PDF(10688 KB)
基于国产立体卫星影像的高原冰湖三维重建与蓄水量估计——以“7∙8”吉隆口岸冰湖溃决洪水为例
曲英杰, 叶松, 王莹, 郑学东, 李芹, 廖博承, 姚正利
长江科学院院报 ›› 2026, Vol. 43 ›› Issue (10) : 137-146.
PDF(10688 KB)
PDF(10688 KB)
基于国产立体卫星影像的高原冰湖三维重建与蓄水量估计——以“7∙8”吉隆口岸冰湖溃决洪水为例
Three-dimensional Reconstruction and Water Volume Estimation for Plateau Glacial Lakes Using Chinese Stereo Satellite Imagery: A Case Study of the 8 July Glacial Lake Outburst in Gyirong
针对高原偏远无资料地区冰湖水量动态精准监测的难题,提出一种基于立体测绘卫星的水量估算方法。以2025年7月8日西藏吉隆溃决冰湖为例,利用高分七号(GF7)与资源三号(ZY3)立体影像重建湖盆数字表面模型(DSM),并结合2021—2025年哨兵二号影像自动提取多时相湖泊边界,通过边界与地形等高线最优匹配估计水位与蓄水量,同时采用ICESat-2激光测高数据验证地形精度。结果表明,GF7-DSM与ICESat-2之间的平均绝对误差为1.37 m;从湖盆几何形态上看,溃决前冰湖最大面积0.69 km2,最大水深53 m,湖盆底高程5 053.4 m,最大水位5 105.9 m;时序监测显示该冰湖存在明显“冬枯夏涨”现象,2025年7月7日蓄水量达历史极值约982万m3;溃决发生于7月7日12时52分—8日12时56分,24 h内下泄水量约424万m3,洪峰传播路径总长41.2 km、落差3 318 m、平均比降80.5‰,巨大势能转化与强水动力条件是导致灾害的直接原因。研究验证了国产立体测绘卫星在无资料地区冰湖动态水量监测中的有效性,可为高原冰湖溃决预警与灾害防控提供数据支撑。
[Objective] The scarcity of reliable hydrological data in the Qinghai-Xizang Plateau has long constrained effective disaster risk assessment in this area. To address the challenge of precise, dynamic water volume monitoring in data-scarce regions, we propose a water volume estimation method based on stereo mapping satellites. Taking the major Glacial Lake Outburst Floods (GLOF) event that occurred on July 8, 2025, in the Gyirong region of Xizang as a specific case study, we validated the application of domestic high-resolution stereo mapping satellites, in the fine reconstruction of glacial lake underwater topography and the inversion of water volume dynamics. [Methods] We utilized high-resolution stereo imagery acquired by the Gaofen-7 and Ziyuan-3 satellites to generate a Digital Surface Model (DSM). A core innovation of this research is the introduction of a Variational Mesh Refinement algorithm. By integrating photometric consistency and geometric smoothness constraints into a unified energy function, this algorithm overcomes the limitations of traditional Semi-Global Matching (SGM) methods in low-texture glacial regions, effectively reducing data noise and achieving the reconstruction of fine micro-geomorphological features. The automated workflow for water volume estimation include: first, multi-temporal lake boundaries were extracted from Sentinel-2 imagery spanning from 2021 to 2025; subsequently, a “Boundary-Contour” optimal matching strategy based on Intersection over Union (IoU) and Chamfer Distance was employed to match the 2D lake boundaries with the 3D DSM topographic contours. Simultaneously, high-confidence photon data from the ICESat-2 laser altimetry satellite was utilized to verify the topographic reconstruction accuracy through a rigorous 5-fold cross-validation process. [Results] The terrain reconstruction method proposed in this study demonstrates superior performance compared to mainstream commercial software packages such as ENVI and Metashape, as well as public datasets like SRTM. The Gaofen-7 DSM achieved the highest level of accuracy; when compared with ICESat-2 data, it yielded a Mean Absolute Error (MAE) of 1.37 m and a Root Mean Square Error (RMSE) of 1.62 m. Crucially, the method successfully restored critical geomorphic features, such as moraine ridges that were lost in other models. Morphological analysis indicates that prior to the outburst, the glacial lake possessed a maximum area of 0.69 km2, a maximum water depth of 53 m, and a lakebed elevation of 5 053.4 m. Time-series monitoring captured a distinct phenomenon of seasonal water volume surges, with the lake’s water storage reaching a historical peak of approximately 9.82 million m3 on July 7, 2025. The outburst event itself occurred within a 24-hour window from 12:52 on July 7 to 12:56 on July 8, during which the water level plummeted from 5 105.9 meters to 5 080 meters, releasing approximately 4.24 million m3 of floodwater. Flood evolution analysis revealed a total propagation path length of 41.2 km, a cumulative elevation drop of 3 318 m, and an average gradient of 80.5‰, confirming that the conversion of immense gravitational potential energy into hydrodynamic power was the direct cause of this disaster. [Conclusions] In summary, this study confirms that Chinese stereo mapping satellites, when combined with advanced mesh refinement algorithms, can achieve meter-level accuracy in topographic reconstruction within unmeasured plateau regions, providing a cost-effective and highly efficient technical solution for emergency monitoring. Geometric matching results indicate that even if the surface water body of the glacial lake appears fragmented, the subsurface water bodies maintain hydraulic connectivity. Based on the reconstruction of the water accumulation processes observed in 2023 and 2025, the study proposes critical risk warning thresholds for this specific glacial lake: immediate early warning measures should be initiated when the water level exceeds 5 104 m or the storage volume surpasses 4 million m3. The method proposed in this study builds a reproducible technical support system for regional glacial lake outburst flood risk assessment and holds significant importance for future disaster prevention and mitigation efforts across the Qinghai-Xizang Plateau.
水量监测 / 冰湖溃决洪水 / 星载立体测绘 / 高分七号卫星 / 应急监测
water volume monitoring / glacial lake outburst flood / satellite stereo mapping / Gaofen-7 satellite / emergency monitoring
| [1] |
Glacial lake outburst floods (GLOFs) represent a major hazard and can result in significant loss of life. Globally, since 1990, the number and size of glacial lakes has grown rapidly along with downstream population, while socio-economic vulnerability has decreased. Nevertheless, contemporary exposure and vulnerability to GLOFs at the global scale has never been quantified. Here we show that 15 million people globally are exposed to impacts from potential GLOFs. Populations in High Mountains Asia (HMA) are the most exposed and on average live closest to glacial lakes with ~1 million people living within 10 km of a glacial lake. More than half of the globally exposed population are found in just four countries: India, Pakistan, Peru, and China. While HMA has the highest potential for GLOF impacts, we highlight the Andes as a region of concern, with similar potential for GLOF impacts to HMA but comparatively few published research studies.© 2023. The Author(s).
|
| [2] |
\n We analyze and reconstruct a recent glacial lake outburst\nflood (GLOF) process chain on 26 June 2020, involving the moraine-dammed\nproglacial lake – Jinwuco (30.356\n ∘\n N, 93.631\n ∘\n E) in\neastern Nyainqentanglha, Tibet, China. Satellite images reveal that from\n1965 to 2020, the surface area of Jinwuco has expanded by 0.2 km\n 2\n (+56 %) to 0.56 km\n 2\n and subsequently decreased to 0.26 km\n 2\n (−54 %) after the GLOF. Estimates based on topographic reconstruction and\nsets of published empirical relationships indicate that the GLOF had a\nvolume of 10 million cubic meters, an average breach time of 0.62 h, and an\naverage peak discharge of 5602 m\n 3\n /s at the dam. Based on pre- and\npost-event high-resolution satellite scenes, we identified a large debris\nlandslide originating from western lateral moraine that was most likely\ntriggered by extremely heavy, south-Asian-monsoon-associated rainfall in\nJune 2020. We back-calculate part of the GLOF process chain, using the\nGIS-based open-source numerical simulation tool r.avaflow. Two scenarios are\nconsidered, assuming a debris-landslide-induced impact wave with overtopping\nand resulting retrogressive erosion of the moraine dam (Scenario A), as well as\nretrogressive erosion without a major impact wave (Scenario B). Both\nscenarios are in line with empirically derived ranges of peak discharge and\nbreach time. The breaching process is characterized by a slower onset and a\nresulting delay in Scenario B compared to Scenario A. Comparison of the\nsimulation results with field evidence points towards Scenario B, with a peak\ndischarge of 4600 m\n 3\n /s. There were no casualties from\nthis GLOF, but it caused severe destruction of infrastructure (e.g., roads and\nbridges) and property losses in downstream areas. Given the clear role of\ncontinued glacial retreat in destabilizing the adjacent lateral moraine\nslopes and directly enabling the landslide to deposit into the expanding\nlake body, the GLOF process chain can be plausibly linked to anthropogenic\nclimate change, while downstream consequences have been enhanced by the\ndevelopment of infrastructure on exposed flood plains. Such process chains\ncould become more frequent under a warmer and wetter future climate, calling\nfor comprehensive and forward-looking risk reduction planning.\n
|
| [3] |
许继军, 朱振熔, 魏丹莉, 等. 西藏高原分区域河流健康评价体系应用[J]. 长江科学院院报, 2025, 42(10): 192-200, 218.
西藏高原地域辽阔、自然条件严苛且河流对气候变化高度敏感,为保障区域水资源安全和生态系统稳定,亟需开展区域化、多维度的河流健康评价,构建一套综合健康评价体系,包含自然结构、生态环境和社会服务功能3个准则层及12项指标。根据西藏的地形和气候特征,划分研究区域为藏南河谷地区、藏北高原湖盆区和藏东高山峡谷区,并针对各区域制定了差异化评价标准。采用博弈论的组合赋权方法对指标权重进行优化,选取金沙江西藏段、年楚河和扎加藏布3条典型河流进行评价分析。结果表明,扎加藏布得分为95.51分,处于“非常健康”状态;金沙江西藏段得分为91.85分,建议加强堤防建设;年楚河得分为81.57分,健康状况相对较弱,主要面临水土流失和连通性不足等问题。同时,基于评价结果提出加强水土保持、完善生态流量泄放以及推进防洪工程等措施建议。
(
|
| [4] |
|
| [5] |
郭少华, 向建明, 孙铁钢, 等. 基于SAR影像的冰湖季节性演化自动提取方法与气候响应分析[J]. 测绘, 2025, 48(2): 108-114.
(
|
| [6] |
陈喆, 向大享, 姜莹, 等. 基于国产高分三号星座的洪涝灾害应急监测模式研究: 以洞庭湖区团洲垸溃口险情为例[J]. 长江科学院院报, 2024, 41(12): 189-195, 201.
针对如何在应急条件下开展快速有效的洪涝灾情动态监测预警评估,提出了一种基于时序国产高分三号星座数据的洪涝灾害应急监测模式,探索汛期洪涝溃堤灾前预警、灾中动态监测、灾后损失评估的监测方法和处理流程。基于该工作模式,针对2024年7月5日发生的洞庭湖团洲垸溃口险情,利用溃口决堤前后逐日国产高分三号卫星影像时序数据,迅速开展团洲垸受灾区域应急遥感监测。根据水体提取结果和洪涝受灾前后变化检测结果,发现:决堤发生前一天溃口周围农田区域水体面积明显增大,2024年7月4日上午10时至7月5日上午11时,堤防渗水过程明显;2024年7月6日上午团洲垸基本全部被淹,仅剩主要干道两侧房屋屋顶在水面以上;7月8日至15日,团洲垸呈缓慢退水状态。监测结果表明,基于国产高分三号星座数据可实现高时空分辨率的溃堤前—溃堤中—溃口合拢后团洲垸灾情变化全过程监测,为溃堤灾前预警、灾中应急抢险、灾后损失评估提供有效的数据支撑。
(
|
| [7] |
|
| [8] |
|
| [9] |
|
| [10] |
|
| [11] |
夏天, 严薇, 赵英芬, 等. 资源一号02D卫星数据应用[J]. 卫星应用, 2021(11):54-60.
(
|
| [12] |
《中国河湖大典》编纂委员会. 中国河湖大典[M]. 北京: 中国水利水电出版社, 2014.
(Compilation committee of “China’s rivers and lakes canon”. China’s rivers and lakes canon[M]. Beijing: China Water & Power Press, 2014. (in Chinese))
|
| [13] |
中国新闻网. 西藏吉隆口岸泥石流致多人失联中尼热索大桥垮塌中断[EB/OL]. (2025-07-08)[2025-08-11].
(China News Service. A landslide at Gyirong Port in Tibet has caused many people to go missing and the China-Nepal Rasu Bridge has collapsed, causing a disruption [EB/OL]. (2025-07-08) [2025-08-11]. (in Chinese))
|
| [14] |
|
| [15] |
Automatic reconstruction of surfaces from satellite imagery is a hot topic in computer vision and photogrammetry. State-of-the-art reconstruction methods typically produce 2.5D elevation data. In contrast, we propose a one-stage method directly generating a 3D mesh model from multi-view satellite imagery. We introduce a novel Sat-Mesh approach for satellite implicit surface reconstruction: We represent the scene as a continuous signed distance function (SDF) and leverage a volume rendering framework to learn the SDF values. To address the challenges posed by lighting variations and inconsistent appearances in satellite imagery, we incorporate a latent vector in the network architecture to encode image appearances. Furthermore, we introduce a multi-view stereo constraint to enhance surface quality. This constraint minimizes the similarity between image patches to optimize the position and orientation of the SDF surface. Experimental results demonstrate that our method achieves superior visual quality and quantitative accuracy in generating mesh models. Moreover, our approach can learn seasonal variations in satellite imagery, resulting in texture mesh models with different and consistent seasonal appearances.
|
| [16] |
|
| [17] |
This paper describes the Semi-Global Matching (SGM) stereo method. It uses a pixelwise, Mutual Information based matching cost for compensating radiometric differences of input images. Pixelwise matching is supported by a smoothness constraint that is usually expressed as a global cost function. SGM performs a fast approximation by pathwise optimizations from all directions. The discussion also addresses occlusion detection, subpixel refinement and multi-baseline matching. Additionally, postprocessing steps for removing outliers, recovering from specific problems of structured environments and the interpolation of gaps are presented. Finally, strategies for processing almost arbitrarily large images and fusion of disparity images using orthographic projection are proposed.A comparison on standard stereo images shows that SGM is among the currently top-ranked algorithms and is best, if subpixel accuracy is considered. The complexity is linear to the number of pixels and disparity range, which results in a runtime of just 1-2s on typical test images. An in depth evaluation of the Mutual Information based matching cost demonstrates a tolerance against a wide range of radiometric transformations. Finally, examples of reconstructions from huge aerial frame and pushbroom images demonstrate that the presented ideas are working well on practical problems.
|
| [18] |
徐涵秋. 利用改进的归一化差异水体指数(MNDWI)提取水体信息的研究[J]. 遥感学报, 2005, 9(5): 589-595.
(
|
| [19] |
|
| [20] |
The lack of ground control points (GCPs) affects the elevation accuracy of digital surface models (DSMs) generated by optical satellite stereo images and limits the application of high-resolution DSMs. It is a feasible idea to use ICESat-2 (Ice, Cloud, and land Elevation Satellite-2) laser altimetry data to improve the elevation accuracy of optical stereo images, but it is necessary to accurately match the two types of data. This paper proposes a DSM registration strategy based on terrain similarity (BOTS), which integrates ICESat-2 laser altimetry data without GCPs and improves the DSM elevation accuracy generation from optical satellite stereo pairs. Under different terrain conditions, Worldview-2, SV-1, GF-7, and ZY-3 stereo pairs were used to verify the effectiveness of this method. The experimental results show that the BOTS method proposed in this paper is more robust when there are a large number of abnormal points in the ICESat-2 data or there is a large elevation gap between DSMs. After fusion of ICESat-2 data, the DSM elevation accuracy extracted from the satellite stereo pair is improved by 73~92%, and the root mean square error (RMSE) of Worldview-2 DSM reaches 0.71 m.
|
感谢自然资源部国土卫星遥感应用中心为本研究提供的GF7和ZY3卫星数据。感谢王珺珂在文字校核中提供的帮助。
/
| 〈 |
|
〉 |
