ارزیابی استاتیک شیفت و تصحیح لازم برای کاهش اثرات آن در برداشت برای مگنتوتلوریک

دسته ژئوفیزیک
گروه سازمان زمین شناسی و اکتشافات معدنی کشور
مکان برگزاری بیست و ششمین گردهمایی علوم زمین
نویسنده علی مرادزاده
تاريخ برگزاری ۰۲ آبان ۱۳۸۴

Abstract
Magnetotellurics is one of the natural source electrical techniques that is used in many exploration projects. In this method the role of small-scale near-surface inhomogeneity are very important and cause that the subsurface anomaly appear shallower or deeper depending on the resistivity of near-surface bodies.  This phenomenon, which is independent of frequency, called static shift in MT surveys and its effects on data must be eliminated or reduced someway prior to any numerical modeling.  In this paper, it will first be shown that static shift could be recognized if data are given as pseudosections of apparent resistivity and phase versus frequency along a MT profile and its effects on data can then be minimized by taking spatial average of the apparent resistivity using Hanning window.
Keywords: MT, static shift, near surface inhomogeneity, resistivity, TE and TM modes, spatial average, Hanning window
     
Introduction
The major problem that is encountered in MT explorations is the static distortion of the electric response of subsurface structure due to near-surface lateral variations in electrical resistivity (r).  The cause of this problem is that the horizontal electric field, E, is discontinuous across a lateral discontinuity (Figure ۱a). The boundary condition states that the normal current density, J, must be continuous across the vertical interface and since J=E/r, then the electric field amplitude discontinuity E۱/E۲ across the contact is equal to the resistivity ratio r۱/r۲.  The electric field behavior at a contact may be used to illustrate the effect of a small near-surface inhomogeneity such as that shown in Figure ۱b.  As long as both frequency and resistivity are such that the skin depth in the inhomogeneity is greater than its dimension, the electric field is uniformly reduced within it as shown schematically in Figure ۱a.  This reduction in electric field decreases the calculated apparent resistivity.  This effect is constant for all frequencies such that when sounding data plotted on a logarithmic scale the entire apparent resistivity curve is depressed (Figure ۱c).  This problem is called static shift in MT.
 
Static shift detection
To detect static shift in MT data, they could be plotted as stacked apparent resistivity and phase curves versus frequency or as a pseudosection along a profile.  The latter one is presented in this study.  The tensor nature of MT data dictates that both TE (Transverse Electric) and TM (Transverse Magnetic) mode data to be treated separately.
In a pseudosection of apparent resistivity versus frequency along a MT profile, the static shift would be seen as a series of vertical striped zones parallel to the frequency axis, whereas these vertical zones are not supported by phase pseudosection along the profile.  As an example, the TM mode pseudosection of Adelaide Geosyncline MT data (Moradzadeh, ۱۹۹۸a) is given in Figure ۲.  This figure shows a high conducting zone (ryx <۱۰ Wm) beneath sites YOR (east of DON), STI, WOO (east of STI), and SPE.  This conductive zone is extended for frequencies greater than ۱ Hz for YOR to WOO while it manifests itself over the whole frequency range at SPE.  The phase pseudosection does not support the existence of such a conductive structure, however, like TE mode (not shown) it shows a relatively conductive zone at the near surface ( jyx @ ۵۵۰) under STI to SPE.  Consequently the existence of such a small near surface conductive body has caused the resistivity data beneath these sites to be shifted downward for the whole frequency range.  Although there are a few more vertical zones beneath sites DAW (between OAK and ODD) to FAR in the apparent resistivity pseudosection, they are fairly well supported by the phase data of the corresponding phase pseudosection. 
 
Static shift correction
To obtain an accurate interpretation of MT data, the effects of small near-surface inhomogeneities on the apparent resistivity data must be taken into account and reduced somehow. Unfortunately the amount of static shift for each site cannot be determined directly from the conventional (sparse) MT data sets.  Therefore, in such situations where no other independent data are available any correction of the static shift is based on some assumptions.  Having such assumptions, different approaches, such as spatial averaging or spatial filtering (Berdichevsky et al., ۱۹۸۰; Sternberg et al, ۱۹۸۲; Moradzadeh, ۱۹۹۸b), use of invariant impedance (Berdichevsky and Dmitriev, ۱۹۷۶), and an enhanced spatial filtering of continuous MT profiling (Torres-Verdin and Bostick, ۱۹۹۲) were introduced to deal with this problem. 
The galvanic effects of local near surface inhomogeneities can be taken as random phenomena and consequently, statistical approaches can be applied to remove or suppress such undesired effects on the data (Moradzadeh and Chamalaun, ۱۹۹۷; Moradzadeh,۱۹۹۸a). Based on this assumption many methods of averaging the data have been applied to MT data (i.e. Berdichevsky et al., ۱۹۸۰; Sternberg et al, ۱۹۸۲; Warner et al., ۱۹۸۳; Beamish and Travassos, ۱۹۹۲; Zhang et al., ۱۹۹۵).  We propose an approach that the spatial average of the apparent resistivity data is calculated using Hanning window with the following formula.
                              |x|£w/۲
                                               |x|>w/۲                        
where x is distance,  w is averaging interval or window’s width, and H is the magnitude of the weight given by the window.   
This process represents low-pass spatial filtering, which can smooth static shift to some extent.  To implement such a method to the apparent resistivity data, first arithmetic average of the logarithm of the apparent resistivity at each site (as site average) was calculated then its spatial average was calculated by weighting surrounding sites using a Hanning window of ۱۰۰ km width.  After the calculation of the site average and spatial average for each site, the apparent resistivity curves were shifted by a value given by the difference “spatial average-site average”. The apparent resistivity curve each site was then shifted by a difference derived from spatial and site averages. 
The evaluated static shift values of each site and MT polarization mode are plotted in Figure ۳.  A negative static shift states that the measured resistivity has been depressed from the value it would have in the absence of a distorting anomaly.  From the data in Figure ۳, it can be concluded that the static shifts are small and sum of them (in log scale) for both TE and TM mode are close to zero (۰.۰۴ for TE mode and ۰.۱۳ for TM mode).  These plus other statistical parameters (such as mean, median, skewness, and standard deviation) suggest that the static shift is very close to a random phenomenon.
In the next step the calculated static shift factors have been applied to the apparent resistivity data of each mode and the corrected data are shown as pseudosection in Figure ۴ for TM modes.  From this figure it can clearly be seen that the data improved significantly so that sub-surface resistivity structure given by the apparent resistivity data are well supported by the phase data  (compare Figure ۲ to Figure ۴).  This means that although the static shifts appear small, they are nevertheless significant.  The good agreement between resistivity and phase data in Figure ۴ clearly indicates that the static shifts in the data have been smoothed out by proposed procedures.
Conclusion
The near surface inhomogeneities are able to depress or enhance the electrical response of deep sub-surface anomalies in MT soundings. They cause that the apparent resistivity curve of a MT sounding shifted up or down in all frequencies so that without considering them the results of survey could mislead interpreters. This study shows that static shift could be detected when the apparent resistivity and phase data are plotted as pseudosection versus frequency and their effects could effectively be reduced using spatial filtering of resistivity data by Hanning window.    
 

چکیده
روش مگنتوتلوریک یکی از روشهای اکتشاف ژئوفیزیک الکتریکی با چشمه طبیعی است که در بسیاری از پروژه های اکتشافی بکار می رود. در این روش نقش ناهمگونیهای سطحی در سونداژ الکتریکی بسیار مهم است طوریکه بسته به مقاومت ویژه الکتریکی که دارند ممکن است نتایج کل سونداژ و مقاومت ویژه را به سمت بالا یا پایین شیفت دهند و ابهاماتی را در تعبیر و تفسیر آنومالیهای عمیق ایجاد کنند. این پدیده که مستقل از فرکانس اندازه گیری است شیفت استاتیکی نامیده می شود و اثرات آنها روی داده های سونداژ الکتریکی قبل از هرگونه تعبیر و تفسیر و مدل سازی عددی بایستی حذف و یا حداقل کاهش یابد.
در این مقاله در ابتدا بحث خواهد شد که اگر داده های مقاومت ویژه و فاز مربوط به سونداژهای یک پروفیل در شبه مقاطعی که محور افقی آنها فاصله و محور قائم فرکانس باشد،  نشان داده شوند استاتیک شیفت به خوبی از روی عدم هماهنگی این دو نوع داده معلوم خواهد شد. سپس اثرات ناهمگونیهای سطحی روی داده ها بکمک متوسط گیری فضایی  داده های مقاومت ویژه به شیوه مخصوص به حداقل خواهد رسید.

 

کلید واژه ها: مگنتوتلوریک مقاومتویژه ژئوفیزیک سونداژ سایر موارد