The coordinated development of the Beijing-Tianjin-Hebei has become a national strategy with Beijing and Tianjin as twin engines driving the regional development. However, the Beijing-Tianjin-Hebei region has suffered dramatic ground subsidence during last two to three decades, mainly due to long-term groundwater withdrawal. Although, annual spirit leveling has been conducted routinely in some parts of Beijing and Tianjin, and InSAR technique has also been used to monitor ground subsidence in some local areas of the region, there is a lack of a complete survey of ground subsidence over the whole region. In this paper, we report a research on mapping ground subsidence in the Beijing-Tianjin-Hebei region over a long time span from 1992 to 2014. Three SAR datasets from four satellites are used: ERS-1/2 SAR images from 1992 to 2000, ENVISAT ASAR images from 2003 to 2010, and RADARSAT-2 images from 2012 to 2014. An improved multi-temporal InSAR method, namely “Multiple-master Coherent Target Small-Baseline InSAR” (MCTSB-InSAR), has been developed to process the datasets. A unique feature of MCTSB-InSAR is the adjustment process useful for wide area monitoring which provides an integrated solution for both calibration of InSAR-derived deformation and the harmonization of the deformation estimates from overlapping SAR frames. Three maps of the subsidence rate corresponding to the three periods over the wide Beijing-Tianjin-Hebei region are generated, with respective accuracy of 8.7 mm/year (1992–2000), 4.7 mm/year (2003–2010), and 5.4 mm/year (2012–2014) validated by more than 120 leveling measurements. The spatial-temporal characteristics of the development of ground subsidence in Beijing and Tianjin are analyzed. This research represents a first-ever effort on mapping ground subsidence over very large area and over long time span in China. The result is of significance to serve the decision-making on ground subsidence mitigation in the Beijing-Tianjin-Hebei region.

Interferometric synthetic aperture radar (InSAR) mapping of localized ground surface deformation has become an important tool to manage subsidence-related geohazards. However, monitoring land surface deformation using InSAR at high spatial resolution over a large region is still a formidable task. In this paper, we report a research on investigating ground subsidence and the causes over the entire 107, 200 km² province of Jiangsu, China, using time-series InSAR. The Sentinel-1 Interferometric Wide-swath (IW) images of 6 frames are used to map ground subsidence over the whole province for the period 2016–2018. We present processing methodology in detail, with emphasis on the three-level co-registration scheme of S-1 data, retrieval of mean subsidence velocity (MSV) and subsidence time series, and mosaicking of multiple frames of results. The MSV and subsidence time series are generated for 9,276,214 selected coherent pixels (CPs) over the Jiangsu territory. Using 688 leveling measurements in evaluation, the derived MSV map of Jiangsu province shows an accuracy of 3.9 mm/year. Moreover, subsidence causes of the province are analyzed based on InSAR-derived subsidence characteristics, historical optical images, and field-work findings. Main factors accounting for the observed subsidence include: underground mining, groundwater withdrawal, soil consolidations of marine reclamation, and land-use transition from agricultural (paddy) to industrial land. This research demonstrates not only the capability of S-1 data in mapping ground deformation over wide areas in coastal and heavily vegetated region of China, but also the potential of inferring valuable knowledge from InSAR-derived results.

Sentinel-1 synthetic aperture radar (SAR) imagery, with its wide-swath imaging capabilities, provides a rich data source for large-area interferometric synthetic aperture radar (InSAR) surface deformation monitoring. Precise image coregistration is a prerequisite for InSAR data processing and is one of the most time-consuming steps. Currently, image coregistration uses open multi-processing (OpenMP) to achieve algorithm parallelism for interferometric image pairs. However, owing to the large swath and substantial data volume of Sentinel-1 images, as well as the increasing number of time series images, the efficiency of this algorithm in time series image coregistration is inadequate for practical production needs. To address this issue, this paper proposes a multi-segment parallel coregistration method for Sentinel-1 time series images by combining OpenMP and message passing interface (MPI). During the geometric coregistration process, OpenMP and MPI are used to realize algorithm-level parallel processing and multitask parallel processing, significantly improving coregistration efficiency. In the enhanced spectral diversity coregistration process, all slave images are segmented for coregistration on the basis of the temporal baseline. In each segment, the first image is processed serially, and the other images are processed in parallel. Thirty Sentinel-1 SAR images covering the Tianjin–Tangshan area (plain region) and 27 from Zhejiang Province (mountainous region) are used to validate the proposed method. The results demonstrate that 1) when the number of parallel kernels is 10 or 12, OpenMP can achieve optimal efficiency for algorithmic parallelism in geometric coregistration and enhanced spectral diversity coregistration. 2) The combined OpenMP and MPI parallel method can achieve optimal efficiency with 16 parallel tasks for geometric coregistration and 24 parallel tasks for enhanced spectral diversity coregistration. 3) In the plain region, the proposed method improves the processing speed by 8.2 times and 6.7 times in geometric coregistration and enhanced spectral diversity coregistration, respectively, compared with OpenMP alone. In the mountainous region, it improves the processing speed by 9.7 times and 4.5 times in geometric coregistration and enhanced spectral diversity coregistration, respectively.

Over the past several decades, synthetic aperture radar interferometry (InSAR) has revolutionized surface deformation mapping from space. However, tropospheric artifacts—manifested as turbulent delays (TDs) and vertically stratified delays (VSDs)—remain a critical bottleneck limiting the accuracy of InSAR measurements, as a consequence of the spatiotemporal dynamics of atmospheric properties (e.g., temperature, humidity, and pressure). Current tropospheric correction methods primarily target either TDs or VSDs in isolation, which significantly impedes the complete mitigation of tropospheric noise in interferograms. To address this limitation, in this article, a joint spatio-temporal correction (JSTC) method was proposed for mitigating both TDs and VSDs. In this method, first, a dual-scale temporal low-pass filtering (DTLF) strategy is employed to suppress TDs, leveraging their high-frequency characteristics in the temporal domain. Improved quadtree segmentation is subsequently applied to account for the spatial heterogeneity of VSDs. The JSTC method is validated using both simulated datasets and real InSAR observations in the rugged Hengduan Mountains of southwestern China. The experimental results demonstrate that the JSTC outperforms conventional approaches such as Generic Atmospheric Correction Online Service, global linear, and DTLF, reducing the phase standard deviation by 83.82%, 62.07%, and 84.40%, respectively. This study provides a robust solution for increasing the reliability of InSAR-based deformation monitoring under complex atmospheric conditions.

Time-series interferometric synthetic aperture radar (TS-InSAR) has been extensively applied to ground deformation monitoring and disaster risk assessment. In this process, coherent point (CP) selection and connection are two essential steps. Typically, CPs are initially connected using methods such as Delaunay triangulation, followed by the estimation of deformation parameters for each edge. However, the removal of low-quality edges often fragments the initial network into multiple subnetworks, and reliable deformation estimation requires reconnecting these subnetworks into a single, fully connected network. Despite its importance, this problem remains challenging and has not been fully addressed. In this article, we propose an efficient subnetwork connection method, termed KD-tree with iterative expansion (KDIE), which integrates a K-dimension (KD)-tree-based distance searching strategy with an iterative expansion scheme. KDIE efficiently identifies connection edges using the KD-tree and progressively extends the search distance (SD) up to the atmospheric correlation distance (ACD), thereby avoids missing any connectable subnetworks and preferentially connects nearly all subnetworks with short edges. Compared with the existing multilayer subnetwork connection (MLSC) method, two experiments concerning C-band Sentinel-1A images over a mountainous landscape and X-band TerraSAR-X images over an urban area demonstrate that KDIE connects 1.84% and 0.25% more CPs than the MLSC method, while reducing computational time to only 5%–8% of that required by the MLSC method.