and calculating their density per unit area, and (9) performing the logistic regression model using Terrset software.

Results obtained by applying logistic regression model showed that the most important factors affecting slope instabilities in the Tehran-Soleghan highway area are distance from fault and climate. 27.14% of the Soleghan highway area possesses medium to high potential for instabilities, within which 86.26% of the instabilities have occurred. Furthermore, 4.57% of the Soleghan highway area shows very high risk in terms of instability occurrence, encompassing 61% of the occurred instabilities. According to the prepared maps, the middle and southern parts and a small section in the north of the Tehran-Soleghan highway area have the highest potential for instability occurrence. The high value of the ROC index and its proximity to the end value of 1 indicates that instabilities strongly correlate with the probability values derived from the logistic regression model. Additionally, the assessment of the instability potential map by the SCAI index showed that there is a high correlation between the prepared risk maps and the occurred instabilities, which have been confirmed by field surveys. The obtained results are in a good agreement with the general opinion that SCAI decreases especially in high and very high risk classes and indicates a high correlation between the prepared risk maps and the occurred instabilities and field surveys in study areas.

Finally, it can be mentioned that the logistic regression model is suitable for preparing the zonation of the probability of instability occurrence along the edges of the studied highway. As a final conclusion, it can be concluded that in addition to natural factors, the- human-made factors and particularly unsystematic highway  construction can play an important role in the instability occurrences on the slopes overlooking the highway  and in order to reduce the relative risks and increase the stability of the slopes, it is necessary to avoid manipulating the ecosystem and changing the current land use as much as possible, in addition to policy making for constructions in accordance with geomorphological and geological features of the area.