Scholarlink

<p>Digital Elevation Models (DEMs) are used in a wide range of applications including, but not limited to geoid modelling, natural hazard monitoring, hydrologic modelling and water resources management. Comparison of the heights implied by DEMs on precise levelled control points is the traditional way of accessing their performance. In regions with sparse or no spirit levelled heights this can cause a challenge, since this type of surveying is expensive and time consuming. In Kenya, for example, the height datum in some parts of the country is not everywhere connected by precise levelling to Mean Sea Level (MSL). In addition, systematic errors can occur due to differences in vertical datums. On the other hand, space technology using Global Navigation Satellite Systems (GNSS) is able to provide three-dimensional coordinates in a more efficient and cost-effective way. Evaluation of DEMs using GNSS measurements can be an alternative method to the evaluation using spirit levelling heights in Kenya. In this study, six open source DEMs, namely ASTER, SRTM, AW3D, GTOPO30, TANDEM-X and VF PANORAMA, covering the Kenyan territory are evaluated using a set of GNSS points, by converting into ellipsoidal heights the different heights implied by the DEMs. A total of 312 GNSS points, covering the territory of Kenya with a spacing of approximately 70&#xa0;km, are utilised. Standard and robust statistical methods are used to analyse the elevation errors of the DEMs, both with and without outliers present. Outliers are detected by performing normality tests on the datasets and computing Z-scores and R-scores. Data with Z-score &gt; 3 and/or R-score &gt; 2.5 are considered outliers. Results indicate that, in the absence of outliers, the TANDEM-X90, with a vertical accuracy of 3.414&#xa0;m (95th percentile of the elevation errors), performs best of all other DEMs in Kenya, while GTOPO30, with a vertical accuracy of 19.549&#xa0;m offers the worst performance. The other DEMs perform fairly well with vertical accuracies of &lt; 10&#xa0;m, which is sufficient for applications where relative accuracy is more important than absolute accuracy. For regional geoid modelling requiring the calculation of terrain effects on gravity over the whole country, AW3D30, TANDEM-X90, SRTM and VF-PANORAMA may be used.</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

Evaluation of space derived open-source digital elevation models over Kenya

  • 期刊论文

  • 发表日期：

    2026年1月12日

查看全文
  • C. J. Nyoka,
  • A. H. M. Din,
  • W. Maritim,
  • M. F. Pa’suya

摘要

Digital Elevation Models (DEMs) are used in a wide range of applications including, but not limited to geoid modelling, natural hazard monitoring, hydrologic modelling and water resources management. Comparison of the heights implied by DEMs on precise levelled control points is the traditional way of accessing their performance. In regions with sparse or no spirit levelled heights this can cause a challenge, since this type of surveying is expensive and time consuming. In Kenya, for example, the height datum in some parts of the country is not everywhere connected by precise levelling to Mean Sea Level (MSL). In addition, systematic errors can occur due to differences in vertical datums. On the other hand, space technology using Global Navigation Satellite Systems (GNSS) is able to provide three-dimensional coordinates in a more efficient and cost-effective way. Evaluation of DEMs using GNSS measurements can be an alternative method to the evaluation using spirit levelling heights in Kenya. In this study, six open source DEMs, namely ASTER, SRTM, AW3D, GTOPO30, TANDEM-X and VF PANORAMA, covering the Kenyan territory are evaluated using a set of GNSS points, by converting into ellipsoidal heights the different heights implied by the DEMs. A total of 312 GNSS points, covering the territory of Kenya with a spacing of approximately 70 km, are utilised. Standard and robust statistical methods are used to analyse the elevation errors of the DEMs, both with and without outliers present. Outliers are detected by performing normality tests on the datasets and computing Z-scores and R-scores. Data with Z-score > 3 and/or R-score > 2.5 are considered outliers. Results indicate that, in the absence of outliers, the TANDEM-X90, with a vertical accuracy of 3.414 m (95th percentile of the elevation errors), performs best of all other DEMs in Kenya, while GTOPO30, with a vertical accuracy of 19.549 m offers the worst performance. The other DEMs perform fairly well with vertical accuracies of < 10 m, which is sufficient for applications where relative accuracy is more important than absolute accuracy. For regional geoid modelling requiring the calculation of terrain effects on gravity over the whole country, AW3D30, TANDEM-X90, SRTM and VF-PANORAMA may be used.