Farmland Semantic Change Detection (SCD) is essential for cultivated land protection, yet existing benchmarks and models remain insufficient for fine-grained farmland conversion monitoring. Current datasets often lack dedicated "from-to" annotations, while visual change detection models are easily disturbed by phenology-induced pseudo-changes caused by crop rotation, seasonal variation, and illumination differences. To address these challenges, we construct HZNU-FCD, a large-scale fine-grained farmland SCD benchmark with a unified five-class farmland-to-non-farmland annotation protocol. It contains 4,588 bitemporal image pairs with pixel-level labels for practical farmland protection. Based on this benchmark, we propose a large-small collaborative SCD framework that integrates a task-driven small visual model with a frozen large vision-language model. The small model, Fine-grained Difference-aware Mamba (FD-Mamba), learns dense change representations for boundary preservation and small-region localization. The large-model pathway, Cross-modal Logical Arbitration (CMLA), introduces CLIP-based textual priors for prompt-guided semantic arbitration and pseudo-change suppression. To enable effective collaboration, we design a hard-region co-training strategy that supervises the CMLA semantic score map only on low-confidence pixels. Experiments show that our method achieves 97.63

Remote sensing images change detection often faces the problem of false detections and missed detections due to various factors such as different sensors, seasons, and weather conditions when acquiring bi-temporal images. To tackle this challenge, we propose a remote sensing images change detection network based on FastSAM. First, a siamese network structure is employed as an encoder, utilizing the vision foundational model FastSAM to enhance the generalization ability. Then, in order to reinforce the semantic features within the regions of change, we propose a Differential Enhancement Adapter (DEA) module. Finally, a Full-Scale Skip Connections (FSC) is adopted to synergize the deep semantic features from varying scales with the superficial semantics, thus bolstering the model's capacity to discern finer details. This model has achieved significant results on the Jiashan County land use change detection datasets, and has also made significant improvements compared to other advanced models.

Hynobius amjiensis is a critically endangered species endemic to China, classified as a national first-class protected animal. Understanding the current wild status of Hynobius amjiensis is essential for conservation and regulatory efforts. However, due to the species’ unique ecological habits, direct field tracking is challenging. This paper explores the feasibility of underwater high-definition video surveillance and deep learning to monitor Hynobius amjiensis in the field, and evaluates the impact of data sets on model accuracy. The results indicate that: (1) Variations in time periods, behavioral states, and occlusion levels of target samples significantly affect the model's detection accuracy and recall rate; (2) Incorporating samples of similar species improves the overall performance of the model, enhances species classification precision, reduces the false detection rate, and increases the model's robustness and generalization capability. The optimal model, YOLOv8x-ABCD, demonstrated superior performance across all evaluation metrics and exhibited the highest accuracy in detection results. However, non-living objects in the water, such as aquatic plants, shed egg bands, and branches, were occasionally misidentified as adult Hynobius amjiensis. Experimental results suggest that increasing the number of negative samples effectively reduces such misdetections, lowers the model’s error rate, and enhances detection accuracy and generalization.

In recent years, with the continuous development of digital twin technology, its application scope has become increasingly widespread across various fields. Currently, static WebGIS applications built on the open-source 3D mapping framework Cesium fail to effectively showcase the dynamic process of model changes and struggle to seamlessly integrate models with environmental elements, thereby limiting the flexibility and applicability of digital twin applications in practical construction. To address this issue, this paper proposes an engine integration method based on depth buffering, aiming to achieve a depth integration between Cesium and general 3D rendering engines, thereby effectively enhancing Cesium’s model rendering capabilities in WebGIS application development. Furthermore, through in-depth exploration of graphic rendering algorithms, we have designed scene visualization algorithms applicable to different environmental parameters, achieving effective integration of models with environmental elements and successfully simulating the process of model dynamic changes.

Hynobius amjiensis is a critically endangered species endemic to eastern China with important scientific and ecological values, but its population reproduction is not optimistic. Currently, the mainstream research method is to assess population reproduction through human survey of oocyst number, which is difficult to observe individuals, and human intervention can easily damage the local ecology. Therefore, we used deep learning target detection to detect Hynobius amjiensis underwater. There are many difficulties in target detection in the actual underwater environment, such as poor image quality, many disturbances, small target of Hynobius amjiensis, easy to be occluded and abundant posture changes, which bring great difficulties to target detection. To make the target detection model of Hynobius amjiensis more robust, this paper proposes a target detection model of Hynobius amjiensis based on improved YOLOv8x, which introduces three innovative methods. First, the CBAM attention mechanism is fused with the C2f module to form the CBC2f fusion enhancement module, which enhances the ability to extract feature information and capture key features. Secondly, an enhanced BiFPN network is used to replace the PANet network for feature fusion to improve the information fusion capability between different layers. Finally, a shallow feature layer is added to improve the scale range of feature extraction. The results indicates that our proposed YOLO-XN model achieves greater performance compared to the original YOLOv8x model on the Hynobius amjiensis dataset. Specifically, the precision, recall, mAP, and mAP scores reach 94.6%, 90.1%, 94.6%, 70.1%, which are increased by 1.6%, 2.3%, 1.3%, and 2.0%, respectively. It demonstrates the superior performance of our improved YOLO-XN model in detecting Hynobius amjiensis and provide an effective auxiliary method for intelligent wildlife monitoring.

The precipitation structures and microphysical characteristics of predecessor rain events (PREs) over the Yangtze River Delta area and related tropical cyclones (TCs) from 2014 to 2019 were investigated using Dual-frequency Precipitation radar data from Global Precipitation Measurement (GPM) for drop size distributions (DSDs). Results showed that the total mean rain rate of PREs was larger than that of TCs, primarily due to higher convective and stratiform rain rates, with enhanced fractional coverage of convective rain in PREs. Examination of microphysical characteristics revealed that a greater quantity of small-sized droplets and a smaller quantity of medium- as well as large-sized droplets contributed towards PREs in comparison with TCs. The conclusion still holds when partitioning DSDs based on different precipitation rate categories. Further investigation of DSDs using gamma functions illustrated that precipitation in PREs had lower average mass-weighted diameters (Dm) and enhanced normalized number concentration (Nw) compared with TCs; partitioning precipitation into convective and stratiform components illustrated a larger Dm and lower Nw in TCs than PREs. The analysis of microphysical and thermodynamical processes using the reanalysis data indicates that relatively intense convective activity with drier conditions may be favorable to enhancing raindrop growth through collisioncoalescence processes, as a result of larger Dm in TCs than PREs. The empirical relations (Z-R algorithms) applied in different rain regimes (stratiform, convective, and total PREs) revealed significant diversities, relying on weather conditions and geographical locations. Plain language summary: The Yangtze River Delta area is an important economic belt in China. Under climate change, observed frequent occurrences of weather extremes of heavy rainfall and tropical cyclones (TCs) exert adverse effects on economic development in this region. Thus, a deep understanding of the mechanism of TCs torrential rainfall in the Yangtze River Delta area is urgently necessary. In this study, we investigated the precipitation patterns and microphysical characteristics of predecessor rain events (PREs) in the Yangtze River Delta region, and their association with TCs in the South China Sea-Western North Pacific Ocean (SCS-WNPO) area from 2014 to 2019. We found that PREs had a higher total mean precipitation rate than TCs. Further examination using gamma functions demonstrated that PREs exhibited lower average mass-weighted diameters (Dm) and higher normalized intercept parameters (Nw) than TCs. This pattern persisted when distinguishing between convective and stratiform precipitation components. We believe that our study makes a significant contribution to the literature because these results provide valuable insights into the distinct precipitation characteristics of PREs and TCs in the study region and contribute to a better understanding of tropical cyclone-related rainfall patterns, and act as a scientific basis for disaster mitigation.

With the continuous development of digital twin technology, it has been widely applied in various fields. The application of digital twin technology in the water environment is becoming a hot topic, bringing new possibilities for water quality monitoring and early warning. Digital twin models are the core elements of digital twins. Currently, existing digital twin models of water bodies are mainly implemented in a static manner, making it difficult to express the dynamic changes of water bodies, integrate with environmental factors, and achieve realistic water twin scenarios. This paper proposes the construction of water body twin models through parameterization, and by integrating environmental parameters, accurately and intuitively expresses the characteristics and dynamic changes of water bodies. This has reference significance for promoting the development of digital twin technology.

Monitoring the state and trends of bird diversity in ecosystems is a major challenge that requires widely applicable machine learning-based bird song recognition algorithms. When dealing with the spectrum data set, some traditional classifiers have some limitations in overcoming noise interference and improving the stability of recognition algorithm because the spectrum data show various and peculiar features, the choice of kernel function is limited, and the feature scaling and noise are sensitive. Aiming at the problems of bird audio noise, high similarity of feature extraction maps, and irregular curvature of image features and low recognition accuracy, this paper adopts YOLOv8-BEM based on progress learning to detect and identify birds by using voiceprint. According to the characteristics of voiceprint, this paper proposes a bird voiceprint extraction adaptation module (BEM). Special convolution (DSC) and attention channel (MLLA) were added to the CF2 module in its backbone network, special convolution was performed on important channel data in the branch, and the extraction gradient flow of full-scale jump connection was added to rebuild the network feature extraction ability and improve the model detection accuracy. The proposed model is verified on five bird datasets, and the precision, recall and F1 score are 95.02%, 90.35% and 92.59% respectively, which are better than other comparison models.

Predecessor rain events (PREs) in the Yangtze River Delta (YRD) region associated with the South China Sea and Northwest Pacific Ocean (SCS-WNPO) tropical cyclones (TCs) are investigated during the period from 2010 to 2019. Results indicate that approximately 10% of TCs making landfall in China produce PREs over the YRD region; however, they are seldom forecasted. PREs often occur over the YRD region when TCs begin to be active in the SCS-WNPO with westward paths, whilst the cold air is still existing or beginning to be present. PREs are more likely to peak in June and September. The distances between the PRE centers and the parent TC range from 900 to 1700 km. The median value of rain amounts and the median lifetime of PREs is approximately 200 mm and 24 h, respectively. Composite results suggest that PREs form in the equatorward jet-entrance region of the upper-level westerly jet (WJ), where a 925-hPa equivalent potential temperature ridge is located east of a 500-hPa trough. Deep moisture is transported from the TC vicinity to the remote PREs region. The ascent of this deep moist air in front of the 500-hPa trough and frontogenesis beneath the equatorward entrance region of the WJ is advantageous for the occurrence of PREs in the YRD region. The upper-level WJ may be affected by the subtropical high and westerly trough in the Northwest Pacific Ocean, and the occurrence of PREs may favor the maintenance of the upper-level WJ. The upper-level outflow of TCs in the SCS plays a secondary role.

利用2001-2020年嘉善地区13个监测断面的4项水质指标监测资料,采用水质指数、累积距平法、Mann-Kendall突变检测法、克里金插值方法,分析该区域年度水质变化特征.结果表明,嘉善县近20年水质指数(water quality index,WQI)总体表现为波动变化.2001-2013年,WQI呈现波动上升趋势;2014年开始,WQI呈下降趋势,嘉善县地表水水质得到明显改善.从年际变化来看,氨氮年际变化最大,总磷(TP)次之,高锰酸盐指数(CODMn)变化最小.嘉善县地表水水质污染的突出特征是氨氮、总磷.年内变化分析结果显示,3月份水质最差,10月份水质最好,WQI春季＞冬季＞夏季＞秋季,且氨氮污染指数最高.空间维度分析结果显示,嘉善县北部水质较好,东南部水质较差.

High pressure on urban drainage systems caused by extreme precipitation events would lead to an increase risk of urban floods. Across China, future changes in urban drainage pressure (UDP) and its response to global-scale climate mitigation and local adaptation, have seldom been studied. Here, based on climate projections from the Coupled Model Intercomparison Project Phase 6 (CMIP6), we assessed UDP changes from 2020 to 2099 under different scenarios in 285 cities across China. Under the shared socioeconomic pathway (SSP) 5-8.5 scenario, 30% larger increase of UDP relative to the baseline level (1985-2014), would occur in 22.80% and 79.65% cities over 2020-2049 and 2050-2099, respectively. Under climate mitigation (SSP2-4.5 scenario), UDP in northern China would decrease by 10-30% over 2020-2049. On this basis, 10% enhancement of underlying surface retention capacity (LID10% scenario) would reduce UDP by more than 10% particularly in northern and north-eastern China (23.51% cities). Pipe enlargement adaptation (Pipe10% scenario) would benefit UDP mainly in eastern China (46.31% cities), by postponing the first decade with 30% larger pressure relative to the baseline level by 1-3 decades.

利用 1960 年、1970 年、2000 年、2016 年、2018 年、2020 年 6 个时期嘉善县姚庄镇水系相关数据,建立姚庄镇水系结构与连通性评价指标体系,分析姚庄镇近 60 年来的水系结构与连通性变化过程.研究结果表明:1960-2020 年,姚庄镇区域内水面率下降了 30.77％,湖泊萎缩了近 60％;1970 年以前,农业生产活动导致湖荡萎缩,使姚庄镇水系结构与连通性呈下降趋势;1970-2000 年,水系连通性受到农业生产活动和人为水利工程的影响,连通性总体呈增加状态;2020 年之后,城市化进程中对部分河流清淤连通,河流连通性得到增强.

合理快速评价矿山重金属污染修复的效果对于矿山生态恢复与重建治理工作具有重要意义.以德兴铜矿为例,根据野外实测植被光谱,分析矿区内主要植被的典型光谱特征;根据实验室化验的植被叶片内重金属含量,分析其重金属含量与光谱特征参数红边位置的关系;利用 2003 年和2009 年2 景Hyperion高光谱卫星数据计算矿区植被的红边位置,推断矿区植被富集重金属的情况,进而评价矿山重金属污染修复的效果.研究结果表明,在典型复垦区1 号、2 号尾矿库四周重金属污染修复取得了较好的效果;与 2003 年相比,2009 年重金属污染修复整体上取得了一定的成效,大部分区域被修复,但仍有部分新增污染区,需采取修复措施.该方法可快速、合理、大范围地评价矿区重金属污染修复的效果.

Two-dimensional (2D) and three-dimensional (3D) cloud-resolving model (CRM) results from the Tropical Rainfall Measuring Mission Kwajalein Experiment (KWAJEX) were applied to analyze the diurnal cycle of cloud development in the tropics. Cloud development is intimately associated with the growth of secondary circulation, which can be analyzed in the budget of perturbation kinetic energy (PKE). The ice and liquid water path (IWP+LWP) is a fundamental parameter for estimating clouds, with the analyzed results suggesting that (1) the ice and liquid water path (IWP+LWP) and PKE values attained in convective regions were higher during the nighttime than during the daytime and that the maxima of IWP+LWP and PKE occurred at midnight in the lower troposphere in the 3D model run, and that (2) the IWP+LWP and PKE values in stratiform regions were much higher in the afternoon than in the morning, while the maxima of IWP+LWP and PKE occurred in the afternoon in the middle troposphere in the 2D model run. Further analysis demonstrated that both the high IWP+LWP and PKE values in the lower troposphere at midnight were mainly associated with the warm–humid lower troposphere in convective regions. However, those in the middle troposphere in the afternoon were primarily linked to the dry–cold upper troposphere and moist–warm lower troposphere in stratiform regions. The results further revealed that (1) both IWP+LWP and PKE exhibited shorter time scales in the 2D model runs than in the 3D model runs and that (2) the maximum IWP+LWP values occurred in the afternoon in the 2D model runs and at midnight in the 3D model runs.

为探究HY-1C CZI与Sentinel-2 MSI对海岸带植被信息提取差异,以杭州湾、莱州湾和大亚湾为研究区,将研究区划分为城市和密林两个区域,计算归一化植被指数NDVI和植被覆盖度FVC,对比CZI、MSI的NDVI统计信息和相关性,分析FVC分级面积占比和时空分布异同.结果显示:CZI与MSI在大亚湾与莱州湾的NDVI均值十分接近,但CZI在杭州湾有大量近饱和像元,城市区和密林区分别相对MSI高估0.189和0.183;CZI与MSI在不同样区的NDVI均值线性拟合效果优良,R2最高达到0.98;CZI的FVC与MSI的FVC在不同覆盖等级占比、植被覆盖水平差异、高-低分布趋势三个方面呈现出相似性,在杭州湾地区具有相近的时间分布特征,但CZI的FVC在所有城市平原区及中北部海岸带的高山密林区均有高估现象.综合来看,HY-1C CZI整体可满足我国中等尺度海岸带植被监测需求,在植被变化监测方面有较大潜力.

The three-dimensional Weather Research and Forecasting (WRF) model was used to conduct sensitivity experiments during the landfall of Typhoon Fitow (2013) to examine the impacts of cloud radiative processes on thermal balance. The vertical profiles of heat budgets, vertical velocity, and stability were analyzed to examine the physical processes responsible for cloud radiative effects on surface rainfall for Typhoon Fitow (2013). The inclusion of clouds reduced radiative cooling in ice and liquid cloud layers by reducing outgoing radiation. The suppressed radiative cooling reduced from the ice cloud layers to liquid cloud layers. This was conducive to reducing instability. The decreased instability was associated with the reduced upward motions. The reduced upward motion led to a decreased vertical mass convergence. Consequently, heat divergence was weakened to warm the atmosphere. Together with suppressed radiative cooling, these effects jointly suppressed net condensation and rainfall. Furthermore, the reduced rainfall due to the cloud radiative effects were mainly associated with the reduced convective and stratiform rainfall. The reduced convective rainfall was associated with the reduced net condensation, while the reduced stratiform rainfall was related to the constraint of hydrometeor convergence.

利用1970—2019年的气象观测资料,采用线性倾向估计法、累积距平法、M-K突变检测法和Morlet小波分析等方法,对安吉小鲵栖息地年度、季度气温和降水的变化特征进行了详细分析.结果表明:50年间,安吉小鲵栖息地的年平均气温及四季气温均呈显著上升趋势,20世纪90年代后气温升高趋势逐年加剧;降水量除秋季呈弱减少外,春、夏、冬季和年降水量均呈弱上升趋势;区域内年平均气温存在28年的强显著变化周期;降水存在22、15、11年的变化周期,强显著周期为22年;安吉小鲵栖息地未来几年内气温仍呈现偏高趋势,降水则向偏少趋势发展.

以杭州湾南岸为研究区,利用1980年9月20日的Keyhole遥感影像与同一时期的Landsat MSS遥感影像,通过融合处理,获得既具有高空间分辨率又有多光谱分辨率的历史遥感影像,填补历史高分辨率遥感影像的空缺.研究结果将历史土地利用类型变化监测时间序列推前的同时,提高杭州湾南岸土地利用变化监测的精度.研究中将使用历史融合影像对杭州湾土地利用类型变化进行监测,同时结合1990-2020年4景Landsat遥感影像,获得杭州湾南岸地区近41年的土地利用情况,辅以前人的目视解译结果图,得到分类精度＞90％的土地覆被利用分类结果.从面积变化、类型转化、年均变化率3方面分析讨论了 1980-2020年间杭州湾南岸地区土地利用的时空变化特征.结果表明:1980-1990年,杭州湾南岸库塘呈减少的趋势,耕地逐渐向海岸地区扩张;1990-2020年,城市的扩张面积不断增加且呈现向海岸扩张的趋势,库塘的面积增加明显,主要表现在人工养殖场继续向海岸扩张.城市的发展是影响杭州湾南岸土地利用变化的主要因素,并且在杭州湾城市规划的同时应该充分考虑其生态环境的问题,加强自然湿地保护,严格控制自然湿地的开发规模,坚持可持续发展,在保护湿地的基础上进行合理的开发利用.

从光谱尺度和空间尺度综合考虑了ASTER与WorldView2的优缺点,把经过地形校正、大气校正和精确配准的两种影像重采样至同一级别分辨率,使WorldView2 pan与ASTER SWIR进行融合,融合后的影像与WorldView2的8个波段进行组合得到两种数据的协同影像.以新疆塔什库尔干县老并地区作为研究区开展蚀变信息增强和提取工作,实验结果表明协同影像的信息量更加丰富,光谱分辨率和空间分辨率较原始两种影像大为改善,在铁染蚀变信息增强和提取方面具有显著效果.

This study investigated cloud microphysical processes, and the upward and downward movement of precipitating particles with the Weather Research and Forecasting model during Typhoon Fitow (2013). The Purdue‐Lin single‐moment scheme and the Milbrandt‐Yau and National Severe Storms Laboratory (NSSL) double‐moment microphysics schemes were utilized. The area‐mean rainfall simulated by the three schemes shared similar magnitudes and variations. The area‐mean rain rates were insensitive to microphysics schemes, as they were mainly determined by large‐scale water vapor convergence. However, the local rainfall intensity was sensitive to the schemes and graupel and hail parameterizations. This indicates great impacts of microphysics and terminal velocities of graupel/hail on the local rainfall intensity. Upward‐moving graupel had a stronger net source compared to the downward‐moving graupel over the spiral rainband in the Milbrandt‐Yau simulation due to the low terminal velocity; whereas the net source was much weaker for upward‐moving graupel than for downward‐moving graupel over the eyewall in the NSSL simulation due to the large terminal velocity. Decreasing the terminal velocities of graupel and hail (if hail exists) in the mid and lower troposphere significantly reduced the maximum local rainfall intensity via increasing the sources of upward‐moving graupel. This is because the upward‐moving graupel is prone to remaining in the clouds being advected out of the updraft region instead of directly converting to raindrops as the sinks of graupel. The microphysics and the terminal velocities were together responsible for the local rainfall intensity.