Methods to Identify and Reduce Potential Surface Stream Water Losses into Abandoned Underground Mines TERRY E. ACKMAN and J. RICHARD JONES Pittsburgh Research Center United States Bureau of Mines Pittsburgh, Pennsylvania 15236, U.S.A. ABSTRACT / Methods to identify and subsequently seal surface water loss zones in stream channels were tested by the United States Bureau of Mines at Staub Run, a first-order stream near Frostburg, Maryland, that partially overlies abandoned coal mine workings. Conventional stream gauging was conducted to establish discharge patterns before and after stream sealing. Electromagnetic terrain conductivity surveys were performed within the stream channel to identify zones
Introduction
Fractures in the strata overlying underground abandoned mines in the viciniity of streams frequently result in a loss of streamflow as water migrates along fractures associated with underground workings (Hobba 1981; Hollyday and McKenzie 1973; Williams and others 1986). Such stream infiltration is considered to be a major contributor to the overall volume of underground mine water production (U.S. Environmental Protection Agency 1979). Multiple zones of infiltration (both natural and/or induced) usually exist in stream channels overlying mined areas. Stream loss zones are usually not apparent from visual surface observations. It is also difficult to locate surface infiltration zones from underground mines (assuming accessibility), since subsurface water flow paths may deviate considerable distances through the overlying bedrock fracture systems. By reducing the volume of water that enters underground mines from the surface, the volume of acid mine drainage (AMD) treatment at the mine site can be reduced (Ash and Whaite 1953). Efforts to reduce stream seepage in the anthracite region of Pennsylvania using wooden canals have had limited success (Ash and Whaite 1953). Plastic membranes, clay, and rip-rapping have been used elsewhere. The long-term effectiveness of such liners is questionable during times of drought or intermittent flow, since vegetation, burrowing animals, and insects can damage the integrity of artificial channel bottoms. EnvironGeol Water Sci Vol. 17, No. 3, 227-232
of increased relative water saturation to depths less than 15 m. Zones of increased conductivity were generally found to be associated with areas exhibiting statistically significant (P <~ 0.05) gauged flow losses. Conversely, zones that exhibited declining conductivity delineated areas where between-station flows were not significantly different. Using this information on potential loss zones, an experimental grouting procedure was applied by injecting an expandable polyurethane grout to a depth less than one meter into the alluvial streambed over a 180-m section of the stream channel. Before grouting, the study section exhibited a 24 I/sec flow loss; first-phase grouting reduced this to a 14 I/sec flow loss; with a second-phase grouting the losses were only 3 I/sec.
The U.S. Bureau of Mines, in cooperation with other federal and state agencies and with private industry, implemented a research program to identify and selectively seal sections along a stream channel that are loss zones that potentially infiltrate into mines. By accurately locating these loss zones, the length of stream channel that must be lined can be reduced considerably. This article presents the methodology used to identify zones of subsurface loss and reports the experimental results of polyurethane grout sealing at Staub Run, located near Frostburg, Maryland.
Environmental Setting
Staub Run is in northwestern Maryland within the Appalachian Plateau Physiographic province approximately 8 km south of Frostburg, Maryland (Fig. 1). Staub Run is a first-order tributary of the Potomac River. Annual precipitation averages 105 cm. The area is part of the Georges Creek Basin syncline, the northern extension of the Potomac Basin. Staub Run has an average gradient of 0.2 m/km. the section of Staub Run under study is approximately 0.9 km in length, averages about 3 m in width and approximately 1 m in depth. Staub Run flows perennially to the point at which the coal seam directly underlies the alluvial stream channel. Only intermittent flow is observed beyond this stream channel section, which overlies the abandoned workings. A number of Pennsylvanian coal-bearing strata un© 1991 Springer-VerlagNew York {nc.
228
T.E. Ackman and J. R. Jones
B •~
Stoub
5 ~
N
.
14X~:'/
-4-----
,2~GS5 ~
~ f
\o
n
d
~ u
c
t
i
v
i
t
~
44
#6
LEGEND
Gagingstation
y
station
*Welt locations
~
r[1Tm~Secfions chosen for grouting
GSI
~'~ ~ " ~ ....
~
._~Road o
2oot,
il~
GSI5
Figure 1. Staub Run, Maryland study area. Well 7 o
-
Well 6
Well 5
:~:o--.c
~.%..%~ 5"I(i. W Q
Alluvium
IO-
I ~ I Coal :...........,..
Sandstone I ~ Shale
15-
Figure 2. Stratigraphic relationships at Staub Run (see Fig. 1 for well locations). derlie the Staub Run area. The coal units extend downward from the Pittsburgh Coal at the base of the Monongahela Group to the top of the Mount Savage Coal, which marks the base of the Allegheny Group (Toenges 1949; O'Hara 1990). The cumulative thickness of these coal-bearing strata averages 235 m. The Pittsburgh Coal crops out in the upper reaches
of Staub Run. The coalbed dips to the south at 0.5 m/km. The outcrop is arcuate along the stream channel, such that the outcrop extends along both sides of the mountain valley after exposure within the stream channel. The middle to lower reach of this stream overlies an abandoned turn-of-the-century coal mine (Carlos Mine). The thickness of overburden to the Pittsburgh coal for the 914-m-long test site ranges between 4 and 10 m (Fig. 2). Stripping activity (pillar recovery) has occurred on both sides of Staub Run. In one portion of Staub Run, stripping operations have mined through the stream, necessitating a diversion of approximately 305 m of stream channel (Fig. 1). The alluvial channel bottom consists of sands, silts, clays, pebbles, and cobbles. The thickness of this alluvial material, based on drill records averages 4 m at the study site (Fig. 2).
Methodology Stream Gauging Fifteen stream gauging stations spaced at approxi-
229
Surface Stream Water Losses and Mining
BIO
lO0
w
I
'
'
~
'
T
,
~
-
12
-
I0 " "~ -
90
n.-
< :1:
GAGING STATION 5 6 t =
,
8o
Depth (m) o 7.5 + 15.0
o
7o
600
I00
i
i
I
I
I
~
| _~"~ [ 1
~+
'J"
"4
l
200 300 400 ,600 600 700 DISTANCE FROM STATION I (rn)
3. P r e g r o u t i n g
Figure
/
,~,%t*\ +~ \
8 i
~_ 6
8 I
.f* .-
7 i
discharges at Staub Run
800
900
(October
1986-September 1987). See Figtlre 1 for gauging station locations. 01
, .......
I
I .........
I0
Depth (rn)
IO
5
6
7
8
l
I
I
I
Figure 5.
. ++%.
o 7.5
.
f
~.: 15.0
/
I ........
40
50
Conductivity plots for low-flow conditions.
stream sealing efforts. Gauging was intensified, beginning in March 1987, to focusing on the upper study site by routinely monitoring gauging stations 1, 5, 6, 7, and 8 (Fig. 1).
~.
'+,
/
~ 8
I .........
.30
CONDUCTIVITYSTATION
GAGING STATION
12
I .........
20
E
E 6~ >t--
Conductivity
b 4 D O Z
8a ,~,,,,,I
....
I0 Figure 4.
i~,,,l
.........
I .........
I
.........
20 30 40 CONDUCTIVITYSTATION
50
Conductivity plots for high-flow conditions.
mately 60-m intervals were established along Staub Run (Fig. 1). Stream gauging began in October 1986 and continued through May 1989. Discharge was measured using a portable flowmeter equipped with an electromagnetic sensor, following standard procedures established by the U.S. Geological Survey (Buchanan and Somers 1969). Initial gauging efforts focused on establishing a flow profile for the study area. The profile included routine monitoring of gauging stations 1, 5, 8, 14, and 15 (Figs. 1 and 3). Stations 1 and 5, outside the influences of mining, served as the control stations while station 15 represented the last gauging station in the study area. It became apparent from the gauging data that losses of stream flow occurred in the upper portions of the study site (Fig. 3). Evaluation of these gauging data (Fig. 3) targeted the zone(s) for future
The Geonics EM-34 electromagnetic ground conductivity survey instrument was used to confirm stream loss zones implied by gauging efforts. Apparent conductivity observed at 10-m intervals employing intercoil spacing of 10 m between transmitter and receiver coils was taken within the stream channel (Fig. 1). The relative positions of the coils was 1 m above the surface. Two series of conductivity measurements were taken within the stream channel between gauging stations 1 and 8. The EM-34 electromagnetic ground conductivity meter was used to obtain nominal depth measurements at about 7.5 m (vertical dipole) and 15 m (horizontal dipole). One series of measurements was taken during a high-flow period (Fig. 4) and the other during a low-flow period (Fig. 5). The details of instrument operation are given by McNeill (1980). Theoretically, the magnitude of apparent ground conductivity should increase with a corresponding increase in saturation. Consequently, water loss zones should be zones of higher relative conductivity. Thus, the conductivity data and gaging data should exhibit an inverse relationship. Stream Sealing
Experimental grouting was conducted between
230
T.E. Ackman and J. R. Jones
conductivity stations 2 0 - 2 6 and 3 2 - 4 5 on 1 October 1988 (Fig. 1) because these areas exhibited inverse relationships between conductivity and gauging data with respect to flow loss (Figs. 3 and 4). Subsurface injection of an expandable polyurethane grout was utilized to contain water within the stream channel. Using this approach, a shallow and relatively impermeable barrier was placed in stream segments that have been identified as loss zones by conductivity and stream gauging data. The polyurethane grout serves as a shallow subsurface curtain at the stream channel sediments-bedrock fracture system interface, inhibiting potential flow to the underground workings. One hundred thirty-six hollow steel rods 1 m in length were used for injection of the polyurethane material into the alluvial stream sediments. The rods were 1.9 cm in diameter with hardened steel points. Holes (0.3 cm) were drilled through sidewalls near the lower portion of the rod to enable grout diffusion. The top of the threaded rods incorporated a hardened steel cap to withstand blows from a sledge h a m m e r and for mechanical injection of the polyurethane grout. The rods were manually driven into the stream sediments with a sledge h a m m e r to a depth of 0.66 m beneath the stream channel and placed at 3-m intervals along the stream axis. T h e number of rods placed across the stream depended upon channel width and varied from two to four rods. A measure of 19 1 (20.2 kg) of a two-component polyurethane grout (9.5 1 of each component) was injected into each grout rod. Injection pressures ranged from 28 to 88 kg/cm 2 with the in situ conditions of alluvial material dictating the specific pressure. Grout was injected at a maximum pressure obtainable until a surface leak developed. After the leak occurred, the injection pressure was reduced to 3 kg/cm 2 until the leak self-sealed. T h e pressure then was increased slowly until another leak occurred. This procedure was repeated until the 19 1 of polyurethane grout was depleted.
Results In most natural streams there generally is a downstream increase in the volume of discharge. To test if the between-station discharges along Staub Run followed such a trend before grouting, a series of onetailed t tests was computed for the gauging station sets of 1 and 5, 5 and 6, 6 and 7, 7 and 8, and 5 and 8 (Table 1). The one-tailed t test is used to determine if a significant difference (P ~< 0.05) exists in discharge measurements between the successive gauging stations. Also, the test is directional, in that one can test if
Table 1.
t test comparisons for discharges
A. Pregrouting Gauging stations Differeces 1 5 5 6 6 7 7 8 5 8
Average flow (l/sec)
Number of paired gauging events
89.0 77.8 48.9 43.3 43.3 36.0 93.2 97.7 87.1 64.9
29
Yes
4
No
4
Yes
5
No
19
Yes
B. Pregrouted and phase I grouted (relative to station 1) Average flow Number of Gauging stations (l/s) gauging events Differences Pregrouting Phase I 5 Pregrouting Phase I 6 Pregrouting Phase I 7 Pregrouting Phase I 8
5 6 7 8
- 8.2 -4.9 - 7.3 -6.6 - 15.2 -14.2 - 32.3 - 17.8
42 35 6 30 6 28 27 24
Statistically significant
Statistically significant No No No Yes
the discharge at one gauging station is significantly greater than that at another gauging station. Since flow rates should generally increase in the downstream direction, the successive downstream station discharge should not have been significantly less than the corresponding upstream station discharge. If it can be demonstrated that the reverse occurs, i.e., the upstream station had a significantly greater discharge, then it is assumed that the stream segment between the two gauging stations is experiencing a flow loss and thus represents a zone of significant subsurface infiltration. The t tests were applied only when measurements were recorded for comparative stations on the same day (Table 1). T h e results of the t tests showed that significant pregrouting downstream losses occurred between stations 1 and 5 and stations 6 and 7 (Table 1). Stream segments located between stations 1 and 5 and stations 6 and 7 represent zones of statistically significant (P ~< 0.05) subsurface infiltration. Although the downstream flow rates were less than corresponding upstream flow rates between stations 5 and 6 and stations 7 and 8 (Table 1), the differences were not statistically
Surface Stream Water Losses and Mining
IO
"~ .J
I
I
231
I
0
(,9 n- - I O
1" o G') o
KEY Pregrouting (Oct 86 - Sep 87)
~ -30
Phase If Grouting (June 8 8 - May89)
z -20
Phase I Grouting (Oct 8 7 - M a y 8 8 )
l
- 40 Station 5
I Sta~'ion 6
Figure6. Change in station discharges (relative to station 1) for the three comparison periods.
I Station 7
significant. To further test an overall flow loss pattern, a one-tailed t test also was computed between stations 5 and 8. The t test results illustrated a significantly (P 0.05) higher flow at station 5 as compared to the flow at station 8. This confirms that infiltration is occurring between stations 5 and 8. As noted in Table 1 and illustrated in Figure 3, a significant discharge loss occurred between gauging stations 1 and 5. Based on local stratigraphy and tectonic history, the coal could not have been extracted from beneath this section of stream (Fig. 2). Examination of stereo-paired aerial photographs of the study area (1:1000, 1983 series) revealed an apparent linear transecting the stream in the vicinity of gauging station 5. The area of the apparent linear has, however, been altered significantly through surface mining, timbering, and the construction of residential and other structures, so field reconnaissance of the area failed to confirm its existence. The conductivity trends under both low- and highflow conditions between gauging stations 1 and 8 showed similar patterns (Figs. 4 and 5). There was a general increase in conductivity about 200 m downstream from gauging station 1 that peaked between gauging stations 5 and 6. The conductivity measurements within the stream channel decreased from gauging station 5 to gauging station 6. Between gauging stations 6 and 7, there was a small increase in conductivity until about mid~Tay between these two stations. The conductivity trends under high-flow conditions showed a general decrease in conductivity between gauging stations 7 and 8 (Fig. 4), while a slight increase in conductivity was observed under low-flow conditions between these gauging stations (Fig. 5). Given the overall trend of the conductivity
Station 8
measurements, it appears that an increase in conductivity is assodated with flow losses as measured by stream gauging. In order to test the effects of grouting on the discharges at stations 5-8, comparisons were made between the pregrouting and phase I grouting flows occurring between October 1987 and April 1988. Maximum discharges did not exceed 333 1/sec during the phase I grouting gauging period, so flows greater than these were excluded from pregrouting gauging measurement sets so as not to bias flow comparisons. Pregrouting and phase ! grouting discharges at station 1 (the controls station) were not significantly (P 0.05) different. The discharge change between pregrouting and phase I grouting flows for stations 5, 6, 7, and 8 was standardized relative to control station 1. Figure 6 illustrates the station-to-station change relative to station 1 between pregrouted and Phase 1 grouted discharges. Comparison of the pregrouting flows with the phase 1 grouting flows shows a general reduction in the station-to-station discharge losses (Fig. 6). The ungrouted stream reach between gauging station 1 and conductivity station 20 is some 200 m (Fig. 1); it appears likely that some water has infiltrated through the stream bed prior to reaching the grouted section. This suspected infiltration zone upstream of gauging station 5 appears to be confirmed by trends in the conductivity surveys (Fig. 4). Phase I grouting upstream of station 5 and between stations 6 and 7 did not make a significant difference. Examination of the stratigraphic well data shows the alluvial stream channel between these stations to overlie the coalbed directly (Fig. 2). It is believed that the proximity of the fractured coalbed beneath the
232
T.E. Ackman and J. R. Jones
Table 2. t test comparisons for pregrouted and phase tl grouted discharges (relative to station 1)
Gauging stations
Average flow (1/sec)
Number of gauging events
Statistically significant
Differences Pregrouting 5 Phase II 5 Pregrouting 6 Phase II 6 Pregrouting 7 Phase II 7 Pregrouting 8 Phase II 8
- 8.2 +6.2 - 7.3 + 5.4 - 15.2 -7.5 - 32.3 -3.9
42 26 6 27 6 25 27 24
Yes Yes No Yes
ConcLusions The approach used to identify stream loss zones in this study has demonstrated that areas of high conductivity are generally associated with significant surface stream losses as determined by conventional gauging. The grout injected into the alluvial stream channel effectively reduced stream losses during the time of observation. Comparison of the pregrouted stream discharges with those of the postgrouted discharges showed an average reduction in loss from 24 1/sec to 3 l/sec.
References Cited stream channel (<5 m) is the cause of this continued loss. T h e only significant difference observed in discharge occurred between pregrouted and phase I grouted station 8. This is likely to be the result of a cumulative discharge gain along the entire grouted stream section between stations 5 and 8. A second-phase grouting effort was deemed necessary in an attempt to further reduce the losses occurring between station 6 and 7. On 1 May 1988, 40 additional grout rods were emplaced at 3-m centers within the phase I rod pattern between stations 6 and 7. Grout injection procedures followed those previously described. Stream gauging continued through May 1989. Statistical differences in flows at control station 1 were not observed between the pregrouting and phase II grouting periods. Comparison between the discharges of the phase II grouting period relative to station 1 are illustrated in Figure 6. The series of t tests between the pregrouted and Phase II grouted discharges show that statistically greater flows were realized at stations 5, 6, and 8 (Table 2) after phase II grouting. Although the flow losses at station 7 continued, the phase II grouting effort did reduce the average loss from 15 1/sec to 7 1/sec (Table 2).
Ash, S. H., and R. H. Whaite, 1953, Surface-water seepage into anthracite mines into the Wyoming Basin Northern Field: BuMines Bulletin 534, 30 p. Buchanan, T.J., and W.T. Somers, 1969, Discharge measurements at gauging stations: U.S. Geological Survey TWRI Book 3, Chapter A8, 65 p. Hobba, W.A., 1981, Effects of underground mining and mine collapse on hydrology of selected basins in West Virginia: U.S. Geological Survey RI-33, 72 p. Hollyday, E. F., and S. W. McKenzie, 1973, Hydrogeology of the formation and neutralization of acid waters draining from underground coal mines of western Maryland: Maryland Geological Survey, RI-No. 20, 50 p. McNeill, J.D., 1980, Electromagnetic terrain conductivity measurement at low induction numbers: Technical Note 6, Geonics Limited, Mississauga, Ontario, Canada. O'Hara, C. C., 1900, The geology of Allegheny County: Baltimore, MD, Maryland Geological Survey, 159 p. Toenges, A.L., 1949, Investigation of lower coal beds in Georges Creek and northern part of upper Potomac Basins, Allegheny and Garrett countries, MD: BuMines Technical Paper 725, 100 p. U.S. Environmental Protection Agency, 1979, Dewatering active underground coal mines: technical aspects and cost-effectiveness: Washington, D.C., EPA-600/7-79-124, 123 p. Williams, R.E., G.V. Winter, G.L. Bloomsburg, and D. R. Ralston, 1986, Mine hydrology: Society of Mining Engineers Inc., 169 p.