Scholarlink

Abstract <p>Stream-burning methods are widely used to improve drainage delineation in digital elevation models (DEMs), yet they frequently distort terrain topography, alter natural flow pathways, or create unrealistic longitudinal profiles. These issues undermine the accuracy of hydrological modeling and reduce the reliability of DEM-derived parameters such as slope, drainage networks, and watershed boundaries. To address these limitations, we present the gradient-based optimized flow enforcement (GBOFE) algorithm, which selectively modifies only the cells with the minimum downslope gradient required to achieve hydrological consistency. This targeted adjustment preserves elevation and slope integrity while ensuring effective flow enforcement. We evaluated GBOFE against five established DEM-reconditioning methods (ANUDEM, FillBurn, AGREE, Normalized Excavation, and r.carve) in the Recio and Venadillo basins in Colombia. Performance was assessed through drainage representation, elevation fidelity, and watershed-boundary accuracy. GBOFE demonstrated superior hydrological performance, achieving an Intersection over Union of 0.87, a Dice coefficient of 0.93, and a Probability of Detection of 0.88. It also maintained high topographic fidelity, with elevation RMSE below 7&#xa0;m and slope errors below 4%. In contrast, FillBurn and AGREE produced substantial terrain distortions, while r.carve and ANUDEM were less effective in accurately delineating drainage networks. Overall, the results show that GBOFE outperforms widely used flow-enforcement techniques by reducing false positives and avoiding artificial channel profiles or significant topographic alterations. By balancing hydrological accuracy with geomorphological consistency, GBOFE offers a robust alternative for DEM reconditioning, supporting more reliable watershed delineation, hydrological modeling, and geomorphological analysis.</p> Graphical Abstract <p>This graphical abstract introduces GBOFE (Gradient-Based Optimized Flow Enforcement), a selective stream-burning method designed to hydrologically condition Digital Elevation Models (DEMs) while preserving topographic integrity. GBOFE converts a vector drainage network into a raster, assigns topological connectivity identifiers, and adjusts only the cells necessary to ensure flow connectivity with minimal downstream gradient changes. We assessed GBOFE’s performance in the Recio and Venadillo basins (Colombia) against ANUDEM, FillBurn, AGREE, Normalized Excavation, and r.carve, using metrics for drainage representation, topographic fidelity, and watershed boundary accuracy. GBOFE outperformed these methods by delivering superior hydrological accuracy and preserving landform morphology with minimal false positives, avoiding artificial longitudinal profiles or significant terrain alterations. The result is a hydrologically conditioned DEM optimized for watershed delineation and hydrological modeling. Overall, GBOFE balances accurate drainage representation with minimal topographic distortion, establishing it as a robust tool for DEM-based hydrological conditioning and enhanced flow delineation. The proposed algorithm has potential applications for reducing uncertainty in water resource studies, hydrological modeling, and geomorphological analysis.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Impact of Flow Enforcement Methods on DEM Hydrological and Topographical Consistency: Introducing a Gradient-Based Algorithm

  • 期刊论文

  • 发表日期：

    2026年2月11日

查看全文
  • Henry Rubiano,
  • Cristian Guevara-Ochoa,
  • Miguel Barrios

摘要

Abstract

Stream-burning methods are widely used to improve drainage delineation in digital elevation models (DEMs), yet they frequently distort terrain topography, alter natural flow pathways, or create unrealistic longitudinal profiles. These issues undermine the accuracy of hydrological modeling and reduce the reliability of DEM-derived parameters such as slope, drainage networks, and watershed boundaries. To address these limitations, we present the gradient-based optimized flow enforcement (GBOFE) algorithm, which selectively modifies only the cells with the minimum downslope gradient required to achieve hydrological consistency. This targeted adjustment preserves elevation and slope integrity while ensuring effective flow enforcement. We evaluated GBOFE against five established DEM-reconditioning methods (ANUDEM, FillBurn, AGREE, Normalized Excavation, and r.carve) in the Recio and Venadillo basins in Colombia. Performance was assessed through drainage representation, elevation fidelity, and watershed-boundary accuracy. GBOFE demonstrated superior hydrological performance, achieving an Intersection over Union of 0.87, a Dice coefficient of 0.93, and a Probability of Detection of 0.88. It also maintained high topographic fidelity, with elevation RMSE below 7 m and slope errors below 4%. In contrast, FillBurn and AGREE produced substantial terrain distortions, while r.carve and ANUDEM were less effective in accurately delineating drainage networks. Overall, the results show that GBOFE outperforms widely used flow-enforcement techniques by reducing false positives and avoiding artificial channel profiles or significant topographic alterations. By balancing hydrological accuracy with geomorphological consistency, GBOFE offers a robust alternative for DEM reconditioning, supporting more reliable watershed delineation, hydrological modeling, and geomorphological analysis.

Graphical Abstract

This graphical abstract introduces GBOFE (Gradient-Based Optimized Flow Enforcement), a selective stream-burning method designed to hydrologically condition Digital Elevation Models (DEMs) while preserving topographic integrity. GBOFE converts a vector drainage network into a raster, assigns topological connectivity identifiers, and adjusts only the cells necessary to ensure flow connectivity with minimal downstream gradient changes. We assessed GBOFE’s performance in the Recio and Venadillo basins (Colombia) against ANUDEM, FillBurn, AGREE, Normalized Excavation, and r.carve, using metrics for drainage representation, topographic fidelity, and watershed boundary accuracy. GBOFE outperformed these methods by delivering superior hydrological accuracy and preserving landform morphology with minimal false positives, avoiding artificial longitudinal profiles or significant terrain alterations. The result is a hydrologically conditioned DEM optimized for watershed delineation and hydrological modeling. Overall, GBOFE balances accurate drainage representation with minimal topographic distortion, establishing it as a robust tool for DEM-based hydrological conditioning and enhanced flow delineation. The proposed algorithm has potential applications for reducing uncertainty in water resource studies, hydrological modeling, and geomorphological analysis.