Evaluation of the efficiency of Analytic Hierarchy Process (AHP) to specify the areas with urban flood risk potential, case study: the central part of Torbat Heydarieh

Introduction
Runoff velocity is high in urban basins with smooth man-made channel and impermeable areas. As residential zones and various commercial and industrial activities are concentrated in urban areas, the potential of the flood losses are increasing in the cities. The first step in flood management is prepare flood risk zoning maps to show the probability of flooding in agricultural, residential, commercial and industrial land parcels. Multi-criteria evaluation methods are widely used in environmental science. One of these methods is Analytic Hierarchy Process (AHP) often used by many of researches to provide the flood risk mapping. In these studies, we have used several parameters to flood zoning such as distance from channels, depth of ground water, slope, population density, drainage density, runoff coefficient, land use, building and unbinding area, and etc. Several floods were happened in Torbat Heydariyeh in last year. The catastrophes caused damage and losses to local communities. Thus, it is necessity to study the flooding in this city. The aim of this research is to provide the flood risk mapping for the central part of Torbat Heydarieh by AHP method based on field observation.
 
Materials and methods
Torbat Heydarieh County with semi arid climate lies in northeast part of Iran, in58˚ 41' and 60˚ 07' E and 34˚ 59' to 35˚ 51' N. Mean Annual precipitation and temperature of Torbath Heydariyeh are 274 mm and 13.7˚C, respectively. The central part of the city has flooding problem, so the area with 352.3 ha was selected as study area.  In this research, the Analytical Hierarchy Process was used as one of the most comprehensive methods of multi criteria decision making. This process is considered as various options in decision making and it is able to do sensitive analytical method on criteria and sub criteria. The values of credit decisions are related to quantity and quality of questionnaire. In this research, we have identified the 7 parameters including elevation, slope, drainage density, land use, distance to channel, capacity of drainage channel and the area covered by each channel. The questionnaires have been completed by 14 experts. The matrix of binary comparison criteria was calculated by geometric average. To estimate consistency and the weight of criteria and sub-criteria, the final matrixes of binary comparison   were entered into Expert choice software. The weight of criteria and sub-criteria have been calculated and added to each map in GIS. The zoning map of flood risk was drawn with four classes: low risk, moderate risk, high risk and very high risk. To compare the results of AHP and actual situation of flooding, the depth of flood were measured in the index points in three shower events of 4/3/2014, 11/5/2014 and 1/10/2015. Then, to identify the accuracy of AHP flood risk zoning, the percentage of the adaptation between measured point and flood map risk was calculated in each shower.  
 
Results and discussion
The results of pairwise comparison between study criteria showed the classes including the slope lower than 2%, lower than 40 meters distance from main channel, lower than 1350 meter elevation, commercial and ministerial land use, drainage density upper than 325 m/ha, channel capacity lower than 0.5 m³/s and the area covered  by each channel; the weights of the factors are 0.599, 0.531, 0.379, 0.378, 0.382, 0.615 and 0.571 respectively. The map of flood risk zoning showed 126.79 ha, 118.30 ha, 78.62 ha and 28.42 ha of the study area in low, moderate, high and very high risk classes, respectively. Most of the measured flooded points were placed in moderate and high risk classes. The results of sensitive analysis showed distance from main channel as sensitive factor.
 
Conclusion
The results of this research indicated combination of GIS and AHP as effective tools to determine the areas with potentiality of flood in urban basins. The evaluation of potentiality of urban flood risk by AHP represented 67.75% of measured points of flood in high and moderate risk classes when rain intensity was increased. Thus, the results of urban flood risk zoning with AHP was acceptable. In central part of Torbat Heydarieh, the most important factor related to flooding was distance from main channel. In potential map of flood risk, the regions with very high and high risk were mostly near the main channels.