Measurement and Prediction of the Potential Occurrence of Floods Under Climate Change Conditions (Case study: Gorgānroud watershed)

Flood management is a comprehensive approach that includes multiple evaluation indicators in the river basin and an effective and stable but complex method. Flood risk assessment provides valuable information for flood management in order to assess vulnerability and exposure to risk. The method of the current research, according to the nature of the problem and the subject under investigation, is descriptive-analytical and applied studies with emphasis on quantitative methods. In the current research, the regional changes in flood occurrences in the Gorgānroud watershed were investigated using information from meteorological stations (synoptic data) over a 30-year period (1999 to 2018), land use, plant cover, topographic moisture index, slope, elevation, lithology, distance from the river, river density, erosion, soil analysis, runoff, and simulated data of mean precipitation obtained from the HadCM3 model in LARS-WG under the SRA1B scenario for the years 2011 to 2045. In this study, two different approaches were employed. In the first approach, the LARS-WG model was used for downscaling to predict the future climate conditions (near and distant future), while in the second approach, the SWAT hydrological model was utilized to assess flood risk. Based on the percentage of potential hazards in the Gorgānroud watershed, zoning was conducted using the SWAT and GIS software in order to delineate the areas at risk. In this study, based on the SWAT model, the daily precipitation and daily temperature of synoptic stations were calibrated. The results have shown that climate change and the alteration of natural environmental conditions in the region have led to various consequences and impacts, including changes in precipitation patterns, the emergence of heterogeneity in historical data series, fluctuations in river water levels, reduction in agricultural production, changes in plant composition and productivity in grasslands, alterations in groundwater levels, and the occurrence of social and economic problems. Some behaviors in the time series of climatic elements are sudden, intense, and unstable, appearing unexpectedly. Physiographies factors such as slope, soil texture, land use, and rock permeability have caused different hydrological responses to precipitation events in various watershed areas, which significantly influence the occurrence and characteristics of flash floods.
 
Extended Abstract
Introduction
Flood is considered one of the natural phenomena and highly significant and destructive hazards at a global level, accompanied by human and financial losses every year, in various parts of the world, including Iran. Climate change has consequences and effects on the occurrence of the global warming phenomenon, reduction in agricultural productions, alteration in biodiversity and vegetation coverage of grasslands, changes in groundwater levels, and the emergence of social and economic problems. This research aimed to zoning and model flood-prone and risk areas for a changing climate. The research approach is focused on the climate change perspective and its effects on the hydrology of the Gorgānroud watershed. The key objective was to model and predict the risk of flood occurrence in a regional area under climate change conditions using the fuzzy analysis algorithm, hierarchical analysis, and the SWAT model in the watersheds of Golestān Province. Although numerous studies have been conducted on rainfall phenomena in the Golestān province and Iran, none of them have comprehensively examined the climate variability perspective combined with environmental variables.
Methodology
The Gorgān River Basin, with an area of 10,197 square kilometers, is one of the northeastern basins of the country. A significant portion of this basin is located in Golestān Province. The basin is bordered by the Alborz Mountain range to the south, Ᾱladāgh and Gole-Dāgh Mountains to the east, the Atrak River Basin to the north, and the Caspian Sea and the Qareh-Su Basin to the west. This basin is located within the geographical coordinates of 54 degrees and 10 minutes' east longitude and 36 degrees and 35 minutes at 38 degrees and 15 minutes' north latitude. In the present study, the regional flood changes in the Gorgānroud watershed are assessed based on the utilization of data from meteorological stations (synoptic data) of the Meteorological Organization, using a 30-year statistical period (1989 to 2018), land use, plant cover, topographic moisture index, slope, elevation, geological lithology, distance from the river, river density, erosion, soil analysis, runoff, and simulated data of average precipitation derived from the HadCM3 model in LARS-WG under the SRA1B scenario for the years 2011 to 2045.                                                                                                                            SWAT is an extension under the ArcGIS software that is used for simulating major hydrological processes, including evapotranspiration, surface runoff, deep percolation, groundwater flow, and subsurface flows using the simulator model. The spatial input variables for the SWAT model in this study included data layers such as the Digital Elevation Model (DEM), soil layer with soil texture information, and land use/land cover layer with descriptive information. The hydroclimatology and numerical input variables for the SWAT model included daily precipitation values, minimum and maximum temperature values from stations within and outside the watershed, and average daily runoff values from the watershed outlet along with their geographic locations.     
Results and discussion
By analyzing and examining the obtained results, it can be inferred that climate change phenomenon will gradually have noticeable effects on the study area, leading to changes in temperature and precipitation parameters. This will result in a decrease in winter precipitation and an increase in the temperature of the study area. Based on the analysis of the changes, it is expected that with the increase in scenarios, the runoff will decrease. This trend is observed in RCP4.5, wherein, despite an increase in the extreme precipitation value, the runoff has decreased. In RCP8.5 (with higher extreme precipitation), the runoff has increased. The reason for this change is that the occurrence of extreme precipitation in RCP4.5 was during the summer and early autumn seasons, wherein, despite high evapotranspiration and no snow melting (unlike RCP8.5, which coincides with snow melting in the spring), the runoff is less than the subsequent scenario. Considering the final map, 154 villages with a population of approximately 138,548 individuals are exposed to high flood risk. The largest area corresponds to the category with relatively severe flood susceptibility, equivalent to 2,725 square kilometers. The category with high potential covers 2,607 square kilometers, the category with relatively low potential covers 1,654 square kilometers, the category with very high flood susceptibility covers 901 square kilometers, and finally, the category with low flood susceptibility covers 705 square kilometers of the total area of the Gorgānroud watershed. Furthermore, it can be concluded that more than 85% of the basin area are exposed to the risk of severe flooding.
Conclusion
The results of simulation of stream-flow by SWAT: The results for the SRA1B scenario showed a predicted increase in stream-flow by 1.21% for the month of May. On the other hand, the same scenario predicts a decrease in stream-flow by 2.13% for October. According to the SRA1B scenario, a decrease in stream-flow by 9.2% and 63.38% has been predicted for May and October, respectively. Results of modeling and predicting precipitation under climate change conditions: Climate change scenarios based on the HadCM3 model indicated the highest temperature changes in June, August, and February for different months of the year. This indicates a warming trend in the warmer months of the year and even a warmer February in future periods. Comparing the average monthly temperatures between the baseline period and the simulated data using the HadCM3 model in the region showed that June and August will be accompanied by an increase in temperature in future periods, while temperatures will decrease in April and May, which can be examined from various aspects. Comparing the maximum and minimum temperatures between the baseline period and the simulated data using the HadCM3 model in the Gorgānroud watershed also indicates a warming trend in the warmer months of the year, confirming the discussed subject. The results indicated that both maximum and minimum temperatures at the stations will increase in the future period. Results of precipitation trend analysis using the Kendall method: The obtained results of these time series precipitation plots (1989 to 2018) indicated three different types of changes. From the analysis of the U-U' curves, it can be concluded that only in Aliābad Katoul, the two curves have crossed each other outside the range of ±1.96 and moved in opposite directions, indicating a shift. Results of the interpolation section: By comparing the root mean square error with the mean error in the data obtained from the overall validation in ordinary kriging, simple kriging, and universal kriging, it is expected that the mean square error and root mean square error would be significantly higher. Given that the accuracy evaluation results of the examined methods have been provided, the universal kriging method with the lowest mean standard error of estimation (0.006) and root mean square error (23.73) was considered the best method for interpolation.
Funding
There is no funding support.
Authors’ Contribution  
Authors contributed equally to the conceptualization and writing of the article. All of the authors approved the content of the manuscript and agreed on all aspects of the work.
Conflict of Interest
Authors declared no conflict of interest.
Acknowledgments
We are grateful to all the persons for scientific consulting in this paper.