The dike-pond system (DPS) is the integration of a natural or man-made pond and crop cultivation on dikes, widely distributed in the Pearl River Delta and Jianghan plain in China. It plays a key role in preserving biodiversity, enhancing the nutrient cycle, and increasing crop production. However, DPS is rarely mapped at a large scale with satellite data, due to the limitations in the training dataset and traditional classification methods. This study improved the deep learning algorithm Cascade Region Convolutional Neural Network (Cascade R-CNN) algorithm to detect the DPS in Qianjiang City using high-resolution satellite data. In the proposed mCascade R-CNN, the regular convolution layer in the backbone was modified into the deformable convolutional layer, which was more suitable for learning the features of DPS with variable shapes and orientations. The mCascade R-CNN yielded the most accurate detection of DPS, with an average precision (AP) value that was 2.71% higher than Cascade R-CNN and 11.84% higher than You Look Only Once-v4 (YOLOv4). The area of oilseed rape growing on the dikes accounted for 3.42% of the total oilseed rape planting area. This study demonstrates the potential of the deep leaning methods combined with high-resolution satellite images in detecting integrated agriculture systems.

Assessing Land Use and Land Cover Change (LULCC) related with aquaculture areas is vital for evaluating the impacts of aquaculture ponds on the environment and developing a sustainable aquaculture production system. Most studies analyze changes in aquaculture land in coastal areas, and little research focuses on the inland area, where the conversions between agriculture and aquaculture land is primarily driven by socioeconomic factors. This study assessed LULCC related to aquaculture areas in Qianjiang City, China, from 1990 to 2022, using multitemporal Landsat images and a combination of decision tree classifier and visual interpretation. The LULCC was analyzed by the transition matrix. Results showed that the main LULC type was farmland, which accounted for more than 70% of the study area from 1990 to 2022. The built-up and aquaculture land showed an increasing trend year by year. In contrast, there was a gradual decline in forest/grass land from 1990 to 2016, and then its area increased slightly from 2016 to 2022 due to the policy of returning farmland to forest. Water areas were mainly composed of rivers and ponds, with subtle changes during the study period. The main driving forces of LULCC in Qianjiang City were economic and policy factors, with rapid GDP growth and government policies being the dominant factors.

The classification of optical satellite-derived remote sensing images is an important satellite remote sensing application. Due to the wide variety of artificial features and complex ground situations in urban areas, the classification of complex urban features has always been a focus of and challenge in the field of remote sensing image classification. Given the limited information that can be obtained from traditional optical satellite-derived remote sensing data of a classification area, it is difficult to classify artificial features in detail at the pixel level. With the development of technologies, such as satellite platforms and sensors, the data types acquired by remote sensing satellites have evolved from static images to dynamic videos. Compared with traditional satellite-derived images, satellite-derived videos contain increased ground object reflection information, especially information obtained from different observation angles, and can thus provide more information for classifying complex urban features and improving the corresponding classification accuracies. In this paper, first, we analyze urban-area, ground feature characteristics and satellite-derived video remote sensing data. Second, according to these characteristics, we design a pixel-level classification method based on the application of machine learning techniques to video remote sensing data that represents complex, urban-area ground features. Last, we conduct experiments on real data. The test results show that applying the method designed in this paper to classify dynamic, satellite-derived video remote sensing data can improve the classification accuracy of complex features in urban areas compared with the classification results obtained using static, satellite-derived remote sensing image data at the same resolution.

Urbanization-driven land use transitions, especially impervious surface expansion, profoundly affect aquatic ecosystems, yet the regulatory mechanisms of microbial assembly and disassembly remain unclear. Using environmental DNA metabarcoding, this study examined relationships between landscape patterns, environmental gradients, and microbial communities in the middle-lower Yangtze River. Impervious surfaces (source landscapes) increased from 10% (midstream) to 37% (downstream), disrupting source–sink balance, elevating nitrogen and phenolics, and worsening water quality. Eukaryotes exhibited higher α-diversity and more complex co-occurrence networks, showing greater responsiveness to disturbance. Stochastic processes dominated community assembly (>68%) for both eukaryotic and prokaryotic communities. Dispersal limitation led to higher total β-diversity and nestedness in eukaryotes; by contrast, land use intensified environmental filtering on prokaryotes, resulting in lower β-diversity but higher turnover contribution. Community disassembly simulation reveals a scenario in which impervious surface expansion is associated with non-random species loss, reducing predicted functional richness of both eukaryotic and prokaryotic microbes; the preferential loss of sensitive species is associated with a sharp decline in predicted functional richness, which may have implications for ecosystem stability. Our findings demonstrate that land use change is significantly associated with altered aquatic environments and affects ecosystem functions through biodiversity changes, providing theoretical support for ecosystem conservation.

The Yangtze, Asia’s longest river, supports ecosystems and people, but rising stress from development is degrading water quality and altering algae. Using 2016–2021 water quality data from 83 sites and 2021 phytoplankton (55 sites) and periphyton (44 sites) surveys, this study analyzes the spatiotemporal variations in water quality and algal communities in the Yangtze River mainstream and the driving mechanisms via Mann–Kendall trends, random forest modeling, and hierarchical partitioning. The results revealed significant long-term declines in TN, TP, ammonia nitrogen, and permanganate indices, with a rising water quality index (WQI), reflecting effective basin-wide pollution control. Spatially, the Jinsha River segment maintains a high WQI and low nutrient content, whereas the downstream reaches present increasing TN, TP, and ammonia nitrogen along the longitudinal gradient in 2021, reflecting the cumulative effects of intensified anthropogenic activities from the headwaters to the estuary. Phytoplankton density increases downstream and is correlated with nutrient enrichment, with cyanobacteria/Chlorophyta dominance in the middle–lower reaches. The periphyton density peaks in the middle reaches (1.76 × 10⁸ cells/cm²), which is driven by the optimal temperature, nutrients, and substrates. Integrated ecological assessment via random forest models classified 85% of the sites as “good” or “excellent,” but 15% of the lower-reach sites were rated “moderate,” highlighting nutrient-induced ecological stress. Mantel tests and hierarchical partitioning identified TN as the primary driver of phytoplankton community composition, whereas water temperature emerged as the key factor shaping periphyton distribution. These findings underscore the necessity of differentiated management strategies: strict conservation of upstream pristine habitats and enhanced nitrogen–phosphorus control in downstream industrialized zones.