A technique for stream geometry estimation based on watershed morphometric characteristics

Abstract

This study presents an approach for deriving stream geometry from watershed morphometric characteristics, addressing data scarcity for hydrologic and hydraulic modeling. The approach was tested in Puerto Rico and involves subdividing the study area into homogeneous regions and developing region-specific stream geometry predictive equations using ordinary least squares regression to correlate morphometric characteristics with the stream width and depth for a 2-year return period flow. The resulting equations were verified and validated by applying them to generate cross-sections for eight Storm Water Management Models. Results showed a correlation between drainage area and channel morphology, attributed to the presence of areal/shape characteristics in all the specific-region equations. Verification with an independent watershed sample demonstrated the predictive equations’ robustness and reliability. The study found that subdividing homogeneous regions improved accuracy compared to previous studies. Validation demonstrated up to 45.9 % (∼1.29 m³ s^-1) differences in peak flow modeling, corroborating their applicability for modeling purposes.