GPHYS a) Geo-Electromagnetism James R. Wait Inst. Geophysit u, Meteoralos as gfe Universitit Kan 5 Kain 41 ‘Alborius-Mpnus-Pate 5948 Seen aioe sy Francis Silo Pauley Swdnew “Tokwn Terame‘Spm © 92 v Aeasene Paes, Ie EE AINS, va fal do ten aed SOE ES RE oon ur ‘bay of Cones Catia Notice ite spon enne nse etrennenane Contents Profs is pte 1 Earth Resistivity Principles Intrticion Poti abou a Single Coren Eleroe for a Two Layer over Fr Elctrode Army 7 TireLaper Made! Frain" 9 Disccon of TeesLyer Cte TT Mmnereal and Ariel Approaches 1 Ells of dnistrpy 8 Cant Soce Eeition of Layered Ansari Region 17 Gurren Source snoop: Medi and Exton 20 Ouaereping Vere! Dike Model 23 Warble Ress ih Dapt--Espeniniel Model 20 arleResy wih Dath—Geteral Case andthe Keel Function 28 Rest for a Welge Sle 28 Lining Cae of 180° Wee 32 Image Rel fr deat Wedge 35, Sphere Prien 3d ‘espns of Homogencnt Convcing Splere dt Carew Pot Sanet and Cier Prolene 48 rv Elcom eo Fidel Borcole 47 Grn Eletrode in Wal of Layered Tuma 42 Lining Cae for Extra! Coron Exctation ofa Cable $2 Cable Lected n Rese Half Space 32 Spleroia! Problems 36 Resstace of Grom Eberads 6 Reforncer a“ ‘coments Chapter “Theory of lnduced Electrical Polarization (Quasi-Statie Approach) turin os ot Tae Reese 70 tr Pome largo” 7p Ie tre et Bed Cte Spee Pree 7 epi of ae Porte ad oe mtn reece aes Rein an Ptah 8 Tn Carte She 9 ep of hb Ser Nategor Pace hte” 56 (Chapter I letromagnetic Induction and Loop-Loop Coupling Imran 108 Magne Pl ofa Car Loop of Curent 101 Greate Lap oe Twosaner Earth 104 Exemion wo M-Layer Medan 18 General Orenain of Sauce Diple 110 Expl Fld Esprton fo Veta ond Marzo Magna Dipole Sates 13 Analog de Reig 114 Low-Fregeey Reynce 116 Apmpione Lining Fara 19 ‘Mined tpedence Permeation a4 Homoperea HolfSpce Model 133 netson of Diplacoment Curents Day General Soe fr @ Hercouel Magnerie Dipole 129 Etecrmapt Inet na Contin and Prmecbe Sphere 138 Bandy Conti Cant Sl end Co Banda 135 Qhapier 1V Electromagnetic Coupling for Grounded Sources Invatucion 149 Some General Conidertions 44 Antrpe Bedi Conidae 145 zontal Elecric Dipole Excitation 14g Harizmal Eerie Dipl vr WLaverAniomope Mdm 133 Atl Inpedance Formation 189 Refract wu copter V arth Conduction Induce by Vertes! Curent Sources lnratason 186 Ta Sobtion or Rais ple 16 ‘Quote Lint nd Psi al Space 168 Fecal Grodan Souee Ht Garnet Sauce Btte 172 ann Coping Breen Grd Fine Cin 181 Reece as Chapter VE “Magnetoellaie Theory Imroducion 184 Bane Plane-Wowe Mods! 18S Cuerent Shes Source 198 Tie Source Mode Romar 198 Sogmened Overbodon Model 197 Verabe Subsurface Inrace 202 Cnc Roms 208 Rerocer 208 Chapter VIE Inductive Transtens Ieratucon 200 Pie Excel Lp Layered Gromd 210 te hi Ct S Honogevene Earth oH Space 2 ‘ied Wane Eft Be Soo en 2 ‘Sesrg frm Barel args 2 Bund Spon Haya 23) Fal Comment on nd aM 302 Reece 22 Append A Electromagnetic TM and TE Spherical Waves Invoivcion 238 The Detye Porntale 236 General Surface npedance 239 External Excaian of Sphere 242 Conconcaly Layered Model 243vu con The leave Sain 2a Dipole Exctan 249 The Quas-Sta Lint 296 A firtar Evtemsion 36) Parnes 261 Appendis B Note on Bessel Functions Integrts 262 Index 68 Preface “The purpose of this volume isto provide a survey of the analytia! bases of electra! prospecting methods. Although the emphasis is on theoretical oneepts, many applications are discussed. This book presents derivations that ae for the mos pat sal-containad, at consistent and uniform leve! of dificult, As a primary prerequisite the reader should have an under- sanding of state and dynamic elctcity a the undergraduate level. All fhe same, many of the basic electromagnetic concepts and the special ‘mathemalialfunetios so common in this subject are eeviewed in the tex. ‘So tha the student will not get Tos in mathematical detail, many ofthe cxamples are shown in graphic form. These run the gamut from diect- ‘current earth testing Wo rade pulse probing ofthe cath’ surface. A special tffor i made (o show the unity of various eleteal and electromagnetic methods of geophysiesl exploration. Related topics in mining technology ‘nd (elecommonications in subsurface environments are also include. "The material is sufiient to occupy two full semesters at the First-year sradvate level, but the book i also suitable fr seniors who have taken or fare taking an undergraduate course in fields and waves. The exercises that fare sprinkled throughout she text have been clasetested by seniors and Tirscyear graduate students in both elecrical engineering and geosciences Ip addition, each chapter is supplemented with references to the current journal iteratur Tram grateful to Jeri Bacon and Joann Main for typing the manuscript, to Bob MeDonald ang Martha Dempsey for drafting, and to my protes- Sonal colleagues Dave Hill, Don Duuley, Allen Howard, John Sumner, Minne Nabighian, Jeff Hughes, Ken Spies, Charlie Stover, Jef! Lytle, Catherine Levinson, Khalid Nabulsi, Margarete Ralston, David Chang, ‘and Roy Mallson for hep slong the Way. 1 also wish to thank Academic Press for its expeditious handling of the manuseript and for important editorial suggestions and commentsonaper 1 Earth Resistivity Principles INTRODUCTION ‘The interaction of electromagnetic waves with the earth ean be very ‘complicated, even for ail simple geometries, Fortunately the situation is tnuch easier to handle when direct current (ds) excitation i= used, Thie Timing ease of 2210 frequency is an excelent starting point for our analysis or reasons thal are discussed later, we shall deal with a four-lectode system, Tha i, the curont 1's injected into the earth via two electrodes {ex metal stakes) and the resuling Gels ofthe medium are detected by two adjseenteletrodes (eg, metal pins) (Fig. 1), Nermally the voltage between the two potential electrodes is mensured by a detecior whose intemal impedance is much greater than the impedance between the sleerade contacts and the earth medium. "The quantity of interest isthe wansfer impedance, defined as Z= ¥/1. In the ease of de, Z becomes R, the wansfer resistance, and has dimen ions of ohms, As willbe shoven below, is proportional to the apparent resistivity’ 9, of the earth medium, That, R ~ F,, where Fis a factor that has dimensions of length. The mame of the game is to interpret the ‘measured p in tem ofthe geologzal suructure Rarely is such an interpre- {ation straightforward, but much sight i gained by dealing with idealized ‘models, atthe same ime bearing in mind their inttions. ‘When dealing with static or de current flow, the resistivity p is the reciprocal of the conductivity 3. In what follows we shall deal with the later. For example, we shall specify that the eonduetivity a(x, 9.2) 6 @ function ofthe coordinates. Furthermore, we sal assume thatthe earth i inotropis then Ohm's aw lakes the form Saat, a2 1 Ee Rest Peas iz gb Armnganct far arch sity ig, where J isthe vector curent density in amperes per square meter and E is the vector elect field in vols per meter. Now the electric fields can be derived from the gradient of a scalar potential y. Thats eon a ‘Anotter basic fat is that sivs=0, ° which ols evenywhere except atthe source itself, ‘Asan exercise he reader can easly show that Pe FY Af ab ae | ax ax A case of some importance is that in which the conductivity as) varies conly with depdh z. Then the terms containing do/3x and @0/3y jn (4) ‘atih By introducing eying eourtinates (2 wherer = 4), WLM, wre ae tt He gee HS Heme oO sujet to azimuthal symmety, This equation may be solved by sparaion of varies To th rd we wie 72) = AUz(eh wich ca prot of «function of rand a function of ©, Then (3) separates into two equations that must be said simulaneoualy, = eR e © and #241 ato o here A isthe separation constant, which up to this pit is arbitrary, ‘A general solution of (8) can be wetten in the form ¥- [rokacnzaca, @ Potentials abot» Single Curent eco 3 whete FQ), function ofA can be chase tft the oundery conditions as Ineded. Also, at we shall se Inter, the coatour over‘ can be taken along the real axis POTENTIALS ABOUT A SINGLE CURRENT ELECTRODE FOR A TWO-LAYER EARTH MODEL 'As an example of the previous formulation, consider a two-layer earth model (Fig. 2). The resistivity (0)! or ay is constant for O< 2 0. Here 100) he garden vay ofthe Rss anetion ofthe fist ‘ope of oder aso and denotes the intvationvaable By no accident, A can be rested to the separation constant hal vas inoduced eal Equaion (12) can be verfied by noting that bot sides satis Laplace’s equation ad that theres fo fhe int 1-90 isan elementary fot Returning to (6), note that Rr i ealy J) oY). But ince the tee, the Bestel fnctio of the second types aint for r= 0. exanat appear inthe present problem. Thus we are ld To wie the resalant atta athe upper lye ber [Lore ™ + Berean (a) for the region O-< : 0, The righthand side of (13) clearly susie Laplace's equation v=o, “9 88 il should for a position-independent resistivity py, over the limits O<2 Ws cosy een ta have the form ‘oy ® wre the ulplyng factor i chosen for convenience. The fanetion C{R) is yet to be detemined. Note thatthe factor «"™ cannot appear i the potential is to be finite at 2-00, ere wo B [Peay raya, (15) Potente abou 9 Single Caron Eecrede 5 [Now apply boundary conliions to determine the unknown functions 4, 1a, ard ©. First, the vertical curtent density must be ze al = =0 for de r F/O 09. ey) Using these values for A and B, the potential (2) anywhere inthe upper layer is thus given 88 an integral with a specified integrand, This solution ‘vas apparently Tis given by Stefanes ota, [2 (For related eeferances, see [5-10},) "An important special eas is thal in which the observer ison the earths surface (ie, 2= 0) Then 4.0) = (In./229GU0 K) 2) where Gey 14240 f (6) Cleary, if the thickness of the upper layer becomes suiicenly large (Le, nso rh then GU, K)-> 1. Another special ease of some importance is that in ‘which the uppelayer thickness is very smal (<7. Then Genk) t+ PRE (Ahr dd en‘ 1. ah Reiter Pence ‘The integral here is equal to 1/r. Thus G(r K)A1+2K/(1~ KY = psf 8) ‘The potential Y(,0) is thus given by fpy/2er, which is the appropriate value for 3 homogeneous halfspace of resistivity py. Although the integral in (26) can be evaluated by numerical means, itis useful lo abn series form, To this end, we expand the facto inthe integrand as @) Then (2) can be writen as GK) = 14247 F Ky, 09) svbere “ 1 oh [en pa In obtaining this ceslt we have interchanged the order of integration and summation, This is permisble heccuse (29) is an absolutely convergent ‘The series form for the potential has a clear physical interpretation, As indicated in Fig. 4, the potential at P(r,0) results from the total series of Jmagesourees located ai = (n+ 1)2K, and the seength ofeach ofthese is 2K" ‘Aa altemative and posibly more general image picture is to lcale the images of strength &* both above and below the eath’s surface, In this fee, the observer at P(r,z) could be anywhere within the layer 0 2 < h The proo! ofthis statement is left tothe reader ig 4. The image epsntaton fr 2 singh cutest pis soe ate aie of wel cas wn tenon a te ace, i | | i ‘canara Four eceade ray 7 GENERAL FOUR-ELECTRODE ARRAY In practical situation, as we indicated in the introduction, the eusent source of +/ amperes must be accompanied by a current sink of —T amperes. Also, we nocd wo potetialeletcodes, ag indicated in Fig. 1, The voltmeter then measures the difference # between the potentials (4511. “To allow some generality, let the current and potenti eectodes be located arbivariy as indicted in Fig. 5, in which a plan view ofthe wrfuce = Dis shown, The resultat potential y a P, 6 toy 6 Geran low 2 ca *} oe wee ry (+) ad n= ag JV Th el pole ty ai ve ey where rg = (33-4 yD and a3 = [lay 99? +93, “The volige V Between the eloeirodes is naw given hy V= W519) ~ i099) (4) Now. ithe ground were fully homogeneous (12. p: = p, and/or hc), early the Gs in (32) nd (33) would be I. tn this ease Jovy vary Bb | 05) But if the ground is inhomogeneous, we ean define an apparent cesitvty 1, by weting (35) shaming he ton fhe ee sre an snd the paca decodes Pd Py8 1. at Resi Poneos “Thu, a he rik of being pedanti, we se that rw K) Geer) , Genk) ~hode on ” {san explicit expesion for computing the apparent resistivity as a fonction ofthe geomeny of the uray. {An immediate application of the foregoing is the Wenner array, in which the four electrodes are arranged in a stright line with equal spacing. AS indiested inthe inset in Fig, 6, the potenal electrodes Py and Py a seaddled by the two current electrodes. In this case we see that ry = a9 Ind rig fai = 28, whence pa/1=26(@,K)— G(22.K}. ‘where the function Gis defined by either (26) or (30). The function o4/a, {or the two-layer madel is show in Fig. 6 os'a funeton of the relative spacing 4/1 for various values of the faetor K. As indicted, the apparent resistivity, i greater than the ressvty ofthe upper layer when K > Ty this corresponds ta the casein which the resistivity py ofthe lower layer fs eter than p,. The converse situation applies when ps less than py. In both cates, p, approaches pa the upper-iyer thickness hecomes much Fly. & Apparent restive cur for twotayer enol sing Wenner method, ‘hvso-Layer Mosel Formulations ° 20) 4 [ne Fn 7. Eales amon ne mol or twolar esti polem. greater than the spacing a. Such curves ean be used in a rudimentary Interpretation of electrical sounding data when the geological srustre is simple 2.3.48) Tis seo rewrt the son 25) ofthe woryer problem as ln fy pm = po-2 (0) = 5er GK) = Fe (SRE ayaa, 8) where = — K=(9,~ 9/9, 99). Bat we ea a0 te R= (4) n9/ Gay ngs where m= ‘ons. A transmission in (G85 and! 3, Ap, have dimensions of impedance in nalogy then suggests ise? if we Farther rewrite 0a, a aie ee aye me ead = aad + ag, MAHON «0 Tan Fa a ere 2,0) ean be interpreted a the input impedance of a transmission line of charters impedance and length that tertinates in an imped- ance of 2 olims (Fig. 7). The propagation constant A of course has ‘imersion of reciprocal meters. Note that the impedances 9, and, are reilly pure resistances in this cor, a least forthe static limit in which py “np ave real. The propagation constant ofthe Hines also cel ZQ)= dt eRe ‘THREE-LAYER MODEL FORMULATIONS ‘We now generalize our model to thee layers as ustated in Fig. 8. The -lasscal solution might proceed as follows: For the region 0 <= < h, write Yr in Mlentically the same form as that in (15). For the region hy <<10 1 re Reet Pras ay hy ie bo B [Teme + veyed aande ay in place of (16) since both types of wave Tunetions are permite in the bounded intermediate layer. tn the bottom semi-infinite resion = > hy a. In to= Bt [eae aaa « “The boundry sonon nom equi ht Mg a reo, 8) are hy «sy «sy 46) a) “These led to five linear algebraic equations to be solved forthe unknown coefficients 4, 8, C, D. and E. The solution, though tedious, i staighfor- ward. The solution forthe potential y(r.0) is identical 10 39), but now 2+ yanh(ah) ZG, F Zan) * 4s) where sao mytanb\) Ao MeO) 9) ee ne nenee Discusion of TreoLayer Cane 1" Roe Lenn Lmy—! 9. Eon iomemon ine cna for rene sy problem “The equivalent transmisin line cteuit is shown in Fig. 9 n which the two sections of uniform transmission are connected in series, The propagation constant is A for both sections. The characteristic impedance im, for the Fis section, nor the second sectin, and the terminating impedance 3). DISCUSSION OF THREE.LAY! “To be more explicit, we eam express the threedayer solution 36 er) ape 77 0 3 [ afi -al} 6 were Reo , a ke «6 where — pao 5) (2) ere we have again used the eeu [llanant ry “The expression for Q(r) in integral form ean be computed since the integral in the latter form is convergent when hy > 0, Examples are found in two classic books, by Mooney and Wetel (3] and by AVpin ea. [5] that include many curves. Howover, in many cases the formal integral expressions ean be grealy simplified. For example, ax shown by Sunde (7, 53], the ease ofa thin intermedi layer allows ws to use the approxima2 et Resiy Prices Fg 1, Theor rescue tion exp(—2N,) 21 ~ 20. The (Sh simplifies y= 03= (1+ /osh l= ef oDoaky . 7 DEPT /03KT= w/o: 0c i) ‘Te corsspondng pote function Q(r) is rested to the apparent fess p= SerQUe) for the two secre array (Ley second erent fd potent codes ae at “ining” As an example y/p ib shown loted a unstion of Fi. 1), The parameters are chosen sich that foo I0p and = 10h, Ths we ae dealing wilh 3 moderately resve fowtom layer, whereas the thin intermediate layer cin be either highly Teisve ogy conductive relative fo the lop lye. evident in Fig. 10 that the hah ressiveinermesinte layer more difiult to detest than is th highly condetv layer. nth alr case, the intermediate yer “channel” te urent very effectively. Exercise, Show that forthe ease in which fy hy and p> py and py We fave te fre simpleton a f= 0—Doshe PO mals ©) here py the total transverse resistence of the intermediate layer. Ml [Exetese. Show that forthe casein which hy © end psp, and py, we would get 1 = 05 ~ Nol Ris ase hal (55) omer end Ala Asproaches a where 0) = 1/ay 6; 1/5. and a5 = 1/95 ae the corresponding conduee Ui, Here af isthe longitudinal conducianee ofthe intermediate layer. = NUMERICAL AND ANALYTICAL APPROACHES Unfortunately the integral representations for the potential funetons (ror p, do not end much insight when he layer configuration is general “The numerical data ean be cst in form tha allows general conclusions 0 ‘be drawn, bu tis bouts the seope ofthe present exposition. Here the reader iy referred to Mooney and Wet! [S} and Alpin et al 5). More recent developments are deseribed by Kone [8 and by Koefoed {10}. who ulin an efficent digital filter technigue forthe mumerieal evaluation of the integrals. The limate numeral representation of resistivity is made possible by the computer program Ghandling up to 12 layer) that was eveloped by Caine and Lytle [9 "AS We shal how Tater, mh ean be done to evlist the integrals in series form when one or mace ofthe physieal parameters allowed 0 be either smaller or larger in some relative sense. We shal indicate this in te Context of the theeetaver model under she folowing restitons (2) poy and py (highly conducting lower layer, ©) Iyer he, thin intermediate layer, (6) ps0 ie highly restive mide aye) ‘Then we find without ifcuty that R= (1 Aayiior +n, “) ‘where 9, = ay the transverse resistance. (Actually n, hat dimensions of ‘ohms per square meter) Now wesc that (50 simplifies to va 1 =[(e = ny/(o + ng je“ 1+ [(e = 2a) /(0 +) the subcripts 1 on p and having been dropped. The integrand i now txpanded in pom sere 30 that ann $ (fd But note also that an-£ Jedd, (8) Jom }eana p=, t= eve) el om, 2S TramJe) | *? 2.)“ Sc si con u/s aT i ew ean a( Me (FR) = (5 “where neglected terms are of order (Am,/a)". The result given by (61) is now Basaran 2607) ? wy saunane 2S (1) (“eM vynn dd 2m Sayre (*er™asyan 28S cua ftemnsonan (oie (em (a2) “Tes ntpae canbe eve in lose form. Fist not that [lorraine ry ere Which isthe same a (12. Here a and b ate two positive real constants ‘Suosesive differentiation with respect to a yials ihe wo needed forms [reserve wa os) and [Pertonran 288 ws) Using thee esa, wie (8) 38 27000) hey 2S -1y" 1 _, > TO) Tame “a MS Lindon (eB TS aire 2(2hn/ n= 12 Denier tye) ays + 2) S(-y (5) 0" Tr mehy eh tens a ok. The it wo ms i exes Neglected terms are of orer (a2 The Fit wo ers fii expanson tat af Antony 6 ‘bottom layers perfecy conducting, The succeeding terms inthe series con ‘be Hentfied as corrections to account for the tia resistive Film er coating. ‘Obviously such a series expansion is most useful when «rp: then only the linear term a (nr need be retained. ‘Actually there isan ineresting image interpretation ofthe above results ‘Te ilustrate, eetura (0 (59) and note thet (Faxe)==(-2) o ‘ot sides of (67 yield the right-hand side of (61) up to the order indicated, ‘The integrals in (9) are then ofthe form fre mf2(e + 3) Jone ” Tanti (nde } (68) ‘This result shows thatthe images have the sme character as that for perfectly conducting lower layer, But thet locations are modified. They fceur at = = 2n(+ (ay/p rather than at 2 EFFECT OF ANISOTROPY More often than no, the exeth media ere not isotropie conductors. In a macroscopic sense this means that Ohm's law as given by (1) should be ‘modified. In the usual situation the conduetviy 2 in the horizontal Uiretion ie dfferent from the canduetvty oy in Ue verial direction, THUs for present purposes we write J ne or Bao, and of, oF d Bee wince and p, ae the corresponding resistivities. Te wil be usefl to digress here to consider a simple phenomenological ‘model for the anisotropy. To this end, consider a cubic meter of the fubsurface at a depth and location at which the eurent flow is reasonably teiform, Within ths elemental volume the medium is made up of N parallel sheets of thickness T with an isotropic conductivity e, and thin Intervening layers of thiekness W with an isotropic conductivity 0 (Fig 1, Fist, apply a volage E, s0 that current flows from right to lft6 1. at Rei Petes ==. == — Fig Vann of on ans cond flayed moti, Clearly, the total current flow is N(ayT + 01), amperes, where N i the ‘umber of layers This eutent flow isthe same as No\(T + W’)By, where ithe elfectve or average conductivity in the horizontal direction. On ‘ating these results we deduce that = H/o = (pT + 9b )/(T + W) @) and P= aed T + W/ eT + 0) ) [Now insta verte! eurent J, tha flows from bottom to top. Clearly, the tol voltage is MoT + o:W/\J,. But this must be the same as NaAAT + Wh, whore pis the average o lletve vera! resistivity. Then 2. = 0, = T+ OMIT) a snd thus mod T+ W)/leT + 0,0) m In general, we see that 6,» 0, oF my py- Alto, itis an interesting exercise to show that T= Wi, ete Oahu ALS" 8) 3) where 8 01/0. pal co) ‘The quantiy on he sighchand side of (3) is never less han 1 tfc, iis soles fc to show that oa + ([@-l+8)-2] TWH Wy). as) ‘Again, this ratio is never ess han wait; i sequal to unity when 3 = I, oF when T= 0, of when W=0 (ie, in each ease the region is rendered isotropic). Ay ‘urent Souree Exctaton of Layers Aiscopic elon ” (CURRENT SOURCE EXCITATION OF LAYERED [ANISOTROPIC REGION. Now consider the following problem. As indicated in Fig. 12, current saree or elect pole of araperes is located above a ewoayer anisotropic fonductor. Ta terms of our usual coordinate system (F.2). the upper Inotropic region of resistivity occupies the space : <0, The intermediate layer occupies the space 0 < £ it we replace Y by W's Py By ae and p, by (Of course, in the upper isotopic region = <0 we replace y by Yy and bot ay and By ay ee aa mA ie 12, Caen pe suc in opi ron ne woe nop ln.1" 1. anh envy Pree ‘To posed, observ that sltions of 8) have te form eo =I2]sAr. where = (o/og"® (80) Taking ee fo he eles warp aye stature i flows tha he stable epreentaton for the ptental faneto are ton feet fr oweie sara for <0 (aly af [eme + Daye ™]yy0¥)ad tor O<2 —z and the pls sian isto he used when =< ~ Since the integrals areal ofthe same form, the boundary conditions can be applied in straightforward manner. These boundary conditions are that the potentials and the normal current density be continuous at the interfaces. Thus wow 6) and vee Le week ay When these are applied 10 (81), with the help oF (84), we find that (ot /4aye4 Dy= C+ D, 7 = Ddl/oe)™ CU/9)~ BUW/B. (88) co De Ce, 9) CCA = DUA fae = CL/pIE™, 490) ‘curant Sues Exciaton of Layers Anionic Reson 9 hese B= (nan) 8 (ogi)? and he = kh = (af). The solution ofthese four linear simultaneous equations for Dyy C.D, and C’ is pure lpebra. For example, i isnot difielt 1 show that y+ Reng" Phy + Rey where R=@-pyp+y (92) ‘When this result is inserted into (81), we have an explicit result albeit in integral form, for the potential in the region z <0, The percepiive reader Wil ote that this result i identical to tht for the eorrespondig problem Shown in Fig, 13 when the intermediate layer is of effective thickness, With an equivalent isotropi resistivity f. In this analogous problem of ‘Stuation the bottom region has an equivalent resistivity Tn other words, the situation shown in Fig. 13 is the same as that shows in Fig 12 insofar san observer in the isotropic region: < is concerned. ‘We now specialize our resul inthe sense that z, tends uniformly t9 zero ‘Thus the cirrent source 1 isin the interface. The expression for the potential is now wnt [00+ Bae Nand for 230, 63) where Bud) = —(4/ pel \Dud)- (04) ean Re Rey 1+ yay = 202 Be + Re (95) pe BO Bes Re) he Sa a a tot - (EC nit wan nc rpm roe tos mo ih20 1 anh Resstity Pres “The next sep is ole ng botome very large compared with and 1a Reo eNO da. wo Be 6) (nthe nets ul we would ave H¢r9) = (IB/2=n1GUe.R eo This of course is completly analogous to (98) In similar fashion, an Apparent resstily p, cam e defined fortis case By palB= 2erQ(7/. 8) where G0) [fanaa . See “This presupposes thatthe second current electrode and the second potential slecfode are both a infinity. ‘Our transmission fine a om 1 follows immediatly if we write L (agi a, wr)= (100) where 20) = mA — Re + Re), 01) As indicated in Fig. 14, Z(\) is te input impedance of the equivalent vansmisien line whose characteristic impedance is, and for & termina- {ion impedance tn this ease, Ra -a G48 where 9 = AB and = a 02) ‘rig Mh. Fevivalent tamsminson ne eireuit for a two ayer azotopic model, Ccuran Sure in Aol Mesum ord Extensions a (CURRENT SOURCE IN ANISOTROPIC MEDIUM [AND EXTENSIONS {s useful to consider dhe spacial ease ofthe layered structure considered in the previous section. This is illustated in Fig. 15. in which we now have too half-space regions. The currest source it located at 2 =~ 2p in the ‘soeopie region. inthe two regions are ettersn 4 BOPe gon Bom peneoal add tor 2<0, (103) ana BE [reoenanan for 230. (104) In (103), the sign is to be used for #4 290. Here B= (yp,)"?? and = (./o)", Tse results can be checked by nating that y ~'Yo at 0 and (1/19/82) = (1/yNhiy/¥2) at == 0. The integral in (108) is oF the same (ype as that in (12) Thus we may waite oi 1 . 105 RG [Perey 0 ‘A interesting ming ase iso let 24-90 and 9-2. Then i gee 25 Pee for +> 0. This situation i illustrated in Fig. 16, ia which we have flipped ‘he = axis upward, Cleary, this situation must be identical to the casein which current source of 27 is located in full space characterized by p, and (105) fee Cornet sean in tnt ne,2 1. anh Rest Priciae faa Fe 16 Ansa altace onde yam so AR AA Fa Pallant fy Thus we may assert that forthe Ful anisotropic space, us indicated in Fig 17, de potential is given everywhere by con Although this expression can be derived by other methods, this evolution fs saying ‘To continue, we now extend our analysis tothe cas of tree anisotropic regions as shown in Fig. 18, The primary potential fn the region = <0 is now given by B fremerangeya, Where the % sign is (0 be, used according lo zo fe20, where & = Gd) and V2" Here py and, are ihe horizon and ‘eral resistivities forthe region 2 <0. TE AA am ‘Fie. Current point source in aniztrogle region over a two-tiyeranliotople region ‘ouerepnng Vera! Oke Mec! 23 et-suganan, nytt Bey ce «sy where = t= (oy/p4) and R= 5 — BXG +7)", Similar expressions ‘an be dbtained for the potental functions in the ater reeions. OUTCROPPING VERTICAL DIKE MODEL ‘The geometry of she problem in Fig. 18 sugassts thatthe seslls are relevant to borehole resistivity measurements, The = axis inthis case the axis ofthe borehole. The results tha we have derived are decty appica- ble the intiuence ofthe drill nole fluid does nat have a major effect (We shall deal with this type of problem later) An interesting property gf the problem shown in Fig. 18 i thatthe feds are fly symmetrical about the plane x= 0 when th solution is expressed in terms of rectangular coordi- ates (x, ).2). bere (2 +9) “The transposed problem is shown in Fig, 19. The expresion for the potential in the region = < Oi identeal to (108) if 1/4 is replaced by 41/22 Tor all values of ¥ and J. OF couse, ow py the resistivity im the (Gerical) x direction and (horizontal yciveetion, whereas p, isthe resistive ity in the (horizontal direction. ‘Working formal And numerical examples for vertial dike sirgtures Frischknecht [II], AS these authorscy 1. anh ResetyPrisies indicate the apparent resistivity curves have discontinuities as the cece lwodes are moved across the vertical contacts, VARIABLE RESISTIVITY WITH DEPTH EXPONENTIAL MODEL As indicated in the Introduction, the laterally uniform earth mode! may be analyzed even when the resisivty or conductivity i ¢ continous function of depth. The relevant Z(Q.2) funedon is a solution of (1). A ticularly interesting case in which this situation aries that is occuring een inthe earth's erst or basement layers, where increasing temperature can be associated with continuously increasing conductivity. Such 8 prob em was considered by Stoyer and Wait [12 whose model wil be described nel. ‘The discussion here will fer to the model shown in Fig. 2, with the ifference thatthe resistivity in the lower layer is special fonetion of depth; specifically, Ate) = (ou/e)exp[—B(2-M)] for he z 0 that we imposed carer Thus we are led to write =F lomen ca ty tal a nae LAL B a 7 wir (aly26 1 cro Resseey Pres where we have used the integral formulss (see Appendix B) [Prenat (2) p [lors eiyanmeb+ "ora (ay a wont (fem s Bla aan whete the upper limi is not specified, The result can be expressed conve~ bietly in terms of the exponential integral E(u), of argument u, defined by B= [* (235) “Then it follows easily that noes 2 Bl 8 we Ie pr] + ont (126) “This esul ie valid for any positive or negative real value of 8. “The logarithmic term in (126) ells for some comment; itis only present wen fis negative. In fal f — Ar I this term dominates and the result for gy becomes eA = Ln oh ne sons, oxy where f,= —1/ is positive, Now aote that the coresponding tangential etc fla ie E,= ~84/ir = Io /(2ah). 2) “The corexpondng rail cutet a distance rit 1/07) amperes pet tier where h/t ete stage conductance Of coe 710g erm in (126 does ot appear, the Fonetina dependence with site leer Te we ate (18) nthe form $6.0) = Un 22g (3) re Ge fs tnd hen the spparent resi pcan be computed irom (foray parcilarfourlestade array. For Sample, we mit Comiger the Wenner ray, whish asthe gooey show the nse in Fae 6. The snares reat p, mormalicd by she resitivity ry of the Vert Rett its Deth—Exponenl ode! 7 Fg 20, Nomad apparent sity vere gms ste for Wenner a. vanaf arson on caves (2 psi “upper lyer is piven by by =2G(0) - G(20), 30) where ais the common electrode spacing. The numerical work is simplified ifthe variable of integration in (114) is changed to.2 dimesioales Ferm, For example if & > 0, we replace A by A>. Thea without difficulty we find form of (130) reads (rene [Jena ~ Je(2nBa) Je, aay [Bm ith + (+ DA Results, based on Stoy’s calculation for = 1 are shown in Fis. 20 and 21, in. wich p,/oy is poled asa Tunction ofthe normalized spacing fio for various vauos of te parametcr fh The data forthe limiting case [Bh =0 (ie, absence of overlying homogeneous layer) are consistent with “80 ig. 21Necnaized appre rsty ws nomad seat or Weer ay. alisaf Prac shown tm cues 8928 1 art Ress Preps ‘he potential form given by (126). As indicated in Fig. 20, the apparent resistivity generally decreases with increasing electrode spacing when 8 >0. However, the elfet of an overlying homogeneous layer isto delay this eventual decrease. Coresponding remarks ean be made about the ‘curves shown in Fig. 21, in whieh 2 <0. VARIABLE RESISTIVITY WITH DEPTH—GENERAL CASE. AND THE KERNEL FUNCTION; ‘The next obvious extension is 1 allow the earth conductivity to be a eo ‘model shown in Fig. 2 and allow the resist to be general in the ‘eon f-2 2 , we have nay (Ar) da, Sr awena ow No atthe Foner Soo instore es. i P= f° FRI 00a, aay e.0) hea J0)~ ["Feys.0vyde ay fer Boel aretions J fhe ist ind and order Thus if we tke m= 0, ites tom (40 thas By IE Seer oteanar Fay 7 tag berry. (043) ‘Ths rent shows that at es in pipe an interaton ofthe measured Polen ver yells the kare function 1/fA). Tis fonetion of oui elated tothe ess depth nein (2) Farther dseasion of the concept of direct interpretation ofthe Kernel fureton found in the oral pope By Sle 7) and more recent fone by Pasa 1, 19] and Roma 1 RESISTIVITY FOR A WEDGE MODEL When the resistivity ofthe earth varies in the horizontal direction, the uniformly stated models are not valid ina strict sense. However, these ‘model may often be applied i the herizonal variation is sufciny slow, ‘but there will elerly be eases in which the lateral nonuniformity is signfeant, For insight ito this eas, consider an idealized two-dimensional30 1. anh Rossy Pras ty Fig 2 Careat pit oer J jen nto weg fan fom oper us aed cio rote ar pit sare Ben Har fa wa of ope 3 ‘mode! in which the top and bottom interfaces of an overburden layer are not parallel lk, to facilitate the analysis, assure chat the bottom region i ‘ther an ideal insulator (case 1) or an ideal conductor (ase 2), ‘The situation i indiested in Fig, 222, which shows a wedge region mode ‘With respect to cyingsieal coordinates (7 ,), te upper surace is a $= 0 for all values of» and = and the lower bounding sutface is at ¢= —y for all ylues of rand =. A eurrent electrode that injects a current I ino the ‘wedge region of resistivity pis located at (90,0) As stipulated, the wedge is bounded at the top (@~0) by an insulator. Thus the situation i ‘lectrically equivalent 1a fll wedge fegion of 29, as indicated in Fig. 225, where the current injection is now 21 inthe bisection plane ala distance ry feom the apex. In temas ofthe same cylindrical coordinate system (F.), the current source 2/ isa (70,0) and the bounding surfaces ofthe wedge are al = 2. The boundary conditions at these surfaces are 3/39 = 0 for ease { and ¥=0 for case 2. Hete V = V(r.9,2) is the potential which ‘must suisfy Laplace's equation everywhere within the wedge region, That terdst Zee Pea ae for crema, Pepe Hand Fos regs & [aoiemntcans einen Fe (14s) where 4) is an unspecified function of A. This representation for V has Finite behavior as r-»0, and it vanishes as ¢—> 0, ait must. Also, Vis continuous atthe surface r= ry. The souree conclion ean be sucinetly (sles Hea) (146) aay [Effewreer oe “(2.404 (w+ 8) 8] EfPeordcad case? (148) where en [} me met Tha tor maha ‘The above states that the eurent flowing out of the volume bounded by By+8> p> fy—AS and —¥-0 and +0, In constructing this represent: tion we have used the indicated spectral forms for the unit impulse Funetions 6(@) and 8(2). Now, because J, = ~(1/o\9V//8x) is the radial eursent density within2 1 arth Resststy Petes the wedes region, me may deduce that Ey (Haas Hla-s) LE [Pra oleae) Krys] cosh orga, “) where Kin) = dk (3) an 1s)» ds), The quant in sare brackets the Wronskian and i equal to —L/vy. Ths the right-hand side (of (149) is egal to 18 (4:ay Lewenean| OY!) {ease 1), PEL 4s eee ar deere one 1 this dual quantity i to be identical to the respective right-hand sides of| (147) and (148), we must have Au) loti (e580) and Ay(A)=2p/mb (ex8e2). (130) FFor ease |, explicit forms for the potential thus read rageyn 8B 50 (| hyo 4/4) 1 SE ete tentang pomiean on ses ronan B32 x aes A000) Kine cr r)eos[ m+ §(0/8)8] Kener sero Moe sanereiDeOs[ + $(6/998] xeoshedh (132) In both (151) and (152, the upper forms are to be employed for > ry and the lower forms are tobe used for 7 ry LIMITING CASE OF 180° WEDGE An interesting limiting ease is that for which y= = and for which the resistive wedge becomes a resistive halspace, Then clearly, $0 &, pal Ini) Kur) 08m 608d Heda) Bef ( (i) Ky (Ar) cosmp eas dA, Ken)hglde)cosmpcosde ds, (153) Lining Cate 0 100" wpe 3 Ova) Ket infr)easine tt eosde dR, (Mme 2H) E05(m +f} ORE AD, (54 for ease 2 TFor case 1 in the Hinting case y= we ea actully show that the potential, as given by (153), i given by = fp/228, (135) where fa (b+ P= meme 2) fx the distance from the source to the observer a (7,442). The equivalence Srihe tignchond sides of (83) an (199) can be contre by using Standard Bessel fncton etites Tn the comet of the presea problem, theft of those fs [20-23] Sg, { fslodKathrd Sx-(eoennon athe the upper form sto be used for > rand the ower form for r-< Tria watement ofthe addon theorem for modified Bese onetions of integer oder. The other esl [20-235 Bb 'As indicated in Fig 2 for ease 1, the potemtial at (F.9,2) is now merely the same as for a source J on the suface of a homogeneous resistive coupe INE + B= 2reoe8)7]- (18) L z as) 13~ trysoss)” oor dh My.0 Wha p 2 Limiing cme lo which wedge age 6 soins Ht haps (ad34 1 ah Resitiy Prins oa parole FREE ee @ termprot Srpody w oF ip. 2 ining ene wich wedge ample J, showing oth lanes) an futspace) forte 2 i, evening the ae in heh a pect coming alpen tia tere or (9) scm ate ste 2 halFspace, here defined by 0 > @ > —z, or its the same as for a source 27 Jinan infinite eaistve space —7 O. This is ‘equivalent to the casein which a current source 21 is located acento the perfectly conducting hal-pane sheet in a ful infinite resistive space —2'<$<=. Ths dual situation is shown in Fig. 24 Although the poten tal funtion Tor this situation, as given by (154), looks rather unwieldy, it can e expressed in seatvely simple form. The equivalent situation was worked out by Lebedev [2], who dealt with the electrostatic problem of a point charge adjaeent to a conducting sheet of se contest of the peometry of Fig. 24 Vigsest) = of 2aR + Ve, ass) where 2rancasts/2) }) a “The original derivation of the conducting half-plane problem in the pres- ence of a siatie point source goes back to Sommereld (21) who used a conformal transformation technique. AS he indicated, the final result can De intepeeted in terms of a primary souree and is image in the lover sheet of a doublesurface Riemann space Ths facts not too suprising since we ‘ee, even from (154), that the angular functions have a basic period of 4z, nage Resists or Aussi Wed 35 IMAGE RESULTS FOR ACUTE-ANGLE WEDGE} 11 is useful to specialize our problem such that the wedge angle is yae/M, where Mf + Le, the wedge angle is a submmltple of Now (151) and (15) rede to 1m tS a fan A Kua) cos{mitgyeonedh (281) (10) and fol Bay] VE DL herr Wad Kme ras) xecs[(o+ 1/2)Ma]eosdedd —(case2). (161) ‘Those forms are valid for r> fy Inthe ease in which ¢< ry, we merely replace by ry and 7, by r. ‘We now find i useful to exploit he denis SE aeotmue)= 5, Stoofele- 22] coe us SE meotins ie] -> Be vreof mle J] eer coy “These can be established by noting thal exe) S! xo(- 238 9) + exp = ir) 3 exn( 200 p) ra 2 J _exolin’g) _L-exn(~ en’ 2 T= exp{= Bem 4p Cig) 1 = exten) I Tepe = Meostn'g) iL m= Mm (m= 0,1,2, 0) -o otherwise, (iss Soe ferences 24-28)36 1 Ga ResoeySrnae and ko eS rea(- sa 2 tf eon s258 2a p) _ s304in'g) “Sel 2 28 + t}op) 7 “_ nccinrs) I mee ming) 1= x= in'/ A+ tet) Teal Om7 Te Dy {in'g) 1b exo(— eam) ST Te expen Fay mim (m5 1/2) = Mooants (165) (On witizng (156), (162), and (163), we see that (160) and (161) ean bo iv Yer {e ‘were the plus signin (= IY to be used for ease 1 and the minus cease 2. Wit reference to Fig, 25, we see immediatly that (166) core sponds toa finite set of image contributions. These are ilusieated forthe ‘ise in which y= 7/6 and the eurent source i on the upper surface. ee eee | | nape Rass or Aout Wedge 7 “The images all have the same sign when the lower surface ofthe wedge is bounded by an insulator (Ke, case 1). On the other hand, the images iterate in sign when the lover surface of the wedge is bounded by 3 perfect conductor. The sme image pictare would apply to the fll wedge of [ingle 24 for the case in which carrent source 2/ is on the bisector. The images in that case are all nultpied by 2. This ease isnot ilustrated, but the reader can surely visualize the situation. Exercise, Verify that (166) satisfies the boundary conditions. 11 is evident from the preveding development that asthe wedge angle becomes vanishingly small (Je, )>e2), the number of image contribu- tions becomes very large. The solution inthis ease becomes identical to that “derived forthe planar geomet Exercise, Repeat the analysis fr the situation shown in Fig. 22, but allow the souree J (Ge, eurtent electrode) to be located at =~, where 0 fy < os thatthe source coordinates ef (fy, By, 0). A “The image representation for the potential function as ven by (166) ean the employed in a very straightforward manner to deduce the apparent resistivity for any Foureleetrode aray [24,25]. The physical restition of ‘course are that the fur eletrodes must all be of the suface ofthe resistive ‘wedge and that the angle of the wedge must be a submuliple of =. An ‘example is shove in Fig. 26, in which a Wenner army is located in the transverse plane (Le, perpendicular tothe wedge apex). The results forthe apparent resistivity py a8 tormalzed by the resistivity py ofthe wedge are plotted as a funetion of the common interelectrode spacing normalized by the depth of the Fist current electrode C, to the dipping interface. The snlculted data were taken fom the excellent paper by Macda [25}, which fives a very thorough boundary value analysis that includes the case in Which ps eould be other than 0 or zo. Such a general anclysis was given tales by Skalsays [25], but Maeda’s procedure seems mare dizeet and isa fell-coniained teatment For an interesting discussion of Macda"s paper ‘anda eral evaluation of easier work based on image methods, see Van [Nostrand and Cook [26 As they point out ina cogent fashion. the image epesentation is not generally valid for the inclined interface unless the Tower lier is a perfect conductor or a perfect inwlator and unless, Turdhermore, the wedge angle (as we have defined i) i subrmultiple of sx Nevertheless, a these authors confim, the image eprerentation sven for 2 finite resistivity contrast ean be a good approximation if the dip (Le, ‘wedge angle 4) is sufficiently srl (27. [n such cases the image sequences are truncated in an appropriate manner. This question has been diseused38 1 cay Rescovy Paneer p26. Weer aay et perpen A a Nee. a sparen ey by Ralston and Wait (28), who proposed a novel electrde array to probe {he conductivity steueture of roots in coal mies. ‘SPHERE PROBLEMS} ‘An often-used model for an inhomogeneity in the earth is a sphere. We will consider the ease, but first we need (o consider a sultable represents tion for a current point source in spherical coordinates. “The geometry ofthe situation s shawn in Fi. 27a, The cureent electrode, injgting a curren into the medium, is located at? = =, onthe axis ofan (3.2) coordinate ster, The resistivity ofthe whole space isp. Therefore ‘he potential atthe observer point Pi given by ym pleeR), ast where R= [s+ 98+ (2 2)!" Ie terms of spherical coordinates (7.8 8) with common origin and polar axis # = 0 coinciding with the = axis, eames (20-31) sonar Prosiems 39 Fg 7a Corea! soune faa iit coniocing mato, 68) ig 27% Cans urs in lo fey p abide np of shy we have Rea (P+ = 2ryc0s0)!”, (168) where = 24 y24 2? and cos ses not appear, and infact 3/3 = 0 ‘Our objective sto express 1/ Rin terms of powers of ror of 1/r. Fes, ‘consider the ease in which r [2 we find | | | | Responds of @ Homogsnsaus Concitng Spare a that the potential y at (7,8) iether (176) «am ly. But thee expressions are mathematically identical. Thus P,fcost) £0) omy Remembering that : = reos6, this result i relly generating function for the Legendre polynomials P,. tn fact, an equivalent definition is the Resrigues formula Ld atiyt Pay he Bee (9) where w= coe “The Legendre polynomials P, satisfy the differential equation Hla) + n+ DP, =O, (180) which is well known in the theory of spherical harmonics (31). Another [Legendre polynomial designator, Q, (cos), lso satisfies this equation, but iis not well behaved at 9 0 and at = = when w is an ineger. For hhoundary value problems involving the Full range of @ from O to 2x, only the P,(coet) are of se RESPONSE OF A HOMOGENEOUS CONDUCTING. SPHERE We are aow i a potion to dal effectively with ase boundary value problem evolving aphoral rge The stun stated i Fp. 7b. ‘The sphere, bounded by r= a. is homozeneous with reisviy py. The current souee 1 feted at = zp om the as ofthe spheric eooedinats sysem (7.84). For the problen a defined, symmetry preva, 20 that ajae “The potent 40) il say VY =O everywhere excep ight a the source. That cst)a 1 ah Reituly Pees ‘im the chosen spherical coordinates. Solutions that are relevant to the preseat problem are easly found by the separation-o-varabls technique They ere P.lco80, 182) rnp (cob), eo where isn integer fr single-valued solutions. Fit, choose the soure to be outside the sphere; thus zp > a. The desired ‘eopreseattion forthe potential y for r > 2 ces om et Saat matcoro), (8) were A isa oie and where R= (+24— Yegeosd)!” cy But of course we ean utilize (171) to express & harmonies so that Earsoost) + $4, pyfcost), (185) which is valid inthe region « < r< zy, For the interior region of the sphere ( 0) = eer Treas 24 co and that y must be even about = 0, write 4 ["0rKuirjcontea, (204) whee 0) yt to be determined ip 29, Cynder tem wl caret iat ou1 Een ess Prine (Our source condition i clearly Barada lag B(2M (205) tere J, the urtent density normal tothe slender eylindcal surface of radius ry that encloses the point source of current. Using the result 1 oy dan h te (206) andthe spectral form forthe impulse fonction 42) i en we see that (205) is equivalent to in, 2m F106 raRAreosr: (208) Ba ig, [ok] =, 208) and ths we deuce fom (205) hat JO) = tof Qx*) Lea | ‘heron he desired form i * eS Reon cosrean, any his devlpment could be eatded as dest dration of the ine san Bf kercrncan, any ery ih own ne tery of Boel eons ing all or > Beahe, Comer he cin which te cian deol of ae logit 2! bata of Iain! Show in he pola a [* SOM cog a 2 493 fPman Monee en) whee we have asm hl he ant ngs sf oe teeth of the electrode —fa is homogeneous and has eesistvty py. The feureent point source 1 is located at the origin. The problem has obvious symmeuy about the polar axis. Also, we assume thatthe eslinger is of Infinite lena "Appropriate expressions forthe interior and exterior potentials are (Usan+aanneryjenea ais) and [raenapeysican any tees Hote A) and 80) can be chai th aa manner by ‘curing tat the pte aad caren fw be conan sn te Bovnday src. at rma 16) (= Kapka) Tika) Reha) * eledayR a) ary Fe 3, Cent pin one ona onder fry py rand "See race BEIT,e 1. ah Resbuly Pees and fa a) 0)~ SalpaiiaayRetha) + a fthay Kita] Hore we have made ws ofthe folowing properties of Bessel functions Misi fis= 1s), aKlx\faem — Ki and HAIR EN) + HORA) = (1/2 The last of these equations (the Wronskian relationship) can be ilized farther to write =A 1)Kia) Kia) AO" Tea R= thal a) 19) and 80)= Tena Tyna) Ra)" aay) where K'= p/p The problem is now fully analogous to the elaceostatc problem considered by Smythe [35], whose results refer specifi point charge of g coulombs located atthe center of an aifiled hole ia 2 ilecrie whose relative. permitivity is K. The eaeulations made by Smythe’ colleague A. E. Harson (see (35) are thus dtetly applicable to ‘ur resiivity problem, We depict these cesuls in Fig. 31, where the ‘eqipotenilsurfscesaze relative to the -vot surface shown, The caleula- tions pertain to K’= 8,0r :/09= Sin our pre natn ap wt for pit sr fee on ni a ee Curent cro a Wal Layered Tanna a3 (CURRENT ELECTRODE IN WALL OF LAYERED TUNNEL Another example of «cylinder problem is tht of the curtent flow from a point electrode in a conducting medium adjacent to an airfilled cunnet [G2] Here the tunnel is of radius a, and the adjacent homogeneous region af resistivity p, extends out to adits ay. Beyond radi a, the resistivity is dbx The situation is iustrated in Fig. 32 with respect to a epindsieal Coordinate sytem (n 4,2). Te air-wal boundary i located at ~ ay. We row locate the current source at r= ry where ay < p< ds, DUE ater Wwe an allow ry 4 (i the source atthe wall) “The immediate task isto find the potential Qin the region a, < r hr Breast 2) (223 “ite,0,0) Fig. 32 Ailes sone wth tered woth and carte pn source in walkWwhete Tis an operator defined by PHB (Secs ysoomedead, Gy where Gy ff =2 for m= 12,5. ‘Appropriate solutions forthe concentric conducive region are then y= (1oy/49)T (Kran) + An(AMAN)* Bal d)KCA) (225) for ay << ry = (In/Aa)T (Lala) KoOr) + AalAYaAA + Bal¥)Ka(W)) (226) for ry as ‘The boundary conditions are clearly sell 28/970 ar yam aan 0a va er ace, {©)_(1/o,908,/a7 = (1/p3100/07) a r= a, (nol cute flow is continuous at concensc interface) (0) = 8, atm oy (potential is continsous 4 concentric interface). lis now a straightforward task to determine te unknown cosffcints Ag Buy 405 Cy By applying these tree boundary conditions. Explicit orm are given elsewhere. Here we will consider specially the special ease in which the cucent slectrode approaches the tunnel wall and at the same time the potential elevirode (Le, the observer) also approaches the tunnel wall, Then, on ‘making use of the Wronskian 10) (80) ~ Iya) Ka) = 1/20, 28) we obtain He) [ as) + Qo 2) 0) KGa) | Tare Oo/Z Te) ~ Peay [Rasy Wei7ZVR Cas) Ra) | Ta) + OZ) 3) ‘where we heve dropped the subscript m on Fand £ everywhere and where Ka) Kz) Ka) ae a RT In the present form, the result i ficult wo interpret although the numeri cal evaluation i cesiaty posible Here we dcuns ony the iting ease of a plane boundary. We uilze the asymptotic limiting (Debye) form ‘ven Tor example, by \ Kino eave]? Mite Relay) ~ Kay) Wy where w= 0+ 2!7 and fa, = m. This aymptoic behavior i sppropr= le sien both order and argument of the modified Besel function are Tange 1 Sars, 30) Fie) im a3 ste Tai) a ‘Within the same approximation we would find hat on, ex) where s = 4; ~ ay. Using similar Hmiing forms for the other Basal Fanc- ion atie, i i ten not ificul to show that Q, as given by (229) is approximeted by P she oa Mp t Lek a ro) vere R= (o.— a)/os4 70) 34) 3 Form it also appropriate to replace the sume tion over mr by an integral over @. Thus Suc Jeane Eo gemenf{C emer (235) ‘which becomes more accurate as gyro. Thea, on ientilying a, with the Tranmere coordinate y, we soe thal LT ES Bre = 6 (rarea3, ven ‘YJamiean 2 apicosde dh apeaacin82 1 an Ree Pricias Here we can make use ofthe spectra definition (1) kustsy7]-4ftdemeran amy wihere w= (PEND! we to 5 1 a= | 1 a2 $e —_1__.}. em Bye a [Pee tam) “This i clearly the eoreet limiting form (a, compace with (25) and (0), where here (y? +25)" is the distance from the curreat clectrode to the potenta electro). LIMITING CASE FOR EXTERNAL CURRENT EXCITATION OF A CABLE losely related problem involving cylindrical conductors isthe ease in which the current source i in the extemal homogeneous medium (Fig. 3). Here asuine thatthe eylinder of rméivt a is perfectly conducting. The fuerent source if located at (r.0,0) in the external medium (r > a) of resiivity p, The distance from the observer at (7643) 10 (70.0) = amcor + 28/2, Now write the expression forthe external poten inthe form roe) EZ (pt TAW&u]} 239) fora- yy The region external to the spheroid (enn > m) is homogeneous with resistivity p “The primary potential st P(n. 3.9} given by the usual inverse distance formula Jem V0 /teR, (256) 3 ing 8.01 Me 6 gale hod dy fen yn Bomapees etre of si o58 1 ath esebiy Pras where A isthe linear distance between source and observer. We now apres V7 asses of spheroidal harmonics this [Zee ee OPPO) nS 's the appropriate superposition. The eoetfiient ro imply our development here we we the definitions 2d"P(a) Pasjcosme, 257) seit yet to be deter. Pre (9 = ee» (58) na ronnie 222 om Frey 260) tearing in mind thal y> Land —1.< 8 < 1 Ao we noth ine 021) Song at = iis pt wed Bee ‘WE noe we oor Sure coaiton in the form Bie 8 net He an “This canbe eect y mung bth sie by hyd and inept Mone Oem tta'tt and over tom Dio Ze pag sah ete, Seco the petal ome ayalnes ah sey= 4 Secom, = sie gee 8rd Sans 8.) Pm 5-8)- 3 om arCagee(s) — 8) to eat te ihn side of 6, On thor an, we note ta {-a[¢ PALL BH Jars HEB [Pr engaereny — OngPernd] X PIB) cos. (264) sonar Probie 59 “The square-beacketd term in the above i the Wronskian and is equal to [ (nemp | os pe eee nos mF ip, ceorowrs) BEF ex ‘We are now in position to solve the posed two-tegion problem. We set, v= 3S [nattinn+ a .O3"(n)]P2(8)osmp (266) for ny <9 < and Vi SS caaPttay Ans cosme en for 1<9.< me: The continuity of potential and ngrmal current flow atthe spheroid boundary says that ay Sree iy vers 7 oem 68) Solving, we find that (> PAP. He Pe") TB O50) Pee) — OOS) a) es, Tax ter nc ee hr aa = et ES mariner aycomme em) forall values of > ny ‘When the current point source is inside the spheroid, we proceed in the same fashion: we express the primary potential Jp,/42R in the sume form 35 257) but now p; replaces» in the expression fo 4, ziven by 255). Potential problems in oblate spheroidal eooedinaes (48,0), as shown in Fig. 39, an be handled as a stnightforvard transformation. We lt 9» it and let My ‘As in olher problems of this ype the results apply To a resistive half-space when the spheroid (prolate or oblate) is bisected in a plane in Which the normal current dessity is zero. An example is the oblate hhemispherodal sink that was eated exhaustively by Cook and. Van Nostrand (39). Some oftheir results are shown in Figs 40 and 41 in which the apparent cesistivty pi expressed as ratio tothe resistivity 9. In Fig. 40 ar oblate hemispheroidal sink is chosen in which a ithe interelectrode spacing and the minor axis is 32/8. The Wenner array i located with i tus alang the diameter of the hemispheroid As indicated, there are cusps inthe apparent resistivity curves as the eletcodes cross the veri con- tacts In Fig. 1, the results are shown for the casein which the sink it chosen tobe prolate, oblate, and spherieal with «common axis of rotation, ‘The Wenner army is chosen to be along this axis. The radi of revoltion of oF fet ga Profle appara ety for 8 Wenmer ary of cing ta tenes the dameter ef obese ete n= Sy Pesetance ot Grounded Sacvecte a he 7.41. Annet ev for Wenner sy oer rn anole phe nko salty 9 Nowasens edn cant pws p= Sh the prolate hemispheroi, the sphere, and the oblate hemispheroid are (053, 1.00, and 1944, respectively, relative to the sphere radius b “The results shown in Figs 0 and 41, ofcourse, apply equally well if the array C,P)P3Cy is vertical and the surrounding fegion is fully symmetric ‘This realy is the same geometry that was considered by Seige! (40) in his pioneering paper on ore body size determination, in which he presented ome interesting caleulations forthe ease for which the electrodes straddle fan oblate spheroid conductor that is assumed to be coaxial with the borehole, The important eoaclusion was tht the lateral extent of the body eould he estimated clearly trom the measured apparent resistvit data RESISTANCE OF GROUNDED ELECTRODES ‘An important aspect of any resiivly measurement is the contact, resistance between the electrode and the soil oF other media in whieh the tlectrodes are placed, Although in principle the ranser resistence deduced from # fourelesirode array is independent of the cont rasstances, the Inter do influence the accuracy ofthe measurement. For example, a a dry borchole it may be difficult to inject sufficient current into the medium to obtain satisfactory vollage levels at the potential probes.‘We shall now examine several idealized models of electrodes to illustrate some ofthe basic principles and limitations involve. Hemispherical Electrode The simplest model i to represn the metal lst 8a hemisphere Sow of ny. ras na homegncus hop eat. To make the problem slighty more ire, we alow forthe pretece of reave sheath of ety py wwe tra sb (ig 9). Accent 7 ited to he sesrode that ssumed fo hve fine Condos. The Objet to dtr the aestance Raby definton ite at Yo, whee Vy the vaape or penal at'r~oyr Dig the rad Sorinte” Atal Vth fees n poe btvuen i esate Sn the oter decd "ny tha the caren si. Now ge he owes 7) at ay point r> @ fa sluon of Lapa equation, we Pay immediately wie the sperally ymin sles VU)R ARB for ac red, en and Payer! for rob, en) where A, By and C ave yet tobe determine. "he toundary sonidos a= ae oni of 1) and (am V(a) = Vo. Thus we find that Pe eeeaeae A Bb" = Cb, ny =p tBbb = =p, am AA Be Voy (75) ‘Those ean be volved for A,B and. in partial, we find that ¥en et B= on Bis a ‘The total curren to the eesiode must eros the surface r= b, Thus we may write -LAL| enn = 280, em Pee, 2 Hemera diode wih eine ste, ‘where 2nb? isthe area of the hemispherical surface. This is equivalent to BeB(i Hem em hich the died re hone eve seth i absnt Ge, #= 3 of coun, becomes singly 9rd) When the sheath hick = baa, we ae Cat watts es) ier 1 = (8 ~ ey the nee esac it cms pr square mit. TREY tlie tat eect wl conto he cc eho of Be hese Tiss pruclary te cane whe he eel proeses thes inc nih the meta eet Sisay een af he pent models consider ta the Beni sphtsal eetode ef radios « has a cong of outer fads ah BiStny onan forte lye of ase wih esp, For? the restuly agin, Now the eee reste gen by op (L yy MeL b Renae tie(b-2)*3(a7 4} em “te scone on the igh night be cre represnaion of he est BPS Goelapes rein nthe crt sounding te cect that has teen modified by soaking the repion with conductive saline solsin, ‘Geary the objective is to maximize e and t0 minimize [Eyereise, Derive (280) using the same method as that used for (278) ML x, Prolate Spheroidal Electrode ‘A more realistic shape for an electrode would be a slender eylindrical rod of firitelnat [4 Theis nota very simple shape o analyze unless some Strong assumptions are made about the curent leakage distribution, How ber, we iealize the electrode as e prolate spheroid of perfect eonctuctiv~ fig. an analysis ean be made rather easly. Also, bearing in mind hat the lestrode might be located in a borehole, we allow for a fluid layer or sheath that i Laken 10 be confocal “Choose spheroidal coordinaes (1, 6.4) as shown in Fig. 137, As indi cated ia Fig. 6, the metal electrode i the surface 9 np and the sheath ‘r coating liquid layer i bounded by n ~ ne The resistivity ofthe sheath is ing while the extemal medium >, bat & resistivity p, The semifocal Aisne i Evers Refering tothe digram in Fig. 142, assume thatthe potential lover the surface r= is Vgcosd, where 0 = O isthe vertical axis Show thatoa 1am Resivy Panes the potential in the region > bis 3eigcond [20rd + + Roane 7ay it @=5/2'sazoro-potental ground plane. Bl ven) [Now assume that 9 = ny isan equipotentil surface tht is Hy relative co 4 point at iniity. The coresponding potentials in the two regions are of the form Fon) K+ BQdn) for m me 32) where Qa) is the Legendre polynomial of the second kind of onder zero. ‘The constants K; 8, and ae found from the boundary eondions in a manner fully analogous to that of the soltion for the hemispherical lectrode sytem, we require that M(q) and Jy(a) be continuous at yo ye and that Ving) = Vox we realy deduce that Veko/ oo) Qua) voy= for >a 8 GD sha) + [(0/ 00) = 1} QobM6) ) The total current J supplied to the electrode must eross the surface = Me Thus we may write nL pipet ey, reek He raont| om Heol that Pala) = 4 Ina + B/C 1) (285) and Qn) /dg = —1/ (7 (286) “Ths we fad ha 4Vie ou es) Gia #[ew/oe)— FOG) “This permits us o write two equivalent forms forthe electrode resistance as Sees (2-1) Set om a, eatance of Grounded Bacod9s 6 a al a 2 ig. rae phere motel of ela wh ene seh of VY my surly hance mn 0] aw [s[aa3—]} em ‘To relate the parameters to the electrode dimensions shown in Fig. 43, note that and ou R dee [t= ca/oyry (290) n=[l-(esoaty ayy c=(e a) =a? (202) Here @ and b ar the semiminor and semimajor axes ofthe metal electrode, whereas a and 6, are the semimajor axes of the confocal conducting region TT p= pg, which corresponds to a fully homogeneous conducting environ- sent, we sce that the term in square brackets in (288) is unity. This reduces {o the well-known result that has been quoted, for example, by Sunde (7. ‘Ashe indieates, a very similar expresioa is obtained forthe case b> a by sssuming that the electrode i a log slender rod with an assumed constant leakage eurent tothe surrounding medium, ‘Acualy, fr slender prolate spheroids suck that both b/a and 6/4, l the esl given by (288) and (289) ean be simplified. First, note thal in this cue: 0 thatthe lengihs of both the electrode and the sheath region ate approximately 2b. Also, as Hlustated in Fig. 43, the sheath in this cas is approximately in the form ofa concentric cylindrical region of ‘uler radius a, snd Snnet radius a, the same as those ofthe electrode. In this situation We argue that the resuls or the resistance R, should. be representative of such a rod-shaped electrode located in a botshole filled with electrolyte id, [Noting that foray Ig + Dlg ~ 9 2/tt9— N= (2b/ a poln(26/a) oe _,)!8028/2,) a (Li) aaagay | ob)6 1. ah Restity Prine and plni2b/a, ib ohm, (209) i ps (auf) Tne /a) ‘As indicated by (293), the input resistance is characterise of an electrode jmmersed in an infinite medium of resistivity py only i the contrast p/n) — 1} small. More typically, as seen by (298), the input resistance appears to be determined mainly by the resistivity p of the external tedium. I however, yep, the effective diameter ofthe electrode is 2a, father than 2a, Thus ip general there is Some merit in having the electrode immersed in the deil-hoe Ti. REFERENCES 1. GN. Wason They of on Faction, 46, Camber Un. Res 2 SE Sdiccn 6 Suntan Mt Satta Sr sien ut Sonic tape Pg, Ria 52-4010. 3, HIM Mooney ad W, Weta Te Meta tot 9 Pot Bera ad pare Ren Cfo ada # Lae Earh Uso Mines Pres, es 4 Le Wstand A.M Conds On he menemen of woud oad VLF EEE Tam drone Pop, APSO 53-27 5. UML Afpi MN, Bouchet, G. Vedran and A. Mz Zap Die io eg Ear Cont Emp sy .V. Kets Cahn: Baru New Yr, (Gv, Boia peoeing oe iQ Co So, Mine 68 (2 1-289 19685 1D, Sie Eek Caren ifn Trani Suns. Doe, New oe. TG. kine, Pier of Dirt Caren Resiy Propecia. Csr Dos, Sak 9, Er Prisor and. Lye comput progr four probe ety eerste Tea bortnity lee exch IEEE Tr, Go Eaton, GEM (4), 32235976. 10. 6, Reso Gang Pop, Rony Sang Maanremet: Amer Eheve 1. CIV Reler a F.C, FRsHRnGh. heel Maat fe Gopal Prpecing, Peano Osler, 1986, 12, CTL Shyer and LR: Wak, Rett bing af ax "enone er 13, Rohan Te ea no he rc pn fr oly el uh, Gays 188-30 14, HP Fryer me cae forth opal inert of pode rviy meuemenssone 2 twos te Gp Prope. 71-38 (isa. 15, Danse. Senna a BF Pal. Aparato = mad ah uh ue fig ctponetly yn eet, Ga, Pape 2 3452 am eterencos a 6 Ran, Sent an Pa Rey wong apr at ‘caring tansion ers Gaps. Pret, 2.150739, oo RSET pte ey ig met ro De Paw A merc compuston procure fr the Sis inrrein a go ‘hse sundogs Sas Prager B36 19, Pa The uate errata oun by mean hei Irlr aneton, Gp Props 2 86.180, 21, IR Wat Elecrommgene Rar ow in Str Pega, sr a © Syscan9 sn. EATS Fis nd sin Sin Das Nee ok hove ald pane Te Py USSR IB 142-1284 (18, 24 NC Use Apprentice for dion bls Gps I, 16-137 (1950) 25. K Min Apparent rs for iping bs Gaps 38 123-139 (1955) 238. RG. van Noun nd KL, Cook Appen Fty Cr dpi Bas Bs Shy ase 2, 160-17 (55, 2, Tee hap dine hed fr seta wounding cod on ver a nena InracePie pl Gp. 8190-14072, 28. MA alsa FR Wai Ther f weequeny candy odie of ot ‘rae In al mins Ra 1 1) 118-1109 Dl pele, De No" sttanavnib an metic fom Arca Geophys Union, 200 Faia ‘Aes NW Wahiaghn DC 2008 SU pea) 29. WE B_ MacMiin, Th Tony of he Pa Doves New Yok. 98 M0. J. Stton leomine Pay Sein 38 era New Vek, 18 SLT At coer Sao Fama Se Repo Oxerd 196 32 J.P Wat Some arth esi pubes olan oleae, . Cl. Sk, Miner Srna, fon Pare ip pi 34 FR Wat Tif pre pra scene, Pa S14 45-51 (1860, 3S. Wek Sythe, Scan Brome Ear de MeGriti New York, 96h 3a. 1a Wat The pte fe creat pot wesc homoge Sing fre pho Ph 348-91 (S. a, S°Gamno aed Rau The lat of and pon inthe ole on poms of Arai fe cath iy rape of ao i708 9) typing oak Gs 2-84 (97 SMart Sloan Eee f emaon niazopy on rive lung Imerremens Gasper 4 126-8619, RIL Cook and RG, Van anand (Grier 161-7098 1G See tose aie dec preci, Shit 1 TTL Wat Resanceof akclatads Ekon,La98),90-9 (97D. eprom of revi soe lt ns.cvsot UH Theory of Induced Electrical Polarization (Quasi-Static Approach) INTRODUCTION ‘A natura extension of the resistivity method of geophysical exploration ict allow for the posibiliy thatthe resistivity its isa complex funetion of frequency [1-8]. The basic idea is very simple, and it can be seen lnmmoditely if we write Ohm's law for feds tht vary harmonically with ya) = Ee 0 here the time factor i exit), Ella) i the elected vector shown plc asa fonction of ox Ji) is the carzen density vector, and ia i {heresy 1s understood of eoutse, thatthe components of Pio) and 5X) are pinsore Ths for example, the real pial quantity 60) Gy ihe component) relted tothe complex phasor by ex(1)= REE Cine = [B,loox(et +4). @ where 6, = phase of 6, ‘Then, of cour, o and “ where a lnosetion cy ‘The implicit assumption made her is that p= is) &§ not a fonction of 4, Jia. In other words, we are dealing with Tnesr media in the sense ‘hat E, i proportional to J, even though the ratio may be a fonction of fresueney. tis rather obvious that we may write lie) =[ (Hu) + teri], o where, without loss of generality, a(ia) and (la) are defined as the real ‘conductivity and the real permitivity of the medium. (It seems redundant snd undestable to allow and ¢ to be complex, as one finds in recent wwitings) Equation (1) is actually a precise description of the relationship of the lect fleld and the current density at any point in the medivm for timesharmonic fields. When dealing with general time-varying Feld, itis bepfel to use Laplace transform rotation (6. In this case we may formally ‘deny to wih the Laplace tansform variable + that can be generally ‘complex. Thus to generalize (2) we would write etn [Bape ds oF the equivalent form el) = ay [Pe (neta a ‘The contour of integration in the s plane hasbeen shifted toa ine just to the right ofthe imaginary axis (Le. ¢ > 0), In accordance withthe Laplace transform theory, (4) the wansform of e,() A slalement ofthis fact is 2oy= [Ceoreunte ® ‘A symbole statement is Ble) = Let o ‘whereas the inverse transform given by (7) is abbreviated to e() = LE) (19) ‘The complex resistivity pis) now is continued into the x complex plane and is designated p(). This ote} Lota) + se(3)] ay where, indeed, a(s) and «() ate generally comples, but they are real by ‘definition along the imaginary axis of (3) (ie, for real values of a),nm 1 They of ndaad acta Polreston “The corresponding transform relations, forthe eutent density, are oa)= BAO (2) and Jt) = EY, (3) BASIC TRANSIENT RESPONSE ‘As an example of relevant transient problem we might stipulate that {Jat is an impressed quantity, Then we wish to predit the resultant form Tor £0) sven the complex resistivity (io) for all frequencies. One possibility is thatthe data allow us to use the representation is) Ile = 3 where Vis chosen to be sulfiiently large to represent lo and &® overall el Irequencies. Here 4, and a, are etal eoeficiets that are chosen to allow ‘he best fit ofthe experimental dats [7] for positive real frequencies Infact, it easy to eee that fr real 460) = 0%(—), ‘where the asterisk denotes complex conjugle, In other words [an even fonction of real w and is am ode function of real at “ "We now wile we Sachs cy a he dared analytic form fo 9). Then because Eso)= Hat9) 09 in the ple i flows that (= EWI) where J3)= Hite) 08) is known, ‘To evaluate (17) we need to indicate explicit current density. To illusiate the pot, choose GAO) = So) 19) he frm for the impressed asc Trane Response n where wt (om ib {is the unit step function at s=0. Then clearly hey hiatnenfteran SB Now we can write en Here note dat 1 Aeeay oy ‘Thus QI is given by ae alo E Stevo, ‘We se tat 6:(0)=0, whereas & ay esl) = Eon Jo @3) is the limiting cloctric field for sufficiently long times. The qualitative behavior of the response funtion e,)/ Ey illustrated in Fig, I Por most earth material, «,(0)/ Ey ses smoothly and approaches | at saffciatly Tong times. The quantity <,(0)/E, which we gexignale as A(?y is the normalized stepfunetion response of the medium at a speifed lation. For isotropic conditions it would also be applicable (© the (0) and (0) ‘components. Note that AC) foo "o(s)/5) 4) ig. 1. The nature ofthe eecre fed response fra sepfunetion currantn IL ‘Teor of nace Elation Polarization ‘where py = o/ J. Now (17), for a general excitation, ean be writen as ext) = (Hoa Mosy/s)[ 4). 25) were (ey 1 EO 2 t-yen, 25) ad where 0 a{ailng ean be called the de resistivity. “Thus, in accordance withthe convolution theorem for Laplace transforms, swe ean write, for > 0, ene al Laerver— ner + acrnion| en of fa nnterer+ av.0] Cy mond [touts - ye ° “Those are relly equivalent slatements to the eect that the response «(7 ta the source j,(f) can be reparded as the superposiion of the unit step-funcion responses nthe various exploration schemes that exploit the induced! polarization ‘phenomenon, it oflen convenient (9 use an on-off current scheme with fltemating polars’ (8-10). For example, feom the time ¢=0 that the ‘current sequence ie stared, we would stipulate that jad) = SL o(e) — (r= T= als T= 4) + ME T= = TIF] 0) Bat) = HLBU0) ~ 8 T)~ B= T= 49) BEIT yb 6 where 8(¢) isthe unit impulse fonction. As indicted in Fig. 2, Tis the length of the “eurenton” portion, whereas fis the length of the “eurteat- +ALL F.2 Onl arent weveeem, Bosc Transont Response n Fig. ected pot for on-of ereat wears. ‘off postion. As ean be seen most leary trom (28), the resltnt response e()= ELA()— AUT) AU T~ t+ A= AT) co) where, remember, A(0) = 0 for 1 <0. The qualitative nature ofthe response isillstated in Fig 3 we 18 obviows that after a sufiiently ong time the response waveform xt) becomes repetitive if there are no other time-varying factors involved ‘OF parcular interes here isthe decay portion of the function 80) fst alte the cessation ofthe eureent charging portion of length 7. To facilitate the discussion of this portion, we ait the ime origin = Oto the end of @ charging portion ofa paricuar eyee, as indieated in Fig, 4. The respons, refered to the new tine eign is lesly given by (1) Eyl) ACH T) FALE THA) AUF ITH Ho] @) fer d<1< 4 “Actualy, this is finite sequence because we continue the terms in ‘braces until we arrive at vo PACH PT pla) assuming thatthe “switch” was turned on atthe begining of a positive current charging portion at T+ tay eA Shing tin orn.™ 1h ‘Theor of rduced ects Poarizeton where pis designated an integer. When p is sulfiienly large, the waveform ‘becomes truly repetive. THis evident that when the chasping portion 7 is sufiienly large, the response given by (3) i simply 2()= EM) for 051 < ty, where My=t- ay 4 can be called the sine depandent chargeabiiy. OF course, (33) ean also be (0) = EL M(Q)= MEET) MCE TB) MHIP G) He) 63) for 0< 1 < ty remembering that Me) =O for <0. (OTHER PARAMETER DESCRIPTIONS Other characterizations of induced polarization in geological msi exit, “The relationships among these are confosed and vague because in many cases nonplyseal dseriptionsaze employed. Firs, let us return to (14) and wit tin the form 00) = 3 Grae, tho 6) here eesily G,= ay/4, and C,=1/dy. We have added an iafite JFrequeny sesiiviy ttm fo permit lal be nonzero as o> 2. We make ‘his nonpiysieal adjustment to facilitate the discussion “The equivalent circuit for our model is shown in Fig. 5. The input Impedance of this ercuit between Uhe two terminals is pla) and the de resistance (12, 2-0) is clearly given by Ao So, + Pao: (Gr) ie 5. Exile cent for apparent comple resin } 4 | | ter Paar Dascrons 6 “The time-dependent chargeability i easily found to be Go") 5 38) o (OF course we ean alo write May o, hich i valid for 0< 1 < eo (On the other hand, the chargenbliy nepal is dined according (0 imi fae “y Ue fis sufficiently small and #3 suficienly large, this eould detine an imsfator (2) A frequency effect FE has been defined acconting to FE= (0)~ Pa)/Pa: 9) where py i designated by pyc and py i designated by p,q. A percent Frequency effect PFE is then PFE = 100FE, 4) ‘The proponents [12] of this concept described py, and py. as appareat resides measured at direct eurtent and very high frequency, respec- tively. "A metal factor MF hs also been proposed [13], Essential iis defined by MF = 2X 105(25~ P.2)/PP a (45a) although other aritons occur when ied unis ate we (eg restivty in ohne per fot ‘Avery popular description 14-16] of the compen resitviy (ss 14a) na* Alla 655) vier pi the sil infinite requeny behavior A, and ae constant Ind G'S an empireally ajsted nnintger This eprsntation was ‘Szinaly proposed by Cole und Cole [17] to deseribe ima effects in6 1 Theory of Indoed let! Ptiaton ae Te 6 avi sei hae on the Cl-Coe lation mos, ectolytic conductors. Numerous modifications [15,18] of the basi Cole Cole oem have been adopted to allow for better iting of observed data for the frequency dependence of p(t). ‘An equivalent form ofthe Cole-Cole mode! used by Pelton and collea- ues [15] is given by io) = nf! moft—[1 + cant] “6 which i the same as (456) except for notation and arrangement of the terms, Here, +onfoa)'s ey here py represents or simulates resistance ina blocked pore passe. The ‘Sguivalet ere is shown in Fig. 6, in which Z (9+ a) / (Ger. sy Here note that al infinite Frequency the blockage disappears (ie, 2-0) and = ha OPa/(0* Pa 9) In some of the early experimental studies [1 2] of induced polarization it was observed that the imealependent chargesbility could be approximated by the form M()=1-a~ Bloe (59) for ines ¢ varying from few tens of milliseconds to tens of seconds. The formal trasformation tothe Frequeney ple, assuming that A/(e) holds for all > 0, leads to lis) /aq= a — 05778 ~ Binw— SB. on ‘This predicts that [pl varies approximately as Ine and thatthe phase angle ‘of pis approximately independent of Both of these features rein accord with mich of the data (for the fequeney range 0.1-1000Hz) for the Spocilly prepared laboratory samples used by Collet [7] many years ago ‘and for the more recent work of Collett and Katsube [19]. Costes Sprerisl Panil 7 INDUCED POLARIZATION MODEL BASED ON COATED ‘SPHERICAL PARTICLES. ‘A full understanding of induced polarization (IP) phenomena in geologi- cal media i yet to be achieved. Cleary, a comprehensive model should inclide electrochemical, elecitokinelic, and electromagnetic considerations Recognizing the complesty of the phenomena, we fall back on an idealized ‘conception that represents the heterogeneity of the composite medium 25 & loading of spherical conducting particles in a matrix of resistive material that might be called the baekgrowe. To account forthe interfacial polaizax tion, the conducting particles are taken te bave an impedance that relates the normal current flow and the voltage drop atthe surface of the particle (3,20, 20) Fist we deal with a single particle, and then an ensemble is considered so thatthe effective properties may be deduce. {A single parte of ressvty py, radius a, hasan interface impedance of ‘ae ohms meters squated (Fig, 7), The ambit or suspending medium has & sitivity p. The applied or primacy fekd Zy is uniform and is directed falong the polar axis ofa spherical coordinate system (7,8) that is centered At the particle. The time facor is expla). The frequency «/2s is sufi ‘ently low that Laplace's equation governs the fields both inside and ‘outside the particle. “The imposed condition atthe particle Boundary amount to saying that the voltage drop e is equal tony at r= a where J, isthe normal eutrent density, It ie understood that My is & complex function of frequency in seneral ‘Suitable forms for the potetials are ye Ayreosd for O< rea 2) and = = Byrcosd + Ar Pees for r >a. ) Here do and A are constants 10 be determined from the boundary condi- lions. The latter are that J, is continuous and that the eiference in atta Fig. Sing spheres! particle with complex interface impedance.n ML Theory a nduce Eats Paaton potential isthe voltage drop: ie, 18.1% ae yay eM yy pa pa ME a oH Solving for dy and 4, we got y= ~3E [205 + 9+ 2y/a)]! (85) and A= Eya"[0~ 0 ~ (Mm/ A) ][20) + 2+ Ory fay)" (88) eres Consider the more general problem shown in Fig. 8. The parte le of radius a has « resistiviy p; she coating has an outer radius b and a resistivity The external mediurn bas resistivity p. Dedace the resultant Fields when the applied field is uniform and has stength &. Show that in the case in which pis much less than both p and p,, we can wile the potently for > Basin (33), but now pl Oral = 2°a¥ 4) T+ aq/e) + (20°/a+ dy) = ey/e) | where d= b= a. IF now d->0 and pyr co Such that ry = i 8 Finite, show tha (56) s obtained for oy ~ 0." Actually, (56) can be writen A= ol -8)/(1+28)]Ew en) vere B= pi/e + 19/0. (58) “The general form forthe potential external othe partici thus b= = Eyc0sd + axe Feond, (9) where x=(1~8)/(1 +28), ) ’ ig. Poi wh reve canny | | | | Costes Spain Parlor a {e.8, Ean at dental hry noms vent a phi ssp al ity ented shone penn hao ‘To deal with an ensemble of parts, imagine that they ae located within, a spherical region of radius r (Fig. 9), The total number within this volume vill be N. A uniform field By is ali. The otal ela Pi the resin fet ofthe induced dipoles of ll the particles Ifthe interaction between ‘he partes i ignored one ile fo wre the polenta Pas yo Foren + 038, oy Whete the spherical coordinate systems (44) chosea foreach particle all have their polar axes (Le, f= 0) aligned with the applied field E, Now iis clear that ata sufficient distance (ie, rg. and ‘ve can approximate (6) i the manner = Bur cos + (Nay )o090. @) ‘Then if we regard the spherical region as @ continuum with an effective resistivity , We should be able to write % ~Fiesons 63) (On equating (62) and (63) we see that x= (P— H+ 20. wy where » = Na?/r is the volume of particles per eubie meter. This leads to the impoctan result plo (1 rxy/(l +2} 6) ‘The formula i lly analogous fo the Clausus-Mosoti relation in dilee- tric theory as described, for example, by Feiich (22) In that ease we ‘would be dealing with dielectric particles of, say, permitvity e, in a hort ‘mediom of permittivity, The ume formula appies if p/p is replaced by «eq. here ¢, is effective permittivity and 6 ~ #/e, in the expression for x‘When the volume loading "<1, (65) can be written in the approximate form pefpm 1 —3rx= 1341 —33/(1+ 28), (69) whee B= fo nals) /o0. cn ‘where nas) isthe general Fequency-dependent interface impedance hat characterizes the bounda‘y of the spherical conducting particis and the ljaceat electrolytic medium of ressivity p. Of course both p and the Particle resitiviy p, themselves can be frequency dependent, but inthis ‘dscossion we are focusing our allention on the interfacial or membrane polarization. Furthermore, we will henceforth assume thi p/p, which fs certainly reasonable if we are dealing with metallic particles. Thus Ban) (68) we now sexe the frequency dependence 10 the Form (ia) = 1/2bary, ) we mma eee the Cole-Cale form pris) =nt—ay( [1+ ont)" my pomoll sine and My= 99/2439) ‘Not tht according to (68) and (9), the intesice impedance ix alia) = (9a /2(1/i ay (0, f we adopt the od!) = (ia that fis the experimental dats [23] (whee ap ian ajuted conan), we deduce that +> (p0/209)* ™ isthe empiri ime constant i the Cole~Cole forma in oder fo discus te physical igfeance ofthe model, we now choose 4 lossy-condenserdesription ofthe interface between particle and elecio- Iyte hos medium (2), To this end, we weite where (a) /pa = 9e/(a-+ i), co) where 4 w/oa, w= (0). Here ny i the lowsfrequencyo¢ static characterization ofthe interface and fis the relaxation constant for the interface, Furthermore, if the volume ‘estes Spherion Pariles a loading ris smal, he effective complex eistvity, in acordance with (6), san be writen edit) ayy 3am a wing Fi Where we have assumed that p/p: Now clearly the infinite and 2er0 Frequency limits are m~ P= 0s) = pl ~ 37) 0) and Po = 0} = a1 — 30(1— 9)/(1 + 24)] (76) is then a simple mater to deduce thatthe time-dependent chargeail- ity Ae i given by MQ) I= AW) for 1> 0, ™ where a1 ede acy ts “0 = 78) is the deptuneton respons, Alter manipulating (9, we se hat (78 i ‘equivalent to mn=& Seth MO" 5 age Tree 2 Zeit - act +291] ” 3 pak cargabliy sven by m= 2 MO)= m= PL. (60) “This is consistent wih sm (00 Pao co) Actual the ratio p/ny a8 it appears in (79) can be replaced by unity since v has steady been assumed stall. Thus we ean write Mego) @) where ime) =[24/(1 + 2g) enol ~ a(t +291) 6) Here m(e isthe cha volume loading. xbity as normalized by a factor proportional ta the1s interesting to note hat he chargeably r decay response exis iss maximum valve ara function of q when anaemic 4san 6) Accordingly, when the measurement is made near the bottom of the decay eure Gc, a> 1). the maximum valve of mit occurs when 92 24+ Vand 6) ‘On the other hand, at very short times n= 1/[2e0)"*]. (6) ‘The Feaures af the decay response curves for thi simple ossy-condenser ‘model of the paricle-eeetolyte interface are shown in Figs. 10 and 11. Of ‘course, te responses on a log-linear plot (Fig. 10) are sraght Lines, bu the Slopes are dependent oa parse size. Ten for a fixed ‘decay response, the function m(e) has a maximum parameters changed (Fis. 1D, ‘We can conelude from this model that a medium loaded with small particles will decay mare rapidly than a medium loaded with large parti- flex Furthermore fora given point ia time on the chargeability curve. there will be an optimam particle size in the sense that the response is ‘maximized fora fixed volume loading THis recognized that the foregoing concusions and remarks are based on the simple lossy-condenser desertion atthe pariele interface. Neverthe. less, the quaiiative Behavior ofthe response under more realistic assump tions is not changed. This point has been amply confirmed by the experi= ‘mental work of Colet 7}, Sett nd Wes (24), and Bertin snd Loeb {10} 0s 11s 20zs unc fine for ed parle ae ese oni 10 33 TENS, ey 2 00 os} ‘or 03 13 10 30 ‘PARTICLE SIZE, 1/4 eth, Dey pone a ft pre fr ad pit is ie “The most recent work by Wong [25 which dicuses the electrochemistry in gent deni doesnot really ehange the piture presented hee. AS Was Indicated many years ago, the Fequeocy-dependen interface impedance ‘ia) cncompastes all the tlevantleceochemisry at the partile= Uectroyte boundary [Ij OF cours, am impertent proviso i that we ae Seaing with near phenomena (Le, current denstesseross the interfaces sold be les than A om) asthe developments ofthe costed pontcle model have been deserted by Hsiverson, ef (23) In thai “modified Wait model” they essetally ted our orginal formulated interaceimpeéance description sling ‘(= cons)", oy ‘and demonsated that the Cole-Cole dispersion formula for so) was Achieved. This (inverse frequeney)* dependence was considered (0 be a 00d dscripiion of the observed data tthe pyite-lectrolyte imeface, a Teast over lmied frequency range. When k=, the Warburg inpedance ‘elatinsip was need whereby 1a) is propotonl to 1/ a. This behav for is characteristic of diffusion dominated process the interac, Asscited by Maden and Cantwell (12] and also by Wong [25], However, the oe ofthe Warburg impedance and is sificanee have reenly been, quesioned seriously b) Fink [26) RESPONSE OF A LAYERED POLARIZABLE EARTH MODEL We ate now in a position 10 consider the response of a layered earth ‘model whersin etch ayer or siratum is characterized by a complex resistv~ ity, For purpose of illustration we consider a two-layer flat earth model ‘The current e™' is to be applied to point electrodes C, and C; on thecy 1 Tory of aoe Elta Polarization Aww EEE CEE Bal ig Fours ay ne o-er poll ota, surface ofthe upper layer, and the resultant voltage Ve i to be measured betwen thet potential eletrodes Py and P,. The situation i lusiated jn Fig. 12, in whick the Gy are functions of the individual linear distances, between the electrodes, (f the upper layer of thickness thas a (complex) ressvity pa) andthe ower Inyor has (complex) resistivity) then Sf exe) pais I 1 Gy 22] pare) 8) [orem] ‘The mutual impedance a8 a function of frequency is thus specified by the complex resistivities of the layers in the sume manner as the mutual, fesistance is specified by the de resistivities of the layers in the (eal) resistivity problem. ‘in this time-harmonie problem, the apparent (complex) resistivity pl) can be mitten as ln) = 2nZ i] 69) oa GR GR GR Gh In the ase of the Wenner array, in which the electrodes are arranged in & stright tine in the onder C,PyPs Ca, we choose Cea GPn a and Cha P= 20 and therfare tin) = 2002 (te) om) [Now conser the transient response ofthis wo-ayerpoarable round imadel [Bi The time-dependent chargexbiy Tuacion MC) can be ob- {ano from the averse Laplace vansform operation pean Uf 269 Maya [3( om)h bad Utes of eueny 4/2, hs ioral these ELTA - SG) (0) di. on | sons This can be converted to @ more convenient form for ealeulation and Repaid) sinus a mn [Except in special cates, the integration over must be eared out ‘eumerically forthe given Trequency-dependent pla) and ple) for the layers. IF we adopt the idealization that pio)—>p.. the integral in (93) rns into Convergent cffclis, To remedy this, note that lle) = (0s) ~ Pas 0 whence my=t 5) where f= 040) and pa ‘The peak response as seen by the ara i given by m= M(O) = (Oy ~ Po) 5) Another useful iit the complete charpebility integral sm(coy= fim [Mia “ Asin sing the fil yl them for Lape wasorns We can sede at vncoy etn MOAED Pa (=a, = alo], Bl ais) z Be J (98) ‘The elegance ofthis simple formula belics the difficulty of obtaining the limiting form ofthe derivative as the frequency lends to 210. Exerelse. Assuming hat 060)~ 0408) 8 dy Toi) RR = MUN, Where 4, and a, are constants, show that Sot 3 Ay,“Then confirm that m0) = ‘mend is given by OB. HL GENERAL FORMULATION FOR HETEROGENEOUS MEDIA ‘As indicted by (88), the apparent resistivity pis) ean be expressed as 3 Tuneton ofthe resistivities p(s) and p,(a) forthe individual layers. Ia & seneral functional form, this could be Written 2 his) = pia) 000)] ©) Now without loss of genera, we em st in) = o(0)[1 + 84)] (100) and (0) = 0,0)[ + 5.4}, on where we feard 8 and 3 a8 beng the (comes) frequney-dependent Separtres rom the de eso the iso content anti ‘We now develop (9) into Mactan seis 2] 9 at PaLe(l +) + 8) 102) Here we have made an obvious slight contaction ef notation. For example, jn (102) we mean P= 1h = 10 5180), 82 By), (103) Aoglodt + 801+ 63)) ay ‘Now we write (9) inthe equivalent form (ia) = (O)[1 + 5,¢54)], (104) “Then, in accordance with (102), we must have Blu) = 8, 4105) ere (is the (comple) departure ofthe (comple apparent isis from i lo-frequeney asyeptoi value. We have acy assumed here that pie) is cel in salable region about the real axis Now clay (105) asthe form Bi) = Bebo) + BAH) + BSI) 4 Bashi) + BysBiyBalle) +>. (108) where the indicated bcoeficients are independent of frequency. In patieu- Jae note hat 4 36s D0, a «aon 7 7 Tina, and (108) ste the so ealed dition factors tha appear in many papers onthe subject 11,28, 9} Az defined, they ae real functions and can be obtained from fe resatvty formulations. ‘The real double-subseriped coefficients By, tt. involve double differentiation and again the information i available in the de formulations, We call them -dazorion ectrs sine, for example, functions 8a) and 8,0) that ate Ucar in ine can yieid an appareat lia) that is nonlinear im In. Tn tos eases these second-order effec are aeglible. A notable exception would be the eas in which the resoltaat firorde (ity B+ 8, response is itself very small. 'A useful property of the funetion .(o,.m) is homogeneiny, used in the rathematill seas [27} In fat a seen by (68). we ea wile \04-Ao3) = Na 03) (109) where Ais any constant In oer words ifthe resistive of both layers areebanged from py and ps to Noy and Ap, the resultant resistivity changes from p, Xp, Clary this sa property enjoyed by potential theory teks that saisty Laplace's equations inthe constituent media. In our e25, hhomogencous fonction of order one onthe variables p, and. ‘We now clifereniate both sides of (108) with respect 19, and set X equal to 10 get, (110) (On examining (107) and (108), we see immediatly that yt Bot cay Furthermore, if (109) i differentiated twice with respect A before is set equal to 1, We deduce that fo dosing dn 8g oo any On dividing through by 9, We see imme th Bt Bist BO any Such a balancing relation would hold for higherorder eneicient, pro- vided that the original functions py and their derivatives with respect to Dispos» ae homogeneous functions of dewee ene Tn most eases, only the lading tems in (106) ae needed and thus, in view of (1), we ean write 40) ~6,(0) + BAB) ~ Bie)], cy whore Bing a0 Be, (1s) ‘of course is independent of frequency. Now the coccient Bean easily be ‘deduced from the purely algebraic expression for he apparent resistivity function p, in terms of the ressivies py and py of the two constituent medi, ‘A concrete example serves our purpose, Consider four-leetrode Wenner array of electrode spacing over a twoshyer earth model (Le, = P,C,= a in Fig. 212). The explicit formula for the appar. Ka iert ns a= alt “The ratio p,/oy was plotted in Fig, 6 as function of a/h for various values of K” Now clearly, any “The function By & shown ploted in Fig. 13 asa function of the relative os So MWK Fg 0) Dilton fee fe foe sti Woe any ene e505 TD ESE to ine wma > 0°05 Dion ier acorn (Sore pale owed =p? ® tenet0 Theory ees acral Ptzaton spacing a/h between the closirodes. Various values of K are shown oe the ‘curves. Av indicate, the carves all appraach I asa becomes arg. Also, the function Bis generally decreased when [K{—> | (ie, high ressiviy contras), In the cise in which the overburden is nonpolarizable (ie, Sys) =O), we ave simply (ia) = ByBlio). This yeles avery clear interpretation of 8, the dlaion factor. The ordinates in Fig, 13 are then the normalized or felative IP response fora polriable substratum with & nonpolarizable overburden. RESISTIVITY AND IP ANALYSIS FOR A THIN ‘CONDUCTIVE SHEET Another useful model ie thin conductive sheet of thickness 1 embeded in an infinite Homogencous medium of resistivity phe resistivity of the sheet isp. Both , and pcan be complex Funetions of frequency (Fig, 1). Exercise, Consider a single curent electrode P in a half-space of resis lay a 8) over a hall-space of resistivity pt + 8) (Fig 13). Show chat ‘he apparent resi p(l-+ 5) at P is piven by A(t + KRo/ RK), (ug) were K= (02— i)/ (02+ Pi andy B84 mG—6) for Bubet, (188) wee dina fe . 119 pea tey ® aes hb a_i, 7 z Fg, Me Thin sina sos atl showing the caret cade a (44) af a cond coortinne sem (egcth eset and I Anaya for Tin Concuctive Shes ” Exereke, Inthe previous exercise, assume 8, = 0. Show that 8B, + 8383 (1%) where terms beyond 82 are neglected. Show that 2eiel By ales oye can Nove that Baa/Be= ~rafloa oi). Exerce. In Fig. 1, i the observer Pisin the lower hall-space, show that the resus in both previous exerizsremsin unchanged if (R/ R) is merely replaced by I. Expressions or the potential y(72) were given in Chapter I fr the pure de problem in which the reistivities ae real. For the pretent problems fn Which the essivity isthe same above and below the set, we find that Joy cL nao vt le apg ian cay for 0> > =. and hat =m pee any Cn bet ettainan ox for 2 < ~d. These results ean be checked by noting that ys continuous through the shet. OF course, normal currems to the shet are discontinuous ‘oy the amouat of change of tansverse curtent, For exsmple, the vertical curren into an elemental ring 2rd f= ta LE] tal] where 31/2 isthe hal-thickness oft 1% Fit, 5. Single curs elesoge oer an inrface betwen the poral regions2 1 Theory of nde Eectical Plaeaton “The horizontal current that exists from the annular volume is egal to 4 i) L Efe + dey Dor (y= 2e LB Bale dry k Bert de) ter LBC =I re ER) ow. But the later term in parentheses is ~2')/0=*, Thus the required bound condo AE AB], 9 ® “Te integrals give in (122) and (128) are no ina form for cosed-fom ‘aleation taser sltion ean be developed By simply expanding the denominator bizomialy. Thos we find that Aey- Ceranan- (°{ MPL er mrsayanyan nf Ba ley for0> => —dand y= (ean [CB esaanan Tors < —d. el “The fst integral is of the Lipsitz type, and the second and third are derived by differentining under the integral sign once and vive, respec tively. Then we find without dificult hat (126) and (127) are (aa oA (229) MPV oyaryans => ay 4128) + (29) esi ano Anais fx Thin Contetne Sst 8 ‘where Ra(eeey “These series converge quickly when lpyt/2o,|< I, where is «typical linear dimension ‘As an example, consider the case in which the single potential electrode fs located at r= a and 2 ~0 (ie, two electrodes parallel to the shest and Tocated at a height above it). The complex apparent resistivity for this casei sen to be 2 mat-( 2 bi aaa ajay +1)" [Mafo'=t]ed/ar [xavar +1} (130) It we write lie) = o(O)[1+ 8,00)} i0y= nO)[1 + 8%4)} fie) = o0)[1 + 8(4)], ‘we can deduce direclly the required expression for (ls) by using the prescription denoted by (103). [a the ease ia which the rmbient medium is Tonpolarzable (ie. 8, 0) we have simply 8,(ia) = BaSs{ia) + By3[8(ia) |. a3 were wd nog oe In partly, note hat (ni (Post d/h, (3 ware xl) = 2a f(a? +1)" (133) is goometrical form faclor. Ths simple limiting form for By, ofcourse, is valid only if yt /@pud)a 1 Tory of aioe Eletee Pelton RESPONSE OF A POLARIZABLE SPHERE ‘The simplest finite body that ean be analyzed by our method is the spherical target (Fg. 16) The current point source i loeated ata distance s above the center ofthe sphere that (without los of generality) isthe axis of A spherieal coordinate sjstom (°,0-9). The potential y at (7.8.0) is Cute a de pide iy pen Ts gn e+ 3 acc rtcon). cy where Ing Ot nos SF mae Das aa In Fi. 16, CP= a isthe spacing between the to eletrodes. Both py and ray be complex. -The apparent sess can now be writen a psp) pet 2 arta aes Fae Tp ncord). (136) “Then asl if we designate the complex resitvies by (Ot +89] and p01 + AH], we se that the apparent complex resistivity is (OI +8) where 5,(ie) = B,{ia) + BY[ By) ~ 8,(4s)} «an where 138) Fa 16, Coren source Cin an nou mages me of ety ya oni aa pew feo as py CP = sponte ofa Poaizable Sphere 95 Here», ~ (0) and p= g(0) are the usual de resistivity values. Ii sel to note that in general pelarzation and geometrical Tactors are not separs- ble. These effets are intermingled, as evideneed by the different. n= dependence of the individual terms inthe Iwo summations given by (136) and (138). However, if te spherical target is sfficiendy removed from C and/or P, only the leading term in the summations may be retained. Within this approximation we may infact write By inthe form $v ppb aed 20 ‘where V= 27 isthe volume of the target and dis the perpendicular stance fom the in through CP tothe enter of he sphere. Here Fis a purely geometrical form Taclor tht war obained by Seige! [II]. who {sumed thatthe sphere was load ffestvey in uniform primary field atthe outst Tn ur ae, 38) = (ad), (140) ‘This quantity s sketched in Fig. 17, using Seges data (11) "Note from (139) that high resistivity contrast Ge o)< py oF mp) ‘means thatthe IP response is weakened. Also, itis possibly significant that 3, may be negative bees, as indented in Fig. 217, the form factor can indeed be les than one os p17 Fon cof os ge afc of ahr x ee ace rom inne de 10 De perp om espe aes w= 096 1h Thwor of read Sect Petaenton MULTIREGION POLARIZABLE MODELS. ‘The preceding development can be generalized to multiple homogeneous regione without difficulty. Aeain we are inspired by the carly work of See ‘Asume thatthe region under consideration hes comple resi pfloh where j= 1.2,--- J for he J regions. Again we write afin) = 9(0)[1 + 5(H0)], cat) where the frequeny-dependent departure forthe de value, Now our bjeve is to deduce the apparent complex esstivity pl) fr a particu lar leczode configuration. Ar before, we assume tht al reevant elds are Laplacian in nature. tn other words, the els E(u) in each region are Aervable fom (ia) = when the potentials satisfy erad (io) rye. «ey again we restiet our atention 1 fnear mei, we ean define uniquely an spparent complex resistivity (ia) bythe relation (a) = Ho tai), ‘where ou) isthe voige or diferenee of potential between two electrodes That results from a euren/0) supplied to two additional electrodes. The frequeneyndependent factor H depends oaly on the geometry of the amy. Ta functional sense ln) oto) 93()-089).----2y(t8)), (188) ‘whichis merely a statement tht the complex resistivity as seen by the Tour-leerode array isa funtion of the complen resistivities of the constiae ent medi. Now we write sis) = py(O)[ 1 + 8,(44)]. (144) Then it follows that 54iay~ Fate) iss) ‘where the dilaton factors Bare obtained from 804 Bn, 4 Tre, 146) Aeferences 7 Here we have neglected terms containing higher power of (a). Also the contraction of notation used here for 8 isthe same as that in (103). ‘We again exploit th homogeneous property No.l Pas <= #2) = PaQMP Abas ABs) «sy 1o show that (148) In the ease in whieh the TP decay curve i being measured, the charge- ability Tanetion A4(0) is televant, For the individual homogencous regions swe my write yom ak fTibkeeeeten 8, 6) wire £! the inven Lapse tnsform peat. veut se efter apason Bem hy (14), i alow hate sparen or esultan shares ee min= Baan “This sll also holds, af course, forthe peak ofthe decay curve at ¢= 0; thats, (9 540%. where (Ma) i the peak chargeabity ofthe jth homogenous region. This is the result obtained by Seigel based on a purely static approach. It has been the basis for many subsequent studies by applied geophysics groups ‘Actually, one or more ofthe 8, may change sign and dominate the other terms, This negashe /P can odcur for aspherical target, as we have indicated. Ako, Nabighian and Eliot [28] have shown tht it may aceur in certain (ypes of layered models Related publications [30-S8) on the subjects discussed in this chapter are Included it the reference ist that follows. REFERENCES: Ma ath age Roatan Gp! pins, Pn. Oe 2 LR Wok, Duss on seo stay ina eri partion, Geph> era aecissassey1 Ther af indo Sects! Ptaeaton 1A. Wa, 0. Sig LS. Cate W.& faa 980. TS Pimps of bed Pin fr Gp Eatin. Se ST Eig WA. Sac and. Sent, Conpason a ner an pase ‘neces ince prion, Gap Pres 3 6-68 (97). S'Gckinan. pay Tonio ant Bena Pans. Dost Now Vek 1972 LC Cate Eabrty owntaicn of revags Ia Oligo wd ‘Gove! tei Wat pp 30-0 Peon, Ol. 188 XRitan abe A otto US Poet Wo 261,00 (85 ‘X-Atam, Euan of esd pte ine tne Son, tn. Go. 2 Sacto $b) nh Ei od Tal of Intel Pat, 2 Wo Sige Manos eemiatin aod Ope crs for nde pon, epic, 51-565 (999, $F Eitdec eT Cte nd plato. ever Mi Gp. 23340 ie. SFR Malden and. Cane potion, & ress a Mig Gap, Yot2 Sos of Expoatoe Gaps. Ta, Osha, 1 CR ot ta epee eke Ph Ps and Chef Sinton Mite (R G1. Som, bhp el-28 Why, Landon 1575 Aik Toon, 8 Wor FG Hall, W. RSI and. Nebo. Mine criniton and evi onan epig wi mtetey F Ger Seay 76, Halt dW 1H Poon The env af ind oaping ero ct {ue Seca of xpioaton Csiphycss al erg Haste, Tes Noes ferte-an 15 ean, sco and Rr ae Digeion an sortie. Chom Pp 9 Breast ca. {G19 Yan Vous F Neon and TL Deke Compl revi spe of Py capper mies, Caper 38-928 DE hated Katte Bec gunmen of ets in deveoing pnp ‘nga Copy 3, 530.385 (195 STRIWan inc ten esa cons of 0 tow een, Pre sR Bean casas eects, i edi: Pho Di. Oo Uni Pr, Lan sd New Yr, 98 N.0:tahenon 0:6 tne E'O MA En 6d WC Yates vee ‘anger Rasy oa cone pp, 25-36" Une of Anson, nuny 3 138 Gay of Aina, Tan ‘Wii. oor sel GF Wit nce pletion f cad ih ety et ‘Nc see ns tnt pln ph e ‘to leo, capt 1245-268 (OD ‘ek crochet of ned plan i eee Pon ‘nl Cages aay oe sree. HB-B Ussesty Ato, uy 37, 1S. Saat, Add Cae McGaw New Yr. 538 NCN'Nabahin aE" or Nepiv dn pstmt om sd te: Gps 135-55 Ean int Lan nde polnion eget of hoon mi Seed ehh oi soe Ge, Ppe 38-95 (97) Skah Em pe sop i CM Retr and F.C, Fibhaech. Et Mahe Gop Papen apr Fron, Oxo Se Manoel 4 Testa Eric and Magen Qo, Vol. 37, Oxted Ui Pes (edo, Undo snd New Yar spd by Der Now Yor IRB, Moor, J.W. Di Jy am KL Caok. til proper of hte ‘ars x otc 31.590 9 Vian. De Restn Academe Pres Now Yo 196. 1. Tan Pin of Die era, Msn 75 {Soar A tayo te bw ty eer doprion fai paris i scrote sono Phe Chon 20258 (em Dib faye Amt tng te anys i pln 7105 STR Denl Trechmaniona iy an des gobioton as bd cke- ono: as nla CT. ICA Compe dee erm xs Amst forts mse ad Satanic 0-98 (5. BRecke ko cpt or nas plsaionsersl sens (ne hay Gopi Proper 65-57 (9) EL Bogs J Wyo Ret one and pean in ope iy ‘avr, opr S184 Ryan HM in Ma seo some apt of sigan meme potion tar ts ye eh; Gap Plans a 0-75 Dts and. Shon” Canaan taba one dominant een donnie nde portato Rompe mene, Ces Pare 28 Sea tam ™ Ev Puthomsho, Ei! Pp of Rats: Pm New Yok 1. 5 tesand Gran Te eth eto tts rg ques glade diol trv seine son eatin 18-18. PPG, Halo Te LF phone mewnsersnt tl nde supa. Gentes 38, osc ‘A'De wn HF. Morton lsmsene opine avn and ie doin ince tation aren oper a ahead vr, Ges 30-108 C73. CAS ie and Ge nase Des EM Cop ea ae fs ema wdc cata coming peony 3: €. Wm and EZ Cys dh 30 037) HBS Segc Te preted pos thd MIP, Gaps 39, 21-29 ti. ‘AW Hoshasen D1 hr nd HO. Seip. Some een magne Inaced tution deepen. To: Gels 8 3-3 OO I:R, Wal On he ccrommrete pone os one pee tia Cupp 2569-6590 IE iyon en Reb Zone, Besos coupe, Gyr: Pap 28,295,100 1h ‘Theor at rduese eects Polson x. IR Wai Mul coaplag flops ing on the ound. Gyr 19, 2. “ o [GaTI Nie ht estos or 8) athe ie tp ssl be dhe ws oh {eo temo the hs mtn fewnce On pk Eg (16, he sy aS Sosa wera) i CW Hedman, #8 Kinng 6.0. the menunanent of tae! onto om 3is TC Wa aap 23 ad 2nd Thay (Colin a FJ Zaee E'B Sind a Canc fic i Trin Stes. Dor, New Yok 1968 1%.G:Onnnder and et Zang Compe sity mene shige Bean ‘oer Zane xcs, Toso, Aine ag 0) EA Komar Theta fines polriston met forthe xl of ‘edo En Gel Rep. (eo Ss an} N28 (18, 5.1, Ward sana coaster HH Electromagnetic Induction and Loop-Loop Coupling INTRODUCTION [Nongrounded or inductive excitation of the earths crustal layers can be ' potent probe of the subsurface layers, The principal parameter that i measured isthe electrical conductivity o (i mls per meter). Tae objective {sto deduce the distribution ofthis parameter from a suracs-based system, We presunt here an exposition of the relevant electromagnetic theory, bearing in mind that we are dealing with time-varying fields. Without loss of generality, we may adopt a time-harmonic variation eccording to exp(iat), where w is the angular frequency and (isthe time. If YG) presen the complex phasor, then the etal physical uti wil be the teal past of V(ia)e™. Tn this chapter we shall also be dealing primarily with sufficiently fow frequencies so that the free-space wavelength will slays be much larger than any sigeieant physical dimension of the problem. The meaning of this statement should become clear in what Fallows. [AS in other chapters, an attempt wil be made to present the theory in & reasonably selcoatained manner. References 1-14) are considered to be basic source material. Related studies can be found in (15-35 General background materials given in textbooks [36-42}. MAGNETIC FIELD OF A CIRCULAR LOOP OF CURRENT 1c will greatly help our understanding if we begin with a very basic problem, Specifically, we wish to determine the fields that are associatedse iL scremagnte Induction sn L00p-L09p Couping Plead e 7 gt Cel ap of an with 4 cirelar loop of eleic current Mic), where the surrounding, ‘medium i assumed tobe Fee space 'As indicated in Fig. I, we introduce a eyindrcal coordinate system (rapt) with te eurent loop defined by r= a, 0 < 2e, and : =O. The Frequency’ suffice Tow so that the magnetic field) M(ia) may be derived from (jo) = — gad B(4), w whore isthe salar potenti. Also, because div H= 0, we mast have div grad = 0 2 Here we drop the explicit indication that H, ®, and the causative current 7 are functions of i. ‘Now because dat some point P(r.) external tothe Toop isa solution of Laplace's equation, we are led to write = (7 f. Apere(=[ePvaA AD a (Clearly this form is Raite evecywhere along the axis r=O and decays appropriately as: 20. The Function fs yet to be determined, but J) tobe used for = > 0, is aot dhe same asf), o be used for: <0, ‘We now digress for the moment and say thatthe souree current is a ‘woudimensional distribution with density (7) amperes per meter in the plane #=0, We now waite ddr= [P5004 a 6 0 tha according (othe Fouier-Bese theorem 50) = [rrr 6 grate Fido Chole Lao of Caran 103 But inoue ease, J) = Bera} o where 3(r~ a isthe unit impulse funeton, and thus 58) ~ 10,0 o [Now Ampere’ erren aw tls us that Lig tn +0) ~ Hie. e) Al 6 Also est iss that lig (7 46) He. -9) =0. ° “The later condition tos us that Lo) ~f-0)= 104 (9) “Then els, 2f*famsonrar= [*[tandayprrrar a ns old for lr. Therefore, 50) ~ (14/2400). (ay “Thus the potelal at Pis given by a(n) = = (*1Ra\syarie Man, 3) where the pls sgn is to be used for => 0 and the minus sign for = <0. Here it proves convenient to introduce another sealar function Fr2) ‘thats related to the potential & by o=-aF/os «4 such freon is Fanay~ He [7 soryerien as) and it ay be called the magnets Herts potent. Th fie ofthe loop ‘ould then be obaine fom ~ Ze er "aes ae 09, A special eas of some importance i that in which the ras f the loop, becomes vershngly sa. Then on sing the approximation J,Qa)=2a/2, a7) Ed and108 1. Seeman ndtn ae Lee-L209 ong we se that Flee) = (18) Rae ey? and dd = at itesimal) area ofthe loop. The eorresponding expression fo the potential is 4r,2) = Haz /40R° (9) “This potential leads o the Following expressions forthe ied eomponens of the magnti dipole, tat ee ae om and ey Exercise. Show thot the fel #.(0,2) on the axis ofthe circular loop is ven by ‘magnetic diple CIRCULAR LOOP OVER TWO-LAYER EARTH. ‘To come to grips with an induction problem of some practical impor- lance in geophysical prospecting, we locate the cteulae loop of radius @ at height over a tworier earth model (Fig. 2). Again, we have chosen ‘lindrical coordinates, bot now the loop is centered at 2 = and r=. ‘The ait teplon 2 >0 ie astumed to have freespace properties with magnetic permeability, The upper layer (0> = > = d) is tken to have conductivity « and a magnclic permeability py. The boltom layer or (cular Loop over Twoayer Est 105 Fg. 2 (a) Cea op ott tor hls. (6) The ttre ofthe ectomspsc haltspace (2 < —d) is taken to have a conductivity permeability "Now, Beeause we are sill dealing with slowly varying fields, we ean ase in Ge els in he eo => Oar dale om a potent at ipa solution of Laplace's equation, Ths fo (7.2) = 00% rc) 1 and a magnetic where wns B [PH Qayslivye "dN 3) is the primary potental (asing the upper or lower signs according to svhether £> fof 2 < h, respectively) and & i the secondary potential. "Now because 0°" = 0 forall 2 > 0, we ae led to write w= [Pade a, ey whore PO) is yet to be determined "As an interim step, we might imagine that the conduct Lively infinite so thatthe boundary al =~ 0 is impenetrable insofar as the ‘magnetic flux is concerned, THe boundary condition would then be H. = 0 For 20 from posiiv valu, An equislem statement is 2ereay.o ey108 I. ecromagnete Induction ae Loop Covling Apolication ofthis conition eads to PO) ~ $fad (aye (26) ‘The preceding development suggests that we write the potential expres sion for the upper hallspace (2 > 0) in the Form Ow He, where is given by #8 (7100) R,asorpe MAR, & sr Rf) ito be termi In he iit), and dary A> ‘To deal with the Fields in the conducting layers, we must resort 19 Maxwell's equations In the region O> + > —d, these would take the Ferm inet, 2) co Es 29) bearing in mind that 2/89-=0, Alo, note that E, = E,~ H,=0. "Now it sa simple matter to eliminate , and , 0 get en of this equation suggests that we introduce a scalar Herts potentat via the definition oF, 5, = wilt, @ where Fslstes he Hell equation (==, eo where and o for the region 0> # > — a. Exercise, Show that H, = -(1/2)(@/AnKrOF,/9r) for 0<2>—d, ‘Geel La09 over TwosLayer Earth 107 In the same manner, the feds inthe region < ~d are derivable from es and ar 5, 6) where F satis (PB Ano. or ‘where = io, 440. The basi solutions of (33) and (37) are now exp(e.2) Ideh, where = 02+ 12)" f= 12, We define Rev, > 0 ‘Suitable expressions for the magnetic Hert potentials in the three regions are now easily set up. For => 0, ve Ii) 6 age Meh Ao Bf LOD ess eM, — 8) 0) 20 (oO aes + Hem ye @) Fyn (A oerssannan «) where M, N, and Q ere yet tobe determined. “The boundary conditions to be applied requir that the normal magnetic flux density yA, and the wegen! magnetic field H, are continuous at the planar ay (ay 8) ) “These conditions ead othe four simultaneous equations ofl Ryle = (ME HY sy) 46 by( Mer Me™) = Dea, @) (Me™ = Ne" yu, = Oe (48)108 In, levonagnets Insucton an Leop-L2ep Couting 1ip.3 ql oiete nemios Ene eos meyer eh mo ‘We can immediatly eliminate Af, ¥, and @ 10 allow us to write RolX) = (No~ Ya/(Ma* Yas 9) wee Nam A/ Cisse) (0) ana + Naahd Ye Ngo Raabd” co) where Ny = u,/(ijiya). As a matter of interest we may verify that YBdena™ His. (52) where the surface admitanee. ‘An equivalent transmission line circuit (Fig, 3) immediately suggess ital. The line of length d has a characteristic admitance N, with props Sion constant u, and itis terminated in an admittance N [Exerelge,Verify that for the e-layer model Ni 6) EXTENSION TO ALLAYER MEDIUM ‘We can easily generalize ou results forthe circular loop if itis located ‘ver an Aflayer medium (Fig. 4), The loop of radius o carries a current I that again is assumed to be uniform. The half-space now consisis of M homogeneous regions, The general mh layer possesses conductivity. and permeability ys The magnetic Here potential Fin this layer could be faken to have the form oi Fyne [PRO agent Bye Suarnaa, (58) where 4, and Bare o be determined from the boundary conditions a the layer boundries This esd to linea stom of 2A equations to solve for tenon W-LayrMacen 109 ett aw Fig 4 Cielo ve Maer nape ‘the unknown coefficients, The coefficient 8}, =O becouse only outgoing or ‘decaying fields are permitted as >> —n in the bottom region. As inthe solution earied out for the tworayer model, the coefficients AyeAyy—1 Byclyany 24 can be successively eliminated until we determine the coelicient Ry, Thus forthe region = > 0, we ae able fo writ 8) were RA) = (Mo ¥i)/(Mo* Yo) (56) wee 4 Myeanhind nem W, + Yotanhyd * au) 54 Nyashind Soe Tahoe” es) (9)10 8h evemapnetie Induct an L20p-L299 Coupling and so on, until Nyt May aah ty ey Tr May Mya my (60 Bxerise Show that if say, the mth layer has «vanishing thickness dy and an Fofinitely large conductivity then Yay Yas + onthe where ayy ithe finite yer conductance in mhos. Also show that if all layers were thin in hi ena, 2 yt Saude GENERAL ORIENTATION OF SOURCE DIPOLE ‘A basic Fenture of quasistatic eletromagneti induction problems is tat the Fields in the ar region are solutions of Laplace's equations. This mesns hat fils in the ae are derivable from a sealar potential. This fact allows ts to extend the previous development to account forthe ling ofthe axis ofthe souree magnetic dipole In principle, the same argument would apply for a source loop of finite size but the Formulation would be cumbersome, ‘Ths we will jus eal here with an infinitesimal loop or magnetic dipole. ‘We choose a Caresian coordinate system (x, y2) ith the earths sue face defined by == 0 and the air eyion by + > 0. The earth occupies the ‘space 2 <0 and it can be idemiiod quite generally as an Ab-layerstracture (Fig. 4). We begin by locating our small loop source or magnetic dipole With vertical axis at (03,0), where A> 0, The ansmission line analog fs ‘how in Fig. ‘Using the results in the previous section, we ean imediately write the Following espeession forthe magnetic Field in the ar eapion > O, =~ nde, 6) iil ‘earl Otoinion of Souce Ose am whee + fraarrererrscaryan} GM, (62) were Pe[seo-a] hefese-my — @) and the subscript vindztes hat we ae dealing expiily with a verteal magnetic dipole source But now it lear that we can ao ite (2) 0 the tenn eedr whee Te HAL A [Ranyere* asaya) «@) ‘The Function T can be identified asthe potendal of a magnetic pole at (O,sf) This fact follows immediately from the result obtained when we iniggrate bot sides of (64) wi respect toh from hto 2. We may assert that the potential for a horizomal magnet dipole (or small loop with horizontal axis oriented inthe» direction) at de same location is aegr (66) where 1 is given by (65). 'A simple generlizaton is to allow the magnetic dipole or leap axis to subtend an ange « with the vertical = axis but tll ie in the (=) plan, ‘The magetie potential for this ase, agin fr => 0, would be Ts nail cose Oe 4. ing EE ‘1 v a ey “This is quite general because the Cartesian coordinate system ean always be chosen such tht the = componeat ofthe dipole moment is 20. xereise. Show that RA) in (65) ean be writen as RIB) = = V/A Ye) = XO Ysingshao 12223 for m= 1,2... Me Then show that where 1 where Rp =[P+ (244 2/yQ°1V% The image is located at a complex Gapth 2 (42/4).1 1, Elecomegnetie inducer and Leop-Laop Couping EXPLICIT FIELD EXPRESSIONS FOR VERTICAL AND HORIZONTAL MAGNETIC DIPOLE SOURCES ‘The rns in the previos setion actually enable us to dedues explicit ‘expressions forthe magnetic Feld components for verteal and horizontal ‘ipoles located anywhere inthe region : > O fr the layered earth structure in the region = <0. This amounts to performing the gradient operations indicated by (64) and (66). Now. ater doing ths, we can, without loss of tener, st ¢= 0 o that the mageetic dipole (vertical or horizontal) i focated at height 2h on the = axis, The equivalent satemenis are as follows: for vertical magnetic dipole at (0,0), the magnet field H, for FD Oiis H= ead, «) MA ALLE fray %msgnya]. 68) R For a horizontal or y-irested dipole, also located at (0,0, the magnetic field H for >> 08 H= grad, (70) wee BLE femeemnana}. an In the above, R= [x7 + y? + (y= WYP? and rm (x? + 7 "Now it tars out to be convenient to introduce o Seale length 8 (in ‘meters to permit the nondimensionaiing of the various quantities. For ‘example, the integration variable 2 is then ehanged to g via g8=)r or Ne g/B, where B= +/8, Then we designate Rd) = Ral 8/7) = — Re) my Also, we define As Wf. “The choiee of 8 is somewhat arbitrary, but we choose it to be defined as o-{ecstane|"=(adsl” Using the above resus, the formulas for the magnetic Field components ‘re suntmarized as follows pot rd Sereno for Mogae ie Sources ‘The Vera Dipole 1. Primary tds Hpmsce(e— n/m, p= dey = m7, Dy where C= Jd te. 2. Secondary fields r= (C1 M/E Hes (CIP IINT WSs (CLP Yo \ “The Horizontal Dipote | 1. Primary feds Hymaca7/R* HE =3G2/R°— C]R, HE = 30 — /R es me § (336-9) “Te three base integrals are Tom [Raye Vek 2B de, T= [Rang PY 3a. Tym [RC yge eB ee 13 @ 4 ay 79) ™ m8) @) 0) ) @) ®) 4) sy 86) en14 1. Ecromagnatic induction and Leop-Laep Cousing ‘The integrals Ty, Ty and T, contain all the information tht is needed «© prodiet the performance of any exploration system that uses inductive Coupling ons The integrals may be evaluated numerically as they stand. ANALOGY WITH de RESISTIVITY [A special casei that in which we allow the conductivities ofthe various layers to vanish, Then the expression for Ry) reduces (0 dd) = y/o Co) where ++ tad YO ey anh, o + mgt an Doe naam | 0) Tey anh, o and soon. wil sy myth hay Na yA” on where iy = Mi) for m= 1,2,--<. We can see that the analogy with the de resisviy problem f complete if (ye) here is replaced by py, the resstviy of the Imi layer. The characttistic admittance 9, here is analogous to the ‘hraeteristie resistance my, in the de resistivity problem, ‘Of course, when dealing with the purely magnetic problem in which ‘afd is tine by (38), its desirable to eanee the A/ to factors so that Rad) = (3a ~ 5/0" + Sh (2) where cana So Fan Ad, "anh Se Fyn 3) 4) and <9 on, until ji + wil tanh My ai + a tanh fn Dosoay with ce Resaivty 15 ‘To bring the analogy to a logical conclusion, we define the magnetic courant 10 the pole by K= fagul da, which has units of volts, Then the ‘magnetic potential of this ples given by -[Amemanal 66 (FP + (2 — WP? and Fx? + (9 — 9712, This of course it jn accord with (63), in which the souree pole i located at (0.x, inthe (4.3.2) eoordinae system ‘We ean envisage A as being the elatie curent J per unit length flowing sizcumferentially in a slender solenoidal col. Here she cols terminated at (0.5.8) (Fig. 60). The cross-sectional aes of the coil isd “hat to be perfect expli, we show the electric analogy in Fig. 6b. Now an electric eurtent designated here by I, ows othe electrode which isthe pole location. The resistivity of the mth region % py, where m=O, 1, @ 9 Fig. 6 Magna yose este pole (b) over nerd halt puce2,.+2,Ms The electric potential at (x y,2)forz > 0is given by el or aaag($-Lamenmaanal, 67 where AA) = (ri 2/01 + Sr (98) here fet aah Pa F Zann” os) spp fe ata ic and son, until PH aga AS ey LOW-FREQUENCY RESPONSE When the frequency is sufiienly low, there is unoer ype of approxi- ‘mation that i useful [S, 26} Specialy, consider the case in which the sleetromagnetc induction ean be expresed as an integration or superposi- Sion of the elementary secondary soures. The nate of tht approximation 'S best examined by specializing 9 a nonmagneti two-layer earth model, ‘Thus we set dy = o> and j= Hy = fy inthe gene formulation. Then Ry can be wiiten as (=) #0) + tn — He OF HN Faye A a aK aye R) + (103) vere wy Oe' meer eay Yim Toy ee, = fost. Here we have assumed that disphcement curenss are negligible, Le, e490) and «x0 03, Now the frequency is saiciently Tow, we may EAS AA and edt A/D over the significant range of the integration variable A. But this to the statement that yff? and [y,| are muck less than 1, SRE ett beaks pereoneet eee eee. Under thse condos itis er that the mmeratr of {ton i approximated by 0 98 a> 0, Then ~ a] + const 1, [Prete 1 [P+ enemy (ny «tay ats) (16) ary }. 18) & x thy 2h teen a Scan etae]” [Pah sem] 9) [[Omar te HA]? 2g 4 2g ~ [0h 24a +7]"}, (20) Wher hy =O y= yes ty A y= In the case of a three-layer earth model we would have od {te Vina pel F 1 timer om , { ryt eth ch ] Al PPeaaeeeiny? [Peer ay] aateth caer [Pe eat + 24 ayy” 2, 42d, 2h 2 [Pe eae 2d re my m and too [tex nto] 426,-[oeraszaytee]"] a ee B([eenszat+ aye a [eet a2 #26384 7] ] +B (feonsags rater y?—24-26). 0 ‘As can be clearly observed, the normalized field expression yf. and fy sre geometiclly weighted contributions that, individually, are propor: tional to the layer conductivity ASYMPTOTIC LIMITING FORMS. “The integrals Ty, Ty and T, ean be approximated in an asymptotic sense ‘whee the parameters and B are large. To this end we develop R() in a120 1, @ectenagnts Induct and Loee-Leop Couling ‘Taylor expansion about ¢=0, as Reg= Field (124 where 12 milageF)), ns) “The integrals o dea with are then of the Form. Pf ate vo BB as (2, sod Om [are 9B bs. «7 ‘Then, o8 sng [lerrasa yay (sy ve eal dee that ahs _ 0) OE GaP and (3 Consequeny, (sy = wee 139 BOTT Oe ay om oy) ‘We sess that such results are asymptotic in the sense that a east one of vend B snust be very large compared with I. Thus the resulls have Somewhat lite tity but such limiting forms can provide a check on the ‘ata generated by numerical inlegration of the Ty, Ty and T; integrals ‘ua impedance Fomultion a Application of the asymptotic approach to 2 ewolayer earth model were we set dam 26, dy ~ dy ~ Ha ~ ay is quite straightforward, Using (125) and (72) we deduce that yt, Ay = 2/1", Ay = 10% where (+R +0 €)er0[ -(29'0] «34 GFR? =0- Fhe9[- 27] where bm (e4 ego, + kya) and D=24/3, ‘where 2 Ww ere (Of course, if displacement currents are neglected in the earth, the parame- ters E, D, and 8 are real, ie, f= 3y/4, ad 8 = [2/(6, ny)" ‘Retaining only the frst toe oH terms les to est (H)"2 wig? 2dt= nese ny + BF ana where e+ Wy/B and B = p/8. MUTUAL IMPEDANCE FORMULATION. {In most applications the source cam be represented by a circular loop of area S, with Ny tums. The field ofthe souree loop is then observed by the induced voltage in the oecving loop that has area S. ané Ni turns. Under the basic quasi-suuic assumptions that the separation r between the loopsve W. elocromagnaticInaucton an L20p-Loop Coupling Me ei sale — Fig 7. Two snl cpa op aed ie re pas be small comparcd with a wavelength and large compared with the Toop Giameters determination of the mutual impedance between the loops becomes relatively simple "When the loope are coplanar and are situated in fre space, as illustrated in Fig. 7, we ean deduce the reference mutual impedance Z, in the Following manner. Ifthe source loop carries a current of J amperes, then the dual magnetic Feld atthe receiving loop loeation wil be HS —C/ear, (38) where C= INS, is the current moment (ie, elfective 144). Now by Faraday’s lay of induction, the vollage o induced in te receiving loop wil be ‘om ~ fof total fx threading the loop % number of turns] =~ aHl53] (39) nthe mutwal impedance 2p is thus sven by f= (HaN35,5,/74) x 10-7 ohm, (i) where e have noted that py 4% 1077, We denote this value by 2 in what fllows. By def [Eyercise. Show that for coaxisl leaps ~9/ ‘conventions as Tor coplanar loops in ree space. Ml 22, using the same ‘The mutual impedance Z between pales of loops above a layered round can now be expresed conveniently i terms of the integrals Ty, Ty and Ty ‘The four eases of special interest ar listed here. In exch exse the loops are ata height above the ground and are separated by a distance r (Fig. 8) System L, Coplanar loops with axes vertical 2= 2414 2) can System Il, Perpendicular loop i, the Fist Joop has its axis vertical and the second loop has is axis parallel (o the ground and along 2 radius from the fis loop Ze", (aay 123 Fg. Fourie confit ens in he ntl npn eons. System IIT, Coaxial loops with axes horizontal: z= -2AL1+(B2/2\T,~ BT) (43) Sytem IV. Coplanat ops with axes horizon: Zn 21+ BT). «ay HOMOGENEOUS HALF-SPACE MODEL ‘An important limiting case i when the ground is homogeneous and when the magnetic permeability of the ground isp. Thus we set dj = oo fad e, = @ inthe general formulation. Then 3), ven by (58), reduces Simply to Rady= = M/+ my (14s) where = 0+ ge).‘The basic integrals are now given by ve (2 +24)" Tym (PST per ea gB ds, 046) Opens = pe-e44 98), aan te ys a ee 148) (e425 o whore = + n/8= Cong 9/2, any! ‘Tae appropriate expressions for he mutual impadance Z between 10 loops for sists I-IV are given by (141-048), spectively ‘Its useful to aoe that the limiting ease of zero heights for source and ‘observer yields closed-form expressions forthe mutual impedance expres sions. The results (1 6, 7,5} ae as follows: ‘Case. Coplaner loops ying onthe ground Z/Ze~[2(PA\]9~ (94+ 9 4+ PEM], (149) where Zo Bos Y= Cin" Case Il, Perpendicular loops: 2/2om (UD POK,~ hy J+ 4yr(Ke loki) + 161 4). (130) where 2 223 Jy fy Ky Ky ate modifod Hess unetions with argument we: Case I. Coil oo: 2 [a= (e™/7P\I2+ Lay + 57+ YP) 42— (12/72) 15H) where Zy= 22, Cam 1V, Coplnae lope wil nes pra [Bm lev /PP\R+I+ Pe) G/rr)}. (152) were Zs 2 Peete 6 toatl eat pct cnn ge on veel tea "These formulas are normalized so that ZZ appronches I es yr on a least for eases I I, and TV. OF course, in the perpendieslr Toop arrange- sent (ue. ease I, Z tends lo zee as yr—¥0, as it most "The magnitudes ofthe ratio Z/Z ae pled in Figs, 9-11 forthe four ceases. as functions of the parameter (61) “The foregoing closedform expressions are quite useful in generating mutual impedance data, butof course the results are limited to the assume teh tna ingens ose cope op wh vee re rei ote tt ¢3n/6= ep16 1. eaciromagnatic induction and Loop-Laep Couing Fg TA imposes tare I cas! ps or vss eis. omogercous half-space model. However there ace ese in which even a layered stl ieffectvely conducting hal space, s we ingen he folowing example (1 In Fig 12 we show a sounding curve ovr a layer of highly resistive lava ‘nist about 122 mere thick ha covers molten lava layer at feast 100 reer thick The fel data fit the eoplana® loop response ealeuated curve ‘ery well on the astumption thatthe cust has 2 negligible conductivity From the interpretation ofthe reference eurves such as are shown i Fi {ivi i sugges that 24/8039. Then since r= 6l'm We see tat = {2'm. Alo noe that for f= 10 kHa,r/B = 84, and thus the condctviy of the molen ava estimated lobe 0.50 mhos/. sata a ig 12. Feqscecy wang mage on te Kia Lave Lake, Haw by F.C. Fh ned tnd Lhe Ani singsopanar lap (rate comma. Fel di a Incusion of placer Cents wr INCLUSION OF DISPLACEMENT CURRENTS In many applications to remote probing by electromagnetic induction, the displacement currents oth in the air snd in the ground cannot be ‘ignored, We consider this problem here. ‘We first return to our static formulation of a small lop located in an ‘unbounded homogeneous medium (Fig. 1), ‘The loop has, sty, area dd and cares a total circulating current J. Now ‘we have shown thatthe static magnetic fields of this loap can be oblsined from 4, os and ve 4.= 55+ (1S4) where FstiesLaplae’s equation vireo 9 everywhere exept th ste Inet we fund ha Lad «1a. (*yyayye-n Fe dl oH (ranean, (1 where & = (72+ 2315 Now we wish to generalize this to time-harmonic fields where the frequency is high enough to worry about displacement curent and other related propagation effects. To discuss this problem in a general context, et the medium have electromagnetic constants, and a that do net vary ‘throughout the region, In this case we can deduce fom Maxwell's equa ‘ions that now ae ara" sn) (ass) Fg. 13. Small wire loop o¢ naga pote source i128 1, Elsromagnede necten en Loop-Loop Coup ‘where F now sass (Po yFeo. (59) “This lads ue to wate the dma slation as Fe Nl goto HEA (ayaa a, (0) where «= (024 72)! and y= [(0-+ iesiqus2. The expression given for FF reduces to the nati form a5 >0 and it clearly satisfies the Helmholtz ‘equation as it should, Furthermore it decays appropriately 26 Ro. "For thie raler idealized example we can see clearly how the static solution emergce as the limiting case of the dynamic solution 36 |yR] ‘approaches zero, When we are dealing specifically with feds in ai, the conductivity can beset equal to zero, whence Y=iB with B= iGmU=Ie/o, where Ay the freespace wavelength, This tells us that the stati solution i 2 good approximation if AR-« Lor, equivalently, if Ry, This is one of the jusieations for neglecting displacement currents in the air at the ‘outset and dealing with Laplacian fields in that region. Exercise. Show that in spherical coordinates (R29) the dynamic fields ‘ofthe smal loop source ina homogeneous region are y= (14d [2eR\(L+ YR ye" cos and Hy=(LdA fay + YR + PRIM SiN. If we wish to consider mutual coupling of loops over a stratified ground ‘modcl in the general ease in which the loop separation is not small compared withthe freepace wavelength, we must reformulate tbe prob- lem. The procedure for doing the follows the basie work of Sommedeld [08]. The ease of « horizontal electric dipole is only sighuy more compl- fled than that of the vertical dipole orientation, and the same remask applies to magnetic dipoles. For general media the mathematics is actualy the seme for both clecsie and magnetic dipae types. ‘We outline only the general solution her and refer the reader toa recent text [39] hat summaries the site of knowledge ere we will consider first a z-diteced magnetic dipole (ie, small current caring loop of moment 144 with vercal axis) located al fver the #laer half-spice, The station is basically the same as that Showa in Fig. 4 but now the dsletie constants gy (m= Q1,2,-. 4 Af) ‘re considered. Canara Sohaton fr Hetizanal Magnets Doo 29 ‘The magnetic fields in the air (2 >0) are now obisined fom nage «ast sa (16) tee re 08) wee ty (+ Ye) (64) for m= 0.1.2, ..65M, Ra Ue + 14, and r= Yay + Nata 2°), As before, Yn MR Fae ND Ha, (16s) for m= 1,2,.2-4M = 1,and of course Yi Nae (166) ‘The explicit expression for the input admittance Y, at £=0 isthe same as that in (57) but now to =Limolen + hye) Which allows generally for displacement curenis ‘The reflection facior (is here designated with « subscript L to indicate that the electric field (ie, £,) is perpendicular wo the magnetic Field components H, and H.. Actually R, is furctionally the same as the reflection coefficient fr a plane wave with E (perpendicular to the plane of incidence) parale tothe boundary. Other equivalent designations for this case are TE (Iransorse elec) 0: horizontal polarization. kis quite clear for this verGeal magnetic Gipole (VM) excitation that the earlier quasistatic results are raguined i, = O evenpwhers, GENERAL SOLUTION FOR A HORIZONTAL MAGNETIC DIPOLE ‘When dealing wit the horizontal magnetic dipole (HMD), the situation {sa bit more complicated. Now we regard F as-a vetior that has ronesro ‘components F, and F, when the dipole is oriented ia they direction. For130 ‘i. etectmagnticIndueton an L20p-Loop Coupling any of the piccewite homogeneous regions, we now have Eo ipacule «en a Hea (-y+ anda 08) Here api sep atte sous w -a( =o (169), tere teats 1 a ar ay om, ‘The boundary vale treatment ofthe problem now involves matching E,, 5, H,, and HM, Field components at each of the interfaces. The fai ution forthe region = > O can be written a5 + [Ra@ewemaeyal am and 144.1) pn asm IUD FT aay Ri@y]siane seman Hat a. po ROE RO poryentrban, my roars Exercise, Show thatthe vertical magnetic field of the horizontal magnetic dipole for 2 > O is given by saa _@ [eon nse onyAZAN) 14s | Game BO an). (73) Which does aot involve Ry). Tn (172) we note that RQ) = (Ko~ Z)/( Ka Zh amy were Ky= taf tegss Bau Mal (n+ Het) for m= 1,-0.0. Now ow x 2astt Kalah tale ee a orma L (178) {¥ ndvetion na Conducting and Permeable Spare 131 and Zy~ Ky (76) Funedonally, Ry isthe same as the reflection coefficient fora plane wave with H (perpendicular tothe plane of inidenee) parallel to the boundary. Equivalent designations for this este are TM (iranserse magnetic) ot vetial polarization “Then, for example, we easily deduce that om [Pasay ern 4 el 06) wit Ersoy nayjen 02 aah a7 ‘When diplcement currents are mggected nth i, = 0, andthe above formula fr the divergence reduce o awe ELLE (mayemrvgaeyea), (78 where yA) is defined by (56) except that 9, 1 NOW ay + iad for m= 1,2,+-- AM, Le. dgplacoment eurens inthe earth ayers ae incded here, ‘The coresponding potential Bin this quasistatic mit i, of course, aiven by $= —divE, which is fully consistent with (71). ELECTROMAGNETIC INDUCTION IN A CONDUCTING AND PERMEABLE SPHERE In most exploration ventures the target i « bounded body, even though the background response could be that from an extended region (sen & layered half-space). To give some insight ito this problem we conser the ‘most idealized of all bounded bodies, chat is, the sphere. A. quatesiatie formulation is adopted that sratly simplifies the analysis yet Fully accounts for the induced eddy currents. (See Appendix A fora general analyst of the sphere problem.) Also, we allow for the nonuniformity of the exciting field, which eertinly is the ease when the source is & loop eaerying an coscilaing current ‘To allow further Flexibility and generality we also allow the sphere te have a concentric coating in the Toem of thin conducting shell [13},12 ti, elecvomagnatic induction and Leop-Laep Coupling Pin) ig Soi spe enn y conan hel xe by mags pole ‘The geometry of the situation is sbown in Fig. 14, for which polar coordinates are choten wih te slid sphere of conductivity @ and perme= ability bounded by r=a. The thin concentcie shell of thickness d and conductivity o, i located at r~ b, The source is considered to be a rmagnctic pole whose suength & varies as exp(iu). It is located alr = and =o the polar axis ‘The primary magnetic potential ofthe source is given by Up= K/(erinR a) where R i the distance of the magnetic poe from the observer at (78). In terms of spherical harmonies, K_ Sy) Fahicat 2,7) Palos (180) hich i vid for r< This expression is uly analogous to the core $pondigg veut forthe elete potential of a cament sure (elec Pe) n'a homogeneous medium of resstvity “The magnet eld components asoiated wih the magni ole source se obtained from He — pad cst) or, explicily in spherieal coordinates, avy 4-28 (182) and avy = — 2 183) ‘As understood in any quasi-static representation, the significant distances Ge, R. fete) should be much less than the wavelength in the external 4 neon in 8 Conauing and Peoabie Sphere 133 region. Inthe present coment this means that [yo I, where a isthe propagation consant Th the subsequent analysis i will be useful introduce a moeified potential orsteem fanction yy related 19 Uy by 1 Be 4-- (1s Rego only that fy vanish > 2, we easy deve fom (180) that = ESF) eT ele) (285) forr-< 1 Miao a simple matter to deduce that dy ssc he equation 22 18 ay Sa + aha H(i) -0 (186) ‘Asa consequence, we may write dy ES sy 188) ‘which are fully equivalent to (182) and (163) in a potential sense ‘We are now in a position to deduce the appropriate form of the Field solutions in the repions external to the solid core. The cesullant in the reglons a > band oo 3 [ac +8} cco) ow fora < F< b. The coeicients 4, 8, and C, ae yet be determined.434 IM. ectromagnati Induction end Leop-Loop Coulng ‘We now need to investigate the form of the solution forthe solid core. Hee we do not make the quasistac assumption because ya, ware + (ou! is nota small parameter, Bul, as indicated, we do not neglect isplcement currents in the core since in effect they have already been lenored inthe extemal reson. ‘eeause of the geometry of the problem, we can write Maxwells equ ions forthe core region F< 28 Eytsnde,)= ~iner nt, oy 2 (re) = matty (193) ana Levy Sth = rey 099 Here ote that 11, =0 and 3/0 ‘We now define our stream function aecording 19 1% 195 & 0 “This eoie then requis that Bt hy Blom) 196) forr0) it quite signifieant, particularly forthe imaginary part of, where the response is enhanced Although We have considered this problem in the context of a single ‘magnetic pole excitation, the rvuls are valid fora solenoid of finite length it we merely imagine the later to be the superposition of equal and ‘opposite poles located at the ends [29 A special cate is that af a magnetic ‘ipole of any orientation that ean be considered as the ease in which the solenoid degenerates o a smal oop ‘The shll-sphere model i, ofcourse, highly idealized, but i is a conve nient scheme for giving insight into the nature of the cectomagnetc response of a massive ore body that has a eondveting caver or shield. Such ‘models have been used to discuss the rather controversial proposition that a ‘conductive overburden wil actually increase the response of a. buried ‘massive ore body. This neasvesreening was discovered by Negi (20, but itsefect can be atributed atleast ply othe way in which i define.18 i, Bestomagnte Insveton and Laop-Loap Couping REFERENCES 1. 1.8 Wal The mage inde oer he uioty ea earth, Cot JP 3%, sesh 2, Tf Wt Mata indotins of sic moe ah Ca J hy 380-482 ‘os 3 oR Wi, adatin by am esilains manip over «woe gama, Apt Sat Re Surf) 77-80 (9S ls te sumer a Ey. (0a om, pe seh ace — Hd 4. Jat Tani sd refleion of eeromegate wns ne pie ‘aia moi J Ran Na Br San 5) 4 Oh 28-2929) Tipper es rl ston bth eal avs etl apt ioe oe a Se ‘hifi mem testo heme as Kero’ 196 pe (6h 5 LR Wot anne on th setenagnte expos of stale eth Gaps 2. rie sumer ef thea the igh ie Gapin 90-67 19. 1 Riva Muna casemate oping f hp nr # Aamegeeis eon Gopinath c50 Ratu bys fossa ning gnc pl ner» eondeing omnes Pane Gop a 3 19 (00. 5. ER Wal Wav Poppin Tc Rep, Oxo. 18 0 Kf Spos tnd ue Doc Stil povnd consents fon he ta ‘etn acl owe HE To darn Pop APSO wh 31-32 tm. 11 JER Wa Fh ts oat pl ors ier th, EEE Tne “mca oe NP 90-93, ta RW in A, Faler aed pmsenatons 0 he meu decom “eupine ee o's ores oh cece 22-22 (TL 13 Sa Bama ni ee hs dy = ‘omnes ha sc Gaye 3038-790 14, Fre snc Pes sh on ang see pl ae woe cary {lotion ond alse detonate CaS Ms 6 Geena ts, stb ans it, Wad deve wanting a «yee cart wi «hoo ‘aureedok, Gites 3 60-09 1970, 1 PRT A lao tod freon te etomagnte Fl toe enol iis, ed US5R apy Sa No 2919) 17. R'R'Sain Comply se El yo ed ers may der ‘Sees Gap EDITED 1s, RES: Osco span nn of mae ari by eta fete ma Can ie eas 1s. Orkut b Phd. Paoan Comput of pe cine fo eta tsp nen soins tamestng ly mesa dal emer Be 20. RR Vo td ©. Kote. Ante onthe Iap ee metodo sane Sstomptc ung caves ep Prope 270-1 (TD 21, KR Se apd Us ole trespass fling mae is ‘had ant mle onding erat Ss Cat Pp No THOS 1-8 CD. Petronces 139 [Mga 7) and Tyme eid wie ri eee as, Ako et i "i. (shuld be sn. The can appre me Wate 188 paper wre the Ge ‘os esate along th at we 6 A Meme Mn Te Eso i Sng Ip, Pn. [LR Wallan L Campbell Fs os mogeete ple nena smi oniueng mn, Cp er 8 167 1953. 1, Mec hp po Be Coe ope 1, 20-26 $y, HF Marion, a. H. Wan,Eletomasac esaoat lop sue of ‘en, 8-96 (190, SD. MeNailElevomspnts teen omdsiviy meaner atl indo rember, Tent Net TNs, Gata i Ming: Ontarin Caan LET 1, [si (MeN! eae an inte pce fo Re wd he ey nd ip sprain i allen wah hut IDM. VO, an 1) ae a Pe porch was actly erred by Wal [in 9) BUR" Falr and J Wate High unc sosomsgntic coping bewcen sa ‘pla ops vera nhaogencont pn, Gir 3 9971ON (TD. 24, 17 Dunit, GV Ker aed acto, Comp aoe inteataton of En ». ‘lcm sauaige vers permet in psc 35-745 10 Ite Wak A conecng permeate ee he rc of 9 a Sry ing aren Coe 31, 0°08 (8 TG" Neat Eeronspne acenng dae 10s donald pba zone over « ‘sien ope Bh ie al sg mip. per 1, 66-672 SV. Want. A meio for menting he eat condos of damon dl oe Cece . ates en Sren iseena) nun Gans Pope E-hating AF es nM. Pg Othe asin a elen by ‘ale ad Campoc he fei ton ola opel mane inti cooducing medi Gay. Reset 36) ‘AK Sis. Bectomaene lr a an cing mayne pel onan aioe ‘iho 6 IER Wet enone We Shafi ob, nd Pra, Oxo, 170, WAL Linker forage spi of te How Me os 170 AN Siecle oP ft SU Wi. Characters smnmu oy th in Ata Tew (RE Cali nd Fo acter eb MeCram Hl Ne Yor, 1 FR Boer usta rome fii U3 Navy Rep US. Noy New GIN kr nd FC mh Hl Ml Gna! Pat. ERR Eton ig a rg i, Ma. St 7 1onper IV Electromagnetic Coupling for Grounded Sources inTropucTION In earlier chapters we considered various aspects of low-frequency cut- rent flow in conducting media. When dealing with the conductive eneriza~ tion of the medium, we assumed thatthe [elds could be derived from the ‘gaient of scalar poteaial in an anallesl sense. This is justified when the wavelength is sufficient large, 0 that the Helmholtz equation (ov Wa0 for the fc funtion y ean be replaced by the Laplace equation VY) =O in roureeree cepions. This is what we called quasistatic theory, and its validity isa separate issue that depends on whether or not displacement currents in any of the regions can be neglected. “As we indsated in Chapter HI, the key parametr is, where y isthe ‘complex propagation constant and / i a scale length in the medium, Clearly i [yl sulficienly sal, quasi-static descriptions are vali. Tt is our purpose hete to analyze the fields for various (time-varying) source configurations that allow for both conductive and inductive enegi- zation, Kealized homogeneous of layered earth models will be considered However, to allow some useful generalization ofthe results, the electrical properties of the layers wil be assumed 19 be anisotropic. ‘The prevent Lnalyas is considered to be a suitable bass for dealing wih eletromagnetle ‘mping iv induced polarization methods of geophysical exploration ‘Asin other chapters, the development will be self-contained. Related saterial of a general nature i found in eferences [1-9 specific citations wy also be mede ns appropriate. ome Gane Considerations at ‘SOME GENERAL CONSIDERATIONS: ‘When dealing with conductive enegaation of the medium, we need to understand the basic concept ofan eletie dipole source. Ii rutful here to first deal with a small current eletent located in the region under consideration, “The situation is shown in Fig. 1 in whic the current element at length 7 4s located atthe origin of a spherical coordinate sytem (7.8.8). A uniform current I flows from the bottom end at C, to the top at Cy. The entire current / is made to Flow from C, into the medium and then Back into C,, In other words, Cy and C; can be described as the souree and the sink, respectively. Ifthe medium is leally unifom, the source current emanates radially from C, and the return current flows radially towards C, Its thea easy 10 show thal, provided 7, the resultant curvent densities at (1.8) seis) J,=(leos0)/2279 0 and (sind ne @ ‘The comesponding elect field components are obttned from Ohn’s aw, which we write as Ja (94 eolE, a ‘where @ and car the real conductivity and permitvity, respectively, for an angular frequency «Here we can identify (@ + ew)" wth the complex resistivity pio) used in Chapter TL Also, forthe momeat we treat @ and ¢ as being isotropic, Then it follows that E,~(Ils0s0))/2n(a cay? 4 and Eye (sind) /Ae(a + fay? 6 Fig Shor canetening det nee ns homens ntae 19, Sacromagreie Cousin or Grounded Sources are the resultant electri Fields produced by a current source at and a ‘eurent snk aC, separated by a distance / that s small eompared with Tis now wsefl to calculate the magnetic fel al (7,8) for any point from 0 to 2. Symmetry ofthe problem tls us thatthe only component of the vector magnetic fad is H,. Furthermore, acconing to Ampere's ln, Ey which is mefely a statement thatthe Une integral of the magnetic field {round the erie at (74) is equal 1o he total current passing though the ite, The righthand side of (6) integrates to ({/sin'0)/2r. Thus we ‘oduce thatthe magnetic fed associated with the current source and sink (Ge, euret element) is aersindit,= doe, 6 H,= (Ihsin 40% o iis important to realize that we have obtained expressions for the clestrie and magnetic ils that are rally vali oaly a the frequency tends fo ero, In general, the feds must satisfy Maxwells equations. For the present configurtion these can be written in spheriel coordinates as Tawar yy 8 (a ioyrsnd 6, ° BH) = -(0 + ieayrb (19) “These are valid everywhere excep ight ar the source current element. Here seis the magnstic permeability of the medium. ‘Now anne ttt the soroundng medi is homogeneous everwhere and lia , ond, rom (00) oder to ota fe 7, ® G+ & be a Zeng & aby Heaney vey a Wwe now se y= fein, @ ese tat Sen-y= v4 3) Solutions ofthis gestion ate of the Form fa(aeyle ne a4) Some Gere Consiserntons 13 ‘where ip constant, Clay, only the soltion withthe upper signs is retained if iso vanish as p> (ah ms for Rey > 0) ‘We now can deduce tat the reguired expression forthe magnetic Feld that is vid tall distances is igo 08 are Had 1+ 1 as, This reduces othe dst form (7) ly], yeti als satis (1) for all values of yr “The coresponing expressions forthe elect ill components are most easily dedoceddiretly from the to Maxwelian equations (9) and (10) {sing (15). Then we Hid that E,=(Ieos0/2oP\ 1 + ye (19) and Ey (Isind Ane \(L4 ye PAE “ [As they shoul, these expression reduce tothe required forms given by (1) land () as yr->0. This tansiion back w potential theory has been iseused in detail elsewhere [8] i is quite common in electromagnetic theory to introduce aurilary potential funetons. The electric vecoe potential A it good example. {0 the present context ofan electric dipole, located a the ergin ofa Gartsian coordinate system and oriented in the zditeeton, A need have only a = component, Thus, for example if we choose A= (Ul/Aare™, «gy we recover (15) i aA, n= ~ Bsns (9) ‘This might be regarded as @ heuristic method for introducing such an auwiliary potential fonevion 4, that hat the ment of being tiple (Le, spherically symmetric). Funthermore, as can red be verified, 4, safies (29) en18 IN. Bactomagnte Couping fr Grounded Sources [Now if the eureat clement JI were located at te ergin but oriented the x er y direction, then clearly the vector potential would have eompo- nenis 4, or 4,. eapectvely. In either ease, they would be given by the riohthand sid of (1, "We are now able to-wite a general expression forthe vector potential A {or any distribution of driven currents with veetor density I(x y2). Thus a fff where the integration is ovr the region containing the source, Here Rafer ath een ev)™ @) isthe distance from the source point (x ys2") 10 the observer at (x. ‘As can readily be verified, the Cartesian coordinates of A. indviialy satily (A, A tae, en Ate d= remembering, of cous, that we are dealing with homogeneous that we are excluding the source poins). Furthermore, given the function [A fora specified distribution of eects current sources, we may determine the corresponding fields from H= teu, 2) and =~ igo ~ grad, (5) where v=-stgava a isthe salar potential. The coreciness of (25) and (26) ean be demonstrated by applying the results tothe electric current located atthe ogi, in ‘which eae the appropriate forms such a5 (15)~I7) are recovered. 1 i8 ao Useful to note tht the penta! in this ease reduces fo the expected form Va (seos8)/Ae(o + ieay™ 8) any ‘Actually, if we had postulated (25), which relates H and A, then the subsequent result for E follows from Maxwell's equations and could be ‘se Gane Consort 145 writen as (e+ mayb (7 + gradaiya, 2) where the quantity on the left i the diven current density "To round out our dieussion of bate concepis, we might now write ‘Maxwells equations inthe general ease that ineludes source regions as cu = — jut 0) and cul = (0-4 ee )E + J, en, where J is the electric current density of the active soures, Equations (30) and G1), of course, apply w any inhomogeneous sotropie media, But if we ‘assume tht the region is homogeneous, at leat ina finite repon, then we ean assert that (3) and (31) hol, and it follows that (Ae ns, cy whece aguin we regard Vas dhe Laplacian operator that is applid to the Cartesian componens4,,4,, and 4,..Now ifthe medium is both homoge- ‘ous and unbounded, te sction of (3) i given by (22. OF course, more Sophisticated methods eit for obtaining solutions of (32), but the pretent approach is physically sesfying. verse, Verify hat ae ras by working out the vector operations in Cartesian coordinates In many applications, we sall be dealing with cureatearrying wites of, finite length. The station i illstrated in Fig. 2, in which the impressed or source current is (2°) along the 7 axis. Here 2 is the souree coordinate, ‘where we bea in mind that ("ip general isa function of: Fg 2 Leer eretcarying slant tel in hoger ondetn: pce46, 1v. fteevonpnatc Coun for Grounded Sources [A te risk of confusing the fsue, Cartesian coordinates (x. y.2) spheric cal soorinntes (70,4), and eylindtial coordinates (p.2) are shown in Fig. 2. The radi coordinate i used here to avoid confusion with . OF cours it should not be confosed withthe restivity ofthe medium. Because we are regarding the curent aa filsinen, the volume integration indicated in 2) reduces fo the line integral expression For the veetor potenti at: @) where vy and the integration is cacred ovt over the full extent of the finite current Clement. OF course we cin also regard (33) as the superposition of & Sequential ine of elementary dipoles or current elements of moment Tied", But the important thing to realize is that the expression for (33) fceounts forthe curent injected into the medium at the terminal points ‘=z, and =; a5 well at forthe induction from the eurent along the wit ‘The currents atthe ends of terminals are H(z) and 123) It 1G')= 1 const throughout, then the eurents flowing into (or out of the mediam ave equal and opposite mn eylindrical coordinate, the electric ils ofthe finite curren element Pelee eee sy] =o PA (0+ eae = ae fe and (0 tay (a a 35) ‘where A, is given by (33), The sole component of the magnetic Feld is cbisined from 36) ANISOTROPIC MEDIA CONSIDERATIONS In moving the electial properties of geological strata, it is often convenient lo characterize the effective resistivity in the horizontal direction by py and the effective resistivity in the verieal direction by p,. In Chapter Aniscopie Madla Ganscoratons var | we indicated the physial basis of this macroscopic description in the contest of de resistvicy. The same ies cary over here for which AoW py and p, become complex resistivities. They are related to the real conduce ‘ities and permiivies by = (ont fo) on and Pam Cou lel! 8) In what follows, however, we will contret our notation and replace 14+ logs by dy and a, + Jew by & OF course, and é, are now complex tnd Frequency dependent in gener ‘Obr’s law, relating the complex phators for the curent density J and the eleatie Feld E, is now given by Ln 4En Lm hb, and La 86, in terms of rectangular coordinates chosen so thatthe = axis is vertical. In tensor form we ean write I=@E, 9) where 00 wala ao (40) oo 4 isthe tensor conductivity that has simple diagonal form. Maxwell's equations fora source-(re region can be writen curl = (3)E my and . curl ~ gl (a ‘Asin isotropic media, we inode a vector A such that cul 3) and 2 Scalar potential y according to B= = igyon grad. “y It we choose diva +g", 5) ‘hen, at least for homogeneous media, we find that (46)V4, — 7-0, “ where ti ‘A slghly more complicated equation for 4, is Pam a-dog buen Fvann ah ‘hich reduces to Helmbola’s equation only in the ease for which & = 6, HORIZONTAL ELECTRIC DIPOLE EXCITATION We now wish to dea witha problem that i evant tothe electromag netic sounding of the earth in which time-varying fields and grounded ecirodes are employed [10-24} Specifically, consider a short electric “pole of eurrent moment Ic thal located in horizontal plane a height over ® homogeneous half-space model of the earth. Later we shall eneralice this a Tayered model ofthe eat ‘The situation is shown in Fig, 3, in whieh Cartesian coordinates (x, 342) are chosen a0 that the uppe insulting region, having electrical properties {fs occupies the region = > 0.7he source dipae / dislocated at: = b fn the = axis and is oriented in the x dizetion. The lower homogeneous half-space possesses a horizontal complex conducvity 4, and a vertical complex conductivity 4,. To simplify the analysis we also assume tht the ‘magnetic permeabiity Of the lower half-space is, which isthe same as that ofthe upper region. Tn order to sats'y boundary conditions, the vecor poletial A must have ‘ component inthe direction ofthe diple (Le, ,) and one in the dizection ‘rormal to the boundary (ce, ,). wef, Crt fo Stn ete le al stp stig Using a certain amoung of hindsight, we wouK! be led to write forthe lower hall-space Aga Ld [Feylaboyan 9) and Ante 2 (Po aaaedr, 0) where patty" and where F and G are functions yet to be determined. These particular representations immediately allow us to write ivan Mey Be tie (oly 8 nt on A, sii 1, my =e yyrao, (2) “Then, sing (8), wea find tat G mast saisy BE —orer rena n¥f. 3) where K = 64/6, "The term on the HpAchand ide of (3) is important. 1 coupes the solution for the 4, to that for A, via the anisotopy. This coupling was ignored by Wynn [23 ‘Using elementary methods, we now dedce that solutions of (52) and (63) must have the Ferm FO.) = Aye" (54) and G02) = BAe” — (W/O, 65) where w= O24 y7)/7 and o= (K+ 3), Here we define Rew and Reo >0; thus Fand @ vanish as 2-» ~o. “The general forms forthe solutions inthe upper isotopic half space can be considered as special cases ofthe form inthe lower anisotropic region in Which 1 and y°-0 78 ~egugu2 However, now we must account for ‘ho eouree singularity. Thus, ac 2» and p> 0, we 208 that Tis g-yat, em, (56) Aa? Ab where R= [p? + (=~ W711, and where supersciptp again denotes primary:150 IW. ieoromagnsc Covsng lr Grounded Sauron ‘The appropri tal epson or his primary pete i pom Lt (7 A gondem ny anal fo hers naainytr om wee y= e+ 8) Weare ow ale wite the ee ine epresentaions fre vo vegan for at Ha (4 ye" Ace Hit [AQyetsenyad (8) Eaoreaaoren 68) mA eM He Ry yea) (6 al Rye" aod — (60) ed Eliseo “The common factor /ds/4e has been included for convenience jn all the above expressions. The functions 7,0}, SQ, A and B(A) are sil unknown, "Expressions forthe felé components for = <0 can be ebained from Ee) @) () (65) (66) and orzomal Becic pols Exatation 451 ‘The sume expressions apply forthe upper half-space => 0 i Ags Ass dy tnd" are replaced by doy, Aa fey and 7, respectivly “The boundary conditions at = 0 require that the tangential Field com poneats (Le. Ey, H,. B,, and H,) be continuous, An equivalent require ‘ment is that for? do Aes (6) Bon | Ody tee @) At Aa 7) and m Bear in mind that the lefchand sides of these four equations ate the limiting forms as 20 from negative values. On applying these boundary equations to (58)-(61, we really deduce that 4ay= Aa Rane” ™ nye @ Sa)=[ 8a) aey]em™ (74) and [dete Leta We now have te following explicit forms for the components of the vector potenti for z <0: 80) Ha. ay 79) wer vu Fin “These exact itera forms can be used a the stating point for a numberof studies of wave propagation along the ait-esrth interface, In our stationwe 1. leoranagnte Gousing fr Grounded Sources ‘we are primarily interest in th case in which significant distances (Le 4 and fare stall compared with the fre-space wavelength, In other words we neflet displacement currents in the ai it we set yp» 0). Then Slay i) and (7 simpy fo 1d (A puget Aga Bee (A exer saman ~ and ante reget anan a9) ‘A case of special interest is that in which ->0, corresponding 0 the relevent situation in which the dipole is right in the air-earth interface. To "= in the above two tnd d, inthis case can be expressed in terms of modified Bessel functions vith comple argoments. However, these rather complicated forms are not needed if we restrict atention to the fields in the interface. Thus if both F=Oand (80) aod peodt. whee Aen Bie [A aman cn ad aa 04 94 (E+E). ®) The uae pti wt ELS 2 tone where b= (2K + 7) and = (2+ yn spite o its appearance, the integrand is well behaved at A=0 when we note that o~u—>0 as 20. ‘We ean aot decompose (4) into two parts such that oot de 5) Heczntl Eeare Dipole Beatin 153 were Ide BA. dee Hoa, Bb Bea? to and # Fle-orenan® 7 Hey apps if te htpac opie, = 40). By making use of the integral identity (e225), [rotenone 2 68) itis not diffielt co show that we wits [et yw) K"exp(—yeK"")]. (69) ‘The integral expression for 4, given by (82) etm easily be evaluated in simple form by nosing that M+ w= Na ~ Afr? ©) ‘Then, using ony elementary forms, we find that y= (Lds/2myp'V[1= (1+ whe] en “The resting fils, and E, inthe itera = = 0 are thos oblained from (80) and (81), using the preceding explicit forms for ¥ and a. In particular, we look a, which can be expressed inthe orm cy where 3A = GK yexn(—vK-"*)— (+ oyexp(— ye}. (93) Cleary, in the isotropic limit, K= 1 and A vanishes. Ite evident that A iso disappears for the anisotropie ease if yl > I. On the other and, the low-frequency limit yields cy), 4) inere terms of order (yo) and higher have been neglected. Inthe de limit, in which yp-0, we recover the expected reall fora horizontal dipole on resistvily p=158 Iv. acroragnate Coxaling lor Grounded Sources 35) 30 p<] Zz 20 $ oe ots it. irel 4 Then urine perenne sett ncn cote pee R= fee In dealing with the ae case, we now neglet displacement currents inthe round, Then 8, = 2, and 4, = a, are considered purely real, Thus yptrelet/* and yl =(oyngo)'P. and, of course, K = 4/, i real. To illsale the nature of the Frequency ‘Jependence of the iansfer impedance 6,/7 for this ease, the empltude tnd phase of [1-+ A] are ploted in Fig, 4 as a function of [yp (The inset Shows the geometry in plan view.) The values f the anisotropic ratio Kane Shown on the curves, Ax expected, when A is equal to or near uniy, the frequency dependence of I-A is very small. On the other hand, values of XK tha fer fom {ead to significant frequency dependencies, 2s clearly ‘evidenced in Fig. 4. Ax suggested earlier, 8 measurement of the frequency ‘dependence of the ratio E,/7 should yield useful diagnostic information bout the anisotropy of a conducting half-space (16, 1] lot var WL aerAnsoveple Madam 185 HORIZONTAL ELECTRIC DIPOLE OVER N-LAYER ANISOTROPIC MEDIUM isnot cific: fo generalize the foregoing wo a stratified model of the ‘earth [19] Fig. 5). The horizontal dipole of curent moment ds is again Tocated at = with respect to the earth's surface at 20, The lower half-space consists of W piecewise homogeneous regions with inlerfaes at erty -nvty-y- The thickness of typical Iyer it fy and it 8 bounded by 2", and z = z, In this mh layer, the comples conductv= ity tensor (4,) has the diagonal fore re 00 @=|0 oO}. 0) o 0 @ where dy, and 4,, ace the complex conductivities in the horizontal and vertical dictions” respectively. ‘The upper region = >0 has a permitivity «that later is allowed to ‘vanish when displacement currents are neglected: The whole space i laken fo have a magnetic permeabiliy jy that can assume the value for Tree space ‘We now proceed to solve the problem in the same fashion as forthe homogeneous anisotropie halspace, We shall deal with the equired form Ae ce See re ey Fao ae Fig Hort! deci dl over an yea algae (The gn ge isntined by = abotforthe fields in the mth yer. Thus we are led to write H=ala, 6) an E, = —inqwa, ~ grad dy on wee Ay and te contin By ‘eA, oat, om Ween fn a calf) <0 % and ‘ Py Goa Wee Ag= (100 wee YE ino, In the upertattspace (Pvp =0, oly ore j= x9 onep a the suse ie He 6 eon When dealing witha horizontal dipole we can sel Ay and yg eu 10 0. For the mth layer we write a= [CF RV 02) and Aan & [PGA vaanya, (103) where F, and G, are yet to be determined. They isfy equations with the fume forms as those of (52) and (53), but of course we now add the Appropriate subscript that ranges from I to N- In the present case, the general solutions fr F, and G, have the Forms = Ag) END 2) + en(N) EXPL — 42) (104) and G,, = BylA)EXP( Oy) + BA) EXPL — 83 = (HPPA eHOC 42) + (HP? _AAV EAPC), (105) w= Qty)! (106) and 5 (0 (107) where Ky By / ba ‘These forms hold forall W ayers, except thatthe coefficients by()) and ‘ay() ae O because the Fields must vanish as: > ~O are then ‘oblained from (43)-(45). Simplifications in various Uimiting cases ae posi- ble in numberof situations such a6 tha ia the feild eepion in which p is lage compared with a freespace wavelength (.e, yg). In such cases the integrals can be evaluated by the modified saddle point method Here we shall be more interested in the near-field resin. Fin, we mole that the fed componess inthe interface = source dipole also in the Interface i, = 0) are ebained from for the Ela Me fat and en -e (2) Without approximation fo hi case cm Hie [2 (04 Ransddoyed (an and Bag 5 Mae We [3 ars HES BL" + B.0d era Ra fone. om Equation (122) can be reduced to 1a a Zan | 100) ate Clete ae where we have made use of (105) and (114) and the fact that (08 = 22) y= 1/0) = Fn. (2%) We can now apply the quasi-static Hatin which dplaement currents in the oppe balfapace (es thea) are nepeced. The tem feu, he ‘denominator of (123) drops out, and elsewhere i epaced by A. Then We Find that Ligon (25) Las a. of _ te we S20 [ter and = hie p= ac Hie [Moth (2%) ‘These reduce to the required forms given by (82) and (84) when the lower halspae i enely norm Gs in which case mgt and 2, o7 Act, we hove aot cored throh the arash for he magnetic fied componcns nthe nae These lad fo smd neg forma, OF these ie simples cae that forth vereal mage et for! tis chained fom Pos Me. = «ay where dg, is given by (126), Ie 8 sciking but not surprising Chat Pa. 8 independent of 8, [MUTUAL IMPEDANCE FORMULATION Here we wish to derive expressions for the mutual impedance between grounded circuits of finite length that le on the surface of the layer half-space. A special ease, of course, is that in which the hallpace is Ihomogeneous. The approach follows that of Sunde [7], along with certain limitations that we shall now discuss Essentially, we start by rewriting (119) and (120) i the equivalent forms ental rips SOO roy and = 51 OO, a)160 Iv. tecromagnete Covping fr Goundes Sauroos where, in aecordance with (125) and (126), rye Be [hp sate ox) 200 {2 reir, ho «3p Intec of tomageeoes al-pce, we dade tat reg= BE [A saanen oy er al 3) and 20) ry [| 2B ronan ona [F 0S Ho} (134) Resuming to the general case, we sow conser the situation show ia Fig, 6, The source dipole fi shifted from the origin toa distances along, the eal axis. The component of the electri field at P(x) elearly za1a|-P)~ ah 0} 035 where A rele +98] Fp 6 Geomery fr out hee wo pale eis ah of ee eth, on he ste of he pret ape ‘ut trpecarce Fermultion 181 {In writing this form we have wilzed te fac hat aor _ 494 _ 2009) ar AF)" ae [Now we deduce the resuking volag dV etwesn the potenti electrodes Pry and Py that are lente at (sy) and (3,70. recive. Cleary, ‘thin the context of quasisati eld theory, d= = [E810 as (136) I we now consider the volige V between Py and P, fora uniform current Flowing from Cy(s,0) to Casy0, the tesultant mutual impedance Z between the two currents is fee 3 oo Jace = fe forerteie ote 9-22) OC 8) vers naa[Os Ha)" raalos nanltn—ar toi nucleate] “Te other obviows configpraion i tat in which the wo sounded cireits are at eight angles (in plan vow) (Fig. 7}. The potential electrodes fare now located on the y axis, and thos the coordinates of Py and Py are a \ 7, Geos or oun etch tiie sal re tly pepe182 Iv. etecvomapratc Counting for Grounded Sources @.y,) and (0,9), respectively. In avery straightforward manner, we find ‘that the mutual impedance between thes circuits is given by 2u2 [Pf adem Bir) + Ors) 2643)~ Oru) (19) ‘where now nan (ots rae (T+ 8) Be ay” ‘The functions Pe) an QU) tht appear in the mesial impedence formals gen by (38) aod (39) forthe pall aad the rghtanele rays repeat ar defined y (132) an (13. The obiows seer toa ow ol theo cet to hve ay geal bat nonin onourt. The station I tasted in Fig 8 Tae variable itance'F Between th element ds and is nde. Ao, the variable arte ¢ subtended bythe estoy ofthe elements cd and dS shown Now the diane messed from C, along the curve extending from Cy 10 Ce Corresponding, Sess from Py lon the curve extning rom B to, Then he tal peda ake rm Fa) 2a [fe[Pevemes Seas as = [S[fPineoewas HOLY + OCP) ~ QC Pa) (CP). (140) Fg 8 Geomotyfor coup ween oot of hires contour onde He dpa ail imedanee Formuaton 163 where CPs, ete i the linear distance between C, and Py, Bear in mind here thatthe angle «i «function of both the distances s and $ along the contours, OF couse, the fator cove vanishes whenever the elements de and a are a right angles. ‘Another interesting special case of (140) is that in which P, and coincide, which in effet means chat we have a cleed loop in the receiving ‘lrcut Tn tis ease the mutual impedance formula becomes 22 [Sneed cay where, a5 indicated, the contour ef integration over S isa closed circuit ‘The disappearance ofthe Q functions means thet mutual impedance dees rot depend on the vertical conductivity parameters dyy for any of the ‘eons. In this sense we may desrbe the coupling as purely inductive. 'A further special eases that in which the curren circuit is closed. Tiss the limit of (140) when C, and C, coincide. Thea we could write Zanifrieonctas Ty ‘where both (anticlockwise) contours ae closed ‘Cumbersome as they may appear, the foregoing mutual impedance Formulas can be vsed to predict the Fequeney dependence of the mul ipod for various four-electrode arrays over a multilayer earth model, ‘OF course, considerabie simplification ean be made in certain eases. Some ‘of these were made by Sunde [7] on the basis of his earlier staies with ‘ollborators at the Bell Telephone Laboratories, Other related work is ‘eserbed in review aril [8 ‘A special case ofthe general Formulation leading jo (140) i & wo-ayer ‘earth model in which the thickness of the upper layer ish This ease was ‘onsidered in some det! by Hohmane [21] and by Dey and Morrison (20) In particular. the later authors considered the collinear dipole-dipole array (Fig 9), Here, the euttent elects C; and C, and the poveataleeciraes 2, and Pare euch spaced by a distance L, Then the distance between C, and P, is NL, where W isa specified number. In the paper by Dey and “Mocrison [20] the layer conductivities were isoiropie and were designated nae % Z REELS % 'e-9. Calne dle aay woge condcing earth oe164 1. Blosronapnatc Causing for Grounded Sources one Fy Pec gues ts (PE) or = an = 6/1, hay 0, and og, respectively. A. percent froqueney fect (PFE) was then 0) can be obuained via a vector potential that has only # = component 4. Tavs aw @ and ° where yp egn)e. ‘Asa further consequence ofthe axil symmetry it follows that E20 and Hyy= Ha. 0. ‘Using omy alte hindsight, we are led to write for = > 0 Age = (1dh/aery[ em" /R + fle.2)], a where R = [p+ (= ~ A]? and fig,2) fo be determined. mn (4) the fst af the terms in square brackes has the correct singularity for reproducing the feespacs fields of the dipole, The function f(p.2) i= thus chosen to bea solution of (P= )flez) = 0, oy100 . ex eaneon with te proper behavior any. These consideration lead us to wre = 1B (A feg(-ude— a + RAVe| ade + HBA, © where p= (92+ 9)! and where yA) is an unknown cocticent. For the lower half-space the fields can also be derived from a vector potential that has only a = component 4,. Thus fr 2 <0 1 Fa, o ® and ° Asin, because of symneto, E20, and Now A, satis Ka, 8 4 Mh yao, 10) 9 wpe ae Ty a where and = “Thus suitable aera form for 4 is 1h (rye 4d (rayesssoydd ay where w= Q2HC 4 2) and where TO) is as yet unknown, ‘We now invoke the boundary conditions that require the continuity of the tangential fields al = = 0. Spectcally, Ey, = E, and Hy,= Hat = 0. {sing (8 (Oy, an ea ois tha ay— vie _ Ro (2 and = EN exp wh 13) T= . 3) ‘une Sate Lint an ae win Halspace 169 Cleary the function Ry() can be identfed as a reflection coeficint for «plane wave incident onto & homogeneous anisotropic hall-spac. In this fase the electric vector is parallel () to the plane of incidence and Ae ~ryysind, where 8 is the (complex) angle of incidence. The Fureton TT) cat also be identified as «transmission coefficient that includes the factor exp(— up) ta account forthe height of the dipole source above the interface. ‘This completes the formal exact solution of the problem as posed. The resulting integrals cannot be expressed in closed form ness some simplify {ing assumptions are made. For example, in applications to radio eommini- cations, itis olen permissible to invoke far-field approximations that are jostied when [ypl >I (ce. at Borzontal ranges that are large compared ‘with the freespice wavelength). Since this case bas been discussed clse- where [7 in considerable detail, we shall not dwell on this point here. Instead we shall examine the nearield, or quasistatic, region that felevant inthe geophysical context [8 [Bvercise. Consider that the dipole is located at height over an Naver anisotropic half-space. Show that the expression for y., given by (6), bas the right form provided that R,0) is as given by Eq (114) in Chapter TV. . ‘QUASI-STATIC LIMIT AND FIELDS WITHIN HALF-SPACE ‘When the quasistatic assumption i made, we cam say that Pe In ower for this ob vali he integrands of (6) and (1 Iajorconbutions inthe Fane in wich Pol> bh should have thee ‘This in tun implies that bro, ie, the horizontal range should be small compared with the free-space ‘wavelength, As a cautionary nole, we should point out that is proceeding to the quasistatic limit one should fst carty out the resivatve oper tions oa the vetor potential before sting 2 ‘AS an example, we shall alculale the magnetic field in the lower anisotropic conducting half-space forthe verial sesric dipole located at [eight Hin the upper insulting region. Using (9) and (11) and then settingv0 enn Condueton 1380, we obtnin the quasistati form y=" [n oe MAN for 20) wire one P, Ka 6/6. and P= yap Since neffest we have aged daplacerent cutens in thea, the result for H, piven in (14) is valid if both (~2) and hace small compared the fdespace wavelength, However, daplacement curents inthe ground (ies the lower half-space) are tll retained. When these currents are negligible, both 4, and &, could be replaced by their real couaterpars 0 and 9, respectively [xercise, Show thatthe correct quai state imi for => O is given by a —_* . oe a a es eT) al ieeom yy [eter ‘nse eheckon (IH) so examine th init == A=; whence = Lh (*, = Lah tn,= He Ceaaonan= Li 9 which is the expected form (ie, vce the broadside free-space ld Unfortunately, (14) isnot an elementary integral form; however, the case for = O (uty vertical electri diple just above the interface) does lead toa elosed form known as Foster's integral. The explicit fick! expressions 0, however, are very cumbersome combinations of moved Resse, functions Jy, fj, Ky and Ky with complex arguments. These are discussed tleewhere for both the isotropic and the anisotropic eases, bul & speci ‘as, worth noting ere. We'set = i= O and examine the E, component fn the interface: dh a(jy apd Sc agitias (Rob gr) tee soe of eed kom 1 ers) 5) on TC rol 1, IoKy= KY?/(ym)s (9) i | | | | Vera! Ground Une Source am then 0) Tor the ease 2 = # = 0, where 1 Cs)! ina isthe surface impedance. ‘Of course, we ean obtain this limiting form more simply but less rigorously if we stat with 5, 4a (Peayampernan ay and then argue that ey i [|i suficindy lage. "Then for <0 we have {Ge approximation ca Hal (aan “This has the same form as does (20) in the limiting case in which 2-0, ay VERTICAL GROUNDED LINE SOURCE ‘An interesting extension ofthe point dipole results is to examine the ease jn which the source isa semi-infinite eurent ine grounded atthe bottom fend (Fig. 2). The magnetic ek at the subsurface point (2) obtained by Integrating the dipole response with respect to I from 0 to v2. Then, of fssuming that the current 1 is uniform throughout the entre length, it fallows that for 2 <0 EfCroneran @ -e “pat Fe 2 Sentai veri ie uc ve acti soc apcwm ‘anh Cancion “To achive his ret wehave made se of mean [evant eH ‘To del wih @3), is elghening to change the scl of the verte coordina Ths for expe ot that, 7 park + vy F= RECE Y, @) where ye V/ KP = Gg g)'* is by definition, the effective propagation constant in the vertical direction ‘Then (23) can be writen as HE LOrsaoye oan es) for ~1.> 0, here weQeeyy? and sen ek% Here, the wale ert coordinate As inte in Append B the Incgon ovr ht 6) ca be ated bt gv ke eloeorm ea A= [aoe + Fem), en where ao We shold bear in mind ere that 2, 4 osive quantity for te ‘bsuce obser Hee the disanc rom the gin ete Baton hd ofthe na sour) othe bene inthe sel stem. Alo work sang a Wee 2) 1a the limiting cas in which or zens 10 er, we have simply He= 11020). ch whic sient to dhe magnetic ie that i circumferential to an infinite Ivte exeying& uniform eueene 7 Thus in this iit the semi-infinite line source is being perfectly imaged by the conducting half-space "The elects Held components ofthe semi-infinite vertical current source can most easily be abiained from #, by using Maxwel's equations. That i, OT Vera! Gounded Line Source 173 fore <0, 0) and ey ‘Then, on setting 7p = 0, we obtain the appropriate quasi-static expressions inthe lower halespace. oe and oy abo = Eee RM ME = SEI tke oe) “The limiting values of these expresions for 2 z_->0 should be noted. Thus then oa fliegen} es ‘Then fo] > 1, we so that ay 66) Where H, is given by (29). Alo, the expression 4,E., in accordance with (G4), is zero as 2,0, except at the origin, To check that we have the Proper singularity there, we perform ihe flowing surface integration ofthe normal curent density fof aes. conn oyws ¥. Seth Condueton Provided that a> 0, the integrated quantity tends to J amperes 2s ~2, tends to zero, Here o can he interpreted as a disk-shaped are of radius ¢ ‘vith the euzent source at the cone "The vera ine source shat terminates atthe at-ground interface might ‘be a model of a lightning stroke in some idealized sense. Thus it is of interest to examine the transient response of the subsurface Fields. In Laplace tanaform notation we replce a by « and rewste (27) inthe form e nga t o Here Hays [Cemnnare 1109) 40) Js the Laplace transform ofthe time-dependent current 2() that is aken to be zero for <0. Then yey= [CEM a= LNG 4 where H,(2) isthe coresponding time-dependent magnetic Field at (p, ‘Our objetve then isto evaluate the iwerse transform naga tingaya gh [etme 2) where «> 0 Tn the present context yo[Bteren]® and k= A81/448), which, in general could he quite complicated Functions of s, However. strate he slestronagetic aspects of the problems, we ascme hat (2) and 4) ae replaced by fied constants, and 6 that ae independent of Then, ofcourse, re (onsy!® and Km ayloy ‘We shall io simplify the problem and say that 1) = tutes @ ‘where ut) (= | for > 0, = 0 for <0) isthe unit sep funtion, That is the curent beng trac on oy amperes at = O and remains constant forever fe. ‘Now forthe stepfupction exciton 10) = L(t) = belt Vertes Grounded Line Source 15 and thus for <0 fo 7 3, 14> yal [esol nne=] + evofcon]). (8) We now we the standard res Feo) = esl /20" 0) 5) where te 6) ez)= 3 fe ‘complement of argument Z. The time-dependent field {ito}? R 2822 Juo isthe error funet response is then cone: soho [a] 80 “ where R= ig! + Ke). As indicated by (4, the step-funeton response ‘0 begins at 0, and as lime # approaches ifnity we tend tothe de limit dined by Heo)= 38 wer «) “The normalized steptunction respons, defined by the efthand side of (4 is plotted in Fig 3 a8 a fanetion ofthe dimensiones tine parameter iH). The two eases (a) p/2= 1 and (9/2 3 ere shown in the figure. Tree values of K Ge 1,2 and’) are shown. A indiated, the magnet field response buds upto it de vale in » monotonic manner. ‘These curves exhibit the easy proved propery thal the dependence on the anc ra Kino eevee tn he Soph ae ji ‘athe simple extension th previous analysis to ake the exciting current tobe impulsive in ature. Then, or exarple, we would specify that 10) > 10600) 0) Where fy is the coll charge in ampers-seconds, oF coulombs, Here 3(2) ‘denotes the unit impulse function, which has dimensions of seeands. Now 1) = L100) fy 50) and the corresponding transform fr H,(2 is entice to (44) if f/s i nowce oy? Fig supine ld pot or eet ver tthe snoopy ark replaced by /y, The required transform pair is then —K eng u(t) tataeen(- Sem eH 2p Porm (62) 300)" ‘The normalized impute response defined by 2epaff0/lo, where te ‘= 9,049" 8 plowed in Fig. 4 for p/2-—1 and for'K'= I, 2 and 3. AS exo fa!) Hay I 2mplotigitvie S2ees5s5 ry roy?) Fg Ipuefildrapae fr semis ine os sara vs te soapy ose Ks er ft=s} = indicated, the response rises to a maximum before reseding to ze 3 the time parameter (yy increases. Exercise. Verity by diferetiaton that (0) he impulse response given by (2h is equal to (ly/I) da Hr the ep tancan respons en by coe (CURRENT SOURCE IN BOREHOLE ‘We tm our atention to the ease in which the source dipole i located within the anisotropic half-space, The actual current x length moment of the dipole is ais ve dipole is leated a an actual depth as indiated in Fes. ‘The expeced form of the fel singularity very near 8 dipole located within an anisowopie medium follows fom the sealing la fr the vertical Coordinate. This suggests that in place of (4) we aaW write forthe lower Fig 5. Vera sete dple rena apne,178 V. Set Conduction half-space (ie, 2< 0) «s) i a function of p and = that has not yet been determined however it remain finite atthe soure location “To confirm that ($3) has the correct sta behavior near the source, let IyeRI0. Trea 4) 09) fora eurent element of length di. This is equivalent to cy = — Hdd 565) Pe Fte507 Bh “ ‘The de potential near curent point soure at egy IB = [ova 1) v= [ran 2 “ where B= (o0,)" isthe mean resistivity and where Aen[o'+ t= hoa)” xercks, Consider the de problem in which the cutent electrode injeets ‘srrent into the anisotropic conductor al? = ~ hy. Show thatthe vertical feurrent density of the primary fed is (e+ hg"? Tek “Then show that the verzal current nthe region O-< p< agen by pare firdtden KML e+ Um oy 1 Ol ena” Note that i =(5)~ hy hen in he it c+ e-r0 wehve Iosif? forall a>0. © ‘Cuant Source in Borehole 179 We should ow retonably comfortable tna gen by) ae th igh ind sng Bhar neat ne sauce Now xe te ‘ice shoud ly (1) for Sc eet sein ih eon exept as = yd pw coy tte feo 4.) tha spp n() a aes (9) esther for: 20 ting th Sommer egal peter tea aces inGytwe enw ed town ore Ape EB PERL evertha RyayerM]oydd, (58) ‘where A is yet to be determined. For : > 0, we have simply Auge AK [> aye™yanoyad, 9) waete #0) is not yet determined. On matching tangential Fells at z =O, wwe dee that fey (Vent + Raye" (0 and 6) ‘A quantity of special interest here isthe tangential electric field at the sutface ofthe conducting half-space, Thus for = 0 1 ate were we note that SE 2 [OMI Ailednoyan, 62) n+ (defray @) Now inthe quasistatic approximation, ‘This is justified when [yo]. In this case (62) ean be evaluated to give the closed-form re 4) where Ye= cu) and Re (ph KY?Fe. & Ozona for auting te ie pound dei Done und suas “The form contrived for (66) is vory convenient fer deatng with coupling between finite circuits {IO} As an example corsider che anrangement shown in Fig. 6 Imagine tha the souree is uniform current that extends from = —/h, (0 ~ hy 08 the = ais Physically, this corresponds tothe ease in which an insulated cable is grounded at the endpoints C, and Cs, which mney he identified as the current electrodes, Th diference of potential © between the two surface electrodes P, and P, eas then be measured by an insulated cable connected toa voltmeter, These potential electrodes Py and Pe are located al p= p and ps espectvely, for =O. "Tne desired expression for the transfer impedance is clearly afm 0 oo0k acre “ ‘We designate v/7 by 2G) to exhibit the frequeney dependence. Then, of exot= rll) Ra eared) 6 vere Ryy= (OF + KEY, [AC sulficienty low teguenis, Zia) bacon the tansfer resistance Ray which reduces to Ran 20) ‘The closed-form representation for Zi) is remarkably simple. The only approximation involved isthe neglect of displacement currents in the upper halspace (ie. the ai). A further convenient definition is the apparent (Complex) conductivity o,(i) that it eelaied to the measurable Z(io) via 1 a [ge 20 (6) wa Clearly ou) a8 deine isthe apace condctviy of the sauvet omogencowe tape bsiespace. Tos inthe limit of eo requeney we sett 4,(0) = (aye)! (e (oie my 8 TR a ‘And, of cours, in the limit K= 1,40) = #4 = ey ‘TRANSIENT COUPLING BETWEEN GROUNDED FINITE CIRCUITS. ‘There is some interest now in examining the cas in which the exciing current isa step funtion for the ray depicted in Fig. 6. For example if T= feud), en iis 2 simple matter to deduce thatthe tine domain response o() inthe potential eitcut is given by ete 7) oeWw Y. ents Conducon 20] os ip. Test wpm fo alge in onl alt 7s fr stp an canon asm Cy whee mele se hy)? three Toe were 4 Tne odo * Rin) “Th rel funtion o() can easly be computed hy using lbulted values of the ero futon complement. Alenatsy, 2 computer program can te writen Ths was Joo by Kini abu for the casein whith y= 0 prada hyn and ha de and for value of K= 2 and 5. The ‘han actaly plod in Pp? he ala 07 fa, here am Rg KR m™ 1 Ba, isa refernoe ohmic revitance, These parameter for the curves in Fig.7 's 1/9, 140, which of courte i dimensionles. [tis clear thatthe curves increase monotonically and tend to the asympiote K'* A dstneteharac- tert of the buildup is tha for lager values of K (Ley > 0, the inital tise is delayed somewhat Exerege. Derive (70) from (6) making use of (45) and assuming that a ‘=, are frequency independent, Fen | Bre ‘Pat communication, Unde af Arians, Tween Peterencor 183 REFERENCES: “ip Seu ee Se 821359 Mmenenmn oneness 2, YR. Par an 1B Wat A pad ol nthe eh, Tp ly Ps W 238-270 78 5. WP. King and ©. Sith dom in Mater, MIT Pox Combe Masache 4. EY. Rg On he ow fle eet ie sei {amin Ser, 3.257271 8, RUN Edwids A powned vera agate nem for ple wie magnetic al mong, Ger 193-159 (8 65 Jn Wat Camron on A grosndd ete! gic nat fr pine wave MT lp ong” Gp 34935 (18 1. R'WattBecromauncte i of 3 ale oe an aint alps, Cm Phys 44 387-290 8 TR Mi hi erste Pg et son Ps Bole, Car, 9. OCaml and RM. Fe, Farr ae fr Pal yan. Van Non New Yat {0 DIA Hil and R. Nok Substroe sticFlé fx unde able tie eth for both eqns nd time domain Pep Gin I 34-233 (3),cramer WI Magnetotelluric Theory} INTRODUCTION Conventionally, in electromagnetic probing methods, the source fells ae provided by atficial means. However in principle there is no reason ‘why natural ils cannot be exploited, For example, the geomagnetic field of the earth fluctuates in @ rather complicated manner, due primarily 10 solar influence on the earths jnized environment. Also, lighting die ‘charges act as anteanas thal produce electromagnetic feds that pervade the aumosphere at distances up 10 the antipode. Although there is no clearcut distinction between these source mechanism, it fs now believed that the geomagnetic variations that have periods above one or tvo seconds fre not lightning induced bul that shorter period variations, partewary in the audio portion of the radio spectrum, are produced primarily by the lightning discharges occurring throughout the word. ‘A areat deal has been wien about aatural electromagnetic Feld varia- tions, whether they be casified as noise or as an available source for cophysical prospecting. The author [I] has prepared a very comprehensive review of the subject with an extensive bibliography that covers papers published up to about 1962. A more recent feview, which also includes ‘many references, was provided by Filloux 2) “The basic theory has many aspects, bul there are two topics that we shell 0) 1 be tropic with conductivity e,. permiviy g, and magnetic permeability, The atmosphere is aken to have the constants eg And go re space. "The incident wave is polarized such that the magncti field vector is posal to the interface. Thus we choose our Cavesian coordinate system it Such a manner thatthe magate Feld vector as only a» component H"" ‘This is writen inthe fort HP = Hgexp(— ye08te)exp(— Ysin 0), oy whore @ is the angle of incidence and Yo ilyt)"78= iu/e, The ree ected field wi have the form Hp = esp yeoste)49(—Tosn ds, ° wher, by daisiton, Ri the reflection cote On combining (1) and (2), we can wrt forthe resultant magnetic field 1H, forthe region = <0 Hm tferme Reme ems, ° where A= —iyysind = (u/e)sin® i8 the horizontal ware munber. Also, because H, satisfies the Helmholt equation (= YH, 6 ners, in the present ease, c 8 we oe that 1g 1. Oba fein af pln ple ware rom homoge aap1 ML. Magntotetre Theary “The corresponding expressions for the electric Feld in the air (Le, = <0) are oblained from an, eget = o and au, ing, = He o tn particular, B= Hoke" ~ Romeo, ® whore Kom t/a) = ees, whore 0 (1/9 1205 ohm, ‘Now iis easy show that in the conducting hallspace (ie => 0} Hem acre, o ‘whore Aiea constant and QI! yy =Limwta + hoy) “The elec eld components ae absind from an, o,+ ieuy8, =~ (9) and an, (ou eyes Fe » In parca, for => 0 we see that Eye Kyler, (2) where Kym us Hloy fe). ‘The houndary conditions requite that E, and H, be continuous dough 8. Then, on using 3), (8 (9), and (12), 1 flows easy that Be (Ky> KM(Kot Ki) co) and A= HK (Ky + Ky (4s ane Pano have Mads 187 ‘Thus for the specified incident wave we hive now determined the resultant file components everywinre ‘OF special interest in the present contest ithe surface impedance Z,() interface, defined by 2Q)= EM oy ayy For the homogensous half-space model that is considered here, ZO) ~ Ky = (+H) Hast Heh, 6) which ilustats the explicit dependence on A the horizontal wave number. Now itis evident hati h*/yile Z=1 an whore am rida + ices) = [me/(0y + ies]! In this limit, 2, is independent of the parameter A. Clearly, this isa desirable condition if measurements ofthe tangential fields £, and 7 are to be used to deduce the earth constants 6, «3 To shed further ight onthe subject, ote that ay at 2eyan(ie des By.) 18 sayan(i4 A+ Bee) 8) which, sti, sont valued if AY/9 <1 In an operations sense we can idetty X wih the dervave 18/8, Le, B/eeam— Ax) ~ esp {ks}. Ths pes us wo wre the impedance tsa deane operator zajon(i- 2, #43, a } 19) OA Sat ant * Bef Be © Or, be more exp (20)= (1 & « etedmaltash Bed Boo). 0) This is probably the most rudimentary argument for indizatig that the surface impedances souree independent i the lateral vations of the fies are sutictenty slow. Surprisingly, his hasbeen the cena! pot of such conroveny eer since Cait (3) published hit much-cted paper jn 1953, OF course any general statement about the source effect requires Consideration of tang eter aspects but if noteworthy that Caghiard invaly formulated the problem for» noemally incident plane wave (he, forthe source atthe zenith{In formalatng the above problem, the magaeti vector H was chosen to tbe perpendicular (othe pane a neigencs, oF, in otter words, H i parallel to the ground and remtins so even after vflection and transmission into the earth, The ahr ease, of sf less interes, i that in which the electie vector is abways parallel tothe ground. For example, our incident plane wave mould have the assumed form BS = Eyexn(—Yaconle)exp(~ ysis) en “This case is completely analogous to the problem that has already been worked out In this ease, however, the quantity of inlerest isthe surface fdmittance YA), whichis defined by 100) = elec a To be exci, 740), = 00+ 1)" /ie @ “then we ind that nay — 2 (1 Jaro 9 “The consusions rin the ame: the rai of ke tangent lds ae source independent only i the lateral variation is sfiiently slow. (CURRENT SHEET SOURCE Some exensions to more gener! mods hate ben made [8 18-20) Jn particular we should meson the sas in whieh the source is not @ {Sencoming pane wave butt siren sheet a once eight. Forth tnore, js make the problems more reaiae and more interesting, the ‘pall variaons could be in bo and y (8) “he tation biisted in Fi 2 which asin a Ceres cooeic nate sytem is chosen wih the ssa infest == 0. essed thatthe earth is homogeneds as eeczaleonsunts 0, cand py aNd cups eh space +> 0 The gion © Dsmwme t be reespaee Sovinapermitivn yt peesbly ny, Now me ave 9 nur eon at the ves uy ida eh pAb the sas). hi eon we Shame tha a special componenof the cunt inthe 3 dieston 1 Jreenteftseepee 9) smpere pre meter Statin. the special como- fc in they ction i exp ifsben(—y) A sua, the common ts tr bag, wr te gue guy, aed tnd are the spl wave umber forthe spect smmponen ats bed considered. ms 7 nd a Fig 2. scat of hamapens hallapas hy oe cuen sst sem “The objetve is to deduce the relationships between the tangential levine and megnetie field components atthe interface z = 0, These results lire then used (o draw inferences about the source dependence inthis more fener formulation of the magnettelurie problem. Tn the present problem iis convenient 1 represen the fick components fn terms of two sealr Guantities, Those are the ¢ components of electric fand magnetic Hertz veelors, and they are denoted by 1 and TI", respec: tively. Thus, fr the homogenous earth region £ > 0, we may Write Ba(-it °5) and a(S) es vite I= 0404+ H) pt ih tthe soe and Ts ear Helo euatn (=f) 0 ey wie vee isthe Laplacian operator, From Marvell's equations we may deduce that 2, ipl pa en) Hw =i BE — (0, Heh °) =~ + woo, 0)190 WL apni Tey and =a 4 180, + few) y= 0A ipo, + tt ey “The patil Fields associated withthe TT poteaial are often designated TM (wansverse magni, and those associated with the TIP potent are designated TE (transvers electri) Th the air region 2 <0 the field components can be derived from the potentials Tf and TI in the same fashion, but we must replace + few by Fess and By ti, everywhere, Sometimes in dealing with eletromagnetic Induction problems in the earth, geoptyiiss discard the TM field a the futset and make assertions that such fields are not needed) even for Stbitragy source configurations, ‘The inadequacy of this asiumplion is ‘obvious when one deals with a vertical current source in the atmosphere (ee, a return lighting stoke), ‘We now return to our general problem, For the ar region 0 > swe write (on adding a subsexpt zo) ” Ty = aexp(aye) + Been(~ m2) 2) and Tip = ot exp(uns) + BF expC —tae). (0) where m= Oe BPs ay and , by a, and 6 are cosfcens yet o be determined. a (82) and (3) and in sulequent fied expresins we omit the common facor exp (Pipe = fh tat). We really contin that caf) =0 eH For the sri :< = We ave the coresponding ors T= eexptes) es) and 1 ‘where © and e* are coefficients. ‘We now iavoke Ampere's lw to tll us that Hg —24~ 0) ~ Hoyt = 20 0) oe Hagl—24~ 0) ~ Hol £0 +9) = = 8) Furthermore, because we ave postulated only an eleire curent source, * expat) 36) Curent Sas Sourca 191 ‘he tnngeatial electric fields are continuous al == ~2o. That is Fa(—F0+0)~ Eq (29 0)=0 0 and Eq 59 +0)~ Ba. ~29~0)~0. 0) Subsequent application of the source condition lends to the following ‘olations for the coefiients = Rb, ay oan, 1) em [R= expen]. reo and eR expan] uy where R and RY are ratios of the coctcents that have nol yet been specified. To dei wih he Field within the earth Gs, => 0), we write T= dexp(—1) 5) and T= areap(—iey, 1) where again the common factor exp(—ifx ~ ky + at) is understood and and d® ae coefficients that areas yet undetermined, To solve for d and and at the same time to deduce # and R*, the remaining step is to ‘enforce continuity of the Four agen fed components atthe boundary 0. Ta partiular, we find that Re (Ky Ky /Ks* Ky) « and Em (No MMe Nip 8) where Km tl ego), Ky = Hf + fee) and Nom wince, My= H/C) ‘Thus for he region 3) <# <0 we have exoicily oe egal + BP) s ene 4 Rentenal) «)192 _ Magntotetre Theo and 10.= BY a+ By Tis pethaps useful to note thatthe factors fj, +2, and Gy, — Bf) are equa spectively, to div and cur where js the strlace current density tz = —2, wi components j andj, Explicit expression forthe ‘angel electric feds and tangent mag- nai feds (bt, Eye ay, and Hg) i the teplon — zy <= ¢ Oare obained by operating om Ti, wah Tg in acordsnce with 28)-G1). Ia the mit 20, we find that quit generally a Rhenten2 m= (6) Bac Zula + Zyl sn) and Bay = oy Hay + Zyy Hoye 2) wee ot 3) Trier ee -- ett 4 2 it men a and aw eee eee 59 Bam —2y ~~ ee Heap, ca whee y= QP a+ 66), We may sow ast hat yt Ze (7) where n= h/t, provided tha 8/7? and [P< To within this fwer|2,,/n| and [2m] a8 of he onder of [N/a thos ey ay te negotd n this ae "AL this sige it might be comment that (S1) and (2) ae often povtulted a¢ being general imposanee description of magnetic feu, bot ts asulyovetooke thatthe cements 2. ZZ, 88 Z, ave spatially dispersive. ie, they depend on ff and/or on Ain oneal ‘Curt Sheet Source 193 Following the same Tine of reasoning as that in the previous one- = 0: (e228 emit 60) Pag Flag ae Hay = 2S [+ term in yr ete. co) “These particular results are in agreement with a sinilar analysis cortied out bby Monteath [21} They show in avery ele manner that the leading terms Ge, y= )Hy, and Bey =—"\Hye) are valid only if the horizontal ‘varisions ofthe tangential els are sulfiienly slow. However it is noted tht the fst correction terms involve second derivatives only. This linear ‘variations of the field do not have & major elect on wpreting the simple proportionality between the tangential fields tha sometimes goes under the ame Leontvich boundary condition. ‘The problem as formulated and solved can be generalized without siffcutyto-an W-layerhalspace model ofthe earth [8]. The geometry is the sume as that in Fig 2, but now we mst deal with N interfaces between the various homogencous regions. The solutions for the Hertz potentials i the air region (—29 < 2-<0) have the same form as jn (49) and (30), bu row the reflection coefficients Rand R® are defined by R= (Ky Z)/(Kot Z) (2)198 Vi. Magnttatute Theory wd (Ma Y/(No* IY) (@) were nd ang /az gat | oy Hore Z, and ¥, ace the input impedance and admittance, respectively, for the equivalent transmission line cicuis of W sections in tandem, Explicit formulas for these impedances ae found in Chapter IV (se also [18). In the above discussion we have dealt entirely with specal component ‘ofthe Feds hat varies as exp(— Bx ~ Ay + Ja. Actually, there i no loss fof generality: for example, by Fourier synthesis the actual x-directed Time-dependent source current could be writen as nessnoy= abe fffiaeaen ee) A similiar fr ats th ata eee ll amponet Es Fee ee Eph wih whch oe chase ae. Aca, i the oateton ie enti he wonton ale fi spel dobetne "8/8 LINE SOURCE MODEL RESULTS. ‘To show explicily the influence ofthe soure location on magaetotellurie fields, we choose @ concrete two-dimeasionsl mode [19] (Fig. 3). two- layer earth model is considered in which the upper layer of thickness das a conductivity a, and the lower layer of infinite depts has 2 conductivity tr, The source fs now taken fo be line source Jy of constant current (ie, aja: ‘All displacement euerents are neplected and the effective magnetic per rmexbiliy it assumed to be jy everywhere. Ia the context of our general Formulation (rested eatin this chapter) for @ general curent sheet source al z=, we are in effect saying that LG = BEY (y Une Source mod! Rests 195 Fi 3, Une source ection of 2 tote a «ma fr examining the so ait asthe depndees se he ne! ace pda and Hey) = 0, (66) Where 5) isthe unit impulse, or Dirac function, ty =0, Here fyi the acival fllamental current, given in amperes. This we have a two @ corresponds othe stuation in which the acmal surface impedance is the same as the impedance exhibited at the ects surface for a plane wave normally incident From the Zenith, The question a nse is how much A differs from Q when $ isa Finite (i.e, noninfnite)parameer. "To shed some lish onthe problem, we use numerical integration dats for or the lmiting eae in which the source height x tends 1 210 (or, if not ero we are effectively saying thal zy, the horizontal range). This would be the geometry that is relevant lo controlled source magnetotlluries ‘Segmented Overurden Mex! 197 ‘atone eaeeto ae ° Fg Ample and pao suse impedes parent and @ 2 faction of ans oper era a two-sedom apace mol orn sours chao (CSM), A sample of some of tose results (19,20 8 shown in Fig. 4 in Which the amplitude and the phase of both’ A and Q ate plotted as Factions ofthe thickness parameter D. For these curves, the range parame- ler Sis Fixed at 10 and the conductivity eaio K i either 10-2 or 10 ‘An important observation concerning the cess in Fig. is thatthe and curves are indistnguishabe for the case in which K'= 03/0, = 100. ‘On the other hand, there are significant differences in the cate in which K= ox/0,= 001, even though the parameter $is quite large i, $= 10). ‘The conclusion that reinforces our eatier contention is thatthe source Aependence of the surface impedance for megnetteluie Fields is very marked when the subsirtum is relatively resistive, Tn such eases the rameter must be exceptionally large (eg, S > 10? or 10) if the actual surface impedance i to be insensitive to the source locaton. ‘SEGMENTED OVERBURDEN MODEL ‘As mentioned in the introductory section of this ehpler, there are sa aspects of magnetotelurie theory that are central to oor discussion. The source effects have already been considered, The other topic has to do with18 ante hay Tea] %] [| [or playa [el ls Paolo Md gS. Some ove mode fr examining the fun of vrs cone om sie rakd sce peo the nonuniform stricture of the overburden or basement layers. To deal with this subject in a general manner is ouside the scope of our tsk, but ‘ve shall adopt some selatively simple models that Hustrae the effects of Iateral varaion of the rasta layers “The most bese example of « nonuniform structure is the vera eontaet tat separates two homogeneous quarter spaces. This situation Was ana- Iyeed by C'Ereevle and. Kanete (Il) and by Weaver (10) under the sssomption that the surface magnetic field is everywhere constant and Sirected along the strike, Amore interesting aad only slightly more compli- fated ca If what we have called a segmented overburden (22). The Sivation f shown in Fig. 5. The overburden consists of any aumber of homogeneous segmetss, all of constant depth J but with variable wid Gedgye vray sor edhyey. The conductivity and permeability of the mth Section a og nd py respectively, where m ranges from Oto Mf, “To analyze the problem, we need to make some sather exleme sssump- tions, Spesifelly, me aarume that the magnetic Feld has only ay compo- rent that is pail to the stike of the structure. This means thatthe Gece field vector has only x and y components. Of particular interest is the discontinuity of, atthe verizal contacts. We note here that forthe ‘other polarization, where the eletic Feld is parallel to the stike, che Imagneti eld vector would have = and y components. Analogously, the #7, Component is discontinuous al the vertical contacts by virlue of the Contrasting magnetic permeubliies at each segment, This case is not onsidared hore Returning to the station in which the magnetic field H has only a y component H,(s,2) ve further simplify the problem by assuming that 1,30) seat’ at constant Hy. Also, H, (x) i assumed to be zero. ‘The later condition correxpords (0 having’ an insulator in the botiom region with no soures. For the mth section, we then we H,(%2) = Hl ) Pal (9 —— Segnanies Ovburdon Mace 109 such that (0) = Hy, Pal, 0)=9, Hyhym 0 Pyl. i Furthermore, we choose H, {o be a solution of the Helo equation (3-a)n en foes 0. The coefficients Any and By ae to be determined from the boundary require the continuity of H,(x,2) and E,(x,2) atthe vertical contacts. In addition, note that for the seivafinie end segments (Le, m=Oand m= A) Aes™ Bua? 7) Otherwise the fee would not he bounded as |x|» eo ‘At the verdeal contact x xy, the Boundary conditions take the form Pract + Hayes = H+ Po (8)and uct BPact By BP ae « ae © Now we ely dete tat Hl 2) = HoSinb Yuh 2)/sinb lh (59) ‘Thus forthe x ~ 9 contact it proves convenient writ Gu asin, co) Hy, where abs (ame) )[ + (om 7) ] 1 effect, (88) and (89) lead to a system of 2(A/~ 1 linear equations in order 10 solve for dye and Byy in term of Cy. The procedure it liscused in more det in [22)'snd explicit result for the ease of three Segments (ie, A= 3) are given there “The quantity of special iterest in the presen: context i the surface impedance at the surface z= 0. Within the mis segment it would be ‘blained from oy 22) = ECO) H(80} @) org) $< (= O12, 25 Me In explict Form, this is written 36 Zl3) Cite YS 4) where 8, colt (ogtat)"7h) Sof aaccr| (0? tata F] ah 6 Here Sy cin be desde the normale ere impedance of the Strate fn view ofthe fat tha g/g the Enns impedance ofthe ‘mth medium. ae ms (Cee + Bg to[ (+ att) | | wn Fig.6. Normale nace impen for #-p ielvea tesepet ol of ae sb The saa he oy he Fate shows te gal ol ck ssn. 1 we had ignored the effects ofthe vertical contaes atthe outset, the deduced value ofS, would be mecely the fist erm on the right-hand side fof 5). The summation term thus account Tor the inuence of the lateral change of properties across the contac. "An example of some numerical results that are available (22} for this probiem is shown in Fig. 6. The amplite and phase of the normalized Impedance S(x) are ploted as a distance x from a vertical contact. AS indicated by the inset, three-sepment model is assumed (ie M=3). For this example we have chosen aj/a, 0. $0 thatthe end portions are felaively good conductors. Also, we CAKE py = Hy everywhere, meaning that there sno magnetic contrast. For thi example, (0,19)""h= 1 and =f. In Fig, 6 we als show the constant values of |S] and the phase of $ that ‘would apply if the effect of the vertical contacts were ignored at the ous ‘Geary, the neglect of the vertial contacts in computing the local surface Impedance is not justified, OF coorse, so te liming eases im which the vetial eontct is more than several skin depths away from the point of ‘Observation, the surtace impedance assumes is lea value as deduced for 3 uniformly layered structure.VARIABLE SUBSURFACE INTERFACE ‘The vertical contact ea rather extreme example af lateral nomiformily ‘There are many svation in which lst drat description fs desired. An example would be a layered earth model in whisk one or more of the Interfaces is nonparalel. A general discussion of tis problem is given by ‘Mann (13, who advocates a perturbation scheme that is viable when the structure doesnot depart too much from uniformity. A similiar philosophy ‘was adopted by Katfman and Taborovsky (IS) We ilstate a simiae approach that uses relatively simple perturbation scheme, "The model that we adopt i llusieatad in Fig 7. The earth is represented by a two-layer half-space, bat the interface between the upper and lower regions is represented by afuretion of the form ay} sine = f(s), (96) ‘where we have chosen Cartesian coordinates (x2) with the earth's surface at 2=0. The structure is taken to be invariant in the y diretion (es 3/8) = 0} The upper and lower regions are tken to have conduetivi- lies, and 0, respectively. All isplacement currents are neglected, ad the ‘permeability everywhere i assumed to have the freexpace vale ay. ‘We treat here the H-polarzed ease wherein the magnetic field vector H has only an H, component, The corresponding results for he E-possized ‘ise willbe mentioned in connection with the nirerie! exemples “The indcing Field i tken to have the form expt ~ i), where 20/2 ie the spatial wavelength, As usual, the time faelor adopted, But not shown expicy, i expat. In the conducting ayers. stisies ~r hea ” where foc <2 < f(a), for =>) co) Fg.7. Twoler rth model witha wary ste intrae ‘We now write our solution inthe form Hsi2)= DM syenrrn, %) where in general, includes O and all postive and negative integers. This {orm of the assumed solution is pesioic in x with the exception that there is a uniform translation according to exp(~ hx). ‘On substituting (9) into (97), we realy fin that 2)~00H 42). (100) where r= tga, (m+ for O< 2 < fx) (IDI) and roa eBps —ingaos— (mw NP for z> fla (102) ere, we define Im, <0, Thus in the upper region we write HE) = g(a) a HPCs?) 19) and, inthe lower region, Hala) = 0, 3030(— Bn where b,. is assumed to be O 1 permit outgoing waves only as “Expresions fr the elecsc field components i eliber af the condu regions are obained from Maxwel’s equations as £,~- (195) snd (108) ‘The boundary conditions at the subsurface interface == ftx) are that Components of H, und the tengeatal composes of E be continuous The Inter condition i equivalent tothe statement that coset E.sing 10% be continuous, where ang aiisy/ae (8) ithe slope or inclination ofthe interface "The implementation of the boundary conditions and the messy lgebra are deserted by Hughes and Wait (23, 2). Expressions ae given thre for pe Bay, ayy Tt aso Shown there thas ye € and wy208 Vi Manette They where 2 isthe amplitude of the sinwtoidal variation of the depth to the interac, only sx coefficients need be eorsidered. The solution, as given in (23,24), was facilitated by osing the addition theory for Bessel Tunetions te F eomtnszy (109) “his canbe simplified whe the argument of the Bese functions i smal Gare was taken ta se that neglected higher-order terms didnot materially smosiy the final result ‘A final step in the computation proces is to deduce the actual value of the surface impedance Z(x) a he earth's suaee. I is defined by Zs) = 2.09/20). (119) ‘which isa function of both X and v, but for specifi discussion in the ‘agnetoteluri context, we ean set A= Oi the inducing eld has a spatial ‘aration that i slow compared with the period of the undulation of the subsurface interface [25] ‘Actually, o ply the results toa ease of geophysical interest, we need to allow for 8 profile of nonsinusoidal shape. Thus, for example, Hughes and ‘Wait [23] ease a function ofthe form feoy= sot S sini. any ‘Then, according to 24}, we need only consider the fist AF terms in the fed expansion ifthe following conditions holg ry & 1. a} ©, LS mS, and (1, A> aij. The fist expansion allows us to truncate a series expansion for Bessel functions. The second ensures thatthe feds induced {80 not propagate upward into the air The ater condition is synonymous ‘with neglecting displacement currents in the ar, which we did at he outset “The prevwaly described method for the H-polaized ease (Le. H paral le to strike) cares over diectly tothe E-polaized ease (i, parallel to the sre) OF course, although the formalism is unehunged, questions of ‘convergence are not the same, bul in general either polarization model ‘works if the amplitude ofthe subsurface structre is small. ‘Some numerical data from Hughes and Wait [23] are showa in Fig § Hire the surface impedance evaluated at 2 =0 is shown plotted as a Function ofthe distance in kilometers fom the center fine of a basin with sloping sides. Three hundred terms in the Fourie eves were used to model ‘region 10 ko wide. The inducing field was assumed to be uniform (ie, =), Sign periods 22/a of 1, 1, and 100 seconds were chosen 0 tencompast the variation of interest inthe magnetotellric range. As indi- cated in the figure, both E polarization and AY polarization were coasid> VaristeSupeurtace norace 205 EPobraten etait snece MeDANCE Tad | ie qn zo 6 i, red 2 ‘in & af com 1aholn x a 1p 8_Actl or ead saree impedance for sino dawnt dps sion a ihe site nertos Ito) Tet nw te ec a the ‘med see ot End pases are wn for we pe 1s and Ibs. no ics symm seat =O pol sro. Sne al crete, of ons, ym abut x #0 nly one is shown in each case. 2 iam ‘he cow shown in ig eit fr snr othe bine Rey Ran [7] 1 4 ann hy ieTHe so ed psy mimersl ac apt by Sei i bed on newer sig Ar our sad fh vei cnta he plato i cof talon hows a more mae ests the she pases over the se Ofte sacra Tus in socad wth then tte coe ow 8 Seg or by i en es d's peer206 WL Magntteutie Theory CONCLUDING REMARKS ‘Actually, we have touched only lightly onthe vat storehouse of analyte cal and numerial data available for magnetoteluric sounding The theore- jes! methods, using integral equation approaches, have potent application to more realisti structures. In this connection, the works of Hoan (26) and Howard [27] are most noteworlhy ever though they Jo not apply ‘Specifically w magnetotllurics, We might also call attention to the sugges: tion that rmagnettelluries may have a counterpart in magneteslencs 28] Ge, on the moon). "Noteworthy textbooks on the magnetoteurc method inelude those of Porstendoret [29] in a slim, highly condensed, but very atractve format with references up to about 1972, and a recent massive offering by ‘Kaufman and Keller [3] that heavily emphasizes Soviet eonttbutions. The Tater includes an extensive catalog of type eurves for various structures, snd il lio has numerous derivations presenled in a unigue siyle that follows the Soviet schoo. Finally, [31-45] are relevant co magoetoillric theory. REFERENCES 1. JR Ws Thea of magneton ee al BS SO ED) [LIL Fon, Magnets en electra aves seohce es Ge. Space ge 372), 22-2949. 1. Cosentino te magnetotcre maths of phn pope, Cac hs (185 dean by Re Wat el or re Gps {sri 350, 4 UsStrourInpainon of nduonanomashoveson-unfors alice jes Ge 6-5 7, 5, TE aver Ecmogrt indacon in est consi anomalies at the ‘ele of he ete, Pra EEE 6, (7) 144105099) (6 KC Mbit Mancre suing on hye eh ith oso Hoady, Cape Pp 14-757 1. CTH Soyer Comeguence need platonic ireen Pu dp Gop Ha 85-49 90) 4. 2 Wave kt in gre! pe ah by unl and sono ware fo Ee Pa fe. 286281 (1979). 9. Vet Drier ad MLN Behera. The ional mode of mayrioslue ovning re EEE (7 1038-1084 (179) vend wie {a eomeanie mitopttons nears cosine Ca J Ph 4h 48990 ‘sap als applied pops Gol 27, 1-86 92. 12, D Renkin The mazar ec ont die ipa 2, 66-6 (ED, eteenees 207 33. 5.6 Mane Maan ty of esl nei Cp Re ME he Sue Thoretil mpnosc and rin rape fo n-ne SShononiie Gay 38 38-5 1 ts, AIRTREdiman and faba Pew of Mpal: Sang Med spa we: a Ao aka. 16, Wi oer Te panna aed Se ae ah dina oe asin, ‘ 17, Rely dD. Rtn: Ohne mace a pis Sent Gin Er $9, Se OD 1s, ema Btecomarni apn or sgt gn essa, Ina e150) BLE (0 19, 1 We ad PSs Raa eso of he sates inp nd ae hor “arnt esha EEE tava Po AP 9 K.P Sots Semcetoabarver dice depen of th mace Tid Ses 11) ae ar Nana ay, ise Snel is te ea, o fenton age Rap Pip 72 Peg ‘om Ox tai New York. ieee 2, JR. Walrabd RPS Magneto fale (or epmendvebede, J as. ocr 260-1898). WWM and JR Wate Be wave and ace impede nt il lnhoyenous rye eh fa S10 (1 10-108 (978, 24 Woops 8 econ one = cer th i Snail ceorden, Pre pe Ge 1. 81398195, 2s, TT Waie ane B.C Chang, They of dctmagate seating trom lyre ‘edi wi ay varying ae eS 10h 21-29 OF 26 G.We Hohmann toms setae estar air le ou of arent, ‘Gprer 2D C 27, AQ" Howard, Eesromapie fle of 9 wbieranen indi! itamogery, ‘ony 3 97584 (97 1M ATACE Homa, J.B Ae ad ¥. Nokona. The magnetic to vs seit cai mon Bera Epo eM 1 2G Teenie Pcp of Magri Prin. GbR, Be 30. AA Kastman and 6. V. Kells The Maps Sending Matin. ave, 3 MA, Gali and D. W. Suungwan, Attn eqncney magne wth (pound ec pd sour, Cac 4, 69-0 9, 1, T'Ribate Eucrmaynin sth B's Pron. Ever, Aseria, 6 3B. 1 RNa En mtr nptin or el rc Pe MAU Beiceoy and. Brat, Thos pri of magsocelise poe il tod 361808 Gena. sheet af these as andomiy cous crt, Ges, J. Rtn Sa. 3% ‘ass 9,Vi. Magntotetre Theory cnt 5.0. pie and W. Do, The interne of bran cntoing Souls wom eth Gp Prge 38-09 4. 70. Opinace. Ekemmageate respons of embod fer fin ewsent ‘tion, pats a0 252008 Shoo Lacan re ess 09 age te 381-39 (9 Eee apsoutrs potig oF een ae Ar: Gp tap ser od Ses. The depen on fe magn peas Shc inene on he leon poem, Gp dR ato. Se 6 (oi, s Shao. Te dpe af te sss npn of eth ose loaf heap ld sore Ph, ad Ea 1 (6) 4-4 CH, I Rump a TR Meee hn he a en ‘Sura nf na ef Aton oh SAE sua 6. obo nd hop he lescegpetl ma velar onsen Gn 17-188 SP. tismom Eimer ncn the ath by moving anh eae re Ce 1-8, 3,373, BE Linen’ sna or ge meget nasn, Gap 6 thst ope ste appre sisies J | onaoter VII Inductive Transients ITropucTiON 4 electrical methods of exploration, one i always Faced with the need 19 etect a secondary field that is orders of magnitude smaller than the primary field. When dealing with timebarmonic fields, produced, for example, by sinusoidally varying curens, this separtion of primary and secondary influences can be very difficult This Ft le to the Kea of using pulse excitation and then measvring the secandery Fes of the buried body oF largetafler the cesation of the primary Tels. A sight variation of tis theme f to use a repeated sequence of souree eurent pulses and then restrict the measurements of the secondary to the interval between the source pulses, In this sense, the procedure is similar to that used in microwave radar as developed during World War TI, but actually the approach was first used in the 193Ge by Applaton [I] and is olleagues in Brin in probing the ionosphere from ground-based sounders, Of course, ‘coustic eehosounders were vsed by naval frees in World Wat T for sulbmarine detection, o thatthe basic principle is very ok. Por many year, ‘eophyscits have also used transient elastic waves to probe the deeper ‘rasa layers ofthe earth Unfortunatly, the direct application af the radar principle to the eleteo- ‘magnetic sounding of the eari's subsurface environment isnot a simple alles. The difically resides in th fat that we ae dealing with highly dispersive medium. Tn other words, the alenuation rate amd the phase velocily of plane waves are strong functions of the operating frequeney ‘The consequence is that the so-called primary fists may linger for considerable time after the cesaion of the causative source currents bsiously, this impedes the measurement of the secondary fields that result {rom the presence of the target. Another complication, and one that is oftena i Insucve Tasin ‘ignored, isthe inherent frequency dependence of the constitutive properties ofthe geologes! mate. Iu sour purpose hereto olin the theoretical prineiples ofthe transient dectromagnetic methods used in geophysial exploration (2-11). We Shall resect our atention to indvetive or nongrounded techniques that are ‘propriate for airborne devices. The subject has been developing rapidly ‘over the past several decades. The current trend i toward the use of purely ‘numerical methods that are capable of generating extensive compendia of ‘esponse functions for both two-dimensional and the=-dimensional coafig- trations. However, ia order to provide analyteal and physical insight, we ‘hall dwell primarily on the tractable geometries that yield to simply Understood result in erm of Known functions. ‘We consider fire the tsesiont response of ungrounded loop-type sources, located over layered structures that ae laterally uniform. Then we Consider esentil resis forthe simplest examples of bounded bodies that fate homogencous and inhomogeneous spheres. In site of the gross ideal ination of such geometries, a aumber of important physical principles are exhibited, Extensions 10 more complicated situations are then reviewed, land we soon dicover hat the analytical complexity is forbidding, The necesity of using numerical methods then becomes obvious. We mention Some of these Investigations and desribe salient features of the results 112-19 Following welhestablshed tradition, we present our resus for earth ‘models that ave characterized by frequeney-independent conduetivtes. We alo tenon, however, various investigations that have altempted tc include the uoually nepecied dispersive effect, including the role of dis placement currents in the ait andthe ground. PULSE-EXCITED LOOP OVER LAYERED GROUND (Oar starting point i6 a statement (Crom Chapter I) of the quasistatic solution forthe feds of an oscilating magnetic dipole Iocated over & layered ground, Specifically, the configuration consists of «small loop of aren at carrying a current I(ls)e™ that is Toeated at height hover a ered haltspace. The invocation of the quasi-static assumption atthe fouteet means that we restrict significant linear distances of the problem to be small compared with the Treespace wavelength. Thus for a given Angular frequency , we need to impose the eondiion w/e, where is 4 typical linear dimension and e = 3 % 10" m/sec, This statement presup~ poses that when dealing with «spectrum of frequencies de condition wal be violated only for frequencies that are sufficiently high t be of no Pusesctee Loop eer Layered round an fe Veta) an hn) mages Up sal cea loners eed sane hres bse cto fg eeey Sa consequence, In any cae, ce justification of the quasi-static assumption in the contest of transient excitation will be addressed in what follows ‘We choose a reclangular coordinate sytem (Fig. 1) in which the source ipo located at (0,04) and the earth's surface is the plane 2 = 0. Two eases are considered: the dipole, oF small loop ha ts ex ether vertically ‘9 horizontally oriented. Appropriate superposition ofthe results will of ‘cours, give the desired forms of the field expressions for an arbitarily oriated 009, ‘Omitting the time fucor exp(ien, the timesharmonie solution for the vector magnetic field, for: =O, is obtained from He —gadd, Here @ is « magnetic potential (as discussed in Chapter II that is given by Mead (afae\ a ay a ei) = - (alle=f Redisje™*“yaysda]. (1) whece my =p (2 ~ Wt and p= (329% The upper desivaie and the pla gn before he intel ed fr the vera pole oven aon, ‘hres the loner derivatives the mins sigh ae eed forte na ipole. The relecion coeTicient Ria, where the fo dependence ot22 Vi, lnovte ransots indicate explicitly s identical to the for forthe function Ry) siven by ‘Ba, GO in Chapter Il for an W-layer hal-space. Te the exse of wo-hyer conducting earth, where the upper layer has a thicknes d, we have (uta — ae! =(m= WHF) Raho eae Fa) where and where axe, an 4.02.09 aze the electrical constants of the upper tnd lower layers especiely (ct [] 0” Chaper UI) ‘Now, as we indiated in Chapter V, the transient feds H(i) for source current I} ean be obitined by applying Laplace transform theory. Tv the present context, we can Write Huy = 2nd 0, ° where a= EE) “ and where ©) i defined by (I), with replaced by s, Furthermore, the transform ofthe source current is obtained from Ho) = Lato)» [jet © wnete we choose Res > 0 to ensure convergence. Here L isthe Laplace transfer operator "We now ake the source eurreat to be impulsive. Thais, KO= 10 o where 80 isthe nit impulse funtion at = 0. Them Hay~ ho 0 and becomes wn (solsa)a) 222) CL Rantesnina 6) “This isthe formal expression forthe transient response ofa smal oop that ‘carries an impulsive current and is located over a layered ground. Puce Loop over Layard Ground a3 ‘We will now adopt « noarigorous asymptotic approsch to obtain an spproximate solution that is valid atthe shorter times To thy end We ie the expresion| Rha) = Saar ° where am dale) oy (29) assuming that all derivatives exist at A= 0, ‘When (9) is inserted in (8), we obtain the clase of integrals that ean be evaluated as [rename OZ) (Pama “(hy ay oF “Thus we deduce that had aq - at [Pol fe Le (2) So ayn att Ban) Sect a tf (2) where q(t) = L7Mq(s) and a3 In analogy to our development in Chapter IIE for the Ty. Ty and Ts ‘tera i folows that (12) i asymptote the senae that f(s +30] shouldbe large compared with one, where and are the conductivity and ‘exmitivty, respectivey, of any ofthe layers that may bein contention and / sa typical linear dimension ofthe problem. Thus when we perform, the inverse teansform ven by (13), the final reslt for the transient response i valid only at shot tines. To iusrate this approach, we deal expicily withthe twoayer model shown in Fig 1. Then RyA) i defined by 2), so that afer some tedious algebra, 9) leads to Facet ney Aaey= =, “ Ago) = lanes O(0) 0) AS) = ~ (9098) O%(0), 116)where on EHV Aap -24/009*] 1° TERA) RM x0 —24/ 00097] \ exo| - 2a |, 1 Yea} where ke BTS nd = ot He “The task remaining now isto evaluate the timesdomain functions ay(). Cleary, the case for m=O is trivial and we have a(t} = —8(0) This cancel the fields at 2 ~ 0 forthe vertical dipole and doubles the elds for ‘he horizontal dipole "The inversion of the Laplace transforms form: > 0 is more difficult. To facilitate the evaluation, we neglect disphcement currents so that and K'= a,/a) are eated as real parameters, independent of 5. Thus, for n=, wehave “12 a(n =h ) (= b3 ) [-tnsy 2m} 0 Ra APY + KP) = (oh oFA/(0l* + of) ‘These transforms are ofa standard (ype and we can immediately write * Jo 0s 0 for #0. Ina simi example +25 Rrenp(-2am')], (20) where a (oyo)'/"d. This is equivalent to aye - be (+43 4 3, Roexp(~2ans") +45 Sarvoolmeeen] 2 A, [te43 nete( ah) +43, Swerve SP Nun, an where ee(2) she complement ofthe err faction of argument Z “Te eapone fonctions (0) fr higherordr ferme ean be fxd nthe same nanreraknugh the bse compleiy becomes fridding However, in inportan obtain tat the aniter lop and the echng ieop are sticemy high So that (2 hyope/ °° the teas 1) teyond m= | ean be epee Tins the Cana respons at flows them impele at 0 i harris by » Md” Peg PM @ were nary= alta (154 a ‘ san {%)"" Te ctgs etme a nie=(%) Furthermore, since we have neglected displacement currents he result for ay) is valid only if >> ¢,/ and 1° e/6, in addition to the restriction rated above "The inpulse response ie ploted in Fig, 2 as 2 function of the time parameter T for varias values of the conductivity rato X. Tes important fo note thatthe waveform P(7) does not depend on the Tosaton of the roaree and receer loops, provided thatthe aforementioned sesitions are mel. The etse K-= 1, of course, corresponds 10 the homogeneous ground were the response varies as TY or 1-1. Note that, for the highly Conducting lower layer the rorponse has @ mare rapid decay. Convery, i the lower layer i very poorly conducting, the decay is somewhat slower. ‘Actually, the time. parameter T in Fig. 2 con be read diccly in milliseconds if we choose od?=10"/4s; of, f= 10" mho/m, then16 Yi. Inductive Taser al 00 e4 ey E ‘or io 7% ig. 2 To one fhe mips fran inguin erent nthe sour ap and for wry eth wo = 10°/0.4, and thus d= 90 m. In order for a(0) or P(T) tobe a good ‘approximation for the impulse response over the time interval in Fig. 2, the sum + of the loop heights should thea be greater than about 200 or 300 ms, “The inherent advantage ofan airborne pulse system that operates inthe asymptote short-time region i tat the response waveform is independent of the geometry and of the rcative loop orienations. Unfortunately, however the absolute sensitivity is small LIMIT OF THIN CONDUCTIVE SHEET “The magntic field responce for impulsive current excitation under more eneral conditions usually requires mimerieal evaluation ofthe integration ‘over the wave number A and/or the (aasform variables. Much effort has ‘sen expended in thi conten, and we shal evew some ofthis work later ‘bu ther is another case that yields fo an anaiytiel approximation. To this end, consider the same geometry as that illustrated in Fig. 1, but let the Tower layer be highly restive such that [y+ 2) or [yr 1s whence ma ptry” ch Also, in addition tthe inequality o,/2,7 1, assume the upper layer is sulfiienly thin so that yd. On the eter hand, tanned eed yd is finite parameter, Then (2) can be approximated in the ia complex plane by Raine—14 p= Brgy: en ee ar whe st nae span BERttte as tention 05 ‘where we have writen 6, in place of + cs In other words, the upper layer is represented 25'a thin conducting. sheet whose ‘conductivty~ thickness product is od. The effect of any induced currents inthe regions ‘extemal to the conducting sheet are ignored. “The resllant potential for the impulsive current is the source loop is thus given by wo Cale he rian) vena) (26) ‘This is an elementary inverse transform, and the integral in the above expression is thus writen as 2 [pee ee +2 a ht s+ 21/einad Sond ph hee +21/ound] Iv is now simple mater to obain explicit expresios forthe field ‘components inthe region z > 0 by peforming the derivative operation Hcy =~ ard. As indicated, the response waveforms for eich component are genealy Sieent nd they ane each dependent on the geomet ofthe probe. 8 to etl to not tht we an wre the expen forthe impulse reponse jin the form * 7) 8)ae i, Insts Tansee where 4 (ala? ag hO- where, a before, na[e sons] and now rave [et + + + 2r/oyned ) @ ‘The Function Y(?) of course, i the response of the potenti for a step- funeton eurent inthe souree loo. “The time-dependent (tt aumercaly equal o the distance from the observer (at height and radial distance p from the source loop) to the image point st s= —(r- 21/0) directly below the source loop. The image, of couse, recedes from =~ (0 —co a8 f ranges from 0 Lo oo 'An interesting ease to examine is that in which the reesving lop is & cireule wire of radius b that i concent with the source Toop of ares dt ‘The vokage induce in the receiving loop fora step-Tunetion current Ju(f) inthe source loop is obtained from Faraday’s law a8, cine mf Lott ndoa Here we can weite ‘Thus, for? =, = os where [ee/hd) + 2]0/4, Thayaay +2) + (o/h eo [lernay #2] + w/a) omegensous Earn or Ht Space 210 oF Os iF 16 20 g-3. Voluge rapa ss haat el pf sa sg or wi cencive seat how conc ckner Padu af The eure 4 sl gy Senet by tap fanetnn caren where A= 6, ayd/2. The quantity Ais plotted in Fig. 3 a8 «function of the time parameter r= tb =21/nh for various values ofthe ratio bt. As indicated, the rate of decay depends ‘onthe ratio p/h, Also note that for suficiently long times (se 1 > ye) the response varies as 27, HOMOGENEOUS EARTH OR HALF-SPACE When the upper layer or overburden becomes sufficiently thick, the expresion forthe reflection factor, defined by 3), simplifies to Reha) = A= M/O+ uy, 4 “PN and 7 fangs The sobsript 1 has been dropped ere = fev “The impulse response is agin given by (8), where now Rh) = -1¢ A 05) A+ PP + ons)0 Yi, Indetne Tanarts “Then, as shown in [12] the impulse response of the poten is so=-(22](4=A)ao=10} 6 where Hey~ Be [7#( ME )saamye ea 6 were BUD) = TFL ~ (9707 ee 7") 08) Integrals ofthe type given by (37) are convenient for numerical integration, 5 cemonstated by Olt and Wat [13 Tei uefa to note that if we ust rein the Fs two terms inthe series expansion for FUT), given by AD) [(&)" Pee 12 Pa Iyane san « rit 9!r, 9) then Jo= “Those integrals ate ofthe familar standard form, so that Zu Ft mn ay “This result is consistent with our easier results for (0) and o,f) in the limit of a homogeneous half-space, As we see from (19) and (3), inthe limit d—> 2, (0) = 2am)" !Fu(1) ay and 40) = ~2(050)" "0 ) ‘The complete time history of the impulse response funetons requires a rumericl evaluation of the integral formula for f() and for its derivatives. ‘Thus if we deal jut with the verieal magnetic dipole case, the impulse response ofthe vertical magnet field at (p.2) is given by Lames omegeneous Eat or Hal Space ar am & ee 1 T Fig 4A mage fl Beit eer omogenns lps For an inp sent in Panic lop es nn fly where 4) and T= 1/(oj4p") The fonction H,(7) is shown plotted as a function of the time parameter T ip Fig 4. Here the observer is lated at 2~ hand reverses sige for valucs of R smaller then about @/2¥ (is, R=3/2}, the decay curve is monotonic. The behavior ofthe response ‘curves in Fig 4 8 consistent with the expected li x OR HAT)~ and HAT)}~ ~ saga We can alo show that te lee time cesponse (raring a 1-9 for = 0) is folly consistent with the tere tansfons of ie fase time Sermons ) tor te. aysolution for the mutual coupling batween two coplanar ops Iying on the Surface of a homogeneous half-space. “A closely related transient problem has been diseased by Lee and Lewis (20), who restited the results to the one-loop system. Lee [21) ako considered the question of sign reversals, “The preceding results dealt specifically withthe magnetie field waveform {or an impulsive current is the source loop. Such results are equally valid for a stp-funetion excitation if one regards the reponse asthe waveform ofthe induced voliags inthe reccving loop. GUIDED WAVE EFFECTS: ‘As already indicated in the preceding analytieal formulations for the transient response, all displacement currents were neglected, This is wsually jstified on the bass that the conductivity a is much greater than the ‘roduct of the permitivity «and the angular fequeney «. Or, lo be more ‘efnite, signifieant angular frequencies in the reevant specicum should satisfy the condition o¢ a/c. The corresponding restriction in the time domain is crudely 17> ¢/a, where + here refers to the ine following the leading or trailing edge of the source puss. Typically for earth materials 2106, (where ind ge 10"* thus «/es10-” sec ‘Thos it would be permissible to neglect displacement currents in wonsient response ealeulatons if were somewhat greater than ~100 nsec; however, ‘here are some other problems, as we indicate below ‘As noted in Chapter I, displacement curreats in the air region play another role. At frequencies such that R/e is no longer small, we must worry about propagation effets. (Mere A is the incar distance between source and receiving loops and c=3 x 10", Thes in the coresponding Uransient response problem, should be much greater than Re. We discuss {hs problem ia relation tothe transient response ofthe fields in air for a smal loop on the surface ofa two-layer earth motel. Our analysis follows ‘hatin the paper by Mahmoud eta. (19 ‘The specific problem we deal wth here isto determine the Fels fo the configuration shown in Fig.l, in which for the sake of simplicity, we let ‘n= 0. Thus in effect, the small source loop of area da is lying on the cart’ surface, Then, using the general lime-harmonic solution in Chapter IL, we may write the Laplace teansforms forthe field quamttes in the region : > 0 a5 162)= Fee (a (- He +S )ro. cy c40)= ne Fo © es 108 [fern naaane] A npranya. 61) In writing this result we have used the ideniy JeQho) = HSM 00) + HSA0)= HE" Ao) + HE"oy, (52) ese Hg!” and {2 are Hankel funetons of zero order and of te fist tnd second kind, respeetvely. Tis convenient to write the reflection coefficient inthe form ya) + Rythesye Rytusye T+ RQ) RAOs)e > where R03) = (Wo~ Ha) a * Hy hes) = (m= n/t a) and where «=A? yf, f= ate Peak es(es + 0)). and 92 = pastes + 03) Here R, and Ry canbe interpreted 35 ‘ellection coifcients atthe inleroces = =D and z = ~d. ‘We can readily verily that we recover the quasistatic solution when lisplacement currents inthe ai are neglected. Then Rs)» Ris) 38 Sefined by (2), mating that here p->A. However, we-do not make this assumption in what follows, {A indented, the contour of integration in (51) i along the ea axis of ‘The integrand has branch points at k= ~ fy and A= —ys. The Fst of these i just elow the veal aus, Bearing in mind that, basa nonzere seal pert The second branch point is displaced somewhat further down the ral axis of} because yy has a substantial real part. Te integrand in (1) also ‘nus. a number of poles that correspond tothe zeros, in the complex A plane ofthe denominator of (53) ‘The next stp in the rather tortuous analysis to deform the contour into ‘half-plane in such a way that it flows a semicircle of indefinitely 5). Detours up and over die svo branch points are also shown, In adtion, we must include the residues ofthe poles that result Irom clockwise encrclemeat. ‘In the casein which we look atthe vertical magnetie fel component at28 Yi, lndtie Trane ma fed | Theconplet pln hoi hori contrition le he a sol ered ot Wath caine po consi cepueacre 0, an asymptotic evaluation of the branch eut integration about A= gy leads tothe result searda _ Qi Hi = T+ Arie? # ewe I]eT (56) 2a where Le Reuse T Raye is stratification factor, This result js valid subject 1 the condition that lye 2lta/ v0 >2/0,¢ra/e,)"2. A similar expression is obtained for the contribution 17°), te result of the branch cut integration about N= “iva, which is dominated by the exponential factor exp(~+,p). For (his reaton is contbution (othe (lal Held is insgninat. OF cours, it cold be significant ifthe bottom region ware an insulator, but this s not the cae here Ta dealing with the pole conrbutions, we ace to Find the complex rots Avot the equation “ 1 Ry AsiRyOsperp[ —20,0.8}d]. (38) where the fonctions dependence on ands is expiily indicated. Here we ‘ote that if isa solution, then so fe ay. AKO, if, and 2 ae solutions for a given 2, then wf and —X* are solutions for s*, where the asterisk denotes complex conjugate. Heace u; may be arbitrarily chosen so fong as this choice ensures thatthe waves (Le, waveguide modes) are attenusting inthe pose p direction. In the case of most practical inlrest, we can | sasume that |yi/ ls hen for the weakly atlenuating modes [A$/+] ‘= LAs consequence, RQ) = 1 2/28 —e99(—Ba/) 9) and Rates) (15 — wablln + 12)=(01!2— o8)/(ol!* + 64). (60) Wit these simplifying approximations, we can deduce that Ami am) 6+ 9216)! (sy where n= 1,2... the onler ofthe waveguide mode and ae [Qn ~ By = 2°/(Aa'ip) @) b= Ane — 8/2) @) where cy and Be O or», according 1 whether @, > a oF a < oy respectively, Also rote thatthe plus or minus signs in (6 ae chosen according to whether + isin the lover or the upper hall, respectively, ofthe complex «plane ‘The cortrbution of the waveguide mode then obtained by summing the residues ofthe poles at = A, The approximate form ofthis expression as np = He ne Ser Kala woUe + 0 ty) (65) ‘where Ky is the modified Bese Function of the second kind. “The final (and by no means Wrvial) sk i wo evaluate the inverse transforms to get the wansient response H,(). Symbolcally, Hay = L[H%a) +H) + HEE) HOO) HE + HOMO, (66) ‘Thus on taking the inverse transform of (86), we find that the impulse sesponse of the branch eut contnbetion ke dA 20 no #20) [Berm rt sake nant Sri 70 +fane- sa} oy26 MIL Inte Taine where eo ey ramet ox) 2220 a Ror ra) = antl | aaah (68) Here r= y/ey is the normalized time snd 7s pun)! 0y/e) is the normalized time delay. Using (57) and expanding its denominator, we Find (hat army e 19 4S ante) @) for > ra where the normalized depth is D = doi to/e)” “The contribution of the sesond branch point it the time domain can be handed in a similar fasion, but, as indicated earlier, plays a minor if nak insignificant role inthe (al transient ‘The transient response of the waveguide modes can be important, partieuaely at early simes. Mahmoud ea [19] have sbown that salta(aamS oven HAy= LNA 40,7) e-™sint, 0) where am Salter ives nlen(-p) 0 where T= 41/(0,px6*) and gp + 1) isthe Euler psi function of argument pe ‘As explained by Mahmoud ef of. {19} the impulse response of the ragnetie field H°(0) for the impulsive current TAC), given by (70, i mathematically identiesl tothe cell obtained fromthe singularity expan son metiod (SEM). In that ease, we readily invert the ofdar of the hand the s integrations. Then we first ook forthe poles 5, of Rs) defined by ulded Wave Ete zr (53), inthe complex s pane, These ace approx ately given by =a, 51h, B/C H905) , + where a, and B, are defined by (62) and (63), Following’ this SEM approach, the contour is closed by an infinite semicircle in the lef-hand half of thes plane. Then we must also consider the cesuling branch cu integrals assornted with points inthe s plane at which up and uy equal 0. Thus ia spite of the different approach to the ‘iret method ojiized above, the SEM leads to equivalent final results. ‘As we have indicated, the magnitude ofthe field response H(¢) forthe ‘urrent /,(7) in the source loop isthe same asthe time derivative #,() for 4 cutent fyu() in te source oop. Thus the response H,(") for a retangu- lar source current fu(t)— u(t = 5) f durations, seconds is obtained by siperposios Lew Fg thas) ~ Hle~ 4p 4) ‘An example of such eaeuaton is shown in ig 6, ia which the total eld response 7. as funeton of the normalized delay r~ ry, posted along vit the separate contibatins from the branch line integra andthe ple for waveguide residues. The spproprate late time response tbat was consid- ered eater [12,18] is indieated by a broken core. The ondinater are normalized by the factor 9/54 /(270, Tes evident from the namencal example sbove and from oter salen tions that the waveguide cortnbution to the teasiet response doesnot phy a role except for relatively short tines. Roughly speaking, the time Farameter — 1y= (= p/eley/e)s Where e= (461 shouldbe very Farge compared with 1 i we are to neglect the waveguide contribution wo my Fig. & An cramps of herons ofthe tne detainee gif for 8 seciguas sure ant pale fer Tbop ing the nace of layer eufona Ie tse n/a eee =oa Vi, Inevtne ranorte Since the ratio a)/¢; i of the order of 10°? seeonds, the branch cut contribution i of overriding importance unless we were dealing. with Towrloss dielectric layer structures, ‘BURIED SOURCE PROBLEM [Another related trensent problem deals witha buried source loop that is exited by a polted curent, The solution of such an idealized problem gives ome insight into how the observed fields on the surface are moiied by the intervening medin, This really te fist step that one should take in solving the fal-biown problem of determining the seaterng fom a buried target fora fansent source on the surface. ‘The specific situation that we are considering is shown in Fig. 7. We adop! homogenous half-space model forthe ear, with conductiviy ‘The loop, buried at depth His represented by a magnetic dipole with & vertical axis, Thus the dimension pang orf shoulé be atleast five cimes the radius « of the loop. The loop current is taken to be Knut), where tui) I fort > O and u(?)= O for # <0. The quantity of interest here is he Yollage n() induced in a small lop (with vertical axis) that i located at (2) the upper half-space (ie inthe at. "Tae analysis for this problem proceeds in a fairly straigforward fash- fon; fn fat, he inverse ansforns are ofa type similar to thse considered ‘elit in this chapter. The specific details of the derivation are given tleeer [15-1], When the oberver is a =» 0, the results can be writen in the form oye ee sare [rete 5) Fg 7 Asm centering ep base in = Reopens balls ute Source Probe 229 here we have indicated the dimensions ofthe various factors. Here ¥ isa ‘imensionless function defined by VDP y= gi e(— Ab) [+ s— rev asd je 05) where P= x(ougl?)~! and D= p/h. In 0), the area of the 22ues lop of mia = number af usin vores loop, = ates of seceiving loop (of infinitesimal sie), and N, = aumber of turns of weet we and N,= sumber of turns of eeciving As indicated, (75) is convolution integral that involves the impulse response Y(D, 7). When the curent I() is 8 sep-function [ut where , isa constant, it follows thatthe vallage ° ociye| AMAatety at |B ae om wore ¥(D.T)= Lye. Pyar, (73) where T= #/(opeh Using straightforward numerical integration, the dimensionless function Y¥(D.7) i plowed in Fig. 8 as a funetion of the dimensionless time Parameter T for several values of D. Note tha the time parameter T ean be read neatly in millseeonds if 4noh®~ 10' i, for example, o= 10" tos/m and fh 283 m. Of couse, we should remember that ur results are valid only when ¢ is somewhat greater than e/a seconds because lisplacement currents were neglected a the outset ost 2 el Nomit wai roe + ena ese o>230 Vik ndustve Taste Similar caleulations were eatied out by Wait and Hill (16] for the voltage induced ina smal loop in te air the axis of which i horironal. A similar distortion of the pulse shape was obtained. Such results sugges hat ‘waveform measurements onthe surface are highly diagnose of the relative location of the source that is wansmiting, Such deletion and ranging techniques are believed to have application to mine reseue procedures; teapped miners could use a rudimentary shock-ecied! loop as an elarm signal. ‘An immediate application of the reciprocity principe els us thatthe role of tie two loops in the above extmple ca be switched. Thus the Fesulls in Fig. 8 apply equally well tothe ease ia which the transiting Toop is on the suvface and is excited by a stopfunetion current. The ordinate Ys then proportional tothe voltage induced inthe receiving loop buried at depth hand displaced by a distance p SCATTERING FROM BURIED TARGET “The next fvel of complexity in dealing with transient problems is to tocate a scateing object in the medium. Probably one ofthe most basie configurations is & source loop oF magnetic dipole on the surface, wih the buried setlering object eepresented by a passive loop and the resehing loop also on the surface (Fig. 9). Agnin, we idelize the excth as a Ihomogencous ballspace of conductivity o. The magnetic permeability of the whole space i yy being the value for fee space (ie. = 42% 10" W/m). ‘is clear that the present problem is really a superposition of the wo ‘arir problems, Thus the ist step isto deduce the voliage induced inthe feallering loop. Knowing the intemal resistance or impedance of this Fi, 2 Gee for the sou an sie pet the ce nd rng op a8 scp st et dae shows ures Seber in Ha Space at scattering loop, we then calculate the secondarysouree current wavelorm that in turn propagates signal to the surface receiving loop. The full etsled analysis for this problem is contained in the paper By Fill and Wait [17 These authors chose a mumerisal example in hich the time constant (Le, seltindctanee resistance) of the scattering loop vas 630 sec and showed that the scattering response at the receiving loop was ‘erainly nonexponeatil inform, with a much slower decay for late times. ‘BURIED SPHERE IN HALF.SPACE “The most obvious target, of cours, is a conducting sphere that sould be located within the hallspace or even within a layered medium. The ‘Simplest (bat rather waive) approach i to assume thatthe target sphere is located n'a locally uniform magectc field. Then we may eall upon the ‘reat volume of published results that are availble Tor homogeneous and radially inkomogencous spheres that ean have contass in both condutiv= ity and mageetic permeability. A shortlist of papers on this spesalized topic is included inthe references for this ehapter [2, 21-27), "The principal dificuly, however, is thatthe assumption of a locally uniform field limits the usefulness of the reslls so ealelated. In fact, higher-order induced magnetic dipoles and als elacrc dipoles will coatib- tte to the strength of the secondary source. in a complicated manner, the receiving loop detects the resultant signal that is offen very dificult to incerpret A recent analysis of seh « composite problem has been earied ‘out By Lee Gee Lee [28] and Lewis and Lee [29), This prolife theoretician has atached great significance to the relative importance of the pole and braneheline singularities in the complex frequency plane. As Lee points ot, the influence of the finitely conducting host roek ean completely change the character ofthe intrinsic ransient response characterise of the target, in Tocaly uniform field. The ler is described fully by discrete res ances associated with the pole locations inthe complex frequency plane; ‘but the influence of the adjacent conductive rock is to smear out the clean-cut exponentablke signal from the target [29}. This is 2 manifesa- tion of the importance of the braneh-line coniribuions, This point was ‘made very emphatically by Singh (24 in a careful analysis of the transient response of a sphere for dipole excitation, even though he assumed a Ihomogeneous host medlum of infinite extent (Le, he ignored the elfect of the air-earth interface). Anticipated numerical results based on Lee's [8] intricate analysis should provide much-needed insight into this eruc question,22 i. tnductve Tain FINAL COMMENT ON IP AND EM ‘A final and sil unresolved question in ansient electromagnetic sound- ing of the earths crustal layer i the sgnfieance of induced polarization Even with an ungrounded inductive system, the frequeney dispersion ofthe ‘complex resistivity of the constituent medi wil influence the time-domain Fesponte 21, 30) The untenglng of the IP and EM respons isa task that ‘het nol yet Been completed. Ultimately the dual cole ofthese phenomena in any electromagnetic exploration system should be exploited as a source of Fich information about the environment. A selection of rested papers [51-47] on this topic i given ia the ist of references. [REFERENCES 1. EN: Avplen and G. Bde Wis shasta ly, Pro Mh ae (Lonon) Sets 2. TR Wales contocing stein 4 tine waning maenec Fld. Gl, 16 esr sh Ti paper nai a ey dcr as ii pe Cinvenpl the case kay sarc Wee the parame 8 se ae board 3 Seal. US. atest No 2 75, 950 Ferry 2,156; ed Deb 23. 1958 {te pet speiely doe te ipso omission wae he ep bg Ie dete eel twinge wat pao ent) 4. TAC Wai ‘Te fst eta pgecine meh capyng te varying ie Fie, tee Univer form Maoh 191m pence ep se (rome atone lai wave cae wear 4. TR Wart th monte pt nee heron aed eth Com. Pp. 28 Sram ish Tat ine hrsne and ann oan ein homoge ‘rapes mess econ] 6 WAE'Dolm.Gropgsl anne ot eri Brel spe Holes in wouter ‘pos in Maing td Croan Water Cop, Esnome Gly Rapa No. ‘Ghia Suny of Canna Ota 9. 1 AoA Kaonn Esto nts bse on aeaaonary eomagnte Ha thence sae fr de Set USS Pye Sl Ea No, 83-9 (9, 4, Chr wets Sd: Drain, Gr Maza, aod tN. Nabyhin Expraton ‘hh News! EM pc apr pest a the te spel Earoein AS vet Expnion Gepipuckn(EAEC) sete Debi Jone ITE 9. Giana Oe SROTEN a ox ee toe {rat mera Ball ts Soe Egor Geir 82-8770) 10. ic Parma R Math, onsen soap Chplr Sin Gong Pips, 1s, LE ReadMe aera stomagnte doer of the Dou ne- r> a= dIn particular, they hold as ra; whence Z and Y are designated Z, and Y, [GENERAL SURFACE IMPEDANCE AA thi sage, i is useful define a surface impedance matrix for the (otal fields (8) a8 y= ally + ZH A By= Zale ZapHy|, 69 for a given spectral order m and a. Here the 2s are the elements of a sre impedance matrix. By making use of (32) and (34) and bearing in mind that y= Ezg-+ Egy, ele, it follows hat (36) is equivalent to [llag~ Zan = Zag] +[¥7 Bay ~ Zalan ~ Zl e7) [2th Zatas* Zaye) +[ YC Maa Zee Zap] =O (38)200 ‘open A. Electorate Spherical Waves Here all quantities are tobe evaluated at the limit r= a. tn order that (37) ‘and (38) be satisfied for any general combination of TM and TE waves, it isevident that each of the four square-bracket (rms must be individusy ‘equal {0 0, Note alga tha for a given spectral ordet 9) Zt (Kay? ate a) We have obtajned here the exact form of the surface impedance matrix for a concentialy uniform model of the spherical earth for a spestal component (ie, texseral harmonic) that varies 3s. (cos8)exp— ‘The radial Impedance Z, and the radial admittance Yas they appear in (4), ave obisined from (23) and (35), where it understood that they are defined atthe earth's surface r= a. OF eaure Z, and depend on a, the ‘order ofthe tesertl harmonic, but not oa its degree m. However, because a ‘dependent on both n and m, the elements of the surface max (i Zy, tic) alo depend on both m and m. Indeed, the surface impedance is Function of the lateral variation ofthe causative or source Hold “There is «great simplifeaton in the ease of azimuthal symmetry. Thus, for example, if m=O, we see from (00) that Zyg=Zyy~ = Yy' and ym Zee "Anotier obvious simplification is posible when the earth i effectively & ‘horogencous sphere, Cleary, his willbe an adequate representation wien Irv Is where di the thickness ofthe upper oF outermost Iyer. Thus we deduce from (33) and (35) that Zila Yon heen Acnas ‘where /, i «spherical Bese function defined more completely below). It ‘an be Showa [8] that the right-hand side of this equation can be approxi tated well by 1, provided that [ya (n+ 4). Then of course Z;—y tnd Yj, which i esence i 2 statement ofthe Leontovich boundary conditon (9 ; “An alternative Jevelopmiet [8] indicates that if (ye) # (n+ 2) i large parameter, then it isnot difficult show that 24/1 =Y =~ SA where S,= i(n-+$)/(ya))- Here S, ean be interpreted as the ‘cosine of the complex angle of xtra Excation of Spare 2a EXTERNAL EXCITATION OF SPHERE With the basic framework indicated above, we can now deal with the extemal excitation ofthe sphere. A the surface r= awe stipulate tat for flecirieor TM modes of order n Ea ZH, and Hy \ ashen a “ where 2°" is the surtace impedance foram individual TM mode, Core- Spondingly, we can write forthe magnet or TE modes of order Ig Ty and Bye Thaty (0) where "isthe surface admittance for an individual TE mode For the region r> 2, we now need to postulate the required forms for the Debye potentials U and V. If this external region is homogeneous, with fectieal constants oy 444 ue, We ean Write for a <7 < > BE [Anwhlle)+ Mauthidoyje MP0s) 43) and = BS [Gah te) + Dual (kr)]e-M™PE (e088). (44) The factors in square brackes in (43) and (44) are the radial solutions, They sisty GBI} where y= — k= jpato, + lee). Jy and H, are wo Independent solutions of this equation, These are Schelkunof T's spherical Bessel functions (10, defined by Sse) = 2/2)" Sysuyaf2) (5) and Auge) = (92/2) "(Seve a) (46) in terms ofthe conventional Bessel functions [11] and ¥ of onder n+ “The costicens fa. and Coy in (43) and (4) can be specified from source conditions, whereas By and Dy ae determined by the boundary ‘conditions, as we shall indicate below. The summations over m in (43) and (44) are quite genetlly for integers from n= Ota = 2, whereas ranges fo nse Fom =n fo inching ne Ge not that P= for im > In writing the forms of U and # for (4) and (44), espectively, we have resisicted attention to the region a << ry. Here 7 is ehosen, atleast for the present discussion, to exclude any ative sourees from this region. iY,2a Anpandix A, Cheomageete Speticl Wows In view of Scheksnot?'s normalization [10}, we have the very simple symptotc behavior fr kr JIe)~sinke a and Ar ~exe(— ik) ) In fac, for n=O these are exact re for all fr. tm any ease we ean ent the terms associated with H,(kr) as outgoing waves, whereas the terms associated with J) are related to the exiation ofthe shee BY sures ator beyond the surface = 7 "We ow wih to obiain areatonship betwen she unknown coefficients by exploiting the boundary conditions 12} Here. fr example, we nts tht Fula 3 (200), and oot a Ha RE FV ae (80) “Thus it is lear that the iat equality of Es (4 ssid it (Bah tary + Beth hey) = ZOMG, + ef Apa KO) + Bgl Ke) (st) ere we may solve 98 By rms ne 0 ik As(a) ~ ibeh (ka) Latte a tose | Cy 3) 1m ih/( 0+ eas) rt ea), “The dimensiones quantly 4, isthe srfsce impedance for TM modes ‘rormalized bythe nasi impedance ofthe external medium. The prime fn the functions J and fin (32) indicates differentiation with respect to the inetd argument, ie, boy ~[ahtev/ ey oH tis eny to demonseate thatthe second equality of Eas. (1) is ako used by (43) when 2, is siven by (52). In similar vein we ean work Conceneaty Leer Mode! 243 With the TE oF magnetic modes and apply (42) and (44) to deduce that yw | ilhed= tthe =e Ai(ka) — 18,48, (ea) lena 6) where ae vey, (55) ‘The dimensionless quantity 4, is the surface admittance for TE modes ‘normalized by the intrinsic admitance 1/9 ofthe external medium ‘CONCENTRICALLY LAYERED MODEL, he ee Cee ery hora ests eee Sl pe ae ee enn er ace eee Te (atop # yey] and n= [igure + 0) Ne now cir spat fom of he tons fo the Fels within the pth concen region tar Wounded by rm, and r> gee (Fig. A2). * i ‘To ils, we dea withthe etic oF TM odes only, The come sponding rest Tor the magneto TE made low by enalon)2a everdin A. Flromagnatc Sparc! Waves Fe. A2 apa ieptence 2, irscon afr a rina indo 2 87 op fr he ‘itn he eto gs <7 ee may wr Um BS [aehlagl + BaaRayenmrra (ob). (7) “The factor in square brackets isthe radial solution. I satsies (31) with ym yp Here fy and, are independent solutions and, agin following Schelkunoff, they are defined by Lea)= 2/2)? co and 2) = 22/0)" Bercreh2) - in terms of the conventions! modified Besel fonctions of order + §. The coefficients day aD Fy, 48 a8 yet undetermined. ‘When 2m we have the simple asymptotic approximate forms Rce)mean(—2) co) ana heaymsians, 6 ich, fact are exact for = 0, Isao worth noting that (ata! © oer @ and Ct nial @ H@ey which are exact finite series representations that hold for any integer, The summations extend overall inlegers sTrom 0 to-m. Finally note thatthe ‘The Rerave Staton 245 Following dents hold: cr 5) 66) o “The “moditied” spheres! Bessel functions on the leehand sides ofthese four equations are used as smaller of convenience when the region is highly dissipative, We choose these forms when dealing with the concenic ‘pherieal regions for r< a. (We propose here that J, and &, be called Schelhunoff fonctions in reeogiton of the uemendous contributions of Secpei Schelkunoff to the mathematics of cletial wave transmission and radiacion in dissipative media) ‘THE ITERATIVE SOLUTION ‘We now return to (57) and aot that forthe pth segion the radial salution| fo the th order TM mode has the form l= aah) * Beat 63) ‘This form ofthe solution will hold for any concentric reyion: for example, in the outermost region or surface layer, we merely replace 1, by 1 mi two new sets of eoeticens ee 398 yy TO Be explicit, aL = abel 09 Pray VP) @) forthe region a > F> ay. When we dea wth the homogeneous core agin (Ge, Orc op the soln ataly has the Form ALY = etaahelter «70 bic is fit as r—>0, The infinite behavior of (yr as 60 reais hatin effet yes be Went 0 ‘We now noe that, proeeting from the Pth region out the fist region (is,p=) we have 2P~I counts Now the objective so ota an xpesion forthe surface impedance 2)" ata, 2, This defined in Scerdance wit 3) by PAM aah CHe) + Paakel re] OF He ytd 11* Bay Hrl 2 m)286 AnpansixA. Eusctomtgnetc Spherical Waves caluated at r= 4, An equivalent statement is £1001) (Baa! od Rie) "Te * (Bal inal RoCe ‘The boundary conditions tellus that the tangential electri fields and the {angental magnetic fede are continuous at each ofthe spherical interfaces OM P= Spud py eo dyes Bata: These cor [Fawn], [7 aho],.. i) 70. ~ and (ett Help elllpoid (96% eer HY ‘Thus we have 2? —2 linsar equations to sole forthe 2~2 unknown cenfficents In procedural sense, we would fist apply the boundary Conditions at the rap interface to obtain an expression for the ratio trBao/ pert The Process i epested atthe r= dp iMlerOce and is Sontined to The r= ay interface, whenee the 30 |B ne 6 fly ‘Spied. Thus the desited form for 2{" is obiained from (72). "The algebraic process described above is puely iterative and requires n0 tedious mai inversion, In fac, wees ew on an aralogy wit ansmis- Sion line theory” The one complication, however is thatthe transmision Tie s nonuniform inthe sense that the propertis ofthe line depend onthe radial distance along th ie. "To explain our anlogy, we need Lo introduce some defiiions, We weler specitcly tothe pth secon a illseated in Fig. A2. Fer a, > r> dx the characteristic impedances of the equivalent transmission line section for the nth TM mode are HE np Mo FOI ED) 5) nd Re Cyr = HB Raloen «76) ‘The supersrinls plus and minus refer specifically to the inward and ‘outward directions, respectively?” For example, the invartelooking, wave Tunetion forthe nh mode is (70, = haere Pe (cos), m "No cof hs rea rom wing he sabe Kan few lpn snd ‘fora Sena! oso Te tate Selton ‘Then the associated inward-looking impedance is defined by aru far me which leads to (75), An analogous development leads o (76), ‘We also inrodice the transmission ratios Ko = BG /Mog= Bal Wa Rit) Aittyed Fe Rerpteen) Leogag Rec) brood At Qo) Hyer) a7 roy oy, (60) for a mode of exdern. For example, can be intepreed as the product of (i) the ratio ofthe electri field Eon Eat ‘and (i the ratio of Egy or Eq, ate to tat ar yest tata = Spas The exresion for §,_ has 2 corresponding intepetation in term of the tangentst ‘magne fics Using the above-detined parameters: we ean write (71) in the explicit + rade Zi = KS ed REE + where ° pw MASK cn MTR R (re) ana yyw 2a Ke ‘Also, in turn, we ean write 28 = Ke 0 San For any value ofp we have _ (/ZI2)~ OK Oy ap00) C72) OTR Oppo) ny 2 @) co) @)28 Appanda A. lacroragnetie Sparel Waves and yw BBE R ed a FER Odea)” but of course, forp= P. etn) Tornoe) “The determination of the admittance function Y}" for the TE modes proceeds in a similar vein, Here iti convenient to regard the problem as the dual of the TM situation, Thus impedances become admitances. For example, the characteristic admittances for the pth section are written as 3° “) fegn, soy t HG 8 MOP" 5 Foe « forthe ingoing wave and Lay) - 1 ys Mz typy= — 2 eed 5) HO Teen o forthe outgoing wave, whee co) “Then, i analogy to (81, we have the following expression for the input admittance forthe TE modes of order at r= a Left 1 = Mn) on where 4, and Saf a defined by (79) and (80), but now ge — PERU ond) a and 3) For any vale ofp fy, and follow analogously from (8) and (6), Exercise, Verify by a ditet solution of (73) and (74), for a tworsyered sper, that the required form for Z, as gives by (BI) is obtained when Pd Dial Eton 248 Solution, We make an obvious contraction in notation and note that from (@ Jiry= Anny + aR) os fore, >7> ay and feo) = Cl 05) for ra at @=0 with reference 10 the spherical coordinate system (70,6) at the eenter of the sphere. Our first tisk isto determine aa appropriate expansion or the primary fields of the ipo soure (12, the Fields that would exist in the absence of the sphere), Because ofthe obvious symmetry, we ean see from (23)-(28) that 2+ 2 Jour wo vB) um, cot) and am are the only nonvanithing (eld components, The superscript phere, for these pure TM waves, denote pray. "We now wish to consruet& form for UP that eegenerates the desired Fields ofthe dipole, For example with reference to a spherical coonlinate Deal extn 251 system (9) centered atthe dipole, we have the wellknown form: Ids on ost me Msi (103) winere sind = (rsinO)/R. An alternative form ie ma =48 a(t ie (Fe «19 eping in mind that Ra[B +P 2brcos8)"”. (On comparing (102) and (104) i is immediately evident that ba “ Ula ab R U3) is the desired form, We now make ese of « wel-known aon theorem (11 12 shat nour Sask Forsnean| omlene by ‘sing ths elt in (105) provides he dese frm of he primary Debye potential ‘The resultant potential in the region r >a is given by unureu, (07 (106) wre 0 isis (yuo (008) an ge rie to otng mesa 72, Becaesinahal amet Tals (ie, 3/0 = 0), solos for U" mt be ofthe fort RMP (6058 tere n= 041.2 WS tre mow toile rte resultant Dye poet fo the eo 13a the coments carson pepe rhe on alt, Fane Tepes kaye 209 | Kacroyhrn) Feet) Ric oa +8009) rican, (109)22 opendix A Elcemagrats Spee Waves where the coeficint 4, is yet to be determined. The rule indiested in (106) Spples herein dstermining which pair of functions in the square-rackeied (erm are to be used. For example, if a < r< b, our solution bas the form un Su (110) were rum cont x Retry) + ARn|Pscoe®). (11) “The appropriate boundary condton fr these TM modes scaly Beeiye tet iene erty)» a were 2 isthe surface impedance defined by (81). The subscript 1 has ‘bea dropped. ‘When (112) applied 1 (111), we immediately determine that Ke (ray 2" La) pny ae [Ee les “me om how) 0) 6 Ke (va) Tow aay ot Boo as Rta)" where 7= 7/(o+ le) inw/y isthe intrinsic wave impedance of the external medium. ‘When (113) ford, i incerted int (109, we have a complete solution of ‘he problem of the radial electric dipole located at r= b, We could just as easily have worked with the unmodified Schelkuroff functions Jy and #, as Sefined by (45) and ($6) Iti naeuetve to repeal the Jerivtion for this ‘eae, but certain obvious steps willbe mite. “The adion theorem, as piven by (106), s equivalently writen eas = Aykoyhtery — (°) (119) where k= —iy has a vanishing imaginary part for a good insulator DeoleExetation 5 ‘We now wt the slo for he esiant Daye poesia i the eon > ain the form ‘ere a = alte UF ea toy Senn by dh «(eae tigen| agent a whichis to be compared wit (109). The eoetficent 4, (nt the same as d,) ‘is determined by the boundary condition (112) We rendily deduce that | KtGh 20 Rigo ze | tn Later 5 yy, KE (iy + 2 | Fs ckay a8) where (ka) Bs¢kay thay ltt) gS Ke (ikay "Fda "Thay (119) and - Riciho) _ (hay (a= Ca "Rika ~ "Fi ckay (29) Here 2°") = Z{" is again defined by (BL, which we elect to express in terms of the modified Schelkunoff function because the inlerior of the sphere (i r-< a aj) i normally the mediam with high loss, We now use (117) to obtain an explicit expression for the horizontal lect field component Ey at (72,6) for r= > a. Using (25) we se that = eon Feo tava BOD ~ (ee ral ‘ ete RO A canytictry| (C080), Hk )5(b0) Fay ay Pecos) = 4 P(cosd) and R, isthe square-bracket factor in (118),258 ‘Aopandx A. Bectomagnte Shwe Waves [Asan interesting aplication ofthe reciprocity theorem, we ean now use (0 for e(0,7) to determine the radial elec field a (79,0) For a horizontal electric dipole sat r= 6. Thus £,(0.0.0) = e(7.) oy “Then, bearing in mind that (8,9) most stisfy (P+ RYE =O, (2) ‘ve can immediately write that the radial electric field at (70,9) forthe Iovzone Spas at r= bs gven by =i bn On Tene Bey qh tsk } 1 x {fi + RED cary] n2eos0 core, {Ayia | Bo Fihay HABE Bolen cont (a4) where = [R ike) = ZY/[ Ke (ko) +29], (28) [Now we ean observe that for @ < ¢< b (12) is equivalent to Jeo En +(e (+ B)E cootser+ saith] ecco core nett) SOY aaah + Bathe |reesb ease. (126) Ths an and Bus / Ani = Rol da(Ko}/ Hk) (128) We now must obtain an expression for H, of the same horizontal elects ipole Here we procesd more diretly. Not frst that te primary fields of the dipole can be obtained from a Herz vector that has only an x | ipa Exton 255 component TI. Thus we write T= le 2) ‘Then, of couse, H= (9+ eayeas tl) an, ori) tg Seinen), 30 were N= -Msing cay aad 1, =T1,cosBoase om Then we dade that for <8 = = - sin, ar 999 = Le Sens nieonteyietysng a(e+ 2 )ern, (e+ Bon wy sere rem Sasdanriient ng oy provided th i(2n+ ty fds Ft kb) i204 1) fds tk) ry ae [Now we can write the appropiate form forthe resultant Debye potential V for the horizontal cece dipole at (7.8.9) for the case in which acreh: hn) + Btyth(hry]P(c0s0)sing, (136) ‘where Rp(Iq kay Ha), can286 opands A. acomagnai Sphere! Naver where = [ME (ike) — HY [Chay +), (38) here Mg (tay = = 4 ets) 19) ike) — 5 Fay (39) and Mg (tay = & Patt) (14) 1 Fo) chieved a complete derivation for the resultant Debye potentins extemal to the concentrically layered sphere for both radial and tangential dipole excitation. The corresponding soluGon for a ‘magnetic dipole follows from duality. We merely make the wansformations EH and H-> ~E, everywhere. The eleciric eurent source /ds becomes magnetic evrent sources that i, 16> ld, where ifthe toil ‘arrent in the elemental loop of area d4. Other transformations to secure {he duality are uw >a-+ few and 0-+ ise ie, everywhere, The speci case of a homogeneous sphere in the presence of radial and angen ‘magnetic dipoles was worked out by March (13. ‘THE QUASLSTATIC LIMIT ‘We continue our discussion ia the context of electri dipole exc ‘The quanthies of special interest are the rajlection fete for TM. modes and Ry for TE modes. These are defined by (125) and (138), respectively Tn many geophysical applications, we deal with wavelengths that ae long ‘compared withthe dimensions of the spherical body. Thus for example, fer, where k= 2e (wavelength in exteral medium)". Also, under sich condone nal Fy ok whee he propa Son connie ny ofthe cones rego Ta prea ney tat [Ji 1, where y, is the propagation constant of the outermost oF Cr layer Under Hach conden ts not fia to show tat Ran lyn cay 1 tht he sphere ating at «psec conduetig largest asi ‘TM mogés are concerned. For example, note that ‘sre a) ip = yh = Tan) (1a) | ‘The Ouse Lint 2sr ad : 1) n oO ya 1 “on "3 Co in the cs in wich I ‘The ieteqvency ov asisiac Unit forthe TE modes i more inresng eh ee wed hae f= : ; : 144 (fen Talay TA Gy a wee a= 9/42 (0) 45) and a= yay = yo Abo note hati p/p were sient args ne=(n y/n 046) Jn analogy to (141). But this limit would be realized seldom, if ever. I fat, for many cases of geophysical interest, we need 10 examine (144) for a range of values of the parameter a tnd j/. This sk i simplified if we neglect cisplcement currents, Als, we resiiet attention to the dipole mode (ie, n= 1), and furthermore we se Q, = 1, eoeresponding to & fully homogeneous sphere. In this atherexieme special case we write Rpm +N), «4 where AF and Nae ral functions of (9)! and wy / ‘The functions AY and, so defined, ae plotted in Figs Ada and Adb, respectively, a function ofthe normalized fads for various permeability fatto. This particular notation is consistent with my earlier work (14) on the subject (in the early 19505). The reader should be cautioned that in his Comprehensive review Ward [IS] cheaged the definiion and denoted (AT iN) by A= AN. ‘As indicated, forthe case ny = pin Fig. Ad, the in-phase response 3 ‘vanes from O to T when the normalized radius ranges fom small to large ‘values. For the same range, 3W rises [rom Dand reaches a maximum before ‘decaying to D for large valves of the normalized eadius. The main difference For te cases in which p/p > 1 is thatthe iephase response is nogative Tor small values of the relative radius. Tais low-frequency limit actually core- Sponds fo magnetostatcs, and i snot difficult to show that sig 3M 2( 4 fC +28) 48) sand that 39 "To illustate the nature af the external feds inthe quasistae limi, we ‘can write the explicit expressions forthe secondary felsexternal to the258 poperdix A. Electromagnets Sparc! Waves Ca v2, o Cr 05 al #031 ol oF oe 8 00 Boo ° opel Fe Ad mph and guna () componsns of iad ple momen for concn arte phos fase fh nae adn ‘phere. The electric dipole source, as indented in Fig. A3, may have any’ inclination angle g. Thus we deduce that Lessing cos ar(a eo) 1ascosy Zo"! nt p co HER S pion Hebe) 0) ve (eos) ‘The Quast State Ln 259 Tor the electic o¢ TM modes and peu - Hsinysing 3 ine (30) As indicated by (149), the induced electric modes have no essential fe- ‘quency dependence. They ean be induced by both the radial component ‘Of the souree dipole (.e, 1d¢con4) and the tangential component (i, Tdssing). On the other hand, the induced magnetic modes have a strong ‘and inherent Frequency dependence because ofthe factor Re. In particular, this produces a significant quadrature response that is illustrated in Pig, |Aab forthe case n= | ‘Te illustrate more effectively the eature ofthe solution we now further reset the situation tothe casein which a®-< rb, In this case only the Fist ‘n= | or dipole term ie needed inthe summations in (149) and (150). Then wwe Find that : Deingsing Re 404 1m fi [Bungano ns 40-0), sy jx lise” (o- asin yin corp + doosycond], erm ge tte Ta [0—taitsngconpsdconpense), (10) j= Adee | — Bsinysingcond At + singsing + yx tite Lainyringond SE singsing so), 15) p= pli [0+ sinyconteosy + 2eosysind]. (15 Foe + aye Idsa* Esing cose RE _ sinycos cose + 2sin pcos pm He [—Banacoe hl —snyouten ngs (158) ana need = singsing + Fafa kayo [O~ sinvsing + 0) (136) ‘The factor in square brackets in each of the above six expressions is ‘wetlen af the sum of thee pars. Te fst ofthese tthe influence of the induced magnetic dipole and is proportional to siny. Tt thus reaches a ‘maximum fora purely tangential dipole, but it vanishes for w purely eal Aipole. The second and third pacts inthe square brackets ure proportional to siny-and cosy, respectively; chus dhe can be identified with the electric Aipoles induced in the » and + directions, respectively260 Aopondi A. Elesromagnate Sphacl Waves Actually, the particular forms of (1S1)-(156) ate not resected to homo defined by (148) form = 1, must be [A FURTHER EXTENSION ‘The model of the spherical body just described consists of discrete homogeneous regions Simulation of a body with continuous variation of the properties can be treated 25a limit problem. That is the continuous radial profile is replaced by an arbitrarily large number of smell steps. [a dealing with lossy fi, conducting) media, the convergence of the process is usually quite satisfactory. However, an altemative approach to suck ‘Sivations ist allow the electric and magnetic properties (0 be specified as ‘continuous funstions of In particular, we may wish (o generalize oor ‘concentrically stratified mod! seh tha one or more ofthe layer rete a being radially inhomogeneous. We now outline briefly how this type of problem can be handled. ‘Consider a radially confined region in which the local propagation constant (is defined by ao 1) = Laer! Locey + ergo]? cast) and the coresponding intrinsic impedance is ne ner) =[inine]/1(7) (138) where the properties of ofr (7), and p(r) are continuous. For TM modes of order n, we can then consiuct characterise impe- dances that have the form KE) ATO TA) 59) Kone AIBN Ho. (160) Wwhete T, and JY, are independent solutions of the equations (i) om anna ‘The corresponding characterise impedances for the TE modes are of the I I otonees 261 form Me On) = Eean Tatar) 2 and Me tory = aan os (ss where F and W ave independent solutions of the equation (este Hodge (Zao oon Of course when 9() eran a) ae constant wi espet to over the sange of values being conileredy 7, and, reluce tothe Schakunelt function fy and W, and, reduce foe Schlunot fenton Ke Note thatthe primed factions that our i (159), (10, (163) and (163) are defined in secordance with a0) iam, (165) ‘bearing in mind that y i r dependent. Ton REFERENCES: 1. 1. lOve decomp rp cndsig ohio & ae el _Geoplysies 25, 649-658 (1960). me “Mike ero on he ot ir 4 bumann Ty of Ge mpl mated fr lel coe, Gepion tren Be TT 4 okie 68 ae Tin fa teseneusrn 5 Site nd Kea Hh eens detest ep fhe mn Sopie b9, cFSSD © CRet tems sn te ms: Gh Se Gro 1 Atha gr Tc Pas {nip 50 rte tn Pptabne tons eG oa cea ee come ye ES aber ncensens pare re eras osetia tee262 Apsondi A. cromagnete Sparc! Waves 12. 1. R Wal, Become etre om ly inom phate gp Se Tex Sot B18 1190195) [Noe tate gh ard ie of Ep (oh and (49) soa all fe ped wate. ec De ean (a) $c) aS eect ml sate ht on he hand is “dh Bel ens el he cde th hi 18, HLM, Neh Th fd apne plein Ie pene + cong phe, Cee 661-8195) 1 TER Mi Compe monte remeailiy fst paris, Prox Rai Eng. ssa 8s, 1, SIH. Mind The stomogetc mtd in Ming Gp Vo, 2-272 So. xploaton Grossi, Oxon, 887, [Noe Now aah npr) Shaul west The Tw ae a ae Ag) 3- ay Saat et Aue "ea e-Sih nan % 3-3 ts mo ovens BB Note on Bessel Functions Integrals ‘Often we need an integral representation for the fields of « dipole oF other source singolatity. Here we discus how a numberof such represent tions ean be derived ab ii ‘We begin with the function R~'exp(~ R), where in eylindsicalcoordi- nates R= (2+ 2°)" and y= a + if, where a and fare real. Now clearly (FR exp(— YR) =O, This geese that FH = LP save "savy aa, a where the egh-bad side i also a solution ofthe Helmholtz equation i Ae 7). Weal regard both p and = a greater than 0 Inte static Tite, 0), we then have afar non ar. @ In view of the Lipschitz identity) have the limit this limit But from (1) we also rayne. ° “Then, according to the FourierHesel transform theorem, a) pe LY. ferssanae “ where w= 2+ 75)M% Asa consequene, So [odesyanan, ° hich is our desired result. This integral relation was used originally by208 Depend 8. Nete on Base Funston negais ‘Arnold Sommerfeld in 1909. An excellent account of dhe theortcal basis nd application s found in his famous lecture note [I ‘Several other basic integral formulas ean be found dretly from Som= ‘merfelds integral. For example, consider the following operations: [lemsonane 1 [eB sparen == Lenigan|” 2 [ran deman pessoal” 3 [u0n® a Lee genet], o here apuin : > 0. This reduces to the required Form when 2-0, te, [ianae ‘Another useful esl is obttined if we integrate (6) with respect to = from = tooo, Then we find that [oZs0na- Lene ‘The ing fom for £0 isa noeworty c LO) Sit note fom icine i we diferent (6 wt pet wo = and Sense? Oy [Penonane 21+] REFERENCE 1. ALN Sumas oi Dien! Egan Phir. Aexdeic Pres, New Yok, | | | | Index a Aitorne eeoagnetc tod, 12-15 3a a6 niseoy “ite er allan, 6, 12 for electromaynei fils, 146 ‘lent bone mek 1 uly rein, 58 ‘Dare pobiem 31-33 ol nonin 36 Bare Rempnea Bir eure prose 28 Burd po 23 orl eet 20 © crete Eee eS ‘Come spherical parle 77 Gee denon. 15-16 Compe condi Compe ime (Comper mai Tit aig ‘hy reine 34 Smee aan, 3 Conte sure agstaaris (CMT, Debye poten 26 29,241 Dalene caret 2138 2225 Ei sat Erie cpt 1,16 6-19 Braud an to Fowrclode a, fer 1-2 teil mod, ‘eeevomapete EM) euplg, MO-16¢ ‘Exgucnil eos me 3-28 F forsale er moe 13 fp rod | rene fe FE. 18 6 enptrel mel 2 Pane oper oll foe mete 3 peace feel spe 5 forinuetve spt 22 sumptaie oms octagon fi 6 maul mplenr os, 1 Aten teat, 210-230 tata inpence 7 “ne fered medi 92-108 plc storm 6.18 22 (Gerd me rsa an 15-2815 Stratis 7 Tntaones tains 195 ‘reso ee 9, 1 twolape mou 318 18. apne plyonit42.57 eter bediycondiin 18,240 pte and map 18 opm pes, 22 ‘eu 1 fepune 122 pepe, 12 ue rns og. (26 M agate tiple Siw 10 Nest ec atid Wa model 69 itso posable mel 96 fr ope te ‘Meine ince planation 9, 97 ‘Nepinesenng 7 ° Omen | ‘Ontrpping vr dike. 25 P Pak charpy. 75 son ese fet PEL 5 emene sper 13 kc pe 3.15 ‘eg pe Te conucting medi? ‘eu eloaode ° Quist, 26 (taste erento, 12, 169 alan fo pte e229 baste feral cab, 3-55 foe a Ter epiner rion 45-54 ‘meaomet 2 ferdn 21 Tee apere poses 3-45 ‘arbi wap 38 207 (or wedge en, 20-38 eve a lee sonar 12 (eet are 7 s Sette enstons 248 ‘Spree ie EM ean, 22 ‘Sei cuenta Spi emer me ea Shean. 263 Sieroter ecm hye ota 2 Spi EM econ, 29 fue! Set pate, 8,259 Inde magnet cpl 36.58 nel ne phe stl mode 1S Shel pdt ered ck mde 6-61 ‘cor 3 sdearce mote S + ‘sis reponse 7-9 seme 29 [erapent! boos 29 {orale ae ners, 299 t Tan set mde in hvomapsee vans 26 1m idued yarn 50. in revi 10 ers roomed et, 1 ‘a plaabis media 70 {ore ie os 76 ‘Trani ne ‘eter 33“Tate see (TE) waves 35 Tenia pane ted 1-52 are eit ep 2 “Cowra ern funn, 2-29 ae suse tc, 202 ‘ew paeal, Wolpe oii ode 29-40 tine st ese 38