基于地质大数据理念的模型驱动矿产资源定量预测

在大数据科学成为新的科学范式的背景下,基于地质大数据理念,提出了模型驱动的矿产资源定量预测评价的新方法,以及模型流程建模技术贯穿整个矿产资源预测评价过程的新思路,以地质理论指导地质大数据分析和计算机技术实现地质大数据挖掘2条主线展开研究,实现了面向地质大数据的数据挖掘与矿产资源的定量预测评价。结合青海祁漫塔格铁铜多金属矿床、山东焦家金矿床、云南个旧锡铜多金属矿床等不同地区、不同成矿类型和矿种开展了应用研究,完成了找矿模型工作流的设计与实现,进行了有利成矿信息的挖掘,取得了较好的效果,为大数据时代数字地质研究提供了新的思路。     

Haines Structure: a probable Eocene complex impact structure in the offshore Canning Basin,Western Australia

A seismic structure imaged in a single 2D seismic profile from the offshore Canning Basin, Western Australia, is interpreted to be a possible complex impact crater on the basis of its seismic character. The feature, herein referred to as the Haines Structure, is symmetrical in two dimensions over ～2.5 km and comprises a centrally uplifted basal surface, a depressed upper surface, a highly deformed intervening package and an overlying horizon that is ‘sagging’ over the depression. The possible impact structure lies within carbonate units of Eocene or Early Oligocene age as determined from seismic correlation to petroleum exploration wells. The structure has not been drilled; therefore distinguishing characteristics used to define an impact origin, such as shock metamorphism, are not available. Comparison with other features in the Neogene succession of Australia's North West Shelf that have previously been interpreted as impact structures highlights the presence of key elements in the Haines Structure that characterise known complex impact craters, and the absence of seismic features related to alternative processes, such as a link to deeper structures that would be expected beneath a volcanic pipe.     

油气勘探开发数据体平均值代表性分析

油气勘探开发涉及各类数据体,同一数据体可得多种平均值,目前尚无明确而有效的方法以判断何种平均值能客观反映数据体的典型水平。运用数据统计平均分析方法,对勘探开发实践中的孔隙度、渗透率、产量、成本等数据体进行系统分析,确立了加权中位数计算公式和平衡中位数法则。加权中位数计算公式适于分析不同领域正常有序数据体的基本特征,加权中位数为正常有序数据体的平衡点;平衡中位数法则适于确定有足够大数据容量、能满足统计分析基本要求、能选择合理权衡指标的正常有序数据体的典型水平;有明确物理意义的加权平均值也可确定数据体的典型水平。     

基于熵权-正态云模型的岩爆烈度分级预测研究

岩爆是深埋地下工程主要的工程地质灾害之一,其烈度分级预测是必须解决的岩石工程重大问题。针对岩爆烈度分级的不确定性,选取岩石单轴抗压强度与抗拉强度比σ_c/σ_t、切向应力与岩石单轴抗压强度比σ_θ/σ_c、弹性变形能指数Wet和岩石完整性系数K_v等定量化指标建立评价指标体系,应用熵权法确定各评价指标权重,结合云理论建立了岩爆烈度分级预测的熵权-正态云模型。以国内外12组典型岩爆工程实例对所建模型进行检验,并与逼近理想解法、高斯径向基函数法的判别结果及实际情况进行对比。为进一步考察该模型的有效性和实用性,利用该模型对终南山公路隧道工程竖井的岩爆实例进行分析。研究结果显示,基于熵权-正态云模型的岩爆判别预测结果与实际岩爆情况较吻合,表明熵权-正态云模型在研究岩爆烈度分级预测中具有良好的实用性和可靠性,可为深部地下工程岩爆烈度分级预测提供一种新思路。     

UAVRS影像空中三角测量实验研究

无人飞行器低空遥感系统(UAV Low Altitude Remote Sensing System简称UAVRS系统)适合于在阴云山区低空获取高分辨率垂直和倾斜影像,但获取的影像存在像幅小、数量多、基线短、重叠度不规则且倾角过大等问题,而这些问题,都是国内外现有的数字摄影测量系统面临的新问题。本文提出以带权观测值光束法平差解决UAVRS复杂影像区域网的空中三角测量问题的方法。其技术要点在于,从提高单模型定向的精度出发,以单模型定向元素为基础构建自由区域网,为带权观测值光束法平差提供有效初值。经试验表明,空三加密点的平面精度为0.386m,高程精度为0.525m,基本能达到1∶2000地形图测图的要求。     

超宽频海底剖面仪实时采集显示系统EchoViewer的设计与实现

超宽频海底剖面仪实时采集显示系统EchoViewer是为新研制的超宽频海底剖面仪开发的配套软件。该软件从超宽频海底剖面仪的超宽频率和多频段分频合成的特点出发,同时兼容同类系统的需求进行设计,根据声纳在海底介质中的传播特点实现TVG算法和海底跟踪算法,结合遥感图像处理方法实现分频合成算法,并采用软件工程的管理方法实现整个系统的开发,具有系统控制、导航定位、数据采集并成图显示、数据回放和数据后处理等功能。该软件在超宽频海底剖面仪多次海试和国家海洋局908浅剖专项调查任务中得到成功应用。     

Model studies on the effect of nitric acid vapour on cirrus cloud activation

The effect of elevated nitric acid concentrations on cloud particle number concentrations and sizes of mixed phase cirrus cloud (containing both supercooled liquid and frozen ice particles) is studied using a detailed multicomponent condensation model. Our model calculations suggest that high nitric acid volume mixing ratios (VMRs), corresponding to upper tropospheric measurements, can increase the ice crystal number concentrations substantially especially in the case of freezing of non-activated solution droplets (as opposed to freezing of supercooled cloud droplets). This is due to increased droplet sizes caused by nitric acid absorption and the resulting uptake of water.     

Nine thousand years of upper montane soil/vegetation dynamics from the summit of Caratuva Peak,Southern Brazil

Biodiversity loss, climate change, and increased freshwater consumption are some of the main environmental problems on Earth. Mountain ecosystems can reduce these threats by providing several positive influences, such as the maintenance of biodiversity, water regulation, and carbon storage, amongst others. The knowledge of the history of these environments and their response to climate change is very important for management, conservation, and environmental monitoring programs. The genesis of the soil organic matter of the current upper montane vegetation remains unclear and seems to be quite variable depending on location. Some upper montane sites in the very extensive coastal Sea Mountain Range present considerable organic matter from the late Pleistocene and other from only the Holocene. Our study was carried out on three soil profiles (two cores in grassland and one in forest) on the Caratuva Peak of the Serra do Ibitiraquire (a sub-range of Sea Mountain Range – Serra do Mar) in Southern Brazil. The δ¹³C isotopic analyses of organic matter in soil horizons were conducted to detect whether C₃ or C₄ plants dominated the past communities. Complementarily, we performed a pollen analysis and ¹⁴C dating of the humin fraction to obtain the age of the studied horizons. Except for a short and probably drier period (between 6000 and 4500 cal yr BP), C₃ plants, including ombrophilous grasses and trees, have dominated the highlands of the Caratuva Peak (Pico Caratuva), as well as the other uppermost summits of the Serra do Ibitiraquire, since around 9000 cal yr BP. The Caratuva region represents a landscape of high altitude grasslands (campos de altitude altomontanos or campos altomontanos) and upper montane rain/cloud forests with soils that most likely contain some organic matter from the late Pleistocene, as has been reported in Southern and Southeastern Brazil for other sites. However, our results indicate that the studied deposits (near the summit) are from the early to late Holocene, when somewhat wetter and warmer conditions (since around 9000 cal yr BP) enabled a stronger colonization of the ridge of Pico Caratuva by mainly C₃ plants, especially grassland species. However, at the same time, even near the summit, the soil core from the forest site already presented the current physiognomy (or a shrubby/elfin or successional forest), indicating that the colonization of the neighboring uppermost saddles and valleys were probably populated mainly by upper montane forest species.     

数字地质调查中多源地学数据在造山带构造单元 划分和大地构造演化研究方面的应用 ——以青海民和地区为例

多源地学数据包括遥感、地球化学和地球物理数据,介绍了利用多源地学数据进行造山带构造单元划分的方法。遥感数据在确定区域构造边界及活动断层方面的应用非常广泛,遥感影像在解译线形构造即断层方面有非常明显的效果,可以根据不同构造单元的影像差异,区分不同的地质体、线性构造及活动断层,同时用遥感数据叠加三维地形数据分析线性构造可以更加直观地解译线性构造。地球化学数据在确定大的构造边界方面具有一定的指示意义,可以根据水系沉积物的地球化学特点,运用因子分析方法确定大的构造边界。地球物理数据提供的是地质体及构造边界在深部的延伸情况,可以为研究断裂的运动学和动力学特征提供证据。     

正规化矩阵正定时半参数估计量的统计性质

对于包含系统误差和粗差的观测数据,本文将混合Cook距离引入到半参数模型中,实现了粗差的定值定位。首先,构造补偿最小二乘函数,利用泰勒展式,根据均值漂移模型与删失模型的等价性,导出了观测数据中剔除第i个观测数据前后参数分量和非参数分量相应估计值之间的关系式,为粗差的定位分析奠定了基础。其次,将混合Cook距离作为诊断统计量,进行粗差定位分析,得到了参数分量和非参数分量的诊断统计量的简洁计算公式,为了提高粗差定位的准确性,给出了混合Cook距离参数 Q 和C的一些常用形式,通过合理选择相应参数,计算参数分量和非参数分量的距离函数,实现粗差的定位,并将系统误差和粗差从观测数据中区别开来。最后通过模拟算例和实测数据验证了本文方法的正确性。