Journal of Spatial Analysis

---

---

 Relationship between hydrogeomorphic features and suspended sediment load under Kashfarud basins
 
Introduction
As a stressful stimulus, river sediment is the most significant threat to aquatic ecosystems. To prevent or minimize the damage, three stages of the erosion process should be investigated (Naseri et al., 2019: 83). Determining the amount of sediment transported by rivers is important from different aspects. Sediment carried by water flows is considered a factor effective in shaping the geometric structure and geomorphic characteristics of rivers (Tashekabood et al., 2019: 282).
Data and methodology
To estimate the amount of annual suspended sediments, the flow and sediment statistics of hydrometric stations (8 stations) and meteorological stations (13 stations) were employed (Figure 2). The research statistical period is 25 years (1993-2017). The altitude, area, and perimeter of the basins were obtained from topographic maps with a scale of 1.25000. To investigate the correlation between independent and dependent variables, the normality tests of Shapiro-Wilk and Kolmogorov-Smirnov were performed in SPSS16 software. To extract the geomorphic features of the basins, the digital elevation model was used. Then, ground surface corrections and pretreatments such as removal of hydrological pits were performed and ground drainage pattern was determined.
Stepwise multivariate regression
In the present study, stepwise multivariate regression was used to reduce the number of independent variables and determine the effective factors in the sedimentation of the basin. This method investigates the effect of several independent variables on a dependent variable (Zare Chahuki: 2010). In stepwise multivariate regression, the independent variable that has no more significant effect on the dependent variable is removed from the analysis, hence excluded from the equation. The general form of the stepwise regression equation is:
Equation 1                                                                            Y= a + B1X1 + B2X2 + …… + BnXn + e
Data description and interpretation
The principal component analysis method was used to determine the most effective characteristics of sediments as well as their grouping. In principal component analysis, variables that have a high correlation and are distributed in a multidimensional space are reduced to a set of non-correlated components, each of which is a linear combination of the main variables. The obtained non-correlated components are called principal components (PCs). Prior to component analysis, the KMO coefficient was used to ensure the appropriateness of the data for principal component analysis. This coefficient fluctuates in the range of zero and one and if its value is less than 0.5, the data will not be suitable for principal component analysis and if the values of this coefficient are between 0.5-0.69, The proportionality of the data is moderate and if the value of this coefficient is more than 0.7, the data will be quite suitable for performing principal component analysis.
Regression analysis results
In this study, the sediment weight of the basin was considered as a dependent variable and other parameters as independent variables. The variables of slope, precipitation, basin length, Elongation Ratio (R), circularity coefficient, and unevenness of the basin have a higher correlation with the amount of sediment production in the basin than other variables.
An eigenvalue was used to determine the number of factors. The minimum eigenvalue for the selection of final factors is 1, and factors with an eigenvalue bigger than 1 are considered final factors. The results showed that the three factors of circularity coefficient, compactness coefficient, and basin form coefficient have an eigenvalue bigger than 1.
Conclusion
The results showed that geomorphic parameters have a high correlation with the amount of annual sediment. The results showed that seven factors of slope, precipitation, basin length, elongation ratio, circularity coefficient, unevenness coefficient, and form ratio of the basin were the most important in estimating the amount of suspended sediment based on the principal components analysis method. The average of special sediment varies from 134 tons per year in Dehbar basin to 16 tons per year in Kardeh basin and also the average annual sediment varies from 261.6 tons per year in Golmakan basin to 156.7 tons per year in Shandiz basin. Evaluation of Bartlett's test of sphericity tests and KMO values is 0.9. Therefore, the data is suitable for factor analysis. The percentage of variance explained by each factor indicates that the circularity coefficient with 50.71% of the variance explains all the research variables. In total, three factors of circularity coefficient, compactness coefficient, and form ratio of the basin could explain 82.6% of the variance of all research variables. Therefore, the results are consistent with Lu et al. (1991), Sarangi et al. (2005), Tamene et al. (2006), Zhang et al. (2015), Salim (2014), and Ares et al. (2016).
Khorram Dareh sub-basin with heavy rainfall (504 mm) has the lowest specific sediment, which is due to the geological structure of the region. Based on the calculated indicators, most of the studied sub-basins are elongated. The form ratio of the basin is less indicative of the elongation of the basin. The highest branching ratio of the basins is in the vicinity of faults. Also, high circularity values indicate points prone to sedimentation. River sections up to degree 3 are located in more subdued areas and have a steeper slope. Golmakan, Khorram Darreh, Zashk, and Dehbar sub-basins have a high potential for sedimentation. Regression equations of sediment measurement curves are usually used in sediment load estimates. The most important reason is the ease of application of these equations. According to the research results, it can be concluded that the integrated use of principal component analysis, cluster analysis, and multivariate stepwise regression has a suitable and acceptable efficiency in estimating suspended sediments. Testing the regression model concerning different climatic and hydrological regimes of Iran’s watersheds to achieve an efficient pattern of using these equations can be fruitful in estimating sediment load in different regions.
 
Keywords: Hydrogeomorphic, Sediment erosion, Kashfarud basin, Stepwise multivariate regression

---

Determination of flood-prone areas using Sentinel-1 Radar images
(Case study: Flood on March 2019, Kashkan River, Lorestan Province)

Introduction
Although natural hazards occur in all parts of the world, their incidence is higher in Asia than in any other part of the world. Natural phenomena are considered as natural hazards when they cause damage or financial losses to human beings. Iran is also one of the high-risk countries in terms of floods. Until 2002, about 467 floods have been recorded by the country's hydrometric stations. In addition to natural factors such as rainfall, researchers consider human impacts such as destruction of vegetation cover, soil destruction, inefficient management, destruction of pastures and forests, and encroachment on the river are the most important factors for the occurrence and damage of floods in the country. One of the most efficient and emerging tools in flood surveys is the use of radar images. SAR images and flood maps produced by radar images provide researchers valuable and reliable information. Moreover, maps obtained from SAR images help officials to manage the crisis and take preventive measures against floods. The Sentinel-1 satellite is part of the Copernicus program, launched by the European Space Agency, and is widely used in mapping flood-prone areas. The contribution of Sentinel-1 to the application of flood mapping arises from the sensitivity of the backscatter signal to open water. This study aims to determine high-risk and flood-prone areas along the Kashkan River using Sentinel-1 radar images.
Data and Methods
 The study area includes a part of the Kashkan river from Mamolan city to the connection point of this river to Seymareh river, after Pol-dokhtar city. The average annual discharge of the Kashkan river is 33.2 cubic meters per second based on the data of the Pole-Kashkan Station. The length of the river in the study area is about 100 km. To investigate flood-prone areas, we applied pre-processing and image-processing steps to each flood event including SAR images belonging to March 25^th, 2019, March 31^st 2019, and April 2^nd, 2019. SAR images were acquired from ESA Copernicus Open Access Hub. climatic data was downloaded from power.larc.nasa.gov. To create meander cross-sections, the Digital Elevation Model of the studied area was utilized. Cross-sections were created using QGIS software. Pre-processing steps include: applying orbit data, removing SAR thermal noise, calibration of SAR images, de-speckling and topographic correction. In image processing, we applied the Otsu thresholding method to distinguish water pixels from land pixels. In thresholding methods, the histogram of each image is divided into two parts according to the amount of gray composition. The higher the amount of gray (i.e., the pixel tends to be darker), the more pixels represent water, and conversely, the lighter-toned pixels (i.e., pixels that tend to whiten) represent land. The Otsu thresholding method is a commonly used method for water detection in SAR images. It uses an image histogram to determine the correct threshold. The most important feature of the Otsu method is that it is capable of determining the threshold automatically. The Otsu algorithm was applied to all images using MATLAB.
Results
According to the flood maps, on March 25th, 6.51 percent of the study area was flooded, while on March 31th, only 3.96 percent was flooded. This is mainly due to less precipitation on the 31st. On March 25^th the average daily precipitation was 47.46 mm while on 31^st of March the average daily precipitation was 31.64 mm. On April 2nd, however, there was no rainfall, on the day before more than 63 mm of precipitation has occurred. This massive amount of precipitation on the previous day has led to more than 25km2 being flooded in the studied area.
Conclusion
Results showed that meanders and their surrounding areas are the most dangerous sections in terms of flooding. The meander's dynamic and the river's hydrologic processes are essential factors affecting flooding in those sections. Generally, various factors affect flooding and the damage caused by it. This study aimed to determine flooded and flood-prone areas (according to flooded areas in previous events) using new methods in a short time and with high accuracy to use this tool for more accurate zoning and efficient planning in the future. The results showed that radar images are practical, robust, and reliable tools for determining flooded areas, especially for rapid and near-real-time studies of flood events.
Keywords: Floods, Radar images, Sentinel-1Satelitte, Kashkan river