高分辨率数字高程模型（DEM）是地形分析、水文模拟和地质灾害评价等研究的重要基础数据。然而，受云层遮挡、地形阴影和立体影像匹配条件影响，HMA 8-m DEM在祁连山区存在数据空洞和局部高程异常。本研究利用HMA 8-m DEM、SRTM DEM和ICESat-2 ATL08测高数据，基于ICESat-2高程残差先验引导的DEM超分辨率重建方法（PAG-SR），将SRTM DEM重建至8 m，并对HMA 8-m DEM的数据空洞和局部异常区域进行填补与替换，生成祁连山区8 m融合DEM。采用200,000个未参与训练的ATL08测高点进行验证，结果表明，融合DEM的均方根误差（RMSE）、平均绝对误差（MAE）和平均偏差（MBE）分别为3.9657 m、3.1216 m和2.4351 m。该数据集可为祁连山区地貌分析、水文模拟、冰川变化和生态环境研究提供基础高程数据支持。\n
HMA 8-m DEM 由 GeoEye-1、QuickBird-2 和 WorldView-1/2/3 等高分辨率商业卫星立体影像生成，数据从美国国家冰雪数据中心网站(https://nsidc.org/)下载。本数据集采用的 HMA 8-m DEM 空间分辨率为 8 m，其源影像获取时间范围为 2002 年 1 月至 2016 年 11 月。该产品具有较好的地形细节表达能力，但部分区域受云层遮挡、地形阴影和立体影像匹配条件影响，存在数据空洞和局部高程异常。SRTM DEM 来源于美国国家航空航天局（NASA）等机构联合实施的航天飞机雷达地形测绘任务（SRTM），数据从 NASA Earthdata 网站(https://earthdata.nasa.gov/)下载。本数据集采用的 SRTM DEM 采集于 2000 年 2 月，空间分辨率为 30 m，具有覆盖范围广、数据连续性较好的特点。ICESat-2 ATL08 数据由 ICESat-2 卫星搭载的先进地形激光测高系统（ATLAS）获取，数据从美国国家冰雪数据中心网站(https://nsidc.org/)下载。本数据集采用的 ATL08 数据时间范围为 2018 年 10 月至 2025 年 10 月。\n
（1）对HMA 8-m DEM、SRTM DEM、ICESat-2 ATL08数据及辅助边界数据进行投影转换，统一至WGS 1984 UTM Zone 47N投影坐标系；（2）对ICESat-2 ATL08数据进行质量筛选，计算其与SRTM DEM之间的高程残差，并结合高程、坡度、坡向和地形起伏度等地形因子，采用XGBoost模型生成连续的高程残差先验；（3）利用高程残差先验辅助PAG-SR模型将SRTM DEM超分辨率重建至8 m；（4）以HMA 8-m DEM为主体，采用PAG-SR重建DEM对湖泊水体区域以外的数据空洞和局部异常值进行填补与替换，采用经三次卷积插值重采样至8 m的SRTM DEM替换湖泊水体区域，最终生成祁连山区8 m融合DEM。\n
从经过质量筛选的ICESat-2 ATL08数据中选取200,000个未参与训练的测高点，并将其与对应位置的融合DEM像元进行匹配，采用均方根误差（RMSE）、平均绝对误差（MAE）和平均偏差（MBE）评价融合DEM的高程精度。验证结果表明，融合DEM的RMSE、MAE和MBE分别为3.9657 m、3.1216 m和2.4351 m，整体具有较好的高程精度。此外，对最终数据的空值、异常高程和空间连续性进行了检查，结果显示数据完整覆盖研究区，不同数据源交界处未发现系统性高程突变或明显的人工分块接缝。因此，本数据集可为祁连山区地貌分析、水文模拟、冰川变化和生态环境研究提供基础高程数据支持。", "ds_source": "", "ds_process_way": "", "ds_quality": "", "ds_acq_start_time": "2000-01-01 00:00:00", "ds_acq_end_time": "2025-12-31 00:00:00", "ds_acq_place": "祁连山区", "ds_acq_lon_east": 103.14, "ds_acq_lat_south": 35.769999999999996, "ds_acq_lon_west": 93.47, "ds_acq_lat_north": 39.85, "ds_acq_alt_low": null, "ds_acq_alt_high": null, "ds_share_type": "apply-access", "ds_total_size": 5461814022, "ds_files_count": 0, "ds_format": "*.tif", "ds_space_res": "8米", "ds_time_res": "", "ds_coordinate": "WGS84", "ds_projection": "", "ds_thumbnail": "6f5ece04-e5eb-4119-b107-a022ac30d065.png", "ds_thumb_from": 0, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "None", "ds_from_station": null, "organization_id": "52b7b79b-860c-49a5-9083-9a70cf8bed5a", "ds_serv_man": null, "ds_serv_phone": null, "ds_serv_mail": null, "doi_value": "", "subject_codes": [ "170.4510", "420" ], "quality_level": 0, "publish_time": "2026-09-07 09:30:44", "last_updated": "2026-09-07 09:30:44", "protected": false, "protected_to": null, "lang": "zh", "cstr": null, "i18n": { "en": { "title": "8 m Fused Digital Elevation Model Dataset for the Qilian Mountains", "ds_abstract": "
High-resolution digital elevation models (DEMs) are essential foundational data for terrain analysis, hydrological modeling, geological hazard assessment, and related studies. However, due to cloud cover, terrain shadows, and stereo image matching conditions, the HMA 8-m DEM contains data voids and localized elevation anomalies in the Qilian Mountains. In this study, the HMA 8-m DEM, SRTM DEM, and ICESat-2 ATL08 elevation measurements were used. Based on an ICESat-2 elevation residual prior-guided DEM super-resolution reconstruction method (PAG-SR), the SRTM DEM was reconstructed at a spatial resolution of 8 m and used to fill data voids and replace localized anomalous areas in the HMA 8-m DEM, thereby generating an 8-m fused DEM for the Qilian Mountains. A total of 200,000 ATL08 elevation measurements that were not used for model training were used for validation. The results showed that the root mean square error (RMSE), mean absolute error (MAE), and mean bias error (MBE) of the fused DEM were 3.9657 m, 3.1216 m, and 2.4351 m, respectively. This dataset can provide fundamental elevation data support for geomorphological analysis, hydrological modeling, glacier change, and ecological and environmental research in the Qilian Mountains.\n
The HMA 8-m DEM was generated from high-resolution commercial stereo satellite imagery acquired by GeoEye-1, QuickBird-2, and WorldView-1/2/3, among others, and was downloaded from the National Snow and Ice Data Center website (https://nsidc.org/). The HMA 8-m DEM used in this dataset has a spatial resolution of 8 m, and its source imagery was acquired between January 2002 and November 2016. The product provides a good representation of terrain details; however, some areas contain data voids and localized elevation anomalies due to cloud cover, terrain shadows, and limitations in stereo image matching. The SRTM DEM was derived from the Shuttle Radar Topography Mission (SRTM), jointly implemented by the National Aeronautics and Space Administration (NASA) and other organizations, and was downloaded from the NASA Earthdata website (https://earthdata.nasa.gov/). The SRTM DEM used in this dataset was acquired in February 2000 and has a spatial resolution of 30 m, with extensive spatial coverage and good data continuity. The ICESat-2 ATL08 data were acquired by the Advanced Topographic Laser Altimeter System (ATLAS) onboard the ICESat-2 satellite and downloaded from the National Snow and Ice Data Center website (https://nsidc.org/). The ATL08 data used in this dataset cover the period from October 2018 to October 2025.\n
(1) The HMA 8-m DEM, SRTM DEM, ICESat-2 ATL08 data, and auxiliary boundary data were reprojected to the WGS 1984 UTM Zone 47N projected coordinate system. (2) The ICESat-2 ATL08 data were quality-filtered, and elevation residuals between ATL08 and the SRTM DEM were calculated. These residuals, together with terrain factors including elevation, slope, aspect, and terrain relief, were used to generate a continuous elevation residual prior using the XGBoost model. (3) Guided by the elevation residual prior, the PAG-SR model was used to reconstruct the SRTM DEM at a spatial resolution of 8 m through super-resolution. (4) Using the HMA 8-m DEM as the primary elevation dataset, the PAG-SR-reconstructed DEM was used to fill data voids and replace localized anomalous values outside lake areas, while lake areas were replaced with the SRTM DEM resampled to 8 m using cubic convolution interpolation. Finally, an 8-m fused DEM of the Qilian Mountains was generated.\n
A total of 200,000 quality-filtered ICESat-2 ATL08 elevation measurements that were not used for model training were selected and matched with the corresponding pixels of the fused DEM. The root mean square error (RMSE), mean absolute error (MAE), and mean bias error (MBE) were used to evaluate the elevation accuracy of the fused DEM. The validation results showed that the RMSE, MAE, and MBE of the fused DEM were 3.9657 m, 3.1216 m, and 2.4351 m, respectively, indicating good overall elevation accuracy. In addition, the final dataset was checked for missing values, anomalous elevations, and spatial continuity. The results showed that the dataset completely covers the study area, with no systematic elevation discontinuities or obvious artificial tiling seams detected at the boundaries between different data sources. Therefore, this dataset can provide fundamental elevation data support for geomorphological analysis, hydrological modeling, glacier change, and ecological and environmental research in the Qilian Mountains.", "ds_time_res": "", "ds_space_res": "8 m", "ds_projection": "" } }, "submit_center_id": "ncdc", "data_level": 0, "recommendation_value": 0, "license_type": "https://creativecommons.org/licenses/by/4.0/", "doi_reg_from": "reg_local", "cstr_reg_from": "reg_local", "doi_not_reg_reason": null, "cstr_not_reg_reason": null, "is_paper_in_submitting": false, "belong_to_nieer": false, "allow_update_data": false, "ds_topic_tags": [ "数字高程模型", "DEM超分辨率", "数据融合", "ICESat-2", "祁连山区" ], "ds_subject_tags": [ "地球科学", "地理学", "自然地理学", "测绘科学技术" ], "ds_class_tags": [], "ds_locus_tags": [ "中国", "祁连山区" ], "ds_time_tags": [ 2000, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2018, 2019, 2020, 2021, 2022, 2023, 2024, 2025 ], "ds_contributors": [ { "true_name": "张彦丽", "email": "zyl0322@nwnu.edu.cn", "work_for": "西北师范大学地理与环境科学学院", "country": "中国" }, { "true_name": "饶科", "email": "3034836915@qq.com", "work_for": "西北师范大学", "country": "中国" } ], "ds_meta_authors": [ { "true_name": "饶科", "email": "3034836915@qq.com", "work_for": "西北师范大学", "country": "中国" } ], "ds_managers": [ { "true_name": "张彦丽", "email": "zyl0322@nwnu.edu.cn", "work_for": "西北师范大学地理与环境科学学院", "country": "中国" } ], "category": "遥感及产品" }