.png){kind=icon}
.png){kind=icon}
Return
Improving urban catchment delineation through comparative DEM accuracy assessment and bias correction techniques
.png){kind=icon}
OA
AI
A
A
P
R
DOI：10.1016/j.srs.2026.100463.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Accurate terrain representation is essential for hydrological modelling, yet freely available global Digital Elevation Models (DEMs) often contain substantial vertical errors, particularly in complex urban environments, because they represent surface elevations rather than true bare-earth terrain. This study evaluates the accuracy of four widely used DEMs, ALOS PALSAR, SRTM, ASTER GDEM, and CARTO, within the Assi River catchment in northern India using 1695 high-precision ground control points (GCPs) using an independent train–test validation framework. The uncorrected DEMs exhibit distinct error characteristics: ALOS PALSAR and CARTO show large systematic vertical biases (∼–60 m), SRTM displays moderate bias, while ASTER GDEM is dominated by high random noise. To address these limitations, a hybrid two-step correction method was applied, combining linear bias correction with a kriging-based residual surface. This approach reduced RMSE by up to 90% and eliminated global vertical offsets, yielding near-normally distributed residuals for all DEMs. Hydrological analysis using the corrected DEMs revealed substantial changes in the delineated catchment area. ALOS PALSAR increased from 15.78 km2 to 16.39 km2 (+0.61 km2), SRTM decreased from 17.20 km2 to 15.06 km2 (−2.14 km2), CARTO expanded from 14.11 km2 to 22.90 km2 (+8.79 km2), and ASTER GDEM increased from 1.56 km2 to 3.46 km2 (+1.90 km2). These results demonstrate that vertical biases in raw DEMs can substantially distort watershed boundaries and hydrological interpretation, whereas the proposed correction framework significantly enhances DEM reliability for urban catchment studies. Overall, this study underscores the necessity of DEM preprocessing for accurate flood modelling, drainage planning, and terrain-based hydrological assessments in low-relief urban landscapes.
Journal
IF:
5.2
Papers:
457
Citations:
980
Organization
Cited Papers