The Effects of Land Use Changes on the Maximum Flood Discharge in the Songhor Watershed

At present, land use changes in a basin over time affect many processes of soil erosion and sediment production and cause loss of soil quality and fertility (Ildoromi et al. 2017). Studies have shown that land use change and its effect on hydrological processes can play an effective role in managing water resources and floods and reducing the damage caused by it, which has been considered by many researchers in recent years. Kazemi et al. (2018) in the study of the effect of land use change and drought on runoff in the Central Zagros Basin stated that the most land use change in 25 years in the field of natural resources, especially in rangelands and increased runoff and floods has been in the basin. Foroutan et al. (2019) investigated the effect of land use change and physical development of the city on changes in Asadabad urban flood runoff. The results showed that with the increase of urban land use area, surface runoff has increased by 350 m3, which is a considerable and sometimes dangerous volume of runoff in a small city. Gomindoga et al. (2015) investigated the effect of land use change in Ethiopia's Jilljal Abai watershed and concluded that reduced rangeland and forest land use increased the maximum flood. The present study aimed to investigate the effects of land use change on maximum flood discharge in the Songhor watershed using the Win TR-55 model and Landsat 7 and 8 satellite images for the years 2000-2015 using ENVI software.

The Songhor watershed in Kermanshah province is a part of the Karkheh watershed with an area of 63.17 Km2 and has a cold semi-humid climate. In this study, the main variables were 24-hour rainfall, concentration-time, flood coefficient, basin area, and slope and land use change area. To study and prepare land use change maps during two periods, ETM and OLI sensor images of Landsat satellite from 2000-2015 were used. After performing geometric correction and band compositions with the help of ENVI software, the adjusted plant difference index (NDVI) was prepared. To classify the images, the classification method was supervised and the maximum probability, kappa index, and general accuracy were used for the correct evaluation. In the next stage, land use changes were classified into five land use classes including irrigated agriculture and gardening, rainfed agriculture, rangeland and forest, residential areas, impenetrable and rocky lands. Finally, land use maps of the Songhor watershed in two time periods of 2000 and 2015 were drawn in ArcGIS. In the next step, the maximum flood hydrograph under the influence of land use change was estimated using the Win TR-55 model for the years 2000-2015. To analyze the peak flow sensitivity of the basin, the values of the canal slope and to evaluate the results by WinTR-55 model, two statistics of correlation coefficient (R2) and Root Mean Square Error (RMSE) were used.

The study of hydrographs showed that in all sub-basins, according to the amount and intensity of land use change, the peak discharge increased in 2015 compared to 2000. This amount decreased in sub-basin S1 due to fewer land use changes and concrete-mortar watershed structures in 2015. It showed a decrease of 4.11% compared to 2000 and indicates that the structures have reduced the speed and volume of discharge from the basin. In other sub-basins, due to the slope and topographic condition, the amount and intensity of land use changes, especially the sharp decrease in rangeland lands and the existence of residential and impenetrable areas, peak discharge values increased. The highest increase in discharge occurred in the S6 sub-basin with 24.88% and the S8 sub-basin with 22.44% in 2015 compared to 2000. Examination of the basin outlet hydrograph showed that the peak discharge rate in 2015 increased by 12.2% in the total basin. One of the effective factors is land use changes and subsequent changes in the basin CN values during the 15 years under study. The results of the sensitivity analysis of the WinTR-55 model showed that the CN parameter is very important and causes high sensitivity. The results of verification and calculation of the estimated error of the model in the return period of 2 to 100 years showed that the percentage of flow estimation error by the model in the 100-year return period has the lowest value.

 Land use changes in the watershed were influenced by factors such as the suitability of rainfall in 2000-2008 and the tendency of some farmers regarding dry farmland to increase the extent of their agricultural land to increase their income. The population of the city has increased by 7.5% from 2005 to 2015 and the development of urbanization has made agricultural areas around the city become residential areas. Droughts in recent years have also destroyed some grassland species, reduced soil moisture, and increased runoff. The velocity and flow rate were the peaks. Output hydrograph survey of the basin showed that the peak discharge in 2015 increased by 12.2% in the whole basin. In general, the effective factors in increasing the discharge in the Songhor watershed are the land use changes followed by changes in the CN values ​​over the 15 years studied. The results of WinTR-55 sensitivity analysis showed that the CN parameter was an important parameter in the model. It showed that the model a has high ability to estimate the maximum flood discharge for this type of condition in the Songhor watershed and indicated the high accuracy and efficiency of the model in investigating hydrological fluctuations