Ruren Li

Land subsidence, the superposition of effects from natural environments and human activities, has posed substantial threats to sustainable development in mining areas. Monitoring land subsidence is particularly crucial. However, most monitoring models ignore the combined effects of subsidence‐induced factors, making it difficult to accurately capture the spatiotemporal heterogeneity of subsidence. To address this challenge, this study proposed a GeoDistillBERT‐BiLSTM framework for time‐series land subsidence monitoring by fusing multimodal features. Firstly, subsidence patterns from 2020 to 2022 were assessed through Small Baseline Subset InSAR (SBAS‐InSAR). Then, the Pressure‐State‐Response (PSR) framework was used to screen representative factors influencing subsidence from three aspects: human activities, natural environment, and subsidence velocities. Subsequently, a GeoDistillBERT‐BiLSTM framework was developed that combines a geological modal attention mechanism with a modality‐aware adaptive loss function. It could dynamically assign importance coefficients to different factors by generating spatiotemporal weight matrices, facilitating the fusion of multimodal features for time‐series monitoring, thus revealing the spatiotemporal changes of subsidence. Based on this, Geodetector is introduced to explore mechanisms behind subsidence dynamics. This study used the Dagushan open‐pit mine as a case study; it demonstrated that the GeoDistillBERT‐BiLSTM framework improves the fusion accuracy of multimodal features by establishing spatiotemporal weight matrices. Compared with conventional statistical models, the loss function L _total reduced by 2%. Compared with other factors, human activities and their interaction with NDVI played a key role in subsidence mechanisms. This study provides a novel approach that integrates multiple factors and spatiotemporal correlations for time‐series subsidence monitoring, helping to prevent geological hazards in mining areas.

The advent of precision agriculture has highlighted the necessity for the careful determination of crop phenology at increasingly smaller scales. Although remote sensing technology is extensively employed for the monitoring of crop growth, the acquisition of high-precision phenological data continues to present a significant challenge. This study, conducted in Youyi County, Shuangyashan City, Heilongjiang Province, China, employed time-series spectral index data derived from Sentinel-2 remote sensing images to investigate methodologies for the extraction of pivotal phenological phases during the primary growth stages of maize. The data were subjected to Savitzky–Golay (S-G) filtering and cubic spline interpolation in order to denoise and smooth them. The combination of dynamic thresholding with slope characteristic node recognition enabled the successful extraction of the jointing and tasseling stages of maize. Furthermore, a comparison of the extraction of phenophases based on the time-series curves of the NDVI, EVI, GNDVI, OSAVI, and MSR was conducted. The results showed that maize exhibited different sensitivities to the spectral indices during the jointing and tasseling stages: the OSAVI demonstrated the highest accuracy for the jointing stage, with a mean absolute error of 3.91 days, representing a 24.8% improvement over the commonly used NDVI. For the tasseling stage, the MSR was the most accurate, achieving an absolute error of 4.87 days, with an 8.6% improvement compared to the NDVI. In this study, further analysis was conducted based on maize cultivation data from Youyi County (2021–2023). The results showed that the maize phenology in Youyi County in 2021 was more advanced compared to 2022 and 2023, primarily due to the higher average temperatures in 2021. This study provides valuable support for the development of precision agriculture and maize phenology monitoring and also provides a useful data reference for future agricultural management.

Ground subsidence often occurs over a large area. Although traditional monitoring methods have high accuracy, they cannot perform wide-area ground deformation monitoring. Synthetic Aperture Radar (SAR) interferometry (InSAR) technology utilizes phase information in SAR images to extract surface deformation information in a low-cost, large-scale, high-precision, and high-efficiency manner. With the increasing availability of SAR satellite data and the rapid development of InSAR technology, it provides the possibility for wide-area ground deformation monitoring using InSAR technology. Traditional time-series InSAR methods have cumbersome processing procedures, have large computational requirements, and rely heavily on manual intervention, resulting in relatively low efficiency. This study proposes a strategy for wide-area InSAR multi-scale deformation monitoring to address this issue. The strategy first rapidly acquires ground deformation information using Stacking technology, then identifies the main subsidence areas by setting deformation rate thresholds and visual interpretation, and finally employs advanced TS-InSAR technology to obtain detailed deformation time series for the main subsidence areas. The Turpan Basin in Xinjiang, China, was selected as the study area (7474.50 km²) to validate the proposed method. The results are as follows: (1) The basin is generally stable, but there is ground subsidence in the southern plain area, mainly affecting farmland. (2) From 2016 to 2019, the maximum subsidence rate in the farmland area was approximately 0.13 m/yr, with a maximum cumulative subsidence of about 0.25 m, affecting a total area of approximately 952.49 km². The subsidence mainly occurred from late spring to mid-autumn, while lifting or subsidence mitigation occurred from late autumn to early spring. The study also analyzed the impacts of rainfall, geographical environment, and human activities on subsidence and found that multiple factors, including water resource reduction, overexploitation, geological characteristics, and the expansion of human activities, contributed to the subsidence problem in the Turpan Basin. This method contributes to wide-area InSAR deformation monitoring and the application of InSAR technology in engineering.

Aquatic vegetation is an important component of aquatic ecosystems; therefore, the classification and mapping of aquatic vegetation is an important aspect of lake management. Currently, the decision tree (DT) classification method based on spectral indices has been widely used in the extraction of aquatic vegetation data, but the disadvantage of this method is that it is difficult to fix the threshold value, which, in turn, affects the automatic classification effect. In this study, Sentinel-2 MSI data were used to produce a sample set (about 930 samples) of aquatic vegetation in four inland lakes (Lake Taihu, Lake Caohai, Lake Honghu, and Lake Dongtinghu) using the visual interpretation method, including emergent, floating-leaved, and submerged vegetation. Based on this sample set, a DL model (Res-U-Net) was used to train an automatic aquatic vegetation extraction model. The DL model achieved a higher overall accuracy, relevant error, and kappa coefficient (90%, 8.18%, and 0.86, respectively) compared to the DT method (79%, 23.07%, and 0.77) and random forest (78%,10.62% and 0.77) when utilizing visual interpretation results as the ground truth. When utilizing measured point data as the ground truth, the DL model exhibited accuracies of 59%, 78%, and 91% for submerged, floating-leaved, and emergent vegetation, respectively. In addition, the model still maintains good recognition in the presence of clouds with the influence of water bloom. When applying the model to Lake Honghu from January 2017 to October 2023, the obtained temporal variation patterns in the aquatic vegetation were consistent with other studies. The study in this paper shows that the proposed DL model has good application potential for extracting aquatic vegetation data.

The spatiotemporal variations in suspended sediment concentration (SSC) in the lower reaches of the Yellow River exhibit significant variability and are influenced by reservoir operations. Understanding the spatiotemporal distribution characteristics of SSC in water holds crucial implications for environmental protection and reservoir operation management. Based on daily-scale SSC monitoring data from four hydrological stations in the lower Yellow River, this study established an SSC remote sensing model applicable to Landsat series satellite data. The independent variable of the model, Rrs(NIR)/(Rrs(G) + Rrs(R) + Rrs(SWIR)), demonstrated sensitivity to water bodies with different SSC values. Distinctive spatiotemporal characteristics in sediment transport were observed across the lower Yellow River. Spatially, the SSC values in the Sanmenxia and Xiaolangdi reservoirs were notably lower than those in other river sections, averaging 1008.42 ± 602.83 mg/L and 1177.89 ± 627.95 mg/L, respectively. Over time, the majority of the river sections (96%) exhibited decreasing trends in SSC during 1984–2022, particularly in the downstream Xiaolangdi reservoir, with average SSC values of 4265.58 ± 1101.77 mg/L in the 1980s and 1840.80 ± 2255.15 mg/L in the 2020s. Seasonal variations in SSC were prominent, with higher summer concentrations, averaging 5536.43 ± 2188.77 mg/L (2020s summer) and 814.11 ± 158.27 mg/L (2020s winter). Reductions in SSC during 1984–2022 primarily occurred in summer, weakening its seasonal variability in the lower Yellow River. Water discharge emerged as a critical factor influencing suspended sediment transport, with SSC increasing in high-water-flow months. Following the construction of the Xiaolangdi reservoir, the relationship between SSC and water discharge at different stations underwent notable alterations. This study enhances our understanding of the spatiotemporal dynamics of suspended sediment transport in the lower Yellow River, providing valuable insights for utilizing long-term Landsat series data in the dynamic monitoring of river sediment transport.

Integration development of urban agglomeration is important for regional economic research and management. In this paper, a method was proposed to study the integration development of urban agglomeration by trajectory gravity model. It can analyze the gravitational strength of the core city to other cities and characterize the spatial trajectory of its gravitational direction, expansion, etc. quantitatively. The main idea is to do the fitting analysis between the urban axes and the gravitational lines. The correlation coefficients retrieved from the fitting analysis can reflect the correlation of two indices. For the different cities in the same year, a higher value means a stronger relationship. There is a clear gravitational force between the cities when the value above 0.75. For the most cities in different years, the gravitational force between the core city with itself is increasing by years. At the same time, the direction of growth of the urban axes tends to increase in the direction of the gravitational force between cities. There is a clear tendency for the trajectories of the cities to move closer together. The proposed model was applied to the integration development of China Liaoning central urban agglomeration from 2008 to 2016. The results show that cities are constantly attracted to each other through urban gravity.

Street greening, an indispensable element of urban green spaces, has played an important role in beautifying the environment, alleviating the urban heat island effect, and improving residents’ comfort. Vegetation coverage is a common index used for measuring street greening. However, there are some shortcomings in the traditional evaluation methods of vegetation coverage. Part of the vegetation coverage cannot be determined from a two‐dimensional perspective, such as shrubs and green walls. In this paper, the Sentinel‐2 image was used to extract the street fractional vegetation cover (SFVC) and the Baidu street view panoramas were used to extract the green view index (GVI). To overcome the lack of a single perspective from the street vegetation coverage evaluation, the above two indices were merged to construct a comprehensive street greening evaluation index (CSGEI). The research area is the Longhua District of Shenzhen city in Southern China. All three indices were divided into five classes using natural breakpoint methods based on previous research experience. The results showed that Baidu street view panoramas could effectively identify shrubs and green walls that were deficient in the Sentinel‐2 image. The GVI is a supplement to the street vegetation coverage. The SFVC and GVI were divided into five classes, from L1 to L5 implying a gradual increase in the percentage of the vegetated area. The result has shown that the SFVC was in the L1, accounting for 53.68%. After index merging, the process of accounting for the L1 decreased to 31.29%. The multiperspective integrated CSGEI could comprehensively measure the distribution information of street greening and guide the planning and management of urban green landscapes.

Urban greenness plays a vital role in supporting the ecosystem services of a city. Exploring the dynamics of urban greenness space and their driving forces can provide valuable information for making solid urban planning policies. This study aims to investigate the dynamics of urban greenness space patterns through landscape indices and to apply geographically weighted regression (GWR) to map the spatially varied impact on the indices from economic and environmental factors. Two typical landscape indices, i.e., percentage of landscape (PLAND) and aggregation index (AI), which measure the abundance and fragmentation of urban greenness coverage, respectively, were taken to map the changes in urban greenness. As a case study, the metropolis of Wuhan, China was selected, where time-series of urban greenness space were extracted at an annual step from the Landsat collections from Google Earth Engine during 2000–2018. The study shows that the urban greenness space not only decreased significantly, but also tended to be more fragmented over the years. Road network density, normalized difference built-up index (NDBI), terrain elevation and slope, and precipitation were found to significantly correlate to the landscape indices. GWR modeling successfully captures the spatially varied impact from the considered factors and the results from GWR modeling provide a critical reference for making location-specific urban planning.

The ever-rising concentration of atmospheric carbon is viewed as the primary cause for global warming. To discontinue this trend, it is of urgent importance to either cut down human carbon emissions or remove more carbon from the atmosphere. Grassland ecosystems occupy the largest part of the global land area but maintain a relatively low carbon sequestration flux. While numerous studies have confirmed the impacts on grassland vegetation growth from climate changes and human activities, little work has been done to understand the driving forces for a reduced carbon uptake (RCU)—a loss in vegetation carbon sequestration because of inappropriate grassland management. This work focused on assessing RCU in the grassland of Inner Mongolia and understanding the influential patterns of the selected variables (including grazing intensity, road network, population, and vegetation productivity) related to RCU. Neighborhood analysis was proposed to locate optimized grassland management practices from historical data and to map RCU. Ordinary least squares (OLS) and geographically weighted regression (GWR) models were applied to explore the driving forces for RCU. The results indicated that the human-related factors, including stock grazing intensity, population density, and road network were likely to present a spatially varied impact on RCU, which accounted for more than 1/4 of the total carbon sequestration.

This letter proposes a distance-constrained (DC) zonal analysis approach to quantify how much more carbon could be further sequestrated by vegetation in mainland China based on multiple data sources. Our approach first segments the area into homogeneous landform–vegetation–soil (LVS) zones. Good land management practice (GLMP) corresponding to high sequestrated carbon (target carbon level) is identified at the locations in the same LVS zone. The target carbon level is set as the 90th percentile of the historically sequestrated carbon using the proxy of net primary productivity (NPP) at the locations within the LVS zone. When GLMP is realized over the entire LVS zone, more carbon could be sequestrated. Our results show that on average about 1/4 of more carbon could be added to the existing amount given the selected “good” land management practices are adopted by neighboring locations where lower carbon sequestration levels exist. The carbon sequestration potential for different land cover types differs significantly.