基于深度学习卷积神经网络的地震波形自动分类与识别

发展高效、高精度、普适性强的自动波形拾取算法在地震大数据时代背景下显得越来越重要.波形自动拾取算法的主要挑战来自如何适应不同区域的不同类型地震事件的分类与筛选.本文针对地震事件-噪音分类这一问题, 使用13839个汶川地震余震事件建立数据集, 应用深度学习卷积神经网络(CNN)方法进行训练, 并用8900个新的汶川余震事件作为检测数据集, 其训练和检测准确率均达到95%以上.在对连续波形的检测中, CNN方法在精度和召回率上优于STA/LTA和Fbpicker传统方法, 并能找出大量人工挑选极易遗漏的微震事件.最后, 我们应用训练好的最优模型对选自全国台网的441个台站8天的连续波形数据进行了识别、到时挑取及与参考地震目录关联, CNN检出7016段波形, 用自动挑选算法拾取到1380对P, S到时, 并与540个地震目录事件成功关联, 对1级以上事件总体识别准确率为54%, 二级以上为80%, 证明了CNN模型具有泛化能力, 初步展示了CNN在发展兼具效率、精度、普适性算法, 实时地震监测等应用上具有巨大潜力.

     

Finite Prandtl number 2-D convection at high Rayleigh numbers

Finite Prandtl number thermal convection is important to the dynamics of planetary bodies in the solar system. For example, the complex geology on the surface of the Jovian moon Europa is caused by a convecting, brine-rich global ocean that deforms the overlying icy “lithosphere”. We have conducted a systematic study on the variations of the convection style, as Prandtl numbers are varied from 7 to 100 at Rayleigh numbers 106 and 108. Numerical simulations show that changes in the Prandtl number could exert significant effects on the shear flow, the number of convection cells, the degree of layering in the convection, and the number and size of the plumes in the convecting fluid. We found that for a given Rayleigh number, the convection style can change from single cell to layered convection, for increasing Prandtl number from 7 to 100. These results are important for determining the surface deformation on the Jovian moon Europa. They also have important implications for surface heat flow on Europa, and for the interior heat transfer of the early Earth during its magma ocean phase. Electronic Supplementary Material is available if you access this article at . On that page (frame on the left side), a link takes you directly to the supplementary material. Electronic Supplementary Material is available if you access this article at . On that page (frame on the left side), a link takes you directly to the supplementary material.     

Temperature dependence of CO2 solubility in high pressure quenched glasses of diopside composition

The temperature dependence of carbon dioxide solubility in glasses of diopside composition, quenched from 20 kbar, has been investigated using a combination of high-temperature mass spectrometry and Raman spectroscopy.CO₂-charged diopside glasses were synthesized in a piston-cylinder apparatus. Because of diffusion of hydrogen through the platinum capsules, significant amounts of H₂O, CH₄ and CO were detected along with CO: in the diopside glasses. All three carbon species show a bimodal release pattern in the mass pyrograms. The CO₂ solubility shows a linear and negative temperature dependence. We do not observe any maxima in the solubility curve as was reported previously (Mysen and virgo, 1980a).None of the additional bands observed in Raman spectra of CO₂-charged diopside glasses compared to those in the spectrum of diopside glass can be assigned to molecular CO₂. These bands are caused by CO^?2₃ ions and indicate that the physical solubility of molecular carbon dioxide is negligible. The bimodal release pattern observed for CO₂ in the mass pyrograms, is consistent with the Raman data which strongly suggests that CO^?2₃ ions are present in at least two distinct sites in the glass.     

Pleistocene oceanographie changes indicated by deep sea benthic foraminifera in the northern Indian Ocean

An attempt has been made to understand the Pleistocene bottom water history in response to the paleoclimatic changes in the northern Indian Ocean employing quantitative analyses of deep sea benthic foraminifera at the DSDP sites 219 and 238. Among the 150 benthic foraminifera recorded a few species show dominance with changing percent frequencies during most of the sequence. The dominant benthic foraminiferal assemblages suggest that most of the Pleistocene bottom waters at site 219 and Early Pleistocene bottom waters at site 238 are of North Indian Deep Water (NIDW) origin. However, Late Pleistocene assemblage at site 238 appears to be closely associated with a water mass intermediate between North Indian Deep Water (NIDW) and Antarctic Bottom Water (AABW). Uvigerina proboscidea is the most dominant benthic foraminiferal species present during the Pleistocene at both the sites. A marked increase in the relative abundance ofU. proboscidea along with less diverse and equitable fauna during Early Pleistocene suggests a relative cooling, an intensified oceanic circulation and upwelling of nutrient rich bottom waters resulting in high surface productivity. At the same time, low sediment accumulation rate during Early Pleistocene reveals increased winnowing of the sediments possibly due to more corrosive and cold bottom waters. The Late Pleistocene in general, is marked by relatively warm and stable bottom waters as reflected by low abundance ofU. proboscidea and more diverse and equitable benthic fauna. The lower depth range for the occurrence ofBulimina aculeate in the Indian Ocean is around 2300 m, similar to that of many other areas.B. aculeata also shows marked increase in its abundance near the Pliocene/Pleistocene boundary while a sudden decrease in the relative abundance ofStilostomella lepidula occurs close to the Early/Late Pleistocene boundary.     

Nature and the Singapore resident

Singapore has witnessed rapid urban development in the last thirty years. Its landscape has changed from one of slums and squatters in the immediate post-war years to a city of gardens and high-rise high-density housing. The concept of a `garden city' was introduced to guide the development of a city-state with abundant greenery. In the process, many parks and gardens were created including park connectors which are green linkages between parks and other land uses. This paper presents the findings of an ongoing research into park connectors. It reports on that part of the survey results which examine the views of residents towards nature provision in park connectors and suggest some design options to enhance the planning of park connectors. This revised version was published online in July 2006 with corrections to the Cover Date.     

Noble gases in ureilites: cosmogenic, radiogenic, and trapped components

Abundances and isotopic compositions of Ne (in bulk samples only), Ar, Kr, and Xe have been investigated in 6 monomict, 3 polymict, and the diamond-free ureilite ALH78019 and their acid-resistant, C-rich residues. Isotopic ratios of Kr and Xe are very uniform and agree with data for ureilites from the literature. The measured ratio ³⁸Ar/³⁶Ar showed large variations due to an experimental artifact. This is shown to be connected to the pressure dependence of the instrumental mass discrimination, which for ureilites with their low abundance of ⁴⁰Ar is different from that of the usual air standard. This observation necessitates a reassessment for the recently reported ³⁶Ar excesses due to possible decay of extinct ³⁶Cl in the Efremovka meteorite.Trapped ²²Ne in the range of (1.4-2.5) × 10⁻⁸ cc STP/g is present in bulk ureilites. A Ne three-isotope plot for polymict ureilites indicates the presence of solar Ne. ²¹Ne-based cosmic ray exposure ages for the 10 ureilites studied range from 0.1 Ma (for ALH78019) to 46.8 Ma (for EET83309)All ureilites may have started with nearly the same initial elemental ratio (¹³²Xe/³⁶Ar)₀, established in the nebula during gas trapping into their carbon carrier phases (diamond, amorphous C) by ion implantation. Whereas diamonds are highly retentive, amorphous C has suffered gas loss due to parent body metamorphism. The correlation of the elemental ratios ¹³²Xe/³⁶Ar and ⁸⁴Kr/³⁶Ar along the mass fractionation line could be understood as a two-component mixture of the unaffected diamond gases and the fractionated (to varying degrees) gases from amorphous C. In this view, the initial ratio (¹³²Xe/³⁶Ar)₀ is a measure of the plasma temperature in the nebula at the formation location of the carbon phases. Its lack of correlation with Δ¹⁷O (a signature of the silicate formation location) indicates that carbon phases and silicates formed independently in the nebula, and not from a carbon-rich magmaThe elemental ratios ¹³²Xe/³⁶Ar and ⁸⁴Kr/³⁶Ar in carbon-rich acid residues show a decreasing trend with depth (inferred from carbon consumption during combustion), which can be interpreted as a consequence of the ion implantation mechanism of gas trapping that leads to greater depth of implantation for lighter mass ionThe similarity between trapped gases in phase Q in primitive chondrites and the C phases in ureilites—for both elemental and isotopic compositions—strongly suggests that phase Q might also have received its noble gases by ion implantation from the nebula. The slight differences in the elemental ratios can be explained by a plasma temperature at the location of phase Q gas loading that was about 2000 K lower than for ureilite C phases. This inference is also consistent with the finding that the trapped ratio ¹²⁹Xe/¹³²Xe (1.042 ± 0.002) in phase Q is slightly higher, compared to that of ureilite C phases (1.035 ± 0.002), as a consequence of in situ decay of ¹²⁹I, and becomes observable due to higher value of I/Xe in phase Q as a result of ion implantation at about 2000 K lower plasma temperature.     

Fluid inclusion study of the Higher Himalayan quartzitic pelites, Garhwal Himalaya, India: Implications for recrystallization history of metasediments

Quartzitic pelites forms a part of Higher Himalayan Crystalline of higher geotectonic zone in Garhwal Himalaya. Quartzitic pelites (locally known as Pandukeshwar Quartzite) in Garhwal Himalaya is sandwiched between high grade metamorphic rocks of Central Crystallines and Badrinath Formation. Fluid inclusion studies are carried out on the detrital, and recrystallized quartz grains of quartzitic pelites to know about the fluid phases present during recrystallization processes at the time of maximum depth of burial. The quartzitic pelite (Pandukeshwar Quartzite) essentially consists of recrystallised quartz with accessory minerals like mica and feldspar. Fluid microthermometry study reveals the presence of three types of fluids: (i) high-salinity brine, (ii) CO₂-H₂O and (iii) H₂O-NaCl. These fluids were trapped during the development of grain and recrystallization processes. The high saline brine inclusions and CO₂-H₂O fluid with the density of 0.90 to 0.97 gm/cm³ are remnants of provenance area. CO₂ density in detrital quartz grains characterise the protolith of the sandstone as granite or metamorphic rock. The H₂O-NaCl fluids involved in the recrystallization processes at temperature-pressure of 430-350°C; 4.8 to 0.5 Kbars as constrained by fluid isochores of CO₂-H₂O and H₂O-NaCl inclusions and bulging and subgrain development during recrystallization processes. The re-equilibration of the primary fluid due to elevated internal and confining pressure is evident from features like ‘C’ shaped cavities, stretching of the inclusions, their migration and decrepitation clusters. The observed inclusion morphology revealed that the rocks were exhumed along an isothermal decompression path.     

Sensitivity Analysis of Coal Rib Stability for Internal Mine Dump in Opencast Mine by Finite Element Modelling

The present paper discusses the impact of the geometrical parameters of the coal rib and the mine dump on the stability of the coal rib. The geometrical parameters such as the slope angle, the height of dragline dump, the height of main dump, the gradient of seam and the thickness of coal rib have been considered as input to the numerical model for the stability analysis of the coal rib. Sensitivity analysis has been performed based on the results of the analysis in term of factor of safety of the coal rib. The input parameters have been classified in terms of significance (i.e. very high significance, high significance medium significance and low significance). The factor of safety is more influenced by highly significant parameters. The height and the slope angle of dragline dump and the thickness of the coal rib are highly significant parameters for the stability of the coal rib. The gradient of the seam is a medium significant parameter whereas, height of main dump and the number of dragline cut dump are low significant parameters for stability of coal rib.     

Geochemical behaviour of LREE,Y and Zr in uranium mineralized and non-mineralized granite from Darshanapur area,in the Gogi-Kurlagere fault zone,Bhima basin,Yadgiri district,Karnataka

Granite core samples (n=14) from the Gogi-Kurlagere fault zone in the central part of the Bhima basin were studied in terms of LREE, Y and Zr mobility during uranium mineralization. LREE, Zr and Y along with LILE (Ba, Rb) and P show behavioral differences in the mineralised and the non-mineralised samples. Average ΣLREE in mineralised granite (240 ppm) is higher than in non-mineralised samples (157 ppm). The average Zr and Y in the mineralised granite are 193 ppm and 17 ppm, while the corresponding abundances of these elements in non-mineralised portion are 148 ppm and 11 ppm respectively. Besides enrichment of U, Th, Ba, Pb and Rb and depletion of Sr are observed in mineralized granite in comparison to non-mineralized granite. Hydrothermal alteration has led to the mobility of these elements, which again dependent on the overall geochemical behavior of the migrating fluid. REE and Y in association with uranyl [(UO₂)²⁺] ion were transported as carbonate complexes like [UO₂(CO₃)₃]^4- and [REE (CO₃)₃]^3- and were later incorporated into favourable structural loci by precipitating minerals like pitchblende and coffinite.     

Upper mantle anisotropy beneath northeast India–Asia collision zone from shear-wave splitting analysis

Teleseismic earthquake data recorded by 11 broadband digital seismic stations deployed in the India–Asia collision zone in the eastern extremity of the Himalayan orogen (Tidding Suture) are analyzed to investigate the seismic anisotropy in the upper mantle. Shear-wave splitting parameters (Φ and δt) derived from the analysis of core-refracted SKS phases provide first hand information about seismic anisotropy and deformation in the upper mantle beneath the region. The analysis shows considerable strength of anisotropy (delay time ~0.85–1.9 s) with average ENE–WSW-oriented fast polarization direction (FPD) at most of the stations. The FPD observed at stations close to the Tidding Suture aligns parallel to the strike of local geological faults and orthogonal to absolute plate motion direction of the Indian plate. The average trend of FPD at each station indicates that the anisotropy is primarily originated by lithospheric deformation due to India–Asia collision. The splitting data analyzed at closely spaced stations suggest a shallow source of anisotropy originated in the crust and upper mantle. The observed delay times indicate that the primary source of anisotropy is located in the upper mantle. The shear-wave splitting analysis in the Eastern Himalayan syntaxis (EHS) and surrounding regions suggests complex strain partitioning in the mantle which is accountable for evolution of the EHS and complicated syntaxial tectonics.