深度神经网络拾取地震P和S波到时

从地震波形数据中快速准确地提取各个震相的到时是地震学中的基础问题.本文针对上述问题提出了利用深度神经网络拾取到时的新方法,建立了用于地震到时提取的17层Inception深度网络模型,在对原始三分量数据进行高通滤波和归一化处理后输入网络直接输出到时信息.整个过程基于神经网络自适应提取波形特征,自动输出结果.通过对100组加了不同强度的噪声数据进行了可靠性检验,相比于其他方法神经网络方法对于噪声具有较高的容忍度以及稳定性,并且与地震目录数据有较高的相似性.相比于AR-AIC+STA/LTA,深度神经网络虽然运算速度稍慢,但整个过程不需设定时窗与阈值,同时具有更高的可用性,并且可以迭代升级以提高精度.此方法作为人工智能方法,为波形到时拾取提供了新思路.     

Basement structural architecture and hydrocarbon conduit potential of polygonal faults in the Cooper‐Eromanga Basin,Australia

A thorough and complete understanding of the structural geology and evolution of the Cooper‐Eromanga Basin has been hampered by low‐resolution seismic data that becomes particularly difficult to interpret below the thick Permian coal measures. As a result, researchers are tentative to interpret the basement fault architecture within the basin, which is largely undefined. To provide a better understanding of the basement fault geometry, all available two‐dimensional seismic lines together with 12 three‐dimensional seismic surveys were structurally interpreted with assistance from seismic attribute analysis. The Upper Cretaceous Cadna‐owie Formation and top Permian reflectors were analysed using a common seismic attribute technique (incoherency) that was used to infer the presence of faults that may have otherwise been overlooked. Detailed basement fault maps for each seismic survey were constructed and used in conjunction with two‐dimensional seismic data interpretation to produce a regional basement fault map. Large north‐northeast–south‐southwest‐striking sinistral strike–slip faults were identified within the Patchawarra Trough appearing to splay from the main northeast–southwest‐striking ridge. These sinistral north‐northeast–south‐southwest‐striking faults, together with field‐scale southeast–northwest‐striking dextral strike–slip faults, are optimally oriented to have potentially developed as a conjugated fault set under a south‐southeast–north‐northwest‐oriented strike–slip stress regime. Geomechanical modelling for a regionally extensive system of Cretaceous polygonal faults was performed to calculate the Leakage Factor and Dilation Tendency of individual faults. Faults that extend into Lower Cretaceous oil‐rich reservoirs with strikes of between 060°N and 140°N and a high to near‐vertical dip angle were identified to most likely be acting as conduits for the tertiary migration of hydrocarbons from known Lower Cretaceous hydrocarbon reservoirs into shallow Cretaceous sediments. This research provides valuable information on the regional basement fault architecture and a more detailed exploration target for the Cooper‐Eromanga Basin, which were previously not available in literature.     

A model‐based data‐driven dictionary learning for seismic data representation

Planar waves events recorded in a seismic array can be represented as lines in the Fourier domain. However, in the real world, seismic events usually have curvature or amplitude variability, which means that their Fourier transforms are no longer strictly linear but rather occupy conic regions of the Fourier domain that are narrow at low frequencies but broaden at high frequencies where the effect of curvature becomes more pronounced. One can consider these regions as localised “signal cones”. In this work, we consider a space–time variable signal cone to model the seismic data. The variability of the signal cone is obtained through scaling, slanting, and translation of the kernel for cone‐limited (C‐limited) functions (functions whose Fourier transform lives within a cone) or C‐Gaussian function (a multivariate function whose Fourier transform decays exponentially with respect to slowness and frequency), which constitutes our dictionary. We find a discrete number of scaling, slanting, and translation parameters from a continuum by optimally matching the data. This is a non‐linear optimisation problem, which we address by a fixed‐point method that utilises a variable projection method with ?₁ constraints on the linear parameters and bound constraints on the non‐linear parameters. We observe that slow decay and oscillatory behaviour of the kernel for C‐limited functions constitute bottlenecks for the optimisation problem, which we partially overcome by the C‐Gaussian function. We demonstrate our method through an interpolation example. We present the interpolation result using the estimated parameters obtained from the proposed method and compare it with those obtained using sparsity‐promoting curvelet decomposition, matching pursuit Fourier interpolation, and sparsity‐promoting plane‐wave decomposition methods.     

Robust dual‐sensor data sea‐surface ghost attenuation for quality control purposes

Three‐dimensional receiver ghost attenuation (deghosting) of dual‐sensor towed‐streamer data is straightforward, in principle. In its simplest form, it requires applying a three‐dimensional frequency–wavenumber filter to the vertical component of the particle motion data to correct for the amplitude reduction on the vertical component of non‐normal incidence plane waves before combining with the pressure data. More elaborate techniques use three‐dimensional filters to both components before summation, for example, for ghost wavelet dephasing and mitigation of noise of different strengths on the individual components in optimum deghosting. The problem with all these techniques is, of course, that it is usually impossible to transform the data into the crossline wavenumber domain because of aliasing. Hence, usually, a two‐dimensional version of deghosting is applied to the data in the frequency–inline wavenumber domain. We investigate going down the “dimensionality ladder” one more step to a one‐dimensional weighted summation of the records of the collocated sensors to create an approximate deghosting procedure. We specifically consider amplitude‐balancing weights computed via a standard automatic gain control before summation, reminiscent of a diversity stack of the dual‐sensor recordings. This technique is independent of the actual streamer depth and insensitive to variations in the sea‐surface reflection coefficient. The automatic gain control weights serve two purposes: (i) to approximately correct for the geometric amplitude loss of the Z data and (ii) to mitigate noise strength variations on the two components. Here, Z denotes the vertical component of the velocity of particle motion scaled by the seismic impedance of the near‐sensor water volume. The weights are time‐varying and can also be made frequency‐band dependent, adapting better to frequency variations of the noise. The investigated process is a very robust, almost fully hands‐off, approximate three‐dimensional deghosting step for dual‐sensor data, requiring no spatial filtering and no explicit estimates of noise power. We argue that this technique performs well in terms of ghost attenuation (albeit, not exact ghost removal) and balancing the signal‐to‐noise ratio in the output data. For instances where full three‐dimensional receiver deghosting is the final product, the proposed technique is appropriate for efficient quality control of the data acquired and in aiding the parameterisation of the subsequent deghosting processing.     

Potential wind erosion rate response to climate and land‐use changes in the watershed of the Ningxia–Inner Mongolia reach of the Yellow River,China, 1986–2013

Climate and land‐use changes could strongly affect wind erosion and in turn cause a series of environmental problems. Thus, the objective of this study was to assess potential wind erosion rate (PWER) response to climate and land‐use changes in the watershed of the Ningxia–Inner Mongolia Reach of the Yellow River (NIMRYR), China. The watershed of NIMRYR suffers from serious wind erosion hazards, and over recent decades, wind erosion intensity and distribution has changed, following climate and land‐use changes. To understand these processes in the NIMRYR watershed, the Integrated Wind Erosion Modelling System (IWEMS) and the Revised Wind Erosion Equation (RWEQ) were used to calculate the PWER under different climate conditions and land‐use scenarios, and to assess the influences of climate and land‐use changes on the PWER. The results show the PWER in the whole watershed had a significant declining trend from 1986 to 2013. The results of the relationship among PWER, climate change, and land‐use changes showed that climate change was the dominant control on the PWER change in this watershed. Compared to the period 1986–1995, the average PWER decreased 23.32% and 64.98% as a result of climate change in the periods 1996–2005 and 2006–2013, respectively. In contrast with climate change, the effects of land‐use changes on the average PWER were much lower, and represented a change in PWER of less than 3.3% across the whole watershed. The study method we used could provide some valuable reference for wind erosion modelling, and the research results should help climate and land‐use researchers to develop strategies to reduce wind erosion. Copyright © 2017 John Wiley & Sons, Ltd.     

Variation in reach‐scale thalweg‐migration intensity in a braided reach of the lower Yellow River in 1986–2015

Thalweg migration of an alluvial river plays a key role in channel evolution, which may influence the effect of existing river training works and biodiversity on floodplains, and cause losses in riparian land and property. The braided reach of the Lower Yellow River underwent continuous channel aggradation during the period from 1986 to 1999, and then remarkable channel degradation in 1999–2015 owing to the state of operation of the Xiaolangdi Reservoir in 1999. Here we quantify associated thalweg migration changes and identify the key influencing factor in the braided reach. Thalweg‐migration distances and intensities at section‐ and reach‐scales were calculated during the past 30 years from 1986 to 2015, in order to investigate the characteristics of thalweg migration in the reach. There was a 47% reduction in the reach‐scale thalweg‐migration distance and a 35% reduction in the corresponding migration intensity after the reservoir operation. It is also revealed that fluvial erosion intensity is a dominant factor in controlling the thalweg migration, based on the investigation into various influencing factors in the study reach. The thalweg‐migration intensity of the braided reach can be expressed as a power function of the previous four‐year average fluvial erosion intensity. The calculated thalweg‐migration intensities in 1986–2015 using the proposed relation generally agree with the observed data. Copyright © 2017 John Wiley & Sons, Ltd.     

Effects of changing climate on ice cover in three morphometrically different lakes

A one‐dimensional hydrodynamic lake model (DYRESM‐WQ‐I) is employed to simulate ice cover and water temperatures over the period 1911–2014. The effects of climate changes (air temperature and wind speed) on ice cover (ice‐on, ice‐off, ice cover duration, and maximum ice thickness) are modeled and compared for the three different morphometry lakes: Fish Lake, Lake Wingra, and Lake Mendota, located in Madison, Wisconsin, USA. It is found that the ice cover period has decreased due to later ice‐on dates and earlier ice‐off dates, and the annual maximum ice cover thickness has decreased for the three lakes during the last century. Based upon simulated perturbations of daily mean air temperatures across the range of ?10°C to +10°C of historical values, Fish Lake has the most occurrences of no ice cover and Lake Wingra still remains ice covered under extreme conditions (+10°C). Overall, shallower lakes with larger surface areas appear more resilient to ice cover changes caused by climate changes.     

Enhanced identification of a hydrologic model using streamflow and satellite water storage data: A multicriteria sensitivity analysis and optimization approach

Hydrologic model development and calibration have continued in most cases to focus only on accurately reproducing streamflows. However, complex models, for example, the so‐called physically based models, possess large degrees of freedom that, if not constrained properly, may lead to poor model performance when used for prediction. We argue that constraining a model to represent streamflow, which is an integrated resultant of many factors across the watershed, is necessary but by no means sufficient to develop a high‐fidelity model. To address this problem, we develop a framework to utilize the Gravity Recovery and Climate Experiment's (GRACE) total water storage anomaly data as a supplement to streamflows for model calibration, in a multiobjective setting. The VARS method (Variogram Analysis of Response Surfaces) for global sensitivity analysis is used to understand the model behaviour with respect to streamflow and GRACE data, and the BORG multiobjective optimization method is applied for model calibration. Two subbasins of the Saskatchewan River Basin in Western Canada are used as a case study. Results show that the developed framework is superior to the conventional approach of calibration only to streamflows, even when multiple streamflow‐based error functions are simultaneously minimized. It is shown that a range of (possibly false) system trajectories in state variable space can lead to similar (acceptable) model responses. This observation has significant implications for land‐surface and hydrologic model development and, if not addressed properly, may undermine the credibility of the model in prediction. The framework effectively constrains the model behaviour (by constraining posterior parameter space) and results in more credible representation of hydrology across the watershed.     

Investigation of the thermal regime and subsurface properties of a tidally affected,variably saturated streambed

Temperature and moisture content in the variably saturated subsurface are two of the most important physical parameters that govern a wide variety of geochemical and ecological processes. An understanding of thermal and hydraulic processes and properties of transient vadose zones is therefore fundamental in the evaluation of such processes. Here, an investigation of the thermal regime and subsurface properties of a tidally affected, variably saturated streambed is presented. Field and laboratory measurements, as well as a forward numerical model, are jointly employed in the investigation. Temperature, soil moisture, surface level, and water level data were recorded in a transect perpendicular to a tidally driven stream. Frequency‐domain analysis of the subsurface temperature measurements revealed the rapid decay of the tidal temperature driver within the top ～30 cm of sediment. Several techniques were used to evaluate subsurface thermal and hydraulic properties, including thermal conductivity and the soil water retention curve. These properties were used to constrain a forward numerical model that included coupled treatment of relevant variable saturation thermal and hydraulic physics. Even though the investigated vadose zone is intermittent and relatively shallow ( 20 cm), the results illustrate how error can be introduced into heat‐transport calculations if unsaturated conditions are not taken into account.