RELATIONSHIPS OF CERTAIN STRUCTURAL ELEMENTS T0 IRON ORE CONCENTRATION IN A PORTION OF THE CARY MINE HURLEY. WISCONSIN The“ for III: Dogma of M. 5. MICHIGAN STATE UNIVERSITY Ben Clifford Trethewey 1958 SUPPLEMENTRY ‘ATE R IL IN BACK OF BOOK LIBRARY i, pupil-frank. $05544" 5?»..riuw3alwwnffli _ . _ 2, I II‘ It. w! 2; a31r73r1v3 f? 3.”?11” gidcs;buxL ‘I ELflh HT3 16 IR“? 07% 0 .us. a r1(m IN 5 "(V ‘ITF ‘ ,‘TI .‘ f‘ 173'" H TT‘,‘ . v1 lix.‘ C a.:‘...) V.L.L'. V .:I..d: 71] '3" “#13" 'T ‘ T 1" \ §Ll't.‘J“'- , J. .- Ren Clifford Trgthcwey ments for the degrne of Easter of Jcience sis suinitt:d in Dfirtial fulfillment of the Ilivan State University. art Lansing Vichigan 195é Figures and tables. Acknowledgment. Introduction... Purpose........ AbSt'raCt. o o o o a o Stratigraphy... Structure...... Develooment 0? Definition and Dre. measurment G 0:31; ' Independert variables. dependent vnriables... Statistical orocecdure..... Interpretation of results.. Conclusions................ Sueqestions for further study. Bibliograp‘I’rl'VOOOOOOOIOOOOOOOOOO Figure 1. figure 2. FIGIHKQS All) FAETIJB Location of the Cary nine....... Geologic column................. Celluloid temolate.............. flranw showing linear regression. Results of conoutations......... Nata used in comoutations....... ii '\ F‘T"T,T II “ -;"T.T"T‘IYT~"‘ £L‘Jl.1'.0'7 .J;"J .'\J,- ::J ‘4 Q The writer wishes to thank Dr. J. Zion, Hr. J. Trow and Jr. W. Stone ouse for valuable suggestions and guidance. The advice given by Dr. 1.0. Eaten, of the Statistics Weoartment, is wreatly anoreciatcd. The writer also wishes to thank Ir. J. Rovce, Er. 3.3. Vennedy, ”r. J. SVarrcr ?nd the Engineering Staff of the Ironwood Oifice of ‘iekands Hatter and Company whose interest and coooeration wade this study possible. Were words do not sufficiently express 4y depreciation for tte long Hours of work Lois Trethewey so unselfishly donated. l more fitting expression of ny aooreciation will 1 I so torttcoming in the near future. 111 Irvmvnnfitxrr Location and Vistory The Jerv mine is located an tIG flotatic Iron flange on tare 1/2 :-‘:~-c 27; N 1/2 Ezrsc 213; L3}; :14? are :72 end :2: 1/72 :5 1/2 ... 'I‘ ',/ o w 1‘ “‘1 . 3 . 1 - —,- -.‘(-TC .723, .r. In!" 3- %.. .he mne mas been one-lifzte-v b" t::e e" 0' Minncsda . ~ . Michigan I | “‘ ~ . 5. ' \‘-- ." Wisconsin ”I” A'" ‘ \ I o Map Shaun"! Location 01‘ fl»: Cary Minb 4, Scale: Una. .-. 7320 fur WISCONSIN MICHIGAN CARyMM/g I money MONTREAL Ej IRONWooo Figure l 1 fennah Iron )onoanv sirmxéiits oneninq date it 1' ‘ickands-”nthor ” Conoanv is sctine enent for tie co uauy. Shinments have been made from the Cary every year since its onening exceot lO32. Total nroduction up to the year 1/56 was lh,665,639 tonS. q The Gonebic range extends from Gogebic Lake in uofiebic Countv, Wichigan, to mtkins Lake, Bavfield County, sisconsin, a distance of some f0 niles. The main nroducing area is the nortion of the ranee between iakefield, Wichigan and Nontreal, “isconsin, a distance of some 16 miles. Ore wns discovered in the Gonebic District in 13A? end the "irst shimment was node in 1§94. In recent veer, the Range has shioped an nverane of a 1/2 milJion tons annually. Total oroduction of the range up to 1950 was 273,939,940 tons. West ore shinmcnts have been made bv rail to tte loading facilities at ishland, 'isconsin, where the ore is trans- ferred to lake heats and shipped to the lower lake sorts. Purnose In the Search for new orebodies on the Genetic Range, the question ”What are the urinary factors controllinq the formation of an orebody?" invariably arises. It is univer- sally recognized that, almost without exceotion, orehodies are found in troughs formed b" the intersection of dikes and imnermeable horizons. Jhat is not agreed upon is the role of such variehlcs as the oresence of faults and the . 3 types of fault movements, irregularities of the footwall, differences in dikes and original porosity of the formation. So far it has been innossihle to differentiate between ore- carrying and barren dikes or to measure the norosity of the iron formation before the ore forming neriod. It may, how- ever, orove valuable to investigate the influence of faulting and the configuration of the footwall on ere localization. he turnose of this study is to devise a method of measurine these variables and to adant statistical oroceed- ures which prooerly evaluate the effect these variables may have on ere concentration. It should he stated that, due to the length of tle method used and lack of data in unpro- ductive areas, ttis studv will he limited to tTe west orebody of the Cary Nine. The writer hooes this naner will oresent a new anC fruitful annroach to the iron ore problem and will lead to further investigat ons of this tyoe. ’ .7..1=-\' “ "1 5 VI iLAI‘L) .L AirLr‘J J. i contour man, showing irregularities in the footwall surface, was constructed b" orojecting data, ohtained from cross-section mans of the Cary cine, onto a datum plane established sub-narallel to the dip of the formations. fault orientation and rate of change of relief or the footwall, per unit area, were measured by means of the number and orienta- tion of contour lines in unit arras. The distance of each unit area to the nearest dike was measured directly down din. Fault orientation (5}, rate of Chan e of relief (r), and nearness to dike (d) were considere* as some of the indenen- dent variaoles effectinn iron ore concentration. Average nerrent iron (l), averane cross-sectional area (Y') and increase in 7e (Y”) oer unit area were consiiered measures of iron ore concentration, t“e dennndent variable. A statistical design (aultiple regression) was used to investi- gate tre relations“ios between the independent and desendent variables. It was found that dikes and transverse faults are the mafior factors controlling iron ore concentration in tne “ortion of the Cary nine studied. A predicting equation T* and standard error of estinate were also comoutel. ~ese may be utilized to predict measures of ore concentration fron denendnnt variables. c 71”» = r-i c) I :4 The oldest rocks in tEe area are the Koewatin sreonSCOnes which lie innediatelv sontn of the iron range. Vost autnors are of the oninion that they are altered basaltic flows. younger granite intrudes the Keewatin series and can be seen in manv nines of the ranve. A long oeriod of erosion end 6 the llrct- can era . The Sundafi quartzite and Tad River dolo'ite were de- “ositcd on the ienenlainei Arcfiean surfacc durine t“a lorer Vuronian. These two formations are fornfi tocnther onlv in the eastern Dart of t‘o range. t is not known wfctVer their 1 absence in the central portions or tee ra*ge is due entirelY ['1 to erosion or, in nart, to non-denosition. Tn the tar? mine a t‘in conrlomerate lvins on the Archean rocks nay renresent ;1‘ T“ the inndar cnartzite. -we gaf Tivcr doloxito is nissing entirelv. Following this erosion neriod, the Talns formation was de"osited. ln tte Carv win: the Qalws corstitutes the foot- wall and is reorusented, from tottom to top, b“ green slates, thin bedded hrown cnartz slates, thick beddoi brown quartz 1 slates and a quartzite. fine deposition of the quartzite was interrunted eitter bk a short ocriod of erosion or a sudfien change fron elastic to chemical sedimentation. Generally, mining men of the Gogebic Rance consider tHis break to te an erosional feature but Allen and Carrott (1”73) favor the change tron elastic to chemical sedimentation theory. Since this stndy is concerned with this surface, CEOLDGIC COLUMN -- g 6065 81C 1R 0N RANGE chuawan Conglomcrl‘l'c, sandstone, quartzite, volcanic Flows. dikes T UPPCV ylfl' Uncon‘formi‘l’y Tylcr slalc Hvrom'an Iron earl-ouch Sla‘te Pabsf frflmmhl Upper Unconlornifj Anvil {crruglnous chcrf and slafe Ironwood Pence {c moinovs Slate Ironwood Uncomforuu'fy (n Norrie ferruginov: duff Lower Ina-“Veal Yale {anyways Sllfcs COM charts Mialch lflurmian Plyhaufll {erruynovs chcrf, 50.15”“! Sla‘l'c. Concrcfo'onory jasper, am! Send, chcrr Um ”formif‘y 0007125 1?, Palms Quar‘lz sla'l’e Lower Unconformif'y 334 River cherry dolomifc Hw om‘a n 50» Jay gum-12.51! ovrch'fiah Kce we fin UnConformify Granite Grcc 01 Sch 5:? Figure 2 both argunnnts nust he considered when interpreting data. The Ironwwod iron Formation was deposited following this erosional or sedimentary break, heginning with the “lvnouth and followed hr the Yale, Norrie, Pence and Anvil members. The “1v“outh and torrie members are oredoninently cherty horizons, while the Yale, ‘ence and Anvil members are pre- dominentlv slaty. Vost of the ore uroduced on the range is ‘nined fron these two chertv rembers. l fragmental hed, believed h~ sone to renresent an uncorforrity, is found at the top of €50 Norrie. Another fragmental horizon, na ed the ”abet, overlies the gnvil menher and is believed to reoresent the unconfornitv T l") tevween the middle and unner Juronian. The rahst has been included with the nooer Vuronian *Vler formation which con- sists of a great thicrness of nelitic and iron carbonate- bearing slates. Unconformahlv overlvinn the Tvler slates are Keweenawan sandstones, lava flows and conwlomerates. A stratigraphic column, used by the fickands Nether Comnany,is shown in Figure 2. STRUCTURE me Uuronian rocks of the Goqcbic Range form n honocline which dins northward an average of sixty-five degrees. Poldind is very minor on the nroductive parts of the range but increases markedly on the extreme east and west ends. The rocks of this ranse have been disnlaced by faulting and r21 1 four fault tyoes are recognized. L793? are, by name, transverse, oblique (Eureka tyne), Sunday Lake tyne and bedding faults.i Transverse faults are the youneest and some displace the overlying lower heweenawan. The die on these faults is near vertical and the strike is nerpendicular to the strike of the iron formation. T"iovenent along major fau ts of this tyne, found in the west near fiellen and in the east near Takefield, is such that the eastern and western portions have been disolaced northward with respect to the central area. Tovement alone minor faults of this tyne was in the same direction and generally, the east side moved north relative to tVe west side. Cblique faults nreceeded transverse faults and strike oarallel to the beds while dinaing at right antles to them. The most widely known fault of this tyne is the Tureka. Vovement along this fault was much like a scissors with the oivot located in the east central Dart of the district. East of the nivot, the upper block moved south relative to the lower block. The Sunday Lake fault preceeied oblique types. Since it is the only one of its kind recognized on the ranso, nothing more will be said other than its strike is anproxi- mately northwest and the din, vertical. The oldest known fault or the range is narallel to the bedding and serves as a ”secondary footwall” upon which, orebodies of the Norrie member are deposited. It is found in or near the Vale member ttrougUout the entire producing nortion of the range. In the Cary nine, tHe north block has moved ”00 feet west and LOO feet up dio. All other fault tvoes disnlace the bedding fault and it, in turn, displaces most dikes. A large sill is found in most of the mines to t‘e west of the Cary. Since it is not found in the Cary, it will receive no further mention. Intruding diabase dikes are connon in all of the mines. ?he intersection of ttc southeastward dinnin: dikes and the footwali forms eastward plunging troughs. Some westward plungins troughs are formed by the intersection of south- westward dinning dikes and the footwall. It is tqouefit that most dikes are ore-cheenawan but possibly of two different ages. tome are undorbtedly post-lower Keweenavan since they intrude the tran ridges to the north. LJEVVIX_K**JTT (3’ CFtfi Virtually all- ore‘oodies lie in t-“e troughs forncd by the intersection of dikes and some innervious Horizon. Iost neolosists arree that these troughs served as sites of enrich- ment and tonether with fracture zones, controlled the ninration of ore forming so utions. ?rom tfiis noin' on, there are as many tteories as ttere are neologists. The ver“ nature of the orininal iron formation is still being debated. Leith and Van fiise (lCll) Favored an iron carbonate-chert type of sediment deaosited in iairly deep 10 .‘ -.’- _ ,-:vwflc qr} ‘ M 6'? if". ‘57-’75?" ' ' "‘ D .' ". ’"u '0 riff-‘0 “.0 if, Iffta’eru J3. o (1 9+ Dropouts send lIOu lOrJaUlOu OJ l1._e1en f21<3:ies mav Have been deosited from the earn sea deoendinn on ‘t?1ce I?1and Ph values at the site of devosition. Thus, in atzreeas where the oxidizing notential of the water was high, t>x~igmarv Hematite would he deoosited. In deep water, where 7311 and ”h were lower, either tle carbonate or sulfide facits VVCDIJld he deoosited. there are many different theories regarding the nature c>j? he concentrating erocess. Leite and Van Kiss (1911) iTEFVored the tWeory that downward percolltinfi, oxidizing, fifleateoric waters leached out silica and oxidized the iron (:rzreonate to hematite. Gruner (1TB?) proposed the teeory tWiat.ascending hydrotFornal solutions Heated dos ending Yneteoric waters which leached silica and oxidized the iron (ZELrbonate. Tyler (17h?) suggested that tWe metamorphic 'Crroducts of the ctert, siderite and greenalite of the origi- 11511 iron formation were leached and oxidized by either *tVdrothermal or meteroric waters or both. It is not necessary to comnent further on the above nlentioned problems since the present study is concerned VVitn factors that nest geologists agree help control ore (zoncentration. Those factors are fracture zones and Ebroxinitv of dikes. ~~ .fi 17"“ 21 ' if - ‘ * ’ " 1 '1‘ *“. ,V ,‘1-‘13 [1.1 n ",I ‘J riL'1 , t a ,y 1 . , f . . .~ . )JJ.‘IP'..L.LlC.‘.‘ L’ J ,11.;u.4'.,1L.-i.'.111 U. z :11; 1"») Independent Varia ole k Before choosing and annlying a statistical nethod, it _— ”$5153 necessary to d ermine the measurable independent xraa:riables which could influence ore localiyation and to (i sflvise a method which would reoresent tie oxantit" of iron t>Isesent. Bo accom1lish t‘1ise ob i: ctivcs, it was first ricecessarv to define the indeoendcnt var'ables to conform to *331e Cweole ic set in? and szlflsc1wntly devise a method of m easurinrt them. The variable (r) was d fined as the rate of C? an“e O.L (alief on tTe footwall surface Der unit area. In this study, In El unit area is one hundred Feet square of the 50 twall and ‘t Exe unit of relief is five beet. ?or example, if the unit 51]?Qa of the footwall contains six fire-foot contour lines, TSPIe rare of chanse (r) will be (5). In :eol ogiCE‘ 1 terms this variazsle would represent either ‘tlie anount of disolacemcnt on a ?ault, the slope of a fold, (Disian irregularity of an erosional surface. Hotchkiss (1919) Ellid others state that foldinv on t}.e Gogebic Range is Jery 1 rlinor. It is unlikly that folds arge enough to effect 318 v-- (‘3' 4 Eiriable exist in the Cary line so this interpretation ”vs JAM that considered. It is, however, aencrally recognized that t7”1e footwall surface represents either a chants fron a elastic L. k . o .L. 4‘0 0 1| ' f43 a cheaical tyne sed1noncao1on or a 811gn; er051onal break ‘- ») ‘__‘ ._. J. Allen and L. T. ferret (1911) believe that this break Cices not mark a break in sedimentation of any fignificance. 1‘: ? & . l2 .’ ' ‘« -:"‘ ~ .--. "‘_.'. . fl “ M . A “ ”r" '3 - T" 4'1" ‘0 '“. " '~ 3‘ car‘. JOun anarler, ,.1 i n 4:-r L01 www-OI unmouny (D r _ t UrJ .J P . O '2“ 3.3 Q {E cnrx ‘tne Gorebic Ianie is o? the ooirior it is probeblv a ssuzz~face of slight erosion {personal communication). Erosional feaélftures could ef‘ect the (r) values and an effort should be rn.:>..-de to recognize them. Recognition of faults as sucl 1 is of great importance in s sstudy 0? this kind. lny movement greater than five feet ‘ is}. a direction nerpenflicular to the footwall surface will er£T7ect the (r) values. Fortunately, this was the RFCdSCSU Cii.rzmtion of.novonont of the two nain fiult typos effecting iirle footwall orebodies. its Lefiflinf Fault, wfiic” is rou:“ly 7?:1rallel to the footwall, dOCc no: displice “he footwall ant 5L! 1 .- - ° 4- "1‘ r- - ‘. 'Wais no inilucnse on log ;~e “neon +, (7.) adult ”id 07': ore <2c>nccntrari01 elonj tNe footwall if nnr, cnnnok Ba measured and will not be (3 iscussed nere. EFe mov:ment of tne Sunday IElke fault is Tfelieved "no be mainly vertical, but it is the (>211fi‘one of its kind known on the rants and does not involve ‘L‘:e Car? 'in?. 31:“ou33 t“? L ' V 1 1‘ . ~ , ‘ . - ~F‘.‘ c tranSJelse nfld OQIIFUC iuult {‘ ‘ — ° , r. .. w. , :. .:, _;__ h .r , 't t~--.‘.r':.“)es did Haws cowwonenos oi novemenu nerdllel to the loot- ‘hnall.surface, the mnin direction of movement was peroendicular ’-1-0 tie slams of tne FootWflll. The varishle (s) was introduce. to.Velo di°ferentintc \ " ' ' '1‘ “‘ ’ f'\ ‘L ‘ ,-. 4‘ '1‘; . vo- . 1‘ . ' '3 A ~' ‘ . ‘ hoween slese two Iault novelenos. L.lS ;nr1asle 15 “brine: $18 the average sloop—strike per unit area of the footwnll I.‘ :»1, Q'- q I ‘. D . -' ~ ‘\ r‘ . 5 P‘. I." V f: ’V ‘ '. Surjace, Lie ..(—:rtn ’slonp-Strlko- nay ’T‘X?)l€‘.7.nfx._ .33, u; 61.1319; Sin analozy between the footwall surface and a ”olird sedi- “Ientarr bee having the same configuration. fhe strike and a j‘ I! In... .4le ‘14.. 'JJJj t . . fut? 13 d i 73 if“? J00 JV." t“: :2: (42’; it '71:} r» " '5 T" " 1' "" "1 '1-1‘ “L1 4' f- . 1t~de a .2, it, o *3 s] MTG-F'Frii03 Haul 53]_c)vwe are to tee footvall surface. It is ole n 510 we striLes o; I") aulted areas of tre footwall will slow the t]?€1<3€ of a fault on the contoured surface if the displacement 11c>1¢uwl.to the footwall is greater than tre contour interval 0:5' five feet. dince the types of faulting so effecting the :Fc>:>twall are mutually oernendicular, the Slflfe-StriKCS will I‘c;iflect this orientation. is a result, slope—strikes help . r5 :1 _ o 0‘ ._ 1“ 4-- . ’ r33.i -erentiaoe neoueen oblique and tranSVerse faults. ine varieole (d) has been defined as the nearness of 51 ‘l'it area center on t‘e footwall surface, to a dike W1€w;sured directly down din. Tron the nreeeedin: discussion it is evident tVat to “Iexasure the variables (r) and (s), a contour map on the < Chatwall surface must te orenarod. Such a map Has been T3I“enared usin: mine naps end drill “ole data. This contour Tn€1fi~nns oreoarcd using methods described in a paner written E37“ “.J.C. Jonolly (1736) entitled “L Contour I.tnod of sevealine Jone Ore Structures". It was first necessary to establisn a datum nlnne unon VILEiCH footwall data might be nrcjected. In order to locate t31is nlane in space parallel to tfie footwall, traces of the T31JUH2YTTP drawn through the coordinates h, 5 and 10, -§ of Efilcn cross-section man of tFe mine. Weasurements were taken <11: :"i'f‘tr foot intervals, from the footnell surface to the lirnce of the datun olane. T“ese data were slotted alwng the k 0 I 4. - D ‘Arnces of tFe cross-sections (numbered from 0 to 23 on late 1) 14 «8J1fi contoured usind a five foot interval. It s*ou1d be r1c>1295 Here that the numbers along the riqht Band border of t;7w€a mrid are sea level elevations when the nlane is rotated t3<> a vertical oosition and t?ose alone tlc too of the Irid ELI?€3 cross-section n meers. Hr. .ennedv of Tickends ‘ T V .6-..1er Conoanv suzgested that tlis man be oriented in the 53‘1’10 V l 35 the mine mans. T*is necessitates viewing the saiiriECe from hence?“ t“e feotwall and derijnt in? the dis- TLEIhCGS ¢:om the s rface to the datum plane as negative \rzrlwes. AltFtufin t'is arrcicenent mev be awkward to tee liriinitiated, tEose familiar with the nine maps will find ‘t‘gis orientation nrefereble. In order to use this man to measure the indenendent \rezriahle (r), it was onl" necessary to count the numbe‘ of <2C>nfeur lines in eac'n of the blocks in the datum grid. m l 0 determine the variable (3) for eact of these blocks, t‘t‘e orientation 0? the contour lines was considered. Since t2>19 two gault tvoes nest common in this nine (oblique and 'tcransverse) make mutuallv nernencicular traces on LAB root- ‘Vnall, it was convenient to assume that contour lines falling VWithin certain arc segncnts renresented a certain fault t?fioe. The autfior realises tTis fss notion is not valid for Elllcases but is true only if the contour line daisity per ‘Thit area is 5154. such an assnnntlon 1s Sustified in t’i U) Q U) (D ‘J ) ’D f.) if; (D H o .3 L f’ (D r 0 » r J CD tical analysis 1hr intcrnctien or m 1‘.) :5 fl. H H- 0 ,3. 01 Ho cf- . .14 ‘1 ’1 H. 8 taken into account. 9: mil-.75.; Eufiuiuitm l1 (Ethnviflu L .. . .2“. 15 celfluloid tefinlabe (fix. 3) was useu to measure t} \ L" ‘ Ll “v1a;riable (s). A half‘ circle was divided int 0 six cmual 30 c3'~:*ree arc se,_m nts The dividihq lines, beginning at the ].<3:ft end of the diameter line and nrocee inq CTQChzwix ‘”“:Hre nimbered l, %3 4 3,2 and l wit“ 10 divisions in each 1?:iTT”e. T*us, slone-strikes (shown an 3133» 3) ‘4 faliinq with- i.r1 the 3-4 ranne were classified as oblieue ¢ault tvers and 13"08e falling within ‘He 1—2 '1 .~. ranfe 1'ore CM) Sid .Ie as trn Kr6=rFe Tenlt tydes glove-strikes faiiih: within the 2-3 “(wqee were not clas.i fied as a fault tyne. The grcat Eltivantare in tYis numericel clasei:Fice .tion 0. fawlt tyfics -'- ° . "A , 1 f o - . 1.8.t at it is amen ole to Stutl eticalana1331r 3. "’1 ig. 3 e3-]u]oi D. d- S9 :3 ’15 ~_: ‘ J I .d to measure slone-strikes. £ng _’2’ 1] mm. , . 1, .. lie Ceoen ent Variables n - ° 1. a. ‘- -'. . 7. -“"-vr ,«J‘ ' oev181ng a 1ot~od of estiqetinq tn; quantity u1 iron Dl"€? sent in '1 block one "‘nndrer? f‘ect "igh and bo"n'°.ed by a. 11f:5_t area was dif?icult hecquse of the lack of data. v#3.11ues o? the iron formation are ronfiilv oE'si 2% a] €1r1"3 sub-levels near 330 footwnll but :- a‘rle a'.-'3.‘.' ”row We worwott-xali. e mnrcen": iron \rf‘_‘w-... f) O '1‘ (I) H a ’3 ' 1) 2L TG‘WI‘Fsent on“: “The wortion 0:" 131- IT) \I iron formation close to Tii‘e ‘ootwall surfoce. Fit in tie orotodies themselv‘s: Elxrerare percent iron values are fair}? core ent but iron off 1?Etoidlv near t“e ovter Boundaries. “oceuse o7 tFiS, it had 13C) ‘ve sss=vu1' thei.’?7e ev~rmtvz13fim1r of‘sz oroBOAV'I1“““ t‘e -!—~‘ I ' " I I '3- 5361‘”311.'27s 1:"n1~sontivdfiwe 0’ 1"1 on“1rma<3rebor¢' . i t f.‘ ' 1‘ . ~, ’ -. y ..j " C" : ‘1 " «1,1 I r, . 1101 ,orlanion ont111r¢ _‘ ore u} FOJullY 3 :‘nsnrefi with a olanincter and volumes can be calculated from - zese mensureme.ts, out 1t 18 UU’PMLUuuule N st or those are ‘ -.- ~ .1 :1 ..-.“.'-' '° ‘1 - h- - «’ " 13 e onlv nortions 01 two 1 rndtlon w ic= FeVo HHOH eniicned ‘two ore erode. bbvionslv, the gxacb vo7unx of org in each 'VUfiit block Cannot be negsured. 1“ is boliovo§ the? EEO marfiin ~ C17 error in calculated volumes is 51311 but until wore data C o r 1"- .‘w _o ““1.“ ‘- H “.I“_ 1 A 5"1‘ fl ,3 ['1 3~S fivfllldule ~10 1r «rent arror C: not w 1:hctou 10 account 10r SO'IQ 0'” this error, 1.11111“; ‘::;-loc‘;lrn1izr'rtor. 1 e v: lucs so attained :‘rere slotted on the footwall datum nlane in their respective 4“ nositions. Thev renrcscnted cross-sectional areas of LNG 4. orsbodv in an east-west nlano neroerdicnlnr to t3w lootwall. .l 30 convert these to volunrs, it was only necessary to multioly 9 br 109 feet, t‘e north-south dimension. Wlocks ”its no measuratle or todirs were assumed to contain 50“ cubic feet .0 o. ore. Ell 0* these v lues were concourcd, usins a lCOO Cubic foot interval. from tnis man, tie depend nt variable (7') for eac“ unit area was calculated. 1, r l i orocoedure, much t1e same as that nsoc to firternine (Y), \fils ussz'to chfizernirx1 (7'). ‘Jhe iJriividlufiL sna]J.z1rtas ‘ witnin tte unit arra wrre me sures L...) O : L—J C7 0 ’- p. m H. 0‘) s 0 CL weisnted averase volume 0? ore was ca tre voliuunto? or: infiaoc FJ'), Howeveru if<1 statistical. conoutations were wide such simnlcr by dividing the writhtefl averafic volume or: values 5“ 1C0. In 3901031031 terms, L dividine by unis factor of 100 simply chances the volume measure to a measure of tEe average cross-sectional area 01 tte ore in a unit block. Both these measures roilect, eruallv well, the amount of ore present in a unit block, r‘1 1‘0 Iron index (3”) can be calculated and has meaning if it is asswmc” t‘et the iron formation, as d'nosited, had a uniform conwosition. For the nurnosos of this study, the iron formation was assumed to “eve an orisinal co"wosition o? 37.5' “e and blocks with no ore ware assunof to “ave on 19 Iron Index 0“ 12 cutie *eet of "e. ”ith the oriqinal comoosition of the formation given, then the increase in iron of the unit block will be: present percent minus original percent iron times volume of the inriched nortion. Admittedly, average oercent iron or the entire block and volume of tne enriched portion can only be approximated. Not in: more is desired or expected than that these factors renresent the geologica.l picutre fairlv accurately. T‘s actual calculations can be nn4 v '53.) :3 Cl. 94 a” ‘. Ci“ 11') l . .' .J :3 O L I (D denendabilitv of these estimates (8V). This met”od ntilizinw two variables, is called linear regression. In this tudv, a method called multiole rearession till he used. 7ultiole regression is very sinilar to linear recression with the excention tlat the effects of more than two variables can be studied simultaneously. Anotter eiventage Linea r Regrcss ion fi’jurc 4' 23 of usins nultinle r~“r'ssion is that more of the variables involved in the nrocess can be used in the nrcd'cting equa- tion. Theoretically, if all oertinent variables could be measured, the nredictin: equation world always give the exact value of the deoenient variable. Instead 0? two variables as was the case in linear refression, this study is concerned with four. Obviously, a *raoh of f\ur variates cannot be constructed as we are limited to t roe dimensional soece. it is not necessary to construct a graoh it the oredictinq equation, correlation coef'icients, and standard error 0“ estinate can be calculated. In this s udV, the orcdictine eouation will take the form of the 7our variate linear function y = a + bd + cr + es. The standard error of estimate as well as a variety of correla- tion coefficients can to calculate? in a manner similar to that used in linear rerression. Correlation coef“icicnts that reveal the degree of relations”io between any conbination of variables err verv innortant. for exenole, one could calculate the correlation 99*ween oercent iron and slots-strike w ilc tVe eTfect attributable to dikes is held constant. These wartial correlation coefficient then, can 570w what variahles are most inoortaxt in ore concentration. f“: owl ihle correla- tion coe ficient Rvdrs reveals the pronortion of tTC total 1 variation 05 (Y) tTai these indeocndsnt variables lave accovnted for. To test whether a correlation coeTTicient is sirnificant, the Paniliar null hvootFesis is se‘ u“. This shit s that there is no relationsh'i Pctween the variables. If the v¢lue for the coefficient exceeds that which is given in a table Found in ”nedecor (lQhO, p. Zflé), showing the 53 an 1% points for r and R, the hvpothesis is rejected and the correlation is said to be sisnificant. TFis means, sinilc, that if the coefficisnt exce ds the value siven 7"or the 11 level, the clances are better t‘&* 170 to 1 that this correlation is not a result 01 s crancc relationSWid. TVe Tollowint discussion will owtline, according to smasdard Worceedurs, how to calcul to these elevents. wore covalete discussion of the dir‘OdS used can be found in Croxtow and Trowdefi (1“37). Tle enuation of the Urodictins line can he ohtaincd bv: l) Solvins these three equations sinultaneously for b c and d. [in/'1. (Egghkjb +134, - MAE—”JG +545- {(42,652}: = id)’ _ (1431! Y} 134'“?———-"L(“’]L+[f"— [$17G +Er5~(""”/f5)]=£ky- Liz-"’5‘ ’ [24: ~%flyg+pr,-flgl c ,[¢,._é%_)z],_zsy , (arty) obta‘ ‘ s 4 d 2) duhstituting the values .2 ’ -- in the cQuTtion a = Y - d - cr ' 93' 3) Jubstitutinfl the values 0) win" £01 a, 0. C 3‘s d in the nredictina efiufitinn v I a + bd + Gr + s3. \4 ' .' * C I‘ *‘ ‘ :‘I ,. “‘- 0. A 0‘ '~ v v I n . ‘ »§ 7 T“ * ‘ x - '3 I. l‘m: standwrvi«arror o- {InhLmlte vhf so inifitz.nr‘ :» orlul.: a. .. ( “A“— “W Q ":5 534,, = Fs-tssgcfiz s—sa— r as .r. . 1|..M‘mmslirtufl4JJufllfldflflhf Inlj- . Nit. fll‘ 25 toe wzltiole correlation coefficient is fiiven bv the Fornula: Aggy- Mgfli-r 65:;- Wis e [“7’ WJ‘ fiyO/VS Z W" t y; ~ (‘4.- ‘artial correlation coefficients may be obtained by tte formulae: ._ t n _._ 4’7): 1944’ _ n- (mum) y ‘ l — 3 bX'K " —L __._..__— V[x-£R&j[2y"‘_mn J W" WK \//"’>3'K "othing would te,gfiined by showing tiese computations ‘ o 1&1" l a") 64- 1‘1 c . rfi 4'1' r) n" . ,‘ «an 0" '~ . 7‘0 "1!’ 1/ g‘ DE?!) L) ,r ptc V). I-‘ '1 ,‘J rx'.’..i( ' Int/t). Ch.) '.I [;_»(. , iv “(2.“? I \‘3 f anv sta dard advanced statistics textbook. The data use? in .l,‘| use calculations are s“own on “latcs 5 and 7 in the pocket. Tor conveience, comout d sums of squares and other data is Show“ in Table l. The comnutnd standard error of estimate, 1“ correlation conflicients, and the predicting ocuation for 1" '1 (x 1. 0‘ 1." ' 7, ' "1 V sac. deoeidant VcIldulQ 18 snown in iaole 2. ’_T-" ~ 'T'Y‘JT) . \r‘1 :1 "‘ w " -‘ ‘_“I H» H] g .‘1’ IL. 7.,511,‘ AL 127‘; L'.!_\ W k) .‘ Ibilzjr' ' is) The correlation coefficients listed in Table 2 show, as mirht he exnected, that dikes and transverse faults are the nain ”actors controlling ore concentration. The rate of chanre of relief has no discernahle cifect on either tenor or awount or ore. A discussion 0? each coefficient for each derendent variable would simolv be rsoutitious. f*c Followina discussion 0? the results obtained ior average owrcent iron {7) will also annlv to the other denendent variables, in; rflflunf .ledlflfl.‘ .._.... “'1 4N? 31. .I a. ”redictina 5v ryd ryr rys “ydrs ”redictins iv' ryr rys iT‘ivdrs 3redictine 3v” ryd TV 1" rys Rvdrs 26 T 7L} W“ I JVITfi ‘I‘a bl e 1 ”ercent Iron (T) ecuation: y = 52.361 - 3.18d + 0.03hr + 1.8558 : 3y 3 6.072 = -O.58h:::::: rVd.I‘ c -O.59[+::::,‘: = +0.176** ryd.s = —O.642** 3 +0010()3:7 rys.d = +0.363** - 0.616** ryr.d +Oofll¥5 Volune of Iron Index equation: y' = 18.131 - h.156d + 0.2l7r + 5.1198 : Ev' - 22.002 = -O.310** ryd.r . -O.291#* - +0.123 rYd.S a -C.36l** = +0.178* rys d = +0.26?** . 0,359:* ryr.d - +0.052 Iron Index caustion: y” 3 296.?73 - O3.705d + 36.931r 500398 - -O.295** -o,273** ryd.r I +0.132* ' ~0o352** ryd.s I +0.15h* rys.d a +0.231** - 0.31.32»:- ./ ryr.d +0. 006 l1 1., I‘I ‘1‘ 0.3.). x ixaiZ a vn .j‘a . U o u o. O r if .for Y' Tor Y" 3(ds) 27 Table 2 DATA 8.”qu ififl+p3?0 1,692.16h 700.922 -o.595 E( Volwne Index 5,u3e.o 254,258.84 129,u*3.127 -2,164.11 -9 one 006 *"///O. '73 ,7011' o 63 9 (5) 1,079 570.5 6,030 1,525.03 1,126.592 151.738 dr) -215.50h 3(rs) 6?.107 % Iron Index 11,953ag 112:h76'0 523,477.73 142.3h1,514-0 19,603,662 83,952,397.7 827.3 9 5.65 1,487.231 793.?57 h1,794.2h1‘ 17,935.965 n - 237 verv sianificani (0.015) nultinle correlation coef“icient indicates that over hal' o? tfie varittion of the amount of ore is accounted for bv the variables studied. 3bcre is then, a definite relationsnio between the inde- pendent and dependent variables. In neoloeical terms, this means {yam about two-thirds Of tDe variation in DCFCQKt iron n‘ 5—, (:3 . 3 is exalained bv dikes and faulti Turnins now to tn; nartial correlation cochi i nts, we can see what tie relationsnibs the infienendcnt variaVles ('1‘ bear to the dencndent variables. Lye re ationship between dikes and tenor is very significant. Even when slone-strikes and rate of relief change are onrnittbd to vary, tWis is true. I7 slows-strikes are teld constant (at their avaraxe value) the relations in between dikes and tenor (rvd.s = .642) is even more nronounced. TVis may be caused bv the inter- action of dikes and slope-strikes in some areas; that is, un“avorahle slone-strikes reduce the tenor in some areas. is first glance, the Bietlv significant value of (rvr = +0.176) for the correlation betwern ra.e of relief 1 nanse and oerccnt iron would indicate that this variable is 0 1 a definite influence on ore concentration. If however, tne dike e7 ect is held constant, we see that = +0.0h5. r yr.d TFis sayS, in effect, that what was evidently 8 rClHtiOHShiP between rate 0? relief chance and tenor is really due to the mask?ne ef’ects of dikes. F"! ine relationshin Between slope-Strikts and percent iron is an imnortant one. Even if tWe e‘fect of dikes 18 cold 29 \ constant, (rvs.d = +0.363) we see tnat the correlation is "'1‘ ' ,. l, l. [“18 flFTflS twat as tre (S) VTlUPS K/ still ver'r significant. increase, tenor increses. Renemherina that transverse faults give *ivh (s) values, this means that they are a major influence on the denosition of ore. The nredicting ecuation for nereent iron, enables one to predict with some degree of accuracy, the actual nercent iron in anv unit area. For ex nnle, if tte values for a white area are d=l, r=2 and s=3 the value of (v) would be ? 5h.9 ~ = 6.07f. Ttis is to say t at, most of the time, this estimated value of (y) will be within (plus or minus) one standard error of estinate of the actual value of (Y) for that unit area. for the above exannle, it wenld mean that tte actual value of (Y) will be between £9.74: and 60.F¢” most of tte time. It is unfortunate that the standard error Las suck a larne value. This, in part, mav be due to the assumntion of avera e nercent iron values for the blocks in which there is no data available. It is interesting to surculate what tte error would be if the real assay valnes for each block were known. Although the nredictina equations do not fulfill their function With the degree of accuracy desired, the basic nrincinles are sound. This is evidenced b“ the fact that the statistical analvsis shows what mining men alrendv knew; that ore concentration is controlled bv dikes and transverse “anlts. Because the method demonstrated that rate of relief 'J: '5! 30 chanse does not effect ore concentration, it served a useful ouriose in a negative way. The ecuations can be usefull iF the indeaendent variables have ontimum values. This, at least, nermits us to say that in the unit area under consideration, there is an excellent chance of "ore grade" iron formation. CCNCLUJIC “13‘ 1 sis of t‘is study, that tie D) It has been found, on the b e?Tect of variables involved in the process of ore concen- tration can be evaluated. is was previouslv known, dikes are doninant in controlling ore concentration. Bikes and transverse faults account for two thirds of the variation in the measure of average nercent iron used in this study. Transverse Faults, arohablv bv directing the migration of solutions, exert a major influence on ore concentration. The Fact that the oost-lower Keweenawan transverse taults influenced ore concentration mav innly that all enrichment was post-lower Keweenwan. There is no discernible relationship between rate of relief chance on the footwall surface and any measure of ore concentration used in this study. It has also been found that it is possible to fire ict From the variables, witVin_a known Narnin of error, some measure of the amount of ore. Wore accurate oredictions can be made when more accurate measures 0? enrichment in areas void of data can he collected. w .mw-u-rnIv-mx; "\2'» .‘IY‘II’W'T'HT' r‘r‘IUDY 'D‘th'fwle. \..’LJ “(a L .l IlLll—xmut 1.. ' In the Cary mine, as well as in other mines on the range, the bedding fault often serves as an innermeable horizon uoon which orebodies may lie. These are called ”hanging wall" orebodies. Since both transverse faults and dikes intersect this horizon in the same manner as they intersect the footwall, it would be interesting to see how ore development is related to these variables on the "secondary footwall". If anv measurable post-transverse fault movement has occured along the bedding Fault, the positions of favor- able transverse fault-dike intersections on the "hanging wall" could be calculated fron these intersections on the footwall. the methods used in this study could be aaolied to vein- like denosits. Conollv (1952) contoured the configuration of an iustralian sold vein and determined the relationshins between gold values and vein contiquration by visual insoec- tion. If the configuration of the surface were broken down into slone-strike and rate of relief change connonents, some innortant relationshins may have been discovered. A study of this type should he made using trace elements as indenendent variables. Predictions of ore tonnascs within a relatively small margin of error might be possible. This aeolies to all mineral deposits and may even have an appli- cation in the oil industry. In anv mineral deoosit where the relationship to structure is not annarent, a study of this kind may he very valuable. . \ pit .N tug-\hlf‘lulig .. n BIBIICQRIVHY Conolly, G. J. C. (1936) A “Contour Ifethod of “ vealing So.1e fire Structures, Econo1ic Geology, Vol. 31, Io. 3, pp.“?39;27I. Croxton, F. E. and D. J. Cowden (1939) Annlied General Statistics, ”rentice-Hall, Inc., NEW YorE. Nixon, J. J. and F. J. Jaseey (1951) Introduction to Itatistical Analysis, Me :raw-GiII Bock Goneanv, Inc., Tew York, pp. I33-172. Gruner, J. J. (1937) Hydrothermal Leaching of Iron Cre, Economic Geology, V01. 32, No. 2, pp;—I?I-IBO. ”otchkiss, T. O. (1919) Geologv of the Gogebic Rzange and. Its Relation to 'ecent inine LeveIopment, En} Ineering and 1n1n; Tournal, JoI. 10", pp. hEI-KSB, 501—5(:7 537- 541, 577 5 9 M mes, H. I. (1954) Sedimentarv_facies of Iron 1ornation, Economic Geolnny, 001. Kg, I». 3: pp. 230-793. Lake Superior Iron Ore Issociatiow Tie (1J52)0£9k0 superior Iron Ores, 2nd ed., illia1 1eather Connany, Cleveland, pp. 33-53 0 Snedecor, G. I. (lGhO) gtatistical Methods, Iowa State College Press, lmes. Tyler, 3. A. (1949) ;‘he Uevelooment of Soft Ir>n (re Fron "etamorleosed Iron 1ormation Geol. Soc. finer. Bull., 1:01.6(T150. 7, "W312!“ Van Tise, C. Q. and C. K. Ieit 1 (101]) The Geolog: of t"e Inks Quogrior_2egion, U. o. G. 3. Ien. 52, pp. 275-250. W', mini.» a I _. , - . H12 zv . . . ‘1‘,— Demco-293 Date Due VOCKeX rub.) Z 7 fl/é‘fé’f : . . . . . .2 , . . . . r. . .. . . .. . . I .1. ivynlfibvgt 0.31.0346.» 61...... 11. 0 :0. . ._. . w , o . . . - . ~ ~ . s (x .. 2.- r 1. .0 . a , a 4 . I ._ . - W. A . .L . P. . . a» O I. 2 3 4 5 6 7 8 9 O I. 2 3 4 5 6 7 8 I I l l ‘ I | 1| 1 81 89 86 67 60 OO 01 O. I. O. O. I. 0. O. 421 811 8 1 312 4 3 .413 5 3 2 rds rds rds rds rds rds rds s 67 68 67 82 84 48 O 321 311 8 l 712 3 l 2.3 173 rds rds rds rds rds rds rds 5 5 5 l _ MI 30 3O 4O 59 52 53 5O 5O 79 71 09 7.9 7...” 7.7” 7.00. 7.2 8 .1.) 817.2 3 l 3 1 3.4.1 331 .421 41.11 5 l 613 9 3 6 2 .452 2.43 332 321 212 s O rds rds rds rds rds rdB rds rds rdB rde rds rds rds rds rds rdS rds _ 5 5 l 5 5 5 5 5 5 . .96 92 OO 10 5O 5O 58 55 55 .44 43 25 29 22 24 21 23 _ o. o. o. I o 00 o. co o. O. 0. O o o. .0 to 00 O . no “.112 8 2 7 3 7 3 7.43 .432 522 612 8 1 312 3 1 262 653 5.43 533 723 613 _ . _rds rds rds rds rds rds rds rds rds rds rds rds rds rds rds rds rds 9 4 2 I _ 5 5 .. 60 63 83 .48 _ .49 .45 45 .48 .45 00 O9 9 O O. 9 8. 4 L2 8 2 7 3 753 ,742 733 623 812 m 2 411 7 3 60.3 7?4 No.4 m?3 MT3 NIB 8 rds rds rds _rds rds rds rds rds rds rds rds rds rds rde rds rds rds 5 .1. l 43 .40 69 41 4O 48 .46 .41 .48 89 O8 97 99 90 O 16 10 2....qu 7 3 7 l 151 24.2 331 622 513 8 1 6 2 5 3 623 .212 3 3 4 4 513 7 3 _ 7 ~rds rds rds rds rds rds rds rds rds rds rds rds Mrds rds rd.) rds rds I 1+!!! H . S . 1 1 , 5 5 5 5 00 00/ O7 _ 30 35 «)1 30/ Q/Q/ 38 58. 1/0 1 172 _ 10 10 1/0 10 IO EK W T _..._11 2 1 2 1 .253 342 233 721 213 3 3 7