Understanding the surface thermal environment is vital for assessing climate change, ecosystem health, and the urban heat island effect, as well as comprehending global surface temperature and long-term trends. Based on eastern Zhejiang, we utilised the multi-temporal Landsat data during 2013-2023 to obtain the surface temperature with the Landsat8/9 Level2 temperature product. We analysed the influence of the surface thermal environment on different land cover types, different geomorphological types, and seasonal variations by the elevation, slope, aspect, shaded relief, NDVI (Normalised Difference Vegetation Index), NDBI (Normalised Difference Built-up Index), and MNDWI (Modified Normalised Difference Water Index). We analysed seasonal variability, geomorphic units, shadow terrain, and correlations between landforms and the surface thermal environment using a regression model. Elevation and slope showed significant negative correlations in forested and shaded areas, especially in autumn and winter (R2 = 0.56 in full map), and were stronger in mountains than plains. Seasonal solar radiation caused slope-temperature differences, most pronounced in winter very correlated with temperature in mountainous areas except summer. NDVI was negatively correlated (R2 = 0.19 in full map), highlighting vegetation's cooling effect, while NDBI was strongly positive (R2 = 0.45 in full map) due to impervious surfaces and reduced ventilation. MNDWI showed a cooling effect (R2 = 0.14 in full map) linked to water and humidity. (c) 2025 The Author(s). Published by Elsevier B.V. on behalf of COSPAR. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).

Fine-grained hyperspectral mineral classification remains challenging due to spectral homogeneity among minerals with different morphologies, severe spectral mixing from intergrowth, and high dimensionality. Existing methods rely on spectral separability assumptions, which become insufficient when spectral differences are subtle and spatial-structural ambiguity is high. To address these limitations, we propose S3AM-ECA-3DCNN, a spectral-structural collaborative feature learning framework. It uses a 3DCNN backbone to jointly model spectral-spatial features with progressive spectral downsampling. The spectral-similarity-based spatial attention module (S3AM) performs spatial purification by suppressing interference from spectrally mixed neighboring regions, and the efficient channel attention (ECA) module adaptively recalibrates discriminative spectral bands to enhance fine-grained representation. This establishes a spatial-first, channel-second collaborative optimization paradigm. To improve generalization under limited training samples, adaptive global pooling and a lightweight classification head are employed to reduce model complexity and mitigate overfitting. Experiments on a 146-class hyperspectral mineral dataset (covering silicates, carbonates, and sulfates) show that the framework achieves 93.424% overall accuracy, 91.099% average accuracy, and a Kappa coefficient (×100) of 93.368, outperforming mainstream methods. It significantly reduces misclassification among spectrally similar but morphologically distinct minerals, demonstrating strong robustness and discriminative capability for large-scale fine-grained tasks.

Hyperspectral remote sensing technology, with its ability to acquire continuous and fine-scale spectral information, shows potential in mineral classification but faces technical challenges such as spectral variability within the same material and spectral similarity among different materials. In previous work, we explored rock classification using various hyperspectral classification methods; however, many minerals were found to exhibit inherently similar spectral signatures, which limited the classification performance of these methods. In this study, we aim to address this issue by constructing a preliminary mineral classification system using laboratory hyperspectral data to redefine categories from a spectral perspective. Specifically, minerals are reorganized into new classes based on their spectral curve characteristics rather than traditional mineralogical definitions, thereby establishing a spectrally driven framework for evaluating separability among spectral groups. A total of 250 mineral specimens were imaged in a darkroom environment using shortwave infrared bands to obtain high-resolution spectral data. After preprocessing, minimum noise fraction (MNF) transformation was applied for dimensionality reduction. Based on their spectral curve characteristics, the 250 mineral specimens were reclassified into 33 spectrally defined preliminary groups. Then, support vector machine (SVM) and random forest (RF) were applied to compare the classification performance of the original individual mineral identification task and the spectrally defined group task, using accuracy, Kappa, and metrics calculated for each class. The experiments demonstrate that the classification scheme based on 33 spectral groups improves classification performance among groups and partially reduces the spectral confusion occurring in the original individual mineral identification task. This framework offers new insights for laboratory-scale mineral research and provides a preliminary experimental basis for future testing under field and airborne/spaceborne conditions.

To address overfitting due to limited sample size, and the challenges posed by “Spectral Homogeneity with Material Heterogeneity (SHMH)” and “Material Consistency with Spectral Divergence (MCSD)”—which arise from subtle spectral differences and limited classification accuracy—this study proposes a deep integration model that combines the Adaptive Boosting (AdaBoost) algorithm with a convolutional recurrent neural network (CRNN). The model adopts a dual-branch architecture integrating a 2D-CNN and gated recurrent unit to effectively fuse spatial and spectral features of rock samples, while the integration of the AdaBoost algorithm optimizes performance by enhancing system stability and generalization capability. The experiment used a hyperspectral dataset containing 81 rock samples (46 igneous rocks and 35 metamorphic rocks) and evaluated model performance through five-fold cross-validation. The results showed that the proposed 2D-CRNN-AdaBoost model achieved 92.55% overall accuracy, which was significantly better than that of other comparative models, demonstrating the effectiveness of multimodal feature fusion and ensemble learning strategy.

Bakanae is a fungal rice disease that is threatening global rice production, causing severe yield losses. The plant microbiome plays a significant role in plant stress resistance, but its high-dimensional characteristics have not been fully exploited. Therefore, we integrated the microbiome and machine learning (ML) to diagnose bakanae disease in this study. We found significant correlations between Gammaproteobacteria and Bacteroidia and the severity of bakanae disease. We constructed different diagnosis models based on random forests (RF), support vector machines (SVM), and convolutional neural networks (CNN) on 88 biological replicates with an independent test set. We found that the RF model demonstrated strong performance across four taxonomic levels, with an accuracy of 88.9% and an F1 score of 94.1%. Notably, a Bray-Curtis dissimilarity-based extraction method was proposed to rapidly screen practical information from the original microbial community, which can enhance the model performance to a certain extent. According to phenotypic data, the disease severity of infected samples was classified into two levels (high and low infected levels) using the K-means clustering method. In the diagnosis of infection severity based on the family level, the model's prediction accuracy reached 77.8%. Collectively, these findings highlight that the combination of microbiome with ML can advance diagnostic strategies for bakanae disease, providing new avenues for precision agriculture.

Geothermal resources are renewable and clean energy sources with diverse applications. However, it only accounted for less than 1% of all thermal energy use worldwide in 2021. Successfully locating and drilling geothermal wells is challenging. Our study aimed to establish an effective model to evaluate the geothermal potential in the Hangjiahu Plain of China. We incorporated lithology, distances to faults, surface water system, seismic points, magmatic and volcanic rocks, and remote sensing data, including land surface temperature, normalized difference vegetation index, and slope into the geothermal potential evaluation system and for the first time, compared the performance of three machine and one deep learning models: support vector machine, adaptive boosting, light gradient boosting machine (LightGBM) and TabNet for predicting the geothermal potential in the Hangjiahu Plain. Results showed that TabNet had the best performance, achieving an accuracy of 95.95%. LightGBM closely followed TabNet, with an accuracy of 91.89% but exhibited ten times higher efficiency than TabNet in training complex geothermic data. Compared to other models, the map of areas having high geothermal potential generated by TabNet fits the distribution of geothermal drilling points with high outlet water temperatures, which further indicates that TabNet has good predictive performance.

Hyperspectral remote sensing is a cutting-edge technology in remote sensing,which has the characteristics of multi-band and high spectral resolution,so it is increasingly widely used in rock identification and classification.In the current study of hyperspectral rock classification,many rocks are easily confused because of their similar mineral composition,and the classification accuracy is not always high.In the study of high spectral lithology in a wide range of field conditions,there is a lot of interference from the external environment,such as ground cover,pixel mixing and so on,so the spectral characteristics of rocks need to be further studied.The rocks with similar spectra are reclassified.In this study,from the point of view of the laboratory hyperspectral remote sensing system,the HySpex hyperspectral images of 81 common magmatic and metamorphic rock samples were taken as the research data images,and the images were preprocessed such as reflectance correction.Combined with the spectra of rock samples measured by ASD to verify the extraction of corresponding sample spectral curves in the images,the spectral information representing each rock sample was extracted,and the spectral similarity was classified.Finally,the preliminary classification system of 9 large and 28 small categories based on 81 rock samples is obtained.The initial classification system has similar composition properties and spectral characteristics of rock samples in large classes.The spectral characteristics of small classes are more similar than those of large classes.In order to verify the effect of preliminary classification experience on computer lithology classification,the follow-up study is based on the initial classification system of rock samples,and the minimum noise separation technique is used to extract the feature information of hyperspectral images.Finally,the computer classification algorithm model uses the maximum likelihood method and random forest classification,and the training samples set each rock as a single rock book and each subclass in the initial classification system as a sample.Complete the hyperspectral image classification of common magmatic and metamorphic rocks.The experimental results show that the accuracy of maximum likelihood method and random forest classification based on traditional model is 83.21％and 83.63％,while the accuracy of maximum likelihood classification and random forest classification based on initial classification can be improved to 85.46％and 89.39％.Random forest classifier is superior to the traditional maximum likelihood method,while the rock primary classification system has some advantages compared with simple original rock classification.It can be used as a reference for future rock classification work.

The issues of the same material with different spectra and the same spectra for different materials pose challenges in hyperspectral rock classification. This paper proposes a multidimensional feature network based on 2-D convolutional neural networks (2-D CNNs) and recurrent neural networks (RNNs) for achieving deep combined extraction and fusion of spatial information, such as the rock shape and texture, with spectral information. Experiments are conducted on a hyperspectral rock image dataset obtained by scanning 81 common igneous and metamorphic rock samples using the HySpex hyperspectral sensor imaging system to validate the effectiveness of the proposed network model. The results show that the model achieved an overall classification accuracy of 97.925% and an average classification accuracy of 97.956% on this dataset, surpassing the performances of existing models in the field of rock classification.

利用 HySpex高光谱成像系统,对 125 块岩石样本和 250 块矿物样本进行近红外高光谱测量,提取岩矿的高光谱影像数据集,将岩矿光谱数据制作单种岩石矿物光谱卡,以岩矿影像高光谱数据集的方式进行编排和整理.基于光谱特征对岩石样品进行分类实验,分类效果较好.

Geothermal resources are one of the most valuable renewable energy sources because of their stability, reliability, cleanliness, safety and abundant reserves. Efficient and economical remote sensing and GIS (Geographic Information System) technology has high practical value in geothermal resources exploration. However, different study areas have different geothermal formation mechanisms. In the process of establishing the model, which factors are used for modeling and how to quantify the factors reasonably are still problems to be analyzed and studied. Taking Hangjiahu Plain of Zhejiang Province as an example, based on geothermal exploration and remote sensing interpretation data, the correlation between the existing geothermal hot spots and geothermal related factors was evaluated in this paper, such as lithology, fault zone distance, surface water system and its distance, seismic point distance, magmatic rock and volcanic rock distance, surface water, farmland, woodland temperature and so on. The relationship between geothermal potential and distribution characteristics of surface thermal environment, fault activity, surface water system and other factors was explored. AHP (Analytic Hierarchy Process) and BP (Back Propagation) neural network were used for establishing geothermal potential target evaluation models. The potential geothermal areas of Hangjiahu Plain were divided into five grades using geothermal exploration model, and most geothermal drilling sites were distributed in extremely high potential areas and high potential areas. The results show that it is feasible to analyze geothermal potential targets using remote sensing interpretation data and geographic information system analysis databased on analytic hierarchy process analytic hierarchy process and back propagation neural network, and the distribution characteristics of surface thermal environment, fault activity, surface water system and other related factors are also related to geothermal distribution. The prediction results of the model coincide with the existing geothermal drilling sites, which provides a new idea for geothermal exploration.

The purpose of this study is to investigate the relationship between the natural environment and orthopedic diseases based on remote sensing in Zhejiang Province, China. The Landsat 8 OLI images were employed to extract environmental factors such as the vegetation, water, and urban indices. Combined with the distribution data of patients of different ages diagnosed with different types of orthopedic diseases, derived from the Second Affiliated Hospital, Zhejiang University School of Medicine, we analyzed the spatial distribution relationship and evaluated the natural environmental factors around the distribution sites of the patients. The results showed that the vegetation index (NDVI) is negatively correlated with the prevalence of hospital visits, whereas the water index (MNDWI) is positively correlated. And urban index (IBI) is positively correlated but unstable. The analysis and evaluation of the impact of natural environmental factors related to the patient on orthopedic diseases show that the risk of orthopedic disease might be associated with less vegetation and more water, and are not related to urbanization. And this conclusion is stable in the four seasons.

Rice exposure to cold stress will be alleviated by the decrease of cold extremes under climate change. The disappearance time of cold exposure and its response to different potential scenes in the 21st century, should be particularly important to make long-term adaptations in China. Given these issues unresolved, here we assessed the decadal changes of rice exposure to cold stress in China throughout the 21st century, and compared the patterns under different combinations of Representative Concentration Pathways (RCP) scenarios, rice phenology shifts and cold tolerance enhancements. Under RCP4.5 with no changes in rice phenology and cold tolerance (i.e., the baseline scene), cold exposure would disappear mainly after the 2030s, 2040s, 2060s and 2080s in Liaoning, Fujian, Jilin and Hunan, respectively. But in Heilongjiang, northern Yunnan, southern Sichuan, western Hunan and northwestern Zhejiang, cold exposure would remain until the end of this century. Compared with the baseline pattern in these provinces, RCP8.5 and cold-tolerance enhancements would make cold exposure disappear 2–3 decades early and reduce the intensity by more than 40%. In southern China, when rice thermal-sensitive stage is moved 10/15 days forward, the disappearance of cold exposure would be 2–4 decades early relative to the baseline pattern. Comparison of the patterns under the baseline scene and other potential scenes suggests cautious optimism about the effectiveness of adjustments in crop phenology and provides quantitative support to the enhancement in cold tolerance. This has broad implications for the research community relative to adaptation planning under climate change.

社交媒体数据可以为台风灾害追踪、灾时救援和灾情评估提供及时有效的信息.现有研究常采用主题建模和情感分析等技术对台风期间社交媒体平台(如新浪微博等)舆论话题和情感变化进行研究.在省域范围内以小时为时间粒度的多维度有效性论证尚有欠缺,且在舆情分析时未能区分用户群体差异.本文以台风"利奇马"为例,在浙江省域范围内,以新浪微博数据为研究对象,首先从词频分析、台风关注度时空变化以及特定灾害事件响应3个角度探讨了微博数据对台风灾情响应的有效性;其次采用隐含狄利克雷分布(Latent Dirichlet Allocation,LDA)主题模型技术挖掘微博文本主题信息,并根据Louvain算法对主题社团进行划分;然后开发了一种基于自定义情感词典的情感分析方法用于情感指数计算,与SnowNLP相比情感倾向性预测精度得到了提高;最后分析了台风期间官方和民众在新浪微博平台上的话题关注以及情感演变差异.结果 表明:①在省级范围内,微博数据能有效反映台风动态和灾害时空分布;②台风事件微博文本的主题变化反映了灾情不同阶段舆论关注点的动态变化;③官方微博文本比民众微博文本具有更明确的主题社团结构;④台风事件相关微博文本中的消极情绪在台风登陆后显著增加,其中民众微博文本对台风灾害的情绪响应更及时,官方微博文本中的情感表达始终相对积极.

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.

In addition to human activities, this study found that topography is also an important factor affecting land surface temperature (LST). In this paper, based on Landsat 8 OLI/TIRS remote sensing images, a radiative transfer model was adopted to retrieve the LST, and a maximum likelihood method was used to remove artificial environmental interference factors, such as water bodies and built-up lands. This paper aims to analyze the influence of topographic factors, such as elevation, slope, aspect and shaded relief, on the LST of Hangzhou. By means of a statistical analysis, we obtained the quantitative relationship between these factors and constructed a multiple linear regression model of terrain factors and LST. The research revealed the following findings: (1) in the study area, elevation and slope are negatively correlated with LST, and all the factors have linear relationships with LST. (2) The relationship between aspect and LST is not significant, and high values of LST are found on the southern, southeastern and southwestern slopes; the lowest values are found on the northern slopes. (3) There is a significant linear relationship between the values of the shaded relief map and LST, and the more shadows there are, the lower the LST value will be. (4) After comprehensive analysis of the influence of the abovementioned topographic factors on the LST, it is found that shaded relief has the greatest contribution and is positively correlated with LST. The influence of shaded relief on surface thermal environment should be paid more attention in the process of surface thermal environment work. The assessment of the influence degree of shaded relief and surface thermal environment should be the premise and basis for many other studies.

本次研究中使用1964年的锁眼卫星Keyhole影像并结合1990年、2000年、2010年的Landsat TM、ETM+影像和2018年的高分一号GF-1遥感影像,对杭州闲林闲湖城矿山研究区进行矿山环境变化分析研究,结果表明:1964年～1990年:矿山的面积增加23.793%,草地面积大量减少;1990年～2000年:矿山废弃,大力建设城镇用地,待建设土地增长较快,城市面积增幅达到5.646% ;2000年～2010年:水体面积增幅达451.951%,城镇用地面积增幅为4.739% ;2010年～2018年:水体和草地的面积持续增加,未利用地类面积减少.总结1964年～2018年这54年间,闲林矿山的面积持续减少,城市的覆盖面积增长明显,废弃后的矿山变成了东海闲湖城的居民地,得到了环境保护利用和开发.

å ¨çƒæ°”å€™å˜æš–ä¸Žäººç±»æˆšæˆšç›¸å ³ï¼Œç ”ç©¶å‘çŽ°ä¸ä» æ˜¯åŸŽå¸‚çš„çƒ­å²›æ•ˆåº”å¼•èµ·äº†å±€éƒ¨åŒºåŸŸçš„åœ°è¡¨çƒ­çŽ¯å¢ƒåˆ†å¸ƒå·®å¼‚ï¼Œå¾ˆå¤šåœ°è´¨æž„é€ ã€å²©æ€§ã€åœŸå£¤ã€æ¤è¢«ç­‰åœ°è´¨åŠè‡ªç„¶åœ°ç†å› ç´ ä¹Ÿæœ‰å½±å“ã€‚æœ¬ç ”ç©¶é€‰å–æµ™æ±Ÿçœä½œä¸ºç ”ç©¶åŒºï¼Œåˆ©ç”¨Landsat 8 OLI/TIRSé¥æ„Ÿå½±åƒåæ¼”åœ°è¡¨æ¸©åº¦ï¼Œåˆ†æžå†¬å­£åœ°è¡¨çƒ­çŽ¯å¢ƒåˆ†å¸ƒåŠå ¶å½±å“å› å­ã€‚ç»“æžœè¡¨æ˜Žï¼Œæ–­è£‚å¸¦é™„è¿‘åœ°è¡¨çƒ­çŽ¯å¢ƒå—åˆ°æ–­è£‚å¸¦åˆ†å¸ƒå½±å“ï¼›å²©çŸ³å’ŒåœŸå£¤é€šè¿‡ä¸åŒçš„åœ°è¡¨è¦†è¢«ç±»åž‹å½±å“åœ°è¡¨æ¸©åº¦ã€‚åœ°è¡¨çƒ­çŽ¯å¢ƒåˆ†å¸ƒä¸Žæ–­è£‚å¸¦ç­‰è‡ªç„¶å› ç´ å­˜åœ¨ä¸€å®šç›¸å ³æ€§ã€‚

使用1993年、2003年、2008年和2015年4期Landsat TM/OIL影像,对杭嘉湖平原区进行土地利用/覆被变化(LUCC)分析,结果表明:1)1993-2015年,湿地面积持续减少,总缩幅高达46.14％,主要受城市化进程影响;2)1993-2003年,城市面积由50.18 km2扩增至118.13 km2,湿地面积同期缩小99.2 km2;3)2003-2008年,湿地退化程度较其他时期更为明显,总缩幅达51.09％,减少的湿地类型主要转变为城市和休耕地;同样城市扩张也最为显著,年均增长近30％;4)2008-2015年,在整治政策的保护下,湿地面积实现正增长,年平均增幅为6.32％.总体而言,杭嘉湖平原湿地格局变化受自然因素和社会因素共同作用,但人为因素占主导,2008年之前侧重城镇发展,将湿地用地变为城镇或农业用地,近年来在政策引导下,推动湿地保护,使其面积有所增加.

In this study, topographic map and MSS/TM, SPOT5 satellite remote sensing data were used to extract the boundaries of urban land of Shangyu city .GIS technology with a measure of different urban expansion indexes gives a hand to analysis Shangyu urban land expansion .We have found that from the 1970 s to the 2010 s, Shangyu urban land had been expanding .The extension strength in-dex maintained a high stability after explosive growth in the early period .The city compact index was continually reducing while the urban sub-type dimension index remained stable after substantial decline .Urban areas have become increasingly complex while the boundaries have become tortuous and broken .Thus, future municipal planning and construction should be focus on urban compact and improving the efficiency of urban service functions .

新疆雅满苏地区铁、铜矿资源十分丰富,区内有雅满苏铁矿、天湖铁矿、黑峰山铁矿、沙泉子铜铁矿等.区内变质岩岩性与成矿相关.采用多源遥感数据信息提取方法,利用ASTER与WorldView-2遥感协同影像数据,结合Fieldspec 3地物波谱仪实测的光谱曲线,选取主要岩性的光谱吸收特征对应的影像波段,并通过波段比值法与主成分分析法等,增强区域内的主要岩性特征.最后选取不同波段比值与主成分分析分量因子,即WorldView-2的波段4/8,ASTER的波段7/6,ASTER波段10、11、12、13主成分分析的第二分量因子PC2,组成假彩色合成影像,增强区域的变质岩岩性特征,使得岩性与地质体边界更加清晰.与研究区已有的地质资料、野外地质调查结果对比,分类结果匹配较好,能够为变质岩区域的遥感地质调查与找矿工作提供一定的参考作用.