Dong Li

Rice extraction in complex landscapes is a challenging issue in remote sensing, particularly in areas with diverse land use types and spatiotemporal variability. To enhance the accuracy of rice extraction, this study proposes a novel approach integrating Sentinel-2 spectral-temporal-spatial imagery with a hybrid deep learning architecture for extracting single-cropping and double-cropping rice. Firstly, a time-series dataset of spectral and texture features was constructed to capture the seasonal variations of rice. Secondly, an active learning strategy was employed to select high-confidence samples, and spectral, temporal, and spatial information was integrated into a unified dataset. Finally, a hybrid deep learning model, Convolutional-Transformer Hybrid Network (CTH-Net), was developed, combining Convolutional Neural Networks (CNN) and Transformer networks. The model incorporates a fusion module to effectively integrate multi-scale temporal, spatial, and spectral features and a residual module to improve gradient flow, mitigating the vanishing gradient problem in deep networks. Results demonstrate that CTH-Net achieved 99.69% overall accuracy in rice extraction, maintaining >96% accuracy for single-cropping rice, double-cropping rice, and abandoned land. It outperformed models like CNN, Transformer, Long Short-Term Memory (LSTM), and Support Vector Machine (SVM) in handling fragmented rice distributions and mixed land types, significantly improving extraction accuracy. This study provides an efficient and reliable solution for rice extraction in complex landscapes, supporting agricultural monitoring and management.

In the context of intense interference from human activities and massive loss of natural wetlands in coastal zones, it is crucial to understand the behavioral ecology of shorebirds for formulating targeted conservation measures. Based on satellite tracking and remote sensing techniques, our research investigated the activity pattern and habitat selection characteristics of the Pied Avocet (Recurvirostra avosetta ) in Bohai Bay, China. The results showed that Pied Avocets are relatively flexible in habitat selection in Bohai Bay. There are obvious individual differences in their residence time (119–210 days) and activity range (core home range from 15.34 to 95.12 km²). Pied Avocets may only move around a fixed breeding site throughout the breeding season, or they may transfer to another location for a second breeding. The mariculture, salt pan, and industrial‐mining land are the main components of the Pied Avocets' habitat, and the proportion of their area in the main and core home ranges is > 10%. The species prefers sparsely vegetated areas near coastlines and water bodies as habitats and has a certain tolerance for human disturbance. Our findings could provide specific management advice for alleviating human‐bird conflicts in the highly developed coastal zones. The study on the activity pattern and habitat selection of the Pied Avocet offered technical and data support for shorebirds habitat protection.

Previous studies of vegetation dynamics in the Yellow River Delta (YRD) predominantly relied on sparse time series or coarse-resolution images, which not only overlooked the rapid and spatially heterogeneous changes, but also limited our understanding of driving mechanisms. Here, employing spatiotemporal data fusion methods, we constructed a novel fused enhanced vegetation index (EVI) dataset with a high spatiotemporal resolution (30-meter and 8-day resolution) for the YRD from 2000 to 2020, and we analyzed the vegetation variations and their driving factors within and outside the YRD Nation Natural Reserve (YRDNRR). The fused EVI effectively captured spatiotemporal vegetation dynamics. Notably, within the YRDNRR core area, the fused EVI showed no significant trend before 2010, while a significant increase emerged post-2010, with an annual growth of 7%, the invasion of Spartina alterniflora explained 78% of this EVI increment. In the YRDNRR experimental area, the fused EVI exhibited a distinct interannual trend, which was characterized by an initial increase (2000–2006, p < 0.01), followed by a subsequent decrease (2006–2011, p < 0.01) and, ultimately, a renewed increase (2011–2020, p > 0.05); the dynamics of the fused EVI were mainly affected by the spring runoff (R² = 0.71), while in years with lower runoff, it was also affected by the spring precipitation (R² = 0.70). Outside of the protected area, the fused EVI demonstrated a substantial increase from 2000 to 2010 due to agricultural land expansion and human management practices, followed by stabilization post-2010. These findings enhance our comprehension of intricate vegetation dynamics in the YRD, holding significant relevance in terms of wetland preservation and management.

Mangrove restoration projects are often evaluated based on the increase in mangrove forest (MF) area, but the reliability of this indicator as a measure of successful restoration is questionable. Considering both numerical and quality dimensions, this study assessed mangrove restoration efforts in the Leizhou Peninsula and Beibu Gulf (LP-BG) in China. The hypothesis was that due to the limited social capital investment in mangrove restoration, there exist hierarchical differences in the urgency of mangrove restoration. Time-series Landsat imagery from the Google Earth Engine platform was used to analyze the MF distribution from 2000 to 2020. A resilience indicator, incorporating resistance and adaptive capacity, was constructed to assess MF quality within identified mangrove boundaries. The results revealed an increase in MF area from 6655.87 ha in 2000 to 14,607.93 ha in 2020. However, the majority (79.6%) of MF patches exhibited low resilience (values < 3), with only a minority (51 patches) demonstrating high resilience. Interestingly, MFs within the mangrove reserve displayed higher resilience, but these areas did not exhibit a significant spatial expansion of MFs. These findings highlight the limitation of relying solely on the net growth in the MF area as an indicator of successful restoration. Instead, an ecologically optimal solution is recommended, focusing on expanding conservation boundaries to include remnant MFs outside protected areas, rather than creating new planting areas. This study provides an assessment framework to evaluate the efficiency of mangrove restoration efforts and offers insights for local decision makers to guide future restoration endeavors.

Multi-temporal InSAR (MT-InSAR) methods have been widely used in remote sensing monitoring of ground subsidence, which occurs at many places around the world. Land subsidence, caused by excessive extraction of groundwater, has always been a problem to be solved in Tianjin, China. Although the subsidence in the urban area has been controlled at a low rate, the subsidence issue has not been effectively solved in the suburban area recently, which should be paid much attention. This paper aims to present two multi-temporal differential interferometry techniques, persistent scatterer (PS) and small baseline subset (SBAS), for monitoring the latest surface subsidence in a Tianjin study area on the basis of 20 Sentinel-1A images obtained from March 2017 to March 2019. Our research showed that the average velocity map obtained from the SBAS method closely followed the outcomes of the PS technique from the perspective of identifying similar subsidence patterns. Subsidence rate gradually increased from the urban area of Tianjin to the suburbs and high subsidence zones were mainly distributed at the junction of the Wuqing, Xiqing and Beichen districts. In the past two years, the annual average subsidence rate in the high settlement area mostly exceeded −50 mm/year, which caused serious damage to local infrastructures. Besides, high-resolution remote sensing images combined with field investigations further verified the successful application of MT-InSAR technology in Tianjin’s subsidence monitoring. Effective ground subsidence control measures need to be taken as soon as possible to prevent the situation from getting worse.