|
Application of Artificial Neural Network and Regression Models in Sediment Yield in Plots Located in Disturbed and Undisturbed Plots in Educational and Research Forest Watershed of Tarbiat Modares University, Iran
|
Majid Khazaii *   , Seyyed Hamidreza Sadeghi , Seyyed Khallagh Mirnia |
|
|
|
Abstract: (15625 Views) |
| Sediment yield prediction is one of the important factors in watershed management. So that many different techniques viz. artificial neuron network (ANN) and regression models have been used to predict sediment yield. However, their performances have not been evaluated. The present study therefore aimed to model storm wise sediment yield from small plots located in disturbed and undisturbed forest areas in three replicates in Educational and Research Forest Watershed of Tarbiat Modares University (Kojour) occurred during October 2008 to April 2009. The modeling was conducted using two methods of ANN and bi and Multivariate regression models. The performance of the models was checked by means of RMSE, R2 and coefficient of efficiency. The results of the study showed that the multivarite regression model with three input variables of rain depth, runoff volume and runoff coefficient performed better than other multivarite regression. The ANN model with the inputs similar to those for the multivarite regression also had better performance rather than other ANN models. In overall, the ANN models with the same input variables performed better than multivariate regression models in the study treatments. |
|
| Keywords: Deforestation, Erosion Plot, Soil Conservation, Storm wise Sediment Yield |
|
|
Full-Text [PDF 406 kb]
(8104 Downloads)
|
Type of Study: Research |
Subject:
Special
Received: 2014/03/2 | Accepted: 2014/03/2 | Published: 2014/03/2
|
|
|
|
|
|
|
| Add your comments about this article |
|
|
|
