Performance assessment of global open‐access digital elevation models in China mainland coastal region
Performance assessment of global open‐access digital elevation models in China mainland coastal region
复制标题
DOI：
10.1002/esp.5677
复制
发表时间：
2023-08
影响因子：
3.3
通讯作者：
Hu Ding;Junhao Liu;Sihang Yang;Junxing Luo;Yi Liu;Xinyi Liang;J. Na;Shuai Jiang;Yingchun Fu
中科院分区：
文献类型：
--
作者：
Hu Ding;Junhao Liu;Sihang Yang;Junxing Luo;Yi Liu;Xinyi Liang;J. Na;Shuai Jiang;Yingchun Fu
数字高程模型（DEM）是海岸带生态系统监测的基础数据集，最近报道了几个全球开放访问的DEM。在沿海地区，尚未对这些数字高程模型进行全面的垂直精度评估。本研究以高精度ICESat-2数据为参考，对中国大陆沿海地区8个开放获取的DEM数据集（SRTM-3、SRTM-1、TanDEM-X、ASTER GDEM v3、MERIT DEM、AW 3D 30、NASADEM和NationalDEM）进行了调查。统计工具，包括平均绝对误差（MAE）和均方根误差（RMSE）的选择来描述数据的误差/不确定性和空间分布。进一步分析了高程差、坡度、地形地貌和土地利用等因素对高程精度的影响，以评价其适用性。评估结果显示，NASADEM和NASADEM数据集具有最高的精度，对于3弧秒和1弧秒分辨率DEM，MAE值分别为1.68和1.88 m，RMSE值分别为2.55和2.61 m。其他精度接近的DEM包括AW 3D 30、SRTM-1、MERIT和SRTM-3 DEM。结果表明，EMAL-DEM优于其他数据集，表明其适用于沿海地区。
Digital elevation models (DEMs) are the fundamental datasets for coastal ecosystem monitoring, and several global open‐accessed DEMs have recently been reported. In coastal regions, a comprehensive vertical accuracy assessment of these DEMs has not yet been carried out. In this study, eight open‐access DEM datasets, including SRTM‐3, SRTM‐1, TanDEM‐X, ASTER GDEM v3, MERIT DEM, AW3D30, NASADEM, and CoastalDEM, were investigated across the coastal region of the Chinese mainland using high accuracy ICESat‐2 data as a reference. Statistical tools including mean absolute error (MAE) and root mean square error (RMSE) were selected to describe the data error/uncertainty and spatial distribution. Moreover, the effects of elevation ranging, slope degree, geomorphogenesis and landuse on vertical accuracy were further analyzed to assess their applicability. The assessment results revealed that the CoastalDEM and NASADEM datasets had the highest accuracy, with MAE values of 1.68 and 1.88 m and RMSE values of 2.55 and 2.61 m, for 3‐arc second and 1‐arc second resolution DEMs, respectively. Other DEMs with close accuracies include AW3D30, SRTM‐1, MERIT, and SRTM‐3 DEM. The results proved that the CoastalDEM outperformed other datasets, indicating its applicability in coastal regions.