Document Type : Research Paper

Authors

• Saeede Jalali ¹
• Kazem Nosrati ²
• Shahram Bahrami ³

¹ Department of physical geography; faculty of earth science shahid beheshti university tehran;IRAN

² Professor Department of Physical Geography Faculty of Earth Sciences Shahid Beheshti University Tehran.IRAN

³ Department of Physical Geography Faculty of Earth Sciences Shahid Beheshti University

Abstract

Keywords

• Chehel-Chay catchment
• geochemical tracers
• geomorphic landforms

20.1001.1.2008479.1402.54.5.7.5

Main Subjects

• soil erosion and conservation

Investigating the contribution of geomorphic landforms in sediment yield by using sediment fingerprinting method

Extended Abstract

Introduction

An important and significant issue in applied research for the management of watersheds is to know integrated drainage sediment yield process to recognize hill slopes and fluvial system. In erosion and sediment study, investigation of form and landform evolution was the earth science study core in the long time. In recently two decades, sediment fingerprinting method was proven as a key method for contribution of sediment proportion. The goal of this paper was to investigate sediment fingerprinting by use of geomorphic landforms in Chehel-chay catchment.

Data and research methods

Chehel-Chay catchment is located in the northern mountain of eastern Alborz ( E to   E longitude and   to   N latitude), covers an area of 256 km². It is a forest mountainous catchment with elevations ranging from 190 m in outlet catchment to 2570 m, and average catchment elevation is 951 m. Mean slope percent is between 35-40%. The study region has population of 14068 with the majority located in Dozain rural (5700). The long-term (30 years) mean annual precipitation data collected at the nearest climate station indicates an average total of   precipitation 750 mmyr^-1 with most precipitation falling between October and March. Precipitation in the upper parts of the catchment is mostly snow.

The Topographic Position Index (TPI) compares the elevation of each cell in a DEM to the mean elevation of a specified neighborhood around that cell. Four classes were defined using the criteria; concave and convex slope, valleys and ridges and slope between concave and convex slope.

After sampling all sediments, soil samples were transported to the geomorphology lab and they were air-dried for determining main particular fraction. All samples mixed with boric acid and grinded. Then, they were altered to solid pile and maintained in a special container and measured by XRF (X-ray fluorescence) method. Concentration of geochemical element: As، Ba، Cao، CI، Cu، Cr، Ni، S، Pb، Sr، V، Zn (in ppm) and oxidation percent of Al2o3، Fe2o3، K2o، Mno، Mgo، Na2o، P205، SiO2، Tio2 were calculated.

For investigating the proportion of each source of sediment in tributary, three main steps were used. First, the non-conservative behavior of tracers and a mass conservation test was performed. Second, a two-stage statistical procedure identified the optimum set of source material properties to use as composite fingerprints. The abilities of individual properties to discriminate among sources were tested via the Kruskal-Wallis rank sum test, and those properties that return a P value >0.05 were excluded. Then, a stepwise discriminant function analysis (DFA) was performed to determine the proportion of samples that were accurately classified into the correct source groups. Third the mixture sampling-Importance-Resampling (Mix SIR) Bayesian model was used to estimate source proportion. The model predictions were evaluated using 9 to 11 sets of virtual sediment mixture for the tributary land forms/use and anomaly drainage network and the steam ordering drainage source proportion was multiplied by the values of the tracers selected as constituent properties in the composite signature and the resultant concentrations used as input to the un-mixing model. The predicted source proportions were then compared with the known proportions to assess the accuracy of the un-mixing model predictions. The outcomes of the virtual mixtures tests were assessed using the root mean squared error (RMSE) and mean absolute error (MAE)

Results

Using tributary landform as sediment source, S, Sr and Fe2o5 were non-conservative and excluded from further tests. According to Kruskal-Wallis H-Test, Cu and Cr were not significant for next test, so 17 tracers selected to discriminant function analysis and then in the final step in this section Ba, Ni, Pb, Mgo, Mno, V and Cao was selected to enter mixing model.

The results of mixture sampling-Importance-Resampling (MixSIR) Bayesian model shows that among chosen morphologic classes as well as surface source sediment in chehel-chay catchment, the Upper slope 60.9 % which located in the end of slope convex form and Ridge (narrow valleys) 28.5% have most proportion in sediment yield. According to comparisons of the predicted and known relative contributions from landform and stream ordering, using the virtual mixtures showed that the RMSE ranged between 1.7% and 19.4% and MAE between 0.6% and 8.4 % in land forms.

Discussion and suggestions

Sediment tracing is successful methods to catchment management and maintaining soil and water. According to the results, concave slope is most contribution in sediment yield in Chehel-Chay catchment. Farm lands are located in this landforms and accelerated erosion process. Utilizing protective measures such as stabilizing plant roots and plotting on convex slopes is one of the effective methods to prevent soil erosion.