Home

USING SURFSEIS - the Kansas Geological Survey

1. Pause A see Section 3 3 1 L Left Click R Right Click 6 2 Dialog Controls Display of Dispersion Curve 6 3 Continued Description Automatically scales both horizontal and vertical axes to Toggle on off display of signal to noise ratio curve Zooms in on particular portion of display Brings up a dialog box enabling various display parameters to be changed see Section 6 2 Presents processing history of current dispersion curve Stops processing available only during Pilot Analysis Pauses processing available only during Pilot Analysis A Action Button D Dialog Button Dialog controls enable changes to be made in display attributes and can be activated by either clicking the Control button or by right clicking the mouse There are four categories in this option Data Points Axis Labeling and Panel 6 2 1 Data Points Display attributes of data points line and mark are important when multiple files are displayed simultaneously A specific file to change can be indicated in the Apply to File box In the Line tab the style thickness and color of the line connecting data points can be changed In the Mark tab the style size and color of the data point marks can be modified You can select dispersion and or S N curves to apply the indicated changes to En Line bin Eyhr Ticks nen l habre 1 Dni f a fi End Cd Oooh poe gg ApphinC
2. 2 Overtone analysis is the best way to observe the dispersive nature of any type of seismic wave without the bias associated with event interpretations Fundamental principles of this analysis Park et al 1996 are quite straightforward This type of analysis basically attempts to construct a wiggle trace 1mage where local amplitude maxima trends represent possible dispersive energy 1 e fundamental and higher modes This is accomplished by examining all possible phase velocities for all freguencies being considered For most dispersion curve analysis itis RWTH TT TT recommended that Overtone analysis be performed right after the Preprocess step Phase Velocity Frequency Method Many times critical information such as phase velocity range and optimum frequency range MIN presence of strong higher modes and any body wave noise may already be reasonably well known MAE 1200 E There are two parameters to define for this analysis frequency and phase velocity Default values determined by the program will usually be sufficient for most cases But you Figure 3 4 1 have access to these values through the right mouse button which displays the dialog box in Figure 3 4 1 The Algorithm operation in the Method tab 1s the same operation as described in Section 3 3 3 Examples from different algorithms Approximate Figure 3 4 2 Normal with Integrity 1 Figure 3 4 3 and Norma
3. a i i LITT flo pa LT hi E LL LLAS da x m n 4 Determines scale of each seismic trace Size Amplitude will determine the maximum horizontal deflection in trace spacing of the maximum amplitude value of a trace or entire record if Normalization option below is Off Gain increases or decreases the display amplitude or gain in decibel Fill Right side positive values Type Fill left side negative values and No Fill neither side of wiggle will be filled Scales entire trace relative to the maximum Normal _ value of each trace ization Eira Coin dil Lm Sum piauda in traca speci 5i Type St3 FW FE He Pil Hume un 5 3 2 Image Tool Bar Dialog Scale Te Beale ciue sE Option Description Vertical Increases vertical size of displayed image arica by the specified multiplier Sal Cato Fado ui Increases horizontal size of displayed Horizontal image by the specified multiplier increases trace spacing Seismic Data Display 5 5 5 3 3 Image Tool Bar Dialog Time and Trace re pre FE Option Description Display Time Selects range of time displayed Display Time Range ms Range Begin r End fn Display Trace Selects range of traces displayed e Range Cesgla Troca Fiano E 5 3 4 Image Tool Bar Dialog Labeling Win scala Wi TAE Cabina Option Description Horizontal Leti y Trace Option fo
4. di i HN E si a i L i i 1 1 x T iL I mf CAER grh EAN a i 1 I LI I Figure 8 5 2 FEE 8 Figure 8 5 3 Figure 8 5 4 Working with SurfSeis 8 25 o IND Fas LE M Ai a 2n al Tr La du 43 fien Ts RAI s a RJT i uc Dispersion Curve a 5 alas al ow ae a n atta Figure 8 5 5 Figure 8 5 6 pr Far gt EDODED s PE EDODED e PIE ii T RARA AAA eS RE Te 53 DM YN DE MPO Le Figure 8 5 7 IE MU battu Figure 8 5 9 Appendices 9 1 Appendices Appendix __ Index Bibliography and Recommended Reading on MASW 10 1 Bullen K E 1963 An introduction to the theory of seismology Cambridge University Press 381 pp Coruh C 1985 Stretched automatic amplitude adjustment of seismic data Geophysics v 50 p 252 256 Gucunski N and Woods R D 1991 Instrumentation for SASW testing in Geotechnical special publi cation no 29 Recent advances in instrumentation data acquisition and testing in soil dynamics edited by S K Bhatia and G W Blaney American Society of Civil Engineers p 1 16 Heukelom W and Foster C R 1960 Dynamic testing of pavements Journal of the soil mechanics and foundations division v 86 n SMI 1 28 Ivanov J Park C B Miller R D and Xia J 2000 Mapping Poisson s Ratio of unconsolidated mate rials from a joint analysis of surface wave and refraction events Proceedings of the Symposium on the Application of Ge
5. Full Size Normalize Gain Print Save Restore Scroll Bar Select Mouse Horizontal Vertical Position Seismic Data Display 5 3 Mouse Button L R Description A Displays zoom of the data selected by mouse D Displays a dialog box for viewing and changing display controls simultaneously Sections 5 3 1 5 3 4 A Displays maximized view of record in current window A Toggle trace normalization on off A A Changes gain L up R down D Prints displayed image after responding to a dialog box based on print setup D Saves displayed image in a bitmap file BMP after responding to an option dialog box A Restores original unprocessed record Available to user only after the data has been processed in some way A Displays both horizontal and vertical scroll bars Both Horizontal and Vertical buttons below are enabled A Selects one or more traces for processing A Clears any mouse generated marks or symbols A A Changes horizontal scale L up R down Enabled only when Scroll Bar button BI is clicked A A Changes vertical scale L up R down Enabled only when Scroll Bar button BI is clicked A Toggle on off displayed location of the position of mouse cursor in trace number and time L Left Click R Right Click A Action Button D Dialog Button Seismic Data Display 5 4 5 3 1 Image Tool Bar Dialog Wiggle Option Description Tau nrd Incu 1 Lnhuling
6. Tolerance option under the computation tab in the Control dialog box Figure 8 3 18 sets the range of velocities to search at a particular frequency Once an accurate phase velocity has been calculated for the reference frequency the process moves to adjacent frequencies At this point the range of phase velocity searched through is updated according to the phase velocity values estimated during the previous calculation The tolerance option sets the upper and lower bounds for this searching range The existence of higher modes needs to be evaluated to properly set this option The tolerance option has no effect on the calculation when no higher mode energy or any body wave event 1s observed within a velocity range with an upper limit at least two times higher than the phase velocity of the fundamental mode It is critical to properly set this option if higher modes are present at energy levels comparable to the fundamental mode and in the same frequency range as the fundamental mode The significance of this option 1s dependent on the difference in phase velocity dVphs between the fundamental V0 and next higher mode V1 observed in the overtone image The following guides describe how to properly choose an option 1 Small if dVphs lt 0 25 x VO 2 Moderate if dVphs lt 0 50 x VO 3 Large if dVphs gt 0 50 x VO Working with SurfSeis 8 14 The Algorithm setting under the computation tab in the Contro
7. al s e Normalize amplitude Normalization spectra 0000 Smoothing Smoothing amplitude spectra Smoothing F Mo x Cercel Chapter 6 Display of Dispersion Curve Dispersion curves saved into a text file DC possess frequency vs phase velocity domain and signal to noise ratio curves This data can be examined manipulated and handled using moduli discussed in this chapter This chapter includes a Displaying dispersion and signal to noise ratio S N curves a Printing curves a Saving images of curves as a bitmap file BMP 6 1 Button Controls Button pen EENE a E Uu Din Display of Dispersion Curve 6 2 Name Open Save Print March X Up X Dwn Y Up Y Dwn Mouse Button L R Description Opens a file for display Saves displayed image as a bitmap file BMP Sends displayed graph to a printer based on print setup Continuously displays sequential files Enabled only when multiple files are open Increases horizontal scale of displayed graph increases Hz inch Decreases horizontal scale of displayed graph decreases Hz inch Increases vertical scale of displayed graph increases Hz inch Decreases vertical scale of displayed graph Mouse Button Button Name L R Te Aui Auto A fill screen Ez um NA Zam Zoom A i Control A Control History p Pause History A Stop A see Section 3 3 1
8. Ch 1988 dispersion law of Rayleigh type waves in a compressible Gibson half space International Journal for Numerical and Analytical Methods in Geomechanics v 12 p 639 655 Waters K H 1978 Reflection seismology John Wiley and Sons Inc Xia J and Miller R D 2000 Fast estimation of parameters of a layered dipping earth model by invert ing reflected waves Accepted for publication in Journal of Environmental and Engineering Geo physics Xia J R D Miller C B Park and J Ivanov 2000 Construction of 2 D vertical shear wave velocity field by the multichannel analysis of surface wave technique in Proceedings of Symposium on the application of geophysics to engineering and environmental problems SAGEEP 2000 Arlington Va p 1197 1206 Xia J R D Miller and C B Park 2000 Advantages of calculating shear wave velocity from surface waves with higher modes Exp Abs Soc Explor Geophys 1295 1298 Xia J Miller R D and Park C B 1999 Estimation of near surface shear wave velocity by inversion of Rayleigh waves Geophysics v 64 no 3 p 691 700 Xia J R D Miller C B Park J A Hunter and J B Harris 1999b Evaluation of the MASW technigue in unconsolidated sediments Exp Abs Soc Explor Geophys p 437 440 Xia J R D Miller C B Park E Wightman and R Nigbor 1999c A pitfall in shallow shear wave refraction surveying Exp Abs Soc Explor Geophys p 508 511 Yilmaz O 19
9. M 250 500 Time xxi hwr gwm a Lerida i 1000 Working with SurfSeis 8 3 3 2 Data Format Most engineering seismographs output SEG 2 natively Other output options will generally include SEG D or in a rare instance SEG Y Processing seismic data requires organization of information within the trace in a very structured form Most seismic processing packages use data in SEG Y or a slightly modified version of SEG Y SurfSeis uses the same modified SEG Y data format used by WinSeis and WinSeis Turbo This modified form of SEG Y is trace by trace sequential and possesses a fixed header length followed by data Conversion routines are available for most seismograph and processing formats Trace headers sample resolution and data order are what designate the format of a seismic data file SurfSeis reads and writes data in the KGS modified SEG Y format This data format 1s unique and requires conversion before anything can be done with or to data regardless of whether it came directly from a seismograph or other processing software Several data formats are presently being used by commercial seismograph manufacturers The engineering standard is designated as SEG 2 Pullan and Hunter 1990 Conversion of data files requires the user to double click the format header icon and select the appro priate conversion routine Because the majority of modern engineering seismographs are using SEG 2 reference will be made prima
10. and draw a rectangular region around the noise to attenuate filter out Click on the image to drag and draw a straight line that specifies a time boundary Unnecessary Unnecessary Double Click Displayed Image Band reject filtering is applied to attenuate noise identified by PAM In the case of Option A in PAM a rectangular region centered at the mouse cursor position with an area one fifth the displayed image window is chosen as noise Seismic energy above earlier than the specified time boundary is zeroed with a smooth tapering at the boundary Gain is applied using a scaling window one fifth the displayed record length Amplitude spectrum of each trace is displayed separately Seismic Data Display 5 8 Figure 5 1 i NA aron A m SR TRW i MM SN dani SALI tora les TR x s s te A j TN ji Me T HOI iN ni mu m zl EL E n LL E ie LEES OW Exe m am m r RN 23 L E a e a a i agni mr n Tr re m i ij il i I ni il Ti F H ru aa aa QRO a n o O OOL 10 L e ei JI Va EIA Tat ur Dr il MH en eer Ale Ti HI HIE Mare ttr HEEL TH IHRE TI AAN E AAA mas dem d m uim ovr Er ram z LEN NI n F n a e B d E a WN n Ly m AAA e lm AAA AAA ee Aa i MWYN la Ne Hi FERRE in ka SL M y aeth uio Hi M SIT Miu Pe ERE us T i Er a mc i ars ST my
11. in the native data format of the seismograph and need to be converted into KGS format using conversion routines in the icon Headers and Conversions Converted data can contain single or multiple shot gathers records depending on acquisition parameters and seismograph settings SurfSeis includes a variety of display options including a Record Checking a Image Display Q Seismic Data Processing Seismic Data Display 5 2 5 1 Main Tool Bar Ss Image Process Mouse Sutton Name Button Description L R Record Record A Toggle to display hide the seismic record tool bar Image Image A Toggle to display hide the wiggle 1mage tool bar AL Process Process A Toggle to display hide the data processing tool bar L Left Click R Right Click A Action Button D Dialog Button 5 2 Record Tool Bar ac gt 5 Mouse Button Name Button Description L R lt First A Selects and displays first record of input data Previous A Selects and displays previous record gt Next A Selects and displays next record Last A Selects and displays last record 7 Jump D Selects user defined record and displays it Info A Displays seismic data information Save D Saves displayed seismic record as individual file DAT L Left Click R Right Click A Action Button D Dialog Button 5 3 Image Tool Bar A 88 21880584 amp E HE S o E ee ee eB Be Name Zoom Dialog
12. input IND and output IVO and LST files will be displayed with appropriate extensions These names however can be overwritten as desired Once an input seismic file is selected and displayed a preprocess step 1s applied to the input data This quick look process provides an approximate range of surface wave arrivals in a display area detected and marked by the program as shown in Figure 2 1 2 This range 1s determined by lower and upper bounds of phase velocity Phase velocities of the dispersion curve will usually fall in this range If the marked zone encompasses most of the surface wavefield the remainder of the process will proceed effectively If the marked zone 1s not consistent with the surface wave arrivals the process may need to proceed through the step by step procedure see Chapter 3 FIf the input seismic file contains more than one multichannel record the first record will be selected for processing IBI TIER in ES pun Fite dor Milch os eod Ll niw 7 DAT eden 1115 rir Guiput Film irw inversion Aora LET a Ll Figure 2 1 1 Figure 2 1 2 Full Auto Analysis 2 3 2 2 Controls Full Auto analysis 1s designed to perform the entire process of generating an S velocity profile without intervention using the parameters determined either during the process or user defined default values Some key parameters can be manually set before processing run begins In any case the def
13. method during the field tests at Fraser River Delta Vancouver British Columbia Rob Huggins and employees at Geometrics were also key supporters of the technique software development Contents at a Glance Chapter 1 Introduction sesion l 1 Chapter 2 Full Auto Analysis oooooccnccooccnnncnocnnnnnonccnnnncnncnnnos 2 1 li NR a 2 2 PM CONO cia 2 3 Chapter 3 Dispersion Analysis ALL IA LL I LL LL Y YLeeen 3 1 Eu o n a sola 3 2 2 SA b A 3 3 Je C OD LLO S ec AAA AAA AWE AA 3 4 AERE TD T I TT 3 8 SPEI ML 3 9 o ROSI 3 10 LI lana 3 10 Chapter 4 Inversion Analysis LL ALL AAA LL LI LL LL i YLneeon 4 1 bd TEE 4 2 A A 4 3 3 Layer Mode als 4 4 di Ribelle 4 5 Chapter 5 Seismic Data Display euu LL AIALL I LL LL LLeeeou 5 1 I Mata TOOL DOE scsi 5 2 2 RECA ToolBar 5 2 3 mate TOO Barn aia ias 5 3 4 Process Tool Bafs 5 0 Dy Processine SEISMIC Data uns 5 7 Chapter 6 Dispersion Curve Display seen 6 1 I Dutton CONOIS pias 6 2 2 Ddos Controls 6 3 Chapter 7 Inversion Results Display seeeeeeeesse 7 1 PEEL ERR 7 2 L on 7 3 3 Poisson s Ratio and Density i 7 3 Chapter 8 Working with SurfsSeis Lee LL LA LL Le iY LLneeon 8 1 Li Data ACQUISILONI sarai 8 2 AAA II GN SO Cy 8 3 3 Dispersion Curve Analysis sissi 8 6 4 INversion Analy
14. pattern recognition Park et al 1996 and is by far the most unbiased way of delineating phase velocity information It 1s probably most useful in identifying surface wave higher modes overtones A seismic record contains surface waves and body waves A good guality record has been defined as a record in which the surface wave energy dominates any body wave However there is always some body wave energy coincident with the surface waves Identification of body wave events and their association to freguency and phase velocity ranges 1s important for the following reasons Working with SurfSeis 8 10 Ba DPI ikki dr 13 sal x a Jas rps YA M ni i me he FFEL Figure 8 3 11 1 An approximate P wave velocity Vp function can be obtained from either refraction or reflection analysis An approximate range of surface wave phase velocities can be estimated from these data using general rules of thumb 2 Correlation of Vp with Vs from surface wave analysis for unique layers is always useful information e g Poisson s ratio on geotechnical investigations 3 Proper identification of body wave events enhances the confidence of surface wave dispersion curve analysis For example it can be used to confirm that the final dispersion curve was not mistakenly obtained from a body wave event Or 1t can make 1t possible to evaluate the influence of body wave events on the surface wave dispersion curve analysis If you s
15. profile whose theoretical dispersion curve best matches the calculated dispersion curve using the root mean square error RMSE as a guide and constraint Xia et al 1999 Each time a theoretical curve 1s calculated based on a layer model and compared with the experimental curve 1s called an iteration After each iteration the theoretical marked as Current curve 1s compared to the experimental marked as Measured curve and the Vs profile for the theoretical curve 1s compared with the initial Vs profile Figure 8 4 2 The frequency axis for the dispersion curve 1s inverted during display to allow approximate comparison of the phase velocity to the Vs profile Iterations will continue until either the minimum RMSE Emin 1s met or the maximum number of iteration max 1s reached whichever occurs first If the experimental curve is fairly accurate or makes sense from theoretical point of view then the RMSE usually drops rapidly during the first several iterations Figure 8 4 2 A high quality dispersion curve usually will be characterized by a consistent and smooth change in phase velocity with frequency For these high quality dispersion curves Emin is usually reached after just a few iterations and the analysis is completed Once a level is reached beyond which no noticeable drop in RMSE is observed Figure 8 4 3 the iteration should be stopped manually by selecting the Stop Ite button in the lower righthand corner o
16. sequence number Once that number is assigned the conversion will proceed During the conversion process the file being converted is displayed When the conversion process has terminated a complete listing of each file and any problems encountered during the conversion are listed The resulting data file will be 32 bit floating point with all traces samples assembled in a modified SEG Y format Working with SurfSeis 8 6 8 3 Dispersion Curve Analysis The following are the most influential factors affecting the dispersion analysis approximate phase velocity range optimum frequency range for depth of interest ease in 1dentifying higher modes and identification and reduction of noise events RE The influence of these factors on the analysis 1s highly dependent on the data quality signal to noise A good quality data set suggests the surface wave event is the most prominent seismic event 1 e the highest S N whereas a bad quality data set 1s usually contaminated by noise In dispersion curve analysis the fundamental mode surface waves are signal and everything else 1s noise Noise includes all higher mode surface waves as well as all body wave events There are a variety of ways to obtain a dispersion curve The following sequence 1s recommended Data Quality Figures 8 3 1 and 8 3 2 are examples of good quality data The data set in Figure 8 3 1 SanJose DAT was obtained at a soil site near San Jose California
17. to control all parameters used during the dispersion analysis Controls parameters are grouped into four types Analysis Type Parameters Computa tion and Curve By clicking this button a tabbed dialog box Figure 3 3 1 will pop up displaying each group of parameters in a tab If the Preprocess button has not been clicked the dialog box will only have one tab Analysis Type This is because all other parameters will be determined after the preprocessing has been executed 3 3 1 Controls Analysis Type There are two types analysis Normal and Pilot Aided Figure 3 3 1 Normal Analysis This analysis type does not use a reference HE El dispersion curve obtained from a previous pro Aaalysis Type Porametore Computation Curve cessing For that reason this type of analysis has minimal bias when calculating phase velocities Harma Plut Alded You can control most of the key parameters and examine the influence of each parameter on the performance Normal analysis is recommended in any of the following situations 1 You have an input data set consisting of multiple records collected in a continuous or X Care FET roll along mode and they have not been 5 i processed to obtain a dispersion curve before In this case you may need a high confidence Figure 3 3 1 dispersion curve to use as reference for generating other dispersion curves 2 Your input records in general
18. 000 by Kansas Geological Survey All rights reserved Information in this document 1s subject to change without notice The software described in this document 1s furnished under a license agreement not sold The software may be used or copied only in accordance with the terms of the agreement It is against the law to copy the software on any medium except as specifically allowed in the agreement No part of this manual may be repro duced or transmitted in any form or by any means electronic or mechanical including photocopying and or recording for any purpose without the express written permission of the Kansas Geological Survey Credits PROGRAMMING Primary Choon B Park Co Programmers Jianghai Kia and Julian Ivanov PROJECT MANAGEMENT Rick Miller and Mary Brohammer GRAPHIC DESIGN Julia Shuklaper and Choon B Park ADMINISTRATIVE AND TECHNICAL SUPPORT Kathy Sheldon Acknowledgments SurfSeis has been developed in association with rigorous testing of the multi channel analysis of surface waves MASW method developed at the Kansas Geological Survey During this period many people and organizations supported this program through encouragement constructive comments and funded projects Firm and continuous support from the entire staff of the Kansas Geological Survey made realization of the MASW method possible James Hunter Ron Good and their colleagues from Geological Survey of Canada provided an invaluable chance to refine the
19. 87 Seismic data processing Doherty S M Ed Investigations in Geophysics no 2 Soc of Expl Geophys
20. Italo makesi Mi Lowest 500 Highest Poot intensa 0 50 possible to change the frequency increment of the curve so that adjusting the number of data points x ceca wg within the phase velocity data set permits the Inversion process to function properly The dialog box for this option is shown in Figure Figure 3 6 1 3 6 1 Usually an excessive number of data points more than needed will simply take more computation time during the inversion process without benefit to the accuracy 3 7 Save Any one of the last three calculated curves can be saved Figure 3 7 1 After saving a curve the program automatically brings up the next record in the input file or moves to the inversion process depending on whether the file being processed contains multiple records or only one record Figure 3 7 1 Chapter 4 Inversion Analysis Inversion of the calculated dispersion curve uses the phase velocity with frequency curve as a reference to estimate the vertical S velocity Vs structure of near surface materials The inversion algorithm in SurfSeis has been adopted from Xia et al 1999 The most commonly used inversion method uses an initial model before actually beginning to search iterate for the answer An initial model consists of several key parameters S velocity vs P velocity v density p and thickness of the layers in the earth model Using this set of parameters the program begins searching for a s
21. LE USING SURFS Ers 2000 For Multichannel Analysis of Surface Waves MAS W User s Manual October 2000 Written by Choon B Park with assistance from Rick Miller Mary Brohammer Jianghai Xia and Julian Ivanov of the Kansas Geological Survey Kansas Geological Survey 1930 Constant Avenue Lawrence Kansas 66047 3726 Phone 785 864 3965 Fax 785 864 5317 Email SurfSeis kgs ukans edu Disclaimer of Warranty The SurfSeis software has been extensively tested and its documen tation including the user s manual has been carefully reviewed However the Kansas Geological Survey makes no warranty or representation either expressed or implied with respect to the SurfSeis program and its documentation its quality performance merchantability or fitness for a particular purpose The SurfSeis software 1s licensed not sold on an as 1s basis and the licensed user assumes all risk as to its quality the results obtained from its use and the performance of the program In no event will the Kansas Geological Survey be liable for any direct indirect special incidental or consequential damages resulting from any defect in the SurfSeis software documentation or program support This disclaimer of warranty is exclusive and in lieu of all others oral or written express or implied No agent or employee is authorized to make any modification extension or addition to this warranty Program Copyright 2
22. MS DOS COPY command or Windows copy procedures The following programs are currently available with SurfSeis Program Format converted from 90002FPT Bison 9000 BISCONVE Bison GeoPro 70002FPT Bison 7000 EASI2FPT EG amp G Easidisk Working with SurfSeis 8 4 SV2KGS EG amp G Seisview GEOF2KGS EG amp G GeoFlex SEG22FPT SEG 2 EG amp G 2401 StrataView ABEM MarkVI OYO DASI and most modern seismographs ABEM2FPT ABEM Terraloc OY O2FPT OYO Models 5 8 OYO160FP OYO McSeis 160 OYO170FP OYO Model 170 OSEISFPT OYO 1500 O Seis SCIN2FPT Scintrex SEGY2FPT SEG Y most types KGS2FPT KGS integer These reverse translators are also included Program Format converted to 9000SEGY SEG Y FPT2SEGY SEG Y SEG2SEGY SEG Y The particular formatting routine necessary for your data depends on the seismograph It was collected on Most new engineering seismographs read and write SEG 2 SEG 2 format accommodates variable header lengths maximizing the flexibility of the trace headers to store information during the acquisition of seismic data however once the multi data files are organized into a single file multi record configuration a fixed trace header length 1s necessary SurfSeis uses a data format that is modeled after SEG Y Raw unformatted data must be copied onto the computer hard drive and will likely be represented by a series generally in some sequential order of similar yet unique file names Each unique name represents a saved recor
23. ace waves over unconsolidated sediments by the MASW method in Proceedings of Symposium on the application of geophysics to engineering and environmental problems SAGEEP 2000 Arlington Va p 1 9 Park C B R D Miller J Xia J A Hunter and J B Harris 1999 Higher mode observation by the MASW method Exp Abs Soc Explor Geophys p 524 527 Bibliography and Recommended Reading on MASW 10 2 Park C B Miller R D and J Xia 1999 Multimodal analysis of high frequency surface waves Proceedings of Symposium on the application of geophysics to engineering and environmental problems SAGEEP 99 Oakland Calif March 14 18 p 115 121 Park C B Miller R D and J Xia 1999 Detection of near surface voids using surface waves Proceedings of Symposium on the application of geophysics to engineering and environmental problems SAGEEP 99 Oakland Calif March 14 18 p 281 286 Park C B Miller R D and Xia J 1998 Imaging dispersion curves of surface waves on multi channel record Exp Abs Soc Explor Geophys 1377 1380 Park C B Miller R D and Xia J 1998 Ground roll as a tool to image near surface anomaly Exp Abs Soc Explor Geophys 874 877 Park C B Miller R D and Xia J 1997 Multi channel analysis of surface waves MASW A summary report of technical aspects experimental results and perspective Kansas Geological Survey Open file Report 97 10 Park C B Miller R D a
24. ault values shown in the controls dialog box represent the most optimum values Figure 2 2 1 Figure 2 2 2 Full Auto Analysis 2 4 2 2 1 Controls Dispersion Figure 2 2 1 Options Frequency Range Algorithm Integrity Curve Smoothing Picking Tolerance Description Frequency range of dispersion curve Section 3 3 2 Phase velocity calculation method Section 3 3 3 Shown only in case of Normal algorithm Section 3 3 3 Degree of dispersion curve smoothing Section 3 3 4 Tolerance in phase velocity calculation Section 3 3 3 2 2 1 Controls Inversion Figure 2 2 2 Options Stopping Criteria Layer Model Description Criteria upon which the iteration of inversion stops Inversion will stop when it reaches either the maximum number Max Iteration or the minimum root mean square RMS error Parameters for the earth layer model Thickness will select the earth model with either equal Equal or depth varying Variable thickness Constant values of Poisson s Ratio and Density will be assigned to all the layers Chapter 3 Dispersion Analysis Full Auto in Chapter 2 provides an effective way to produce one S wave velocity profile from a single multichannel record However it 1s only truly effective when the record has a high signal to noise ratio S N This means that the input data should be as free as possible of body waves and higher mode surface waves A
25. ce waves are weaker relative to near offset traces For this reason the far offset traces need to be examined carefully before analysis Figure 8 5 6 shows a sample data set SanJose DAT with a gain that makes it obvious that the far offset traces are contaminated with body and surface wave noise events When this data set 1s separated into near offset and far offset traces overtone analysis of the near offset traces Figure 8 5 7 reveals a fundamental mode dispersion curve that is better defined especially at low frequencies than that obtained from full offset analysis Figure 8 3 13 The dispersion Working with SurfSeis 8 24 image for the far offset traces Figure 8 5 8 suffers noticeably from the noise problem Even with the noise problem on far offset traces the dispersion image 1s of an even better quality than the image obtained by using full offset traces The two analyzed dispersion curves are displayed in Figure 8 5 9 and show that the near offset curve has a higher S N in general Although higher modes are complicated with dependence on a variety of parameters such as layer geometry and receiver location Tokimatsu et al 1992 they tend to become more significant at further offsets Park et al 1999 This reinforces the need to carefully examine the far offset traces and their effects on the analysis A i a y Hn SP TRUE A Figure 8 5 1 WT TT b HI TIT Pl T I a LI E E ni E LLL LLL ir DI E i z
26. ction and refraction waves acquisition as well This aspect could be beneficial in some situations when both methods surface wave and body wave analysis are required One of the most significant differences in data acquisition procedures with MASW when compared with the conventional body wave survey 1s the enhancement of low frequency energy A sledgehammer is a common seismic source used for a MASW survey although many different types of seismic source can be used see Miller et al 1986 Keiswetter and Steeples 1995 At most of the soil sites a sledgehammer with 10 kg mass usually assures optimum spectral characteristics for a target depth less than 10 m A heavier or lighter one may need to be used to meet the required spectral characteristics as the primary target depth increases or decreases It is also Important to use a low natural frequency geophone for most studies A 4 5 Hz geophone 1s most often recommended SurfSeis is designed to generate a Vs profile using a simple 3 step procedure preparation of a multichannel record sometimes called a shot gather or a field file dispersion curve analysis and inversion The term multichannel record indicates a seismic data set acquired by using a recording instrument with more than one channel and for most cases involving modern seismographs at least 12 channels would be necessary However a multichannel record for use with the MASW method can be recorded in several different
27. curves change somewhat noticeably as the integrity increases The change soon becomes negligible after a certain value e g 3 When data are not noisy a value of 3 1s usually the optimum With noisy data the value needs to be increased proportionally with the noise levels Improvement in the quality of dispersion curves with different integrity values 1s usually guantified by changes in S N curves although changes in the general trend of the dispersion curves should also be considered With good guality data the Approximate algorithm usually produces a dispersion curve comparable to that obtained using the normal algorithm with optimum integrity Figure 8 3 23 Figure 8 3 24 shows dispersion curves obtained using different degrees of Tolerance It demonstrates that the higher mode is picked in the high frequency range when Moderate and Large tolerances are used This 1s due to the larger phase velocity searching range as previously described Once a judgment has been made as to the best curve the curve must be saved before proceeding to the inversion step By checking the processing history History button all the displayed curves can be studied Figure 8 3 25 This step can also assist in Judging which curve should be taken as final representative through examination of a few key processing parameters If only part of the curve is reliable and to be saved the Resample button provides options to edit
28. d For CDP or common offset data each field file 1s generally a unique shot gather The process of file naming 1s usually accomplished during the downloading of data onto the processing computer or at the time the file was collected recorded The conversion programs are within the format header user interface icon To con vert SEG 2 data clicking on that sub icon will be necessary This will open the SEG2 gt KGS conversion window which has a general layout that 1s consistent with operational windows throughout the SurfSeis processing package By highlighting the Files heading input data files can be selected using a standard Windows style file directory server When the complete list of files has been highlighted and the OK button clicked the screen will return to the basic window with all the selected files listed The output file can then be selected again under the Files heading Once the output file name has been assigned both the input and output should be listed in the original window During the conversion process 1t 1s possible to decimate the data resample delete the files as they are converted space saving measure only and or shorten the record length Any of these operations can be selected in Working with SurfSeis 8 5 the Options menu When the input file name and output file name have been selected and any options designated the Convert heading should be clicked on This will begin the conversion process by requesting a source
29. e dispersion curve provided to the inversion step The inversion uses the dispersion curve as the only empirical data with no reference to the original seismic record Confidence in the dispersion curve can be estimated through a measure of signal to noise ratio S N displayed along with the curve The fundamental mode of the surface wave is the signal used for the analysis There are several factors that interfere and disturb the analysis body waves and higher mode surface waves These noise sources can be controlled to a limited extent during data acquisition but never totally eliminated Dominance of these types of noise is common with farther offset distance between source and receiver see Park et al 1999a and 1999b To obtain an accurate dispersion curve it 1s Important to examine without any bias the spectral content and propagation velocity called phase velocity characteristics of both signal and noise waves SurfSeis allows examination of these characteristics through a unique step called overtone analysis Inversion of the calculated dispersion curve 1s performed with SurfSeis in a fully automated manner in which a most probable solution is sought in an iterative mode see Xia et al 1999 Display of data during certain parts of the analysis 1s critical to insuring the optimum solution It can sometimes become a necessary step for the assessment of optimum analysis parameters All the display modules in SurfSeis can be used effectivel
30. eadjustment of Data It may be necessary to preprocess seismic data prior to dispersion curve analysis There are usually two reasons for this First to enhance the signal to noise ratio S N resulting in a broader band dispersion curve increasing investigation depth and confidence and second to help identify more effectively the different types of seismic events within the frequency and phase velocity range of interest This in turn can help optimize estimates of some key processing parameters e g frequency range for the dispersion analysis A strong first arrival event can be troublesome in two different ways First automatic detection can mistake 1t as a surface wave event 1f 1ts energy 1s equal to or greater than the surface wave Figure 8 3 7 For this case a dispersion curve calculated using the default parameters will be erroneous because the dispersive nature of the body wave event in this case a guided wave event will be of equal or greater influence over the surface wave event Secondly the first arrival can alter the phase relationship of surface waves at frequencies dominated by first arrival energy Again for this case the calculated phase velocities will deviate from actual to an extent proportional to the degree the first arrival energy dominates A strong first arrival event can be effectively removed using the Mute process see Section 5 5 Underwater data shown in Figure 8 5 1 have significantly more pron
31. ecting the Initial model Current model iteration Final model and R M S shear velocities Toggle on off to display phase velocities When this button 1s on a separate panel appears Inside the panel are check boxes for selecting curves for the Measured Initial Current iteration and Final phase velocities and the Signal to Noise Ratio S N curve Toggle on off to display P wave velocities based on selected Poisson s ratio L Left Click R Right Click A Action Button D Dialog Button 7 3 Poisson s Ratio Pos and Density Figure 7 2 Petes Pali F Piotta Roi and Donate en Wr a pd Dwi E AL F Li om E Carmel E Frei n i Display of Dispersion Curve 7 4 Mouse Button Button Name L R Description Toggle on off to display Poisson s ratio for the layer model When this button is on a separate panel Pos A appears Inside the panel are check boxes for select ing the Initial Current iteration and Final Poisson s ratios Density A Toggle to display density curve based on user selection L Left Click R Right Click A Action Button D Dialog Button 7 3 1 Animated Zoom By pointing dragging and drawing a rectangular area see Figure 7 3 which includes part of the displayed curves a selected zone can be zoomed in on using animation mode The original display can be restored by double clicking any place within the curve display a
32. ection of a constant background color Display of Dispersion Curve 6 4 Figure 6 2 4 Chapter 7 Inversion Results Display Inversion results are saved into two different files One file IVO contains the final results of inversion The other file LST contains the final as well as intermediate iteration results The module displaying the entire inversion process 1teration by 1teration through all the defined layer parameters can be instrumental in confidence determinations Included with this module are a Theoretical layer model display at each iteration a Theoretical dispersion curve display corresponding to the layer model a Printing and saving BMP displayed image Display of Dispersion Curve 7 2 Figure 7 1 7 1 Main Mouse Button Name L R Description Saves displayed image as a bitmap file Save A BMP l Sends displayed graph to printer according to Prnt A printer setup t Scale A A Changes HM scale of displayed graph L increase R decrease Ite A Changes which data iteration is displayed L Left Click R Right Click A Action Button D Dialog Button 7 2 Velocities Mouse Button Button Name L R Sene a Vs A Phe velocity Phs A PAN ENE P A Display of Dispersion Curve 7 3 Description Toggle on off to display S wave velocities with layer model When this button is toggled to the on position a separate panel appears Inside this panel are check boxes for sel
33. elect the Overtone button after Preprocessing overtone analysis will be initiated with the default parameters defined for frequency and phase velocity range These default parameters are chosen such that the dispersive nature of surface waves can be most effectively delineated whereas body wave events are excluded from the image The default values are much broader than values determined automatically during preprocessing The reason for this increased range is to assure a wider coverage during pattern recognition For that reason resultant images usually contain the remnants of noise events as well as signal patterns Figure 8 3 12 Once this initial image has been examined and evaluated another overtone analysis can be run with ranges and parameters more focused on the fundamental mode energy only Figure 8 3 13 The image in Figure 8 3 13 was obtained by using a narrower range of phase velocity and wider range of freguencies than the image in Figure 8 3 12 To achieve a higher resolution image a smaller increment 5 m sec of phase velocity was used than in the initial analysis run 15 m sec Body wave events can be identified by enlarging the frequency and phase velocity range beyond that optimum for surface waves during overtone analysis Usually body wave events will have freguencies and phase velocities several times higher than surface waves Working with SurfSeis 8 11 For example the direct refraction and reflection ev
34. elected a dialog box appears allowing selection of dispersion curve file s A single file or multiple files to be processed in succession can be defined Once an input file 1s selected an initial model is created using phase velocities in the input file and the preset parameters The initial S velocity vs model is displayed along with the input dispersion curve Figure 4 1 1 The dispersion curve is also displayed but with an inverted frequency axis so that general trend in the S velocity vs model and dispersion curve match in an approximate sense The P velocity v model and other parameters e g density can be displayed using appropriate buttons and tabs A detailed explanation of the display features can be found in Chapter 7 Actual parameters values can be viewed and changed using the Layer Model option Section 4 3 Buttons available at this stage of the process include Run Controls Layer Model and Back Figure 4 1 1 Figure 4 1 2 4 2 Controls Inversion Analysis 4 3 The Controls operation allows changes to be made in the iteration stopping criteria type of initial S velocity vs model and weighting of phase velocities Figure 4 2 1 Stopping Criteria The inversion 1s halted when one of the criteria 1s met the root mean square error RMSE or maximum number of iteration Imax The RMSE default setting 1s 3 0 but the optimum value may change based on the input dispersion curve A more de
35. ents at most soil sites usually possess frequencies that range across a couple hundred Hz e g 50 300 Hz An air wave event usually has frequencies even higher than direct refracted or reflected energy e g 100 500 Hz Phase velocities horizontal propagation speed of body waves are generally in the range from 300 to 5000 m sec They are however usually non dispersive with just a slight dispersive character observable on some refraction events Clayton and McMechan 1981 Back and side scattered body waves can also appear dispersive when they have more than one azimuth angle with respect to the survey line Park et al 2000 There are two types of body wave events generally evident on shot records recorded to optimize surface wave energy air wave and the first arrival refraction or direct events Since alr wave events can possess a significant amount of energy and a phase velocity 343 m sec in 20 C close to or sometimes within the phase velocity of many surface waves phase velocity calculations of surface wave energy can be unreliable unless the air wave event is identified its influence assessed and removed if possible When the frequency range for the surface waves being analyzed 1s lower than about 100 Hz the air wave event will not likely affect the surface wave analysis because air waves generally have freguencies higher than 100 Hz On the other hand refraction events usually have velocities many times say five times hig
36. ep will usually come up with optimum values However occasionally the highest frequency selected may extend beyond the overtone frequencies For this case the value determined automatically needs to be manually changed to a lower value determined using the overtone image On the other hand the lowest frequency automatically selected may be so low it includes frequencies in which the phase velocity becomes either non dispersive or unstable e g 5 10 Hz range in Figure 8 3 13 The nondispersive range 1s represented by a flat curve An unstable range 1s represented on the dispersion curve by an irregular image This usually occurs within the frequency range suffering from near field effects Stokoe et al 1994 In other words the surface waves in this frequency range were not developed sufficiently to become well behaved plane wave events This usually indicates longer offsets between source and receivers need to be used during acquisition to avoid this effect From the overtone image in Figure 8 3 13 the optimum frequency range would be selected to be 11 21 Hz After properly setting the frequency range the phase velocity needs to be checked at a reference frequency Figure 8 3 5 Automatic selection usually assigns values to these parameters within 20 deviation However a deviation as much as 100 may be acceptable as long as there 1s no significant higher mode energy observed at the reference frequency Figure 8 3 13 The
37. ers that define the frequency range for Parameters filter i e band reject filtering The shape of the filter operator is a Hz solid graph on the displayed spectra Scale Changes horizontal scale of displayed spectra left click increase right click decrease Print Prints displayed spectra Save Saves displayed spectra image BMP Seismic Data Display 5 11 5 5 2 Process Mute Controls Mute Parameters Option Description Defines the portion above Top or Mute Type Mute A a Liga Top Bottom Type below Bottom the drawn straight line that is to be zeroed Tapering Determines degree of smoothing along Degree the edge of cutting boundary Weak tapering will result in a relatively abrupt edge while strong tapering results in a very smooth ramp Actual length of TU i E smoothing window can be manually specified C Weak bid aa aa Tapering Teper Widow mal x Cericel 5 5 3 Process AGC Controls AGL Paramotora Option Description LEE I Defines the length of the automatic gain C Weak Mild Strong control AGC scaling window The window length under the Mild option AGC AGC Window ms Degree is equal to 20 percent of displayed time 2 axis Actual length of the window can be manually specified by changing the value in the spin edit Amplitude Spectra Parameters NEM Option Description
38. erspective If this RMSE is greater than the minimum RMSE Emin specified in the Controls dialog Figure 4 2 1 the inversion algorithm will automatically modify the Vs profile see Kia et al 1999 and repeat the procedure by calculating a new theoretical curve Each round of this searching procedure is called an iteration and iterations continue until either Emin or maximum number of iterations max is reached The theoretical dispersion curve Current that is being compared with the experi mental curve Measured and its associated Vs profile being compared with the initial profile Initial are displayed for each iteration Figure 4 4 1 The changing trend of RMSE is displayed in the upper righthand corner of the display If the experimental curve is reasonably accurate or makes sense from a theoretical point of view the RMSE usually drops dramatically during the first several iterations Sometimes later iterations may reach a level beyond which no noticeable change in RMSE can be observed see Section 8 4 In this case the iterations can be stopped manually by clicking the Stop Ite button in lower right hand corner of the display rat ae iia SA pi LU ddau Tm Ma TI Figure 4 4 1 E m ya a mala ann 1 Ca t Se herw Depas 2 kamaa Kingi Mba 7 Ferniorm Chapter Ss Seismic Data Display SDD Multichannel seismic data collected in the field are
39. et clicking the Run button starts the calculation of the phase velocities followed by the display of one dispersion curve Based on the displayed curve guality a few trials may be necessary with different parameter settings Figure 8 3 22 Process No 1 in Figure 8 3 22 has been obtained by using Approximate algorithm Process No 2 by using Normal algorithm with integrity 1 and Process No 3 by using integrity 3 For demonstration purposes the highest analysis frequency has been mcreased to 35 Hz from the default selection of 17 5 Hz A slight difference can be observed between all three curves in the lower half of the analyzed frequency range The most prominent difference occurs in frequencies higher than about 23 Hz The quality of a dispersion curve 1s judged according to two criteria signal to noise ratio S N and general dispersion curve trend A high S N indicates high confidence in the obtained phase velocity S N higher than 0 5 1s usually acceptable The S N curve will in general show a roughly down going trend This means that S N will generally decrease with frequency which 1s due to the relative increase in noise events at the higher frequencies Park et al 1999 This down going trend in S N should be examined in much a broader perspective than possible with dispersion curves done The trend of a dispersion curve 1s generally down going phase velocity as frequency increases Unless
40. f the display A prolonged inversion usually occurs when the measured dispersion curve trend does not make sense from theoretical perspective Xia et al 1999 either in whole or in part This type of dispersion curve usually has an abnormal trend where phase velocity increases instead of decreases with frequency across a significant portion of the curve or the curve flattens at low frequencies The dispersion curve shown in Figure 8 4 3 represents an example of this abnormal trend at both high and low freguencies The inversion process did not drop below the specified Emin 3 0 even after the specified max 30 was reached as a result of this trend Whenever the inversion 1s prolonged calculated Vs profiles usually Working with SurfSeis 8 20 contain one or more velocity reversals These reversals are noticeable in this example at depths of about 3 m and deeper than 10 m Figure 8 4 3 If the reversals are a result of prolonged inversion they should be considered suspicious and could be numerically exaggerated In Figure 8 4 4 the inversion of a dispersion curve with a reverse dispersion trend phase velocity increases with frequency in a consistent manner is typical for seismic data acquired over pavement therefore nothing is wrong with the curve The inversion analysis still had a large RMSE even after 30 iterations due to this abnormal trend This type of reverse dispersion curve needs a separate inversion alg
41. h rn CE BR rir uri lr Freire p Ph A BI Gala Te mx Fleer a herre Uc a Working with SurfSeis 8 18 Figure 8 3 23 Figure 8 3 24 Figure 8 3 25 Working with SurfSeis 8 19 8 4 Inversion Analysis Once a final dispersion curve has been selected and saved as a DC file the program will prompt to proceed to either Inversion analysis for the saved curve or Dispersion analysis to move on to the next seismic record depending of course on whether the input seismic data contains only one or multiple records For Dispersion analysis the analysis for each record would be similar to the previously described procedure As well with the Dispersion analysis option the previous dispersion curve can be used as a Pilot curve Moving one step further the Pilot Aided analysis type in the Controls dialog box see Section 3 3 prompts automatic analysis for all remaining records in input file Figure 8 4 1 displays the beginning stage of the inversion analysis Two other button options are available besides the Run button Controls and Layer Model The Controls button enables the changing of some parameters that relate to the inversion process e g stopping criteria see Section 4 2 The Layer Model button makes 1t possible to change individual items in the initial earth layer model see Section 4 3 The Run button begins the inversion process searching for a Vs
42. have a low S N 3 You have a good understanding of the processing parameters to use regardless of the S N This means that shot records are obtained along the same linear survey line moving source and receivers in a regular fashion so as to maintain a consistent source to receiver offset and spread geometry e g a roll along mode in CDP survey Pilot Aided Analysis This analysis type uses a reference dispersion curve usually obtained from a previous process In contrast to normal analysis pilot aided analysis has the greatest bias during the Dispersion Analysis 3 5 calculation of phase velocities This type 1s recommended for any of the following situations 1 Your input data set consists of multiple records collected in a continuous or roll along mode and at least one of those records has been processed to calculate the dispersion curve using Normal analysis as described above 2 A reference dispersion curve exists that is known to be typical from the area your current data were collected If this type of analysis is chosen SurfSeis will prompt you to choose a file with a reference pilot dispersion curve 3 3 2 Controls Parameters Options Description Frequency Frequency range and Range dispersion curve increment Apparent Phase Upper and lower bounds on Velocity phase velocity which will be used by the phase velocity calculation as an approximate reference Values assigned immed
43. her than surface waves in the same setting This large difference 1s in part due to a much greater penetration depth of refracted energy compared to surface waves at the same source In the case of direct wave events however the velocity will generally be four or five times higher than surface wave velocities since the direct wave 1s sampling the same overburden material the surface wave is Figures 8 3 14 and 8 3 15 show overtone images obtained using frequency and phase velocity ranges appropriate for body waves It is sometimes important that these ranges overlap those of surface waves so that both types of waves can be examined on the same display in a comparative manner Overtone analysis shown in Figure 8 3 14 focuses on examination of an air wave event whereas the one in Figure 8 3 15 targets the first arrival events A frequency range of 5 250 Hz with 1 Hz increment and a phase velocity range of 50 450 m sec with 5 m sec increment are usually optional for air wave analysis The phase velocity range 1s enlarged to 50 6000 m sec with 50 m sec increment for first arrival analysis In Figure 8 3 14 the normal air wave event 1s clearly non dispersive with about 345 m sec velocity at freguencies higher than 100 Hz Dispersive air wave events that were either side or back scattered or combination of both from surface objects near the survey line are also identified It 1s apparent that air wave events do not interfere with surface wave analy
44. iately following Preprocessing which have been determined by the program to be the optimum values 3 3 3 Controls Computation Reference A reference phase velocity 1s necessary and represents a starting point for the program to calculate phase velocities for all frequencies within the Frequency Range as specified in parameters discussed in Section 3 3 2 The phase velocity will be calculated at this particular frequency using the specified phase velocity as a starting point to begin the downward progression through each frequency At each new frequency the calculation will begin again using the phase velocity calculated from the previous frequency as the reference The program Diso mywn Tp Pares Camputaton Sua anal zznt arm meme Hanga e Lowest PAO Highest 775 Inorement 10 50 Apparent Phage Valocity Lowest M Highest pas i kima 3 Figure 3 3 2 Tolerance mall r Moderar Reference Phase Veloso het Eeer Hz u Ju Algorithen sa dicis 10 AMonmal A FS m l mr DEC Frezza Figure 3 3 3 Dispersion Analysis 3 6 will continue doing this until it reaches the lowest frequency At that point it will move on to the frequency immediately higher than the reference frequency The reference phase velocity and frequency displayed in this panel of the dialog box Figure 3 3 3 are the values the program determined automat
45. ically and considers to be most reliable references with which to begin the calculations Tolerance This operation establishes the range of phase velocities to examine when calculating the phase velocity at a specific frequency Intuitively Small sets a narrow range and Large represent the biggest range possible If a certain portion of the calculated dispersion curve appears to be influenced by the inclusion of higher modes the Small option will provide the most desirable results Algorithm Two types of dispersion analysis algorithms are available Normal and Approxi mate Fundamental principles of these two algorithms can be found in Park et al 1996 Normal is most accurate and is used whenever the input record has a high S N or all the key parameters e g reference phase velocity and frequency frequency range and incre ment etc are properly set either automatically or through your own determination When this option is selected another subparameter called Integrity 1s initiated Integrity con trols how intensely the program tries to establish the most probable value of phase velocity A value of 3 1s a good choice for most situations Increasing it by one usually doubles the computation time Approximate is best when S N is not as good or you are not confident in or unable to select the optimum parameters The advantage of this algorithm 1s that any perturbations i
46. l dialog box Figure 8 3 19 provides the opportunity to select a couple of slightly different algorithms for overtone and dispersion curve analyses Selecting the Approximate algorithm option will usually provide a good preliminary representation of overtone images This is an especially effective option when working with noisy data This option can effectively delineate higher modes when more than one higher mode exists within a given freguency range during the overtone analysis Figure 8 3 20 The Normal algorithm option on the other hand focuses on the mode with the highest energy levels Figure 8 3 21 Because of this approach the Normal option will usually produce an overtone 1mage with greater resolution than the Approximate option In Figure 8 3 21 the fundamental mode clearly dominates energy in the 10 20 Hz range where the first higher mode dominates energy 1n 20 40 Hz range Once the optimum frequency range for the fundamental mode has been identified 1t is always best to use the Normal algorithm The Integrity option in normal algorithm controls how intensively the phase velocity search is undertaken at each calculation fre guency Optimum integrity can be obtained by successively increasing the integrity value and examining the phase velocity change and 1ts effect on the calculated dispersion curve An Integrity value of 3 1s generally optimum Run After all the parameters have been properly s
47. l with Integrity 3 Figure 3 4 4 demonstrate the functional range of these options The seismic data input into these overtone analyses Figure 3 2 1 were obtained by using 4 5 Hz geophones From these examples it 1s evident that the optimum frequency range 1s around 5 20 Hz This range 1s estimated by noting that the higher mode possibly the first overtone starts to dominate at frequencies higher than 20 Hz with the low 5 Hz of the optimum frequency range chosen consistent with the geophone natural frequency This range can be wider than suggested here depending on the closeness of source to the recelvers on near surface materials and frequency characteristics of the source Dispersion Analysis 3 9 O O S3095 ws E a lt Y mp o DU e E E ES 62252 f ESS o er e O Un el e RES o Sao 8S gt 5 aor ep 5 PE UN E 8523 ao E ts ao AE Eeo SE ac 3 EX 9 BESEBS 5 2 o 2 en 83825 S gt o bb Y O sa og o 823 rH ESOOcC 2 Ae n 25288 ven JD pumi YS 3 bb t zx CE qu Ss POY Y 5023 aeb nai Oor AGE QoS Sous 2 BEERS Uo amp a O R 50055 e c amp A er ES o ood O 828 Ti Cc So i 355 5 EL cos amp 2 lt A Ou 82 oe 2 cn Azo Coda mn nO o o PES D 5 oO Seog Dispersion Analysis 3 10 3 6 Resample This option makes it possible to truncate Fon ing DRE DEA DITAR fae curve s that seem unreliable
48. lso the full auto option assumes that the main purpose of the analysis 1s to produce an S wave profile from a single input record Although the inclusion of strong body waves 1s usually obvious when a record is displayed Chapter 5 that 1s not true of higher mode surface waves Chapter 8 In addition It 1s sometimes necessary to process multiple records to produce multiple Vs or dispersion profiles from the same processing flow to allow comparative examination of processing parameters relative to previous processing results The Dispersion option in the analysis section gives complete flexibility to determine key processing parameters A complete range of options are available between full automatic and full manual modes It also provides a way to construct dispersion curve Images for all types of dispersive waves regardless of whether they are body or surface waves Contamination by higher mode surface waves and body waves can be clearly examined using these Images Following is a summary of the features of this option a Control of key processing parameters e g frequency range increment etc a Construction of dispersion curve images through the wavefield transformation method a Dispersion analysis of multi record input data Dispersion Analysis 3 2 3 1 Main When the Dispersion button 1s clicked a dialog box Figure 3 1 1 pops up that allows you to select a file containing multichannel seismic record s The file
49. lt in a change in Poisson s ratio S Figure 4 3 1 A change in density R will not result in change in any other parameters Others The total number of layers in a model can be changed using the up or down arrow in the of Layers edit box Currently the maximum number is set to 20 layers The Thick ness Model option enables a choice of Egual and Variable models The Egual option generates a model consisting of all layers with egual thickness and the Variable option generates a model with layers whose thickness increases with depth The Velocity Incre ment option defines the amount values in the edit boxes of Vs and Vp are changed when up or down arrows are clicked Import LYR and Save LYR options import previously saved files containing all previously worked layer parameter values and save the current set of values respectively Inversion Analysis 4 5 44 Run Run starts the program searching for a Vs profile whose theoretical dispersion curve matches the experimental dispersion curve obtained from the Dispersion analysis The match will be evaluated on the root mean square error RMSE between the two curves Xia et al 1999 The inversion algorithm first calculates the theoretical curve using the initial Vs profile along with other layer parameters previously explained then compares the theoretical curve with the experimental curve from the RMSE p
50. may contain one or more records and you can process one record at a time or sequentially as a group Once an input seismic file is selected and displayed the first record in the file is ready to be processed If based on the data displayed this 1s not the record you want to process you can choose another record by clicking the Next Record or Jump button Chapter 5 in the display Figure 3 1 2 shows a panel of buttons displayed on left side of the screen as a tool bar at this stage Only buttons that can be executed currently are enabled and all others are disabled The set of both enabled and disabled buttons will change according to the stage of the analysis Only two buttons excluding the Back button are enabled Preprocess and Controls buttons pe Pie Tru Sali chase Oria DATE Ch AG ico 101 aet Figure 3 1 1 Ya im Save Tas Figure 3 1 2 versions Dispersion Analysis 3 3 3 2 Preprocess When this button is clicked the program tries to estimate the following two data characteristics 1 Optimum range and increment of frequency 2 Optimum upper and lower bounds of phase velocity These operations are performed first by examining frequency spectra of all the constituent traces and then by examining energy distribution in time of each trace The phase velocity bounds are calculated through a best fit method applied to the trend of the 2 D energy distributi
51. n Inversion Analysis 4 4 4 3 Layer Model Lia The initial S velocity vs model is approximated from the measured dispersion curve The initial P velocity v model is defined from this v model by assuming a constant Poisson s ratio of 0 4 A default density value of 2 0 g cc is assigned into all layers default is 10 layers in the earth model The maximum depth Zmax of investigation is assigned based on the longest surface wave wavelength Amar measured from the input dispersion curve A proper thickness model is then created within this depth range with the last layer assumed to be the half space From the Layer Model option a manual change can be made to any of the default values Figure 4 3 1 Parameters A Blob FA Sawa LYFR n B Tack haa Fla der Six kinds of parameters are displayed and can be ia ot Eu manually changed depth Z thickness H S velocity wala remem fT a LE Mae Vs P velocity Vp Poisson s ratio S and density R Depth Z represents the depth from the ground surface to the bottom of the associated layer Changing any depth Z will result in changes in thickness H Since the deepest depth Zmax 1s constrained by the long est surface wave wavelength Amar measured it is not allowed to change A change in S velocity Vs will result in a conse quent change in P velocity Vp in accord with the value of Poisson s ratio S Changes in P velocity Vp will resu
52. ncluding noise introduced during the acguisition of surface wave data are usually averaged out and will not cause the whole analysis to fail The disadvantage to this approach is that the dispersion curve will be an average curve possibly lacking the integrity possible with high S N and good estimations of the optimum parameters 3 3 4 Controls Curve Smoothing Once all the phase velocities in the specified frequency range are calculated the program can apply variable degrees of smoothing to the curve Since the calculated phase velocities usually contain a certain degree of perturbation caused by numerical truncation random noise field and equipment noise etc this smoothing will usually average out the perturbation effect A smoothed curve will usually converge better to a final solution during the 1terative inversion process see Chapter 4 Processing History Keeping Dispersion Analysis 3 7 Dispara Pesa Fon Analysis Typo Baramatara Computation Casa j Curve 3 monty 7 Mans 7 Minar C Medium Lage Processing Mistery Respira Seat C Intermediate Detaled Kiwa vw Figure 3 3 4 This option sets the degree of detail retained in the processing history file The processing history file will be appended when the dispersion curve file DC is output This file can be viewed using a text editor Section 6 1 when the dispersion curve 1s displayed Dispersion Analysis 3 8 3 4 Overtone
53. nd Xia J 1996 Multi channel analysis of surface waves using Vibroseis MASWV Exp Abs Soc Explor Geophys 68 71 Pullan S E and J A Hunter 1990 Delineation of buried bedrock valleys using the optimum offset shallow seismic reflection technique Society of Exploration Geophysicists Investigations in Geophysics no 5 S H Ward ed Volume 3 Geotechnical p 75 87 Richart F E Hall J R and Woods R D 1970 Vibrations of soils and foundations Prentice Hall Inc New Jersey 414 pp Rix G J and Leipski E A 1991 Accuracy and resolution of surface wave inversion in Geotechnical special publication no 29 Recent advances in instrumentation data acquisition and testing in soil dynamics edited by S K Bhatia and G W Blaney American Society of Civil Engineers p 17 32 Sheu J C Stokoe II K H and Roesset J M 1988 Effect of reflected waves in SASW testing of pave ments Transportation Research Record No 1196 p 51 61 Stokoe II K H Wright G W James A B and Jose M R 1994 Characterization of geotechnical sites by SASW method in Geophysical characterization of sites ISSMFE Technical Committee 10 edited by R D Woods Oxford Publishers New Delhi Tokimatsu K Tamura S and Kojima H 1992 Effects of multiple modes on Rayleigh wave dispersion characteristics Journal of Geotechnical Engineering American Society of Civil Engineering v 118 n 10 p 1529 1543 Vardoulakis I and Vrettos
54. olution continuously converging in an iterative fashion on the most probable values The S velocity vs is most sensitive and influential to the surface wave phase velocity Influence of all other types of parameters can usually be neglected as long as they have been reasonably estimated The initial S velocity v model is approximated from the measured dispersion curve The initial P velocity v model is determined using this v model and a constant Poisson s ratio of 0 4 A density of 2 0 g cc 1s assigned to all layers of the earth model The maximum depth of investigation is determined from the longest surface wave wavelength measured from the dispersion curve The thickness or layer model 1s then created by successively increasing the thickness of each layer as its depth increases to the maximum depth of investigation A ten layer model is initially assigned The iterative inversion procedure can continue uninterrupted unless a stopping criterion is imposed Two types of criteria are used maximum number of iteration Imax and maximum root mean squre error RMSE That is the inversion process will stop when either of the two is met The following summarizes the features of this module a Invert for S velocity vs profile with all parameters automatically determined default a Change of the stopping criterion Q Change of the number of layers Inversion Analysis 4 2 4 1 Main When the Inversion button 1s s
55. omputation tab of the Control dialog box Figure 8 3 5 This range of phase velocities 1s called the Apparent Phase Velocity in the parameter tab Figure 8 3 6 This beginning information 1s also critical for the program to accurately track changing trends in the phase velocity at other frequencies If data are of good quality automatic detection of the phase velocity range 1s usually reasonably accurate Figure 8 3 8 If the data have a poor S N this automatic detection can result in erroneous findings Figure 8 3 9 In Figure 8 3 9 a strong body wave event a guided wave within a water layer 1s identified incorrectly as a surface wave event This data set needs to undergo preliminary processing to insure an accurate analysis see Section 8 5 Figure 8 3 10 illustrates how an apparent phase velocity is calculated from a specific linear trend of arrivals As long as the program properly identifies surface wave events the analysis will go smoothly by doing nothing more than selecting the Run command Curves resulting from fully automatic analysis should be taken as preliminary with the potential that they may have been adversely affected by noise events sometimes difficult to detect on raw data Frequency Range Proper selection of the frequency range to analyze 1s also critical Since different types of seismic events e g body wave fundamental and higher mode events usually have characteristic frequency bands imprope
56. on in the time domain These bounds are marked on the displayed record see Figure 3 2 1 The phase velocities in the analyzed dispersion curve will usually fall into this range If the marked zone encompasses most of the surface wavefields the process will usually proceed effectively using these automatically determined parameters This means you can proceed to the next stage of processing Run Otherwise based on your judgment you need to check the parameters and change them appropriately by clicking the Controls button discussed in Section 3 3 By right clicking the Controls button a dialog box will pop up see Figure 3 2 2 showing a couple of control parameters You can instruct the program to either broaden Wide or shorten Narrow the frequency range You can also select the offset range to be used for the automatic determination Near Middle Far and Full options will respectively use only those half traces nearest to source in the middle of the receiver spread furthest from source and all the traces This control should be set whenever traces in specific offset ranges have questionable surface wave quality Figure 3 2 1 Figure 3 2 2 F3 Pre Process Parameters Analyzing Frequency Rang Narow Moderate Wide Offset Range for Auto Detection Near Middle Fa Full x Ces Dispersion Analysis 3 4 3 3 Controls This option enables you
57. ophysics to Engineering and Environmental Problems SAGEEP 2000 11 19 Ivanov J Park C B Miller R D Xia J Hunter J Good L and Burns R 2000 Joint analysis of surface wave and refraction events from river bottom sediments Exp Abs Soc Explor Geophys p 1307 1310 Mari J L 1984 Estimation of static corrections for shear wave profiling using the dispersion properties of Love waves Geophysics v 49 p 1169 1179 McMechan G A and Yedlin M J 1981 Analysis of dispersive waves by wave field transformation Geophysics v 46 p 869 874 Miller R D Xia J Park C B and Ivanov J M 1999 Multichannel analysis of surfaces waves to map bedrock Leading Edge v 18 n 12 Miller R D J Xia C B Park and J Ivanov 1999 Using MASW to map bedrock in Olathe Kansas Exp Abs Soc Explor Geophys p 433 436 Miller R D J Xia C B Park J C Davis W T Shefchik and L Moore 1999 Seismic techniques to delineate dissolution features in the upper 1000 ft at a power plant site Exp Abs Soc Explor Geophys p 492 495 Miller R D J Xia C B Park and J M Ivanov 1999 Multichannel analysis of surfaces waves to map bedrock Leading Edge v 18 n 12 Miller R D J Xia C B Park and J Ivanov 1999 Shear wave velocity field to detect anomalies in the subsurface Abs Proceedings of the Int l Conference on Applications of Geophysical Technologies Dec 11 15 2000 St Louis Miss
58. orithm for a proper handling of layer parameters Nazarian et al 1982 This algorithm 1s not part of this version of SurfSeis however it will be part of SurfSeis 2 soon to be released A prolonged inversion can also occur when the stopping RMSE Emin is set unrealistically low Figure 8 4 5 Although the input dispersion curve possesses good quality the inversion was prolonged because the Emin was too low 0 3 Therefore the observed velocity reversal 2 5 m deep 1s likely more exaggerated than actual The RMSE of each unique layer in the final Vs profile can be viewed by selecting Final R M S Figures 8 4 6 and 8 4 7 This RMSE is a measure of relative error for each layer in comparison to theoretical criteria Kia et al 1999 and can be used as a measure of confidence A prolonged inversion will usually result in a profile with a large RMSE 1 e small confidence for most layers Figure 8 4 6 Small RMSE values i e big confidence should be expected for a normal inversion where the process 1s stopped after several iterations Figure 8 4 7 Figure 8 4 1 EE ET A u i Pa nam T nop 2 jg ym Working with SurfSeis 8 21 Figure 4 2 Figure 8 4 3 Figure 8 4 4 Working with SurfSeis 8 22 Figure 8 4 5 pa ell ml Ci f is m EUN Figure 8 4 6 LE a i SOT Ji Figure 8 4 7 Working with SurfSeis 8 23 8 5 Pr
59. ounced surface wave in this case the Scholte wave energy after the high amplitude body wave events guided and refraction events are muted Figure 8 5 2 With this proper preprocessing of the data the overtone image Figure 8 5 4 of the fundamental mode 1s much better defined across a broader bandwidth than the raw data without editing Figure 8 5 3 The dispersion curve has in general a higher S N and broader frequency range Figure 8 5 5 Another problem source in dispersion curve analysis 1s scattered energy or energy arriving from a source outside the 2 D plane of the survey Scattered energy can come from any place in the survey area Off line anomalies e g voids and building foundations can produce coherent energy with an apex related to the anomaly location Anomalies in line can cause back scattered arrivals which are the sound equivalent of waves on a pond when they bounce off a fixed object in the water Depending on the relative energy of these scattered wave events the dispersion relationship of surface waves can be significantly altered Some times they create a separate dispersion curve usually reverse dispersion on the overtone Image e g the back scattered air wave event shown in Figure 8 3 14 The scattered wave noise problem 1s usually offset dependent such that the noise tends to represent a large portion of the recorded energy at further offset traces due to attenuation where the source generated surfa
60. ouri Miller R D J Xia and C B Park 1999 MASW to investigate subsidence in the Tampa Florida area Kansas Geological Survey Open file Report 99 33 Miller R D J Xia C Park J Ivanov N Geier and D Laflen 1999 Using MASW to map bedrock in Olathe Kansas Kansas Geological Survey Open file Report 99 9 Moore R C 1964 Paleoecological aspects of Kansas Pennsylvanian and Permian Cyclothems Kansas Geological Bull 169 v 1 p 287 380 Nazarian S Stokoe II K H and Hudson W R 1983 Use of spectral analysis of surface waves method for determination of moduli and thicknesses of pavement systems Transportation Research Record No 930 p 38 45 Nazarian S 1984 In situ determination of elastic moduli of soil deposits and pavement systems by spectral analysis of surface waves method Ph D Dissertation The University of Texas at Austin Park C B Miller R D and Xia J 1999 Multichannel analysis of surface waves Geophysics v 64 n 3 p 800 808 Park C B Xia J and Miller R D 1998 Surface waves as a tool to image near surface anomaly 68th Ann Internat Mtg Soc Expl Geophys Expanded Abstracts p 874 877 Park C B Miller R D Xia J Ivanov J Hunter J Good L and Burns R 2000 Multichannel analysis of underwater surface waves near Vancouver B C Canada Exp Abs Soc Explor Geophys p 1303 1306 Park C B R D Miller and J Xia 2000 Detection of higher mode surf
61. oving vehicles and 1s therefore considered noise This data set cannot be processed properly due to a complete lack of any identifiable surface wave signal Some adjustments or enhancements that can help improve the quality of bad data are explained in Section 8 5 Working with SurfSeis 8 7 Sometimes even a data set that appears good during acquisition may need to go through an adjustment based on the results of a more careful examination post acquisition MIN E RECORD SG HN rre B i pos dj el E E iaa a A dd a Si Anf M i Mat i WM A ID As ra 4D UN A cre i eat E Lui LEE P l zi RE i L AF eel EN abl Een bn d iti a ZJE E hu a Vn dms E m E y l u N e n AT P ali i a i i 1 i L pee pr TL mn I EL E i EJE da CU i Tite NERIS B ja E ih i Nem inane ol Ten aris y TD a Jens atk Figure 8 3 3 Figure 8 3 4 Working with SurfSeis 8 8 Phase Velocity Range A good estimate of an approximate phase velocity Vphs range for the surface wave 1s critical Vphs will generally range from as low as 10 m sec to as high as 3000 m sec depending on material type This information 1s used by the program to initiate the analysis by searching within this range for a phase velocity corresponding to a particular surface wave frequency with the greatest coherency throughout the entire range of offset and highest S N This frequency 1s called the Reference Frequency in the c
62. pn o TL E TE RT FELT Lm mum Teri an im ama a es a I aa a a F J I ru e iom Ee dw off mi n aim mm Rm mum I pm um P BNET em Lei Ri un Om i en Ta n rr Hn n yg J no om j m um m ma mmc a Seismic Data Display 5 9 Figure 5 3 TI A A HYF bom jS et uil jl In im M un v i _ mami a RHY UM i en nr Visti m Hi ul IN w 7 ihe TES i iib ULM da r LI A um gr i x I E DR AA O A II i x 4 a a m T al E IL 3 a H E NENENES E z d A n ri E cre z Figure 5 5 dE A apt Pees n Figure 5 4 Seismic Data Display 5 10 5 5 1 Process Filter Controls TI WI Fuld 5 Waak lll Strang AMPLITUDE SPECTRA Amplitude Spectra Display i On meza grad Dialog 1 T 5 10 18 21 28 0 GO 44 BD Frequency P ius Sri Cut Fates Parameters Hz annoiato Xn Pif 3f ap d Dialog 2 X Cancel Dialog 1 Option Description Filter Controls degree of filtering Weak filtering indicates the narrowest Degree bandwidth filter centered on the noise whereas strong signifies the widest bandwidth Amplitude Allows Amplitude Spectra Display dialog box to pop up before Spectra actual filtering takes place to allow the user to change filter Display parameters based upon displayed spectra Dialog 2 Option Description Filter Specifies four filter paramet
63. r annotating trace number Mere ddi Header Options for annotating specified header Cap tO TE Word word data sorted ascending to Site bina TE PEREZ Fre Identifies how often the specified labels Ewmy lin oca RECORD 10 2n ms x quency appear along the top of the image To Dee int Title Title of the displayed image z gt gt m Separation between major solid line and Mm Time l l i CS Trem Dus minor dotted line time lines Title of the time axis Seismic Data Display 5 6 5 4 Process Tool Bar se we s m ec Note A more detailed description 1s listed in a separate section titled Processing With Multichannel Seismic Data Mouse Button Button Name L R Description m L double click starts filtering process E Filter A D R displays dialog box with parameters Section 5 5 1 juta L double click starts muting process Mute A D R displays dialog box with parameters Section 5 5 2 L double click starts AGC automatic gain control SIC AGC A D process R displays dialog box with parameters Section 5 5 3 ET L double click starts spectral analysis process Sperm A D R displays dialog box with parameters Section 5 5 4 Toggle between original and processed records Enabled E Raw A only after processing has been applied to an original record Toggle between current and last processed records E Proc A Enabled only after at least two processing s
64. r selection of this range can result in the analysis of noise events rather than the fundamental mode surface wave event An approximate range of surface waves both fundamental and higher modes 1s usually accurately detected automatically once Preprocess has been appropriately applied Figure 8 3 7 Automatic detection can also be controlled or limited by right clicking the Preprocess button see Section 3 2 Analysis of the amplitude spectra 1s another way of estimating this frequency range Figure 8 3 11 When surface wave events dominate the record the amplitude spectra will predominantly show the offset varying spectral characteristics of surface waves The largest amplitude arrival bands on the traces as illustrated in Figure 8 11 define the optimum frequency range EEC ord ii e iu dl HT ES TEE Teic it ti A A A YN m LL am ED TM cia h i E TrtrtTTT A A AA he ini EM Figure 8 3 10 Working with SurfSeis 8 9 Figure 8 3 6 mala A nox s a a cuit n al ik E gg mM eT TN li i Dot OE BM il i i h Ui id MaNi i WW FATH WL ji A i Tie Corm v 1 M com ha MU M WITHIN i T UK wd wi be I AI ATHE A TA d ug IA a Figure 8 3 9 Overtone Analysis Overtone analysis is a tool for identifying changing propagation velocity patterns with freguency for all types of seismic events both signal and noise events This method 1s based on 2 D
65. rea Figure 7 3 Chapter 8 Working with SurfSeis This chapter discovers practical issues associated with using SurfSeis for specific projects Previous chapters have explained specific parts of the software in relation to their functions and capabilities In this chapter actual demonstration data sets that come with the software can be used to help understand the process and functions of this software and method All the demonstrations focus on included data sets In this chapter the following are addressed a How to acquire multichannel seismic data and prepare it for processing with SurfSeis u How to obtain an S velocity profile from the analysis of a single record a How to appropriately process data before starting dispersion curve analysis Working with SurfSeis 8 2 8 1 Data Acquisition MASW requires a multichannel record with at least 12 traces to produce reliable results A multichannel record can be obtained in either of two methods acquisition with a seismograph with at least 12 channels or acquiring single channel data with the source and receiver incrementally separated a predetermined distance after each trace is recorded and then gather up all the traces in source offset order to construct a 12 channel record Because low frequency components of surface waves increase the maximum depth Zing of investi gation and make the inversion process more stable 1t 1s critical that no analog filtering be applied during da
66. rily to that format The general flow of all con version routines 1s the same so little would be gained by going through each individually Information required from the user includes input file names usually one file name for each shot gather or field file output file name SurfSeis as most processing routines handles all the field files from a particular line as a single file and source sequence number SSN The source sequence number allows files collected with non sequential file names numbers such as a missing file number or with extensive or alpha character file names to be renumbered renamed and regrouped The function of any conversion routine 1s to reorganize and update trace headers in such a way as to allow a specific processing algorithm to recognize and operate on data Formatting seismic data for processing with SurfSeis involves organization of trace headers and data bytes into specific fixed length patterns and sequences The formatting utilities conversion routines are designed to operate on raw data input from and output to the hard disk Getting the raw data from the seismograph s preferred storage media hard drive floppy disk 9 track tape DAT cassette data tape cartridges RAM etc onto the hard disk requires procedures software and or hardware that can be supplied or recommended by the seismograph manufacturer Often the transfer of raw unformatted data to a computer s hard disk requires nothing more than the
67. sis ad 8 19 gt Preadj strientor Data ri ERES 8 23 A Po O Y UY ne DC 9 1 Bibliography and Recommended Reading on MASW 10 1 Chapter 1 Introduction The multichannel analysis of surface waves MASW method was first introduced into geotechnical and geophysical community in early 1999 although earlier development versions came out several years prior MASW is a seismic method which generates a shear wave velocity Vs profile 1 e Vs versus depth by analyzing Rayleigh type surface waves on a multichannel record The method utilizes multichannel recording and processing con cepts widely used for several decades in reflection surveying for oil exploration MASW utilizes energy commonly considered noise on conventional reflection seismic surveys The fundamental mode of ground roll the Rayleigh type surface wave event 1s without a doubt one of the most troublesome types of source generated noise on reflection surveys Rayleigh wave energy 1s defined as signal in MASW analysis and needs to be enhanced during both data acguisition and processing steps Because of this reversed definition of signal and noise in comparison to seismic reflection the method requires slightly different considerations and approaches to data acguisition Acguisition parameters are optimally determined using the approach described by Park et al 1999a It 1s suggeted that in most cases parameter design is favorable to body wave e g refle
68. sis due to freguency characteristics that are uniguely different and in part due to only minor overlap in velocity ranges The frequency and phase velocity characteristics of air wave events can be determined using Filter processing see Section 5 5 Figure 8 3 16 Overtone analysis shown in Figure 8 3 15 focuses on examination of the first arrival A narrow band about 50 60 Hz body wave event with a phase velocity of about 2000 m sec 1s identified This shows a slight dispersion possibly indicating a refraction from bedrock surface The presence of this refraction on the seismic record can be confirmed on the raw record by using the Filter option Figure 8 3 17 i Le PH Working with SurfSeis 8 12 J ve x 1n P yo si i IT a Lil il J EU iyn ai m UN U f ri y D n n E Ux i uc c d MAA li i Wu Figu Figure 8 3 13 i ADAIN i Hl Ju TM T alt a il i Tape M Working with SurfSeis 8 13 Controls After one or more overtone analyses have been completed frequency and phase velocity characteristics of all wave types should be well understood and reasonably well determined Using this information the dispersion curve for the fundamental mode surface wave can be constructed This is done by first properly setting the frequency range in the parameters tab of the Control dialog box Figure 8 3 7 Using automatic mode during the Preprocess st
69. ta acguisition A weight drop specifically a sledge hammer 1s normally employed as the seismic source Although its weight and impact velocity can vary depending on the desired Zmax a hammer of 5 to 10 kilograms 1s usually sufficient for investigations within a few tens of meters of the ground surface Low freguency geophones from an engineering perspective are essential Usually 4 5 Hz geophones are adequate for investigation in the few tens of meters The source offset X7 distance between source and the first closest geophone needs to be great enough X to assure efficient generation of surface waves that extend down to the primary depth range of interest The negative characteristics of the near field effect are controlled by Zmax and near surface velocity To avoid these undesirable attri butes a good rule of thumb is to insure X is greater than Zmay Geophone spacing dX can be determined after the maximum offset Xmax and total number of traces N are estab lished dX Xmax X1 N Xmax 18 usually at a source offset beyond which ambient noise begins to dominate the source generated surface waves see Figure 8 1 1 More detailed information about the selection of X and dX can be found from Park et al 1999 XI dX Xmax je Geophones Offset Source z 92 132 172 212 252 292 Channel 10 20 30 40 50 60 4 No i LL e ni renza di unm d lt i n sr Gv Er lol Ma e KAYA re
70. tailed explanation of the optimum criteria can be found in Chapter 1 Working with Example Data The Imax default value is set to 30 which will usually exceed the number of iterations necessary to obtain a reasonable solution for S velocity v profile Initial v Model The initial v model is usually calculated from phase velocities in the input dispersion curve Dispersion Data option This 1s the only option available 1f only one Stopping F M S Error kiia Ve Mosel kins hermi Dispersion Dato Preis Vs meerted Weight of Dispersa Data CH m E quid PA urp Fis Hermoso X ceat VII Figure 4 2 1 dispersion curve 1s to be inverted However if multiple consecutive dispersion curve files have been selected the inversion results of a previous record can be used as a good approximation of the initial model for the current inversion Using a previous profile Previous v Inverted option as an initial v model a more accurate result is more likely with fewer examinations 1 e iterations If the Previous v Inverted option is selected it is necessary that all input dispersion curve files be selected in the consecutive order of acguisition Weighting of Dispersion Data Each phase velocity in the dispersion curve can be treated with equal confidence Egual option or weighting can be selected in proportion to the signal to noise ratio S N S N optio
71. teps have been completed L Left Click R Right Click A Action Button D Dialog Button 5 5 Processing Seismic Data Seismic Data Display 5 7 Signal processing techniques sometimes need to be applied to displayed records before dispersion curve analysis to enhance signal to noise ratio S N Four different kinds of record specific processing steps are available filtering automatic gain control AGC mute and spectral analysis The following procedure begins a specific process p Click the appropriate button e g BM 2 execute the parameter action by mouse PAM when necessary see table below 3 double click the appropriate part or a specific part depending upon type of process you are applying within the displayed record The actual processing applied to the entire displayed record will be based on automatically determined parameters default values and or parameters determined through PAM This procedure is outlined in the table below To manually change any or all the processing parameters right click the button e g _ m and select from a dialog box containing a list of parameters you wish to enable or disable and associated value change for each individual parameter see Controls section Click Button Parameter Action by Mouse PAM Filter Mute eal Option A Place mouse cursor approximately in the middle of the noise to attenuate filter out Option B Click to drag
72. the curves see Section 3 6 Band Pass Filtering 75 1060 250 406 I HA it de M x DI Te Si zu e i ie es fU MARA Yd i gel l E ME DEDI q m ERA WI et EUH isl V pelles a n M aT Als a pra Figure 8 3 18 e i edid UD de TD HT WA dE ill TR le dq TS pipe Lil i Lon pr wet s til edet WE gu Working with SurfSeis 8 16 Rand Fass Filtering 3050 8120 EEE Rs A CASE Mia gt ci l Log l j i rl JANI AR lis AMARA TTT Mii DA ml Hini in ee a TERI LL me les NES a Es Ne f Id e gum i n ills mari ge de me DES i ig eal MH wiji AAA Sp MT th awe iin s H TEE ES aM sd UU i A E i i qR s i NE Figure 8 3 17 Figure 8 3 19 ME ard i Working with SurfSeis 8 17 lli ME 105 HE TH i jisi PT Hl Drit mal Hn AAA KANUNI Ha INTHE sal BUT i Figure 8 3 21 Figure 8 3 20 Figure 8 3 22 Dispersion Curve Ce e Le nr Fi leno A re a pg Tarif I bri Formia gt EXE Tib Tekh HEH Finch Das t B id Li ri aa I sare Tepp Lara fio sui Tsn bin Ci T 1 A a o en od Ei LM ny Eph ni le ee le A al mu a n Pre Pr rre rn BOE a I Ha md H Are Fili Fobia iy Pl a SE Dos RO ii e de A A ia ra ad siio morgan iri y PU Lira A Corales Hope Capena drei Tipo bia fa gna Free Sarga e d h F
73. the upper most layer of the surveyed area consists of material with a high seismic velocity or stiffness in comparison to materials below it e g pave ment it is unlikely the dispersion curve will contain phase velocities that increase signifi cantly with frequency up going trend Therefore whenever an up going trend 1s observed within a certain freguency range interference by noise must be suspected The most common source for this type of noise interference 1s the higher mode energy When the Working with SurfSeis 8 15 dispersion curves in Figure 8 3 22 are compared to the overtone image previously obtained Figure 8 3 21 it is obvious that the up going trend of the two curves is the result of the first higher mode Although it is tempting to consider the normal down going trend of the other curve Process No 2 in the higher frequency range as good it would be best discarded because 1ts corresponding image is not confirmed in the overtone image shown in Figure 8 3 21 The low lt 0 5 S N curves at these higher freguencies are also indicative of unreliable results Careful examination of S N curves at freguencies higher than 20 Hz reveals that the two up going curves have higher S N than the normal curve This is the result of higher mode energy being considered signal during the analysis Figure 8 3 23 shows the comparison of results from the different algorithms With the Normal algorithm dispersion
74. ume Y apex SM Boh Figure 6 2 1 6 2 2 Axis Attributes of the horizontal and vertical axes are controlled under the Axis tab Upper and lower data bounds for axes can be specified with the desired increment Auto Scale Axes automatically adjusts the scale of all active axes This 1s especially useful when the original scales need to be adjusted after display changes have been applied within the visible window 6 2 3 Labeling The Labeling tab controls the characteristics of title and legend The title of the graph can be manually typed in and changes to title or legend position and font characteristics can be made The legend for displayed curves can be turned off or on by checking the Visible box under the Legend tab A specific curve can be selected and high lighted by clicking it in the legend This is especially useful when several curves are displayed together and a specific curve needs to be distinguished for com parison purposes 6 2 4 Panel The Panel tab controls bevel and color of the entire displayed panel Under the Bevel tab the inner and outer bevel type and width can be desig nated Under the Gradient tab the background color can be changed in a continuous gradual mode from a Start Color to a End Color in a specified Direction When the Visible box 1s not selected the Panel Color option appears which enables the sel
75. ways with its assimilation not limited to the number of channels available on the recording device For example a 24 trace record can be obtained with a single channel device by stepping the shot and or receiver away from each other using a uniform increment By Introduction 1 2 repeating this procedure 24 times following an appropriate acquisition sequence in which the distance between seismic source and receiver change in a consistent manner a shot gather can be generated with 24 equally spaced traces this 1s called a noise analysis see Sheriff and Geldart 1982 It can also be obtained with a 12 channel recording device by making the measurement twice rolling the spread 12 receivers out 12 intervals or moving the source 12 increments away It is important to maintain the same receiver spacing for all the constituent traces and keep good field notes as to distances source offset between source and the closest to source receiver as well as receiver spacing Once a record 1s prepared it 1s ready to be converted into the processing format modified SEG Y KGS and begin the next step dispersion curve analysis The next step begins the process to calculate a Vs profile from the dispersion curve analysis performed on the multichannel record previously prepared This step 1s the most critical because 1t has the greatest influence on the confidence in the Vs profile In other words the Vs profile will have at best as much confidence as th
76. whereas the one in Figure 8 3 2 HardRock DAT was acquired on a limestone bench within a limestone shale cyclothem sequence in Kansas In Figure 8 3 1 highly dispersive surface waves dominate the seismic energy with no other seismic events identifiable A non dispersive surface wave event is also evident in Figure 8 3 2 as well as an air wave event Since the air wave event can be clearly identified as a separate event and represents much less total energy than the surface wave event the analysis will run smoothly with no problematic step s Data shown in Figure 8 3 1 can be regarded as good quality although far offset traces may be contami nated with noise due to low surface wave energy and may require special handling Figures 8 3 3 and 3 4 are examples of poor guality data Data in Figure 8 3 3 were acguired in Kansas where the surface topography was 1rregular and 1rregular voids of various dimensions were known to exist in the near surface Noise waves possibly generated as a result of the irregularity of the surface and presence of voids are obvious on far offset traces about half of the recorded traces This data set needs editing to remove the far offset traces before it is input into the analysis see Section 8 5 Data in Figure 8 3 4 were recorded at a site where heavy construction vehicles were in close proximity to the line and moving all around the survey area Most of the energy on this record 1s the surface waves generated by these m
77. y to provide various types of images that can be used during the analysis as well as preparation of a project report A generalized flow chart for SurfSeis 1s displayed in Figure 1 1 Introduction 1 3 Figure 1 1 Chapter 2 Full Auto A single multichannel seismic record 1s processed to produce one S wave velocity profile when fully automatic mode is selected All the processing parameters such as optimum frequency and depth ranges are automatically determined based on predetermined normal or average data characteristics This option of analysis is best suited for seismic data with a good signal to noise ratio S N The following files are generated and saved during and at the end of process a One dispersion curve DC a One file containing all the processing parameters used for inversion IND a One output file from the inversion process that contains the theoretical dispersion data and layer model a Another output file from inversion processes that contains the processing history followed during the iterative inversion steps Note If the seismic file input contains more than one record the first record in the file 1s processed Full Auto Analysis 2 2 2 1 Main a When the Full Auto button 1s selected a dialog box Figure 2 1 1 pops up that allows selection of a multichannel seismic record s T file All intermediate files resulting from the processing flow dispersion curve DC inversion

USING SURFSEIS - the Kansas Geological Survey

Contents

Download Pdf Manuals

Related Search

Related Contents