Geotechnical Site CharacterizationPROCEEDINGS OF THE FIRST INTERNATIONAL CONFERENCE ON SITE CHARACTERIZATION — ISC'98/ ATLANTA/GEORGIA/ USA/ 19-22 APRIL 1998 Geotechnical Site Characterization Edited by Peter K. Robertson Department of Civil and Environmental Engineering, University of Alberta, Edmonton, Canada Paul W. Mayne School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, USA VOLUME | A.A.BALKEMA/ROTTERDAM/BROOKFIELD/ 1998The texts of the various papers in this volume were set individually by typists under the supervision of each of the ‘authors concerned. ‘Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is ‘granted by A.A. Balkema, Rotterdam, provided thatthe base fee of USS 1.50 per copy, plus USS 0.10 per page is paid diseetly to Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923, USA. For those organizations that have been granted a photocopy license by CCC, a separate system of payment has been suranged. ‘The fee codle for users of the Transnetional Reporting Service is: 90 5410939 4/98 USS 1.50+ USS0.10. Published by A.A. Balkema, P.O. Box 1675, 3000 BR Rotterdam, Netherlands (Fax: +31.10.413,5947) ‘A.A Balkema Publishers, Old Post Road, Brooktield, VT 05036-9704, USA (Fax: 802.276.3837) For the complete set of two volumes, ISBN 90 5410939 4 For Volume 1, ISBN 90 5410940 8 For Volume 2, ISBN 905410941 6 © 1998 A.A. Balkema, Rotterdam Printed in the NetherlandsGeotechnical Ste Characterization, Robertson & Mayne (eds) © 1998 Balkema, Rotterdam, ISBN 90 54109394 Table of contents Introduction Theme lectures WAC Bennett Dam — The chatacterization of a crest sinkhole RA Stewart, B. D.Watts, J.C Sobkowicz, A.G. Kipper & PC Gaffran ‘Trends in geophysical site characterization J.Greenhouse, P.Pehme, D.Coulter & QYarie Geo-environmental applications of penetration testing PK Robertson, T.Luine & L.J.M. Powell Deformation and in situ stress meusurement M Fahey Analysis and use of CPT in earthquake and environmental engineering AK. Mitchell & TL. Brandon Advanced interpretation of field tests. GT-Houlsby Objectives and planning for site investigations QuickSite, the Argonne expedited site characterization methodology L.CBurton & WE. Meyer Site investigation and characterization for deep excavation K.Chung & KYou Aspects of geotechnical site characterization for large international power projects JLM.Clemente & H.Senapathy Precise geological characterization for the wider application of high quality site data TJ. Davies & M.E.Barton Multiple investigation methods for fault rupture hazard evaluation SFranzen, M.Lew & PJ. Eliott xm 23 35 9 99 Hs 12k 127 133 139A knowledge-based system approach to soil property determination HI Fung & LN.Kay Examination and stabilization of the unstable territory L.K.Ginzburg Geotechnical cross-sections re-visited JVHamel Ground investigation for grouting design: A case history RiJameson & G. Pellegrino Use of SID technique to determine heterogeneity of contaminated sites K.G-K.Jayakody, PW.Jayawickrama & KA Rainwater Site characterization for Hong Kong's new airport at Chek Lap Kok LW.Jones, C.S.Covil & A.R. Pickles Characterization and performance at the reclaimed island of Rokko in Kobe MB Karkee & H.Kishida ‘SimSite: Computer model to simulate site investigation for groundwater remediation MC McBrayer, K. EWiley & R.B,Gilbert ‘Acontribution to the geotechnical characterization of large areas for land planning R.Mauro, A. Pellegrino, M.Ramondini & GUreivoli Order in chaos: The geotechnical characterization of melange bimrocks EWMedley Geology and engineering geology in the siting of complex public facilities R.E.Migues Selecting a low-level radioactive waste site for the Central Interstate Compact: ‘Toward a checklist for siting small hazardous waste projects RE Migues Site investigation for Palau compact road CS. Mimura, R.M.Lim & S.R-Carr ‘The high cost of low bid cost site characterization Mirza Site assessment technique on areas with expansive soils APMwasha Geotechnical investigation for a typical power plant — A case study D.N.Naresh Rock type-engineering properties correkation for a site characterization JG. Nunez, ¥.Prashar & S.Ledegna GIS-ASSESS: A spatial analysis tool for site investigation planning and evaluation R.L.Parsons, J.D. Frost, J-LAChamean & ¥.CTsai ‘Characterisation of soft deposits in the Eastern Nile Delta E Pelli & A.Carletti vi 145 151 159 165, 171 177 183 189 195 201 207 213 219 227 233 239 245 251 257Field and laboratory testing program for the seismic evaluation of Tinemaha Dam, Califomia ¥. Prashar, R.S.Tokashiki, GW.Brodt & S.Inel An innovative foundation in difficult ground through detailed site characterization BSQubain & El Seksinsky Site characterization for reclamation project in North Jakarta PPRahardjo Geotechnical investigations to characterize the Upper Quaternary Basin of Venice G Ricceri & PSimonini Geotechnical investigations for the Great man-made river project, Libya NSaffari & S.J. Everton Great man-made river project ~ Phase If: Heave areas and bedding material investigations N.Saffari & H.K, Mhach ‘The site investigation program for the Oslo Airport, Gardermoen, Norway: . Site investigation strategy RBSandven, A.Wam & G.Jd.Westerlund The site investigation program for the Oslo Airport, Gatdermoen, Norway: 11,Use of in situ test results in design R.BSandven, AWatn & G.J.Westerlund Site characterization by means of the natural slope methodology (NSM) T.Shuk Pressuremeter deformation modulus for dynamic compaction at Oresund bridge between Sweden and Denmark SSwedenborg A review of artificial intelligence systems for site characterization D.GTolt Site characterization for a tand reclamation project at Changi in Singapore B.MWin, AArulrajah, V.Choa & M.EChang Drilling and sampling Application of horizontal directional drilling for contaminated site characterization E.NAllouche, K.W.Biggar, S.TAriaratnam & JW. Mah On the use of drilling parameters in rock foundations PColosimo Characterization of the subsoil conditions at Swann’s Bridge, N. eland SJ.Cummings, LG.Doran, V.Sivakumar, D.A, Hughes & A.S Johnson ‘The use of electronics in the management of site investigation and s Principles and applications EFortunati & G. Pellegrino improvement works: vu 263 275 281 287 293 301 307 35 321 327 333 341 347 353 359Detection of cavities by monitored borehole drilling (TMD) A.LGarassino & M.L.Schinelli Effect of sample disturbance on yield stress and compre ZHong & KOnitsuka lity of Ariake clay New method of coring and cohesion determination on jet grouting improved soft clay Wang Jiwang, Wang Jianming & Wang Jiatin Particulatties of geotechnical investigation on loessiat soils AMarcu & R.Stoica Sampling disturbances of soft sensitive clays TS.Nagaraj, N. Miura & A.Yamadera Estimation of soil resistance using rotary percussion drill K.Nishi, ¥ Sueuki & H.Sasao Development of a high-quality undisturbed sand and gravel sampling method LOhara Interpretation of in-situ and laboratory tests on soft clays H.Ohta, A. lizuka, F¥amakami, K.Ohmori & ¥. Morikawa ‘Sample disturbance in soils - Results from investigations in an overconsolidated marine clay R.BSandven & MSjursen Estimation of in-situ undrained shear'strength using disturbed samples within thin-walled samplers TShogaki & ¥. Maruyama Applicability of the 45-mm sampler with two chamber hydraulic pistons on clay deposits TShogaki & S.Shirakawa Importance of instrumented drilling KTani Use of Japanese sampler in Champlain sea clay H.Tanaka, K.K. Hamouche, Tanaka, ¥Watabe, 8. Leroueil & Fournier ‘The use of rotary sounding for new quality control of soil improvement ¥.foukada, H. Mitsuhashi & S.Chida Study of a soil assessment method based on sounds generated by rotary sounding ¥.Tsukada, NAoyama, Yamaki & A.Chichibu Geophysical testing - Seismic techniques Delineation of densified sand at Treasure Island by SASW testing R.DAndrus, R.M.Chuag, K.H.Stokoe, Il & 1A. Bay Quantitative interpretation of steep basement reverse faulting from seismic reflection data R.d.Cumbest & Wd. Domoracki Experimental procedures for detection of underground objects by the SASW test N.Gucunski, V.Krstié & M.H.Maher vill 365 371 377 381 387 393 399 403 419 a5, 439 445, 451 459 465 469SASW control of a vacuum consolidation on a sludge disposal WHaegeman & WEVan Impe ‘The use of seismic geophysics in the characterization of a weak rock site VS Hope, CR.LClayion & LA,Sutton Geophysical surveying methods for soft soil E.C.Leong & TC. Pan Geophysical/ geotechnical characterization of upper Borrego Valley A. Martin, R.L.Nighor, D.M Pendergraft & R.A.Steller GEOSIS: Integrated approach of geotechnical and seismic data for offshore site investigations J ENauroy, JC Dubois, J.L.Colliat, A. Puech, J.P Kervadec & J. Meunier Simultaneous inversion of surface wave velocity and attenuation GJRix & CG.Lai Seismic velocity in cohesiontess granular material deposits MLL Rucker Shallow control for high resolution seismic reflection surveys using piezocone penetrometer soundings FH Syms, RL. Cumbest & D.E Wyatt Experience gained in testing pavements by spectral analysis of surface waves A. Szavits-Nossan, R.Mavar & M.-S.Kovatevié Development of borehole sonar for the evaluation of soil-cement column Sakata, K. Hosokawa, A.Abe, M.Tamura, M.Fujii & K.Komiya Examination of liquefaction potential by seismic tomography after the Hyogoken-Nambu Rarthquake in the reclaimed land of Kobe K.Tanimoto, ¥ Takahashi, ¥. Murata, M. Yamamoto & N.Sugawara ‘Three-dimensional soil stratification using surface waves in microtremors K Tokimatsu, SWakai & H.Arai Liquefaction assessment of mine tailings dams S.DVidich, GH. Beckwith & LR. Keaton Geophysical testing ~ Electrical techniques ‘The establishment and monitoring of expansive soil field sites M.D, Delaney, MAAllman, D.W.Smith & SW.Sloan Monitoring water infiltration under an eatth-fill levee with geophysical techniques D.Doser, DYuan, S Nazarian & M. Baker Dielectric constant and electrical conductivity of contaminated fine-grained soils A.Kaya 473 479 485 491 497 503 509 S15 521 531 537 543 551 557 563A tial of five geophysicat techniques to identify small scale karst M.Long RI-Cone penetrometers experience in naturally and artificially deposited sand M.Mimura & A.K. Shrivastava Characterization of collapsible soils with combined geophysical and penetration testing VA. Rinaldi, J.C Santamarina & E.R. Redolfi Collecting real-time soil moisture profiles in the vadose zone CM Rose, S.C Armstrong, J.L. Walker, G.NYoung & LC Burton Development of a CPT probe to determine volumetric soil moisture content JD,Shinn, I, R.L.Hull, D.A.Timian & R.M.Morey Radio-isotope cone penetrometers and the assessment of foundation improvement AK. Shrivastava & M.Mimura Application of advanced borehole geophysics in site geological, geotechnical and environmental characterization D.EWyatt & Ru. Cumbest Geoenvironmental direct push technology A novel (echnique for detection of contaminated groundwater in situ W.L. Bratton & S.P. Farrington Integrated opto-electronic chemical sensor for BTEX detection in cone penetration testing SEBurns & PW.Mayne Reliability of soil gas sampling and characterization techniques D.J.DeGroot & A.J. Lutenegger Remedial design for DNAPL recovery at a former wood preserving plant, Pensacola, Florida MAGage & LE Young Grouting on retraction of cone penetrometer LT.Lee, PG. Malone & G.E. Robitaille ‘Tomographic site characterization using CPT, ERT and GPR RM Morey SCAPS characterization of VOC contaminated sites GE Robitaille & MA.Ruddle Determining contaminant distribution and migration by integrating data from multiple cone penetrometer-based tools ‘LRossabi, E R.Jarosch, B.D. Riha, B.B.Looney, D.G.Jackson, CA. Eddy-Dilek, R.S.Van Pelt & BE. Pemberton Brine plume mapping using cone penetrometer and geophysical methods ALStrutynsky, RAGlaccum & LCConklin 569 575 S81 589 59S 601 607 617 623 629 635, 641 645 651 655 663Groundwater drinking supply protection using cone penetrometer methods AL Strutynsky, K.Gilmer, C. Osborne, J. Romp & R.Darr Geoenvironmental sampling and testing Monitoring well casing material behaviour subject to different groundwater contaminants KWBiggar & SYong Engineering significance of ground sulphates ABHawkins ‘Tracer methods for groundwater flow and pollution transport characterization J.Hulla, E.Bedndrova, J, Pisko & M.Kostolansky Site characterization performed to support a municipal solid waste environmental recuperation in Recife, Brazil JET.Jucd, S. Machado Santos & E.G.Bastos On-line monitoring of NOx gases with sensors based on oscillating crystals O.0cié & PBJovanié Developing methodology for visualization of ecological accidental situations in ptants O.0cié & PBJovanié Establishment of regional soil gas background values to support waste site screening investigations D.M, Richers & D.EWyatt Author index xt 669 677 685, 693 699 705 709 713Geotechnical Site Characterization, Robertson & Mayne (eds)© 1998 Balkema, Rotterdam, ISBN 9054109394 Introduction In the past 25 years there have been many interesting and informative conferences on topics related to site characterization for geotechnical engineering concems. In addition to conventional drilling, and sampling techniques, there have emerged a mumber of new in-situ tests, including the cone penetration test, piezocone, flat dilatometer, pressuremeter, and other specialty probes, blades, and push-in devices. New drilling and sampling methods have been introduced, including sonic drilling, continuous sampling, combined rotary/hydraulic/ percussive tools, as well as automated devices. A marked increase in the utilization of geophysical techniques has also occurred, notably the profiling of shear wave velocities, ground penetrating radar, resistivity, and other nondestructive measure- ments. As a consequence, a number of successful symposia were held on penetration testing (ESOPF-1, 1974; CPT’81; ESOPT-2, 1982; ISOPT-1, 1988; CPT’95), conferences on in-situ testing (In-Situ’86, PTUK, 1988), symposia on pressuremeters (ISP-1, 1982; ISP-2, 1986; ISP-3, 1990; ISP-4, 1995), vane shear testing (ASTM STP 1014, 1987), chamber testing (ISOCCT:1, 1991), gcophysical testing (STP 654, 1978; STP 1213, 1994), and others (see list to this introduction). These prior conferences have set high standards in the communication of research and practice related to site characterization and in-situ testing. The members of the Technical Committee (TC 16) on Ground Property Characterization from In-situ Testing as part of the International Society of Soil Mechanics and Geotechnical Engineering (ISSMGE) felt that it was appropriate for a major international conference that would link all the different areas related to geotechnical site characteri- zation. Hence, with the assistance of ISSMGE TC 10 (Geophysical Site Characterization) and TC 24 (Soil Sampling) the concept of the First International Conference on Site Characterization (ISC°98) was born. ‘The objectives of ISC’98 are to provide a forum for discussion on all aspects of geotechnical and geoenvironmental site characterization including: planning, specification, drilling, sampling, in-situ testing and geophysical testing. Particular emphasis has been placed on the exchange of practical experience and the application of research results. An aim of this conference has been to enhance the exchange of knowledge between researchers and practitioners ftom countries all over the world and (o facilitate interaction between experienced and younger geotechnical engineers and engineering geologists. A further theme of the conference has been the exchange of novel and innovative ideas, new technologies, and practical applications related to geotechnical and geoenvironmental site characterization, The conference proceedings have been arranged into sessions that cover: Objectives, scope and planning for site characterization; Geophysical testing and remote sensing; Geoenvironmental testing; Deformation and in-situ stress testing; Penetration testing; and Advanced interpretation of field tests, The Technical Program also comprises six ‘Theme Lectures by eminent international experts in the main areas of site characterization, Discussion Sessions on each main Theme include presentations by leading experts as Discussion Leaders and selected presentations from authors of papers. Poster Sessions, a Technical Exhibition, and a Field Denionstration Session have been xuarranged to allow participants an opportunity to see and discuss new techniques in greater detail. The Proceedings also contain a report entitled Pressuremeter Testing in Onshore Investigations prepared by B.G.Clarke and M.Gambin for TC 16. This report represents the first in an expected series of reports on major in-situ test methods. The objective of the report is to describe pressure- meter equipment, site operations, interpretations and applications to guide practitioners in the use and application of pressuremeters. ‘The Conference would not have been possible without the dedicated work and competence of the many authors who have submitted papers and the hard work and enthusiasm of the many individuals and organizations and companies that have provided the basis for the planning and successful implementation of ISC’98. We hope that you enjoy reading and assimilating the 206 papers in these two-volume proceedings and find use for these technical publications in your daily practice and research. Peter K. Robertson Paul W. Mayne University of Alberta Georgia Institute of Technology References (Conferences related fo Site Characterization, given in chronological order) European Sympos Stockholm. Dynamic Geotechnical Testing (1978), Speciat Technical Publication (STP) No. 654, American Society for Testing & Materials, West Conshohocken, Pennsylvania, First International Symposium on Pressuremeter Testing (ISP1, 1982), Editions Technip, 27 rue Ginoux, 75737, aris, France. Cone Penetration Testing and Experience (1981), Sessions Proceedings at ASCE National Convention, St, Louis, Missouri; published by American Society of Civil Engineers, Reston, Virginia (CPT”81). uropean Symposium on Penetration Testing (ESOPT-2, 1982), Amsterdam, The Netherlands; published as ‘Penetration Testing, A.A. Balkema, Rotterdam. Second Intemational Symposium on Pressuremeter Testing (ISP2, 1986) College Station, ‘Texas, USA; published ‘as The Pressuremeter and Its Marine Applications, Special Technical Publication No. 950, American Society for Testing & Materials, West Conshohocken, Pennsylvania, Use of In-Situ Tests in Geotecinical Engineering (In-Situ’86), Blacksburg, Virginia, USA; published as Geotechnical Specialty Publication (GSP) No. 6, American Society of Civil Engineers, Reston, Virginia, 1986. Vane Shear Strength Testing in Soils: Field & Laboratory Studies (1987), Special Technical Publication No, 1014, ‘American Society for Testing & Materials, West Conshohocken, Pennsylvania. Proceedings of the Geotechnology Conference on Penetration Testing in the UK, Birmingham (1988); published as Penetration Testing in the UK, Thomas Telford, London (PTUK). First International Symposium on Penetration Testing (ISOPT-1, 1988) Orlando, Florida, USA; published as Penctration Testing 1988 by A.A. Balkema, Rotterdam. ‘Thied International Symposium on Pressuremeter Testing (ISP3, 1990) Oxford, UK; published as Pressuremeters by Thomas Telford and the British Geotechnical Society, London Proceedings of the First International Conference on Calibration Chamber Testing (ISOCCT-1, 1991), Potsdam, ‘New York; published as Calibration Chamber Testing, Elsevier, New York. Dynamic Geotechnical Testing IL (1994), Special Technical Publication No. 1213, American Society for Testing & Materials, West Conshohocken, Pennsylvania, Fourth Intemational Symposium on Pressuremeters (SP4, 1995) Sherbrooke, Quebec, Canada; published as The Pressuremeter and its New Avenues by A.A. Balkema, Rotterdam. Intemational Symposium on Cone Penetration Testing (CPT"95), Linkdping, Sweden; published by the Swedish Geotechnical Society, 1995. “Advances in Site Investigation Practice (1995), Proceedings of the Institution of Civil Engineers, J.ondon, UK, ‘March 1995; published by Thomas Telford, London. ‘nv on Penetration Testing (ESOPT-1, 1974), published by the Swedish Geotechnical Society, xIVTheme lecturesGeotechnical Site Characterization, Robertson & Mayne (eds) @ 1998 Balkemna, Rotterdam, ISBN 9054100394 WAC Bennett Dam - The characterization of a crest sinkhole R.A. Stewart & RC.Gaflran - BC Hydro, Burnaby, BC, Canada B.D, Watts - Kloln-Crippen Consultants Lid, Richmond, B.C, Canada J.C.Sobkowiez - Thurber Engineering Ltd, Vancouver, B.C, Canada A.G. Kiipper—AGRA Earth and Environmental, Edmonton, Alb., Canada ABSTRACT: In June 1996, a small cavity was discovered on the crest of the 183 m high WAC Bennett Dam in British Columbia, Canada, Subsequent drilling to investigate the condition of the dam core below the cavity resulted in a sinkhole on the dam crest with a surface expression 2,5 m in diameter and 7m in depth. Following this incident the safety status of this very large dam was uncertain. A comprehensive investigation program was urgently planned and executed to characterize the sinkhole. ‘This paper describes the objectives and eriteria developed for the program, the scope of the key activities at the sinkhole, and some important lessons learned during the investigation. L. INTRODUCTION Site investigation and characterization are among the first steps in determining a project’s feasibility and cost. In areas where the ground conditions and the performanee of structures are well known, site chatacterization may be routine. For greenfield sites poor ground conditions and sensitive new , foundation design and site development may be a significant factor in the overall project cost and viability. In many eases a significant component of construction risk may be site development itself, Performance and design requirements are defined by an owmer or owner's representative in concert with his engineering team. Site investigation methods are selected to match these requirements, usually in the context of competitive bidding. The objective is to obtain all geotechnical information necessary for design, construction, and dispute resolution, Due to the inherent variability of ground conditions, the most challenging aspect of site characterization can be determining what constitutes adequate investigation coverage. In effect, the development of the scope of a site investigation is an exercise in risk management, generally done intuitively by suitably experienced. practitioners. The scope of the investigation is balanced among existing information, local knowledge, anticipated variation in ground conditions, sensitivity of the structures to variable conditions, the performance requirements, consequences of poor performance, and the expected or available budget. On large projects with considerable risk (probability of failure x consequence), decision analyses and risk management techniques have been applied to site investigation design (Vick & Bromwell 1989). Another type of site characterizati construction foundation investi subject to distress, The difficulties of this type of work are often compounded by the spatial and logistical restrictions of working in an operating facility. ‘The case history described in this paper deals with the investigation of a sinkhole in a very large dam under an operating reservoir about 170m deep. The overriding objective was to determine the safety of the dam, ‘At the start of the investigation, it was not known whether the disturbed core of the dam was confined to the zone directly below the sinkhole or resulted from a more pervasive mechanism for which the collapse at the surface was but an early warning. in the latter case, drilling to understand the safety of the dam could actually reduce safety through hydraulic fracturing and/or erest collapse. Central to. all investigation planning and execution was the minimization of risk due to drilling, Although the objective was clear, planning was complicated by the necessity to immediately understand the dam safety status, by limitations in the state of practice in drilling technologies able to investigate a dam core under full reservoir, and by an uncertain scope. ‘This paper deseribes planning issues and lessons learned during this challenging project. There is also limited discussion of the investigation techniques and results.2. 1996 WAC BENNETT DAM INCIDENT On June 14, 1996 a passing motorist discovered a 450 mm diameter hole in the asphaltic conerete road surface on the erest of Bennett Dam. Examination of the hole within an hour of its discovery revealed a cavity with a volume of 2 m* and a maximum depth of 2m extending heneath the crest pavement. The top of a vertical steel pipe (later determined (o be a construction survey benchmark tube) was barely visible at the base of the cavity. ‘The following day a careful excavation to 5 m using a backhoe confirmed the presence of the survey benchmark tube (sce Figure 1), and established that the cavity was not evident below 2m. The benchmark tube comprised concentric 65mm inner and 150mm outer thin walled steel pipes; the outer pipe was designed fo act asa sectional sleeve to protect the inner survey pipe. The existence of the buried benchmark tube was not known fo the surveillance engineers who regularly inspected the dam, nor did it appear on any readily available construetion drawings. It was later determined that the benchmark tube extended from the rock foundation at a depth of £15 m to the dam crest. ‘Our immediate assessment speculated that the cavity was likely associated in some as yet undetermined way with the survey benchmark tube. ‘The volume of the surfuce cavity was similar to the open annular space between the concentcic pipes over their 115 m depth to bedrock. This volume balance provided a plausible early working hypothesis for the cavity development, That is, over the 24 years since reservoir filling, core material had slowly migeated into the annular space between the In an attempt to evaluate this initial it was decided to investigate the condition of the dam core at depth around the benchmark tube using a Becker Penetration Test (PT). On June 17, 1996 the first BPT hole was drilled within 1.5 m of the survey benchmark tube, using a elose-ended 140mm OD pipe, driven by single acting diesel hammer. This pipe advanced to 32m with virtually no resistance. While drilling was suspended to attach additional pipe at 32 m the dam crest surrounding the survey benchmark suddenly dropped 7m, leaving a cylindrical, vertical sided hole 2.5 m in diameter (see Figure 2). Coincident vwith the collapse, the water fevel in an open “leaky” inometer casing (Observation well OW-5) 30 m_ away rose suddenly by more than 2.5m. This unexpected local crest collapse constitutes the 1996 incident at Bennett Dam. In early July a second survey benchmark tube, and an associated second more subtle sinkhole, was discovered in the upstream shell overlying the core {just upstream of the dam crest. Bennett Dam is located on the Peace River near Hudson's Hope in northeastern British Columbia, Canada as shown on Figure 3. ‘The dam is a zoned carthfill embankment 183 m high, In 1967, when construction was completed, Bennett Dam was the highest earthfill embankment in the world. It retains the 360 km long Williston Reservoir, the largest reservoir in North America, The underground powerhouse generates 2730 megawatts, which is 30 percent of the power requirements of British Columbia, A layout of the dam is shown on Figure 4. Also shown are the locations of the bo sinkholes, The core of the dam is a brondly graded, non- plastic silly sand with some gravel which wes manufactured from glacial outwash deposits. Downstream of the core are two filter zones and a Figure 2. Sinkhole No. 1.Figure 3. Key plan (after Ripley 1967). coarse gravel chimney and blanket drain for seepage control. The upstream and downstream shells were constructed with sand and gravel, The dam is founded on Tertiary sandstones and shales which dip 5 to 10 degrees downstream. A typical cross section is shown on Figure 5. Details of the design and construction are given in Ripley (1967) and the 25 year performance of the dam is reviewed by Stewart & Imrie (1993). Following the dramatic but localized crest collapse, it was clear that significant additional investigations would be required to determine the safety of the dam, to understand the extent of the Figure 4. General arrangement (after Ripley 1967). damage to the core, and to assess alternatives for rehabilitation as required The collapse occurred during the spring freshet as the reservoir was rising to its maximum pool elevation. In consultation with the Office of the Comptroller of Water Rights, British Columbia's dam safety regulator, and the Advisory Board (Drs. RB, Peck and N.R. Morgenstern) the decision was ‘made to lower the reservoir. Intrusive investigations of the core were delayed unlil freeboard was increased. This would improve the safety margin in the event that another crest collapse occurred, On June 22, BC Hydro opened the spillway gates. Discharges through the spillway and powerhouse, at a maximum rate of 5100 m’fs, lasted eight weeks. Over this period the reservoir drawdown was less than 2 m, During reservoir drawdown, various surface geophysical techniques were employed to investigate the condition of the dam beyond the sinkholes, During. this time the intrusive investigations at the sinkhole were planned, Part of this planning involved trial drilling and downhole geophysical surveys in intact portions of the core at locations distant from the sinkhole. During the early investigation stages, the uncertain safety status of the dam drove the urgency to complete the work. As the investigation proceeded, and the safety of the dam was better understood, the need to complete the investigations was driven by the requirement to remediate the dam so that the entire 1997 spring freshet could be captured. The investigations were planned to accommodate these goals, incorporating dam safety risk mitigation measures to the extent possible, 3. INVESTIGATION OBJECTIVES AND CRITERIA ‘The objectives which formed the comerstone of the investigation were to: © characterize the extent and nature of the disturbed core beneath the sinkhole(s); ‘© establish whether the loose zones and their cause constituted a dam safety hazards ‘+ gather sufficient information for remediation, design and construction; and ‘complete the investigations before the harsh winter to permit return of the reservoir to full service. ‘Although the scope of the work could not be fully defined, the planning was facilitated by the delay needed to meet the requirement for additional freeboard. This investigation delay of 6 weeks was used to develop the criteria and protocols, and to determine the test equipment and techniques. AllZONES O cone @ reansirion @ Farce @ pean @ reevious sueue oroaa. two CREST £6,601 6 e ooo Figure 5. Dam eross section (after Ripley 1967). planning was done at site by a resident team of senior personnel brought together from BC Hydro and consulting companies across Caneda. Experts in specific investigation techniques were retained as needed, Key investigation criteria were as follows: © minimize all dam safety risks during drilling; ‘+ maximize the use of surface geophysical (non-intrusive) techniques to examine at least the upper part of the dam; ‘© use proven technology to drill atthe sinkhole, with modifications to minimize core damage; * use the least intrusive techniques for investigating the sinkholes; ‘develop the program incrementally, building ‘on the early results; * access the global experience of others who have investigated similar problems; ‘© minimize the number of drill holes, recognizing that the information gathered must be more valuable than the potential for immediate or ong term damage to the core; ‘© maximize the amount of information gathered al cach drill hole; © deill 24 hours/day and use multiple drills where practicable; «assess drilling data as obtained and adapt the program as conditions become better understood * test any unproven methods under conditions similar to those expected at the dam; + develop safety and surveillance protocols, both for personnel and emergency response (eg., heavy mechanical equipment and emergency grouting equipment to be in “ready” mode on the dam crest); + develop detailed drilling procedures, maintain a continuous presence of an experienced senior engineer on crest for “first response”, and develop deiailed response plans for all foreseeable contingencies. “The investigation plan incorporated all objectives and eriteria noted above. The following sections discuss some individual components of the investigation program, particularly identifying planning considerations and lessons learned. ‘This paper describes the investigation at Sinkhole No. 1. Investigations at Sinkhole No. 2 and beyond the oles are not discussed. 4, CONE PENETRATION TESTING 4.1L Selection of Method and Objectives ‘The seismic piezocone penetration test (SCPTU) was the principal investigation technique used to define the characteristics and extent of the disturbed dam core beneath the sinkhole. The SCPTU was selected because, of all intrusive investigation fechniques available, it yields the most information with the least disturbance. It is also a very controlled, sensitive penetration method with seal time data output. While considering the seismic piezocone, we were concemed about whether it could be pushed into the core at all, because of the likely presence of inter-layered loose and dense materials. Piezocones, ‘or for that matter cone penetrometers of any kind, cannot be pushed into well compacted soils. ‘The basis for believing that the piezocone could penetrate the core came from the minimal resistance experienced by the Becker Penetration Test at the sinkhole, We were also aware of dynamic cone penetration tests of the backfill materials around vertical riser pipes in the glacial till core at the LG4 development in Quebec (Boncompain et al. 1989). The SCPTU program objectives were to determine the depth and extent of the disturbed zones, and to gather as much data as possible fiom each SCPTU profite.4.2 Deseription of SCPTU Equipment and Procedures ‘The piezocone included transducers for measuring tip resistance, sleeve friction, pore pressure and temperature. It also included a bi-axial geophone for measuring compression wave/shear wave velocity. The equipment and penetration procedures followed ISSMFE (1989) standards, except that the tip area was 1Scm? instead of 1Qem® and the penetration rate was generally slower than 2 em/s. A larger tip was used because of the potential need for greater sensitivity in very loose core. A slower penetration rate was used because of concer that the cone rods would buckle on dense material before the hydraulic ram could be shut down, Both concerns proved to be vali After collaring each SCPTU hole (Sonic, Barber, id mud rotary rigs were used for this purpose) initial drilling was carried out by a Simco 5000 mud rotary drill rig to reach the test level. Procedures for drilling to support the SCPTU tests were similar to those described in Section 6 of this paper, ‘The piezocone was pushed with a hydraulic ram fitted to the Simco 5000 driil rig. Before cone testing, HWL rods and BWL rods were lowered inside the HW casing to laterally support the cone rods, ‘The piezocone was then lowered to the bottom of the casing and pushed into the soil. Testing continued until: ‘© the piezocone refused in a stiff zone, or * bending in the piezocone rods became excessive (due to a long, laterally xsupported section beyond the casing in soft ground). “The piezocone was then removed from the hole, the casing was advanced, the hole cleaned out, and testing continued, Because of friction buildup, the casing had to be nested several times fo reach target depths. Before resumption of SCPTU testing in uncertain ground conditions, a “dummy” cone ‘mounted on BWL rods was quickly lowered to the bottom of the casing and pushed into the soil. On several occasions the dummy cone encountered stiff ground. This trial procedure saved time in lowering and then removing the 1 m threaded SCPTU rods. Ryery 1 or 2 metres, piezocone testing was paused fo allow pore pressures and temperatures to equilibrate around the tip, and to carry out downhole seismic velocity testing. Seismic “shots” were made by discharging a heavy gauge shotgun into a prepared hole, ‘The shot pattern varied, but generally ‘included locations on four sides of the SCPTU hole at several prescribed offsets. 43 SCPTU Test Program and Results Six SCPTUs were completed at Sinkhole No. 1. Typical results are shown on the following figures: + Figure 6 shows the location of selected drill holes at Sinkhole No, f, + Figure 7 shows the standard test results (tip resistance, shaft friction, friction ratio, and dynamic and equilibrium pore pressures) for the first SCPTU hole in Sinkhole No. 1. © Figure 8 is a section showing piezocone tip resistance with depth for a number of SCPTUs at varying distances from the centee of Sinkhole No. 1. ‘Note that in the first SCPTU the tip resistance is remarkably tow (0 a depth of 33.5 m, with values approaching the equilibrium pore water pressures. ‘The cone rods essentially fell under their own weight to 33.5 m, The low tip resistance was first thought to be due to low density but subsequent testing showed that it was primarily a function of very low stresses in the disturbed zone. Below the low tip resistance zone is dam core with rapidly varying. density. The characterization of the dam core is described in a later section. 44 Lessons Learned ‘The first SCPTU test was carried out, of course, without knowledge of the general conditions below the sinkhole. There was much debate about whether there would be a repeat.of core collapse andlor pore pressure response in adjacent —_piezometers, Consequently, an attempt was made to anticipate all possible dam responses, and to develop procedures and contingency plans to address them, However, when venturing info unknown ground conditions, there is always much to be learned, We recognized at the beginning of the program that the piezocone investigation had an inherent imitation. Beeause the piezocone cannot be pushed through compacted soil, the location of the loose zone must be estimated a priori. ‘The piezocone is therefore a characterization (ool, not a general investigation tool in this situation Normal methods of interpreting SCPTU data (Lunne et al. 1997) are based on measured parameters and some knowledge of in situ stress state. In this case, in situ stresses were extremely low. Since this was not known at the start of the program, interpretations of soil type using SCPTU results were misleading. Fortunately, soil classification was not an issue, as material type was well known. On other projects involving unexpected and unusual stress conditions, this issue could be ‘more important,SURFACE EXPRESSION OF SINKHOLE No.1 DH96-24 N NOON ~ CPT96.3 Chives bef DHI6-34 ~y HOT DOWNSTREAM SLOPE, DH96.28 SCALE: 3. 0 3 6 METRES LEGBND © BENCHMARK TUBE @ = SONIC DRILL HOLE TOMOGRAPHY PLANE Figure 6, Selected drill holes al Sinkhole No. 1. ‘The “dummy” cone proved to be a valuable toot for proving out ground after dritlouts because drill rods can be lowered much faster than SCPTU rods. Drilling uid was routinely lost when cleaning out the inside of the casing, as the drill bit approached the bottom of the casing, This issue is discussed later. Despite the above shortcomings, the SCPTU exceeded our expectations in _ providing comprehensive geound information quickly and, most importanily, without reducing the safety of the dam. Indeed, the success of the SCPTU program permitted the use of more intrusive investigation techniques at the sinkhole, 5. SONIC DRILLING 5.1 Objectives In addition to the SCPTU program, we recognized that it would be necessary to drill holes into the dam to probe the extent of the disturbed core, to characterize the condition of both disturbed ‘and —~ ~e NODA pen6a6 7 La) EX ~ DIW630 - aries S 4 UPSTREAM ‘SLOPE BENCHMARK TUBE, N\, \ \ CPI96-4 CONE PENETRATION TEST HOLE, MUD ROTARY DRILL HOLE @4 INCLINED SONIC DRILL HOLE undisturbed core, to define pore pressures, and to install instruments of various types around the sinkhole, However, the spectre of damaging the core by the drilling itself caused great concem, and resulted in a considerable planning effort to select a suitable drill rig and drilling technique. The drilling method would have (0 be capable of investigating intact core and any zones of intermediate disturbance where the SCPTU could not be pushed. Listed approximately in order of decreasing priority, the drilling method should: 1, Not damage the dam core in any significant way. In particular, it should not cause hydraulic fracturing under conditions of low total stress and/or low pore pressures which might exist in the dam core, 2. Not trigger collapse of loose zones (as occurred with the Becker rig). 3, Be capable of ling 10 depts in excess of 120m. 4, Allow recovery of soil samples for logging and laboratory (ests. Asa minimum this should include standard classification tests and preservationQt (bars) Fs (bars) REM) um) ° 200 40 0 1 2 3 0 2 4 - 0 0 2 0 Opa Pesste ic 20 30 = a4 & a = z | 6 70 80 F Figure 7. CPT96+1 profile at Sinkhole No. |. of soil micro-features sch as layering, seams, gradation changes, ete, 5. Allow measurement of in situ density, 6. Permit installation of piezomeiers and casings (for later geophysical testing). 7. Allow proper completion of the hole, including grouting and sealing to prevent migration of water vertically through the drilted zone. 5.2 Selection of Drilling Method A drilling method that could meet the above criteria equatly well would be a rare find, but the main concern was meeting the foremost criterion, that of not damaging the core, This seemed to mile out standard techniques that employed drilling fluids. Air-drilling methods had the greatest potential 10 cause hydraulic fracturing, but methods using water and drill mud were aso undesirable. Hydro Quebee was in the process of drilling several relatively shallow experimental holes into dams with similar cores using very heavy drill mud. ‘The benefit of the heavy mud appears to be lack of mobility along fracture planes despite the increased likelibood that heavy mud will cause hydraulic fracturing compared to water or air, However, this feclnique was in the research stage at shallow depths, and thus not considered suitable for immediate use at Bennett Dam. Drilling methods that drive a casing into the ground were also considered unsuitable, mainly because of the energy required and because they were foo slow or would not reach the necessary depths. "This included the Becker hammer drill and cable tool drilling, One drilling method met most of the criteria given earlier, This method, called “sonic” drilling, had been successfully used to similar large depths in the Fraser River delia. Tt is used extensively in environmental work and had been used previously by BC Hydro in seismic assessments in granular materials in dams. Reviewing past performance of the sonic drill, it appeared capable of meeting most of the objectives stated earlier. ‘The exception ‘+ The objective of not triggering collapse in a moderate to large loose core zone was probably not achievable with a rig that had sufficient power to drill to the desized depth and that transferred significant vibration into the ground.Figure 8: SCPTU profites at Sinkhole No. 1. ‘+ The continuous sample recovered by a sonic drill is slightly disturbed and thus it is not possible to directly measure density of the core. 5.3 Description of Sonie Drilling ‘The sonic drill is truck mounted and accompanied by a support truck which carries casing and drill rods (Figure 9), ‘The rig has a hydraulic head with an osoillator that generates and transmits a eyelic axial force down the drill rods to the core barrel and drill bit (Figure 10). ‘The vibrations at the bit loosen moist soil and liquefy saturated soil, allowing the core barrel to penetrate the ground. ‘The sonic drill altemately advances the core barrel and then the casing, normally in 3 or 6m sections. The equipment is not wireline; consequently the dritl SURFACE EXPRESSION OF SINKHOLE No.1! cpro¢-1 ENO dsm) ; ENG BB BNE | Comb) Tint) panterest | \ i 0 NS Ny 20 Be = 30 N. " Ee) RS 50 e ge tira ¢ B.° E a a -70 LEGEND: 30 og sas Druin our ane &] rpresistance *” aero) “| pisoF SECTION " N oRe rnxstortrosra exawacesTion 1) qecutray seats re D Mees 0 N. = parorrkt— SS DEDROCK SURFACE 16 one = sneer gee) rods and core barrel must be removed from the hole to retrieve the sample, Holes are usually started with a large casing and core barrel, and then telescoped down in size as friction builds up on the casing. Table I shows typical rod sizes, Normally the equipment operates dry, but a ‘Table 1. Sonic drill casing, core barrel and rod sizes. ‘Nominal Size Inside Outside (am) eter, Diameter, omy) nm), Hem Core Barrels 115 96.2 nas 165 2 1619 Drill Rods 90 160 89.8 Drill 40 122.8 1403 Casings 190 146 193.7 20 201.2 219.1 10wre 9, Sonic drill and support truck. modification on this project was to add only enough water to balance in situ pore pressure. To address the shortcomings of the sonic drill, the following were implemented: + Investigations with the sonie drill were started at a distance from the sinkholes, with the intention of drilling in ground that was largely undisturbed, to avoid triggering ground collapse. ‘© The sonic rig was instrumented to measure the following drilling parameters: hydraulie pressure and rate of fluid flow (o the vibratory head, vibration frequency and acceleration of the head, hydraulic pressure driving the torque and downward pressure on the drill rods, and drilling advance rate, This had never been done before, and it was hoped that correlations could be developed between these drilling parameters and ground conditions such as density. Su Testing of Sonie Drilling The capabilities of the sonic rig were tested in sound " core before drilling near the sinkholes. ‘The first sonic hole was drilled in early July 1996, on the right abutment of the dam, to a depth of 68 m. No problems were encountered in this hole. The next ‘wo sonic holes on the crest were drilled in mid- and late July 1996, in the terrace section (Figure 4), (0 depths of 120 m and 123 m. In the first of these holes, the 140 mm OD casing stuck at a depth of 117m but was recovered with a 40 tonne jack, While drilling in the rock foundation, the core barrel stuck and was eventually lost. ‘The second hole was drilled without incident. Even though problems were encountered, the prototype drilling was successful in that it identified the need to use at least 3 sizes of casing when drilling to 120 m depth, and allowed development of drilling, sample recovery, and hole completion techniques while still drilling in non-critical areas of the dam. It also provided the opportunity to test crosshole tomography techniques. 5.5. Production Drilling ‘Thirteen additional sonic holes were drilled from the crest of the dam into the core. Most of these were about 120 m deep, but two holes drilled in the canyon section extended deeper, one to a record depth of 140 m, This was considered to be the depth limit in the core using three sets of casing. ‘The depth from the dam crest to base in the deepest part of the canyon is 183 m. To reach this depth with the sonic drill, one would need to telescope an additional casing size, either starting with a larger or finishing with a smaller diameter casing than is commonly available. As part of the risk mitigation procedures and in anticipation of settlement during drilling close to the sinkholes, all drill rigs were supported on steel beams spanning the dam crest, and the work area was overlaid with geogrid (o protect personnel from sudden drops in the ground surface. In addition, surveillance crews monitored surface and subsurface ‘movement while drilling was underway. After completion of drilling, 70 mm diameter slope inclinometer casing or 76 mm diameter plastic casing was lowered into the hole, then the hole was grouled in slages (lo limit grout pressure and thus prevent hydraulic fracturing) while simultaneously removing the steel casing. In areas of very soft andor collapsing ground, tube-d-manchette sections were included to permit later grouting of the plastic casing. 5.6 Results ‘Two figures are included which show results fromFigure 10, Hydraulic vibratory head on sonic sonic drilling. Figure 11 shows fines content and void ratio (from water content) with distance from the sinkhole, Figure 12 shows a typical plot of drill parameters with depth. 5.7 Lessons Learned Many lessons were leamed throughout the sonic drilling program, first as drilling and sampling procedures were developed, and then later as specific problems were encountered in each hole, “The first lesson was the need to carefully control water level in the drill hole. ‘The water level needed to be sufficiently high to prevent heaving of soil, but also sufficiently low to prevent hydraulic fracturing. Control of water level was even mote important when carrying out falling head permeability tests. Proper control in most cases was obtained only where the piezometrie pressure was known with @ reasonable degree of accuracy before drilling. This formation was gathered from previous drilling, adjacent SCPTU testing, seepage analysis and existing piezometers in other parts of the core. Another lesson was the difficulty of retrieving continuous samples in loose or soft ground. Various types of bits and core catchers were available, but 12 experimentation was necessary to develop a reliable sampling procedure. {In some cases, despite these measures, samples in loose ground were repeatedly dropped, causing severe sample disturbance (e.g., mixing, segregation when dropped through water). ‘The safety precautions taken during the sonic drilling proved to be invaluable. Surface settlements: of up to 1.5 m occurred over an area of 2 to 3. diameter around individual drill holes in or close to the sinkhole, A group of five field inspectors supervised the drilling on a 24-hour basis (2 rigs), monitored instrument response, and Jogged all samples. Factors which contributed to the quality of this work were: written drilling, testing, and safety procedures that were updated as knowledge was gained; specific written instructions for each hole; establishment of standards for logging holes; and thorough and frequent communication between senior engineering, staff and field staff. In regards (o logging standards, we found that in the early part of the program all staff were consistently underestimating the fines content of the dam core samples. The fines in the samples were almost entirely coarse silt, which was difficult to distinguish visually from’ fine sand. A set of standard samples was prepared, and all staff were trained in the visual estimation of fines content, after which time the logging became more consistent and accurate. This issue was important, because other field and laboratory testing was triggered by the identification of genuine fines deficient zones in the core. Drilling parameter data (such as advance sate, head vibration or acceleration, etc.) were collected ‘on most holes. Very extensive and detailed analysis of this data, both on a theoretical basis and looking for trends, provided no correlations with the geotechnical characteristics of the ground. ‘This was a keen disappointment, Completion grouting of the holes proved to be very difficult, Grout was placed in the hole in controlled lifts to prevent hydraulie fracturing of the core, Typically these lifts were 15 to 20 m thick. Since the grout needed to set up before another lift was placed, it was a time consuming process (even with the use of accelerators), one that ideally should proceed independently of the drilling except for the need (0 remove easing from the ground, In some holes where a higher permeability zone was present, it was necessary (o place the grout in lifts only 2 (0 5m thick, which further slowed the completion process. Mixing and pumping of non-standard grout mixes was also a problem (sce Section 7). Keeping sanded grouts in suspension required carefil control of water content, which was also necessary in the maintenance of the mixing and pumping equipment.02 So lg 8 8 3 2 so z a 100 Desiga’ | | Ref Line Ref. Line Ref. Line Ref. Line 50 1 Specs. DUB6-36(¢= 1.2 my DINE34(¢=24 m) 1196-24 (= 11m) DHI96-37 (r= 52m) Figure 11, Fines content (passing No. 200 sieve) relative to design specifications, and void ratio for drill holes located at various distances from Sinkhole No. 1 ‘The use of accelerators under variable temperature conditions caused particular challenges. ‘One behaviour that was not anticipated was the amount of bending and in some cases buckling that occutred in the slope inclinometer and other plastic. casing, and which led to problems with insertion of geophysical probes and the tube-d-manchette packer/ grouting unit. The buckling effect was most pro- nounced close to the sinkholes, but some effect was evident even in holes 10 (0 20'm away, fom which ‘we concluded that there were at least wo causes: © bending of plastic casing during installation and grouting, and + iaterat ground movement occurring after installation of the casing, due to later drilling or other reasons. ‘When the problem was recognized, two actions were taken: we switched to a larger plastic casing (to give greater clearance for the probes), and we modified procedures by keeping a string of drill steel ide the plastic casing during grouting, to reduce casing movement and bending during this process, ‘An_ important non-technical lesson was the necessity for clearly defined contracts with the drillers, even under near emergency response conditions, Particular challenges on this project included obtaining commitment for rigs when requirements changed daily, scheduling of erews when the drilling extended for long periods of time and/or sround the clock, retaining good drilling crews who understood the site, and cost inefficiencies related to use of short term contracts over a long period where production efficiencies 13 should reduce costs, In retrospect, after an initial period in which the site staff became familiar with the drilling method and the drillers became familiar with the site conditions, the sonic drilling method performed essentially as expected. With care, all objectives except items 2 and 5 (see Section 5.1) were met. Use of the sonic rig to set surface casing for some of the SCPTU holes confirmed the potential of this method {0 cause ground settlement at the sinkholes, and hence it should not be used where collapse of loose ground is anticipated. Despite our best efforts, the recorded dritling parameter data were not useful in assessing ground conditions, The parameters which ean be assessed for this method are very limited, but the technique does provide essentially a continous relatively undisturbed sample. ‘The vibrations can cause breakdown of coarser particles and this should be evaluated on a material and site basis 6. UNDISTURBED SAMPLING AND PRESSUREMETER TESTING 6.1 Selection of Drilling and Sampling Methods In assessing the processes responsible for the development of the sinkhole and for remediation design, it was important to obtain samples for gradation analyses and to measure the stress conditions within the sinkhole, Despite its merits, the SCPTU test could not reliably provide thisbe 3 sg ¢ 4 83 5 § ze gf Ee Ba ao g aa BE 24 0og8 G8 #8 dg 4g is 8283s¢e¢3e3 233 28 22.2% 2.98 ss i ss E = 60 60 i i ‘s ‘s Figure 12. Sonic drill parameters for DH96-22 (55-65m). information. Since the sonic drill was also_not suitable for this purpose (because of the potential for farther ground settlement and also because of poor sampling eapabilities in soft ground), alternate methods of obtaining soil samples were considered. ‘A technique for mud rotary drilling in disturbed core (described below) had been developed during SCPTU testing. At this stage in the investigation, mud loss in the sinkhole was less of a concen because it had been decided to remediate the sinkhole with compaction grouting. Mud rotary drilling was therefore selected for the sampling hole. Because of the variable ground conditions, several different samplers were tried: fixed piston sampler; heavy wall tube sampler with core catcher; side wall sanipter; HQ3 core barrel; and a “slush” sampler which had originally been used in the arctic for sampling soft ice and which was modified for use in soft ground. OF these, the shush sampler was the most successful and is described herein. Two pressuremeters were also brought (o site - a standard high pressure “insertion” model, and a low pressure, self-boring model. 6.2 Drilling, Sampling and Pressuremeter Test Procedures ‘The holes were drilled using a mud rotary rig, but 4 the drilling procedure was unconventional. + The upper portion of the hole (undisturbed core, above the water table) was collared, For the sampling hole, a sonic rig installed 194 mm OD casing (0 15 m depth. For the pressuremeter hole, & Barber rig installed 152 mm OD casing to 18 m depth, ‘© The sonie (or Barber) rig was moved off the hole and replaced by a Simeo $000 rotary rig. © PW casing was advanced into the disturbed dam core, alead of the drill bit, in 3 to 6m increments, using a combination of down pressure and rotation with no fluid circulation, «The inside of the casing was then drilled out with a tricone bit and mud circulation, ‘The drill bit was not allowed to extend beyond the bottom of the casing. © The fast (wo steps were repeated to advance the hole between sampling or pressuremeter test depths. ‘© When excessive friction developed on the outside of the PW casing, the drillers switched to HW casing and continued with the same procedure. Test hole DH96-36 was continuously sampled. It was drilled in November 1996 in sub zero conditions. Samples were taken frequently (every 1 to 2 m), mostly with the slush sampler affectionately known’as “Frostic", This sampler is similar to a Tong, very thick wall tube sampler, with an enlarged{ip (Figure 13). It was attached to BWL drill rods, Towered to the bottom of the casing, and pushed into the soil, A plug of soil at the bottom of the sampler was then fiozen by passing liquid CO, from the surface through a 6.4 mm diameter plastic supply fine into an expansion chamber at the tip of the sampler, and returning CO, gas through a similar line to the ground surface. Freezing of the plug typically required 2 to 4 minutes. Once freezing was complete, the CO, supply was shut off, the sampler was withdrawn from the hole, and the bottom of the sampler submerged in hot water to thaw the frozen plug. The sample was then extruded manually, logged, photographed and partitioned for gradation testing and water content determinations, Pressuremeter testing was carried out at drill hole DH97-1_ in eatly March 1997, A comparatively robust high-pressure pressuremeter was used for the first two tests because the pressuremeter operator ‘was concerned about damage to the membrane from coarse material, The limiting pressure at 10% strain was very low in the first test (less than 50 kPa). Because of its low sensitivity, the high-pressure tool was teplaced by the self-boring pressuremeter for the remainder of the tests. A total of 23 pressuremeter tests were completed over a depth range of 28 m to 91m. 6.3 Results ‘The dam core was successfully sampled in DH96-36 from a depth of 32 m to the bottom of the disturbed core at 109 m, using the slush sampler. ‘This feat itself was remarkable, and served as. an indication of the low stress conditions. ‘The lower, much harder and likely undisturbed dam core was drilled with an H1Q3 retractor core barrel from 110 m to bedrock at 115m, The disturbed dam core was tested with the pressuremeter in DH97-1, Liftoff’ pressures were very low, varying from 15 kPa at 28.5 m depth to 220 kPa at 90.7'm depth. ‘The following figures illustrate the sampling and test results: + plots of void ratio vs. depth and fines content vs, depth (Figure 11) for DH96-36; atypical pressuremeter plot (Figure 14); and + a plot of horizontal stress with depth (Figure 15). 6.4 Lessons Learned As anticipated, mud rotary drilling beneath the sinkhole proved to be difficult. ‘The most important consideration was hydraulic fracturing due to mud pressure. The drilling procedure provided some protection against hydraulic fracturing as long as the 15 Figure 13. “Frostie” sampler. Peessueemeter Test 12 BH97-L Depth: 50.0 m,, Elev: 631.7 m. ical f Static Water o 2 4 6 8 0 2 WW Radial Disptacement / Radius (%) 16 Figure 14. A pressuremeter result at Sinkhole No. 1 drill bit was inside the casing, but in almost every run, drilling fluid was lost when the bit approached within a few centimeters of the botiom of the casing. On several occasions there were reactions in nearby piczometers, possibly caused by hydraulic fracturing. Loss of mud in loose ground had been anticipated and accepted as an unavoidable consequence of obtaining relatively undisturbed samples. Itis possible to advance a mud rotary hole in soft ground using the above procedure without° TT Estimated initial » total vertical stress 3 J i 40 Ao i ay N E 100 + Estimated initial total lateral stress 120 0 200 400 600800 1000 Horizontal Stress (kPa) Figure 15. Horizontal stress versus depth from pressuremeter tests at DH97-1. hydraulically facturing the ground, as Long as a plug of soil is maintained at the bottom of the casing. However, if sampling or in situ testing is desired, it is necessary to clean out the entire casing which inevitably resulted in mud loss possibly due to hydraulic fracturing. ‘The second lesson learned was also an important one - that loose ground could be effectively and efficiently sampled with the “slush” sampler, permitting recovery of slightly disturbed material, Near continous samples are possible, although actual sample lengths may vary if ground conditions are variable, Minor problems that occurred when the sampler (designed for soft ice) was used in soil, were overcome with small modifications. ‘This type of sampler has application in many other geotechnical problems involving soils that are too sof for conventional sampling techniques. Grouting the mud rotary sampling hole was more difficult than the sonie holes. The main problem was high geout loss, despite the use of smualt lifts, grout accelerators, and even permitting, the grout to partially set inside the drill easing. ‘This was due to the extremely loose ground conditions around this hole. Significant time and effort were required 10 complete the hole, Some of the performance issues noted above were accepted in this specific application because remediation of the poor ground within the sinkholes was planned. Under other circumstances the performance of the mud rotary drill would not have been tolerable because of the lack of fluid control and the uncertainty of damage to the core. 16 7. CROSSHOLE TOMOGRAPHY 2.1 Objectives A variety of geophysical tests were used during the sinkhole investigation, The majority of these were deployed from the surface to search for potential defects in the upper part of the dam. ‘These surface methods met with very limited success. This section describes two downhole procedures that were used to investigate the dam core around the known sinkholes, over the full depth of the embank- ment. They are crosshole radar tomography and erosshole seismic tomography, both deployed from within cased drill holes. ‘Tomography, meaning measurement along a cut or section, is an imaging process originally devel oped for medical purposes. It allows the spatial distribution of some property within a test zone to be calculated from measurements made atthe boundaries of the zone. Drilling within the confines of the sinkholes was minimized to reduce the risk of precipitating another ground collapse. ‘Tomography, in theory at least, promised the ability to characterize the sinkholes tusing drill holes that were safely beyond the disturbed zones. 7.2 Basic Principles In crosshole tomogeaphy, a signal source is positioned in one cased drill hole and one or more receivers are located in another drill hole. Transmission of energy from source to receiver defines one ray path, Measurement of ray path, travel ime usually forms the basis of the procedure although in some cases other parameters may be measured as well, The discussion that follows is restricted (0 travel time; identical techniques apply for other parameters. Tn a typical application, a dense amay of overlapping ray paths, with a wide range of orientations, is generated, ‘The travel time for each ray path represents one piece of information, Considered individually, this information allows the average velocity along one ray path to be determined. A collection of arrival times from a large suite of ray paths, however, permits the variation of velocity throughout the measurement plane to be calculated. This is carried out by sub- division of the zone traversed by the rays into N cells. ‘The velocity in each cell is unknown but can be calculated from the data supplied by the N arrival times, Data processing is actually more complicated than this simple example suggests. A more detailed discussion of tomography can be found in Wong (1987).Output from a typical tomography analysis gives the signal propagation velocity in each of the cells, under consideration, ‘These results are contoured to produce a tomogram that shows the distribution of velocity in the test zone. The overall accuracy of a tomogram is a complicated fumetion of many factors: + drill hole separation; # number of ray paths; © number of cells in the tomogeaphy model; * number and accuracy of the boundary conditions; number of adjacent and intersecting planes analyzed simultaneously; + orientation of ray paths with respect (0 the target; + ‘wavelength of the energy; «velocity contrasts; * error content of the arrival time data; and * contouring procedures, Although some guidelines on the influence of these factors are available, their combined effects on the accuracy of the final product cannot be quantified, 7.3. Planning And Implementation Vattenfall HydroPower Ltd. of Sweden used crosshole radar tomography to image a sinkhole in ‘one of their embankment dams (Carlsten et al. 1995), On the basis of this case history, plans were made, shortly after the discovery of Sinkhole No. 1, to conduct a similar investigation at Bennett Dam. At this stage use of crosshole seisinie tomography was not anticipated. Examples documenting the successful use of the seismic technique under conditions similar to Bennett Dam were not known, Crosshole tomography requires cased drill holes of sufficient size to accommodate the transmitter and receiver tools. For radar measurements, the casing must be non-metallic and held in place with grout transparent to the signal, Standard geotechnical ‘grout used for drill hole backfill typically has a high ‘water content and also some bentonite. Bentonite’s Jow electrical resistivity causes strong attenuation of radar signals. A number of modified grouts, climinating the bentonite while attempting to maintain appropriate strength, stiffness and shrinkage charactetistics, were tested at Bennett Dam. Ultimately, a standard “machine base” cementitious grout was used, This gront had a high resistivity and so did not significantly attenuate the radar signals. Tomography requires that the start and end coordinates of each ray path be precisely specified, In crosshole applications this requirement can be satistied by installing grooved casing that ean accommodate a slope inclinometer probe. This 7 allows the casing’s position to be accurately determined. Original plans called for a ring of five to eight drill holes to be arrayed around Sinkhole No. | on a diameter of about 10 m. Crosshole tomography was to have been conducted between pairs of opposing Grill holes. This layout, ideal for imaging the sinkhole, was modified for two reasons - one logistical, the other technical, Sonic drilling in the yieinity of the sinkhole and SCPTU testing directly in the sinkhole had to proceed simultaneously because of schedule constraints. With two. and sometimes three rigs working in close proximity it was not possible to position all drill holes according to original plans, Secondly, preliminary testing results suggested that the ‘sinkhole might te surrounded by a zone of intermediate disturbance. This ted to a wider distribution of so called “step- out” drill holes. The resulting network of tomogram planes in the Sinkhole No, 1 area is shown on Figure 6. During the reservoir drawdown period, preliminary crosshole radar measurements were made in two planes to test the feasibility of the procedure, ‘This work involved three drill holes near the west abutment of the dam in an undisturbed part of the core, distant from Sinkhole No. 1 (the prototype test site). However, the decision to proceed with testing in the sinkhole areas was made, by necessity, before the preliminary radar results were available, In the interim, a timited field trial of crosshole shear wave measurements had been conducted in ovo of the drill holes at the prototype test site, The sali mechanical hammer sized to fit inside the plastic casing was capable of generating good quality signals, This result, together with general velocity testing experience, was considered sufficient to warrant full scale ctosshole scismie tomography at the sinkholes to complement the radar testing, 7.4. Crosshole Radar Tomography ‘The term radar is a mnemonic that refers to the use of reflected electromagnetic energy in the radio frequency band, (used despite the fact that in the crosshole tests the energy js transmitted, not reflected). The propagation of radar waves through soil is a function of dielectric constant, electrical conductivity and magnetic permeability of the material. The dielectric constant of most non-clay minerals is 5 to 7, while itis about 80 for water, AS a result, the bulk dielectric constant of soil is strongly affected by its water content, Assuming that other factors remain more or less constant, changes in radar velocity should correlate tovariations in water content (Carlsten et al. 1995). If the material is saturated, changes in water content will correspond directly to void ratio. ‘The fact that radar transmitters generate repeatable signals affords an additional opportunity for tomography. By comparing the amplitude of a signal at its source (o its amplitude at the receiver, attenuation can be determined, ‘Thus a ray path, in addition to having an arrival time, will also’ be tagged with an attenvation value in dB/mn, ‘Tomography works on any quantity that has spatial variability, so calculation of radar attenuation tomograms is possible. This may reveal a localized area within the test zone where signal attenuation is very high, ‘Radar tomography at Bennett Dam was not very successful. Although the sinkholes can be discerned in the tomograms, they are indistinct, largely because the radar velocity contrasts are small. In addition, inconsistencies in the results occur: one sinkhole appears as a lower, and the other as a higher, velocity zone. An identical pattern appears in the attenuation tomogeams. ‘The poor quality of the results raised doubts about the usefulness of radar tomography as @ diagnostic technique in this case. ‘The observed deficiencies likely arise from two factors. It appears that the correlation between radar propagation and water content, at least for Bennett Dam core material, is weak. Attempts to correlate radar velocities extracted from the tomograms with measured water contents from similarly located drill holes samples were not successful The combination of electrical and magnetic factors that affect radar propagation nevertheless produced fomograms that imaged the sinkholes. ‘That these results contained inconsistencies suggests that no single parameter exerts primary influence. Radar response appears to be a complicated function of poorly understood parameters that may or may not relate to mechanical soil properties. 7.5. Crosshole Seismic Tomography Crosshole seismic tomography, based on shear wave velocities, successfully imaged the sinkholes. An example tomogram, between DH96-24 and DH96- 26 passing directly through Sinkhole No. 1, is given in Pigure 16, ‘This shows a narrow vertical zone, with shear wave velocities as low as 150 ms. Undlisturbed core, in areas well removed from the sinkholes, was found to have a typical shear wave velocity of about $50 m/s. ‘The gap in the lower left hand comer of the tomogram is the result of a casing Jam that prevented a full sweep of measurements. The shear wave velocity tomograms correlated reasonably well with available geotechnical 18 50 $00 780 200 650 530 $00 +50 $00 50 ELEVATION (m) ‘SHEAR WAVE VELOCITY (mis) 230 p00 130 S70 0 5 10 15 20 DISTANCE (m) ‘gure 16, Seismic tomogram at Sinkhote No. 1. information and data obtained during compaction grouting of the sinkholes, In general, however, a tomogram can easily appear more authoritative than actually is. We noted, on successive reinterpretations of the data, that the final result could change considerably. ‘It is also possible to skew the results when they are being contoured, This is particularly true if colour contours are used, as visual emphasis can be cteated at arbitrary velocities. Monochrome shading lessens but does not eliminate this problem. ‘Areas of unusually low or high velocity in a tomogram often occur in areas where the ray pathcoverage is sparse. A tomogram should not be interpreted without reference to the ray path pattern upon which it is based. Alternatively, a ray path with an erroneous arrival time or incorrect geometry may be included in the data set inadvertentiy. This can have a sublle or dramatic effect upon the tomogram. Procedures to quickly assess the quality of the tomography data and allow errors to be corrected are not discussed here. The reader is referred to Wong (1987). 7.6 Lessons Learned Some of the lessons learned during the course of the geophysical investigations at Bennett Dam are listed below. Several are specific to erosshole tomio- graphy; others, while applicable 10 tomography, relate (0 geophysical investigations more generally. ‘+ Simultaneous analysis of multiple tomogram planes is essential for success. Drill holes should be located so that a maximum number of adjacent and/or intersecting planes can be included within a single three dimensional tomography model. * Slight casing bends can cause tools with minimal clearance fo jam. Simple procedures to straighten casings prior to and during growing should not be overlooked. * Information that can be used as boundary conditions in tomography analysis (e.g. downhole shear wave velocities) should be obtained and applied if practical. This can significantly improve the quality of the results. © Tomography works best when ray paths parallel to the boundaries of the target zone can be generated, * Bad data for even a single cay path can have widespread effects on the quality of a tomogram, Rigorous quality conteol of all data is required. + Crosshole tomography is slow, complicated and expensive. An experienced contractor is essential for success, Even then, independent review of all aspects of the work is necessary. This should include reproducibility checks of selected results, © Crosshole tomography makes no direct measurements within the test zone between drill holes, All results are inferred from measurements made at the boundaries of the test zone. A tomogram is best viewed as an approximate representation of actual conditions, Reasonable, but nevertheless non-wiique solutions, can be obtained. + Cate should be taken in producing contoured tomograms. Monochrome representations _ are recommended over colour to minimize opportunities for visually skewing interpretation of the results ‘+ Interpretation of a tomogram should not be done without reference to the ray path coverage for 19 that plane. Less credence should be given to results in areas where coverage is sparse. Analytical modelling of a proposed tomography program can give considerable insight into how successful the procedure may be. * Attempt to determine if a proposed geophysical technique responds with sufficient sensitivity to variations in the geotechnical parameter(s) of interest. «Prototype tests are valuable but may not reveal all of the potential problems that can occur in ‘wider application of a technique. * Geophysics is very data intensive. Rigorous data management procedures should be in place at the start of any investigation if inefficiencies and errors are to be avoided. * Communication between geophysical and geotechnical practitioners is difficult; considerable effort is required if misunderstandings are to be avoided, * Clearly defined requirements for data presentation and reporting should be specified in all contracting for geophysics work. * Like many fields, instructive failures are reported in the geophysics literature less commonly than the suecess stories, Even these, we suspect, may be coloured by over optimism, At a minimum, recognize that not all experience reported in the literature may be applicable to a new situation, 8. CHARACTERIZATION OF CORE BENEATH SINKHOLE 81 Geomeny The sinkhole geometry was interpreted trom cross sections of the SCPTU results, primarily the tip resistarice and the pore water pressure response. An example cross section is shown on Figure 8, Itis not immediately clear what the lateral limits of the sinkhole are, due to the spatial variation of the tip resistance Very low tip resistances (less than 1 MPa) were ‘encountered to a depth of 35 m and moderately tow values (1 {0 10 MPa) to a depth of 95 m, The latter are interlayered with zones of higher tip resistance and zones where the piezocone met refusual. There isa suggestion from the date reviewed in 3D thal the most disturbed zone is drifting downstream at depth (reference CPT96-5 and-6). The SCPTU results a short distance from the sinkhole (reference CPT96-3 and-4) show a variable tip resistance profile, but no significant or vertically continuous zones of low tip resistance. This is interpreted as a transitional disturbed zone between the intact core and the sinkhole.‘The results of the sonic drilling were used to extrapolate beyond the SCPTU locations to the intact core. Since the measured drilling parameters could not be correlated with density, the moisture contents from the samples were used to determine void ratio. Figure 11 shows the result for four drill holes at increasing distance from the benchmark tube at Sinkhole No.1. The reference line was calculated assuming an initial void ratio of 0.307, corresponding to the average fill density measured during construction, and using the compressibility of the core material when subjected to overburden stresses. It is clear ftom Figure 11 that void ratios deotease with distance from the sinkhole. For example, the average void ratio below a depth of 30 m is 0.38 in DH96-36 (1.2 m from the centre of the sinkhole), 0.33 in DH96-34 (2.4 m away), 0.27 in DH96-24 (11m away), and 0.26 in DH96-37 (52m away). The average void ratios obtained in these holes are representative of other holes at similar distances from the sinkhole. A linear interpolation of the results would suggest the edge of the disturbed zone at about 5 m from the benchmark tube. This is consistent with the results of the SCPTU which shows disturbed core at 4m from the centre of the sinkhole. Sinkhole No. 1 is a vertical to. sub-vertical feature with a downstream drift at depth. The diameter of the most disturhed material is 2 to 3 m to a depth of 90 m. There are some stiffer zones between the loose zones, which may be continuous, ‘or which may reflect a wander in the vertical sinkhole, Between the inner zone and intact core is a disturbed zone, with an outside diameter greater than 10 m and fess than 20 m, The seismic tomography results shown on Figure 16 support these dimensions, with a vertical column showing very Jow shear wave velocities (less than 200 nv/s) to depths in excess of 70 m with a diameter of 2 to 3 m. ‘The tomography also shows a downstream drift to the vertical feature, 8.2 Soil Properties The SCPTU data (lip resistance, friction ratio, pore pressure response and decay), the 24 pressuremeter test results, and the relatively high quality continuous “frostie” samples were used to assess the geotechnical properties of the sinkhole material. ‘The following summarizes the main characteristics. ‘© Very low stress conditions (less than 10% of the estimated initial total vertical stress) exist to 90 m depth (Figure 15). ‘© The material is very loose, with void ratios of 20 0.37 compared to intact core with void ratios of 0.25 (Figure 10). ‘© The fines content is essentially unchanged from the as-placed fines content (Figute 11). ‘The samples from the sinkhole showed the existence of “wet seams”, similar to those deseribed by Sherard (1973), throughout the full depth. These: seams were 25 mmm to 50 mm thick, randomly and frequently distributed, and were clearly visible as the samples were extruded from the sample tubes, ‘+ Based on the SCPTU pore pressure response, significant zones of “contractive” material were noted in the sinkhole to a depth of 35 n 9. CONCLUSIONS All the objectives of the investigation program were successfully met. ‘The sinkholes were defined and characterized to the extent required to select and design the remediation method, Along the way, and under conditions of severe schedule, technical, political and environmental stress, a great many lessons were learned, Perhaps the most important general lessons are as follows. ‘We did not discover any “good” way to drill into the core of this dam, under operating reservoir conditions. * For a project like this careful planning, meticulous attention (0 detail, high quality inspection, and detailed safety protocols are eritical. ‘* Prototype testing of new and innovative methods is very helpful, but one must remain vigilant for unforeseen problems which are certain to arise. ‘© Tn situ conditions can have dramatic influence on the suitability of various investigation methods, In this case the extremely low stresses in the sinkhole affected the most suitable geophysical tomography methods, permitted pushing the SCPTU (but influenced the interpretation of the SCPTU), and compounded the problems of drilling with fluids. ‘© Consider all risks with the investigation program and build risk mitigation strategies into all the work. 10. ACKNOWLEDGMENTS The success of the Bennett Dam sinkhole investigation was due to the dedication of a large number of individuals drawn from companies across Canada. In particular, the exemplary contributions of Mr. David Hill, P.Eng. of Thurber Engineering. Ltd. and his drilling investigation team are acknowledged. Likewise, Mr. Dennis Diggle of Foundex Explorations Lid. warrants specialrecognition for the innovative drilling and sampling. solutions he developed. Mr. David Woeller, P-Eng,, and his ConeTec Investigations Ltd. staff are largely responsible for the success of the SCPTU testing. ‘The erosshole seismic tomography would not have succeeded without the dedication and expertise of Mr. Patrick Lapointe, P.Eng. of Geophysics GPR Intemational. 11, REFERENCES Boncompain, B., Paré, J.J., and Levay, J. 1989, Crest’ sinkholes’ related to the collapse of loose material upon wetting. Proceedings of the 12! International Conference on Soil Mechanics and Foundation Engineering, Rio de Janeiro, Carlsten, 8., Johansson, 8, and Wérman, A, 1995, Radar techniques for indicating internal exosion in embankment dams. Journal of Applied Geophysics, 33: 143-156. Intemational Society of Soil Mechanies and Foundation Engineering 1989, International reference test procedure for cone penetration test (CPT). Appendix A, report of the ISSMFE technical committee on penetration testing of soils - TC16, with reference to test procedures, Swedish Geotechnical Institute, Linkdping, Lanne, T., Robertson, P.K. and Powell, J.M. 1997, Cone Penetration Testing in Geotechnical Practice. Bla Academic and Professional, London, 312 p. Ripley, C.F. 1967.” Portage Mountain Dam: 1. An outline of the project. Canadian Geotechnical Journal, Vol. 4, No. 2: 125- ML Sherard, J. 1973. Embankment dam cracking. In Embankment dam engineering papers in honour of Arthur Casagrande (Eds RC. Hirschfeld and S.J. Poulos). John Wiley and Sons, New York Stewarl, RA. and Imrie, AS. 1993, A new perspective based on the 25 year performance of WAC Bennett Dam. Proceedings of the International Workshop on Dam Safety Evaluation, Grindelwald, Switzerland, Vol, 1 Vick, $.G. and Bromwell, L.G, 1989. Risk analysis for dam design’ in karst. Jounal of Geotechnical Engineering, ASCE, 115. 819- 835. Wong, J. 1987. Crosshole seismic scanning and tomography. Geophysics: The Leading Edge of Exploration, January 1987; 36-41, aGeotectinical Site Characterization, Hobertson & Mayne (eds) 1998 Balkema, Rotterdam, ISBN90 54109394 Trends in geophysical site characterization John Greenhouse, Peeter Pehme, David Coulter & Quentin Yarie Hyd-Eng Geophysics, Mississauga, Ont., Canada ABSTRACT: Geophysical site characterization (GSC) is a rich and varied discipline. Sometimes misused, often misunderstood, it challenges the creativity and sciemtifie abilities of its practitioners in areas as diverse as archaeology, forensics, environmental assessments, hycrogeology, rock mechitnies and civil engineering. We review trends in GSC by identifying the major forces for change of the past decade and the responses of the GSC community (o those forces. The forces include the rise of environmentalism, the consequent increase in value of near-surface resources of all Kinds, and demands for higher resolution of those resources, Responses to these stimuli include restructuring of professional societies to begin (a integrate and engage the various sectors of the GSC community and their clients, and the development of methodologies, hardware and software to achieve the levels of resolution and visualization required. For the future we see these trends continuing, though perhaps with a shift in emphasis from the developed to the developing countries, There are (cemendous challenges and opportunities for research in many areas of GSC, including organic contaminant detection, infrastructure testing, fractured rock hydrogeology and novel non-invasive ways to probe the earth. | PREAMBLE part lack professional standards and licensing, The results of their work, simple representations of the Geophysics is a very broad subject. The dictionary _ complicated subsurface, are often difficult to explain definition ~ physics applied to planet Earth ~ and justify (0 a client, Perhaps as a result this encompasses a huge variely of scientific and profession lacks widespread credibility. engineering activities. Individuals who might refer to themselves as geophysicists work in areas ranging This paper is not a comprehensive overview of the from the evolution of the planets, through subject. We try simply to give some broader atmospherie physics, oceanography and hydrology, geophysical context to the technical papers that and ihe microstructure of pores. The boundaries follow in this session. We look first at the between geophysics and disciplines such as ‘organizational and technical evolution of GSC asa geochemistry, remote sensing, soil physics and whole over the past 10 to 20 years, We briefly geological engineering are both nebulous and describe the main geophysical methodologies, with transient, For example, is a survey carried out with an emphasis on recent developments, We shaw some Radarsat imagery geophysics? Is time domain survey examples that emphasize the breadth of the reflectometry (TDR) for soils geophysics? Are cone subject rather than its strengths. Finally we hazard peneironieters (CPT) fitted with seismic or electrical some guesses as lo where GSC is heading in the next sensors geophysical devices? few years. Geophysical site characterization (GSC) is also a ‘Our opinions reftect our point of view of the Tich and vatied field, and one providing tremendous profession. We work for a small consulting challenges to the small group of engineers and company specializing in environmental and scientists whose profession it is. This group, hydrogeological applications. Our “bread-and- however, has problems, They are split amongst butter” work is in support of Phase I audits, mainly several professional organizations and for the most leaking tanks and groundwater contaminant plumes, 23Bedrock and overburden mapping, pipeline construction surveys, locating voids beneath office floors or golf courses, mapping unexploded ordinance, identifying graves (usually ancient but occasionally modem), and a host of other small and usually very interesting jobs fill our schedule. 2 INTRODUCTION GSC can be defined as a reconnaissance survey of site, undertaken preliminary to more detailed and usually more intrusive investigations. The geophysical survey images the subsurface in terms of physical properties averaged over comparatively large volumes. Compared (o geochemical sampling or visual core analysis, undertaken on very small volumes of rock, geophysics has advantages of scat, cost and speed. Disadvantages include lack of resolution compared to point samples (but such precision can be deceptive in an earth which is almost always hielerogeneous), and the fact that the data obtained are at least one step removed from the information required. The measured electrical resistivity must be interpreted in terms of porosity, permeability or contamination levels; seismic velocities must be interpreted in terms of rock or elastic moduli, and so on . GSC is undertaken on many lateral scales, ranging from centimeters to hundreds of meters. “Geophysical” site characterization seems to start al depths of « few centimeters and extend to a few tens of meters. Remote sensing for example, with its very minimal penetration , is not usually thought of GSC method, neither is TDR testing of soils (although these distinctions are very arbitrary). While “sites” can extend to hundreds of meters in somie cases, most targets are shallow. GSC activities can be crudely subdivided into three main areas. Two of them are quite traditional: testing (soils, foundations, pavement ete.); and engineering (geotechnical characterization, mechanical properties, rippabitity, elastic moduli, depth to bedrock, basic gealogy, fracturing, voids, tunnels detection, ctc.), The third, environmental, is a bit of a catchall word. Used strictly, it refers to work performed as part of environmental assessments required by regulatory agencies. Increasingly it is used in a broader sense that includes surveys undertaken for anything that pertains to “the environment”, including groundwater quantity and quality, buried hazards, and so forth. We prefer this broader definition because it identifies geophysics 24 © Environmentat GSC WORK @ Enginsering ® Testing Exploration) jeophysicist GSC PRACTITIONERS Figure 1, Schematic showing the various sectors of the GSC community thydrogeology, civil engineering, geological engineering, eophysics, otiers) and the areas of work (testing, engineering, environmental) in which they are likely to specialize. swith some of the major issues of the day, but the result isan ambiguous boundary between what is ermed environmental and what is termed engineering, GSC professionals tend to fall into sectors which have verged on being solitudes. Again generalizing broadly, they are explorationists who would refer to themselves as geophysicists and whose roots are with the major exploration geophysical societies, and ‘geophysical practitioners within other professions, who would more likely refer to themselves as (for example) a civil engineer who does geophysics. These practitioners can be grouped into sector civil engineering, engineering geology and hydrogeology. ‘These four sectors have been. somewhat isolated by virtue of their training, their professional affiliations and preferred journals and congresses, and also somewhat marginalized by the affiliated. The general subdivision of GSC and its practitioners is shown schematically in Figure | Several factors have conspired to change the shallow geophysics industry in the past 10 to 15 years . Some of these are briefly described below. 3 FORCES FOR CHANGE a. There has been a steady increase in the perceived value of near-surface resources such as groundwater, soils, aggregate and archaeological remains. This reevaluation either presaged or followed on from thepublic's new preoccupation with all things environmental. ‘The increasingly litigious nature of North America has dramatically increased the importance of subsurface information, and the value of non-invasive technologies for obtaining it, b. The collapse of the mining exploration sector in the fate 1 980°s (not unrelated to environmentalism) left many exploration geophysicists and geophysical instrument manufacturers looking for work, Some found it in the environmental market. ©. Groundwater contamination, perhaps the dominant environmental issue for the geophysical community, continues (o challenge our industry. The conductive, inorganic plumes of the 80’s were comparatively easy to locate, In the 90's the hydrogeological emphasis has shifted to organic contaminants, a much more difficult target. 4, The growth and - more importantly for us - the decay of infrastructure, highways, bridges, parking lots, etc, roughly parallels population growth and how provides tremendous opportunity for fast and reliable GSC methods that can map concrete and pavement. fe, As military bases are closed down, the need for non-contact methods for clearing these areas of ‘unexploded ordinance (UXOs) and chemical contamination of the ground has grown dramatically, Landmines, although a rather different problem from UXOs, beg for a non-invasive identification technology, £. New-contexts for geophysics brought new sites and new targets. Sites included landfill industrial waste lagoons, heavily urbanized (and contaminated) neighbourhoods, inside buildings, ‘major intersections, airport runways, ex-armry bases, and artillery ranges, New targets leachate plumes, industrial solvents, diesel fuel, steel drums, tanks, mine tailings, unexploded ordinance, and leaking ponds and barriers. These demanded new approaches. 8. The increasing value of near-surface resources to society, coupled with the public’s increasing familiarity with computer technology, has raised the expectation of clients. Specifically they want higher vertical and horizontal resolution, image-quality presentations and even three dimensional representation of their targets. A generation raised on computer games like Quake is unlikely to settle for plane-layered model interpretation! 25 h. Characterizing these sites and targets jreasingly required measurement and presentation in three dimensions of space, and in time, Anthropogenic processes like the migration contaminants in groundwater, or the remediation a landfill, have time scales more nearly human than geologic. Increasingly itis the temporal development of the target, rather tan its location, that has to be. characterized, i, Environmental science and engineering has blurred the traditional boundaries between disciplines (physics, geology, biology, civil engineering, etc) to which we had become accustomed . Environmental projects foster a team approach, in which professionals from many disciplines meet regularly and on site. J. Stringent safety procedures are enforced for hazardous sites, These, along with concepts like quality assurance and just-in-time data production, have considerably altered the geophysical working environment. k. Geophysics has enjoyed an unfortunate reputation as an inaccessible subject, probably due to (a) the rigour of physics compounded by (b) a certain air of mysticism often associated with its interpretations, In the team-oriented environmental area there is a healthy trend towards more open dissemination of methods and their limitations. 1. Truly phenomenal progress in the processing speed, data handling capabilities and miniaturization ‘of computing clearly overprint all these other factors, as they have in all professions. Only slightly less impressive are the advances in positioning. A decade ago perhaps half the time required for an electromagnetic (EM) survey could be taken up by positioning stations, and half the time for processing, n coordinating the survey with local maps, other surveys, etc. GPS and GIS technology clearly is changing that. 4 ADAPTING How has the GSC community responded to these trends? Without trying to uniquely identify cause and effect in each ease, here are some of the changes we have noted. 4.1 institutional. ‘The past decade has seen the opening of a what