Scholarlink

<p>Digital Elevation Models (DEMs) are key datasets for Earth’s surface analysis, yet recent models such as AW3D30 and COPERNICUS remain underexplored. Most accuracy assessments focus on slope, aspect, and land cover, often overlooking other terrain attributes. This study evaluates the vertical accuracy of six DEMs—ASTER, SRTM, NASADEM, ALOS PALSAR, COPERNICUS, and AW3D30—in Morocco’s Rif Mountains, a region prone to soil degradation and landslides. Accuracy was assessed using 10,015 ground control points (313 geodetic points and 9,702 spot heights) from topographic maps, with errors computed as the differences between reference elevations and DEM values. Statistical descriptors and analyses of error distributions were applied. Results showed that, when using geodetic control points, AW3D30 was the most accurate, with a Root Mean Square Error (RMSE) of 7.77&#xa0;m, followed by COPERNICUS (RMSE = 9.09&#xa0;m). ASTER was the least accurate (14.63&#xa0;m), while ALOS PALSAR, NASADEM, and SRTM fell in between (12.92&#xa0;m, 13.20&#xa0;m, and 13.01&#xa0;m, respectively). Accuracy decreases with increasing elevation, slope, curvature, terrain ruggedness, and texture. Larger errors occur in forests, rangelands, water bodies, and landforms such as spurs and hollows, whereas slope orientation shows a weak influence. This comprehensive assessment provides valuable guidance for selecting DEMs and highlights terrain factors that influence their reliability across complex landscapes.</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

How terrain shapes digital elevation model accuracy: a comparative assessment of six freely available digital elevation models

  • 期刊论文

  • 发表日期：

    2026年4月15日

查看全文
  • Mohamed El Amrani,
  • Rachida Azelmat,
  • Afaf Amine,
  • Mohamed Hafedh Hamza,
  • Khadija Diani,
  • Said Courba,
  • Moulay Ahmed Ben Driss,
  • Nabil Mdiker,
  • Aziza Lamchaimech

摘要

Digital Elevation Models (DEMs) are key datasets for Earth’s surface analysis, yet recent models such as AW3D30 and COPERNICUS remain underexplored. Most accuracy assessments focus on slope, aspect, and land cover, often overlooking other terrain attributes. This study evaluates the vertical accuracy of six DEMs—ASTER, SRTM, NASADEM, ALOS PALSAR, COPERNICUS, and AW3D30—in Morocco’s Rif Mountains, a region prone to soil degradation and landslides. Accuracy was assessed using 10,015 ground control points (313 geodetic points and 9,702 spot heights) from topographic maps, with errors computed as the differences between reference elevations and DEM values. Statistical descriptors and analyses of error distributions were applied. Results showed that, when using geodetic control points, AW3D30 was the most accurate, with a Root Mean Square Error (RMSE) of 7.77 m, followed by COPERNICUS (RMSE = 9.09 m). ASTER was the least accurate (14.63 m), while ALOS PALSAR, NASADEM, and SRTM fell in between (12.92 m, 13.20 m, and 13.01 m, respectively). Accuracy decreases with increasing elevation, slope, curvature, terrain ruggedness, and texture. Larger errors occur in forests, rangelands, water bodies, and landforms such as spurs and hollows, whereas slope orientation shows a weak influence. This comprehensive assessment provides valuable guidance for selecting DEMs and highlights terrain factors that influence their reliability across complex landscapes.