Optimized Land Surface Low Point Detection Using the D8 Algorithm in a Geographic Information System (GIS) Framework

Abstract

Hydrological analysis in urban areas often suffers from inaccuracies in Digital Elevation Model (DEM) interpretation, especially in detecting micro-depressions and small-scale surface flow patterns. Previous studies typically relied solely on the automatic D8 algorithm in GIS without manual verification, resulting in flow directions that do not fully represent actual surface conditions. This study aims to compare manual D8-based flow direction calculations with automatic ArcGIS processing using DEMNAS data for Langsa City. The DEM (8.1 m resolution) underwent sink filling, hydrological conditioning, slope and aspect processing, followed by field validation using GPS measurements. The results show that the manual method identified 23 flow paths, whereas ArcGIS detected only 11. The differences stem mainly from micro-topographic variations that the automatic algorithm failed to capture in flat areas or anthropogenically modified surfaces. Field validation confirmed that 8 of the 11 ArcGIS-derived paths matched the actual drainage patterns, while the additional manual paths better represented subtle elevation gradients.This research contributes by offering a systematic comparison between manual and automatic D8 approaches, highlighting the importance of manual verification in low-slope urban terrains. The findings are valuable for micro-scale flood mitigation planning and urban surface hydrology analysis.