Microplastics (MPs) have emerged as pervasive and persistent global environmental contaminants. However, the current inability to robustly analyze these particles hinders a deeper mechanistic understanding of their environmental behavior and ecological consequences. Artificial Intelligence (AI), with its computational power and data processing capabilities, is revolutionizing the analysis of MPs and transforming methodological paradigms. This critical review summarizes key challenges in the field, including persistent methodological limitations in characterization techniques and the interpretation of environmental behavior. We also critically examine emerging applications of AI in MPs domain. We explore the reliability and generalizability of these AI-driven approaches, highlighting that while data accessibility and computational resources pose core operational challenges, the environmental costs and ethical dilemmas of AI also demand scrutiny. Finally, we envision a paradigm shift towards an eco-conscious AI that harmonizes ecological accountability with computational efficiency. Fostering this transition through interdisciplinary collaboration is crucial for guiding MPs research toward a future that is sustainable across environmental, economic, and social consideration.

Microplastics (MPs) are increasingly recognized as dynamic ecological interfaces rather than inert contaminants, yet their role in shaping microbial functional interactions under climate-driven disturbances remains poorly understood. In particular, how climate driven freeze-thaw cycles (FTCs) influence the dynamics of mobile genetic elements (MGEs) within the plastisphere remains unclear. Here, we used controlled microcosms containing sediments from Tibetan Plateau lakes and polyethylene/polyethylene terephthalate particles to examine how simulated extreme FTCs were associated with changes in plastisphere microbial communities and functional gene co-occurrence patterns. FTCs reshaped plastisphere microbial communities, but their assembly trajectories differed between habitats. In the plastisphere, deterministic selection was pronounced at early FTC stages, whereas it did not further intensify with increasing FTC frequency. In contrast, sediment communities exhibited a progressive shift toward deterministic assembly with increasing FTCs. Concomitantly, MGEs were relatively enriched in the plastisphere and co-occurrence with genes involved in carbon metabolism, antibiotic resistance genes (ARGs), and virulence factors (VFs), forming interconnected functional networks. Notably, FTCs were associated with the enrichment of specific carbon metabolizing taxa (e.g., Comamonas and Brevundimonas) within the plastisphere that also carried relatively high abundances of MGEs, ARGs, and VFs. These patterns suggest that MPs may provide ecological niches linking carbon utilization and gene exchange potential. Collectively, this work highlights the interactive influence of climate-related disturbance and emerging pollutants on plastisphere microbial ecology, and provides a mechanistic basis for future assessment of ecological implications in fragile alpine ecosystems.

Anthropogenic nitrogen inputs have significantly influenced nitrogen cycling in river ecosystems. However, the role of headwater streams in influencing downstream sedimentary nitrogen cycling remains unknown. Here, we investigated a headwater stream-river network on the eastern Qinghai-Tibetan Plateau by integrating 16S rRNA gene sequencing, SourceTracker analysis, and random forest modeling to assess how headwater streams influence downstream sedimentary microbial communities and nitrogen cycling. The results showed that headwater streams act as selective microbial filters rather than passive sources, thereby influence sedimentary microbial communities and enhanced denitrification potential. SourceTracker analysis revealed that inflowing streams contribute 13.6 % -27.6 % of downstream sedimentary microbial communities, increased with decreasing distance from confluences. Several denitrifiers, including Hyphomicrobium, Rhodobacter, and Pseudomonas, as well as the gene encoding nitrous oxide reductase (nosZ), showed relatively higher abundance downstream of confluences, suggesting a higher denitrification potential. Random forest modeling further indicated that nitrogen cycling was primarily associated with microbial community composition, while also influenced by environmental factors such as nitrate concentration, water depth, and microplastic abundance. These findings highlight that headwater streams may influence downstream nitrogen cycling through a combination of microbial dispersal and environmental selection, with potential implications for nutrient management in sensitive high-altitude regions.

High accuracy and time synchronous aerosol optical depth (AOD) is essential for atmospheric correction (AC) of medium and high spatial resolution (MHSR) remote sensing data. However, existing high-resolution AOD retrieval methods often rely on sparsely distributed ground-based measurements, which limits their capacity to resolve fine-scale spatial heterogeneity and consequently constrains retrieval performance. To address this limitation, we propose a framework that takes GF-1 top-of-atmosphere (TOA) reflectance as input, where the model is first pre-trained using MCD19A2 as Pseudo-labels, with high-confidence samples weighted according to their spatial consistency and temporal stability, and then fine-tuned using Aerosol Robotic Network (AERONET) observations. This approach enables improved retrieval accuracy while better capturing surface variability. Validation across multiple regions demonstrates strong agreement with AOD measurements, achieving the correlation coefficient (R) of 0.941 and RMSE of 0.113. Compared to models without pretraining, the proportion of AOD retrievals within EE improves by 13%. While applied to AC, the corrected surface reflectance also shows strong consistency with in situ observations (R > 0.93, RMSE < 0.04). The proposed Trans-AODnet significantly enhances the accuracy and reliability of AOD inputs for AC of high-resolution wide-field sensors (e.g., GF-WFV), offering robust support for regional environmental monitoring and exhibiting strong potential for broader remote sensing applications.

Freeze-thaw cycling (FTC), a predominant climatic driver in alpine environments, remains an unaddressed knowledge gap regarding its impact on microplastic (MP) aging and associated microbial carbon metabolism. Here, we used controlled laboratory incubations coupled with multi-scale molecular and functional analyses to investigate MP-mediated carbon cycling under simulated FTC conditions. FTC exposure significantly modified the surface characteristics of MPs, particularly polyethylene terephthalate (PET), by inducing surface oxidation, increased roughness, and microcrack formation. These physicochemical transformations created reactive micro-environments that favored microbial colonization and metabolic potential. Under the combined stress of FTC and MPs, the abundance of sedimentary carbon-cycling genes declined by 30.39 to 31.27%. Conversely, the PET plastisphere exhibited a substantial enrichment of carbon-cycling genes (up to 3.93 & times; 105), effectively establishing a localized carbon cycling habitat. A dominant functional module comprising fixation, composition, and oxidation pathways accounted for 88.9% of carbon-cycling genes within the PET plastisphere, with Brevundimonas and Comamonas emerging as the key functional taxa in here. These findings demonstrate that intensified FTC can transform inert PET MPs into metabolically active microscale cycling habitats with the potential to influence local carbon-cycling processes, and highlight the need for future in situ investigations to evaluate the environmental relevance of these effects.

As global warming intensifies, the frequency of freeze-thaw events increases, significantly impacting microbial metabolism and biogeochemical cycling. However, the synergistic effects of freeze-thaw cycles (FTCs) and pervasive microplastics (MPs) on microbial community assembly and nitrogen cycling remain poorly understood. Here, we conducted a microcosm experiment integrating metagenomic and random forest model to elucidate the co-regulatory mechanisms of FTCs and MPs on plastisphere microbial communities and nitrogen metabolism. Results revealed that FTCs accelerated the environmental aging of MPs, inducing surface cracking and oxidation, thereby creating microenvironments favorable for microbial colonization. In the experimental microcosms, the combined effects of FTCs and presence of MPs increased microbial richness and diversity, promoted community differentiation between sediment and plastisphere, and increased microbial niche specialization. Functional analyses showed that FTCs induced a functional reconfiguration of the plastisphere nitrogen metabolism, with a selective enrichment of key enzyme genes, such as nitrite reductase, which may enhance nitrite redox activity and N₂O emission capacity. In the plastisphere, the contribution of Acinetobacter to nitrogen cycling increased, whereas Nitrospira declined, possibly due to oxygen limitation. Overall, our findings suggested that FTCs may facilitate transformation of MPs from inert pollutants into potentially metabolically active microhabitats, providing critical insights for assessing emerging pollutants and climate change.

The purpose of this study is to investigate the distribution and ecological drivers of microbial communities in sediments of the Wubu River, a high-density cascade-dammed tributary of the Yangtze River. The extensive hydropower development in the region has altered river ecosystems, yet its impact on microbial community assembly remains unclear. This study aims to elucidate the spatial heterogeneity of fungal and bacterial communities and identify key environmental factors influencing their distribution and stability. Sediment samples were collected from multiple sites along the Wubu River. High-throughput 16 S rRNA gene Illumina sequencing was used to characterize bacterial and fungal communities. Environmental variables, including total geographic distance, carbon content, and nitrogen levels, were analyzed to determine their influence on microbial distribution. Microbial network analysis was conducted to evaluate the impact of dam construction on community connectivity and stability. The ecological assembly processes were assessed using the dispersal limitation model. The results showed significant spatial heterogeneity in microbial community structure, primarily driven by hydropower development. Fungal α-diversity declined downstream, while bacterial communities exhibited stochastic variations. Dominant taxa, including unclassified_k_Fungi (12.91–39.24

Microplastic (MPs) contamination has become a critical worldwide environmental issue, causing significant harm to marine ecosystems. In Sudan, knowledge about the extent and impact of marine MPs pollution is still insufficient, highlighting the urgent need for focused initiatives aimed at prevention and environmental protection. This research establishes systematic baseline of MPs pollution along the coast of PortSudan, encompassing two bays and two coasts in Sudan’s Red Sea region. Water and sediment samples from four locations were analyzed for MPs, including color, size, morphology, and polymer, and proposed a comprehensive risk assessment method. The average abundance of MPs in the seawater and sediments were (154.84 ± 4.44 items/m³, and 171.79 ± 2.66 items/kg) for bay and (91.85 ± 1.511 items/m³ and 61.15 ± 1.64 items/kg) for open water, respectively. In general, MPs were predominantly identified in seawater and sediments as fibers (51.84%, 45.80%), with blue and white color (20.7% and 31.0%), 3.5–5.0 mm groups (27%, 44.1%) and were primarily composed of PVC and PE polymers, respectively. MP diversity was higher in bays than in open-water sites (p < 0.05), suggesting that hydraulic retention and deposition processes may enhance particle accumulation within these semi-enclosed environments. MPs concentrations exhibited positive correlations with salinity and depth, and negative correlations with DO and chl-flu, indicating the combined influence of climatic factors and anthropogenic inputs. The risk assessments results revealed that the four sites were exposed to high or very high-risk levels. These findings represent the first study to enhance our understanding and management of MPs pollution risks in Sudan's Red Sea.

The identification of roads from satellite imagery plays an important role in urban design, geographic referencing, vehicle navigation, geospatial data integration, and intelligent transportation systems. The use of deep learning methods has demonstrated significant advantages in the extraction of roads from remote sensing data. However, many previous deep learning-based road extraction studies overlook the connectivity and completeness of roads. To address this issue, this paper proposes a new high-resolution satellite road extraction network called FERDNet. In this paper, to effectively distinguish between road features and background features, we design a Multi-angle Feature Enhancement module based on the characteristics of remote sensing road data. Additionally, to enhance the extraction capability for narrow roads, we develop a High–Low-Level Feature Enhancement module within the directional feature extraction branch. Furthermore, experimental results on three public datasets validate the effectiveness of FERDNet in the task of road extraction from satellite imagery.

Microplastics (MPs) in aquatic environments has been observed globally. However, the ecological risks of MP pollution in riverhead prior to highly urbanized region remain poorly understood. This study investigated MP pollution related to microbiome in sediments, and ecological risks of MPs in riverhead prior to urbanized area over 291 km of Minjiang River (MJR) in Qinghai-Tibetan Plateau (QTP). MPs in river water and sediments were averagely 245±128 items/L and 124±67 items/kg, respectively, over the investigated river range. The MP distribution indicated that MP abundance is low in headwater section and elevated in middle section and down section with increase of urbanized area. The MPs were found mainly in film, fragments, and fiber morphotypes, with size < 500 μm in both river water and sediments. The polymers of MPs were contributed by polyethylene (PE), polypropylene (PP), polyethylene terephthalate (PET), polyvinyl chloride (PVC), and polycarbonate (PC) at 41.7%, 22.7%, 17.9%, 1.8%, and 1.2% in river water, and 32.6%, 15.0%, 29.3%, 3.1%, and 0.8% in sediments, respectively. Microbiome analyses of sediments revealed that the majority of microorganisms were aerobic bacteria, which contained potential plastic-degrading bacterial genes. Ecological risk assessment indexes of pollution load, polymeric risk assessment and pollution risk indicated that MPs in MJR river water and sediments displayed noticeable pollution levels, i.e., river water exhibited medium to very high pollution risk levels, and sediments showed from low to very high pollution risk levels in riverhead. Monte Carlo simulation revealed that PVC and PC MPs were considered as priority control pollutants although they were not the most abundant polymers identified due to their intrinsic chemical toxicity. Compared with risk levels of global rivers, the results indicate prominent ecological risks caused by MPs in MJR riverhead areas, and thus raise a warning sign.

Microplastics (MPs) have an impact on microbial community assembly and biogeochemical cycles in aquatic ecosystem, but the intrinsic carbon cycle within plastisphere remains unknown. Here, we compared carbon cycling in plastisphere of three major MP polymers i.e., polyethylene (PE), polypropylene (PP), and polyethylene terephthalate (PET) versus sediments in a high altitude lake on the eastern Qinghai-Tibet Plateau. The plastisphere was identified as carbon cycling hotspots, exhibiting distinct microbial assembly and functional patterns from sediments. The PET plastisphere demonstrated the most active anaerobic carbon metabolic activities, with microbial communities dominated by methanogenesis and fermentation pathways and methanogenic gene abundance far exceeding that in the PP and PE plastisphere. Sediments contained functionally diverse communities dominated by methane-oxidizing archaea and carbon-fixing bacteria, facilitated by stable redox gradients that drive various carbon transformation pathways. Community assembly in the plastisphere was primarily driven by stochastic processes, whereas sediment communities were influenced by environmental filtering. Homogenizing selection was stronger in PET plastisphere, favoring the enrichment of specific functional taxa. The findings reveal that MPs influence microbial community structure and metabolic function and the plastisphere is an independent driver of carbon cycling. This study provides new evidence of plastisphere-driven carbon metabolism, highlighting its role as localized greenhouse gas generator, and its divergence from sediment-mediated carbon cycling in lake ecosystems.

Microplastics (MPs) in freshwater have been extensively studied on a global scale. However, a deeper understanding is still required regarding the occurrence characteristics and ecological risks of MPs in protected area lakes(PAL). Here, the study investigated MPs pollution in PAL, outside protected areas lakes (OPAL), and ponds (OPAP) in the eastern Qinghai-Tibetan Plateau, and a comprehensive analysis was conducted comparing lakes or ponds from different income regions. The results showed that PAL has a single source of contamination, while OPAL and OPAP exhibited more diverse MP sources. The surface of all samples showed significant physicochemical changes like oxygen-containing functional groups and potential signs of biodegradation. Microbiome analysis identified potential plastic-degrading bacteria on MPs, which varied by polymer type. Ecological risk assessment revealed that OPAL and OPAP face higher ecological risks, particularly from polymers like PVC and PC, while PAL has low risk. However, we should also consider the environmental changes over the past 100 years of history in this region and emphasize the environmental health of PAL. Notably, MPs pollution is more severe in lower-middle-income regions, highlighting the urgent need for stricter controls.

We addressed rising drinking water risks in tropical tourism catchments by selecting Sanya as a representative case and developing an integrated 10–16 m remote sensing framework (Sentinel-2, GF-1) with a fuzzy evaluation, combining NDVI, WET, and NDBSI, K–T + NDVI eutrophication mapping, and event-sensitive RUSLE (30 m DEM, nonlinear LS, monthly NDVI-driven C, localized R). Land use mapping shows orchards at 736.46 km2 (38.37%) and tourism land at 2.64% (mostly golf), with 86.52% overall accuracy (Kappa 0.84). Basin-wide, 91% of the area experiences slight–mild erosion, intensified near reservoirs; relative to forests (FVC > 80%), orchards (FVC 60–70%) have a 3.2× higher median erosion risk (IQR 2.8–3.6, 95% CI 2.7–3.7). On 10–25° slopes during flood seasons, orchard pesticide/nutrient runoff indices rise 28–46%, and in the Dalong watershed, high-erosion orchard pixels co-locate with pesticide residues by 62% (95% CI 58–66%). Tourism is associated with elevated nearshore chlorophyll-a (Chl-a); the area is generally mesotrophic (0.25–0.75 mg/L), with localized nearshore hotspots > 1.0 mg/L; across monthly composites, nearshore Chl-a exceeds center waters by 130–210%, and in the Dalong Reservoir, the shoreline-to-center ratio is 2.3–3.1 (median 2.7, 95% CI 2.1–3.3) during 2023–2024 flood seasons. Overall, this source-to-sink framework supports forward-looking governance of drinking water sources under dual monsoon and tourism pressures.

The Brahmaputra River Basin (BRB)- a dynamic river system due to natural processes and tectonic forces, severely impact regional socio-economic activities and environmental sustainability. In the present study, high-resolution global digital elevation model (GDEM), topographic survey maps (1955), and multi-temporal Landsat satellite images (1975–2023) have been used to study different watershed levels detailed morphometric parameters of the Brahmaputra River, New (NBR), and Old (OBR) in Bangladesh. The study observed multiple channel sections linked to tectonic activities controlled by surrounding fault systems that confirm the morphometric changes over the past 68 years. Moreover, this study further contributes to the understanding of river dynamics by overlaying the river layer with areal changes, five different transects based on width measurements, spatio-temporal migration rate, and analyzing the multiple sinuosity indices variations at spatio-temporal scale for the periods 1955–2023. The present results will help the geospatial planners and policy makers to understand the ongoing morpho-tectonic activities that will help mitigate the associated natural hazards along the river basin.

Semantic segmentation is a critical process in remote sensing image analysis, supporting various applications. The recent development of the segment anything model (SAM), a visual foundation model designed to segment-anything, highlights the potential of foundational models in computer vision. However, SAM generates segmentation results without category labels, and predictions from semantic segmentation models for remote sensing often exhibit excessive fragmentation and imprecise boundaries. To address these limitations, we propose a strategy that integrates SAM with semantic segmentation models, replacing the imprecise boundaries of remote sensing segmentation masks with the more boundary-accurate SAM masks while retaining the original semantic information. Subsequently, a framework is designed and realized to enhance the prediction results of semantic segmentation models for remote sensing imagery by leveraging the raw outputs generated by SAM. This approach requires no additional training, modification to the semantic segmentation model, or changes to the visual foundation model, making it efficient and straightforward compared with other methods. Specifically, experimental results on two well-known datasets, LoveDA Urban and ISPRS Potsdam, demonstrate the effectiveness and broad applicability of our approach. In addition, incorporating recent visual foundation models, such as SAM-HQ and semantic SAM, further improves segmentation accuracy. As these models advance, the potential of our framework to enhance the performance of semantic segmentation for remote sensing imagery will grow. The source code for this work will be accessible at https://github.com/qycools/SESSRS.

Cloud detection in satellite imagery plays a pivotal role in achieving high-accuracy retrieval of biophysical parameters and subsequent remote sensing applications. Although numerous methods have been developed and operationally deployed, their accuracy over challenging surfaces—such as snow-covered mountains, saline–alkali lands in deserts or Gobi regions, and snow-covered surfaces—remains limited. Additionally, the efficiency of collecting training samples for prevalent deep learning-based methods heavily relies on large-scale pixel-level annotations, which are both time-consuming and labor-intensive. To address these challenges, we propose a Texture-Enhanced Network that integrates an object-oriented dynamic threshold pseudo-labeling method and a texture-feature-enhanced attention module to enhance both the efficiency of deep learning methods and detection accuracy over challenging surfaces. First, an object-oriented dynamic threshold pseudo-labeling approach is developed by leveraging object-oriented principles and adaptive thresholding techniques, enabling the efficient collection of large-scale labeled samples for challenging surfaces. Second, to exploit the spatial continuity of clouds, cross-channel correlations, and their distinctive texture features, a texture-feature-enhanced attention module is designed to improve feature discrimination for challenging positive and negative samples. Extensive experiments on a Chinese GaoFen satellite imagery dataset demonstrate that the proposed method achieves state-of-the-art performance.