Compared to the Last Glacial period, climate change during the Holocene is relatively stable, yet millennial-scale climatic fluctuations still occurred. Based on the grain size, mineralogy, major and trace elements of ZK01 sediments in northern Beibu Gulf, combined with 14 C age of bioclasts, this paper extracts the weathering and erosion signals of bedrock in Nanliu River Basin since late Last Glaciation in southeastern Guangxi, and reconstructs its paleoclimate change characteristics. The results indicate that the provenance of sediments remains relatively stable, primarily derived from Indosinian granites in southeastern Guangxi, with a possible minor contribution of volcanic material in Late Holocene. Combined with the weathering index, grain size, paleoclimate and paleotemperature parameters, the weathering of the parent rock area can be divided into five stages: P1 stage (Early Pleistocene), with strong winter monsoons, low precipitation, and a cold-dry climate, which inhibits rock weathering; H1 and H2 stage(Early Holocene): the sediments is primarily controlled by the climate of the source area, this period is characterized by a strong summer monsoon, abundant precipitation, and a warm-humid climate, which enhances rock weathering. H3 stage(Mid-Holocene): the summer monsoon gradually weakens, the winter monsoon is stronger, leading to reduced precipitation and dry-cold climate, the rock weathering intensity weakens, and is inhibited in the later stage. During the early phase of H3 stage, sediments is mainly influenced by the climate of the source area, whereas in the mid-to-late phase, it is affected by both the source area climate and sea-level fluctuations. H4 stage(Late Holocene): the summer monsoon is strengthened, precipitation is increased, and the climate becomes warm and humid, leading to an increase in rock weathering intensity, and sediments is influenced by both the climate of the source area and submarine tectonic subsidence. This study is of great significance for understanding the weathering process and paleoclimate evolution in southeastern Guangxi and even South China since the late stage of Last Glacial period.

By analyzing the geological characteristics and metallogenic patterns of quartz ore bodies distributed within Nankang Basin, this study deepens the understanding of ore body characteristics, mineralization laws, and genetic models of quartz sand deposits in the basin. The quartz sand deposits in Nankang Basin primarily occur in Lower Pleistocene Zhanjiang Formation of Quaternary and Middle-Upper Pliocene Nankang Formation of Neogene System. The ore-bearing lithology is dominated by grayish-white sand layers. While distributed throughout the basin, ore bodies exhibit significant spatial variations. Those in the northeastern sector demonstrate substantial thickness, stable variations, and well-preserved internal continuity, representing the most concentrated quartz sand enrichment zone in the basin. The ore quality is stable, and the washed and deslimed clean quartz sand reaches the first-grade product requirements of siliceous raw materials for flat glass. Detrital zircon U-Pb dating reveals dominant age peaks at 200-300 Ma and 400-500 Ma. Combining with the basin geomorphological characteristics, the distribution of magmatic rocks in the region and their formation ages, it is indicated that the Darongshan pluton of Indosinian age and Ningtan pluton of Caledonian age are the main sediment sources of quartz sand deposits in Nankang Basin. After the weathering of the rock masses, they are transported over long distances by Nanliu River and its tributaries, ultimately accumulating in the basin to form quartz sand ore bodies. The genetic type of these deposits is classified as fluvial and littoral facies quartz sand deposits.

Based on multi-source geological survey data(such as boreholes, sonar, and single-channel seismic surveys) and hydrological monitoring data, analytic hierarchy process(AHP) is applied to develop a comprehensive evaluation model that includes geological stability, ecological sensitivity, and pollution load. The study systematically analyzes the key geological issues in the Maowei section of Qinzhou Bay, including erosion and deposition, submarine landslides, and seawater intrusion, and quantitatively assesses the risk levels of these environmental geological problems. The findings indicate that the region is primarily characterized by sedimentation, with localized deep grooves of scouring. The heavy metal content in the seabed sediments meets standards, and the mangrove reserve has been affected by dredging. The regional geological environmental risks are moderate, with the main threats being erosion and deposition, as well as slope instability. To address these issues, a trinity green construction strategy of “engineering regulation-ecological restoration-collaborative management” is proposed, which includes the transplantation of mangroves from other areas, the use of pollution barriers to block suspended particles, and the establishment of a dynamic monitoring network. 

The surface crack characteristics of soil samples under different drying-wetting cycle conditions are analyzed by image processing techniques, and direct shear tests are conducted on soil samples with different dryingwetting cycles to investigate the effects of root diameter on the crack and shear strength of expansive soil and the relationship between crack and shear strength of expansive soil. The results show that the smaller the root diameter, the smaller the crack characteristics such as crack ratio, crack length, crack width and crack connectivity, and the larger the number of cracks and shear strength indexes. When the crack is within a certain range, the soil shear strength index decays with the increase of surface crack, and in the process of increasing crack, the shear strength of expansive soil containing small root diameter root system is significantly higher than that of other soil samples; when selecting protective vegetation for expansive soil slope, priority is given to vegetation with fine root system.

There are a large number of fully weathered phyllite soil and red clay in Guilin. It is of great significance for engineering practice to understand the strength characteristics of the two soils and distinguish them accurately. In fully weathered phyllite soil and red clay as the research object, the compaction sample are consolidated quick shear tests, moisture content and dry density are studied against the shear strength of the film. At the same time, using scanning electron microscopy(SEM) techniques from the micro perspective to proved the shear strength is different. The results show that the shear strength of fully weathered phyllite soil in Guilin is higher than that of red clay. The shear strength of two types of soil in the same dry density increased with the increase of moisture content. And then the shear strength of two types of soil in the same moisture content increased with the increase of dry density. This is due to the high content of free iron oxide in the fully weathered phyllite soil, which increases the cohesion of the fully weathered phyllite soil. SEM technology can prove that the particles of the fully weathered phyllite soil are larger, and the aggregate formed is larger and more structural, so the shear strength is higher.

Dissolution diseases are common on the surface of Guilin limestone carvings affected by the environment. Based on the local occurrence environment of the stone carvings, Longyin Cliff Rock Carvings and the Xishan Cliff Statues in Guilin are taken as research objects, an indoor simulated dissolution test of limestone with a cycle of 50 times is carried out, and the dissolution patterns of two kinds of limestones at different pH conditions and cycles are compared. The results show that after 50 rain cycles, the mass loss rate of the rock samples is more than 0. 15%, which is as high as 0. 30% after the pH = 3. 5 rain cycle, and the mass loss of the rock samples after the pH = 6. 5 temperature difference cycle is less than 0. 07%. The dissolution effect of the rain cycle on limestone is stronger than that of the temperature difference cycle, and the acidity of the solution greatly promot the dissolution. Longyin limestone is more easily dissolved than Xishan limestone. According to the results, the annual dissolution rates of the stone carvings are calculated. Without effective protection, the handwriting and carvings of the Guilin Longyin Cliff Rock and the Xishan Cliff Statues would be etched flat after 75 and 111 years respectively under the long-term erosion of rain.

Magnesium phosphate cement(MPC)-based grouting material with high water-solid ratio is prepared by compounding light-burned magnesium oxide and dead-burned magnesium oxide, which is used to plug strong leakage water in underground engineering. The orthogonal test is used to screen the best ratio, and the effect of the compound ratio of light-burned / dead-burned magnesia(L / D) and the magnesium-to-phosphate molar ratio (M/ P) on the performance of MPC is further studied. The properties of MPC are characterized by macro indicators such as setting time, compressive strength, underwater/ in-air strength ratio, stone rate and content of suspended solids, and micro-tests such as XRD and TG-DTG. The results show that the best ratio of MPC grouting plugging material is: L / D= 3 ∶1, M/ P = 7, and water to solid ratio(W/ S)= 0. 7. Light-burned magnesium oxide can improve the viscosity of the initial slurry and the reaction speed of slurry under the water to solid ratio; dead-burned magnesium oxide can improve the strength of the sample, and the combination of them can give consideration to both the reaction rate and the strength of the sample. With the use of high M/ P, the stability of struvite can be improved, and the problems such as cracking of samples and easy decomposition of struvite caused by excessive reaction heat release at low M/ P, the dissolution of unreacted phosphate in underwater environment can be avoided.

High-complicated characteristics in urban space near mountain poses higher demands on urban refined governance. Grasping the dynamic changes in urban land use and land cover in urban space near mountain is greatly significant for achieving coordinated development and building a symbiotic pattern between mountains and cities. Therefore, taking Guilin as a case, a model to identify the space range of urban space near mountain and categories is constructed. Based on the first law of geography, the spatial variation characteristics of mountain-city relationship based on distance, and identifing the space range of urban space near mountain is analyzed in this paper. Then, based on multi-geospatial big data such as POI, remote sensing images, building height data, OSM and so on, a RF identification model is constructed for land use categories in terms of POI features, remote sensing features and building intensity features and verifies the recognition effect, with mapping the urban land use categories. Results show that: (1) According to the size of the mountain, the space range of small mountains is 400 m, the space range of medium-sized mountains is 500 m, and the space range of large mountains is 600 m. (2) The features contained in POI data can effectively identify various land use categories, among which Word2Vec semantic information and TF-IDF information play a great contribution. However, the spectral and textural geatures from remote sensing images and building intensity information from build height data cannot achieve the same effect. This method has achieved remarkable results in identifying land use categories in urban space near mountain, providing technical support and basic data information for future monitoring of the development status of urban space near mountain as well as planning and compilation.

Accurate extraction and identification of individual tree structure information is an important basis for the study of forest parameters such as biological characteristics and composition of trees. In this paper, four algorithms are used to separate individual tree from airborne point cloud data. Through point cloud denoising, data resampling, point cloud classification and other preprocessing, the basic data model of point cloud is generated. Then, four algorithms of segmentation based on CHM, direct segmentation based on point cloud, seed points generated by CHM and seed points generated by layer stacking are respectively used for individual tree segmentation extraction and accuracy comparative analysis, and the optimal segmentation algorithm is determined. It provides a reference for the algorithm of high accuracy individual tree segmentation. The results show that the accuracy of the four segmentation algorithms is above 60%. The algorithm using layer stacking to generate seed points has the highest segmentation accuracy of individual tree, and the accuracy of identifying the number of individual tree is 71. 45%(R 2 = 0. 868), indicating that the algorithm is stable and feasible in the segmentation of individual tree. The results show that the segmentation accuracy is related to the density of the study area, canopy height and seasonal variation.

According to monomer ancient building fine 3D modeling of real problem, an application method of UAV orbital and proximity shooting photogrammetry in fine modeling is proposed, according to the structural characteristics of ancient building complex building and the parameters of the unmanned aerial vehicle lens and image acquisition needs computing overlapping degree around the flightline and close to the flightline, automated image acquisition to route for ancient building high resolution image data. The transition images are captured at the junction of orbital and proximity shooting photogrammetric to realize the image matching between the routes, and the image data is processed to generate a refined real 3D model. The results show that this method can effectively improve the integrity, quality and efficiency of UAV’s collection of fine 3D modeling images of single ancient buildings. The 3D model of ancient buildings has good integrity, real details and textures, and high geometric accuracy, which is conducive to the preservation of 3D digital electronic archives of ancient buildings.

Reduction of bacterial resistance via aerobic composting is a key step for controlling bacterial resistance dissemination of organic solid waste in environment, which attracts wide attention from public and scientists. In recent decades, a lot of related researches have been reported, and these findings are urgent for guiding the futural research in this field. Therefore, this study gathered a total of 132 Chinese publications from the China National Knowledge Infrastructure and 464 English publications from the Web of Science regarding bacterial resistance in composting from 2003 to 2023. A bibliometric approach is subsequently performed using CiteSpace software. The results show that: (1) the number of publications gradually increases during the last 21 years, with significant growth occurring after 2017. (2)Chinese research institutions and teams play the leading role to promote this field in the past five years, by contributing the Top 5 the-highest-publication institutions and the highly-cited publications. (3)The current and frontier key research hotspots are swine manure / tetracycline antibiotics/ microorganisms, and sludge / integrons/ horizontal gene transfer, respectively. The future research is suggested to focus on the following keywords including composting-process optimization, targeted microorganism, and gene-transfer mechanism, for building high-efficiency composting technologies to control antimicrobial resistance of organic solid waste.

In order to address the issues of energy consumption optimization and traffic flow stability for connected autonomous vehicles(CAV) in urban road scenarios during rainy weather, a collaborative control and evaluation framework is proposed based on meteorological perception in this paper. Meteorological perception factors are integrated into the classic intelligent driver model(IDM) and Wiedemann 99 model to construct a meteorology-aware adaptive car-following model(MACM). A parameter calibration algorithm is employed to enable the model to dynamically adapt to driving behavior under different rainy conditions. A safety speed decision-making model is constructed using vehicle communication delay compensation, based on the rain stability conditions of Gipps model. A collaborative control strategy for CAVs in rainy conditions(CCSRC), designed for implementation on rainy days, has been developed to achieve significant reductions in speed fluctuations and meet energysaving objectives. A unified performance index(UPI) and energy consumption model have been established for the purpose of quantitatively evaluating the balance between traffic efficiency and energy consumption. The experiment in a closed-loop simulation environment is built with SUMO, the model??s effectiveness is validated through rainfall intensity classification( no rain, light rain, heavy rain). The findings demonstrate that under conditions of heavy rain, the optimized method(MACM-CCSRC) attains energy savings of 61. 11% for CAVs, accompanied by UPI enhancements of 0. 87. The research findings provide theoretical models and evaluation methods for rain-day CAV cooperative control.

To establish a reasonable microgrid cluster optimized scheduling model is crucial for reducing system economic costs and minimizing environmental pollution. Traditional optimized algorithms often lack sufficient search capabilities and are prone to falling into local optima, making it challenging to achieve global optimal solutions. To overcome these limitations, a microgrid cluster optimized scheduling model based on EJS algorithm is proposed. This model incorporates a Tent chaos mapping strategy during the population initialization phase to enhance diversity, and employs Lévy flight steps to update the position of jellyfish during their passive movement phase, thereby improving global search capabilities and convergence efficiency. Simulation case studies indicate that, compared to PSO, JSA, SOA and SCA, the total cost of EJS model is reduced by 4. 57%, 3. 15%, 3. 67%, and 3. 57%, respectively. The EJS model demonstrates superior performance in terms of economic efficiency, environmental benefits, and convergence efficiency, making it highly promising for practical applications.

In response to the real-time and lightweight requirements of supermarket pedestrian detection models, as well as the high rate of false positives and missed detections caused by scale variations, background interference, and dense crowd occlusions, an improved YOLOv8n-based supermarket pedestrian detection model named YOLOv8-OHN is proposed. This method reconstructs the C2f module with omni-directional convolution (ODConv), significantly enhancing the feature extraction capability of convolution, making the model more suitable for complex backgrounds and densely crowded occlusion scenarios. A enhanced lightweight advanced screening feature pyramid network(ELA-HSFPN) is introduced in the neck network, effectively addressing the multiscale challenges brought by fixed camera capture of pedestrians in supermarkets. The CIoU loss function is integrated with the normalized weighted distance(NWD) loss to reduce sensitivity to positional deviations of small pedestrian targets. Tested on the self-built SEPDataset, the YOLOv8-OHN model achieved mAP@ 0. 50 and mAP@ 0. 50 ∶0. 95 of 96. 2% and 71. 4%, improving by 1. 6 and 2. 2 percentage points compared to the baseline model YOLOv8n, with only 2. 0 M parameters and FLOPs of 5. 2 G, reducing by 33. 3% and 35. 8% compared to YOLOv8n. This improved model achieves high-precision supermarket pedestrian detection while also meeting the deployment requirements of low-power scenarios.

The storage and processing of trajectory data for the construction site safety supervision platform are investigated in this paper. Four objects, namely historical trajectory, trajectory segment, position and stay point, are designed to realize trajectory data storage. The background processing logic of storage-trajectory simplification -storage is developed according to the trajectory storage model. Trajectories are simplified by integrating the geometric-feature-based stay point detection method and the Douglas-Peucker trajectory compression algorithm, and comparative experiments are devised to determine appropriate thresholds. It is verified that the trajectory storage method and the trajectory processing model have a satisfactory effect on the simplification of trajectories, which reduces the storage space of trajectories by about 2/ 3 and significantly improves the calling speed of trajectories, and at the same time, the suitable threshold value is selected so that the trajectory compression and the stay point detection method can retain the original form and important information of trajectories, which effectively optimizes the experience of front-end trajectory visualization of the security supervision platform.

Rural tourism serves as a significant engine for advancing national unity and progress in China. Three villages(Shenpo village, Chashan village, and Chenshuang village) in Guangxi Zhuang Autonomous Region are studied, employing the entropy method and fuzzy-set qualitative comparative analysis(fsQCA) to explore pathways through which rural tourism promotes national unity and progress. Rural tourism effectively promotes national unity and progress through diversified pathways, revealing seven configurational pathways that coalesce into three distinct models: modern governance-led model, shared spiritual homeland linkage model, and collaborative model integrating communication-exchange-integration with livelihood welfare sharing. The pathway combining modern governance with advanced ideological leadership exhibits the strongest explanatory power, representing the most accessible route to achieve national unity and progress. Based on these findings, the study recommends establishing differentiated advantage-driven mechanisms, strengthening the dual-wheel synergistic drive of modern governance and advanced ideological leadership, strengthening the transformation and utilization of cultural resources, building cooperative and mutually supportive tourism communities, embedding rural tourism throughout the entire process of national unity initiatives, and facilitating an elevation from surface-level collaboration to deep-rooted identity recognition.

The main urban area of Guilin is a place where form and beauty of mountains and rivers are densely distributed and which boasts rich historical deposits. A great number of landscape poetry preserved so far are important carriers for studying the distribution of form and beauty in Guilin. This study selects 1 130 landscape poems about Guilin as research samples, and uses methods including typical landscape index, kernel density analysis, Ripley??s K-function and social semantic network analysis to explore the spatio-temporal distribution characteristics of ancient form and beauty in Guilin from three dimensions. Research shows that in Guilin, nodes of form and beauty in the Qing dynasty are more concentrated than those in the Ming dynasty. The first peak of creating form and beauty of landscape appears in the Song dynasty, while the Yuan dynasty is a trough period, and form and beauty in the Ming and Qing dynasties reached prosperity and maturity. Regarding types of ancient form and beauty in Guilin, the development of beautiful mountains and stone forests is taken as the guide, and religious and memorial buildings are built based on the site of mountains, which are also near to mountains and caves, while residences where refined scholars gather are built on the edge of water. The type of nodes of form and beauty presents an inconspicuous discrete distribution. The heat sustainability of Qixing Mountain and Diecai Mountain clusters is the best, while Nanxi Mountain, Diecai Mountain, and Qixing Mountain clusters show relatively large fluctuations in popularity. The ancient form and beauty in Guilin is visually interactive, and the nodes presents a typical landscape combination of “form and beauty-mountains-caves-hydrology-architectures-plants-animals-figures-myths-soundscapes”.