Somm-o Qjj^L IMSRPVIII 622.209773 IL6msr no. 8 Application of Surface Geophysics to Detection and Mapping of Mine Subsidence Fractures in Drift and Bedrock P.J. Carpenter, C.J. Booth, and M.A. Johnston Northern Illinois University S*< Mine Subsidence Research Program Illinois Cooperating agencies ILLINOIS STATE GEOLOGICAL SURVEY Department of Energy and Natural Resources BUREAU OF MINES United States Department ofthe Interior LIBRARY. ILLINOISSTATEGEOLOGICALSURVEY 3 3051 00006 0586 Application of Surface Geophysics to Detection and Mapping of Mine Subsidence Fractures in Drift and Bedrock PJ. Carpenter, C.J. Booth, and M.A. Johnston Department of Geology Northern Illinois University DeKalb, Illinois 601 15 xA $oN Mine Subsidence Research Program Illinois ILLINOIS STATE GEOLOGICAL SURVEY Jonathan H. Goodwin, Acting Chief Natural Resources Building 615 East Peabody Drive Champaign, Illinois 61820-6964 Digitized by the Internet Archive in 2012 with funding from University of Illinois Urbana-Champaign http://archive.org/details/applicationofsur08carp 8 1 CONTENTS ABSTRACT 1 OBJECTIVES AND SCOPE 1 SALINE COUNTY SITE 1 Mining History and Geologic Setting 1 Hydrogeology 1 Subsidence Fractures and Geotechnical Measurements 2 PREVIOUS GEOPHYSICAL STUDIES 3 DATA COLLECTION AND RESULTS 3 Electrical Methods 3 Resistivitysoundings 3 Effectoffracturing on resistivity 3 Resistivity pseudosections 6 Azimuthal resistivity surveys 9 Seismic Methods 10 Seismic refraction 10 Effects ofsubsidence on P-wave and waterlevel 12 Seismic reflection surveys 15 JEFFERSON COUNTY GEOPHYSICAL SURVEYS 16 Resistivity Soundings 17 Effect offracturing on resistivity 17 Seismic Refraction Surveys 18 DISCUSSION OF RESULTS 1 ACKNOWLEDGMENTS 19 REFERENCES 20 FIGURES 1 Saline Countystudy site 2 2 Schematic illustration ofthe resistivity, seismic refraction, and common-offset seismic reflection methods 4 3 Saline County electrical survey lines 5 4 Vertical electrical sounding north ofpanel 1 and its layered model interpretation 5 5 Repeated soundingsat Monument 98 illustrating changes in apparent resistivity before, during, and aftersubsidence ofpanel 2 6 6 (a) Sounding curves made 490fteast ofthe monument line along the centerlineof panel 2. (b) Five-layermodels from inversion ofsounding curves shown in figure 6a 7 7 Geoelectrical modelsforpre- and postsubsidence soundings made overthe northern tension zone and the north barrierpillarofpanel 2 8 8 Apparent resistivity pseudosection 490ft east ofthe monument line 9 9 Azimuthal resistivity arrayconsisting ofpermanent electrodes 10 1 Azimuthal resistivity plots atarray B foran electrode spacing of5 ft 11 1 Azimuthal resistivity plots at array B foran electrode spacing of 15 ft 12 12 Seismic survey lines atthe Saline Countysite 13 13 Typical refraction arrival time data and layered model 13 14 Centerlinevelocity models from pre-, syn-, and postsubsidence refraction surveys 14 9 89 15 Velocity models from refraction surveysoverthe northern tension zone ofpanel 2 14 16 Common-offsetseismic reflection section across panels 1 (post-subsidence) and 2 (presubsidence) 15 17 Jefferson Countystudysite showing panel outlines, geophysical lines, wells, and borings 16 18 Typicalvertical electrical sounding fromtheJefferson County sitewith a layered model interpretation based on well logsand borings 17 1 Layered resistivity models from soundings made overthecenterand edge of panel 3 in Jefferson County 1 20 Layered models from seismic refraction surveysoverpanels 3and 4 1 PrintedbyAuthorityoftheStateofIllinois/1995/700 printedwithsoybeaninkonrecycledpaper I ABSTRACT Electrical resistivityand seismic surveyswere used to map fractures and monitorwaterlevels oversubsiding longwall coal mine panels in Saline andJefferson counties, Illinois. These surveys showed that resistivitysoundings can be inverted to obtain minimum estimates forfracture pene- tration depth, and common-offset seismic reflection profiles can be used to identifyfractured bed- rock horizons. Inversion of resistivitysoundings made during subsidence ofpanel 2 in Saline County revealed a shallow high-resistivity layercorresponding to fractured unsaturated drift. Resistivitysoundings, however, could not detect the presence ofthese fractures below 5 ft, nor could soundings map the lateral extentofdriftfracturesor provide information on bedrockfractur- ing. Azimuthal resistivityvariations were consistentwith predicted drift fracture directions over panel 2, and repeated azimuthal surveys indicate shallowdriftfractures along the centerline had closedwithin 8 months ofsubsidence. Interpretation ofseismic refraction surveyswas hampered bythe developmentofstrong lateral velocityvariations in fractured materials; these surveys, how- ever, revealed no widespread water table fluctuations during subsidence. Reductions in both resistivityand P-wavevelocitywere measured attimes overstatic tension zones along the mar- gins ofpanel 2 in Saline Countyand panel 3 in Jefferson County. These measurements, ifcon- firmed, may indicate long-term fracturing and downwardwaterseepage along tension zones. Common-offsetseismic reflection surveys identified disrupted reflections belowthe bedrocksur- face overthe southern margin of panel 1 in Saline County 8 months aftersubsidence. A thin lime- stone ata depth of 125 ft exhibited enhanced postsubsidence permeabilityand may be the source ofthese reflections. Fractures in the subcropping bedrocksurface (at a depth of85 ft) were notdetected. OBJECTIVES AND SCOPE Geophysical techniques were used to monitorthe developmentofsubsidence-related fractures and corresponding watertable changes overlongwall coal mines in Saline andJefferson coun- ties, Illinois. Geophysical techniquesofferpotential advantages overwells and borings because physical property changes overwide areas and volumes can be measured, as opposed to meas- urements atjustone point. Electrical resistivityand seismic surveys were made before, during, and after mining to test the usefulnessof these techniques in mapping the lateral extent and depth ofsubsidence fractures and in monitoring watertable changes. The geophysical methods tested includedvertical electrical resistivitysoundings, resistivity profiles, azimuthal resistivity sur- veys, and seismic refraction and reflection profiles. Postsubsidence resistivitysoundings and seis- mic refraction profiles were also made overportions of three panels in Jefferson County, Illinois, to testthe applicabilityofthese methods at anothersite. SALINE COUNTY SITE Mining History and Geologic Setting The study site in Saline County is a narrow strip several hundred feetwide crossing longwall mine panels 1 and 2 in northwestern Saline County, Illinois (fig. 1). Approximately 6ftofthe Herrin No. 6 coal was mined from a depth of370 ft. Panels 1 and 2 were 668 and 618 ftwide, respectively. They were separated by 132 ft, and both were approximately2 miles long. Panel 1 undermined the study area in December 1989, and Panel 2 in September 1990. The panels advanced at approximately 55 ft/day, and the maximum surface subsidence along the panel centeriineswas 4.5ft(Kelleheretal.,1991). Panels 1 and 2 are overlain by85 ftof Pleistocene unconsolidated deposits and a 285 ft thick sequence ofgentlydipping Pennsylvanian shales, siltstones, thin coals, and limestones. The unconsolidated overburden is composed of loess, Wisconsinan glaciolacustrine deposits and llli- noian drift (Frye et al., 1972; Fehrenbacheret al., 1984). A detailed analysis of the loess and uppersoil layers (Seils et al., 1992) identified abrupt density increases at0.2 m and 1.5 m and a majorshearstrength increase at0.5 m. Well logs and borings atthe center ofpanel 1 indicatethe lllinoian drift atthe base ofthe unconsolidated deposits is mostly sand and gravel, whereas the overlying glaciolacustrine deposits and loess are primarilyclayand silt, with occasional sand and gravel lenses (Van Roosendaal etal., 1992). Hydrogeology Fourdeep drift piezometers, six bedrock piezometers, and one pump well were installed overand adjacentto panel 1; no piezometers orwellswere installed overpanel 2. The Trivoli sandstone, at In Panel 1 DA»OI 600FT i -SurveyMonuments XI Panel2 InstrumentKey: DriftPiezometer MiningDirection BedrockPiezomter SurveyGrid PumpWell TDRCable MPBX Advancing ^Arcuate Cracks f-StaticTensileZone Longwall^ Face 'DynamicTensileZone Grid Location In Figure 1 SalineCountystudysite (afterISGS, 1991). a depth of 175to200ft, isthe principal aquiferin the studyarea. It is medium grained and argil- laceous, and itexhibits an average hydraulicconductivity of less than 10"6 cm/s when unfractured (Booth and Spande, 1992). During and aftersubsidence, piezometric levels in the Trivoli declined sharplyand showed littleorno recovery, although postsubsidence packertestsshowed a slight hydraulicconductivity increaseto about 10 cm/s. Piezometric levels in the basal drift also showed rapid declines during subsidence, butthese lev- els recovered completely within one year. Waterlevels inthe shallowdriftwere not routinely moni- tored overthe panels. Short-term local waterlevel fluctuations, however, wereobserved in the drift in responseto subsidenceoverpanel 1, and shallowwaterlevels gradually recovered to pre- subsidence levels (Darmody, 1990). Waterlevels measured in domesticwells aroundthe panels (Booth and Spande, 1991) and piezometers screened in upperlevelsofthe drift overpanel 1 (R.G. Darmody, personal communication, 1992) indicate thatmostofthe yearthe watertable lies within 20ftofthe surface. Subsidence Fractures and Geotechnical Measurements Surface displacements, strains, subsidence rate, and subsurface movementswere monitored dur- ing the subsidence ofpanels 1 and 2 (Kelleheretal., 1991). Zonesof maximum tension (static tensilezone on fig. 1) and compression were located 44 and 144 ft inside the panel margins, re- spectively (Van Roosendaal etal., 1992). Fractures were aligned parallel to the panel margins overthe statictension zones and parallel to the mineface nearthe panel centerlines (fig. 1). Arcu- ate fractures occurred elsewhere (Van Roosendaal et al., 1992). Fractures aligned with the mine face near panel centersclosed within several weeks as these areas passed from dynamic tension intocompression. Eight months aftersubsidence, tension zonefracturesoverpanel 1 remained open to a depth ofat least 5 ft (Seils etal., 1992). Van Roosendaal et al. (1992) used surface strain measurements with a neutral-axis bending model (Kratzsch, 1983) to estimate a maximum penetration depth of27to 31 ftforfracturesvisible atthe surface in the static tension zones of panel 2. PREVIOUS GEOPHYSICAL STUDIES Few studies have applied surface geophysical methods to mine subsidence, particularly in Illinois. Burdicket al. (1986) successfully used resistivityto locate shallow abandoned mine voids in the Illinois Basin, and Henson et al. (1989) used high-resolution seismic reflection techniques to map sedimentaryfades thatcould lead to roof instability during mining in the Illinois Basin. Most other longwall mine geophysical studies have been in the Appalachian coal fields. Wilson et al. (1988) used seismic methodsto identifyzonesofabnormal P-wave absorption overthe edges oflong- wall mines. He and Wilson (1989) observed velocity decreases over mined-out areas using common-offsetseismic reflection profiling, and Rudenko etal. (1989) mapped the position ofa longwall mine face during mining using changes in P- and S-wave velocity from seismic refraction surveys. They noted a slight P-wave increase ahead ofthe longwall face and majordecreases in both P- and S-wavevelocitybehind it; theyattributed the decreased velocityto enhanced fractur- ing in the dynamictension zone directly behindthe mineface. None ofthese studies, however, have used geophysical techniques to identify subsidence-induced changes in unconsolidated materials, ortocharacterize hydrogeologic changes accompanying subsidence. DATA COLLECTION AND RESULTS Electrical Methods Resistivity soundings An overviewofthe electrical resistivity methods used in this studycan be found in Telford etal. (1976). The basic electrode configuration used was the Wennerarray, shown in Figure 2a. Twenty-five Wenner array vertical electrical soundings were made over selected portions ofpanels 1 and2 before, during, and after subsidence (fig. 3). Seasonal appar- ent resistivity variations were generally less than 2 ohm-m for spacings greater than 10 ft. In response to rainfall and seasonal variations, however, apparent resistivityvalues changed as much as20 ohm-m at a 1-ft spacing and as much as 3 ohm-m ata 10-ft spacing. Soundings were invertedwith a least-squares computerinversion procedure (Interpex, 1988) to yield multilayer geoelectrical models. Most resistivitysoundings overpanels 1 and 2 were mod- eled with fourgeoelectrical layers (fig. 4), although in several cases the existence ofa thin, con- ductive uppersoil layer required five-layer models. The interpretation ofthese layers is based on a soil boring made at the centerof panel 1 and a roadcut approximately 1/2 mile northwest ofthe panels (Frye et al., 1972). The uppermost layerofthefour-layermodel represents moderate-to- high resistivity topsoil and loess (30 to 80 ohm-m, 1 to 3 ftthick). This layeroverlies 15 to 20 ft of relativelyconductive clayey Wisconsinan lake deposits (10 to 30 ohm-m). Water levelsfrom pie- zometerand seismic refraction surveys indicate this unit is probably continuously saturated below a depth of 15 to 20 ft. The lowermost 20 to 40 ft ofdrift is saturated gravellytills and outwash (30 to 40 ohm-m). The underlying Pennsylvanian shale bedrock has resistivities from 45 to 70 ohm-m thatwere confirmed through comparison with electrical logs from the area. In many soundings (such as fig. 4), the lowersaturated driftcannot be distinguished from bedrock. The equivalence range (or uncertainty) foreach layer's thickness and resistivitywas computed by allowing model parameters to vary such thatthe total root-mean-square (RMS) error of the model fit did not exceed the best RMS by20%. Geoelectrical layers 1 to 3 were reasonablywell con- strained for mostsoundings, and resistivity errors averaged about 10% and depth errors aver- aged 35%. Thethickness of the lowest layerandthe resistivity ofthe bottom (usually bedrock) unitwere poorlyconstrained, with uncertainties commonly as large as 100%. Thus the resistivity soundings were not used to identify fracturing orothersubsidence-induced changes in the bed- rock. Soundings were also not used to estimate waterlevels because seismic refraction surveys generally provided better-constrained watertable depths. Effectoffracturingonresistivity Repeated soundingswere made during subsidence overthe centerline of panel 2 at Monument98 and at a point490 ft eastof Monument 98. These sound- ings exhibited major resistivity increases at shortspacings during passage ofthe mine face (figs. 5 and 6a). Later soundings atthese locations showed no significant differencesfrom presubsi- dence soundings. These temporary resistivity increases may reflectshallow air-filledfractures CurrentMeter x Linesof CurrentFlow j? P-wave Source Geophone si X^^^ n n n n n rT n _ n O GroundSurface O. reflector Figure2 Schematicillustrationofthe (a) resistivity, (b) seismic refraction, and (c) common- offsetseismic reflectionmethods.