Efficiency of Power Regression Model for Bankfull Hydraulic Geometry in Stable Canals (Case Study: Tarwal Basin)

Development of hydraulic geometry relationships are of great importance in different applications of stream restoration design and hydrological modeling. Accordingly the main objective of this study was to develop regional relationships for bankfull channel dimensions in Tarwal basin using the drainage area as an independent variable. Several fieldworks, focused on Chameh Sis, Sangeh Siah, and Ozon Dareh tributaries as well as the main channel of Tarwal River, were conducted between 4 and 19 November 2015. Twenty two stable-channel cross sections were used to obtain the relationship among the bankfull channel dimensions and the drainage area. The results indicated that piecewise models are the best models for such relations. Simple power regression models were assessed as not suitable models for predicting hydraulic geometry according to the Relative Root Mean Squared Error (RRMSE) values of greater than 0.4. While for the piecewise models the RRMSE values were less than 0.2 for predicting bankfull cross sectional area and width of stable alluvial channels in basins smaller than 1165 km2. Also the RRMSE values were less than 0.12 and 0.25 respectively for predicting mean and maximum bankfull depth of stable alluvial channels in basins smaller than 100 km2. Nevertheless there were no ideal model to estimate bankfull channel dimensions in basins larger than 1165 Km2. This could be a result of significant changes in rainfall-runoff relationships or the influences of anthropogenic disturbances in large basins. The results provided useful data for stream restoration as well as water structure design.