High-density seismic acquisition technique shows good effects on solving exploration problems such as exploration in low signal-to-noise ratio region, imaging of complex geological bodies, lithological exploration and fine reservoir description and monitoring. The technique has become one of the mainstream techniques in the industry. However, the biggest restriction factor in the technical implementation is cost increase which is brought in the high-density acquisition. Frequency separated simultaneous sweep (FSSS) is a high-efficiency vibroseis acquisition method of separating the conventional sweep signal into independent sub-sweep signals with different frequency bandwidth, and arranging the orthogonal sub-sweep signals for simultaneous shooting by different fleets through encoding. It is easy for the separation of the signal and the noise by using correlation. But the challenges to using such a method are the limitation of the number of sub-signals in DSD and high requirements of field operation management system and communication network for real-time control. This paper mainly introduces the implementation of FSSS through the cascade encoding technology to rearrange the sub-signatures as a new signature and realize simultaneous sweep or independent sweep technique and shows the application effects. Presentation Date: Wednesday, October 14, 2020 Session Start Time: 1:50 PM Presentation Time: 2:40 PM Location: Poster Station 9 Presentation Type: Poster

ABSTRACT A novel approach for the rapid interpretation of magnetic tensor data is introduced. Magnetic tensor fields are significantly more sensitive to local anomalies than the traditional total magnetic intensity data (TMI). Threedimensional (3D) iterative magnetic gradient migration is a stable, solid, and alternative technique to solve the disadvantages of inverse problem. The approach can obtain a stable migration image by direct integral transforming an observed tensor field contaminated with a high-noise dataset into a subsurface susceptibility. Moreover, a focusingstabilizing functional is considered into the migration framework to produce more accurate images of distributionof sharp susceptibility anomalies with the correct depth in subsurface formations. An experimental study and a realistic geology model are carried out to investigate the effectiveness of this new method to ascertain how the noise level would influence the stability of migration results. Results show a stable image with high resolution. Presentation Date: Monday, September 16, 2019 Session Start Time: 1:50 PM Presentation Time: 2:40 PM Location: Poster Station 1 Presentation Type: Poster

ABSTRACT Shallow reflection data processing is a challenging task mainly because of the high level of coherent/non-coherent noise in the data. Non-reflective coherent events are not easy to attenuate and might be present on final image, leading to misinterpretation. Moreover, quite often the noise and reflection energy share the same signal properties that makes it difficult to separate them using frequency-based filters. In general, nearsurface seismic reflections are stronger in a specific portion of the field record that is conventionally called the optimum window. To enhance the reflections, one solution is to mute the record outside the optimum window where the effect of source-generated noise as well as direct and surface waves are more pronounce. Noise muting might increase signal-to-noise ratio and improves processing; however, it is problematic when a very large portion of data is contaminated with strong noise and the optimum window becomes very narrow. This leads to difficult data processing and smeared images. In this work, we propose to enhance data quality not only within the optimum window, but in the whole field record by interpolation and extrapolation of reflections. To accomplish this, we use a processing workflow based on the stacking parameters of the common-reflection-surface method, which is a powerful seismic processing technique for data enhancement. Data applications on a synthetic data as well as a shallow field data set show promising results. Presentation Date: Wednesday, September 18, 2019 Session Start Time: 1:50 PM Presentation Start Time: 2:40 PM Location: Poster Station 7 Presentation Type: Poster

ABSTRACT Radio atmospheric signals or sferics are broadband electromagnetic impulses generated by lightning discharges. The recording of raw or streamed data, as done by CGG during MEGATEM surveys, allows for the extraction of sferic signals, inadvertently recorded during AEM surveys. The spectral processing of individual sferics, excluding periods of background noise between sferic events, allows for the extraction of AFMAG and VLF data in the frequency range 25 Hz–25 kHz, including good signal in the AMT dead band (1–5 kHz). The recording of the three-component AEM data allows for the vector processing of sferic responses, including the derivation and modeling of the tipper data. Conductivity information can be derived from the tipper data with an apparent conductivity transformation and, more rigorously, with 2D and 3D inversions that take into account the terrain’s topography. The extraction and modeling of sferic responses is demonstrated on a MEGATEM data set acquired in South America. The passive EM data were successfully modeled with 2D and 3D inversions, and the derived conductivity structures confirm and complement the information extracted from the active-source EM data. Presentation Date: Thursday, October 18, 2018 Start Time: 8:30:00 AM Location: 213B (Anaheim Convention Center) Presentation Type: Oral

ABSTRACT Poynting vector based reverse time migration () can migrate seismic data into angle-domain common-image gathers (). The quality of common-image-gather is related to the used migration velocity. However, in the seismic data processing, we usually cannot obtain a very accurate depth migration velocity. The difference between the true depth velocity and estimated migration velocity would introduce image artifacts, which are often observed as non-flat events in the . These non-flat events will degrade the final image that is linearly stacked from all angles. To mitigate such image artifacts caused by the inaccurate migration velocity, we can use both quantitative and qualitative methods to optimize the stacking. The quantitative method uses a numerical method to achieve an optimal velocity model for migration. However, there is residual moveout error still left in the gathers. Instead, we propose a qualitative method to assure geologically meaningful result, so-called segmentation method using Moore neighborhood algorithm, which can decompose events in the into isolated signal groups. From this sequestered image domain, signals can manifest easier because of dealing only local signals that do not interfere with adjacent ones. We can automatically align concave events and focus stretching amplitudes at far angles. Additionally, we do not need cross-correlation to measure the moveout of non-flat events. Consequently, we can avoid the mismatching among the events at different depths. We test our proposed method by one numerical dataset. The numerical results show that the can correct the non-flat events in the from small angle to large angle. Compared with the tradition with Laplace filtering and small angle stacking, our new method can produce a superior migration image with fewer artifacts. In practice, this proposed method can reduce the requirement of migration velocity accuracy in the depth velocity building and provide a first-look of image. Presentation Date: Tuesday, October 16, 2018 Start Time: 1:50:00 PM Location: Poster Station 20 Presentation Type: Poster

ABSTRACT Seismic recorded gathers are imprinted with multiples energy. Multiples are attenuated using different established methods in the industry. A careful estimate of multiples and their subsequent subtraction is common practice in preserving primaries and aid in interpretation of seismic data. Most established methods use two characteristics of multiples, delayed arrival time with respect to primaries, and periodicity. An attempt is made to use coherency as the foundation for attenuating multiples. Karhunen Loeve Transform is used on dataset, using multiple velocity. Principal components attributing to multiples are selected and composed back to seismic traces. These multiples are then subtracted from the original gather. The results are compared with conventional radon method. The method offers advantage over radon, because no transformation has taken place, which often results in inaccurate curvature values for gathers in the dataset, and leaves artifacts on real gathers. The method can be applied in windows, with varying principal component filtering criteria, thus providing opportunity to pick the proportion of principal components associate with multiples with different coherency. This approach can further optimize practice of separating primaries from multiples at different offsets, as varying percentage of principal components for each window can be selected. Since original multiples have been selected from the dataset for subsequent subtraction, without any additional process, the method does not entail designing filters or subsequent adaptive subtraction. It has successfully attenuated both surface and interbed multiples. Presentation Date: Monday, October 15, 2018 Start Time: 1:50:00 PM Location: 211A (Anaheim Convention Center) Presentation Type: Oral

ABSTRACT A new-design of air gun has a reduced output at high frequencies, a lower peak pressure and a smoother pulse onset. We have made seismic images of the same geology both using this new air gun and also using standard air guns. Furthermore we recorded the acoustic signal at many ranges using a streamer that had been specially configured to record a high bandwidth and also using near-field hydrophones. We compare the resulting seismic images from the two types of source and we also compare the relevant environmental impact metrics. The seismic images produced by the two types of air gun are essentially the same. At all the measured ranges the new-design of air gun produces a significantly lower sound exposure level (SEL) per shot than does the standard design. The risk of hearing damage zone around a seismic line has a significantly reduced width and a shallower penetration depth when the new-design of air gun is used. Presentation Date: Tuesday, September 26, 2017 Start Time: 3:05 PM Location: 371F Presentation Type: ORAL

ABSTRACT The existences of detrital carbonate distribution in C block-East Java provides new exploration opportunity in East Java Basin. The J-2 well had drilled and penetrated ~400ft gross reservoir thickness of detrital carbonate. Core plug analysis result shown Oligocene mixed with shallow marine bioclast which growing in isolated carbonate buildup. Therefore, it ensure the detrital carbonate in Cepu formed of a reworked Oligo-Miocene shallow marine carbonate Build up which is the main reservoir in C block. Log & petrophysical data determine the detrital carbonate is gas reservoir which has passably properties (~11% of porosity and ~40% of water saturation). The reservoir give strong positive amplitude anomaly in seismic data and reflected onlap geometry to the body of carbonate buildup. The connection between detrital and buildup carbonate will lead to potential hydrocarbon migration toward to the carbonate body instead trapped in detrital carbonate. The conventional seismic data has limitation to determine continuity in detrital carbonate layer and estimate similar lithology. The quantification and relationship between amplitude and frequencies of seismic data used to estimate the continuity and distribution of detrital carbonate facies. The frequencies of detrital had distinguished using spectral decomposition which transformed to continue wavelet. In addition, the impedance was predicted using model based inversion which build from J-2 well. Cross-plot result shown the detrital carbonate in C block has unique frequencies and impedance value. Horizon slice map identified there are no connection between carbonate build up body with detrital carbonate. In addition, CWT and Inversion identified facies changing in detrital sedimentation from carbonate to shale. As conclusion, detrital carbonate have good chance to be good gas reservoir and new play in C block. Presentation Date: Monday, October 17, 2016 Start Time: 1:50:00 PM Location: Lobby D/C Presentation Type: POSTER

ABSTRACT Data acquired by airborne in-loop time-domain electromagnetic (EM) systems, such as VTEM (Witherly et al., 2004), reflect mainly three physical phenomena in the earth: (1) EM induction, related to ground conductivity, (2) Airborne Inductively Induced Polarization (AIIP), related to the relaxation of polarized charges in the ground (Kratzer & Macnae 2012, Kwan et al., 2015) and (3) Superparamagnetism (SPM), related to the induced secondary magnetic field of certain ultrafine to fine grained magnetic materials (Kratzer, Macnae & Mutton 2013, Sattel & Mutton 2015). It is also called magnetic relaxation or magnetic viscosity. Olhoeft and Strangway (1974) predicted that the SPM effect could become relatively more important and permit mapping of magnetite content with an active, more sophisticated EM system. However, as a background effect it could also represent an additional source of signal to further complicate EM exploration. In this abstract, the general characteristics of SPM anomalies occurring over barren rocks in Greenland, in an environment different from Australia and Southern Africa, are discussed and a method of identifying them in VTEM data will be presented. Presentation Date: Tuesday, October 18, 2016 Start Time: 3:45:00 PM Location: 168 Presentation Type: ORAL

ABSTRACT A conventional workflow for geomechanical analyses of shale gas formations estimates three (, , and ) of the five transversely-isotropic (TI) elastic constants from the borehole sonic data acquired in a vertical pilot wellbore parallel to the X3-axis. The other two TI-elastic constants and are obtained from either a cluster analysis of sonic data acquired from wellbores with different deviations or by invoking a rock physics model calibrated against core data. These constants are then used in static poroelastic stress equations to output the minimum horizontal stress gradients and anisotropic mechanical properties that help to identify intervals with higher fracturability. As such, the poroelastic stress equations require transforms of dynamic to static moduli as well as calibrating the two tectonic strain terms so that the predictions are consistent with the measured closure stresses at select reference depths. This paper describes a new workflow whereby four of the five TI-constants are obtained from the sonic data in a vertical pilot wellbore. In particular, estimates of the TI-constants and from the inversion of dipole flexural dispersions provide ratios of the effective minimum horizontal to overburden stresses that are consistent with the estimates from the aforementioned conventional workflow. Advantages of the new workflow are that the four TI-constants are obtained from the same volume of formation probed by a sonic tool and the workflow is not dependent on having either core data analysis or cluster analysis of sonic data from sections of the well trajectory with different relative dips. Presentation Date: Wednesday, October 19, 2016 Start Time: 8:00:00 AM Location: 168 Presentation Type: ORAL

Summary Airborne inductively induced polarization (AIIP™) effects, in the form of numerous negative transients, have been observed in the VTEM helicopter time-domain EM (TDEM) data from Mirny, Russia. The survey area is located in the continuous permafrost zone. Inductively induced polarization (IIP) effects in ground TDEM data had earlier been reported from the same region. VTEM data reflect mainly two physical phenomena in the earth: i) electromagnetic induction and ii) airborne inductively induced polarization (AIIP) related to the relaxation of polarized charges in the ground. Typical AIIP effects in the VTEM data from Mirny corrupt only the early to mid-time channels, from 55 µs to 400 µs. Moreover, the earliest few time channels, from 18 µs to 48 µs, are not affected by AIIP, because the IP effect takes a finite time to build up. Cole-Cole analysis of the AIIP affected VTEM data showed that the frequency factor and the time constant are close to 0.8 and 0.0001s, respectively. Since the earliest time channel data are not affected by AIIP, the resistivity of the shallow ground can still be precisely determined. As a result, the purely inductive VTEM decay can be accurately modelled and used to remove the AIIP effect from later time-channels. Introduction During September 2014, Geotech Ltd carried out a helicopter-borne Versatile Time-domain Electromagnetic (VTEMTM; Witherly et al., 2004) survey over an area near Mirny, Russia (Figure 1), on behalf of the ALROSA mining company. The survey area is located in the continuous permafrost region of Russia. Kozhevnikov and Antonov (2012) reported fast decaying inductively induced polarization (IIP) effects in ground TDEM data from the same region. AIIP effects in other airborne TDEM system from permafrost region in Canada were reported by Smith and Klein (1996). VTEM data are mainly the time-domain electromagnetic responses of at least two physical processes in the earth: i) electromagnetic induction and ii) airborne inductively induced polarization (AIIP) related to the relaxation of polarized charges in the ground (Kratzer and Macnae, 2012 and Weidelt, 1982). Negative transients, attributed to AIIP effects, have been observed in VTEM ???? / ???? and B-field Zcomponent data (Kwan et al., 2015).

ABSTRACT Throughout the world there are vast areas of the earth covered by basalt flow that are under-explored for hydrocarbons due to poor sub-basalt imaging. The very high reflectivity at the top of the basalt and high energy attenuation within the basalt cause the sub-basalt reflections to be very weak and completely masked by different orders of multiple generated above the basalt. In order to attempt to image below the basalt, seismic data with long maximum offset have been acquired. Sub-basalt reflections do become visible on long-offset data in the cone formed between the basalt top (Trap Top) refraction and the refraction from shallow layers as shown in Figure 1. Processing of long-offset data does have challenges associated with it such as mode conversion and anisotropy (both intrinsic as well as long-offset effect). In the present study a model-based noise attenuation technique has been applied to suppress the different kind of noises and VTI anisotropy has been modeled and accounted for in Kirchhoff Pre-Stack Depth Migration imaging. The results are very encouraging.

Summary: This study focuses on the effect of the leachate to the surface and groundwater pollution at Taiping landfill site.mThe site was previously used for tin ore and other alluvial ore minerals mining. Consequently, the soil stratification of the area was extensively disturbed by tin mining activities in the past. Geologically, the area consists of mainly recent alluvium of Quaternary in aged. Electrical Resistivity Imaging (ERI) technique is used in mapping the pollution distribution in the form of leachate plumes. Results show that groundwater at depth of 5 to 15m was contaminated by the leachate from the dumping site as indicated by resistivity values less than 10 Ohm-m in the pseudosections. In addition, leachate plumes appeared to have seeped to about 30-50m in depth. Resistivity of the backfill surface covering material is found greater than 100 Ohm-m. The study has led to the delineation of the groundwater contamination zones within the study area as well as the distribution of the pollutant in the soil strata at depth less than 30m. This implies that the ERI technique is an effective tool in detecting contaminated groundwater zones or layers. Introduction: The existences of dumping sites that are mostly not properly engineered and managed have raised a major public concern related to groundwater contamination especially in developing countries. Pollutants released or discharged from the dumping sites could contaminate groundwater system, flora and fauna which in turn can impacted on people’s lives. The infiltration of rainfall into landfill, together with the biochemical and chemical breakdown of the wastes produces a leachate which is high in suspended solids and of varying organic and inorganic content. If the leachate enters surface or groundwater before sufficient dilution occurs, serious pollution incidents can transpire. This in turn causes groundwater contamination which is an extremely costly operation to remove. Although many landfills today are constructed in permeable layer, leaks can still occur. Colucci and Lavagnalo’s (1999) study for example indicate that on average, 15 leaks per hectare on impermeable layers transpired. Today, although the open dump approach is considered the primitive stage of landfill development, it remains the dominant waste disposal option in most of the countries. The main reason is that it is cheap and easy to operate. However, the disadvantages of this approach include the indiscriminate disposal of waste, limited measures to control operations, and the negative environmental effects to its surroundings. Malaysia is currently experiencing acute problems with solid waste management (MSW) although we have more than 230 landfills (many of them are open landfills). Malaysians generate about 6 million tons of solid waste per year, or about 15,000 tons daily. To accommodate the amount of waste, an average of 1.8 dumping sites per municipal council is needed (Ministry of Housing and Local Government, 2002). Malaysia’s high annual rainfall depth (an average of 3000mm) (DID, 2000), has somehow influences the production of leachate at these landfills. With high rainfall depth, leacheate production is also expected to be very high due to the infiltration of rain water through solid refuse.

ABSTRACT In this abstract we will demonstrate that the application of non-hyperbolic reflection tomography can produce a 3D subsurface velocity model that is smooth for prestack depth migration and at the same time geological for accurate subsurface pore pressure prediction and lithology interpretation. Our non-hyperbolic reflection tomography workflow is very different from the conventional reflection tomography workflow: no automatic volume picking of residual moveout is used. Instead a set of prestack events are interpreted on prestack seismic image volumes. The prestack events carry detailed moveout information that is more accurate than a single parameter fit to the common image gathers. In addition the prestack events carry structure dip information that is needed for accurate 3D ray tracing offset by offset. Examples will be given for both compaction driven and lithology controlled geological environments.

Introduction Summary It is know that surface layers may act as waveguides to the propagation of EM energy. Here we analyze the case where a concrete pavement can act as a guide of EM waves over a higher permittivity soil. This situation cannot be described as a true waveguide as the steady state is never reached. We estimate a phase velocity dispersion curve from the F-K panel of a common-shot profile. That curve is then used in an inversion procedure to obtain waveguide refractive index and thickness through a minimization strategy that seeks models where the numerical dispersion curve matches the picked dispersion curve. Dispersive wave propagation may occur in the uppermost layer if its thickness is comparable to or smaller than the wavelength, and its lower boundary is a strong reflector. In this case EM waves will be trapped in the waveguide, traveling in a modal fashion (Annan et al., 1975). As in all cases the waveguide is overlain by air either the substratum has a lower or a significantly higher permittivity than the waveguide’s (Arcone et al., 2003; van der Kruk et al., 2006). A low-velocity layer of wet soil over a higher velocity layer of sand and gravel is a studied example of the first case (Arcone et al., 2003; van der Kruk et al., 2006). An ice layer over water is an example of the second case (e.g., van der Kruk et al., 2007). This paper deals with the case where the underlying layer has a higher permittivity, enough to provide strong enough reflection. In this case some energy leaks downwards and is quickly absorbed by a high-loss substratum. In such a leaky waveguide total internal reflection occurs only at the upper interface beyond the critical angle. We check our findings against experimental data obtained over a concrete pavement overlying an oxisoil. Numerical Results Consider the case of the transmission of EM waves weakly bound in a guiding planar slab situated on the top of a semi-infinite half-space, interfacing with air at its top boundary. The x-axis is perpendicular to the interfaces with origin at half-width of the 2r thick slab, otherwise infinite. The zaxis points towards the direction of moveout, or observation point. The EM waves are launched into the slab by a dipolar transmitting antenna coupled to the slab’s free interface at z = 0. The transmitting antenna axis can be oriented either along the y-axis or along the z-axis, generating the TE or the TM modes, respectively. The basic geometry is depicted in Figure 1. The EM waves diffract into the slab through the antenna coupling with an angular distribution that depends on the refractive indexes n1 and n2 (Annan et alli, 1975; Arcone, 1995), where where ni, ei , i = 1, 2, 3, and as the slab interfaces with air on its top surface then, n1 = n0. For the slab to be considered as a true waveguide we need to have a hierarchy of refractive indexes as n1<n2 > n3.

ABSTRACT The improved precision of gravity meters, as well as the repeatability of seafloor time-lapse gravity surveying, has made feasible the use of 4D gravity as a new tool for dynamic characterization of hydrocarbon reservoir. The present work proposes a forward model for predicting gravity changes in a compacting reservoir. We study the effects of a compacting reservoir. We study the effects of a compacting reservoir when embedded in a homogeneous half space and when a rigid basement is located beneath it. We find that a 100 m thick reservoir at 800 m depth gives a gravity change of 30-40 µGal. Considering that the present accuracy of 4D gravity measurements is around 5 µGal (Stenvold et al., 2008), these are values which can be detectable.

SUMMARY Knowledge of soil water content distribution and dynamics at the field scale is essential for sustainable and optimal agricultural and environmental management of land and water resources. To constrain full-waveform inversion of time-lapse, proximal ground penetrating radar signals we used a hydrodynamic model to simultaneously identify shallow soil hydraulic properties and continuous vertical electric profiles. The radar model involves a full-waveform solution of Maxwell’s equations for wave propagation in three-dimensional multilayered media. Hydrodynamic modeling was based on a one-dimensional solution of Richard’s equation and was solved numerically using HYDRUS-1D code. The uniqueness of the inverse solution was analyzed using numerical experiments for different textured soils to estimate three soil hydraulic parameters (a, n, and Ks) of Mualem-van Genuchten (MVG) model. The top boundary conditions was time-variable and we assumed initial condition to be a priori unknown. A laboratory experiment was performed to test the proposed method for a transient infiltration event in a homogeneous sandy soil. Radar data were acquired in the frequency domain using a standard vector network analyzer combined with an off-ground monostatic horn antenna. The results suggest that the proposed method is promising for characterizing the shallow subsurface hydraulic properties at the field scale with a high spatial resolution. INTRODUCTION Sustainable and optimal management of soil and water resources requires mathematical models that predict subsurface flow and transport processes. These models depend on the physical system geometry, boundary conditions and soil hydraulic properties such as the hydraulic conductivity and water retention functions. A major shortcoming in the application of these models is the lack of knowledge regarding the spatial and temporal distributions of the model parameters at scales that are pertinent to the management scales. In this respect, ground-penetrating radar (GPR) appears to be promising noninvasive method for characterizing and mapping the soil hydrological properties at the field scale with a high spatial resolution. Recent reviews on the use of GPR in soil and hydrological sciences are given by Huisman et al. (2003) and Annan (2005). GPR methods still suffer from the two major shortcomings. First, the forward model describing the radar data is subject to relatively strong simplifications with respect to electromagnetic wave propagation phenomena. This results in inherent errors in the reconstructed water content images and, moreover, this does not permit to exploit all information contained in the radar data. To overcome this limitation, it is necessary to resort to full-waveform forward and inverse modeling of the GPR data, which can be easily used due to available computing resources nowadays (Gloaguen et al. (2007a); Soldovieri et al. (2005); Lambot et al. (2004)). A second problem is the uniqueness of the inverse solution, which usually constrained by providing additional information through smoothness conditions or via more advanced geostatistical models (Giroux et al. (2007); Gloaguen et al. (2007b)), or by including data from additional sensors (Ghose and Slob (2006); Linde et al. (2006)). A promising approach for time-lapse analyses is to resort to integrated, or usually referred to as joint, hydrogeophysical inversion techniques, where the geophysical and hydrodynamic models contain each other

ABSTRACT Local cosine/sine basis is a localized version of cosine/sine basis with a smooth window function. It has orthogonality and good time and frequency localization properties together with a fast algorithm. In this paper, we present a new method combining the main idea of local cosine/sine bases, multi-scale decomposition and the dispersion relation to form a multi-scale, multi-dimensional self-similar dreamlet transform. Meanwhile, a storage scheme based on the dreamlet decomposition is proposed by using zig-zag sequence. We apply this method to the SEG-EAGE salt model synthetic poststack data set for data compression. From the result, almost all the important features of the data set can be well kept, even in high compression ratio. Using the reconstructed data for migration, we can still obtain a high quality image of the structure. Through comparing with other decomposition and compression schemes, we believe that our sheme is more closely related to the physics of wavefield and has a better performance for seismic data compression and migration.

Summary In this paper we propose a new statics method for very thick near-surface structures with near-surface gravel deposits and a huge sand dune in western China. This method is based on getting the thickness of the near-surface deposits by long-array refraction survey combined with the use of up-hole survey, getting the velocity using the theory of continuous media and tomographic inversion. Finally, statics are computed by time-depth curve variation with space. We present an example which indicates that the new method is helpful for solving statics problems of very thick near-surface structures. Introduction Super-thick gravel deposited at piedmonts and huge sand dunes are distributed widely throughout western China. Its velocity varies continuously in lateral and vertical directions because of the action of gravity and nonhomogeneity. It is very difficult to find a clear velocity interface and its surface is inconsistent. The near-surface investigation method based on a layered medium and the static method depend on the assumption of surface consistency can not be applied in this area because of low accuracy. The surface velocity varies continuously in the vertical direction because of action of gravity; in addition, alluvial fan causes the velocity to change laterally. Many alluvial fans exist at piedmonts. Surface of alluvial fan is generally flat but the weathering layer is very thick and the high velocity interface varies greatly with the features of continuous media. The two T-D curves in Fig.1 show that the interval velocity changes differently at different positions in the alluvial fan. The layering of velocity in the vertical direction is indistinct, showing characteristics of a continuous medium. For a continuous medium, information received from near offset shallow-refraction is contraflexure instead of a refraction wave, which causes major discrepancies when the normal interpretation is applied. So it is significant to use the survey method for super-thick near surface layers with continuous media (Mike Cox, 1999, p.17). Method and techniques 1. Theory of continuous media The velocity of a continuous medium changes gradually and it is the function of depth (He, 1986, p.91), following equation (1). 2. Solution of near-surface thickness Normal survey methods can not be properly carried out in sand dunes and gravel zones deposited at piedmont. So the following two special ways were applied in regions covered with continuous media. If the interface of the high velocity layer can be computed accurately, the thickness of the low velocity zone can be obtained. 1) Shallow-refraction with long spread After inversion of theoretical model and analysis of real near-surface features, we drew the conclusion that an additional spread and a greater number of receivers were necessary to get the refraction wave from the high-velocity layer. The better the low-velocity zone layers were interpreted, the more accurate the thickness of these layers will be. A spread was designed to meet the requirements of continuous media: 48 channels, 454 m length with receiver intervals of 1, 1, 2, 2, 3, 3, 5, 5, 5, 10, 10, 10, 10, 10, 15, 15, 15, 20, 20, 20, 15, 15, 10, 10, 10, 15, 15, 20, 20, 20, 15, 15, 15, 10, 10, 10, 10, 10, 5, 5, 5, 3, 3, 2, 2, 1, 1 (m).

ABSTRACT Time-lapse seismic has proven to be a useful tool for reservoir management and is increasingly becoming a standard practice to monitor production, identify undrained resources, accelerate the production of reserves. Dedicated inversion tools have been developed to obtain time-lapse seismic images to provide in a reasonable amount of time reliable 4D signal in a shared earth model. As an efficient substitute to full 4D inversion, a series of tools sharing the denomination of ‘warping’ have been developed in order to produce relevant 4D difference volumes. These include correlation-based methods, time-shift non-linear inversions and relative velocity (or time-strain) non linear inversions. Standard correlation-based methods and non-linear time-shift inversion are mainly used to align base and monitor surveys and generate time-shifts which may sometimes be used interpretatively but in general have insufficient resolution and stability to be of interest as 4D reservoir attributes. Even after alignment the difference volumes, while characteristic of changes in the reservoir, are difficult to interpret quantitatively, due to the band-limited nature of seismic data. Conversely, when formulated as a non-linear inversion to retrieve interval quantities like relative velocity changes or time strain, the warping obtains information which can be easily integrated in the reservoir models, at least for the relatively numerous cases where the 1D-approximation can be used on post-migration time data, typically from PSTM or PSDM imaging. However this inversion formulation is seriously ill-posed, which thus requires careful regularisation and the integration of geosciences at every stage of the inversion. Conversely the regularisation still allows working trace by trace. This is computationally efficient as it does not require inversion of the 3D dataset as a whole. As an application three examples are discussed. Case A is an HPHT reservoir with very low signal to noise ratio and problematic repeatability close to the platform; in this case dedicated warping allows obtaining stable 4D signal at the reservoir level as well in the overburden. Case B is a turbiditic reservoir where a 4D signal due to gas either injected or coming out of solution under depletion is measured. Case C is a stress-sensitive reservoir where a stable 4D signal inverted at the reservoir level correlates very well with production data.