مدل‌سازي بار رسوب زيرحوضه دو حوضه آبخيز زاينده‌رود عليا با استفاده از روش رگرسيون كمترين مربعات فازي

Modeling for Sediment Discharge in Sub-basin No. 2 of Zayandehrud Olia Watershed Using Least Square Fuzzy Regression Model

برآورد توليد رسوب حوضه‌هاي آبخيز، يك امر مهم در طراحي سازه‌هاي آبي، مسايل زيست‎محيطي و كيفيت آب رودخانه‌ها است. يكي از متداول‌ترين روش‎ها براي برآورد ميزان رسوب، روش‎هاي تجربي است. با توجه به عدم قطعيت ارتباط بين توليد رسوب و عوامل موثر بر آن، روابط تجربي ارايه‎شده در اين زمينه كارايي چنداني در اين امر ندارند. اين پژوهش به مدل‌سازي بار رسوب يكي از زير‌حوضه‌هاي (زيرحوضه 2) حوضه آبخيز زاينده‌رود عليا پرداخته است. روش آن از نوع توصيفي ـ تحليلي مبتني بر روش‎هاي آماري، ميداني و كتابخانه‌اي است؛ بدين‎گونه كه ابتدا منطقه به واحدهاي همگن ژيومورفولوژي بخش شد، سپس نمونه‌برداري خاك از عمق 30-15 سانتي‌متري انجام گرفت و عوامل فيزيكي و شيميايي خاك در هر واحد همگن، تعيين و مدل‌سازي بار رسوب با استفاده از مدل رگرسيون كمترين مربعات فازي انجام گرفت. نتايج نشان داد كه رسوب برآورد شده كل زيرحوضه با روش رگرسيون كمترين مربعات فازي سالانه 625/2 تن در هكتار است كه واحد همگن 10 حساس‌ترين واحد همگن به فرسايش و توليد رسوب است. همچنين نسبت نيكويي برازش بين رسوب برآوردي مدل مذكور و رسوب مشاهده‌اي 73/0 است.

Introduction Estimation and prediction of sediment in the basin was a wide range of issues such as dams, river engineering, pollution migration, the environmental of effects, watershed management and water resources in many parts of the application. Although the direct measurement can be time-consuming and costly, it is more reliable for the measurement. Experimental methods with specific conditions of the countries provided the models and calibration model in our country without the results achieved. Usually the estimation of sediment from a direct measurement of equation or presented in the experimental methods are determined. Beacause the same Fuzzy Regression methods are widely used in model relationships in nature have been recruiting. This is a logical connection methods can input and output between the observational data by. Due to the time-consuming and costly nature of direct methods and experimental estimation of sediment model calibra lted in this study was attempted using Fuzzy methods, modeling of sediment load subbasin 2 Olya Zayandeh Rood basin and estimate the results of this model with the deposition of the comparison view. This watershed is located in South-West of Shahrekord City and inclode 9 subbasin. Sediment discharge from the basin is transferred by Zayandeh Rud River behind Zayandehrud-dam reservoir. Methodology This research is based on a description of the type of statistical methods, analytical field and a library. In order to scratch the area was divided into homogeneous units, geomorphology, soil sampling depth then 15-30 cm and soil physical and chemical parameters in each homogeneous unit and sediment load modeling using Fuzzy least squares regression model the most. Initially the number of soil samples at 37 points in homogeneous units was identified. For regression model fitted using most low physical and chemical parameters, and least square Fuzzy the soil in front of a homogeneous units, view the amount of sediment using principal component analysis (PCA) and the effect parameters of low weights and have in common with the amount of sediment were removed. Finally, among the listed parameters for Calcium Carbonate, percent organic matter, specific gravity, bulk density and higher weights and silt were fitted to the model. Results and Discussion Fuzzy method for the examination in this study, the test has been used well fitted. In this method, the ratio of intersection to union is used to evaluate the models. Calculation of parameters in Fuzzy Regression models for fuzzy numbers that are on the central value of the recipient and also the degree of possible ambiguity. According to these ambiguity values represent that given estimates of sediment per unit can be changed in a homogeneous domain that the width of the left is the ambiguity of smaller amounts of sediment on the recipient and the right signal to large amounts of sediment with more ambiguity. The unit of homogeneous production will have the highest precipitation of 10 that this is due to the high silt, apparent specific gravity. It also has one of the most 9 homogeneous unit of production value in less sediment. According to the findings of the research, the percentage difference between the deposition of sediment and view estimated immediately after the most squares is less Fuzzy Regression analysis. The results of this study show that most little squares regression model Fuzzy Modeling of sediment load is acceptable model. Conclusion The results showed that the estimated total Subbasin deposits by a little Fuzzy Regression model the least squares 625/2 tons per hectar per year, which is most sensitive to homogeneous units of homogeneous units 10 to erosion and sediment production and must be in protective priority programmers. According to the findings of this study, therefore, is a little fuzzy regression model the most squares for modeling and estimation of sediment in the basin of the Subbasin and similar, is an appropriate application. A comparison between estimated deposits of the most squares regression model fuzzy and low in view of the sediment showed that the differences between the deposition of estimated and observed sediment and least squares this little fuzzy regression analysis is that the estimated regression model of deposition of the least squares fuzzy, immediately percent less than the amount of sediment in the view.