The following paper has been submitted to GeoManitoba 2012, the Canadian Geotechnical Conference to be held in Winnipeg in Sept/Oct 2012, as a draft for consideration. The first author, John Danielson, has given permission to post this draft here.
Potential Development at Dublin Gulch, Yukon: Part 2 – Engineering Challenges
J. Danielson1, P. Quinn2, K. Hanley1, D. Welkner1, T. Urquhart3, L. Toussaint4, W. Newcomen5, and G. Hensold1
BGC Engineering Inc., 1Vancouver, British Columbia; 2Victoria, British Columbia; 3Halifax, Nova Scotia; 4Kamloops, British Columbia, Canada
Dublin Gulch, located in the heart of central Yukon, is the site of potential development of an open pit heap leach gold mine. The area has an unusual geological history, possibly unique in a Canadian context. Interesting local geological characteristics, described in a companion paper, lead to a number of engineering challenges. This paper describes some of the more important engineering challenges associated with planned earthworks construction, including: provision of high quality borrow; foundation and cut slope design; construction on ice-rich permafrost, and, bedrock rippability. The paper also discusses the various methods employed to understand and solve these challenges, including detailed reconnaissance and field mapping, drilling and in-situ testing, geophysics, and plate load testing.
Situé sur Plateau de Stewart dans la zone centrale du Territoire du Yukon, Dublin Gulch est un projet de mine d’or a ciel-ouvert en cours d’étude et développement incluant la lixiviation sur le tas comme méthode de séparation. Le faciès local est le résultat d’un passé géologique complexe, probablement unique dans le contexte canadien. Les conditions géologiques particulières au site, décrites plus en détail dans un article associé, impliquent de nombreux défis géotechniques dont certains des plus importants sont liés aux travaux de terrassement proposés; en particulier: l’approvisionnement en matériau granulaire d’emprunt de haute qualité, l’étude de stabilités des déblais et de fondations, la construction sur pergélisol à glace excédentaire et l’aptitude des roches à la taille et à la scarification. Cet article décrit par ailleurs les diverses méthodes employées pour comprendre et résoudre ces défis en incluant la cartographie de terrain, la cartographie géologique, le forage, les essais in-situ, les études géophysiques et les essais de charge sur plaque.
Dublin Gulch, Yukon, is the site of potential open pit heap leach gold mine development. The site’s unusual geological history presents a number of interesting challenges for development, and these are discussed in a set of two companion papers. This paper, Part 2, describes some of the more important engineering challenges and the methods used to solve them. The companion paper, Part 1, describes the interesting engineering geological setting that leads to the existence of these challenges.
The property is located at the confluence of Haggart Creek and Dublin Gulch on the Stewart Plateau, approximately 40 km north of Mayo, and 15 km northwest of Elsa, YK, as illustrated in Figure 1. The site has an unusual geological history, with few comparable settings elsewhere in Canada. Old Proterozoic to Lower Cambrian aged sedimentary rocks have been modified by tectonic processes and altered by regional and contact metamorphism. Unlike most of Canada, the area has not been affected by recent continental glaciations, so a variably thick zone of weathered rock occurs near the surface. Weak, highly disturbed bedrock has been modified by gravitational displacement and periglacial processes. Ice-rich permafrost is widely present, and valley bottom sediments have been disturbed extensively by anthropogenic activity.
The complex geological setting creates several important engineering challenges which have been explored through detailed geotechnical site investigations. The following sections describe some of the more important geological and geotechnical challenges for development, with discussion of the methods used to investigate and solve these challenges.
Provision of Engineering Borrow
Development of the proposed mine will involve sourcing and production of several different important engineering materials, including sand and gravel for engineered fill, sound rock for concrete aggregate, silt or clay for low permeability liners, and durable rock for riprap and rock fill. Sand and gravel are relatively abundant, located primarily in anthropogenic deposits of placer tailings in the Dublin Gulch and Haggart Creek valley bottoms. Much of this material needs to be removed to provide a suitable foundation for the heap leach pad, therefore its re-use as an engineering material is desirable. The primary difficulties in re-use of the placer tailings as engineered fill relate to constructability: selection and processing of similar materials that can be placed and compacted with normal quality control efforts; and, de-watering of the valley bottom to facilitate efficient removal.
An example of coarse placer tailings with relatively significant proportion of cobbles and boulders is shown in Figure 2. In this specific pile, a 75 mm grizzly screen would remove perhaps 10 to 15 % as oversize materials, and the finer materials passing through the screen would be suitable for general use in different applications. The oversize clasts tend to be strong, derived largely from granodiorite or quartzite. The weaker site rocks, including schists and phyllites, have generally broken down into finer components. Therefore, sand and gravel derived by screening can only be used in applications where the potential for further deterioration is not critical, and should not be used as structural fill for support of heavy vibratory loads, for example. Approximately 2 million cubic metres of exploitable placer tailings are potentially available for re-use from sources in the Dublin Gulch and Haggart Creek valley bottoms within a reasonable haul distance of proposed earthworks. These have been classified and grouped by similar grain size as illustrated for the Dublin Gulch valley bottom in Figure 3.
A relatively small quantity of silt or clay is required for construction of mine infrastructure, and is primarily intended for use in low permeability barriers. Adequate quantities of silt appear to be available in alluvial and till deposits; however these are generally frozen and contain excess ice and excess organic content in places. An example of recently excavated frozen sandy silt, with significant organic content, and containing excess ice, is shown in Figure 4. During the geotechnical site investigations, the excavation contractor determined the most efficient excavation method to be a bulldozer with single ripper tooth working in tandem with an excavator fitted with a hydraulic breaker, and the excavator performed most of the work. Care would be required in silt borrow development to allow the frozen silt to thaw and drain as necessary in time for re-use in construction.
The most significant challenge in borrow development is identification of suitable sources of durable rock, for use as concrete aggregate, rock drains, riprap, and general rock fill. Non-durable rock, which is abundant around the site, can be used as a substitute for durable rock fill in some applications; however, it needs more care in placement and compaction, therefore durable rock within a reasonable haul distance is preferred. The volume of concrete aggregate required for facilities construction is relatively small ( 100 blows/0.3 m) within 1.2 to 1.5 m of the ground surface in two tests. Therefore, while the material is a very weak rock, it behaves as a very strong soil, with inferred shear strengths of 45 degrees or better at low confining stresses.
Geophysical testing at 11 boreholes indicated increasing shear wave velocities with improvements in rock mass quality. The mean shear wave velocities for Types 3, 2 and 1 rock were 845, 1080 and 1145 m/s, respectively. The shear wave velocity The most significant challenge in borrow development is identification of suitable sources of durable rock, for use as concrete aggregate, rock drains, riprap, and general rock fill. Non-durable rock, which is abundant around the site, can be used as a substitute for durable rock fill in some applications; however, it needs more care in placement and compaction, therefore durable rock within a reasonable haul distance is preferred. The volume of concrete aggregate required for facilities construction is relatively small (< 20,000 cubic metres); however, it is important that these materials yield acceptable concrete that will perform for the design life of proposed facilities.
Figure 5 shows a typical exposure of poor quality, highly weathered metasedimentary rock that does not yield durable rock when excavated. These materials may be excavated and re-used as non-durable rock fill, requiring placement in thin lifts, watering and compaction control, similar to an earth fill. They are therefore of use for engineering applications, but cannot be placed as efficiently as a durable rock fill, which may be placed in thicker lifts, and can be placed in winter.
Durable rock for rock fill, and potentially for concrete aggregate, may be derived from strong quartzite or granodiorite in the intrusive ore body or its surrounding hornfels aureole. Figure 6 shows strong but highly jointed rock at the contact between the granodiorite intrusion and surrounding country rocks. These materials, when excavated, will yield an angular rock fill suitable for many purposes, and can be crushed to produce potential aggregate. Selection of concrete aggregate requires additional care, considering other factors beyond the scope of this paper, but these materials are potentially suitable, as are the coarser clasts from the placer tailings which are illustrated in Figure 7.
Large durable rock blocks are required for certain applications, including riprap lining of important drainage structures and rock drains below waste rock storage areas. Massive strong rock, which would yield large blocks on blasting, is very rare in the area. It may be necessary to quarry these materials from within the area of the proposed open pit.
Design of Facilities
Development of the proposed mine will involve construction of major fills with heavy static loads, including the heap and waste rock storage areas, as well as erection of buildings and equipment with vibratory loads, including a series of three conventional crushing facilities. Site investigation for selection of foundation design parameters has been hampered by a number of challenges related to site geology. The first challenge has been to select drilling methods that will yield sufficient amounts of engineering data. Drilling activities can be divided broadly into those that are primarily focused on exploring rock conditions, and those that are focused on overburden soils.
Rock drilling has been completed by diamond drillers engaged by the owner to complete exploration drilling. Equipment and tooling were modified to suit geotechnical investigations. This included use of triple tube coring gear (most often HQ3, some PQ3), with provision to complete in-hole installations including standpipe or vibrating wire piezometers and thermistor strings. Drillers were also equipped and trained to do packer testing, and to retrieve oriented core from inclined boreholes.
Even with very careful drilling, and experimentation with different drilling fluids, drilling run lengths, bit types and other factors, recovery of rock tended to be very low, due to the foliated, jointed and friable nature of much of the rock. Recovered rock core lengths were often very short, therefore it was difficult to select representative core samples for field point load testing and laboratory testing for unconfined compressive strength (UCS) and Brazilian tensile strength. Rock mass rating (RMR, Bieniawski 1976) parameters were obtained for all rock, where possible, to estimate rock mass strength and stiffness for foundation design. A core orientation tool was used in selected inclined boreholes to obtain orientation information for bedrock structural discontinuities. This effort was largely unsuccessful in shallow holes, as the core tended to move around in the core barrel, due to the typically low recovery and closely spaced joints along foliation, leading to frequent development of easily displaced disks. A typical example of rock core retrieved is shown in Figure 8.
Overburden drilling met with different challenges. The overburden holes were drilled using both solid stem and hollow stem auger drilling techniques. Various bit types and CRREL barrel coring techniques were employed in selected holes. Standard penetration testing (SPT) was attempted in all formations, but was frequently unsuccessful, due either to the presence of cobbles and boulders or frozen ground. Poor results were obtained in the placer tailings, till, colluvium and weathered bedrock, each of which contains a substantial proportion of coarse clasts that can interfere with SPT testing (Figure 9). Of 77 attempted SPT tests, 46 yielded valid results. The other test results were set aside due to incomplete recovery (and the associated influence of gravel or cobbles plugging the sampler), refusal in frozen ground, or refusal on large clasts or bedrock. Energy measurements were conducted on a small number of SPT tests to calibrate the testing system for conversion of “N” values to (N1)60 values. A small number of dynamic cone penetrometer tests were also made. A number of tests were also conducted with a hand-held wildcat penetrometer, which has been used successfully by the authors on other projects; however, meaningful correlations could not be obtained between SPT and wildcat results for the formations encountered at this site. The wildcat test results were therefore used only for general context to support engineering judgement.
SPT blowcounts are generally considered to be one of the more reliable sources of information for deriving engineering parameters for granular materials, and since these were sparse, other testing was necessary. Consideration was given to other overburden drilling methods, including Becker hammer and sonic drilling. Both methods remain under consideration for possible future use, but in the 2011 investigation the lack of SPT data was compensated for, to some degree, by geophysical investigations (downhole seismic and seismic refraction) and plate load test data. The geophysical testing and plate load tests were compared with visual observations from test pits, retrieved drill core and mapping of exposures in an effort to classify the site materials into engineering categories on the basis of easily observable characteristics. These classifications were then used to determine engineering parameters for use in geotechnical design, including shear strength and stiffness.
For the purposes of foundation design and slope stability analysis, three rock types were defined. “Type 3” rock is usually the first “rock-like” material underlying the overburden soil materials. Type 3 rock is defined as being rock that is highly or less weathered (i.e. weathering grade W4 or better), with an intact strength greater than R0 (i.e. minimum UCS strength 1 MPa). “Type 2” rock is defined as rock with Geological Strength Index (GSI, Hoek and Marinos, 2000) or Rock Mass Rating (RMR, Bieniawski, 1976) of 30 or greater, and core recovery during drilling of 50 % or greater. Alternatively, where GSI and RMR data are unavailable, average Rock Quality Designation (RQD) of 10 or greater is taken as an approximately equivalent criterion. “Type 1” rock is defined as having GSI, RMR or average RQD exceeding 40. Completely weathered bedrock, which does not meet the minimum criteria for “Type 3” rock, is treated as a soil.
A relatively significant proportion of the material encountered at the planned foundation grades for various facilities was classified as either Type 3 rock or as a soil-like highly to completely weathered rock. It was therefore important to develop confident estimates of material properties for these very weak rock masses for use in foundation design.
Hoek and Diederichs (2006) and Hoek and Marinos (2000) provide some guidance on expected strength and stiffness of weak rock masses like those encountered at the site; however, local site-specific data were necessary to establish confident estimates for design. These were obtained for the weakest rock masses through:
– SPT, DCPT and wildcat penetrometer testing in the highly to completely weathered rock;
– Downhole geophysical testing (shear and compression wave profiles) across all formations; and
– Plate load testing on “Type 3” rock and highly to completely weathered rock.
The highly to completely weathered rock was weak enough to permit auger drilling, using a combination of both solid stem auger flights and CRREL barrel coring. Penetration testing in this material met practical refusal (> 100 blows/0.3 m) within 1.2 to 1.5 m of the ground surface in two tests. Therefore, while the material is a very weak rock, it behaves as a very strong soil, with inferred shear strengths of 45 degrees or better at low confining stresses.
Geophysical testing at 11 boreholes indicated increasing shear wave velocities with improvements in rock mass quality. The mean shear wave velocities for Types 3, 2 and 1 rock were 845, 1080 and 1145 m/s, respectively. The shear wave velocity data were used to estimate small strain shear moduli, and these small strain stiffnesses were used to provide an initial estimate of larger strain elastic stiffness for use in settlement calculations, by reducing to about 5 to 20 % of the small strain value. The results of plate load testing (discussed below) were used to increase confidence in the estimated stiffness values.
Measured compression (p) wave velocities showed no clear relationship, with mean p-wave wave velocities of 2850, 2700 and 2900 m/s for Types 3, 2 and 1 respectively. This lack of relationship is believed to be associated with issues with grout stiffness in the annulus between the drillhole casing and the rock/soil formations. Efforts were made to match grout stiffness with the average expected formation stiffness, but the results suggest the grout may have been stiffer than the weaker rock masses, and similar to that of Type 1 rock, therefore the p-wave velocity data are tentatively interpreted to overestimate stiffness in Type 2 and 3 rock. This conclusion is supported by the results of the plate load testing.
A total of 14 plate load tests were conducted at three different test areas, with two tests in highly to completely weathered rock, and one test in Type 3 rock. Tests were conducted with four different diameter circular steel plates: 0.3 m, 0.53 m, 0.76 m and 1.2 m. Most tests were conducted to the maximum reaction load, which was provided by a 25t jack and D9H bulldozer (48t rear axle load). Deflections were measured using three dial gauges attached to 4 m long horizontal reference beams mounted outside the zone of plate influence. Figure 10 shows the typical test set up.
Plate load test results were plotted to produce displacement-pressure curves, from which stiffnesses can be inferred. A typical result is provided in Figure 11. Each test included a load-unload cycle to a nominal working load, followed by loading to jack capacity.
The subgrade reaction modulus, kv, was calculated as the slope of the line tangent to the section of each test curve most reflective of ground behavior after immediate settlement due to some nominal seating load. The rock mass deformation modulus, Em, was then estimated from the relationship:
E_m=〖πd/4 k〗_v (1-μ^2 ) 
Where d is the diameter of the plate used during testing, and µ is Poisson’s ratio.
Em values estimated from plate load testing range from 72 to 152 MPa in Type 3 rock and 15 to 116 MPa in completely weathered rock.
In the completely weathered zone plate load testing, geophysics, auger and diamond drilling were completed in close proximity, providing a broader context for developing engineering judgement from the test results. No material was recovered during diamond drilling in the tested stratum. Recovery from auger drilling consisted of dense sand with some silt, interpreted as completely weathered rock disturbed during the drilling process; signs of oxidation were noted. Geophysical testing indicated an average shear wave velocity of about 150 m/s in the region near the ground surface affected by testing.
Comparison of the various data sources, including borehole logging, penetration testing, downhole geophysics, field mapping, plate loads tests and available literature allowed the interpretation of deformation modulus of approximately 60 MPa, 100-500 MPa, 1000-2000 MPa and 2000-3000 MPa for highly to completely weathered rock, and Type 3, 2 and 1 rock, respectively. These values were recommended for use in design, subject to additional testing at later stages of investigation.
Hoek-Brown strength parameters were also derived for the three rock types, conforming to typical ranges of strength corresponding to the indicated stiffness for phyllite or schist. These parameters were derived by first compiling all of the rock strength and joint strength data from point load testing (i.e. Is50 values) and field classification of all core samples and discontinuities. Statistics were generated for all available data, and mean values were selected for design.
Engineered slopes include both constructed slopes built from engineered fill, and cut slopes excavated into in-place geological materials. Design of major fills for long and short term static and dynamic stability, including design of the heap and waste rock storage areas, is a complex topic, outside the scope of the present paper. This section will discuss engineering challenges associated with development of cut slopes.
The project site has considerable relief, with elevations ranging between about 800 m in valley bottoms to over 1400 m in the uplands. Steep slopes are also common, with slope angles on the hillsides typically ranging from about 15-20 degrees to well over 30 degrees. Development of roads, and construction of level pads for crushers and other facilities, will therefore necessitate the development of numerous significant cuts, some as high as about 100 m. Planned cuts will intersect a wide variety of materials, including placer tailings (i.e. loose, variable fill), colluvium, till, weathered rock, and intact bedrock. In many places, the overburden will contain permafrost, which usually contains excess ice. Slope design will depend on groundwater (and ground ice) conditions, as well as the shear strength of the affected materials. Selection of reasonable shear strength parameters for most of the overburden materials is relatively straightforward, although limited to some degree by the lack of SPT blowcount data or other direct in-situ strength testing. Nevertheless, shear strengths for the various overburden materials have been inferred from various data sources, including the limited SPT data set, downhole geophysics, plate load testing, wildcat penetrometers testing, and visual classification.
The more significant challenge for cut slope design was in the selection of appropriate strength parameters for rock, since bedrock will be exposed in most of the larger cuts, and since rock mass quality tends to be unusually low. It is important to distinguish between cuts that are expected to be governed by rock mass strength (or soil shear strength parameters, where appropriate), and those where failure is expected to be structurally controlled.
Selected cut slopes will be excavated either entirely in overburden, or in highly to completely weathered rock that is treated as overburden in analysis and design. In these cases, the slopes are analysed assuming general failure to be controlled by typical shear strength throughout the soil or rock mass. Most cuts, however, will involve significant proportions of stronger rock (i.e. Type 3, 2 or 1) where failure may be controlled by strength and orientation of discontinuities.
Shear strengths of existing discontinuities can be inferred from field interpretations of rock core, and some information about orientation can be inferred from joint angles in relation to core axis. However, most geotechnical drilling outside the open pit area has involved vertical holes, where three-dimensional orientation of structures is not possible. A small number of inclined holes were drilled, and attempts were made to orient the core to derive joint orientations. As discussed previously, these efforts were largely unsuccessful in relatively shallow rock, to depths of interest for foundation design. In order to supplement these data, a number of existing rock outcrops were examined. A total of 55 outcrops were mapped, as illustrated in Figure 12. These included both natural outcrops and man-made cuts, and included a wide range of lithology and rock mass quality, covering most conditions expected during site development activities.
Compilation of discontinuity data from the outcrop mapping program allowed the development of assumed shear strength properties, and also supported the development of a structural domain map. This map has been used to infer the typical orientation of foliation and other joints sets in areas where site-specific data are not available, allowing interpretation for preliminary designs. Such preliminary interpretations can be refined through site specific data gathering in subsequent phases of investigation.
Construction on Ice-Rich Permafrost
The project site is in a region of widespread discontinuous permafrost, and ice-rich frozen ground is widely present, being observed at varying frequency in different areas of the site. Figure 13 shows the location of subsurface observations where frozen ground was present or absent in test pits or boreholes. Where frozen ground is present, it usually (i.e. 90 % of the time) contains excess ice.
Where ice-rich permafrost is present, earthworks activities can require extra effort if conducted during the seasonal thaw period. Typical thaw-related instability during summer construction is shown in Figure 14. Care is required to ensure effective drainage in order to minimize the duration of poor trafficability if excavation and disturbance of ice-rich ground must be completed in summer/fall. An alternate approach to minimize trafficability issues would be to excavate ice-rich ground in the winter.
Two additional challenges associated with development in permafrost deserve brief mention. First, the frozen ground is more difficult to excavate than unfrozen ground, and may require ripping. Second, excavated ice-rich materials require additional care in disposal or management. Stockpiles of ice-rich material will have very low strength during thaw and drainage, and will therefore not stand at steep angles in spoil piles.
The effort required to excavate natural materials in site development is an important factor in planning and cost estimation. The weak quality of most bedrock at the site is expected to be advantageous for site development, due to its ease of excavation. Shallow bedrock at the site is often excavatable by normal excavating equipment, including tracked excavators or bulldozer blades. This is illustrated in Figure 15, which shows test pits excavated to the limits of excavator reach in weathered bedrock.
It is important to develop an understanding of equipment requirements for site preparation activities, including bulk excavation. It is evident that much of the shallow rock on site is excavatable, but selected outcrops contain competent rock that may require blasting. Some effort was given to attempt to define where common excavation, ripping or blasting would be required, to support capital cost estimating. Four methods have been used in this assessment, each of them relying on different rock description parameters collected during the site investigations. The four methods include:
Caterpillar performance handbook (2011) rippability charts based on measured compression wave (p-wave) velocities;
Pettifer and Fookes (1994), based on fracture spacing and point load index;
Kirsten (1982), based on Barton’s Q index and field interpretation of rock hardness (R values); and
Tsiambaos and Sarglou (2010), based on GSI, fracture spacing and joint condition.
These four different methodologies each approach the issue of rippability from a different perspective, and they yield different results from available site data, sometimes conflicting with each other. A consensus estimate of rippability has been obtained by comparing the results of all four approaches, and weighing site specific observations of machine effort during access road, drill pad and plate load test pad construction.
Based on the four methodologies employed in this work, it is considered reasonable to propose an interpretation of the excavatability conditions in the rock mass within the area of the proposed infrastructure development. Note that numerous assumptions and extrapolations are intrinsic to this assessment, so there are some uncertainties in the associated conclusions. The main conclusions of this assessment are as follows:
Approximately 10% of all rock mass will require blasting, and 50% will require ripping;
Approximately 40% of all rock mass can be excavated by common excavation equipment;
Highly to completely weathered rock can be excavated;
Type 3 rock will most likely be excavated with up to 40% requiring ripping;
Type 2 rock will most likely require ripping, with a portion of up to 35% that could be excavated, and potentially the occasional need for blasting; and
Type 1 rock will mostly require ripping, potentially hard ripping (D10T bulldozer and low production), with probably 10-20% requiring blasting.
The validity of this assessment will be determined during construction, and will be further examined in subsequent stages of investigation. This interpretation leads to a potentially practical method for assessing subgrade conditions for foundations. Where original rock fabric is evident and the rock can be excavated by easy to moderate digging with an excavator, the material is highly to completely weathered rock and can support light foundations. Where hard digging, or possibly easy to moderate ripping is required, Type 3 rock or better is present, and is capable of supporting light to moderate foundations. Where hard ripping is necessary, Type 2 or better rock is present, and is capable of supporting moderate or heavy foundations. Where blasting is required, Type 1 rock is likely present, and is capable of supporting heavy foundations. These interpretations are generalizations, and subject to further investigation and confirmation during construction.
The interesting geological history at Dublin Gulch leads to numerous engineering challenges for earthworks construction at the site. Several of these have been discussed, including: provision of high quality engineering borrow; foundation design; cut slope design; construction in permafrost; and, rippability of bedrock. Site investigation methods had to be adapted to the challenging site conditions to obtain sufficient information to support confident design. Traditional drilling methods were supplemented by detailed site reconnaissance and field mapping, downhole and surface geophysics, and plate load testing.
The writers would like to acknowledge Victoria Gold Corporation for permission to use its data in development of this paper.
Bond, J.D. 1998. Surficial Geology of Dublin Gulch, Central Yukon. Geoscience Map 1998-6. Indian and Northern Affairs Canada, Exploration and Geological Services Division, Yukon Region.
Caterpillar TC Performance Handbook. 2011. Edition 41, Rippers Section, pp. 74 – 79.
Hoek, E. and Diederichs, M.S. 2006. Empirical estimation of rock mass modulus. International Journal of Rock Mechanics and Mining Sciences, 43: 203–215.
Hoek, E. and Marinos, P. 2000. Predicting Tunnel Squeezing. Tunnels and Tunnelling International.
Kirstern, H.A.D. 1982. A classification system for excavation in natural materials. Civ eng S Afr 24, pp. 293 – 308.
Pettifer, G.S. and Fookes, P.G. 1994. A revision of the graphical method for assessing the excavatability of rock. Q J Eng Geol 27, pp. 145 – 164.
Smit, J. Sieb, M. and Swanson, C., 1995. Summary Information on the Dublin Gulch Project, Yukon Territory. In: Yukon Exploration and Geology, 1995, Exploration and Geological Services Division, Yukon Region, Indian and Northern Affairs Canada, p. 33-36.
Tsiambaos G, Sarglou H. 2010. Excavatability assessment of rock masses using the Geological Strength index (GSI), Bull Eng Geol Environ 69, pp.13 – 27.