SBAS-InSAR技术支持下的滇西北宾川断陷盆地沉降监测与预测

Monitoring and Prediction of Land Subsidence in Binchuan Fault Basin of Northwest Yunnan Supported by SBAS-InSAR

滇西北宾川断陷盆地一带新构造运动活跃,滑坡、地震现象时常发生.针对该区域内复杂的地质结构及缺乏大面积的有效监测等问题,采用SBAS-InSAR技术处理覆盖宾川县断陷盆地的58景2019年1月7日—2020年12月15日Sentinel-1A影像,首先获取研究区内的整体地表形变沉降速率场,然后分析了沉降区的时空分布特征及5个典型的沉降漏斗区,并对沉降区的成因进行探究;最后随机提取沉降最严重的A区中的639个沉降点,选取其中的609个点作为BP神经网络模型的学习样本,对剩余的30个沉降监测点进行预测,并将预测结果与SBAS-InSAR监测值进行对比.结果表明:(1)宾川县断陷盆地区最大沉降速率为-117.89 mm/a,位于州城镇第二中学片区;(2)研究区内的地表沉降主要与地质结构、地下水的抽取及人类活动等有关;(3)BP神经网络模型的预测结果与SBAS-InSAR监测值的最大误差为2.451 mm,均方误差为0.18 mm,表明BP神经网络模型能够很好地预测县级城市的地表沉降.

The Neotectonic movement is active in Binchuan fault basin in northwestern Yunnan,and the landslides and earthquakes happen frequently.Aiming at the complex geological structure and the lack of effective monitoring in the area,SBAS-InSAR technology was applied to process 58 Sentinel-1 A SAR images covering Binchuan fault basin from January 7,2019 to December 15,2020.First,the overall surface deformation settlement rates was extracted.Then the distribution characteristics of the settlement area and five typical settlement areas were analyzed,and the cause of the settlement area was explored.Finally,639 settlement points in the most severe settlement area A were randomly selected,the 609 points were selected as the learning samples of the BP neural network model,the remaining 30 settlement points were predicted,and the prediction results were compared with the SBAS-InSAR monitoring values.The results showed that:(1)The maximum subsidence rate in the fault basin area of Binchuan County was-117.89 mm/a,which was located in the second middle school area of Zhoucheng Town;(2)The surface subsidence in the study area was mainly related to the geological structure,the extraction of groundwater and human activities;(3)The maximum error between the prediction result of the BP neural network model and the SBAS-InSAR monitoring value was 2.451 mm,and the mean square error was 0.18 mm,indicating that the BP neural network model can predict the land subsidence of county-level cities well.