无/缺水下地形数据的高原堰塞湖水量遥感估算

Dammed lake water volume estimation by satellite imagery and digital elevation model under unknown underwater terrain scenario

水量遥感动态监测对于高原堰塞湖风险评估、预报预警和处置决策等具有重要意义。针对高原无资料或缺资料区,充分利用空天遥感技术,文章提出了一种无/缺水下地形数据的高原堰塞湖水量遥感定量估算方法。该方法首先通过遥感水域面积提取,获取堰塞湖淹没空间范围;进而采用不规则复杂多边形中线定位算法,确定堰塞湖中心线位置;然后基于河道中心特定点高程信息,结合局部河道比降估算,生成堰塞湖水下地形河道中线约束因子;再根据河道边坡高程信息和水下地形约束因子自适应拟合出局部堰塞河道的水下未知地形;最后通过三维曲面离散积分实现堰塞湖水量遥感动态定量估算。实验以东帕米尔高原的萨雷兹堰塞湖为研究区,展开遥感水量调查与局部验证研究,结果表明:萨雷兹堰塞湖当前水域面积约为89.09 km^(2),水量约为162.49亿m^(3);这一结果与专家预估的水资源量155—165亿m^(3)基本吻合。经局部模拟实验精度对比验证,模拟结果与实际数据动态误差总体控制在10%以内,相关系数达到0.95(P<0.01,双尾),进一步证明了算法的鲁棒性和估算结果的可信度。为无/缺水下地形数据的高原堰塞湖水量遥感估算提供了一种有效的方法,实现了水下地形未知的高原堰塞湖水量遥感快速反演与定量测算。

The real-time dynamic monitoring of water volume has great value for risk assessment,prediction,and early warning and disposal decision-making of dammed lakes.In view of the difficulty in obtaining underwater topographic data of dammed lake in areas without gauged data on the plateau,directly,quantitatively,and timely estimating the water volume of dammed lake by using remote sensing technology is difficult.This study aims to solve the problem of rapid quantitative estimation of water volume of plateau dammed lake under unknown underwater terrain scenario by remote sensing and perform the risk monitoring and disaster assessment of dammed lake.According to the existing remote sensing data and digital elevation information,this study puts forward a remote sensing quantitative estimation method of water volume of dammed lake on the plateau without underwater terrain by fully using remote sensing data.The details are as follows.First,the submerged area of dammed lake is extracted from remote sensing images.Second,the center line of the complex polygon of the dammed lake is calculated.Specifically,according to the water area of the dammed lake,the Tyson polygon algorithm is used to calculate the position information of the dammed lake centerline.Third,through the location of the polygonal center line of the dammed lake,the fixed-point elevation measurement is carried out to complete the fitting calculation.Then,according to the fitting estimation of the middle line elevation and combined with the slope elevation information,the unknown underwater terrain of the dammed lake is adaptively simulated.Finally,based on the simulated underwater terrain and the submerged area of the dammed lake,the capacity of the dammed lake is calculated by 3D curved surface space discrete integration.The dammed lake,namely,Sarez Lake in the Pamirs was selected as the research area.Remote sensing survey and empirical research on water volume were carried out using the proposed method.The research results show that the water area of Sarez dammed lake is approximately 89.09 km^(2),and the water volume of Sarez Lake is approximately 16.25 billion m^(3).This result is consistent with the expert’s estimated water volume of 15.5 billion m^(3) to 16.5 billion m^(3).The accuracy verification of the local simulation experiment shows that the overall dynamic error between the simulation data and the measured data is controlled within 10%,and the correlation coefficient is 0.95(P<0.01,double tailed).This finding further proves the robustness of the algorithm and the credibility of the estimation results.This method can rapidly estimate the water volume of plateau dammed lake,with high accuracy and strong technical universality.It provides an efficient method for remote sensing estimation of plateau dammed lake water volume under none or lack of underwater terrain data scenario.It also solves the problem of quantitative calculation of plateau dammed lake water volume with unknown underwater terrain.