Nonlinear Analysis and Modeling Techniques Applied to Geophysical Systems Posters
Presiding: B A Murray, Duke University; A S Sharma, University of Maryland at College Park
NG23A-01 1330h
NARMAX Approach Based Identification of the Nonlinear Processes of the Magnetosphere
The NARMAX approach allows the user to build concise nonlinear models from data by identifying the model structure and estimating the unknown parameters. The narmax method has been successfully applied to many real systems ranging from brain imaging to the oil industry. We have applied the NARMAX approach to derive a global continuous-time physical model of the Dst reflected magnetospheric dynamics, and have shown that results based on Markov type models and on local linear filters can be deduced from the global NARMAX model.
NG23A-02 1330h
Conversion of Bouguer Gravity Data to Depth, Dip, And Density Contrast With Complex Attributes Analysis Technique in the Area of Greece.
The complex attributes analysis is an operator used in the extracting parameters of the buried structures with susceptibility and density contrasts distributions, which lead to the gravity and magnetic anomalies in the region of interest. In this paper is presented the complex attributes analysis of gravity field filtered for wavelengths lower than 50 km in the territory of Greece. The area o Greece has a complex tectonic history and fault system dominated by the subduction of the African plate beneath the Euroasia. A Low-pass filter is used on the Bouguer Anomaly to cut off wavelengths lower than 50Km in order to delineate the major faults structures of interests at big depths. The complex attributes technique aids in interpretation of potential field anomalies, because it can delineate the edges of concealed targets. In obtaining the source parameters from the complex attributes like the local depth, strike and dip, the assumption of sloping contact for the subsurface model is used. The estimated local parameters are in agreement with results obtained by previous interpretations. They can be used in combination with other method to interpret the anomalous field.
NG23A-03 1330h
Nonlinear Classification of AVO Attributes Using SVM
A key research topic in reservoir characterization is the detection of the presence of fluids using seismic and well-log data. In particular, partial gas discrimination is very challenging because low and high gas saturation can result in similar anomalies in terms of Amplitude Variation with Offset (AVO), bright spot, and velocity sag. Hence, a successful fluid detection will require a good understanding of the seismic signatures of the fluids, high-quality data, and good detection methodology. Traditional attempts of partial gas discrimination employ the Neural Network algorithm. A new approach is to use the Support Vector Machine (SVM) (Vapnik, 1995; Liu and Sacchi, 2003). While the potential of the SVM has not been fully explored for reservoir fluid detection, the current nonlinear methods classify seismic attributes without the use of rock physics constraints. The objective of this study is to improve the capability of distinguishing a fizz-water reservoir from a commercial gas reservoir by developing a new detection method using AVO attributes and rock physics constraints. This study will first test the SVM classification with synthetic data, and then apply the algorithm to field data from the King-Kong and Lisa-Anne fields in Gulf of Mexico. While both field areas have high amplitude seismic anomalies, King-Kong field produces commercial gas but Lisa-Anne field does not. We expect that the new SVM-based nonlinear classification of AVO attributes may be able to separate commercial gas from fizz-water in these two fields.
NG23A-04 1330h
SAR Interferometry: On the Coherence Estimation in non Stationary Scenes
The possibility of producing good quality satellite SAR interferometry allows observations of terrain mass movement as small as millimetric scales, with applicability in researches about landslides, volcanoes, seismology and others. SAR interferometric images is characterized by the presence of random speckle, whose pattern does not correspond to the underlying image structure. However the local brightness of speckle reflects the local echogenicity of the underlying scatters. Specifically, the coherence between interferometric pair is generally considered as an indicator of interferogram quality. Moreover, it leads to useful image segmentations and it can be employed in data mining and database browsing algorithms. SAR coherence is generally computed by substituting the ensemble averages with the spatial averages, by assuming ergodicity in the estimation window sub-areas. Nevertheless, the actual results may depend on the spatial size scale of the sampling window used for the computation. This is especially true in the cases of fast coherence estimator algorithms, which make use of the correlation coefficient's square root (Rignon and van Zyl, IEEE Trans. Geosci.Remote Sensing, vol. 31, n. 4, pp. 896-906, 1993; Guarnieri and Prati, IEEE Trans. Geosci. Remote Sensing, vol. 35, n. 3, pp. 660-669, 1997). In fact, the correlation coefficient is increased by image texture, due to non stationary absolute values within single sample estimation windows. For example, this can happen in the case of mountainous lands, and, specifically, in the case of the Italian Southern Appennini region around Benevento city, which is of specific geophysical attention for its numerous seismic and landslide terrain movements. In these cases, dedicated techniques are applied for compensating texture effects. This presentation shows an example of interferometric coherence image depending on the spatial size of sampling window. Moreover, the different methodologies present in literature for texture effect control are briefly summarized and applied to our specific exemplary case. A quantitative comparison among resulting coherences is illustrated and discussed in terms of different experimental applicability.
NG23A-05 INVITED 1330h
Stochastic Theory of Compressible Turbulent Fluid Transport
We develope a stochastic model for the turbulent transport of passive scalars based on the Fokker-Plank equation for the probability density distribution of the displacements of infinitesimal fluid parcels in randon turbulent motion. Such a theory is the microscopic basis behind semiempirical models of turbulent diffusion, which apparently have only been developed for incompressible flow. Here we specifically develop the theory so that it applies to compressible flow. We apply it to the particular case of stratified mesoscale turbulent transport of tracers in the ocean, and we find that it generalizes the recent parameterization of Gent and McWilliams (JPO 20 150 (1990)).
NG23A-06 1330h
Stochastic Models of Quasigeostrophic Turbulence
In this talk, I present a closure theory of quasigeostrophic turbulence for arbitrary shear flows based on stochastic models. A stochastic model represents the eddy-mean flow interaction through a nonnormal dynamical operator, and parameterizes the eddy-eddy interactions by an effective dissipation and random excitation. In the context of a stochastic model, the main facts which a closure theory of turbulence must explain are (1) the form of the eddy dissipation operator, (2) the magnitude of the eddy dissipation, (3) the space-lag correlation of the random forcing at every point in space, (4) the rate of enstrophy transfer to subgrid-scales. We present a theory which accounts for all of these. The theory gives good quantitative agreement for the structure of eddy fluxes and variances obtained from fully nonlinear simulations, and predicts a -3 slope for the equilibrium energy spectrum. The theory is based on the hypothesis that the eddy-eddy nonlinear terms depend only on the local properties of the eddy statistics. Under a linearity assumption, this theory can be solved for all of the unknown parameters in the stochastic model except one, namely (4) listed above. However, the stochastic model predicts a -3 energy spectrum at asymptotically large wavenumbers, consistent with inertial range theory. By invoking inertial range theory for the functional dependence of the energy spectrum with enstrophy cascade rate, the stochastic model can be closed completely. This theory, which avoids phenomenological mixing-length arguments, can be applied to arbitrary shear-flows and gives reasonable statistical predictions of midlatitude eddy fluxes and variances.
NG23A-07 1330h
The Spatio-Temporal Structure of the Magnetosphere during Magnetic Storms
Input-output analysis of geomagnetic indices and ground magnetometer measurements as the output and the solar wind as variables as the input is used to generate a spatio-temporal dynamical model of the magnetospheric dynamics. The data of magnetic field perturbation with 1-minute resolution ground stations during 2002 are used to study this spatio-temporal structure, mainly on high latitude magnetic perturbations. All of the 57 ground magnetometers are from 3 station group--CANOPUS(13), IMAGE(26) and WDC(18). A technique that utilize the daily rotation of the Earth as a longitudinal sampling mechanism is used to construct a two dimensional representation of the high latitude magnetic perturbations both in magnetic latitude and local time. The data of magnetic field perturbation at the magnetometer stations are used as the output of the nonlinear system driven by the solar wind. The model is used to predict the spatial structure of geomagnetic disturbances during intense geospace storms.
NG23A-08 1330h
Correlation Dimension of Current Reversal Models
Some features of the magnetospheric activity seem to possess low dimensional dynamics. Self-organised criticality (SOC) is often referred to as the statistical behaviour responsible for the observed distribution of the electrojet indices. One of the models extensively used for this objective is the field-reversal model. We apply the correlation dimension analysis to an implementation of the field reversal model to show that, although the field reversal model does exhibit low dimensional dynamics, quantitatively the dimension of the dynamics does not agree with that calculated from the dataset of electrojet indices. Thus, the electrojet indices statistics cannot be explained directly by this model.
http://www.acse.shef.ac.uk/~bates/research/?file=abstracts
NG23A-09 1330h
Stochastic Large Eddy Simulation of Geostrophic Turbulence
Results are presented of (fine-scale) eddy-resolving simulations of different instances of turbulent quasi-geostrophic ocean circulation. A stochastic model for the effects of neglected subgrid degrees-of-freedom in coarse-scale simulations is proposed and the results compared to the fine simulations results as well as with existing models. As a precursor to the introduction of the models, we also study various aspects of the nonlinear rectification of stochastic forcing in quasi-geostrophic models of ocean circulation.