Artificial intelligence in seismology:Advent,performance and future trends

Realistically predicting earthquake is critical for seismic risk assessment,prevention and safe design of major structures.Due to the complex nature of seismic events,it is challengeable to efficiently identify the earthquake response and extract indicative features from the continuously detected seismic data.These challenges severely impact the performance of traditional seismic prediction models and obstacle the development of seismology in general.Taking their advantages in data analysis,artificial intelligence(AI) techniques have been utilized as powerful statistical tools to tackle these issues.This typically involves processing massive detected data with severe noise to enhance the seismic performance of structures.From extracting meaningful sensing data to unveiling seismic events that are below the detection level,AI assists in identifying unknown features to more accurately predicting the earthquake activities.In this focus paper,we provide an overview of the recent AI studies in seismology and evaluate the performance of the major AI techniques including machine learning and deep learning in seismic data analysis.Furthermore,we envision the future direction of the AI methods in earthquake engineering which will involve deep learning-enhanced seismology in an internet-of-things(IoT) platform.     

2019年初浙江罕见连阴雨过程的成因分析

利用浙江省2019年2月地面气象观测资料、FY 2G红外云图资料、NCEP和ERA Interim再分析资料,对浙江省2019年冬末春初典型的连阴雨天气过程的成因进行了分析,得出以下结论：①连阴雨期间500 hPa极涡呈偶极型,主体持续收缩在北极附近,浙江位于700 hPa和850 hPa切变南侧稳定的西南气流及温度正距平中,对流层高层西风急流持续偏强、位置偏北,强度和位置变化超前于降雨;②源自孟加拉湾北部至青藏高原、中南半岛到南海的两条水汽通道持续输送水汽至我国华东地区,浙江同时受两条水汽带叠加影响时雨强达到最大,红外卫星云图分析可对水汽输送通道和天气形势作大致判断,提供降雨预报辅助性参考;③700 hPa湿Q矢量显著辐合区和850 hPa经向风0 m〖DK〗·s-1等风速线重叠的区域易出现明显降雨,集中降雨时段浙江上空存在很大的低层假相当位温和水汽通量的水平梯度,水汽输送带位于梯度区之上。     

Study the elastic properties and the anisotropy of rocks using different machine learning methods

This paper aims to demonstrate that the elastic stiffnesses and the anisotropic parameters of rocks can be accurately predicted from geophysical features such as the porosity, the density, the compression stress, the pore pressure and the burial depth using relevant machine learning methods. It also suggests that the extreme gradient boosting method is the best method for this purpose. It is more accurate, extremely faster to train and more robust than the artificial neural networks and the support vector machine methods. Very high R-squared scores was obtained for the predicted elastic stiffnesses of a relevant dataset that is available in the literature. This dataset contains different types of rocks, and the values of the features are in large ranges. An optimal set of parameters was obtained by considering an appropriate sensitivity analysis. The optimized model is very easy to implement in Python for practical applications.     

基于数据挖掘分析中药汤剂治疗肾病综合征的用药规律

目的：基于中医传承辅助平台分析中药汤剂治疗肾病综合征(NS)的用药规律。方法：检索中国知网（CNKI）、万方数据(WANFANG DATA)、维普中文科技期刊数据库(VIP)、中国生物医学文献服务系统（SinoMed）、PubMed、Embase、The Cochrane Library等电子数据库,筛选2010年1月1日至2023年7月31日关于中药汤剂治疗NS的相关文献,并采用中医传承辅助平台TCMISS（V2.5）分析其药物功效、性味归经、关联规则等。结果：最终纳入78篇文献,方剂78首,涉及中药共140味。高频药物依次为黄芪、茯苓、白术、泽泻、丹参等。药物功效以补虚、利水渗湿、活血化瘀、清热、收涩为主。药物四气以温、平、寒性为主,五味以甘、苦、辛味为主,归经以脾、肾、肝经为主。关联规则分析得出94组药物组合,出现频次最多的组合为黄芪、茯苓,置信度最高的组合为山药、泽泻→茯苓,并演化得到6组核心药物组合及其新方药物组合。结论：中药汤剂治疗NS以利水渗湿、助阳化气、滋阴补血、健脾益肾药物为主,并根据病情辨证施治,可为临床诊疗提供一定的参考。     

Assessing Worldview-3 multispectral imaging abilities to map the tree diversity in semi-arid parklands

Semi-arid parkland agrosystems are strongly sensitive to climate change and anthropic pressure. In the context of sustainability research, trees are considered critical for various ecosystem services covering environment quality as well as food security and health. But their actual ecological impact on both cropland and natural vegetation is not well understood yet, and collecting spatial and structural information around agroforestry systems is becoming an important issue. Tree mapping in semi-arid parklands could be one of these prerequisites. While for obtaining an exhaustive inventory of individual trees and for analysing their spatial distribution, remote sensing is the ideal tool. However, it has been noted that depending on the spatial resolution and sensor spectral characteristics, tree species cannot be distinguished clearly, even in the sparsely vegetated semi-arid ecosystems of West Africa. Thus, this work focuses on assessing the capabilities of Worldview-3 imagery, acquired in 8 spectral bands, to detect, delineate, and identify certain key tree species in the Faidherbia albida parkland in Bambey, Senegal, based on a ground-truth database corresponding to 5000 trees. The tree crowns are delineated through NDVI thresholding and consecutive filtering to provide object-based radiometric signatures, radiometric indices, and textural information. A factorial discriminant analysis is then performed, which indicates that only four out of the seven most abundant species in the study area can be discriminated: “Faidherbia albida”,” Azadirachta indica”, “Balanites aegyptiaca” and “Tamarindus indica”. Next, random forest and support vector machine classifiers are employed to identify the optimal combination of classifier parameters to discriminate these classes with a high accuracy, robustness, and stability. The linear support vector machine with cost=1 and gamma=0.01 provides the optimal results with a global accuracy of 88 % and kappa of 0.71. This classifier is applied to the whole study area to map all the trees with crowns larger than 2 m, sorted in four identified species and a fifth common group of unidentified species. This map thus enables analysing the variability in tree density and the spatial distribution of different species. Such information can afterwards be correlated to the ecological functioning of the parkland and local practices, and offers promising opportunities to help future sustainability initiatives in different socio-ecological contexts.     

Automated lithological mapping by integrating spectral enhancement techniques and machine learning algorithms using AVIRIS-NG hyperspectral data in Gold-bearing granite-greenstone rocks in Hutti,India

In this study, we proposed an automated lithological mapping approach by using spectral enhancement techniques and Machine Learning Algorithms (MLAs) using Airborne Visible Infrared Imaging Spectroradiometer-Next Generation (AVIRIS-NG) hyperspectral data in the greenstone belt of the Hutti area, India. We integrated spectral enhancement techniques such as Principal Component Analysis (PCA) and Independent Component Analysis (ICA) transformation and different MLAs for an accurate mapping of rock types. A conjugate utilization of conventional geological map and spectral enhancement products derived from ASTER data were used for the preparation of a high-resolution reference lithology map. Feature selection and extraction methods were applied on the AVIRIS-NG data to derive different input dataset such as (a) all spectral bands, (b) shortwave infrared bands, (c) Joint Mutual Information Maximization (JMIM) based optimum bands, and (d) optimum bands using PCA, to choose optimum input dataset for automated lithological mapping. The comparative analysis of different MLAs shows that the Support Vector Machine (SVM) outperforms other Machine Learning (ML) models. The SVM achieved an Overall Accuracy (OA) and Kappa Coefficient (k) of 85.48% and 0.83, respectively, using JMIM based optimum bands. The JMIM based optimum bands were more suitable than other input datasets to classify most of the lithological units (i.e. metabasalt, amphibolite, granite, acidic intrusive and migmatite) within the study area . The sensitivity analysis performed in this study illustrates that the SVM is less sensitive to the number of samples and mislabeling in the model training than other MLAs. The obtained high-resolution classified map with accurate litho-contacts of amphibolite, metabasalt, and granite can be coupled with an alteration map of the area for targeting the potential zone of gold mineralization.     

Delineation of management zones in agricultural fields using cover–crop biomass estimates from PlanetScope data

Several methods have been proposed to delineate management zones in agricultural fields, which can guide interventions of the farmers to increase crop yield. In this study, we propose a new approach using remote sensing data to delineate management zones at three farm sites located in southern Brazil. The approach is based on the hypothesis that the measured aboveground biomass (AGB) of the cover crops is correlated with the measured cash-crop yield and can be estimated from surface reflectance and/or vegetation indices (VIs). Therefore, we used seven different statistical models to estimate AGB of three cover crops (forage turnip, white oats, and rye) in the season prior to cash-crop planting. Surface reflectance and VIs were used as predictors to test the performance of the models. They were obtained from high spatial and temporal resolution data of the PlanetScope (PS) constellation of satellites. From the time series of 30 images acquired in 2017, we used the PS data that matched the dates of the field campaigns to build the models. The results showed that the satellite AGB estimates of the cover crops at the date of maximum VI response at the beginning of the flowering stage were useful to delineate the management zones. The cover-crop AGB models that presented the highest coefficient of determination (R²) and the lowest root mean square (RMSE) in the validation and test datasets were Support Vector Machine (SVM), Cubist (CUB) and Stochastic Gradient Boosting (SGB). For most models and cover crops, the Enhanced Vegetation Index (EVI) and the Normalized Difference Vegetation Index (NDVI) were the two most important AGB predictors. At the date of maximum VI at the beginning of the flowering stage, the correlation coefficients (r) between the cover-crop AGB and the cash-crop yield (soybean and maize) ranged from +0.70 for forage turnip to +0.78 for rye. The fuzzy unsupervised classification of the cover-crop AGB estimates delineated two management zones, which were spatially consistent with those obtained from cash-crop yield. The comparison between both maps produced overall accuracies that ranged from 61.20% to 68.25% with zone 2 having higher cover-crop AGB and cash-crop yield than zone 1 over the three sites. We conclude that satellite AGB estimates of cover crops can be used as a proxy for generating management zone maps in agricultural fields. These maps can be further refined in the field with any other type of method and data, whenever necessary.     

确定坡面径流过程曼宁糙率系数的实验方法研究

曼宁糙率系数是用水动力学方法进行流速计算的关键参数。坡面流曼宁糙率系数与明渠流的不同。为确定坡面径流过程的曼宁糙率系数,自行研发了一种包括供水系统、实验水槽和数据观测记录系统的室内可变糙率坡面实验系统。通过87场预实验验证了供水系统的稳定性和准确性。以坡度、实测流量、实测水深、不同糙率板上河砂的平均直径和地表粗糙度为自变量,以曼宁糙率系数为因变量,选用均方根误差(RMSE)和决定系数(R ²)为评价指标,对166种实验场景进行了支持向量机(Support Vector Machines, SVM)训练与预测,发现：① 紊流的训练结果难以预测层流和过渡流的曼宁糙率系数,说明流态不同时,实验因素对水流的影响机制不同;② 若要较为准确地预测曼宁糙率系数,至少需要包括实测水深在内的3种因素;③ 当同时考虑4种及更多种因素时,紊流状态下均可对曼宁糙率系数进行较为准确的预测。     

基于SVM的海浪要素预测试验研究

采用支持向量机对海浪要素中的有效波高进行预测,采用风场和波浪场作为学习要素,对比不同特征向量对有效波高预测结果的准确度。取台湾岛东部海区作为实验区域,使用NCEP再分析的数值模式数据作为学习样本。选用支持向量分类机,建立了4组不同特征向量的模型进行海浪有效波高的预测,并对4种模型的结果进行比较和分析。实验表明,当输入的特征向量过多或过少时,会对模型的预测结果和计算效率产生不同的影响。当使用风场和波浪场共同作为特征向量进行学习时,在该区域预测结果与模式预报结果相比更接近,相关系数将近99%,均方根误差约0.2 m。     

基于珞珈一号夜间灯光数据的广州市建设用地识别

夜间灯光数据和人类活动密切相关,可用于识别城市建设用地。目前主要利用DMSP/OLS和NPP-VIIRS夜间灯光数据进行建设用地识别,由于数据质量原因,这两类数据的识别结果精度较差。珞珈一号夜间灯光数据与比以往夜间灯光数据相比,时间分辨率、空间分辨率和光谱分辨率明显提升,是进行建设用地提取的更理想的数据源。本研究首先对珞珈一号夜间灯光数据进行辐射和影像配准,提高数据质量,然后利用支持向量机(Support Vector Machine, SVM)影像分类方法对广州市2017 年建设用地分区识别,并利用Kappa系数分区、分地类评价识别结果精度。研究发现：① 利用珞珈一号夜间灯光数据识别建设用地的精度明显优于利用DMSP/OLS和NPP-VIIRS夜间灯光数据识别结果的精度;② 广州市中心城区辖区的建设用地识别结果精度较高,识别结果Kappa系数均在0.9以上;外围辖区识别结果精度相对较低,识别结果Kappa系数为0.85左右;③ 城市、建制镇等单个地块面积较大、灯光亮度较高的地类识别结果精度较高,识别结果Kappa系数均在0.9以上;村庄用地、铁路公路用地由于单个地块面积小、布局比较分散、部分路段无照明条件等原因,识别结果Kappa系数相对较低,为0.85左右;采矿、风景及特殊用地夜间基本无人类活动,缺少夜间灯光,难以用夜间灯光数据识别,Kappa系数为0.45左右。本研究证明了利用珞珈一号夜间灯光数据能有效识别建设用地,同时丰富了珞珈一号夜间灯光数据的应用场景。