Cone Penetrometer and Geophysical Methods
Andrew I. Strutynsky
Robert A. Glaccum
Linda C. Conklin
RADIAN International LLC
ABSTRACT: Naturally occurring brines (concentrated solutions of chloride salt and water) are pumped from deep wells and are transmitted by pipeline to industrial processing facilities. An accidental brine release affected shallow groundwater quality at a site in Michigan. The chloride brine plume was delineated using non-intrusive surface geophysical, minimally intrusive cone penetrometer, and downhole geophysical methods. Surface electromagnetic surveys were used to map the areal extent of the plume. Direct push cone penetrometer soundings with soil electrical conductivity measurements were used to profile stratigraphy and detect electrical conductivity anomalies associated with brine intrusions. Anomalous zones were sampled, using penetrometer groundwater sampling, for direct chemical analysis. Geophysical monitoring stations were then installed in boreholes to allow periodic monitoring of the plume's response to remediation.
Soil electrical conductivity (EC) measurements are uniquely useful during geo-environmental exploration of brine contaminated groundwater. Both surface and downhole methods can be used to monitor increases in EC resulting from the migration and diffusion of brine through groundwater. Surface methods have the utility of rapidly covering large areas to identify where brine intrusions might have occurred. Downhole methods are very useful in pinpointing the vertical extent of contamination as well as in estimation of brine concentrations. Downhole measurements can be obtained in cased boreholes, or through use of cone penetrometer soundings which include EC measurements.
Soil electrical conductivity is controlled by the conductance of the soil particles and the conductance of the fluid occupying the soil pore spaces. The ratio between pore fluid and combined soil-pore fluid electrical conductivity is termed the formation factor (Archie, 1942). Clay particles can be electrically conductive due to adsorbed water and ionic electrical charges on the clay platelets, so clay electrical conductance depends on mineralogy, porosity and pore fluid characteristics. Sand grains are typically nonconductive, so sand conductance depends primarily on pore fluid characteristics and porosity. The addition of brine to groundwater greatly increases soil electrical conductance.
A brine groundwater contamination exploration method was developed based on several technologies. First, surface electromagnetic (EM) surveys can be used to rapidly map affected areas, and give reasonably good indications of the degree of impact of brine releases on groundwater, to depths as great as 60 m (200 ft). However, surface EM methods lack the ability to adequately quantify brine concentrations, cannot provide detailed definition of the depth intervals affected by brine intrusions, and cannot easily differentiate between natural electrically conductive clayey soils and zones of elevated chlorides.
Downhole geophysical induction logging can be performed to great depths in suitably cased boreholes. Data quality is high, allows vertical delineation of affected intervals, and allows estimation of brine concentrations. However, drilling and casing boreholes, and the disposal of large volumes of exploration derived wastes from drilling activities, are expensive and time consuming. Further, borehole stability can be a significant problem where thick sequences of saturated sands are encountered.
A high capacity, truck mounted, geo-environmental cone penetrometer system, including downhole EC logging and penetrometergroundwater sampling, can be used to vertically profile subsurface conditions and obtain direct groundwater samples. The penetrometer exploration method is fast and accurate, and provides high resolution lithologic and electrical conductivity data. The direct push penetrometer exploration method requires no borehole, so no soil cuttings are generated, and borehole stability is of no concern.
The electromagnetic (EM) geophysical method determines electrical properties of earth materials by inducing electromagnetic currents in the ground and measuring the secondary magnetic field produced by these currents. An alternating current is generated in a wire loop or coil above the ground's surface. Both the primary magnetic field (produced by the transmitter coil in the instrument) and the secondary field (produced by currents in the earth) induce a corresponding alternating current in the receiver coil of the instrument. The coils are kept at a fixed distance and orientation relative to the ground to simplify data analysis.
After compensating for the primary field, both the magnitude and relative phase (in-phase and quadrature) of the secondary field are measured. The quadrature-phase component, using simplifying assumptions of homogeneous and isotropic conditions, is converted to a value of apparent soil electrical conductivity (EC). This value represents an estimate of the local average soil EC. The depth of measurement is dependent on the instrument's coil spacing, orientation, and operating frequency, and the actual subsurface EC variations. Averaging limits discrimination of thin, high concentration brine intrusions from broader, more diffuse plumes. Multiple profiles using differing coil spacing can be performed to bracket approximate depths of brine affected groundwater. Data quality may be degraded by cultural interference as caused by utility lines, steel fences or other large metallic objects.
Surface electromagnetic measurements were taken by Geosphere with Geonics EM38, EM31 and EM34 systems, using coil separations of 1 m (EM38) , 3.7 m (EM31), and 10 m or 20 m (EM34). The nominal explored depth is proportional to coil spacing. For the shallow, near field EM38 instrument, the explored depth is about 1.5 m; for the intermediate EM31 it is about 6 m; and for the deep, far field EM34 instrument, the explored depth is about 15 m with a 10 m coil spacing, 30 m with a 20 m spacing, and 60 m with a 40 m spacing. Instruments were carried manually, and measurements were digitally logged during profiling at the chloride brine release site. Shallow and intermediate readings in the source area were typically taken at 15 m (50 ft) intervals along lines 15 m apart using the EM38 and EM31 instruments. EM34 (deep far field) coverage was conducted on lines spaced 60 m (200 ft) apart in the western portion of the plume, where the chloride brine had migrated to much greater depths.
Permanent geophysical monitoring stations were installed using hollow stem auger drilling techniques. A water filled, PVC casing was grouted into place to provide an access conduit for lowering downhole geophysical logging tools. Screened intervals are not necessary (or even desired) as there is no need for direct communication with the surrounding groundwater/soil system. In induction logging, the field is induced in the surrounding soil through the casing itself. Electrically conductive (i.e., steel) casing materials cannot be used.
Periodic downhole logging of the geophysical monitoring stations was performed at the chloride brine release site using a Geonics EM39 induction system. The induction technique has been used for decades by the petroleum industry for formation characterization in oil and gas wells. The Geosphere application of these methods to shallow groundwater exploration includes increased resolution due to slim, compact logging tools (3.6x130 cm or 1.4x50 in) and slower logging rates. This permits the detection and resolution of formation conductivity changes across vertical intervals of less than 0.3 m (1 ft). The EM39 induction logger is capable of measuring with an accuracy of 2 mS(milli-Siemen)/m in the range of 0 to 1000 mS/m.
The geophysical monitoring station provides significant advantages over traditional screened monitoring wells at brine affected sites. A continuous EC profile is obtained from the surface to the bottom of the cased hole, where traditional monitoring wells only provide data across the screened interval. This is significant as brine plumes typically sink with time, retreating from screened intervals, which can lead to misleading results. The geophysical monitoring station avoids the costs of installing monitoring well clusters, and eliminates data gaps which occur where monitoring wells are screened in different portions of the aquifer. Monitor well purging and analytical testing are also eliminated, leading to additional cost savings.
Like surface EM and penetrometer CPTU-EC methods, the borehole EM39 system measures the bulk (soil and pore fluid) EC. With suitable assumptions of formation factors, a value of groundwater conductivity, and thus approximate concentration of a predominant ion such as chloride, can be estimated from downhole EC measurements.
CONE PENETROMETER SYSTEM
The STRATIGRAPHICS cone penetrometer system is used during geo-environmental and geotechnical exploration programs in difficult soil conditions. The heavy (240 kN and 300 kN, or 24 and 30 ton) truck mounted rigs are fully self-contained, and include data acquisition systems, dry and wet work areas, water tanks, steam cleaners, decontamination and grouting systems, separate rodstrings for sounding and sampling, optional dynamic rod driving, and heavy duty downhole equipment for use in glaciated terrains (Figure 1).
Cone penetrometer (direct push) systems require no borehole to advance probes and samplers and result in very little exploration derived waste. Downhole equipment is decontaminated during retrieval using an automatic rodwasher, and the open hole is pressure grouted. Most exploration activities are performed inside an enclosed portion of the rig, providing all-weather capability and a low visual profile during operations. Truck mounted penetrometer systems can be very productive, with as much as 400 m (1300 ft) of stratigraphic logging per day, with depth capacity exceeding 60 m (200 ft). As many as 18 groundwater, or up to 30 soil or soil gas samples, can be acquired in a day.
Soil resistance to penetration acting on the tip and soil friction on the sides of the penetrometer are separately measured during CPT. These measurements are accurate and repeatable, and have been used for the evaluation of stratigraphy and geotechnical parameters for decades. Performance of geo-environmental CPT is specified by ASTM Standards D5778, D6067 and Guide D6001.
The CPT tip resistance increases exponentially with soil grain size. Tip resistance in a sand aquifer is typically one to two orders of magnitude greater than in a clay aquitard. For example, the CPT tip resistance varies from about 10 to 40 MPa (100 to 400 tons per square ft (tsf)) in a dense sand aquifer, while tip resistance in a stiff clay aquitard ranges from about 0.5 to 1.5 MPa (5 to 15 tsf). The friction ratio (proportion of friction to tip resistance) is also used as an indicator of soil type. The friction ratio allows discrimination between loose sands and hard silts and clays, where tip resistances can be similar. The friction ratio ranges from about 1% in a sand to greater than about 3% in a clay. Silts have intermediate friction ratios.
High resolution, continuous soil profiling (sounding) for geo-environmental exploration is most often performed by STRATIGRAPHICS using the Piezometric Cone Penetration Test with soil Electrical Conductivity measurements (CPTU-EC). The cone tip and friction resistance measurements (CPTU-EC) are evaluated for soil types and geotechnical parameters (Douglas and Olsen, 1981). The piezometric measurement (CPTU-EC) allows evaluation of soil saturation, hydraulic conductivity, potentiometric surfaces, and soil types (Saines et al, 1989, Robertson et al, 1986). The soil Electrical Conductivity measurement (CPTU-EC) provides information on soil moisture in vadose zone soils and indications of groundwater quality in saturated soils. The EC measurement has proved very useful in exploration for inorganic (metal, brines and landfill leachate) contamination and somewhat useful in non-aqueous phase liquid (NAPL) exploration.
Soil EC is measured by STRATIGRAPHICS using a rugged two electrode array mounted on the tip of the penetrometer (Strutynsky, et al, 1991). A 3 kHz AC voltage is applied to the array to control polarization and contact resistance effects. EC is computed based on in-phase currents induced across the array and a reference resistor. The EC measurement has a resolution of about 1-3 cm.
The STRATIGRAPHICS Penetrometer Sampler can be deployed in groundwater, soil or soil gas sampling modes by using interchangeable components. The groundwater sampler is a heavy wall, shielded wellpoint sampler. The shield prevents cross contamination and clogging of the sampler screen. The shield is retracted to allow groundwater to flow through a 0.5 m (20 inch) long screen into the barrel. Sample can be decanted from the barrel or can be pumped to the surface using an inertial pump. A small diameter pressure transducer can be lowered into the sampler to log the rate of groundwater inflow. Inflow results can be analyzed to estimate soil hydraulic conductivity.
Penetrometer exploration can be used at sites where predominant soil grain sizes are less than about ½ the diameter of the downhole tools, i.e. less than about 3-4 cm (medium to coarse gravel). A small fraction of larger particles can be tolerated, as long as the coarser particles are within a matrix of much finer soil grains, as is common in many glacial till units. Deep penetration can be limited by excess friction on the rod string where soft, squeezing clays are encountered. Rod friction can also be a problem where thick sequences of hardpan silts or very fine, saturated sands are encountered. Thick sequences of very soft soils, such as peat layers, can limit deep penetration as little lateral support is provided to the slender penetrometer rod string in the weak layer, while attempting to push through dense soils at depth. Extremely dense (SPT blowcounts greater than about 100) can be difficult to penetrate. Dynamic rod driving may help in this situation.
Chloride brine production (40% chloride salt solution) from well fields in Michigan is transported to a processing facility through pipelines buried at shallow depths. At one site, galvanic corrosion at a welded pipeline joint led to a release of about 40,000 gallons of chloride brine. The initial response removed over 100,000 gallons of surface water and brine, but a portion of the released brine permeated into a thick sand aquifer, and some seeped into a nearby creek. Chloride brine migration was primarily controlled by differences in density between the heavy brine and lighter fresh groundwater, and by the generally westward regional groundwater flow.
Soils in the release area consist of about 180 m (600 ft) of sands, silts and clays, which overlie shale bedrock. These soils represent a sequence of lakebed and shoreline deposits in a paleo environment when water levels in the Great Lakes were much higher than at present. The sand aquifer is probably the result of dune and beach activity. The aquifer ranges in thickness from about 6 m (20 ft) near the brine release point, to over 40 m (130 ft) at a distance 600 m (2000 ft) west of the release point. An aquitard consisting of interlayered clays, silts, and silty sands underlies the sand aquifer.
The suspected release point and surrounding area were initially profiled to depths of about 6 m using near field EM38 and intermediate EM31 surface geophysical instruments. Within a week, the shallow plume was mapped to an area of about 35 acres. Purge and monitoring wells were installed to control the migration of the chloride brine through the aquifer. Additional geophysical profiling, and installation of more purge and monitoring wells, was performed in several phases over a period of years, as it became apparent that westward plume migration was continuing.
Due to the dense chloride brine sinking to the top of the westward dipping aquitard layer and the westward regional flow of groundwater, a far field EM34 surface geophysical instrument had to be used for profiling the central portion of the plume. At a distance of about 600 m (2000 ft) west of the release point, the plume had migrated deeper than the layer resolving capabilities of the EM34 instrument. Further plume characterization was performed with the CPTU-EC cone penetrometer method.
A series of 33 CPTU-EC soundings, to depths as great as 50 m (165 ft) and totaling 1061 m (3481 ft) of data, and 27 penetrometer groundwater samples, were completed to characterize the deep, westward extension of the plume. A thin zone of groundwater with slightly elevated chloride concentrations was eventually found to extend along the top of the confining aquitard unit to as far as 1200 m (4000 ft) west of the initial release area.
The indirect CPTU-EC data indicated that the sand aquifer was relatively homogeneous, with very few, thin, apparently discontinuous clay or silt seams or layers. In contrast, the aquitard was found to be much more heterogeneous than the aquifer, with interlayers of clays, silts, and silty sands. EC data (both CPTU-EC and geophysical) indicated groundwater unaffected by the chloride brine is low in electrical conductivity, which provides a very clear contrast to groundwater with elevated chloride content. Piezometric testing during CPTU-EC indicated aquifer hydraulic conductivities to range from about 1E-3 to over 1E-2 cm/sec. Aquitard hydraulic conductivities ranged from about 1E-4 to less than 1E-8 cm/sec, depending on whether the testing was in silty sand or silty clay layers. The indirect penetrometer evaluations of stratigraphy, hydraulic conductivity and EC agreed very well with indirect geophysical, direct sampled borehole, and pump test data collected at the spill site.
The CPTU-EC data were evaluated to delineate the extent of the sand aquifer and clay aquitard units. Depth intervals within aquifer soils, where peak EC measurements were encountered, were selected for direct penetrometer groundwater sampling. Sample targeting using the continuous CPTU-EC sounding logs significantly reduced the number of samples and analytical testing required to adequately characterize groundwater chloride concentrations. Near the plume source, where CPTU-EC measurements were greater than about 400 mS/m (4mS/cm), direct penetrometer groundwater sampling was typically not performed, as high chloride concentrations were not in question for these conditions.
A comparison between CPTU-EC penetrometer data and a drilled, sampled and geophysically logged monitoring station is shown in Figure 2. The two exploration points were within about a 9 m (30 ft) distance of each other, with the CPTU-EC sounding slightly closer to the plume axis. Very good correspondence is observed between the soil types evaluated from the CPTU-EC measurements and visual classification of obtained samples. The continuous CPTU-EC logs provide data across unsampled intervals in the discontinuously sampled borehole, and greatly increased resolution of thin layering.
Comparison between the CPTU-EC and downhole EM39 data also shows very good correspondence. The CPTU-EC data provide higher resolution (2-3 cm versus 15-30 cm for EM39) of layering. In many instances, this can be advantageous (for example, during other projects when trying to detect thin LNAPL or DNAPL layers). A disadvantage to the high CPTU-EC resolution is that in gravelly soils where particle size is comparable to EC sensor resolution, a significant amount of noise can be exhibited in the EC data.
The CPTU-EC piezometric data in the example sounding log indicate that the bottom of the aquifer unit is slightly less permeable than at shallow depths. This fact, not apparent from visual descriptions of drilled samples, may explain the somewhat unusual shapes of the EM39 geophysical and CPTU-EC logs, and thus the distribution of chloride concentrations at the bottom of the aquifer.
The peak value of EC (and thus, highest chloride brine concentration) might be expected to fall at the interface between the sand aquifer and the underlying aquitard at 25.5 m, if the aquifer was homogeneous in hydraulic conductivity. Instead, the peak EM39 and CPTU-EC values were measured at a depth of about 21 m. This is the depth at which aquifer hydraulic conductivities slightly decrease, as indicated by CPTU-EC piezometric measurements. EC measurements slowly decrease with depth from this point, and somewhat abruptly decrease at about
23 m. The CPTU-EC piezometric data indicate a thin (<10 cm) silt seam at this depth, not detected during drilled sampling. A very abrupt decrease in EC finally occurs at the aquifer/aquitard interface, at 25.5 m.
A comparison between peak EC data measured in the CPTU-EC soundings and average EC data from surface geophysical surveys is presented in Figure 3. Generally good correspondence can be seen between the two exploration methods. As expected, the CPTU-EC peak data are much higher than the average data from the surface EM surveys. The deep, low concentration chloride affected groundwater in the western portion of the site was only detected using CPTU-EC data. The surface EM geophysical survey was of great benefit where plume depths did not exceed about 20-30 m.
Correlations were developed from comparing CPTU-EC and downhole EM39 data to laboratory chloride concentrations obtained from groundwater samples, in order to allow estimation of chloride concentrations. For the CPTU-EC data, and for chloride concentrations less than about 1000 ppm (parts per million), the following relationship was derived: chloride concentration, in ppm = (0.8 * EC (in uS/cm))-171. For chloride concentrations up to 50,000 ppm, an alternate relationship was developed: chloride concentration, in ppm = (1.9 * EC (in uS/cm))-1029. For a regulatory maximum chloride concentration in groundwater of 250 ppm, the corresponding CPTU-EC value had to be less than about 500 to 550 uS/cm.
It is important to note that these relationships were developed at a site where background groundwater was very low in dissolved minerals, and thus of low electrical conductance. Application of these correlations to other sites is limited - site specific correlations should be performed at each project site.
An exploration technique for groundwater affected by chloride brine was developed based on surface and downhole geophysical, and direct push cone penetrometer methods. Indirect and direct measurements using these methods were obtained and compared to results of traditional drilled and sampled boreholes and monitoring wells. The new technique provided good technical and economic justification for extensive use at the described site. Use of the combined geophysical and cone penetrometer exploration technique was then adopted for use at other brine release sites in the area. These programs met with very good success and acceptance by regulatory agencies.
Archie, G.E.,1942. The Electrical Resistivity Log as an Aid in determining some Reservoir Characteristics. AIME Vol. 146.
Douglas, B.J., R.S. Olsen, 1981. Soil Classification using the Electric Cone Penetrometer. Cone Penetrometer Testing and Experience, ASCE.
Robertson, P.K., R. G. Campanella, D. Gillespie, and J. Grieg, 1986, Use of Piezometer Cone Data, InSitu 86, ASCE Specialty Conference, Blacksburg, VA
Saines, M., A.I. Strutynsky, and G. Lytwynyshyn, 1989. Use of Piezometric Cone Penetration Testing in Hydrogeologic Investigations. First USA/USSR Hydrogeology Conference, Moscow, USSR
Strutynsky, A.I., T. Sainey, 1991. Piezometric Cone Penetration Testing and Penetrometer Groundwater Sampling for Volatile Organic Contaminant Plume Detection. Petroleum Hydrocarbons and Organic Chemicals in Ground water: Prevention, Detection and Restoration. API/NWWA.
Strutynsky, A.I., R.Sandiford, D.Cavaliere, 1991. Use of Piezometric Cone Penetration Testing with Electrical Conductivity Measurements (CPTU-EC) for Detection of Hydrocarbon Contamination in Granular Soils. Current Practices in Groundwater and Vadose Zone Investigations, ASTM.