AEMINV
Contents
1. T T T T T T T T T T T T T 4000 6000 Distance m 0 1000 2000 3000 4000 5000 6000 7000 x0 3551654 Distance m xl 3558660 yO 7098299 yl 7098128 RMS d 0 28319 RMS m 0 02270 Roughness on Magn susc off 1 0 1 2 3 logy Resistivity Qm 312. HMD source and horizontal response component or b horizontal loops i e vertical magnetic dipole VMD source and vertical response component Two measurement configurations are considered a in line along the flight line and b broadside perpendicular to the flight line Figures la and 1b illustrate the coaxial and coplanar loop HMD AEM systems used by the Geological Survey of Finland Note that although AEMINV was originally developed for airborne applications it can also be used to interpret ground EM data and VMD systems such as Apex Maxmin II Fig 1d The EM response of the loop loop system is defined as the in phase real and quadrature imaginary components of the ratio Z H 01 Ho 1 where Hijo is the total magnetic field which is the sum of the primary field due to transmitter loop and secondary field due to currents induced into conductive earth and Ho is the free space magnetic field at equivalent source receiver distance L The data are assumed to locate at the center between the source and the receiver a Rx P xy h Tx e ae E EEREN e loop spacing L Figure 1 Schematic view of three loop loop configrations often used in geophysical EM measurements a coaxial inline HMD b coplanar broadside HMD c inline VMD The EM response is assumed to locate at the center P x y h between transmitter Tx and receiver Rx and the loop spacing L is kept constant AEM data location P x y h
3. 1 ror ELOD Skin oints 00 366 63 22 471 85 0 270092E 01 0 207693E 01 Z Rhol Thil Rho2 NP 118 resistivity in Ohm meters Listed conductance values S are multiplied by 1000 Listed susceptibility values SI multiplied by 1000 Sus2 Cndl fxZ fxRl 0 0000 100000 0000 227 1138 0 0000 100000 0000 313 9914 34 407 13 0 0000 92188 8281 234 1963 292 9494 401 6601 456 0468 fxT1 fxR2 1 0000 1 0000 1 0025 2 2 rgZ 2711 3 5404 1399 rgRl bl To 1 1 I rgT1 H 1 ed ll rgR2 0 150 0 0 0 0 fxS 0 0 0 rgs 4 4 Output files The measured and the modelled data and the data weights can be saved into a column formatted text file OUT The file can be used for example to prepare response graphs using commercial third party software In addition to useful descriptive information a commented copy of the file header is stored before the actual data columns If the comment characters preceding the header lines are removed the file can be read into AEMINV as a preformatted data file Otherwise the format conforms to that of a generic data file The computed data Re_c amp Im_c are located in columns 6 and 7 and measured data Re_m amp Im_m and weights are located in columns 8 9 and 10 An example of the file format is given below Results from AEMINV computed and measured EM data 1 25 000 1 0 freqs loop spacing HMD
4. GUI using a screen aspect ratio that suites either old 3 4 displays or widescreen displays When changing from normal to widescreen mode the program asks for an aspect ratio value 0 7 0 8 is good for most widescreens Ok to Exit item is used to confirm the exit operation On exit the graph parameters current results measured and computed data and the model are automatically saved into AEMINV DIS AEMINV OUT and AEMINV LAY files respectively Errors that are encountered before the GUI starts up are reported in the AEMINV ERR file Inside GUI mode run time errors are displayed on the screen 2 2 Controls Update button is used to validate the changes made to text fields to perform forward computation and to refresh the graph accordingly Note that pressing Update is not always necessary but usually it does not hurt to press it either When starting optimization for example the values of global and local Lagrange multipliers are automatically checked Layers text field is used to define the number of layers 1 3 to be used in the modelling After changing the layer number it is necessary to press the Update button After that the model will reset to 1ts default values Line text field defines the number of the profile line It is active only when multi profile data has been read in Likewise the lt Line and Line gt buttons which are used to swap the active line to the previous and next one will be inactive for single profile
5. flight line Q 0 0 Q O O 0 0 0 0 Oo gt ground level Pi Eq 447o layer thickness t overburden Pa Ey H Ki basement Figure 2 Schematic view of two layer resistivity model used in AEM inversion 1 1 Requirements and setup AEMINV requires a PC with MS Windows Vista 7 8 operating system and a graphics display with a resolution of at least 1024x768 pixels Memory requirements and processor speed and are not critical factors since dynamic memory allocation is used and the computations are rather fast even on older computers AEMINV has a simple graphical user interface GUI that is used to handle file input and output to set inversion parameters and to visualize the measured and computed EM data and the resulting pseudo sections The user interface and visualization are based on the DISLIN graphics library The distribution file AEMINV ZIP can be downloaded from author s web site at the University of Oulu Finland https wiki oulu fi x EoU7AQ The distribution file contains a 32 and 64 bit executables AEIMINV EXE amp AEMINV64 EXE dynamic link library for OpenMP parallelizations LIBIOMPSMD DLL for 64 bit version only this description file README TXT an user s manual AEMINV_ MANU PDF and two example data files EXAMPLE DAT and GENERIC DAT and example result files for 1 2 and 3 layer cases To install the program simply unzip Pkzip 7zip the distribution file onto hard disk and a
6. and convergence will be fast but the model will become rugged Typically the value of LG should range between 0 1 and 10 1 and 1 on the 10 logarithmic scale widget Note that both the data and the model parameters are scaled using the maximum data variation and maximum parameter variation resistivity scale limits and depth limits Therefore the default value LG 1 0 provides rather good compromise between data fit and model smoothness but the convergence may be rather slow Therefore the Lagrange multiplier should be decreased in the beginning of the inversion and increased after sufficient fit has been obtained to obtain smoother model Im scale defines a relative weight WI between in phase and quadrature components By default WI 1 0 the two components have equal importance in the inversion Decreasing WI gives quadrature component less importance in the inversion Vice versa increasing WI reduces the importance of the in phase component This parameter should be used if the other response component is much noisier and lower in amplitude than the other F length defines the filter length FL used to compute the parameter roughness default FL 2 The roughness is defined as the difference of the parameter p from the mean of the surrounding points p X 9p j i FL i FL 2FL where 6 is Kronecker s delta function The longer the filter is the smoother is the resulting Occam model and the less sensitive the mo
7. data To quickly jump from one profile to another you can first give the number of the destination line on the Line text field and then press either one of the Line buttons to change the active line Swap freq button is used to change the response graph from one frequency to the next one The button is inactive in case of single frequency data Note that the actual frequencies in Hz the loop spacing and the nominal flight altitude are defined when the data is read in but they can be redefined via the 4eminv Configuration Set system parameters menu item Swap comp button can be used to show only the in phase real or the quadrature imaginary component or both data components When changing the data component the y axis of the response graph will be rescaled accordingly This allows the user to see the fit between the measured and computed better Lagr sca text field defines the so called global Lagrange multiplier LG which determines whether the inversion tries to minimize the data error instead of the model roughness The horizontal scale widget below the text field allows faster changing of the logarithm of the Lagrange multiplier Important The global Lagrange multiplier is the most important inversion parameter which affects the convergence and smoothness of the resulting model If LG gt 1 then the model will be smooth the fit will be poor and convergence will be slow If LG lt 1 then the fit will be good
8. of both the response plot and the pseudo section since both graphs are equally high The 6 th line defines the 1 minimum R_min and 2 maximum R_max value of the logarithmic EM response axis ppm or 3 minimum Z min and 4 maximum Z_max values of the depth axis m 5 the threshold for small in phase values and 6 maximum spike value Note that the depth axis is positive downwards and that the minimum depth is usually based on the maximum value of the recorded airplane height The 7 th line defines the default values for 1 3 resistivity Qm and 2 4 thickness m of the first and second layer and 4 the resistivity and 5 magnetic susceptibility of the basement The 8 th line defines the default weights for 1 3 resistivity Qm and 2 4 thickness m of the first and second layer and 4 the resistivity and 5 magnetic susceptibility of the basement The 9 th line defines first 1 the measurement system 0 VMD 1 HMD and 2 configuration 0 in line 1 broadside and then the column indices of the 3 4 x and y coordinates and 5 airplane elevation The 10 th line defines 1 the number frequencies NOF and 2 loop spacing and 3 nominal flight altitude The next NOF lines define the column indices of the 1 in phase real and 2 quadrature imaginary data components 3 weight and 4 frequency Please note that the parameters on lines 9 10 NOF are used to store the previous values when reading XYZ files and generic data file
9. save the current layered earth model into LAY file Read disp read new graph parameters from DIS file SS AA save the graph in Adobe s Postscript format ee save the graph in Adobe s Encapsulated Postscript format Save Graph as EPS f save the graph in Adobe s PDF format Save Graph as PDF l save the graph in Windows metafile format Save Graph as PNG Save Graph as GIF save the graph in GIF compressed format Selecting any of these options brings up a typical Open Save file selection dialog that can be used to provide the name and directory location for the file Data files are stored in text format The graphs are saved as they appear on the screen in landscape A4 size The AEMinv menu has following sub menus Source type choose loop loop system VMD and HMD Configuration choose loop configuration and define system parameters Inversion enable data weights and parameter roughness and set thresholds Distance unit define distances in meters kilometers feet or miles Color scale choose between five different color scales Miscellaneous change some other graphical setting Note that usually horizontal magnetic dipoles vertical loops are used in airborne applications and vertical magnetic dipoles horizontal loops are used in ground applications Slingram a k a HCPL or HLEM The inline HMD configuration is a coaxial loop system and the broadside HMD system is a co planar loop system cf Figure 1 Perpendicular sy
10. starting with or characters are considered to be comment lines 4 2 Generic data files If the input data file starts with a or character it is considered to be in a generic column format and the user needs to provide all the necessary header information interactively Any other lines that start with these characters are considered comment lines and they will be ignored The example below illustrates the format of the generic data file Generic data file x y re im h 35516 00 70982 00 27 16 27 35516 00 70983 00 9 1 30 After opening a generic data file the program asks for 18 The number of frequencies NOF loop spacing m and nominal flight elevation m Measurement system HMD vs VMD and loop configuration inline vs broadside Column indices of the x and y coordinates and altitude ICOX 1 ICOY 2 ICOH 5 BR Ww N ma Frequency Hz and the column indices of in phase and quadrature components and data weights for each frequency 4 2 XYZ data files XYZ files or Geosoft XYZ files are column formatted text files that can contain multiple profile lines The LINE directive is used to separate individual profiles The number or character string after the LINE directive is used as an identifier or title of the line Rows starting with character are interpreted as comment lines As with generic data files the user needs to provide the abovementioned header information interacti
11. that for example the default value of 0 01 is actually 0 0001 Thus all singular values that are smaller than 1 10000 compared to the largest singular value will be damped Normally there is no need to change the threshold Increasing MIN SVA above the default value produces stronger damping which leads to slower convergence Decreasing MIN SVA may lead to unstable inversion as the inversion will start to resemble steepest descent algorithms Note All distances loop spacing and flight altitude depth to the top and layer thicknesses used in the computations are defined in metres Since the inversion is performed point by point the actual data coordinates can be anything Distance units menu item only adds correct unit on the x axis below the response graph and pseudo section It does not redefine depths heights and distances in actual computations The Miscellaneus sub menu Show hide labels Show hide layer lines show or hide resistivity labels in the pseudosection show or hide black horizontal lines between layers Show hide skin depth show or hide skin depth curve in the pseudosection oS show hide profile start and end coordinates Show hide fix marks show hide markings for fixed parameters Show symbols lines swap between normal and alternative data scaling Alternative scaling use alternative scaling for the data Reverse data sign Reverse profile direction Swap ppm Edit title lines reverse the sign o
12. Depth m AEMINV Laterally constrained 1 D inversion of EM profile data Version 1 3 user s guide 1000 2000 3000 4000 5000 Distance m 1 0 1 2 3 4 5 log y Resistivity Qm Markku Pirttij rvi 2014 University of Oulu 6000 7000 Contents COMES 0 A A a Male A eee Ned te eae Eies 2 AO 3 AeA Reguirements ANG EP O ARS Bala 5 A A Uat 6 2 O A anna dec gaat deans 6 2 2 COMMONS nannan ienai e eaa a HRI eae ne na ee 9 3 Interpretatiom procedures A eee a ee Sie 14 A A AS 17 e mhi eiie acai eaae a EAE alata a EEA E aN 17 4 2 Generne dita lesa 18 4 2 A O 19 4 3 Layer Meda 19 SADA ae 21 A AO OO e a A a Aa O ces 22 5 Additional information rd 25 RI a dd e 25 5 2 Terms of Use and disclaimer i sssiasssssatassasusaasavsstiaransesacdaresacadazoasergavsssdarsanegaagaraneasias 26 Appendix A AEMINV GUI at StartUP oocoocoococonnnoncnononcnnnoconocnnnnonnccon coco no cono nono ccon anno nacinnos 21 Appendix C Three alternative inversions A a A Ae 29 Appendix D 2 frequency AEM data 0 TA A 30 Appendix Er Discontin ities lid Aa o 31 1 Introduction AEMINV is a computer program for geophysical interpretation of frequency domain airborne electromagnetic AEM data using one dimensional 1 D layered earth model Multiple profiles i e measurement lines can be processed simultaneously The model parameters are the electrical resistivity and thickness of the layers and the resistivity and magnetic susceptibility of
13. E L 5 4 A y y gt e E 2 a nnn E attra a al iiaii a a 0 j j T l T T T T T T T T T T T 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 me L I 1 f l 1 1 f El 1 g 6000 4 ee iz a Freq 14363 0 Hz x Realm 2 A sl Real c d pi ob SRT x Imam 4000 k i Imag c x E ee ena Ei j h 2000 NY 1o 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 Distance m L L 1 1 1 TP L eel 1 i Real ppm amp Ima Depth m 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 ii A E 0 00 1 25 2 50 3 75 5 00 log10 Resistivity Ohmm Height m Skin m 30 Appendix E Discontinuities Top Two layer inversion obtained after automatic setting of discontinuities and manual insertion of additional discontinuities for layer thickness and basement resistivity Middle The same psudosection after resetting the basement resistivity Bottom The same pseudo section after resetting also the layer thickness and overburden resistivity and the resulting fit between the measured and modelled data above it Depth m Depth m Real Imag ppm Depth m 2000 4000 6000 Height Skin 100 200 300 6000 7000 RMS d 0 02518 RMS m 0 02506 1000 2000 3000 4000 5000 1 1 L 6000 7000 RMS d 0 15737 RMS m 0 02506 1000 2000 3000 4000 5000 1 1 i fi PATA A FIET F ponp PAR a
14. VMD 1 2 4 x y amp altitude column 3220 00 8 9 10 freq re im amp weight column Number of lines 1 Total number of data 236 Number of frequencies I Frequencies Hz F 3220 0 Loop spacing m L 25 00 HMD source Inline configuration Units parts per million ppm Line 1 x y dr h m z m Re c Imc Re m Imm wgh 118 number of points 35516 70982 0 000 27 00 0 00 20 34 17 56 27 00 16 00 35516 70983 37 216 30 00 0 00 9 5 6 9 06 9 00 1 00 35517 70983 74 337 32 00 0 00 14 74 14 96 18 00 17 00 inline broadside 1 00 1 00 1 00 4 5 Graph options Several graph parameters see Appendix A can be changed editing the AEMINV DIS file This allows translating the graph texts into other languages for example Note that the format of the AEMINV DIS file must be preserved If the format becomes invalid or if you replace the AEMINV EXE file with different version you should delete the graph parameter file and a new one with default parameter values will be generated automatically the next time the program is started The file format and default values are shown below Aeminv 1 30 parameter file 32 32 24 16 1 1 1 1 1 1 2 1 300 300 0 85 0 67 0 70 0 10000 100000 00 63 00 200 00 10 00 5000 00 10000 00 50 00 5000 00 50 00 10000 00 0 001000 0 00000 1 000 1 000 1 000 1 000 1 000 1 000 LOL Be 3 1 25 00 30 00 3 4 0 3220 000 AEM interpretation l1 Layer model 2 Layer model 3 Layer model Di
15. a and model RMS values Update Layers 2 Eme Eines Swap comp Lagr sca 0 0 Im scale F lenath Iters Z wght Current line zoomed part C Alllines full length Defaults Weights R1 10000 1 000 n 50 1 000 R2 1 000 T2 R3 SB Default Magnetic susceptibility on off Show magnetic susceptibility Resist Depth 0 1 63 100000 300 Zoom Pan Fix Free Roughen Reset fis R1 tgh A1 res A1 fix T1 rgh T1 res T1 fix R2 tgh R2 tes R2 Edit edit L1 edit R1 edit R2 HMD Inline Loop 25 00 m Freq 3220 0 Hz Single profile Points 118 AEM interpretation Line 591 Real m Real e Imag m mag e Real amp Imag ppm eget e a 1 DO HH e DE I T 3000 4000 Distance m Depth m Height Skin 3551654 7098299 3000 4000 Distance m 1 2 3 4 logig Resistivity Qm 7000 xl 3558660 yl 7098128 RMS d 0 02898 RMS m 0 02326 Roughness on Magn suse off 1 D AEM inversion 1 3 c MTP 2012 28 Appendix C Three alternative inversions Top The half space resistivity and depth to the top flight height were optimized whish is similar to the traditional apparent resistivity transformation Middle The depth to the top was fixed and the parameters of the t
16. arrow cursor becomes a crosshair cursor above the graph area The user makes selections pressing the left and right mouse buttons as discussed below 12 Important Zooming and all other mouse editing functions work in a similar fashion To zoom in the user presses the left mouse button at two different horizontal locations on the pseudo section or response graph and then updates the selection by pressing the right mouse button once Pressing the right mouse button again without any selections will end the edit mode and program returns to normal operation When zoomed in the y axis of the response graph will be rescaled accordingly Thus zooming allows the user to see the details of the data better Note that zooming can be used to limit the inversion so that the parts of the profile outside the current zoom area are unaffected in the inversion Zoomed part vs Full length To zoom out to the original extent one enters the zooming mode and presses the right mouse button once without any selections with left mouse button To pan the zoomed view left and right the user needs to press the left mouse button only once and then update the screen with right mouse button The fix R1 fix T1 fix R2 fix T2 and fix R3 buttons are used to manually fix or free the parameters from or to the inversion Fixed parameters are indicated by a white cross above the corresponding element in the pseudo section The editing mode is similar to zoomi
17. del is to data noise Iters defines the number of successive iterations to be performed when the Optimize button is pressed default 10 iterations Z wght defines a local Lagrange multiplier LZ for the parameter corresponding the depth to the top of the model Note that ZZ should be equal to zero in two and three layer inversion in which the topography of the top layer should be fixed to zero level In half space inversion however it should be non zero by default LZ 0 01 Decreasing LZ below 0 01 makes the topography more rugged and increasing LZ above 0 01 makes the topography smoother Optimize button performs a given number of inverse iterations The inversion is based on linearized inversion scheme The sensitivity matrix is constructed numerically using forward differences The linear system is solved for the parameter steps using singular value decomposition SVD together with adaptive and automatic damping 11 The two radio buttons below the Optimize button are used to decide whether the inversion is performed only for the current profile or all profiles one at a time Current line vs All lines Alternatively if only one profile has been read in the inversion is made only for the zoomed part or the whole profile length Zoomed part vs Full length The six text fields below the optimize button are used to define the default or initial values left column and local Lagrange multipliers of the model parameters resisti
18. eal m orks rogs E lectric Layers 2 Loop 25 00 m Real e Line 1 Freq 3220 0 Hz x Imag m Single profile AEM interpretation EAS wap fre Swap comp j z Legr sca 3 Line 591 E El gt Im scale md 1 0 ll 1 J 7 e E FF a z ee e tHe ds O e an att Iters 10 al 3 a A ta rer Z wght 0 000 amp 2000 e a E Optimize oh 3 han F Current line zoomed part El C Alllinesfull length 5 4000 5 i E Defaults Weights 2 J t L R1 10000 1 000 4 ui 50 1 000 4 6000 R2 5000 1 000 7 F r T T T T T T T e 0 1000 2000 3000 4000 5000 6000 7000 an Distance m Default I Magnetic susceptibility on off 50 4 Y 1 l l l 1 I Show magnetic susceptibility Resist Depth PR 0 mn 01 63 3 Z 50 max 100000 200 3 Zoom Pan 100 Fix Free Roughen Reset Aa fix R1 igh R1 res R1 150 fix T1 rgh T1 res T1 200 fix R2 igh R2 res R2 0 1000 2000 3000 4000 5000 6000 7000 Cero a a x0 3551654 Distance m xl 3558660 hR A y0 7098299 yl 7098128 Edit editR1 _ edit R2 F 4 0 1 7 3 4 5 RMS d 0 46321 data ON Roughness on Height logig Resistivity Qm Magn suse off 1 D AEM inversion 1 3 c MTP 2012 Appendix B AEMINV GUI after inversion The inversion result was obtained after 20 iterations 2 runs using the default inversion parameters shown in the left control panel The black line on the cross section depicts cumulative skin depth RMS d and RMS m are the dat
19. f the measured data reverse mirror the profile direction swap between parts per million ppm and percentage change or edit the two title lines of the graph The dynamic range of AEM data can be several decades and hence the low amplitude anomalies are difficult to see clearly Alternative data scaling makes it possible to see better the fit between measured and computed data for small in phase and quadrature component values Reversing the data sign might be useful when reading coaxial data that is normally negative but that for some reason has been stored as positive values Reversing the profile direction not only affects the way it is shown in the graph but will also cause the new data and layer model to be saved in reverse order Usually AEM data are defined in parts per million units whereas ground data are defined in per cents The Swap ppm item thus allows AEMINV to normalize the computed response correctly When alternative data scaling is used the y axis units are displayed as instead of ppm or The Exit menu has two items The Restart wide norm item will rebuild the GUI and swaps between traditional 3 4 aspect ratio and widescreen displays When changing to widescreen mode the program asks the user for a scaling value which is less than one for widescreens equal to one for 3 4 screens and more than one for rotated screens The Exit menu has two items Restart wide norm item is used to close and restart the whole
20. input data file is shown below Line 591 T 2Sa A O nof loop HMD coaxial o L X y eZ 3220 340 fre re im wgh 35496 00 70981 00 49 40 32 35497 00 70980 00 48 6 30 35498 00 70980 00 20 23 30 The 1 st line defines a header text max 80 characters which is used as a secondary title above the response graph The 2 nd and 6 th line are used for comments and can be left empty Usually the header line should define the name of the measurement site and possibly the number of the profile line If the header line is empty the default title in the AEMINV DIS file will be used instead Note that the Edit title lines menu item can be used to redefine the titles and if the secondary title initially set in the DIS file is not blank it will take precedence from that of the data file To disable the fixed secondary title and to enable the title line of the data file one should give blank line for the secondary title using Edit title lines menu item The 3 rd line defines the 1 number of frequencies NOF 2 loop spacing L in meters 3 loop orientation VMD 0 HMD 1 and 4 loop configuration inline 0 broadside 1 Note that loop spacing and measurement system can also be defined inside the GUI so it is not required to remember their codes 17 The 4 th line defines the column indices ICOX ICOY and ICOH of the x and y coordinates and the flight altitude meters If altitude data is not available the corresponding co
21. ischknecht et al 1991 The quasi static solution does not take into account the effect of permittivity i e dielectric constant e amp Non zero magnetic susceptibility k 44 1 is assigned only for the basement layer all other layers are non magnetic 4 4 The Hankel transforms are computed by a convolution algorithm based on that of Anderson 1984 The optimized filter coefficients have been computed using Christensen s 1990 algorithm The parameter optimization is based on a linearized inversion method where singular value decomposition SVD with adaptive damping is used The inversion method has been described in my PhD thesis Pirttijarvi 2003 The lateral constraining is adapted from the GRABLOX program I have made for 3 D block model inversion of gravity data See GRABLOX2 user guide Pirttij rvi 2009 for more information about the Occam inversion method AEMINV was written in Fortran90 and compiled with Intel Visual Fortran XE 14 The graphical user interface is based on the DISLIN graphics library version 10 3 by Helmut Michels http www dislin de Since DISLIN graphics library is available for other operating systems Solaris Linux Mac AEMINV could be compiled and run on these platforms without any major modifications However I do not intend to provide active support for the program and the source code will not be made available If you find the computed results erroneous or have suggestions for improvemen
22. lumn index should be zero ICOH 0 In this case the elevation of all data points is set equal to the value given in Height text field in GUI control panel Note that the x axis of the graphs is the profile distance computed from the beginning of the profile e g d axo tyo If any of the column indices ICOX ICOY ICOH is negative the profile direction will be reversed actual coordinates are preserved If x or y coordinates are missing or only profile distance is available the corresponding column index should be set equal to zero ICOX 0 or ICOY 0 Note that the profile distance x axis must be in an ascending order also when x or y coordinates are missing The profile will be reversed automatically if the profile distance computed as d x x0 or d y Yo would become decreasing The 5 th line defines the frequency F1 in Hertzes of the first dataset and the column indices of the in phase ICI1 and quadrature ICQ1 components and data weights ICW1 Data weights are used to increase wgh gt 1 or decrease wgh lt 1 the importance of individual data points If the data file does not include data weights the corresponding column index should be set to zero ICW 0 If the file contains data from two or more frequencies the frequencies F2 and column indices ICI2 ICQ2 ICW2 of the successive datasets would be defined on the next NOF 1 lines The remaining NOP lines define the actual data in column myn format Lines
23. meter should be non zero to allow the inversion of the depth to the top of the conductive half space When multi frequency data are available the user may want to try to interpret the thickness of conductive targets This however may be ambiguous because real targets are often two dimensional and AEMINV uses layered earth model The skin depth curve should be used to assess if the EM system is capable of seeing through conductive layers 14 Important A previously computed and stored layer model can be opened after data has been read in Before reading the LAY file it is necessary that the number of layers is set according to that used in the data Since the initial number of layers is two the user must change it to one before reading a one layer model The user should then pay attention to the default values of the layer parameters The initial resistivity and thickness values depend on the problem frequencies loop spacing and true earth conductivity and the user should experiment with their values to allow the inversion to start from a good initial guess Normally the initial resistivity model should be a homogeneous half space unless there is some a priori information against this assumption eg more conductive overburden After editing the text fields of the default values the user can press the Default button without pressing Update button to reset the initial model The inversion can now be started by pressing
24. new folder will appear 2 User interface In the beginning AEMINV reads graph parameters from the AEMINV DIS file If this file cannot be found a new one with default parameter values is created automatically The program then displays the standard Windows Open file dialog so that the user can select the file that contains the measured EM data The DAT file extension and mask is used by default for data files that contain only single profile and a predefined header The program then computes the synthetic response of the initial two layer model Finally the program builds up the GUI and creates graphs of the measured and modelled data and the layered earth resistivity model see Appendix A If the user cancels the open file operation the graph area will be blank Note Multiple profile lines can be read from Geosoft XYZ files File Read XYZ data but only after the GUI has already started up Thus when data is to be read from an XYZ file one can choose to cancel the initial open file operation e g pressing ESC key See chapter 4 for more information on file formats 2 1 Menus The main window of the AEMINV application has three menus The File menu open and read measured data from DAT file Read AEM data Read XYZ data open and read measured data from X YZ file Save AEM data save the measured and computed data into OUT file Read Layer model Save Layer model read previously saved layered earth model from LAY file
25. ng the the left mouse button is used to select points However the program stays in the editing mode and more elements can be selected or previous ones can be unselected until the right mouse button is pressed twice in a row 1 e without any left mouse clicks in between Important If a single left mouse click is made outside the profile left to the beginning or right to the end of the profile then all the block elements of that parameter are set free The rgh R1 rgh T1 rgh R2 rgh T2 and rgh R3 buttons are used to set discontinuities into the model These allow but do not necessitate a jump in layer boundary or resistivity at that point The editing mode is similar to fix free status editing with the exception that the discontinuities are located between two model blocks Discontinuities are indicated by white vertical resistivity or horizontal thickness lines between the block elements in the pseudo section As with fix free status editing if a single left mouse click is made outside the profile then all discontinuities are removed from that parameter 13 The res RI res Tl res R2 res T2 and res R3 buttons are used to reset the corresponding parameter to its arithmetic mean If discontinuities have been set in the model the mean is computed between them see Appendix D This option helps the user to create rough categorical interpretation model The edit RI edit Tl edit R2 edit T2 and edit R3 bu
26. o the program please use either the conference paper Pirttijarvi amp Lerssi 2006 or this user manual e g Pirttij rvi M 2014 AEMINV Laterally constrained 1 D inversion of EM data Version 1 3 user s guide University of Oulu Department of Physics lt https wiki oulu fi x PIU7AQ gt The program is provided as is The author M P and the University of Oulu disclaim all warranties either expressed or implied with regard to this software In no event shall the author or the University of Oulu be liable for any indirect or consequential damages or any damages whatsoever resulting from loss of use data or profits arising out of or in connection with the use or performance of this software All in all use AEMINV at your own risk 26 Appendix A AEMINV GUI at startup Program controls at the left graph area at the right with response on top and model cross section at the bottom The measured data are plotted with symbols and computed response with solid and dotted lines The dotted line on top of the cross section depicts flight altitude variations The vertical white lines depict automatically defined discontinuities in the basement layer E AEM inversion 1 3 s e roe t l lelak File AEMinv Exit RP D Works2 _Progs Electtic Update HMD Inline ap R
27. ome point and then the fixing buttons to fix that value during inversion The constrained inversion will then try to adjust the neighboring elements to fit the fixed value to minimize the model roughness Alternatively one can save the current model into a LAY file and manually edit the parameter values and fix free values The user should always test different models and different values of the various inversion parameters particularly the global Lagrange multiplier and the separate parameter weights As often in geophysical interpretations there is no unique inverse model The practice however will create masters Because of dynamic memory allocation the only limit for the total number of data points is the available computer memory However since single profile is considered simultaneous processing of large datasets should be avoided because the graph will get too crowded In principle however it is possible to process multiple lines by putting discontinuities between the lines and invert the whole dataset blindly and use external software to create maps of layer conductivity and thickness 16 4 File formats 4 1 Formatted data files Before starting up the program make sure that your input files are formatted properly Currently EM data can be read in three formats a preformatted DAT files b generic column formatted DAT files and c Geosoft XYZ files which support multiple profiles The format of the preformatted
28. s that do not contain file header 23 e The next line should be left empty e The following lines define various text items of the graph max 80 characters O O O Main title line of the response graph Second title line of the graph used if the data file does not include a title Three possible title lines of the layered earth model graph currently not used at all Title of the x axis defined in the data file usually the distance along profile Two y axis names for the response graph Four legend names of the response graph measured and computed data The label of the color scale defining resistivity Note that the logo label is added in front of the color scale label automatically The vertical axis name of the resistivity pseudo section always meters The legend names for the auxiliary flight height and skin depth curves The label of the color scale defining susceptibilty Special characters and should not be used in any text strings because they define instruction strings that enable Greek symbols Other instruction characters that must be avoided are for superscripts for subscripts and _ for setting the baseline back to its original position Please see DISLIN manual for more information about the topic 24 5 Additional information The numerical computation of the EM loop loop response of the layered earth is based on the solution in Keller and Frischknecht 1967 and Fr
29. stance Real Imag Real Real Imag Imag c Resistivity M2 W M1 m Depth m Height Skin Susceptibility SI 303 e The 1 st line is a comment line defining the file version e The 2 nd line should be left empty e The 3 rd line defines four character heights The first one is used for the main title and the graph axis title the second is used for the axis labels the third is used for the plot legend and the fourth is used for the resistivity labels of the pseudo section 22 The 4 th line defines The remaining parameters define 1 flags for normal vs widescreen display 2 show hide resistivity labels 3 show hide line between layers 4 show hide skin depth curve 5 show hide fix free status 6 flag for ppm parts per million or percentage response 7 index number of color scale and 8 flag for symbols or solid dotted line for measured in phase and quadrature data The 5 th line defines first the 1 x horizontal and 2 y vertical position of the origin of the main graph in pixels from the bottom left corner of the page and then the length of the 3 x and 4 y axes relative to the size of the total width and height of the page eg 0 5 50 of the width or height which is equal to 2970x2100 pixels landscape A4 5 the aspect ratio for widescreen display Aspect ratio is equal to 1 for 3 4 and about 0 8 for most widescreen displays Note that in vertical direction the length defines the relative height
30. stems VMD HMD or HMD VMD cannot be modeled This is because AEMINV uses a 1 D model and interpretation of perpendicular systems would benefit from 2 D modelling The Configuration Set system parameters menu item can be used to define 1 loop spacing 2 frequencies and 3 nominal flight altitude Data weights are used to increase or decrease the importance of individual data points in the inversion The weights are read from the input data file together with the measured data If weights do not exist the corresponding menu item Weights on off will be inactive Likewise Roughness on off item is used to activate or inactivate roughness and to draw or hide the roughness lines in the pseudo section Set thresholds item defines thresholds for small in phase values MIN_RE maximum data spikes MAX_SPK and minimum singular values MIN_SVA The first one is used to free fix free status of magnetic susceptibility in the initial model By default the magnetic properties of all those data points whose in phase component is higher than MIN_RE value are not inverted at all the pseudo section will be white When data is read in in phase and quadrature values that deviate from the surrounding values more than the MAX_SPK are rejected If MAX SPK 0 all data values are accepted MIN_SVA defines the minimum value of the normalized singular values used as Marquardt factors in the SVD based inversion method The value is defined in percents indicating
31. the Optimize button Usually the convergence is rather slow when the default value of global Lagrange multiplier is used and the optimization must be made multiple times before good fit is achieved Therefore user may want to decrease the Lagrange multiplier at the beginning and increase it later to create smoother model To see if stable solution is found the user should pay attention to the data RMS and model RMS in the upper right corner of the graph as well as to any visible changes in the pseudo sections To see finer details in the fit one can swap between different response components using Swap comp button try using the Alternative scaling menu item and or zoom into a section of the profile using Zooming button Without any discontinuities the constrained inversion will create a quite smooth model After few basic inversions one should have a pretty good idea where there might be lateral conductivity or thickness variations Using the editing buttons for model roughness rgh R1 rgh T1 etc one can define discontinuities After this few more optimization runs should be made to see how the model changes Alternatively after editing discontinuities one might disturb the model using the resetting buttons and start the inversion from a new initial model 15 If the user has any a priori information for example about the thickness of the overburden at some point then one can use the editing buttons to set the thickness or resistivity at s
32. the basement layer Depending on the number of frequencies max 10 1 3 layer models can be utilized The inversion can also optimize the resistivity of the half space and the depth to the basement which is equal to the traditional apparent resistivity conductivity and depth transformation The inversion is made independently for each profile point using one dimensional layered earth model see Figure 1 A resistivity pseudo section is obtained by plotting the resistivity thickness values below adjacent data points with colored rectangles Starting from an initial model an iterative linearized inversion with adaptive damping is used to update model parameters so that the data error 1 e the difference between the measured and the computed data is minimized Laterally constrained inversion is achieved by minimizing the roughness of the model i e the variation of the model parameters between neighboring points together with the data error As a result of the constrained Occam inversion a smoothly varying resistivity thickness model is obtained The roughness and or the fix free status the parameters can be set manually to allow discontinuities and to incorporate a priori data in the model Please be noted that because the model is one dimensional the pseudo section resulting from the inversion does not represent true 2 D resistivity distribution Two source receiver systems are considered a vertical loops 1 e horizontal magnetic dipole
33. ts please inform me 5 1 References Anderson W L 1984 Fast Hankel transforms using related and lagged convolutions ACM Trans on Math Software 8 344 368 Christensen N B 1990 Optimized fast Hankel transform filters Geoph Prospecting 38 545 568 Frischknecht F C Labson V F Spies B R and Anderson W L 1991 Profiling methods using small sources In Nabighian M N ed Electromagnetic methods in applied geophysics Volume 2 Applications SEG p 105 270 25 Keller G V Frischknecht F C 1966 Electrical methods in geophysical prospecting Pergamon Press Pirttijarvi M 2003 Numerical modeling and inversion of geophysical electromagnetic measurements using a thin plate model PhD thesis Acta Univ Oul A403 Univ of Oulu Pirttijarvi M amp Lerssi J 2006 Laterally constrained 1D inversion of airborne electromagnetic data Abstract B020 Technical programme Near Surface 2006 Helsinki Pirttijarvi M 2009 Gravity interpretation and modeling software based on 3 D block models User s guide to version 2 0 University of Oulu Department of Physics lt https wiki oulu fi download attachments 20678029 Grablox2_manu pdf gt 5 2 Terms of use and disclaimer You may use the AEMINV program free of charge If you find the program useful please send me a postcard If you publish results computed with AEMINV please provide a notice AEMINV University of Oulu without the quotes When referring t
34. ttons are used to manually edit the position of layer boundaries and resistivity values When thicknesses are edited the vertical position of the mouse click defines where the layer boundary should be positioned Likewise when resistivities are edited the current color scale appears at the end of the pseudo section and the vertical position of the left mouse click defines what resistivity value should be given to that point Single left and right mouse button clicks are used to change model values point by point If two points are selected two left mouse clicks are made then linear interpolation is made between those points If multiple points are selected with left mouse then spline interpolation is applied through those points If a single left mouse click is made outside the profile eg above the color scale then layer boundary or resistivity is reset to the corresponding value along the whole profile and the selected value is updated in the default text field 3 Interpretation procedure When the data are read in also the measurement system VMD vs HMD loop configuration in line vs broadside loop spacing and frequency or frequencies are set The loop spacing should be checked and corrected if needed To validate changes made to the text fields one needs to press the Update button Initially the number of layers is 2 which allows separation of conductive overburden and basement When single layer model is used the Z wght para
35. vely See the example below XYZ file X Y H RE IM LINE 100 35516 00 70982 00 24 27 16 35516 00 70983 00 30 9 1 LINE 101 35616 00 70981 00 29 37 26 35616 00 70983 00 31 11 10 4 3 Layer files The parameters of the layered earth model can be saved into a column formatted text file LAY and The header of the file defines the parameters used in the inversion and the normalized root mean square RMS values of the data and model error Along with the model parameters the file includes the fix free status O fixed 1 free the roughness parameters 0 smooth 1 left 2 right and 3 totally discontinuous Note that to fit the example on single page it has been edited a bit 19 Results fro Lagrange sc Im scale i Filter leng Re Im hide Z elev weig Rho 1 defau Rho 1 weigh Thick 1 def Thick 1 wei Rho 1 defau Rho 2 weigh Rho minimum Rho maximum Thick minim Thick maxim Susc minimu Susc maximu Susc defaul Susc weight Number of 1 Number of 1 Data RMS er Model RMS e X Y Dist Number of p 1654 8299 0 1691 8303 37 1728 8306 74 m AEMINV the interpreted 2 layer model ale m re th value ht LE t ault ght 1t m SI E Thickness in meters ayers Computational parameters 1 0000 1 000 2 10 000 0 00000 10000 0 1 000 50 000 1 000 5000 0 1 000 0 1000 100000 0 63 000 200 000 0 000010 10 00 0 001 1 000 NL 2 ines NOLIN
36. vities R R2 and R3 and thicknesses T7 and 72 Depending on the number of layers the extraneous text fields will be invisible The local Lagrange multipliers are used to control the lateral variations of each individual parameter Increasing the value Li gt 1 0 makes lateral variations smoother and decreasing the value Li lt 1 0 allows larger lateral variations Typically Li values should range between 0 1 and 10 Important Setting the Li value equal to zero will exclude that parameter from the inversion which is essentially the same as if the parameter is fixed for all data points Default button is used to reset all model parameters to those in the text fields above The four text fields below Default button allow the user to redefine the minimum and maximum resistivity values and the minimum and maximum depth in the pseudo section The limiting resisitivity values shown on the color scale 10 base logarithms and the depth to the bottom of the pseudo section are also used as the limiting parameter values in the inversion In other words the inversion cannot make the resistivity of a layer smaller than the minimum resistivity value shown on the color scale and the thickness of a layer greater than the maximum depth Note that initially the minimum depth is defined by the flight altitude Zoom Pan button allows zooming in to a section of a profile When applying the button the program goes into editing mode and the normal
37. wo layer earth were optimized This gives the same result as the first inversion if the resistivity of the top layer is very high Bottom The resistivity of the top and bottom were fixed to 10000 Qm and the depth thickness and resistivity of the conductive middle layer were optimized Depth m 0 1000 2000 3000 4000 5000 6000 7000 RMS d 0 04249 RMS m 0 02052 L E L Depth m 0 1000 2000 3000 4000 5000 6000 7000 RMS d 0 02871 RMS m 0 02813 1 li 1 1 li 1 li 1 1 1 Depth m 0 1000 2000 3000 4000 5000 6000 7000 x0 3551654 Distance m xl 3558660 yO 7098299 yl 7098128 RMS d 0 11614 RMS m 0 03427 23 Roughness Resistivity Qm Mage care off i 1 0 1 2 3 4 5 cesos Height Skin lozig Appendix D 2 frequency AEM data Combined inversion of two frequency 3 124 amp 14 363 kHz AEM data coplanar HMD s using a two layer model Note the presence of conductive overburden at the beginning of the profile Imag gt Real The separation between the overburden and basement conductors could be improved by setting discontinuities at selected points HMD Broadside AEM interpretation Loop 21 36 m RMS d 0 02432 Line 461a Vimpeli RMS m 0 00152 1 1 1 1 el 1 fi 1 f 1 a 3000 y Freq 3125 0 Hz Realm i ie x Real c 20 7 Lx T x x Imagm A F E 2000 4 re Image E g X rate i F i 1000 g e i A ee SR
Download Pdf Manuals
Related Search
AEMINV armin van buuren arm investment aem investor relations aem investingnote ame invoice aminvestment login aminvestment bank berhad aminvest research i3 aminvest fund price aem inventory shortfall arm investment center aminvest market outlook aminvestment bank bhd aminvestment services berhad aem investor