Measuring and estimation of suspended sediment variation using sediment rating curve approach in Mereg River

As a major component of hydrological, geomorphological, and ecological functioning of rivers, suspended sediment has been identified as the leading direct cause of river instabilities. Therefore, recognizing the variability of rivers is essential to manage water resources and environmental issues. As a result, 5 storms were sampled and analyzed for 4 months in Mereg watershed from 2015 to 2016. Sampling was taken in 2 hours interval and suspended sediment concentration was obtained in the laboratory by filtration method. Estimation of temporal variation of suspended sediment during storm was carried out using sediment rating curves. Different approaches were used to prepare sediment rating, including hysteresis pattern, hydrological and seasonal separation. Measured floods have the peak flows from 1.38 to 49.40 m3/s covering a suitable range of low to high peak flows in this river. Moreover, the peak suspended sediment concentration measured in these storms was from 1 to 15.2 g / l. The average suspended sediment load in the study period was 15,919 tons. Measurement of suspended sediment during winter and spring showed that the generation and transportation of suspended sediment in spring is higher than winter in the studied watershed. The fitted relationships on all rating curves was valid and had a determine coefficient higher than 0.5 and estimated error below 50%. The results show that rating curve can provide acceptable estimate of suspended sediment load in the Mereg River without the need for classification or segmentation of rating curve.
 Extended Abstract
1-Introduction
Sediment entrance into rivers causes many problems, including the reduction of reservoirs capacity by sedimentation and the increase of turbidity in water supply systems. In many catchments, short and intense precipitation events are responsible for a large part of the total sediment transport. Therefore, recognizing the variability in different scale of sediment supply, especially in flood scale, is considered very important to manage and understand the river and water structure systems. The lack of proper and adequate data has led to the use of sediment rating curves and regression relations to estimate suspended sediment loads, widely. Due to the lack of data or scattered data, it is very difficult to study suspended sediment in the rivers of Iran. The studying temporal variation of suspended sediment, especially during storms, to achieve sustainable river management is very necessary. The present study aims at investigating intra-storm and seasonal variations of suspended sediment during rainfall and assessing the sediment curve approach in estimating sediment load in storm base in Merg River watershed in Kermanshah province.
2-Materials and Methods
Mereg watershed with 1446 square kilometers area is located in the west of Iran, Kermanshah province. It is a relatively mountainous region with the average altitude of 1524 meters. The average slope of the basin is about 6%. The length of Mereg River to hydrometric station of Khers Abad is 121.34 km.
To conduct the present study and to prepare suspended sediment and flow data, sampling of water and sediment was carried out at Khers Abad hydrometric station in two seasons, the winter of 2015 and the spring of 2016. At the time of rainfall, sampling was done simultaneously with runoff and rising water levels in the river at intervals of 2 to 4 hours. The method of filter paper was used to measure suspended sediment. Discharges associated with sediment samples were also calculated using the discharge- water stage relationship of the Regional Water Organization Station.
Sediment curve approach was used to estimate suspended sediment. Separate curves for rising and falling limbs of hydrographs, seasonal data (winter and spring) and hysteresis patterns were developed and compared with sediment rating curve of the whole data. In addition to the coefficient of determination, the estimate relative error was also used to evaluate the sediment rating curves.
3-Results and Discussion
Five events were chosen from January to May, 2016. Storms had a peak discharge of 1.38 to 40.49 (m3/s), which covers a wide range of low-to-high peak floods in Mereg river. The amount of rainfall is significantly different from 5.27 to 109 mm, as, it led to the average discharges with different values from 0.79 to 10.9 (m3/s). Different rates of discharge have led to different amounts of mean and maximum sediment concentration in these storms; as a result, the suspended sediment concentration ranges from 0.31 to 4.54 (gr/lit) and the maximum concentration from 1 to 15.5 (gr/lit) was obtained.
The development of the hysteresis pattern of sediment showed that two events had the pattern of figure 8, and the other storms had a clockwise or anticlockwise pattern. All of the events were fitted well with a sediment rating curve with a high-level determination coefficient. The highest determination coefficient (96%) was for the event on 04/06/2016 and the lowest (57%) was for 04/13/2016.
Due to poor estimates in its rising limb, the largest uncertainty (75.68%) was for the event of 04/13/2016. The relative error of events is less than 50%, except for the event of 04/13/2016. In addition, the sediment rating curve was established for all measured events in the study period, seasonal events and all the data. They were also evaluated based on the determination coefficient and estimation error. According to the results, the relative error of sediment rating curve of the total data is about 32% which is within acceptable range.
The measurement of suspended sediment during the winter and spring season showed that the sediment transport rate in spring events is higher than winter events. Establishing the sediment hysteresis patterns showed that most of the measured events had a clockwise and anticlockwise pattern, indicating the dominance of one of the sources of sediment (hillslope or waterway) in each storm. Although, the fitted relationships on the total measured rainfall data have a determination coefficient greater than 0.5, the separation of the rising and falling limbs of hydrograph and developing sediment rating curve for them improved the determination coefficient but, it was not able to increase  the accuracy of the rating curves. The calculation of the estimation error indicates that there is no difference, and in some cases, the error is raised by creating a separate curve for the rising and falling limbs. Although the relative error and the difference between observed and estimated values ​​have increased with the hydrological separation of the rating curve, , the relative error is still less than 50%.
The integration of all data has a rating curve with a lower determination coefficient than the rating curve of the each storm. The rating curve of seasonal data cannot affect the accuracy of rating curves or reduce the relative error.
4-Conclusion
During the study period, the hydrological regime of Merg River showed a lot of variability.
Based on the analyzed samples, the total amount of suspended sediment transported during the study period (4 months) was 15919 tons. The sediment transport process in two measured seasons has almost the same pattern, according to the results, without data classification or separation; the rating curve can provide an acceptable estimate of the suspended sediment load.