长短时记忆神经网络在中期地震预报中的探索--以川滇地区为例

利用机器学习方法对地震活动大数据进行挖掘,识别出一些过去认识不到的异常,提高地震预测的准确性,是一个非常具有挑战性的科学问题.本文基于川滇部分地区(24°N—32°N,98°E—106°E)地震目录,采用滑动的时空窗口,选取16个反映地震时空强度分布特征的地震预测因子,建立了长短时记忆(LSTM)神经网络,对研究区域9个子区块未来一年的最大地震震级进行预测.通过设置训练集∶测试集=8∶2和训练集∶测试集=7∶3两个不同的训练测试模型,对过往发生的地震进行了回溯性预报.结果表明:训练集∶测试集=7∶3模型能够利用1970年1月至2004年9月的地震目录进行学习,成功回溯性预报2008年汶川地震;训练集∶测试集=8∶2模型利用1970年1月至2009年5月资料进行训练,回溯性预报2010—2019年间6级以上地震的R评分为0.407,回溯性7级地震预报时准确率高达92.31%.本文还探讨性给出预测意见:2022年2月前研究区西部、中部、东部、西南部存在发生5.1~5.3级地震的潜在危险性.     

涠西南凹陷涠洲10-3油田及围区流三段沉积相研究

为了促进涠西南凹陷涠洲10-3油田及围区流三段油藏的滚动勘探开发,开展了油区沉积相精细研究。利用岩芯、录井、测井和地震等资料,在识别不同级别层序界面和湖泛面的基础上,将涠洲10-3油田及围区流三段划分为2个长期基准面旋回和8个中期基准面旋回,建立了高分辨率层序地层对比格架。在岩芯描述基础上,进行单井沉积相、连井沉积相、平面沉积相分析,认为研究区流三段发育冲积扇、扇三角洲和湖泊3种沉积相,并系统划分了沉积微相类型。探讨了碎屑流沉积的相带归属,将发育在河道中的碎屑流沉积划分为河道型碎屑流沉积,将单独产出的碎屑流沉积根据沉积环境不同划分为水上非河道化碎屑流沉积和水下非河道化碎屑流沉积。最后对各中期基准面旋回的沉积相展布特征和演化规律进行了探讨,并建立了符合研究区的沉积相模式。     

Linking pesticides and human health: A geographic information system (GIS) and Landsat remote sensing method to estimate agricultural pesticide exposure

Accurate pesticide exposure estimation is integral to epidemiologic studies elucidating the role of pesticides in human health. Humans can be exposed to pesticides via residential proximity to agricultural pesticide applications (drift). We present an improved geographic information system (GIS) and remote sensing method, the Landsat method, to estimate agricultural pesticide exposure through matching pesticide applications to crops classified from temporally concurrent Landsat satellite remote sensing images in California. The image classification method utilizes Normalized Difference Vegetation Index (NDVI) values in a combined maximum likelihood classification and per-field (using segments) approach. Pesticide exposure is estimated according to pesticide-treated crop fields intersecting 500 m buffers around geocoded locations (e.g., residences) in a GIS. Study results demonstrate that the Landsat method can improve GIS-based pesticide exposure estimation by matching more pesticide applications to crops (especially temporary crops) classified using temporally concurrent Landsat images compared to the standard method that relies on infrequently updated land use survey (LUS) crop data. The Landsat method can be used in epidemiologic studies to reconstruct past individual-level exposure to specific pesticides according to where individuals are located.     

Concentrations and Distribution of Trace Metals of PM_(10) and TSP Particles Collected in the Qingdao Area

Aerosol samples of PM10 (particulate matter with aerodynamic diameters less than 10μm) and TSP (total suspended particles) were simultaneously collected from April 2001 to March 2002 at the top of Mount Baguan on the downtown campus of Ocean University of China, Qingdao, China. The concentrations of Al, Fe, Mn, Cu, Pb and Zn were determined by means of inductively coupled plasma atomic emission spectrometry (ICP-AES). The monthly variability of the mass concentrations of aerosol particles and the concentrations of trace metals are presented and discussed. The distribution pattern of these metals in PM10 and TSP is also discussed. During the observation period, the mass concentration of PM10 at this site ranged from 13.80 to 306.42μgm-3 , while that of TSP ranged from 31.02 to 568.82μgm-3. Both PM10 and TSP reached their highest concentrations in springtime, while the lowest values occurred in summertime. The concentrations of crustal metals followed the same variation pattern, while those of anthropo     

First‐order exchange coefficient coupling for simulating surface water–groundwater interactions: parameter sensitivity and consistency with a physics‐based approach

Distributed hydrologic models capable of simulating fully‐coupled surface water and groundwater flow are increasingly used to examine problems in the hydrologic sciences. Several techniques are currently available to couple the surface and subsurface; the two most frequently employed approaches are first‐order exchange coefficients (a.k.a., the surface conductance method) and enforced continuity of pressure and flux at the surface‐subsurface boundary condition. The effort reported here examines the parameter sensitivity of simulated hydrologic response for the first‐order exchange coefficients at a well‐characterized field site using the fully coupled Integrated Hydrology Model (InHM). This investigation demonstrates that the first‐order exchange coefficients can be selected such that the simulated hydrologic response is insensitive to the parameter choice, while simulation time is considerably reduced. Alternatively, the ability to choose a first‐order exchange coefficient that intentionally decouples the surface and subsurface facilitates concept‐development simulations to examine real‐world situations where the surface‐subsurface exchange is impaired. While the parameters comprising the first‐order exchange coefficient cannot be directly estimated or measured, the insensitivity of the simulated flow system to these parameters (when chosen appropriately) combined with the ability to mimic actual physical processes suggests that the first‐order exchange coefficient approach can be consistent with a physics‐based framework. Copyright © 2009 John Wiley & Sons, Ltd.     

Assessing the Vibrational Frequencies of the Holweide Hospital in the City of Cologne (Germany) by Means of Ambient Seismic Noise Analysis and FE modelling

Ambient seismic noise measurements were conducted inside the Holweide Hospital (Cologne) for checking whether its frequencies of vibration fall into the range where soil amplification is expected. If this is the case, damage might increase in case of an earthquake due to an amplified structural response of the building. Two different techniques were used: the ratio between the horizontal and vertical components of the spectra recorded at stations located inside the building and the ratio between the corresponding components of the spectra recorded simultaneously inside the building and at a reference station placed outside. While the former method might be preferred because of less equipment involved, the latter has the advantage of producing more stable results and deleting automatically the influence of the sedimentary cover, which might obscure some eigenfrequencies of vibration of the building. An independently performed finite-element analysis of the hospital showed a good correlation between measured and calculated eigenmodes.     

Effects of the number of stations and time resolution on derivation

For at least 30 years now it has been well known that the Dst index can be modelled using the solar wind as input. Since then, many attempts have been made to improve the predictability of Dst using different approaches. These attempts are useful, for instance, to understand which features of the solar wind–magnetosphere interactions are most important in producing magnetospheric activity and how the Dst index would improve the space weather forecast. The Dst index is by far the most reliable and simple indication that a magnetic storm is in progress. In this work, the effect of using more than four magnetic stations and shorter time intervals than the hourly averages used in Sugiura's procedures is evaluated. The discussion is based on the results presented by Burton in 1975 and Feldstein in 1984 considering 4 or 12 magnetic stations and time averages of 2.5 min for a magnetic disturbed period that occurred from February 7–28, 1967, including two geomagnetic storms. The analysis has shown that the global representation of a magnetic storm by the standard Dst (Sugiura) is well preserved either using 4, 6, 12 magnetic stations or using 1 h, 2.5 min 1 min averages. A brief review of the current understanding of Dst has been included to support the discussions. The analysis performed has shown that a more refined Dst index (time and number of stations>4) would be useful to investigate the intrinsic processes and the different current systems involved in the ring current development during magnetic storms; the standard Dst, as it is conceived, is quite adequate to monitor geomagnetic storms and identify their overall features; concerning the magnetic stations normally considered, the inclusion of higher magnetic latitude stations (>35) may underestimate the observed Dst.     

全国矿产资源规划信息数据库及管理系统的构建

随着矿产规划管理信息的逐渐增多和矿政管理部门的需求不断提高,本文根据规划管理的业务需求以及第二轮的矿产资源规划标准,建立了全国矿产资源规划信息数据库管理系统。通过Oracle 10g建立矿产资源规划信息数据库,并通过ArcGIS Engine构建管理系统,实现了全国矿产资源规划信息的整合与共享,为矿产规划管理的科学化、精细化提供了依据。     

《测绘法》适应性浅析

分析了新时期测绘发展面临的新形势,对《中华人民共和国测绘法》规定的有关概念内涵、管理体制、运行机制、准入制度等的适应性进行了探讨,对进一步修订《中华人民共和国测绘法》提出了建议。