<<

kChapter 22

STEVEN G. VIcK

HYDRAULIC TAILINGS

1. INTRODUCflON 2.CHARACTERIZATION AND PROPERTIES

ailings, defined as sand-, silt-, or clay-sized For most metalliferous , tailings represent the Tsolid from -processing opera- end product of crushing and grinding operations tions that were originally produced, handled, and after extraction of the valuable mineral particles. deposited in slurry form, represent a special case of For other ores, such as , coal, or con- slope stability in that they are artificially con- struction aggregates, tailings are produced by sim- structed deposits of manufactured . Tailings are ple washing and screening operations. In either distinct from other types of coarse wastes case, the maximum particle size typically varies or smelter that are dumped or piled under dry from 0.1 to 1.0 mm, with the relative proportions conditions. Tailings deposits share many geotech- of sand- and silt-sized particles governed by the nical characteristics of such other hydraulic fills as. type of processing required for optimum mineral offshore drilling islands, hydraulic-fill water , extraction efficiency. Mechanical grinding pro- dredged material deposits, and certain other indus- duces particles that are highly angular; clay is con- trial wastes (e.g., and papermill sludge). tained in the tailings only if present in the parent These related materials are not discussed here, but , and clay content is usually significant only for pertinent information can be found in two publi- tailings produced by washing operations. Specific cations from the American Society of Civil gravity of tailings solids varies from as low as 1.5 Engineers: Geotechnical Practice for Disposal of Solid for coal to almost 4.0 for tailings derived from sul- Materials (ASCE 1977) and Hydraulic Fill fide-type orebodies where is rejected during Structures (Van Zyl and Vick 1988). processing. Vick (1990) summarized gradation and Tailings are often identified by their association index properties for tailings from a variety of ore with mining and milling operations. Where types. encountered along transportation routes, tailings deposits may affect the stability or safety of these 2.1 Construction of Tailings facilities by (a) causing instability of excavated cuts, Impoundments (b) creating potential landslide hazards from adja- cent deposits, (c) increasing subgrade compressibil- During operation of a milling facility, the tailings ity and settlement, and (d) possessing adverse slurry is usually pumped to a tailings impound- environmental characteristics (high erodibility, ment at a pulp density of 15 to 50 percent solids by. heavy- , and acid drainage). weight. Within the impoundment the solids settle

577 578 Landslides: Investigation and Mitigation

from suspension, and the supematant water collects des carried farther in the slurry stream. This sedi- in a decant from which it is discharged or recy- mentation pattern gives rise to an above-water cled to the process. Central to geotechnical beach of sand tailings (defined as those with less characterization of any tailings deposit is the man- than 50 percent passing the No. 200 sieve) and a ner in which discharge and deposition of the tail- zone of slimes (more than 50 percent passing the ings slurry have been conducted, and conventional No. 200 sieve) extending from the more distant exploration and sampling techniques alone cannot regions or the above-water beach to beneath the substitute for an understanding of these operations. decant pond, as shown in an idealized manner in Usually the tailings, slurry is discharged from Figure 2 2-2. The degree of particle-size segregation the embankment perimeter, typically as shown in on the beach, and therefore the extent to which Figure 22-1. To the extent that a sand fraction may the sand tailings and slimes are differentiated, de- be present in the slurry, these particles tend to set- pends on such factors as gradation of the original tle nearest the point of discharge, with finer parti- slurry and its solids content (Abadjiev 1985). The

FIGURE 22-1 Tailings slurry discharge and Spigots beach deposition Tailings discharge pipe (modified from Vick 1990). Decant pond - — — OMM FONS

Embankment

Dike Decant Spigotted Tailings Tailings Beach\

Dike

FIGURE 22-2 _ Sequential tailings raising and tailings deposition showing idealized sand and slimes regions (modified from Vick Slimes 1 Sands 1990). Hydraulic Tailings 579 transition between zones of sand tailings and slimes 2.4 Typical Strength Characteristics in a tailings deposit is gradual, interlayered, and in- distinct and is governed by the various techniques Strength characteristics governing slope stability and locations of tailings slurry discharge adopted vary according to sand or slimes characterization. during operation of the impoundment. Never- For most types of sand tailings, the drained (effec- theless, characterization of a tailings deposit into tive-stress) friction angles typically range from 33 to regions of predominantly sand and slimes is funda- 38 degrees. These relatively high values for loose mental to establishing broad ranges of physical, materials are attributable to particle angularity. index, and engineering properties and to targeting Sand tailings display virtually no effective cohesion subsurface exploration in key areas. intercept (c') except in special cases involving chemical cementation, such as gypsum tailings or tailings affected by oxidation of sulfide 2.2 Sand Tailings Characteristics (usually pyrite). Vick (1990) extensively summa- Sand tailings derived from beach deposition are rized drained strengths for various tailings types. usually loose, with relative densities ranging from Provided that the nature, origin, and deposi- 30 to 50 percent. Upon initial deposition, tional characteristics of sand tailings deposits are they typically exhibit corrected standard penetra- fully understood, estimates of drained strength tion test (SPT) blowcounts, expressed as (N1 ), of based on published values for similar tailings types 3 to 5. There is evidence to support some increase are often sufficient for effective -stress analyses. in (N1) values with time due to "aging" effects Slope stability is seldom sufficiently sensitive to (Troncoso 1990). ordinary variations in drained strength to justify extensive laboratory testing. For slimes tailings, either drained or undrained 2.3 Slimes Characteristics strength may be pertinentto slope stability, depend- Slimes, on the other hand, qualify as low-plasticity ing on such factors as the degree of desiccation- to nonplastic silts for tailings derived from induced overconsolidation that the slimes may grinding of metalliferous ores. Other ores and have experienced during deposition. For slimes processes, most notably Florida phosphatic ores, ex- derived from metalliferous ores, drained (effective- hibit less favorable plasticity, sedimentation, and stress) angles of internal friction are generally in the consolidation characteristics in their clay slimes range of 28 to 37 degrees (Vick 1990). Slimes (Bromwell and Raden 1979; Schiffman et al. 1988). exhibit little or no effective cohesion. A void ratio of 0.8 to 1.3 is typical of many slimes The undrained strength (s) of slimes is, how- deposits, with (N1) values often ranging from 1 to ever, sensitive to details of the original deposition 3. Of particular significance to slope engineering process. For slimes deposited within and beneath issues is the high specific retention of void water the decant pond, normally consolidated conditions exhibited by slimes. Once sedimentation and con- prevail after dissipation of deposition- induced solidation processes are complete, further gravity excess pore pressures, and the variation of drainage seldom results in appreciable reduction in undrained strength with depth can be character- water content or gain in strength except in the zone ized by stress-normalized ratios (s /a') ranging affected by surface desiccation, which seldom from about 0.20 to 0.27 (Vick 1990; Ladd 1991). extends below depths of 1 to 2 m. Thus, slimes In contrast, slimes deposited above water or on deposits, even in dry climates, have remained soft intermittently saturated portions of the beach or and virtually fully saturated for up to 60 years after otherwise influenced by local desiccation during deposition has ceased and surface water has been deposition may exhibit far more irregular un- removed. This feature, along with the associated drained strength patterns, which may be difficult low undrained shear strength, makes slimes the to assess (Maclver 1961). Although recovery of most troublesome tailings material from the point undisturbed samples of saturated slimes for labora- of view of slope engineering or embankment con- tory testing is possible using fixed-piston samplers, struction. Slimes deposits also exhibit high com- vibration during sample or sample extru- pressibility, although rates of consolidation are sion frequently causes gross disturbance and densi- usually relatively rapid. fication. Draining of sample tubes in the field 580 Landslides: Investigation and Mitigation

before transport and freezing of recovered samples apply to tailings deposits, but more important than have both been used successfully to reduce these the computational procedure is the adoption of effects. Alternatively, Wahier and Associates appropriate strength parameters. Ladd (1991) pro- (1974) and Maclver (1961) provided examplesof vided a thorough discussion of effective stress analy- undrained strength assessment for slimes by field ses versus undrained strength analyses in actively vane-shear testing. However, thin sand seams are operating tailings impoundments and demon- often present, especially in beach transition zones, strated that pore pressures due to undrained shear- and these may make drainage conditions during ing must be addressed in assessing slope stability. vane-shear testing difficult to ascertain. In con- These findings apply to excavated slopes in tailings junction with conventional sampling, the cone slimes as and emphasize the previously dis- penetrometer test (CPT) is very useful for identify- cussed importance of stress history and undrained ing interlayering and other stratigraphic details of strength resulting from the original circumstances tailings deposits. of slimes deposition. Equal in importance to calculating stability of 3. SLOPE STABILITY AND cuts in tailings deposits is evaluating the feasibility ENVIRONMENTAL ISSUES of excavating them. The typically dry and even dusty surface of abandoned tailings deposits may Slope stability problems related to tailings deposits produce a misleading impression. Although sand most commonly arise in connection with dams that tailings eventually dewater by gravity drainage given contain the deposited slurry, and a rich literature on sufficient time and foundation permeability, even tailings-dam stability exists, including work by perched water remaining in sand layers can produce Morgenstern (1985), Kealy and Soderberg (1969), "running" conditions on excavated faces. The same Soderberg and Busch (1977), Klohn (1972), Vick is true for slimes, which may remain fully saturated (1990), and Corp et at. (1975). Brawner (1979) de- only a few feet below the desiccated surface crust scribed several failures and repair case histories. long after the impoundment has been abandoned. Seismic stability of tailings dams is of considerable Both saturation and low undrained shear strength of importance because of the liquefaction susceptibil- slimes seriously impede the mobility of wheeled con- ity of tailings deposits; examples of current practice struction equipment during excavation, and costly in dynamic analysis were provided by Finn et al. and time-consuming dragline or backhoe excava- (1990), Troncoso (1988), Lo et al. (1988), and Vick tion techniques must often be used. These problems et al. (1993). have been avoided during large-scale remining of Although tailings-dam design and analysis tailings by hydraulic excavation (McWaters 1990), methods apply in principle to transportation- but this requires that a new impoundment be related slope engineering situations involving tail- constructed to retain the excavated tailings slurry. ings, important differences exist. Chief among these are the following: 3.2 Use of Tailings as Fill Route alignments on or through tailings deposits Tailings from abandoned deposits have been con- most often involve excavated cuts, sidered for use as highway fill. Dry sand tailings Tailings may be considered as a borrow source have been found to have suitable compaction char- for embankment fill, and acteristics and engineering properties. Pettibone Tailings deposits are inactive and abandoned and Kealy (1971) presented typical engineering rather than actively operated in conjunction properties for compacted tailings fill and described with ongoing mining operations. the use of tailings borrow for construction of a 6-km 3.1 Stability of Excavated Cuts section of Interstate-90 near Kellogg, Idaho. However, in the construction of highway fills Stability of excavated cuts in inactive tailings using tailings, their highly erosive nature must be deposits is ordinarily controlled by the cohesion- considered (see Section 3.4). A roadway fill con- less nature of the sand and the undrained strength structed from tailings at Silverton, Colorado, was behavior of the slimes. Conventional limit- rapidly eroded when a culvert designed to carry the equilibrium procedures of slope stability analysis flow of a small stream became blocked with debris FIGURE 22-3 Erosion of highway fill constructed from

Silverton, Colorado, May 1985. Road embankment was • , . . destroyed in a few o ourswienculvert

va igh U MU Fill acted as - - . temporary dam but 'c. . - was quickly eroded - - - -i --- a' once flowing water overtoppedit ilk Reconstruction of roadway with new - - -'---• culvert took several - - ;,f1' - • - weeks COIRTESY OEThE ' -

•'• H tr'' ; - -

-- fill constructed from

of ;ct cas dis - ly ins EEr for former tailings discharge pipes are visible on slopes. c • - hehind tailings and - .- - - accessto it was disrupted by this - . failure * - - - riu- Mi\ER 582 Landslides: Investigation and Mitigation

during a spring storm (Figures 22-3 and 22-4). The Slope erosion notwithstanding, eroded material blockage was not detected during the night and the from adjacent flat tailings surfaces can affect trans- stream overflowed the fill and rapidly eroded it. portation facilities. Tailings eroded by water can Once started, the erosion occurred so rapidly that quickly plug ditches and culverts, and wind erosion nothing could be done to prevent it. Reconstruc- has been known to cause visibility-related safety tion took several weeks. This road was not a major hazards in dry climates. The rapid erosion of a high- highway but did provide the only access between an way fill constructed from tailings near Silverton, active gold mine and its mill as well as the only Colorado, was mentioned in Section 3.2. This rapid feasible access to recreation sites in a mountain val- erosive failure resulted from a blocked culvert, ley. The erosion of the road fill thus caused some which dammed a small stream and caused it to measurable economic hardship. overtop the highway fill. The adjacent active tail- ings disposal site is shown in Figure 22-4. This site 3.3 Toxicity of Tailings was constructed with comparatively steep slopes. Wind erosion on these slopes produced large dust Use of tailings as highway fill or disposal of exca- clouds under relatively low wind velocities. The vated tailings may present environmental issues re- steepness of the slopes hampered revegetation lated to toxic liability. Potential toxicity of tailings efforts. With the recent final closure of the mine depends on the geochemistry of the orebody and and mill, these slopes are being regraded to lower the nature of the metallurgical processing and can- angles and revegetation is beginning to help stabi- not be generalized even for ores of the same basic lize these tailings. Such revegetation of exposed type. Although probably the majority of tailings tailings surfaces can be complicated by the absence deposits are chemically innocuous, some tailings of nutrients, poor textural characteristics, and may contain high levels of such constituents as sometimes by the high residual metal concentra- arsenic, lead, cadmium, fluoride, or molybdenum, tions in the tailings, which may be toxic to plants. which can be harmful if ingested by humans, Successful revegetation of large areas of tailings grazing animals, or certain plants. Such exposure often requires specialized studies, including field can be enhanced when disturbed tailings are dis- test plots. Smaller areas can be quickly revegetated persed by wind. The solubility, mobility, and poten- by covering tailings surfaces with topsoil, although tial for transport of many heavy within this may be quite expensive. groundwater or surface-water environments are greatly enhanced by the low pH levels generated by 4. FAILURE OF TAILINGS IMPOUNDMENTS sulfide oxidation of some abandoned deposits, espe- cially those rich in pyrite, pyrrhotite, or marcasite Slope failure and breaching of actively operating minerals. Exposure of such tailings surfaces to the tailings dams are frequently accompanied by atmosphere accelerates oxidation and acidification flowing of the saturated tailings deposit. These through complex chemical and biological reac- flows have produced extensive damage and loss of tions. Gadsby et al. (1990) provided an overview of life in the past, but damage to adjacent trans- recent research and practice in this area. portation facilities has historically been limited to temporary closure of railroads and highways. 3.4 Erosion Susceptibility of Tailings Perhaps the greatest potential hazard may be posed by certain types of actively operating tailings dams Environmental issues also derive from the highly where flows triggered by seismic failure may affect erosive nature of tailings slopes, and special mea- lifeline transportation facilities such as highways or sures may be necessary to reduce and control sed- pipelines and impede postearthquake disaster iment transport. Because of their relatively fine and response and recovery efforts. Vick (1991) pre- uniform grain size, tailings are erodible by water sented methods of estimating flow runout distances and wind. Slopes as flat as 3H:1V are frequently in the event of tailings embankment failure. These considered necessary for erosion control and estab- procedures have been used to address the impacts of lishment of vegetation. Erosion considerations potential tailings flows on adjacent highways for rather than mass stability may ultimately be found existing and proposed tailings dams in Utah and to govern slope design for tailings. Montana (Vick, unpublished study). Hydraulic Tailings 583

The potential for flow failures from abandoned REFERENCES tailings deposits is less clear but is the topic of cur- rent research (Troncoso 1990). Dobry and Alvarez ABBREVIATIONS (1967) studied many tailings deposits subjected to strong earthquake ground motion in Chile and ASCE American Society of Civil Engineers docui'nented a number of liquefaction failures and USCOLD U.S. Committee on Large Dams flows from actively operating tailings impound- Abadjiev, C. 1985. Estimation of the Physical Charac- ments. However, with inactive deposits, slope teristics of Deposited Tailings in the effects were limited to deformation and cracking. of Nonferrous . In Proc., 11th Inter- No flow failures are known to have occurred for national Conference on Soil Mechanics and Founda- abandoned tailings deposits that did not retain tion Engineering, San Francisco, A.A. Balkema, impounded water, under earthquake loadings or Rotterdam, Netherlands, pp. 1231-1238. otherwise (USCOLD 1994). ASCE. 1977. Geotechnical Practice for Disposal of Solid Nevertheless, in one recent case involving two Waste Materials. New York, 885 pp. abandoned tailings deposits in Missouri that Brawner, C. 1979. Design, Construction, and Repair were constructed using unusual tailings deposi- of Tailings Dams for Metal Mine Waste Disposal. tion procedures and that retained surface water, In Current Geotechnical Practice for Mine Waste dynamic analyses and flow-failure runout assess- Disposal, American Society of Civil Engineers, ments were performed to evaluate the potential New York, pp. 53-87. for inundation of a downstream community Bromwell, L., and D. Raden. 1979. Disposal of resulting from tailings-dam failure triggered by a Mining Wastes. In Current Geotechnical seismic event (Vick et al. 1993). It was found that Practice for Mine Waste Disposal, American Society a proposed highway embankment downstream of Civil Engineers, New York, pp. 88-112. from the dams could be designed to retain a pos- Corp, E.L., R.L. Schuster, and M.M. McDonald. sible flow and thus substantially reduce the haz- 1975. Elastic-Plastic Analysis of Mine-Waste Embank- ard to downstream residents. Thus, if potential ments. Report of Investigations 8069. Bureau of failure hazards from tailings dams are evaluated Mines, U.S. Department of the Interior, 98pp. and recognized in advance, judicious planning Dobry, R., and L. Alvarez. 1967. Seismic Failures of and layout of transportation facilities can mitigate Chilean Tailings Dams. Journal of the Soil ASCE, Vol. 93, their effects. Mechanics and Foundation Division, No. SM6, pp. 23 7-260. Finn, W., M. Yogendrakumar, R. Lo, and R. Ledbetter. 5. MITIGATION MEASURES 1990. Seismic Response Analysis of Tailings Dams. Proc., International Symposium on Safety and Construction of cuts and fills on or through 'tail- In Rehabilitation of Tailings Dams, International ings deposits requires addressing stability, sub- Committee on Large Dams (ICOLD), Sydney, grade compressibility, construction feasibility, and Australia, BiTech Publishers, Vancouver, British environmental factors. The characteristics of tail- Columbia, , pp. 7-33. ings deposits can be less favorable than those of Gadsby, J., J. Malick, and S. Day (eds.). 1990. Acid natural in most of these respects. Various soil Mine Drainage: Designing for Closure. BiTech improvement methods have been considered for Publishers, Vancouver, - British Columbia, tailings deposits, many of which were summarized Canada, 495 pp. by Mitchell (1981, 1988). However, the applicabil- Kealy, C., and R. Soderberg. 1969. Design of Dams ity of virtually all such techniques is liiuited to for Mill Tailings. IC 8410. Bureau of Mines, U.S. specific characteristics of soil, saturation, or both. Department of the Interior, 49 pp. No single technique has been found universally Klohn, E. 1972. Design and Construction of Tailings feasible for both sand and slimes tailings under all Dams. Transactions, Canadian Institute of Mining conditions. Avoidance of tailings deposits during and Metallurgy, Vol. 75, pp. 50-66. route selection, with due consideration for the Ladd, C. 1991. Stability Evaluation During Staged risks that adjacent impoundments may pose, Construction. Journal of Geotechnical Engineering, therefore remains a primary mitigation strategy. ASCE, Vol. 117, No. 4, pp. 537-615. 584 Landslides: Investigation and Mitigation

Lo, R., E. Klohn, and W. Finn. 1988. Stability of Bureau of Mines, U.S. Department of the Interior, Hydraulic Sandfill Tailings Dams. In Hydraulic 136 pp. Fill Structures, Geotechnical Special Publication Troncoso,J. 1988. Evaluation of Seismic Behavior of 21 (D. Van Zyl and S. Vick, eds.), American Hydraulic Fill Structures. In Hydraulic Fill Society of Civil Engineers, New York, pp. Structures, Geotechnical Special Publication 21 549-572. (D. Van Zyl and S. Vick, eds.), American Society Maclver, B. 1961. How the Soils Engineer Can Help of Civil Engineers, New York, pp. 475-491. the Mill Man in the Construction of Proper Troncoso, J. 1990. Failure Risks of Abandoned Tailings Tailings Dams. Engineering and Mining Journal, Dams. In Proc., International Symposium on Safety May, pp. 85-90. and Rehabilitation of Tailings Dams, International McWaters, T. 1990. Developing Magma's Tailings Committee on Large Dams (ICOLD), Sydney, Leach Operation. Mining Engineering, Sept., pp. Australia, BiTech Publishers, Vancouver, British 1075-1080. Columbia, Canada, pp. 34-47. Mitchell, J. 1981. Soil Improvement: State of the Art Van Zyl, D., and S. Vick (eds.). 1988. Hydraulic Fill Report. In Proc., 10th International Conference on Structures, Geotechnical Special Publication 21, Soil Mechanics and Foundation Engineering, Stock- American Society of Civil Engineers, New York, holm, Sweden, A.A. Balkema, Rotterdam, Nether- 1068 pp. lands, Vol. 4, pp. 509-565. Vick, S. 1990. Planning, Design, and Analysis of Tail- Mitchell, J. 1988. Densification and Improvemet of ings Dams. BiTech Publishers, Vancouver, British Hydraulic Fills. In Hydraulic Fill Structures, Columbia, Canada, 369 pp. Geotechnical Special Publication 21 (D. Van Zyl Vick, S. 1991. Inundation Risk from Tailings Flow and S. Vick, eds.), American Society of Civil Failures. In Proc., IX Pan-American Conference on Engineers, New York, pp. 606-633 Soil Mechanics and Foundation Engineering, Viña Morgenstern, N. 1985. Geotechnical Aspects of del Mar, Chile, Sociedad Chilena de Geotecnica, Environmental Control. In Proc., 11th Interna- Vol.3, pp. 1137-1 158. - tional Conference on Soil Mechanics and Founda- Vick, S., R. Dorey, W. Finn, and R. Adams. 1993. tion Engineering, San Francisco, A.A. Balkema, Seismic Stabilization of St. Joe State Park Tailings Rotterdam, Netherlands, pp. 155-185. Dams. In Geotechnical Practice in Dam Rehabil- Pettibone, H., and C. Kealy. 1971. Engineering Prop- itation (L. Anderson, ed.), Geotechnical Special erties of Mine Tailings. Journal of the Soil Mechanics Publication 35, American Society of Civil Engi- and Foundation Division, ASCE, Vol. 97, No. SM8, neers, New York, pp. 402-415. pp. 1207-1225. USCOLD (Tailings Dam Committee). 1994. Tailings Schiffman, R., S. Vick, and R. Gibson. 1988. Behavior Dam Incidents. Denver, Colorado. and Properties of Hydraulic Fills. In Hydraulic Fill Wahier and Associates. 1974. Evaluation of Mill Structures, Geotechnical Special Publication 21 Tailings Disposal Practices and Potential Dam Sta- (D. Van Zyl and S. Vick, eds.), American Society of bility Problems in Southwestern U.S.—Investigation Civil Engineers, New York, pp. 166-202. Report, Kennecott Copper Corp., Magna Tailings Soderberg, R., and R. Busch. 1977. Design Guide for Dam, Magna, Utah. Bureau of Mines, U.S. Depart- Metal and Nonmetal Tailings Disposal. IC 8755. ment of the Interior, NTIS PB-243 076, 226 pp.