Aquifer performance test report : St. Mary's Miocene aquifer, Camden County, Georgia, September 30-October 6, 1997 / by Bill Sharp, Sam Watson, and Rex A. Hodges

:P.
.:loo.G!4
\ 1
o.'34
AQUIFER PERFORMANCE TEST REPORT
St. Marys Miocene Aquifer, Camden County, Georgia September 30-0ctober 6, 1997
By Bill Sharp, Sam Watson, and Rex A. Hodges
Department of Natural Resources Environmental Protection Division
Georgia Geologic Survey
Project Report No. 34

AQUIFER PERFORMANCE TEST REPORT
t
St. Marys Miocene Aquifer, Camden County, Georgia
September 30-0ctober 6, 1997
By Bill Sharp, Sam Watson, and Rex Hodges
This report has not been reviewed for conformity with Georgia Geologic Survey editorial standards, stratigraphic nomenclature, and standards of professional practice.
DEPARTMENT OF NATURAL RESOURCES Lonice C. Barrett, Commissioner
ENVIRONMENTAL PROTECTION DIVISION Harold F. Reheis, Director
GEORGIA GEOLOGIC SURVEY William H. McLemore, State Geologist
Atlanta
1998
Project Report No. 34

II
ABSTRACT
As part of the Geologic Survey Branch of the Georgia Environmental Protection Division's "Evaluation of the Miocene Aquifers in the Coastal Area of Georgia Project", the Department of Geological Sciences at Clemson University conducted a pump test at the Georgia Geologic Survey's St. Marys well cluster, located in the city park in the town of St. Marys, Camden County, Georgia. A test of an unnamed Miocene aquifer (aquifer X) at St. Marys, was conducted from September 30 through October 6, 1997 using St. Marys 3 as the pumping well, along with St. Marys I and St. Marys 2 as monitoring wells. The observation well, St. Marys I, is located 78.75 feet away from the pumping well St. Marys 3 and is unscreened with an unknown open interval. Due to the pumping response in monitor well St. Marys I, it is assumed that the open interval includes the same hydrogeologic interval as the pumping well (aquifer X). Transmissivity calculations were made using multiple flow rates with flow rate steps of 71.00 gpm (0-24,500 seconds) and 71.40 gpm (24,500-259,200 seconds), and a well radius of 2 inches. A storativity of 0.0001 was estimated and the pumping well was assumed to be fully screened over the entire thickness of the aquifer. A transmissivity of 538 re/day (50 m2/day) with a skin factor of 3.44 was calculated for aquifer X. Data from the observation wells was not used in the calculations because St. Marys I was unscrecned with an unknown open interval and St. Marys 2 was not screened in the same aquifer as the pumping well. The 30ft (9.14 m) effective aquifer thickness of Miocene aquifer X at t"e St. Marys test site yields a hydraulic conductivitiy of I7.95 ft/day (5.47 m/day) and a
'

Ill
permeability of 6.59 darcys. Observation well St. Marys 2, located approximately 95 ft away from the pump well and screened in the Miocene Upper Brunswick aquifer was monitored to detect vertical leakage across the confining unit separating the Upper Brunswick aquifer from the overlying aquifer X (pumping well). No water-level changes directly related to pumping were observed in St. Marys 2, indicating no pumping related leakage across the confining unit separating aquifer X from the Upper Brunswick aquifer. Although the test site is located in a coastal area, the well cluster is located far enough away from the St. Marys River that water level fluctuations in the river (due to ocean tidal effects) had no influence on the water levels in the test wells.

IV
ACKNOWLEDGMENTS
Funding for this project was provide through Geologic Survey Branch of the Georgia Environmental Protection Division. Georgia Geologic Survey (William Steele and Earl Shapiro) oversaw well installation and coordinated logistical arrangements with Clemson University Geological Sciences Department Personnel to conduct the St. Marys Miocene aquifer perfonnance test. The city of St. Marys assisted with security of the site and the Park Service provided use of their pier for gathering tidal data.

v
TABLE OF CONTENTS
Page .TITLE PAGE ............................................................ I
ABSTRACT ............................................................. ii
ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
LIST OF TABLES ....................................................... vii
LIST OF FIGURES ..................................................... viii
CHAPTER
I. INTRODUCTION ............................................ 1
Purpose of the St. Marys Miocene Aquifer Performance Test .... 1 Site Conditions ......................................... 1
Location ........................................ 1 Hydrogeologic Setting .............................. 3 Description of Wells Used for the Test ................ 7
II. METHODS ...... , ......................................... 11
Test Logistics ............ : . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Data Acquisition Methods ........................ . . . . . . . 11
Pumping Well Data Acquisition Methods ............. 12 Observation Well Data Acquisition Methods .......... 12 Analysis Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Atmospheric Pressure Corrections . . . . . . . . . . . . . . . . . . . 12 Tidal Effects Corrections . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Well Analysis Methods ............................ 13
III. RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Duration ofthe Test .... : ............................... 15 Data Acquisition Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Pumping Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Water Level Readings ............................ 16 Water Level Change During the Test ................ 16

VI
Table of Contents (Continued)
Data Analysis Results ........................................ 21 Barometric Corrections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Tidal Corrections ...................................... 21 Trend Corrections ..................................... 22 Calculated Aquifer Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Calculated Skin Factor .................................. 24 Well Efficiency ....................................... 24 Specific Capacity ......... : ............................ 24
IV. DISCUSSION .................................................... 26
Analysis ................................................... 26 Ocean Effects ..................... ; . . . . . . . . . . . . . . . . . . . . . . . . . 26 Leakage ................................................... 26
V. REFERENCES ................................................... 28
Appendix A ...................................................... 30
Appendix B ...................................................... 40

Vll
LIST OF TABLES Tables I. Chart showing times for each phase of test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 II. Static water levels and maximum drawdown for test wells .................. 21 III. Calculated barometric efficiencies ..................................... 21

Vlll
LIST OF FIGURES

Figure

1. Map showing the location of St. Marys test site............................ 2

2. Map of St. Marys showing location of test site, ........................... 4

3. Map of test site showing relative locations of test wells. . ................... 5

4. Geophysical well logs, geologic and hydrologic units and depth of screen intervals at the St. Marys test site. . ..................................... 6

5.

Well construction diagram for pumping well St. Marys No.3 ................. 8

6. Well construction diagram for observation well St. Marys No.2. . ............ 9

7. Well construction diagram for deep observation well St. Marys No. 1. . ........ 10

8. Graph showing flow rate versus time over the duration of the pumping phase of the St. Marys test and flow rate steps used in the analysis. . . . . . . . . . . . . . . . . 17

9. Graph showing barometric corrected drawdown and recovery for pumping well St. Marys 3. . ................................................. 18

10. Graph showing barometric corrected drawdown and recovery for observation
well St. Marys 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
I
I
I
11. Graph showing barometric corrected drawdown and recovery for observation
well St. Marys 2. . ................................................. 20

12. Theis-Jacob analysis for pumping well St. Marys 3 using a multiple flow rate... 23

l3. Plots of pumping well St. Marys 3 measured drawdown and theoretical
drawdown ............................................. ~ .......... 25

INTRODUCTION

Purpose of the St. Marys Miocene Aquifer Perfom1ance Test

Due to the hydrologic stress imposed on the Eocene to Oligocel)e age Upper

Floridan aquifer, the principal water source of coastal Georgia, the Geologic Survey

Branch of the Georgia Environmental Protection Division is investigating the Miocene

aquifers as an additional source of ground water for the region. The test at St. Marys is

the second of seven Miocene aquifer tests (Tybee Island being the first) to be conducted

at selected sites in southeast Georgia. Four of the seven test sites will be located in

coastal counties, and three of the sites will be located inla~d where agricultural ground-

water use is prevalent. The purpose of the St. Marys pump test is to estimate the

transmissivity and storativity of the Miocene aquifer at the test site. Eventually the

hydrologic properties from each of the seven sites will be analyzed to determine if the

t

t

:

\ Miocene aquifers are viable alternatives to the Upper Floridan aquifer for smaller-

!

I

. demand needs such as community water supply, golf courses, a&rricultural (lower demand

1 ,

o

r

supplemental),

small

industries,

and

non-contact

cooling

water.

Site Conditions
Location
The coastal town of St. Marys is located in the southeast comer of Georgia, on the
I
St. Marys River, approximately 20 miles north of Jacksonville, Florida. Figure 1 is a map
Jf Georgia showing the location of the the St. Marys test site, the Tybee Island test site,

2
Miocene Aquifer Study Location Map
Figure I. Map of Georgia showing the location of the St. Marys test site.

3
and the Toombs County test site. The well cluster is located in the city park on Gallop St., approximately 100 feet south of the railroad tracks in the town of St, Mary's, Camden County, Georgia, as illustrated in figure 2. Figure 3 is a map of the St. Marys test site showing the relative locations of the pump well and the observation wells.
Hydrogeologic Setting The Coastal Plain sediments in the study area consist of unconsolidated to semiconsolidated layers of sand and clay and semi-consolidated to very dense layers of limestone and dolomite (Clarke, et al., 1990). Strata underlying the site dips and thickens to the southeast. The St. Mary well cluster is drilled into Coastal Plain sediments ranging in age from Eocene to post-Miocene. The hydrostratigraphy of the site consists principally of five aquifers. From deepest to shallowest, these are the Eocene Lower
Floridan, the Eocene Upp~r Floridan, the Miocene Lower Brunswick, the Miocene Upper
Brunswick and the Miocene to post-Miocene surficial aquifer (Clarke, et al., 1990). The major hydrogeologic units, geophysical well Jogs, and screen depth intervals of wells are shown in figure 4.

4

/

--
Nassau Co., Fla.

\

I

...

0

miles

0.5

Figure 2. Map of St. Marys showing location of test site, from 7.5 minute St. Marys quadrangle.

I

I
! .

5

Rijilrojd

Road

Power Pole Production Well (Well No. 3)
N

Deep Observation Well (Well No. 1)

Abandoned Corehole Miocene Observation Well (Well No. 2)

I

I

0

20 ft.

,Figure 3. Map of test site showing relative locations of test wells.

6
o Geologic Units GR*501

St. Post-Miocene
Unit
~
)
- -?- -
St. Ma ~ Miocene Unit A
......
St. Ma ~

Surficial Aquifer
Confining Un~~
Miocene Aquifer X Confining Unit Upper Brunswick Aquifer

Miocene Unit 8

Confining Unit

I~ Miocene Unit C

Lower Brunswick Aquifer

500'

= = * GR ganima ray, SN short normal resistivity

~ = I= screened interval

open !nterval

Figure 4: Geophysical \veil logs, geologic and hydrologic units and depth of screen

intervals at the St. Marys test site.



:.;.
7
Description of Wells Used for the Test St. Marys 3, screened in a unnamed aquifer (aquifer X), was used as the pumping well and St. Marys 2, screened in the underlying Upper Brunswick aquifer, served as the observation well for the test in order to detect leakage. The Upper Brunswick well, St. Marys 2, was not used as the pumping well (as was the case with the Tybee Island test) because it did not yield significant quantities of water. St. Marys I was not screened with an unknown open interval due to losing a drill rod down the hole but was also monitored. The construction dia!,:rrams for each of the wells are shown in figures 5-7.

0 ft. 50 100 150 200 250

8
St. Marys Well No. 3 Production Well Site coordinates: 30 44'06" lat. 81 33'07".long. Date of construction: 08/97 Ground elevation: I0-15 ft. MSL Screened hydrologic unit: Aquifer X 0-225 ft., 4 in. steel casing
0-215 ft., cement grout
215-225 ft., bentonite seal 225-255 ft., 4 in. stainless steel screen, 0.02 in. slot 220-260 ft., 18/30 med.-fine filter sand 255-260 ft., 4 in. steel sump.

Figure 5:. Well construction diagram for pumping well St. Marys No.3.

,,

9 St. Marys Well No. 2 Miocene Observation Well
0 ft.

50

0-325 ft., 4 in. steel casing

100

0-319 ft., cement grout

150

Site coordinates: 30 44'06" lat.

81 33'07" long.

Date of construction: 08/97

200

Ground elevation: 10-15 ft. MSL

Screened hydrologic unit: Upper Brunswick aquifer

250

300

319-322 ft., bentonite seal

325-365 ft., 4 in. stainless steel screen, 0.02 m.

350

slot

322-370 ft., 20/30 filter sand

365-370 ft., 4 in. steel sump

Bigure 6: Well construction diagram for observation well St. Marys No. 2.
I

10
St. Marys Well No. 1 Deep Observation Well

Site coordinates: 30 4.4'06" lat. 8133'07" long.
Date of construction: 08/97
Ground elevation: 10-15 ft. MSL
Screened hydrologic unit: unscreened (open hole)

0 ft.

--

0-60 ft., 12 in. pvc casing

200

--

60--560 ft. 12 in. (?) open (?) hole

400

--

;

Drill rod "lost" in hole

'

.___

..-

600

- r~

-

-560 ft.-1040 ft., 8 in. open(?) hole

800 1000

r ~
\
' v- ~ -~

Stablilzer Compressed drill cuttings

Figure 7: Well construction diagram for deep observation well St. Marys No. I.

I 1
METHODS Test Logistics A pump test is composed of three periods of data collection: background pretest, pumping, and backb'TOund post test (recovery). Background data are used to determine if the aquifer is in an equilibrium condition and the extent to which it is being affected by inconsistent externaJforces. Background data are also used to determine the barometric efficiency of the monitored aquifer so test data can be corrected for changes in . atmospheric pressure. During the pumping phase of the test, the aquifer is stressed creating a pressure drawdown cone extending radially from the pumping well. After pumping stops, the aquifer is allowed to recover to pre-test conditions. The test took place from September 30 to October 6, I997. An unnamed Miocene aquifer (aquifer X) was pumped using St. Marys 3. The underlying Upper Brunswick aquifer was monitored to detect leakage, using St. Marys 2. St. Marys I, unscreened and with an unknown open interval was also monitored in the test. The test 1 consisted of 22.52 hours of background data collection, 72.00 hours of pumping and 37.78 hours of recovery data collection.
Data Acquisition Methods Water level readings are recorded as pressure changes in meters of water relative to an initial equilibrium static water level condition. For the duration of a pump test \(background through recovery), quartz crystal transducers measure water level changes in
i
!the pumping well and observation wells. Relative water level changes are recorded
I
I
,automatically on the computer data acquisition system at operator-spec,ified intervals

12 ranging from 3 seconds to 5 minutes throughout the test. An additional transducer monitors and records changes in atmospheric pressure, which are used to correct for atmospheric induced changes in water levels in the test wells.
Pumping We11 Data Acquisition Methods The pumping we11, St. Marys 3, is assumed screened over the entire X aquifer from -210 to-240ft (-64.0 to -73.6 m) MSL (figure 4). A 5 HP submersible pump was placed approximately 90 ft below the initial static water level and a 100 psi transducer was positioned 15 ft below the pump.
Observation Well Data Acquisition Methods Observation well St. Marys 2, screened in the Upper Brunswick aquifer from .: 310 to -350 ft (-94.5 to -106.7 m) MSL, was monitored to detect vertical leakage. St. Marys I, unscreened with an unknown open interval was also monitored. The relative screen positions are shown in figure 4. Pressure transducers (45 psi) were placed approximately I0 feet below the water level surface in the wells to continuously monitor water level changes.
Analysis Methods Atmospheric Pressure Corrections The first step in the analysis is to correct the raw well pressure data for atmospheric pressure-induced changes which can mask the small response of an aquifer in an observation well. Removal of atmospheric pressure-induced water level changes makes it easier to detect water level changes that result from pumping.

' .

.. '.: .-.

13
Barometric corrections are made by subtracting atinospheric pressure changes mu1tip1ied by the barometric efficiency (BE) of an aquifer from the corresponding water level measurements. The BE of an aquifer is the ratio of the change in hydraulic head in an aquifer (due to atmospheric changes) to the actual change in atmospheric pressure. A BE of 1 indicates that 100 percent of the atmospheric pressure changes have been transmitted to the aquifer. A BE of 0 would indicate that none of the atmospheric pressure changes have been transmitted to the aquifer. A typical BE for confined aquifers in the Coastal Plain of Georgia is about 0.6, ranging from 0.4 to 0.8.
Tidal Effects Corrections The St. Marys test site is located approximately 4000 feet from the the St. Marys River. The water level on the river was found to fluctuate approximately 7 feet due to tidal effects. Water level fluctuations on the St. Marys River were measured and recorded at the Park Service Pier in the town of St. Marys using an In Situ Troll pressure transducer and data logger. The water level data from the wells showed no influence from the tidal cycle.
Well Analysis Methods Data from an observation well screened in the same aquifer as the pumping well ,can be analyzed to calculate the storativity and transmissivity of the aquifer (see Test Logistics and Pumping Rates). Data from the pumping well are governed by three ~ariables: the transmissivity and storativity of the aquifer, and the skin factor of the bumping welL If one of the three variables is known or can be estimated, the other two I ~an be calculated. The skin factor ofthe pumping well is unknown and could be highly

14 variable depending on well. installation. The storativity of the aquifer is less sensitive than the transmissivity. It is estimated. from analysis of observation well data (if available) or from average storativity values of similar aquifers. This storativity value is then used in the analysis of the pump well data. Variable rate curve matching of drawdown data yields a transmissivity value for the aquifer and a skin factor for the pumping well ~sing the superposition of the. Theis solution (1935,) or)acob straight-line method (Cooper and Jacob, 1946) for variable tl~w rates, modified for the skin factor analysis of Van Everdingen (1953) for confined aquifers with fully penetrating wells. For partial penetrating wells, data are analyzed using the Hantush (1961, 1964) solution for partial penetrating wells modified to account for the skin factor and multiple flow rate. The Hantush solution is used to calculate the transmissivity of the aquifer and the skin factor of the well, while correcting for vertical flow within the aquifer. Hydraulic con' ductivity is determined by dividing the transmissivity by the effective aquifer thickness. Permeability can then be calculated by multiplying the hydraulic conductivity in m/sec by a factor of 104,000 to convert to darcys (at 20 C; Fetter, 1988).
....

15
RESULTS Duration of the Test The pump test, using pumping well St. Marys 3 and observation wells St. Marys I and St. Marys 2, took place over a five day span from September 30 through October 6, 1997. The specific times for each phase ofthe test are shown in Table I.

Data Background Data Pump On Recovery (pump oft)
Total test

No. of hours Time interval 22.52 hours (19:24 09/30/97 to 17:55 10/01/97)
72.0 hours (17:55 10/01/97 to 17:55 10/04/97) 37.8 hours ( J7:55 J0/04/97 to 07:43 10/06/97) 132.32 hours (09:07 03119/97 to 11 :45 03/23/97)

, Table I. Chart showing times for each phase of the test.

Data Acquisition Results Pumping Rates
Drawdown in the pumping well was created by pumping water from well St. Marys 3 using a 5 hp submersible pump. Flow rates during the test were automatically measured and recorded using an Omega digital flow meter. During the pumping phase of the test, an overall variation in flow rate of approximately J. 75 b'Pm was observed as
i illustrated in Figure 8. In the analysis, multiple flow rates were used to account for the
I

16
fluctuations. Flow rate steps of 71.00 gpm (0-24,500 seconds) and 71.40 gpm (24,500259,200) seconds were used, also shown in Figure 8.
Water Level Readings During the test, 1609 water level data points were recorded in the pumping and observation wells by the data aquisition system. Data points were recorded as frequently . as e_very 3 seconds at times of rapidly changing water levels (i.e. at the beginning and end of pumping phase of the test), decreasing to every 5 minutes when water level changes were relatively small. Figures 9-11 shows plots of barometric corrected water level changes vs. time for respectively, the pumping well St. Marys 3, the deep observation well St. Marys 1 and the observation well St. Marys 2, over the duration of the pump test.
Water Level Change During the Test A maximum drawdown of 16.74 meters (54.92 ft) was observed in the pumping well St. Marys 3 after 72.0 hours of pumping (Figure 9). A maximum drawdown of approximately 1.84 meters (6.04 ft) was seen in the deep unscreened observation well St. Marys 1 (Figure 10), 50 minutes after the pump was shut off. Monitor well St. Marys 2 showed no observable changes in water level due to pumping (Figure 11) indicating a relatively effective confining layer between Miocene aquifer X and the Miocene Upper Brunswick aquifer. Static water levels, measured from the top of the casing of the pumping and monitor wells, were taken on 09/30/97 prior to starting the test as illustrated in Table II.

-- -~~~---
i

Flow Rate Vs. Time for the St. Marys Pump Test

73.00

72.50

e 12.oo
-c.
C)
..nQ...s..) 71.50

~ 0

71.00

c;::::

1.40 gpm tlow rate

70.50

0 gpm flow rate

I.

70.00

:

10/1/97 0:00 10/2/97 0:00 10/3/97 0:00 10/4/97 0:00 10/5/97 0:00

date and time

Figure 8. Graph showing flow rate versus time over the duration of the pumping phase of the St. Marys test and flow rate steps used in the analysis.
-....l

i

Drawdown (barometric corrected) Vs. Time for Pumping

I

Well St. Marys 3

i

I

I

I
I
I -..s
I ~ ~
II - -o

j 1.000
-1.000 -3.000
-5.000

I f -7.000

I CD

I ~ I -s -9.000 -11.000 -13.000

I ;

I ~' fb -15.000 -17.ooo I 1 0/1/01

:
1 0/2/01

I
~ '
10/3/01 10/4/01

I
1ot5to1 1otato1

I1I
!
I I
Ii.
1onto1 1otato1

I
l
i
I
I
I
I
I

0:00 0:00 0:00 0:00 0:00 0:00 0:00 0:00

!
I

date and time

I

!

Figure 9. Graph showing barometric corrected drawdown and recovery for pumping well St. Marys 3 (screened in

aquifer X).



00

Drawdown (barometric corrected) Vs. Time for Open Hole Observation Well St. Marys 1

-.. 1.000
.s 0.500
-C'G
~ 0.000
0
.IC.!.D!. -0.500
-..CD
E
CD ::::s
t/) t/)
!
Q.
10/1/01
0:00

10/2/01 0:00

10/3/01 0:00

10/4/01 0:00

10/5/01 0:00

10/6/01 0:00

10/7/01 0:00

10/8/01 0:00

date and time

Figure 10. Graph showing barometric corrected drawdown and recovery for observation well St. Marys 1 (unscreened and open hole).
\0

Drawdown and Recovery for Miocene Upper Brunswick Observation Well St. Marys 2

14.00 , . . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - . , 13.95

-1'i:" 13.90 13.85 0 13.80

.. 13.75

!
::I

13.70

f 13.65

0.. 13.60

13.55

. 13.50 +------+------+----+------+-----+------ii--------1 . 9/30/97 0:00 10/1/97 0:00 10/2197 0:00 10/3/97 0:00 10/4/97 0:00 10/5/97 0:00 10/6/97 0:00 10fi/97 0:00

Date and Time

Figure 11. Graph showing barometric corrected drawdown and recovery for observation well St. Maiys 2 (screened in the. Miocene Upper Brunswick aquifer).

N 0

.-::

21

Well Scra.'l!ed zone
s. M. 3 Aquifer X

Depth of static WL pror to pwnping Mux. depth of static WI. due to pwnping

7.88ft (2.40 m)

62.80 ft (19.14 m)

Mux.druwJown
1
I
54.92 tl (16.74 m

S.M. I Unscreencd

16.21 fi (4.94 m)

22.22 ll (6.78 m)

6.04.!\ (1.84 m)

S.M.2 Up. Br. Aquiler

10.70 1\ (3.26 m)

10.70 1\ (3.26 .m)

0.00 1\ (0.00 m)

Table II. Static water levels and maximum drawdown for test wells.

Data Analysis Results Barometric Corrections Water level p~essure data from the pumping wel1 and monitor wells were corrected for atmospheric pressure changes using the following barometric efficiencies. The barometric efficiencies were calculated using the method described in a previous section (atmospheric pressure corrections). Table IJI shows calculated barometric efficiencies for aquifer X (St. Marys 3), the Upper Brunswick aquifer (St. Marys 2) and the open hole monitor well St. Marys I.

Table III. Calculated barometric efficiencies.
Tidal Corrections The St. Marys test site was located far enough away from the ocean that the well pressure data showed no response to the tidal fluctuations. Therefore no tidal corrections 1 were necessary for the St. Marys data analysis.

22
Trend Corrections Background water-level data was not available due to the malfunction of a downhole datalogger installed in pumping well St. Marys 3.
Calculated Aquifer Properties Data from the observation wells was not used to calculate storativity for aquifer X because observation well St. Marys 2 was not screened in the same aquifer as the pumping well and observation well St. Marys 1 was unscreened with an unknown open interval. Pumping well skin factor and transmissivity of aquifer X at the St. Marys test site were calculated using data collected from the pumping well St. Marys 3 during the 72 hour pump test. Figure 12 shows a Theis-Jacob curve match for measured and calculated drawdown vs. time. Early time data (from 1 to 1000 seconds) was not used in the curve match because of well bore storage effects. Calculated drawdown for the aquifer X pumping well St. Marys 3 was based on multiple flow rates with flow rate \ steps of 71.00 gpm (0-24,500 seconds), 71.40 gpm (24,500-259,200 seconds)(Figure 8), and a well radius of 2 inches. A storativity of 0.0001 was estimated and the well was assumed to be fu]]y screened over the entire thickness of the aquifer. Hydraulic conductivity and permeability calculations are based on the transmissivity and an estimated aquifer thickness of 30ft (9.14 m) for aquifer X (Figure 4). A curve match for the pumping well St. Marys 3 yielded a skin factor of 3.44, a transmissivity of 538 ft2/day (50 m2/day), a hydraulic conductivity of 17.95 ft/day (5.47 m/day), and a permeability of 6. 59 darcys.

-------

...--------

--------------

St. Marys 3 Pumping Well Theis Curve Match

5.00

7.00
--E 9.00
;c::::: 11.00

0
";C: 13.00

c E

15.00 17.00

19.00

100

- __.;, . .1.-- "\. . . ~-

8=0.0001 T=50 mA2fday skin=3.44

1000

10000

Time (sec)

100000

1000000 f----:-----:;-calculated drawdown

!

observed drawdown

Figure 12. Theis-Jacob analysis (curve match) for pumping well St. Marys 3 using a multiple flow rate.
1w--.J

24
Calculated Skin Factor The skin factor is a variable that quantitatively describes the conductive properties of the well itself. A high skin factor would normally indicate a poorly developed well, whereas a skin factor of 0 normally indicates a perfectly developed well. A skin factor of 3.44 was calculated for the pumping well, Mary's 3.
Well Efficiency The well efficiency is a variable that is directly related to the skin factor of a production well. The well efficiency is calculated by dividing the theoretical drawdown by the measured drawdown in a production well, where the theoretical drawdown is the
\ /
calculated drawdown assuming the well 100% efficient (skin factor is set to 0). A well efficiency of73.8% was calculated for the pumping well St. Marys 3 as shown in Figure 13.
Specific Capacity An average flow rate of 71.40 gpm created a 54.92 ft (16.74 m) drawdown after 72 hours of pumping in pumping well St. Marys 3 (aquifer X). This equates to a specific capacity of 1.30 gpm/ft.

- - - - - - - - - ----- ---- ------ -
i,

-------------

Drawdown vs. Time for St. Marys 3 Showing Measured

Drawdown and Theoretical Drawdown

' .

9.000 .H.

.

~ '''

OO -

~

0o .. I

T=50 m"2/day

i
10.000

8=0.0001

11.000

E
c

12.ooo

j~ 1.::3: .000 ...f.i:'; r; .:.. \.\ ;cc;:::Oc)

Qns 14.000

-~

1/ccc"' :'i<.o:':

1-.5.0_00

1 -cal~ulated drawdown with ~-~

j

skin factor of 0

-observed drawdown

:::::: l~-~'j{\'f:Si*~f,~"~f:1f,~:;~:i~11;01:f,g;g;l;;';;J}t{t;;~~21"(f.0~~J'}ij~{-,i;1)J!:i'i'!~~

1000

10000

100000

1000000

Time (sec)

Figure 13. Plots of pumping well St. Marys 3 measured drawdown (all correction factors applied) and theoretical drawdown with skin = 0, T = 2000 m2/day and S = 0.0001, showing well efficiency calculations. A well efficiency of73.8% was calculated but may not
be realistic due to the storativity (and therefore the pumping well skin factor) not being able to be estimated from the data analysis.

.
:#
N 1..1'1

26
DISCUSSION Analysis
A complete obseiVation well analysis using the Theis cuiVe matching method was not possible for St. Marys test because monitor well St. Marys 2 was not screened in the same aquifer as the pumping well and monitor well St. Marys 1 was unscreened with an unknown open intetVal. Since a storativity value from the obsetVation well analysis was not possible, the storativity used in the pumping wen analysis, had to be estimated. A typical value for storativity for a confined aquifer is 0.0001. The best cuiVe match for the pumping wen was achieved using a transmissivity of 538 ft2/day (50.0 m2/day), and a skin factor of 3.44. This yielded a permeability of6.59 darcys.
Ocean Effects Ocean tidal effects in the St. Marys river caused the surface of the river to
r
fluctuate approximately 7 feet. However, due to being far enough away from the river (approximately 3000 ft) the wells at the test site showed no response to the water level fluctuations in the river.
Leakage No leakage was detected across the confining layer that separates aquifer X and the underlying Upper Brunswick aquifer at the St. Marys test site. Figure 11 shows a 8 em overall increase in water level for the Upper Brunswick monitor well St. ,Marys 2,

27 over the duration of the pump test, indicating an effective confining layer between the two aquifers.

; lI

I . ', '.

... ~ ~ '

28
REFERENCES
Clarke, JohnS., Hacke, Charles, M.; and Peck, Michael F., 1990, Geology and groundwater resources of the coastal area of Georgia, Department of Natural Resources, Environmental Protection Division, Georgia Geologic Survey, Bulletin 113, 106 p.
Cooper, H. H., and Jacob, C. E., 1946, A generalized graphical method for evaluating formation constants and summarizing well field history, Transactions of the American Geophysical Union, v. 27., pp. 526-534.
,Dawson, Karen J., and Istok, Jonathan D., 1991, Aquifer Testing, Design and Analysis of Pumping and Slug Tests, Lewis Publishers, Inc., Chelsea, Michigan, 344 p.
Earlougher, Robert C., 1977, Advances in Well Test Analysis, Millet the Printer, Inc., Dallas, Texas, 264 p.
Fetter, C. W., 1988, Applied Hydrogeology, Macmillan, Inc., New York, NY, 691 p.
Freeze, AJian R., Cherry, John A., 1979, Groundwater, Prentice-Hall, Inc., Englewood Cliffs, NJ, 604 p.
Hantush, M. S., 1961, Drawdown around a partially penetrating well, Journal of the Hydrualics Division, Proceedings of the American Society of Civil Engineers, pp. 83-98.
Hantush, M.S., 1964, Hydraulics of wells: Advances in hydrosciences, v.1., Academic Press, New York, pp. 281-432.
Jacob, C. E., 1940, On the flow of water in an elastic artesian aquifer, American Geophysical Union Transactions, part 2, pp. 574-586.
Jenson, V. G., and Jeffreys, G. V., 1977, Mathematical Methods in Chemical Engineering, Academic Press Inc., New York, NY, 599 p.
Kroening, David E., Snipes, DavidS., Brame, Scott E., Hodges, Rex A., Price, Van and Temples, Tom J., 1996, The rehabilitation of monitoring wells clogged by calcite precipitation and drilling mud, Ground Water Monitoring and Remediation, v. 16, ho. 2, pp. 114-123.
Press, William H., Teukolsky, Saul A, Vetterling, William T., and Flannery, Brian P.,1992, Numerial Recipes, Cambridge University Press, New York, NY, 963 p.

--~-. ~~-;:. -.:. . ,..~-p ~

29

Snipes, DavidS., Benson, Sally M., and Hodges, Rex H., 1996, Establishment of a field hydrogeology experimental site final report~ January 15, 1991 -January 14, 1996, Clemson University Hydrogeology Program, SCUREF Task Order 26 and Cooperative Agreement 29-95, 161 p.
Theis, C.V., 1935. The relation between lowering of the piezometric surface and the rate and duration of the discharge of a well using groundwater storage. Transactions ofthe American Geophysical Union, Vol. 2., pp. 519-524.
Van Everdingen, A. F., 1953." The skin effect and its influence on the productive capacity of a well. Petroleum Transactions, Vol. 198, pp. 171-176.

30
Appendix A Correction Factors in the Analysis of Raw Well Data

31
Barometric Efficiency To compute the barometric efficiency of an aquifer, several measurements of the ambient barometric pressure change in time along with th~ corresponding change in water levels vs. time, for wells only screened in the designated aquifer, are req~ired. Figure A-1 (Dawson and lstok, 1991) is a hypothetical data set consisting of hydraulic head versus atmospheric pressure for a well screened in a confined aquifer. The slope of the fitted line in Figure A-1 is the ratio of the measured rate of change of hyraulic head to the corresponding rate of change of atmospheric pressure (in units of pressure /unit' weight of water) over the same time range. This ratio is termed the barometric efficiency and is used as a weighting factor to determine the relative amount of atmospheric change to remove from the well data (Dawson and Isok, 1991 ).
Barometric efficiency= (.6 I 1.0) = 60%

0.6

33

34

35

36

Atmospheric pressure expressed in terms of height of water (ft) (pressure /unit weight of water).

. . Figure A-1. Plot of hypothetical data set showing change of measured hydraulic, head vs. change of atmospheric pressure in the same units (ft) (modified from Dawson and Istok, 1991).

32
A BE of 1 indicates that 100% of the atmospheric pressure changes have been transmitted to the aquifer (normally confined) and that all of the relative atmospheric changes should be removed from the well data. A BE of 0 indicates that none of the atmospheric pressure changes have been transmitted to the aquifer (normally unconfined) and therefore none of the relative atmospheric pressure changes should be removed from the well data.

Calculating the Barometric Efficiency

Clemson University has adopted a graphical method for determining the BE of a

well by visually comparing the atmospheric pressure data to the BE corrected well data.

The following step-by-step procedures pertaining to this method are taken from Snipes

et al., 1996. Through the use of a spreadsheet, atmospheric pressure data and the BE

corrected well data are plotted on the same chart. Changing the BE of a well on the

spreadsheet, automatically changes all the BE corrected pressure values for that well on

b~th the spreadsheet (Table A-1, columns J,K and L) and on the chart (Figures A-2

through A-5). The value for BE is repeatedly guessed until no effects of atmospheric

pressure change can be seen in the well data. Figures A-2 through A-5 graphically show

that a barometric efficiency equafto .700 minimizes the effects of atmospheric pressure

change on the pumping well St. Marys 3 pressure data.
I

The following procedures are followed to correct the raw well data for

I

1.

atmospheric fluctuations:

I

' I '

. '

AI8 I

c

I

D

I

E

I H

J

I

K

l . -L

1

l

,

!

1

i"initials" = avg of 1 minute prior to pump on

I

2

I

i

1

!

I

!initial Pressure !initial Pressure /initial Pressure

3 Columns A- F from raw data

i

i

i

-v- 4 1) Plot atm and t/(BE corr press) varying BE to remove atm effects- for each well

St. Marys 3-PW 1St. Marys 1-0W 1St. Marys2-0W

45.4191

13.5011

13.577

5 2) Positive t's and BE corr pressures - copy to analysis spreadsheet

i

iBE factor

1BE factor

iBE factor

6

i

l

;

!

!

St. Marys 3-PW 1St. Marys 1-0W jSt. Marys2-0W

7

!

!

I

1

!initial atm

0.7001

0.900l

0.400'

8

I

/PW-raw

jOW1-raw

!OW2-raw

!

10.306

i

!



9

i .::..

!St. Marys 3 1St. Marys 1 iSt Marys 2 I

St. Marys 3-PW 1St. Marys 1-0W 1St. Marys2-0W

10 dt

\atm-raw... iWell Pressure !Well Pressure !Well Pressure lDelta P atm. BE corr press iBE corr Press iBE corr Press

11 (s)

I(m)

l(m)

i (m)

i (m)

i (m)

l(m)

!(m)

/(m)

296

-101

10.311

45.42!

13.501

13.581

0.001

0.00!

0.001

0.00

297

-71 .

10.311

45.42!

13.501

13.581

o.ooj

298

-51

10.31!

45.421

13.501

13.581

o.oo!

299

-31

10.311

45.421

13.50!

13.58/

0.001

300

-1!

10.311

45.421

13.50!

13.58i

O.OO!

301

11

10.311

43.131

13.50!

13.581

0.00!

302

31

10.31!

42.851

13.501

13.58i

0.001

o.oot o.oo1 0.001 O.OOi -2.29! -2.571

0.001 o.ool 0.001 0.00/ 0.001 0.001

.. 0.00
o.oo .,

0.00 0.00.

~\
:.

0.001

0.00

,

303

51

10.31!

41.97!

13.501

13.58!

0.001

-3.441

0.001

O.Oqj

304 dQ5 _

81

10.311

_10~1~ ...J0.311__ _

41.341 40.64\

13.501 13.501

13.581 13.581

0.001 o.ool

-4.08/ -4.781

0.00/ o.ool

O.OOj o.oo,

Table A-1. Barometric correction speadsheet, showing raw pressure values and BE corrected values for all wells in the St. Marys test (abbreviated from 1608 data points).

w w

Change in Water Level Vs. Time for Pumping Well St. Marys 3 and

.- 0.04

Atmospheric Pressure change (in m of water) Vs. Time

I... 0.03 ~

BE factor= 0.00

0

.5. 0.02

Ql
1:11 1:
.u! 0.01
:I!s! 0.00

~~ --atm

. r

Ill

l -0.01 J

PW

-0.02



1012/01 1012101 1012101 1012/01 10/2101 1012101 1012101 1012/01 10/2101 1012101 10/3101

. ,. ~ ~

0:00 2:24 4:48 7:12 9:36 12:00 14:24 16:48 19:12 21:36 0:00

date and time

Figure A-2. Pumping well St. Marys 3 background water-level change versus time relative to "pump on" ("BE factor= 0.00) and atmospheric pressure change versus time re~ative to "pump on" (units in meters ofh2o).

w
~

s~ 0.04
... i Ill
~ 0.03
0
.. 0.02
=Cll
r::
'5Ill 0.01
e
:I 0.00
Ul Ul
I!!
Q, -0.01

Change in Water Level Vs. Time for Pumping Well St. M~rys 3 and Atmospheric Pressure change (in m of water) Vs. Time
BE factor= 0.30
f
I
I
l

!-pwl
J-atmj

-0.02 +---.....---;----;------!---~--+--~::::-+-+---1---

10/2101 10/2/01 1012101 10/2/01 1012/01 10/2/01 10/2101 10/2/01 1012/01 10/2101 10/3/01

~E,r

0:00 2:24 4:48 7:12 9:36 12:00 14:24 16:48 19:12 21 :36 0:00

date and time

Figure A-3. Pumping well St. Marys 3 background water-level change versus time relative to "pump on" ("BE factor"= 0.30) and atmospheric pressure change versus time relative to pump on" (units in meters ofh2o).

w
VI

Change in Water Level Vs. Time for Pumping Well St. Marys 3 and

Atmospheric Pressure change (in m of water) Vs. Time

s-;:- 0.04
i Ill
~ 0.03

0
.s. 0.02 '

Ill

1 cCll

Ill
fi

0.01

=I!! 0.00
Ill Ill
! c. -0.01 .

BE factor = 1.00
"
~
II. It":":P\V

' . ~ ............ - ..... -

.1t~_v

I
l

. ----- .. ---- _,_ ..

~ m

~.02+.----~----~--~----------~----~--~--+--+----~--~
10/2/01 10/2/01 1012/01 10/2/01 1012/01 10/2/01 1012/01 10/2/01 10/2/01 10/2/01 10/3101 0:00 2:24 4:48 7:12 9:36 12:00 14:24 16:48 19:12 21:36 0:00
date and time

Figure A-4. Pumping well St. Marys 3 background water-level change versus time relative to "pump on" (BE=J.O) and atmospheric pressure change versus time relative to "pump on" (units in meters ofh2o). Note how the well data has been over-corrected. In the dashed box, the water level data spikes down as the atmospheric pressure data spikes up.

w
0\

...1:' 0.041
s; 0.03 ~
0
.,.5. 0.02
Gl
c
.u! 0.01 I:....!I,, 0.00
I!
Q, ..0.01

Change in Water Level Vs. Time for Pumping Well St. Marys 3 and Atmospheric Pressure change (in m of water) Vs. Time
BE factor= 0.60

1-pwl' j-atm

..0.02+.----~----~---+----~----~---+----~~--~---+--~

:d

1012101 1012101 1012101 1012101 1012101 1012101 1012101 10121il1 1012101 1o121o1 1013101 0:00 2:24 4:48 7:12 9:36 12:00 14:24 16:48 19:12 21:36 0:00

....~; j.

~

date and time

Figure A-5. Pumping well St. Marys 3 background water-level change versus time relative to "pump on" (BE=0,7) and atmospheric pressure change versus time relative to "pump on" (units in meters ofh2o).

if~
~.1~~ ~t ~ !
. 1 .\
. II
. "1$' A. I
>!
I
.
;~
.!;

w
-...l

38
(I) The raw well data and the atmospheric data, along with the elapsed time in relation to the time the pump was first turned on (dt), are entered into the barometric correction spreadsheet (Table A-I, columns A-E). (2) An initial value for atmospheric pressure (Table A-I, cell H-8) (taken as the average atmospheric pressure data points just prior to pump on) is subtracted from the raw atmospheric pressure data (Table A-1, column B) to calculate the relative change in atmospheric pressure (Delta P atm) (Table A-1, column H) from a set datum. (3) The relative well 'pressure (water level) change for a well is calculated by subtracting an ititial well pressure (taken as the average well pressure just prior to pump on) (Table A-1, row 4) from the raw well pressure data. The effects on the raw well pressure due to
atmospheric pressure changes are then removed by subtracting the (Delta P atm * BE
factor) from the raw relative water level changes to give the BE corrected relative water level changes (Table A-1, columns J, K and L).
Trend Corrections Trends in well data occur as a response to periodic fluctuations in the combined effects of recharge and discharge to the aquifer. The main factor in causing water level trends, is the seasonal fluctuations in recharge to the aquifer (with periods on the order of months). Trend corrections are necessary if the test occurs at a time when the aquifer pressure is rapidly decreasing or increasing due to seasonal fluctuations or other factors. It becomes more difficult to correct for if the trend is changing during the pumpon phase of the test. With sufficient background and recovery data, most long term linear
. '
trends can be recognized and easily removed by a simply subtracting the quantified trend

, !. ...

~-, .

~~ .-

39 .. ;. .. .~

effect from the barometric and tidal corrected data. Due to a malfunction in the data

logger used to collect background water level data at the St. Marys test site, no

background data was available, therefore there was no trend correction applied to the

well data.

40
AppendixB The Time Transient Spreadsheet Analysis Method Applied to the St. Marys Well Data

41

Theis Equation To analyze pump test data, Clemson University uses a spreadsheet curve matching method based on the Theis equation. Radial flow to a well is mathematically modeled using the 'Theis equation. The Theis equation for radial flow to a well states that the drawdown (s) in a well is proportional to the flow rate and the well function ( W (u)), and is inversely proportional to the transmissivity (Snipes et al., 1996). The Theis equation for radial flow to a well is written:

s = _2_W(u)

(1)

41tT

where

and W(u)= J ""e -z~ . u z

and

s = drawdown from static water level in the pump welJ or observation welJ, units in lenth [L],
Q = discharge rate from the pumped well, units in volume per time [ L3 rr ],
e T = transmissivity of the aquifer, units in area per time [ IT],

S =coefficient of storage (storativity) of the aquifer, in dimensionless units,

r = radius, the radius of the pumping well or if analyzing the observation well data, the distance from the pumping well to the observation well, units in length [L],

t =time, t = 0 for initial "pump on",

W(u) = infinite series well function, in dimensionless units,

42
u = the lower limit of integration of the infinite series, in dimensionless units. The well function approximations are given below.
for u < 1 for pumping well or a nearby observation well and/or late pump time
where
Ao = -0.57721566,
AI= 0.99999193, A2 = -0.24991055, A3 = 0.05519968, ~ = -0.00976004, A5 = 0.00107857. for u > 1 for a distant observation well and/or early pump time

where

a1 = 8.5733287401 a2 =18.059016973

bl = 9.5733223454 b2 = 25.6329561486

a3 = 8.6347608925 a4 = 0.267737343

b3 = 21.0996530827 b4 = 3.9584969228

The Theis equation for radial flow to a well is valid under the following assumptions:

1. The aquifer has a constant thickness and has an infinite horizontal extent.

2. The potentiometric surface of the aquifer is horizontal and not changing with time prior to pumping.

.. ".

. .., -:-- ..,- ~ .~ .,....

43

3. All changes in the potentiometric surface of the aquifer are due only to the effect of the pumping well.
4. The pumping well has an infinitesimal diameter and is 100% efficient. 5. There is no source of techarge to the aquifer. 6. Water is released instantaneously from the aquifer as the head is lowered. 7. The well is pumped at a constant rate.
Procedure The Clemson University Department of Geological Sciences uses a computer spreadsheet algorithm based on the Theis equation which enables the interpreter to curve
I match a computed drawdown with the measured drawdown to determine an estimate of the hydraulic parameters for a tested aquifer. The time transient spreadsheet analysis method is adapted from the Theis (log-log) type curve overlay method, which also is based on matching a theoretical drawdown curve with the measured drawdown curve. In the Theis method, a type curve is formed by the log-log plot of W(u) vs. 1/u and is matched to the log-log plot of the measured dniwdown curve (drawdown vs. time). Whereas in the spreadsheet method, a theoretical semi-log drawdown curve, created by the computer spreadsheet's Theis equation based algorithm, is matched to the measured semi-log drawdown curve. The semi-log plot used in the spreadsheet method is the same configuration used in the Jacob straight-line method of analysis (Cooper and Jacob, 1946). Clemson University has adopted the time transient spreadsheet analysis method as the preferred method of pump test analyzis because it allows for more flexibility in defining the controlling physical parameters of the test. In the spreadsheet method,flow rate changes are accounted for, as well as skin effects due to well development. The

44

spreadsheet can also be modified to account for vertical leakage across an aquitard and

well bore storage effects. The time transient spreadsheet analysis method is explained m

the following steps (Snipes et al., 1996).

(1) The measured drawdown (all corrections applied) and the elapsed time in relation to the time the pump was first turned on (dt) are entered into the analysis spreadsheet (Table B-1, columns A and B).

(2) The measured drawdown versus time is plotted on a spreadsheet chart (Figure 8-1 ).

(3) A theoretical drawdown for each time step (dt) is calculated (Table B-1, column D).

2

This is based on the Theis equation (l) where u = r S .' The radius ( r ) is briven as

-

4tT

either the radius of the well in the analysis of the pumping well data or is the linear

distance from the pumping weB to the observation well in the case of observation weB

data analysis. For observation well data analysis, values for T and S (Table B-1, cells G l

and 02) are guessed by the interpreter until the best curve match is obtained for the

theoretical drawdown versus the measured drawdown. In the case of pumping well data

analysis, an additional variable (skin factor) (Table B-1, cell 03) is required along with T

, and S. A full explanation of the skin factor and the inclusion of the skin factor and flow

rate cha~ges in the Theis equation are discussed below.

(a) This spreadsheet is designed for 4 flow rate changes (Table B-1, cells A1 through A4), an initial rate for the time the pump was. first turned on followed by 2 possible additional flow rate changes and then a recovery period where the flow rate is equal to 0 when the pump is cut off. The time relative to "pump on", when each flow rate change begins, is entered into the cell adjacent to the flow rate (Table 8-1, cells B 1 through B4). The spreadsheet can be easily modified to account for additional flow rate changes if necessary. The Theis equation in the spreadsheet is modified to account for multiple flow rates in the following manner (Freeze and Cherry, 1979):

s=AP=t Qn-Qn-[w(un)]

(2)

n=l

47tT

where

and

dt- tn Qn- Qn-J

is the time since the last flow rate changed, units in time, is the change in the flow rate, units in volume per time [L3ff].

Therefore, for wells with multiple flow rate steps, a u value (Table B-1, columns F

through I) and a step drawdown (Table B-1, columns J through M) are calculated for

each flow rate change. For the St. Marys pump test, three flow rates where used in the

analysis. The flow rates consisted of an initial flow rate of 71.00 gpm continuing from

,

.

Q1 (gpm)= Q2 (gpm)= Q3 (gpm)= Q4 [gpm)=
RAH 3-10/95 I

71.00 71.40 0.00 0.00

t1 (s)= t2 (s)= t3 (s)= t4 [s)=

0 24,491 259,191 259 191

S = T (mA2fday) =
skin factor= K (m/davl = darcys =

0~0001 : 50;0 / ?3;44~
5.47 6.59

L (ft) = d (ft) = z (ft) = b (ft) = rad (ft) =

20 full pen.=

0.001 I

15 d =distance from top of AQ to top of screen

17.5 z =distance from top of AQ to middle ofscreen

30 I = distance from top of AQ to bottom of screen

0.1667 b = AQ thickness

measured

dt(pumpon) drawdown

(seconds)

(meters)

1 2.29

3 2.57

5 3.44

8 4.08

10 4.78

12 5.36

14 5.88

16 6.34

19 6.74

time (mins)
0.0 0.1 0.1 0.1 0.2 0.2 0.2 0.3 0.3

calculated drawdown
(meters) 9.34 10.04 10.36 10.57 10.72 10.84 10.95 11.03 11.11

residual 259.42 109.11 122.66 105.13 92.55 77.59 66.20 56.39 48.40 41.88

I care u

I I

I calc step drawdowns

I

dt-t_1

dt-t_2 dt-t_3 dt-t_4 pump on

flowd1

flowd2

from Q1

from Q2 from Q3 from Q4

Q1

Q2

Q3

u1

u2

u3

u4

delta h1

delta h2

delta h3

1.11 E-04 not yet not yet not yet

9.34

0.00

0.00

3.49E-05 not vet not vet not vet

10.04

0.00

0.00

2.07E-05 not yet not yet not yet

10.36

0.00

0.00

1.47E-05 not 'l_et not yet not_yet

10.57

0.00

0.00

1.14E-05 not vet not vet not yet

10.72

0.00

0.00

9.30E-06 not yet not yet not yet

10.84

0.00

0.00

7.86E-06 not vet not vet not vet

10.95

0.00

0.00

6.80E-06 not yet not yet not yet

11.03

0.00

0.00

6.00E-06 not vet not vet not vet

11.11

0.00

0.00

recovery Q4
delta h4 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Table B-1. Pump test analysis spreadsheet for pumping well St. Marys 3.

~

\

'rt

~ Vl

Drawdown (barometric corrected) Vs. Time for Pumping Well St. Marys 3

.-s;...

1.000 -1.000 -3.000 -5.000

0

sr! -7.000
Cl) -9.000

I . .. -11.ooo

!
:::J

-13.000

.men -15.ooo

c. -17.000

10/1/01

0:00

I
10/2/01 0:00

I
'
10/3/01
0:00

10/4/01 0:00

10/5/01 0:00

r-

'
I
10/6/01 0:00

'
1onto1 0:00

I
I
I
I
I
10/8/01 0:00

date and time

Figure B-1. Measured drawdown versus time for pumping well St. Marys 3 (barometric correction factor applied).

I
I
I
I
I
I
I
I
I
I
I
!
I
I
I
I I
I
I
I I I
I
I I
"0""\

47

"pump on" .(t = 0 seconds) until the relative time to "pump on" was equal to 24,491

seconds, followed by a flow rate of71.40 gpm and continuing to "pump off' (t = 259,191

seconds, approximately 72 hours) which was then followed by a flow rate of 0 gpm

(recovery) which continued for the remainder of the recovery period. Figure B-2 is a plot

of flow rate versus time for the durat~on of the pumping phase of the test, showing flow

rate steps used in the analysis.



(b) For pumping well data analysis, a skin factor variable (Table B-1, cell G3) must be considered (in addition toT and S) to produce realistic test results. The skin factor is a numerical representation of the head Joss due to the "damage" to the annular region around an installed well. This region, called the skin zone, is a region where the permeability has been altered due to drilling and/or well completion activities (Kroening et al., 1996). The manipulation of three variables (skin, T and S) in the analysis of pumping well data results in unrealistic interpretions, therefore one of the three variables must be held constant. To accomodate for this, it is best to have an observation well screened in the same aquifer whereby the observation well data is analyzed first to determine an estimate of the storativity of the aquifer. This is done by setting the skin factor equal to 0 due to the low groundwater velocity at the observation well: This eliminates the skin factor as a variable in the observation well analysis, thereby allowing an estimate of the storativity of the aquifer to be computed (as well as an estimate of the transmissivity). This estimatetf storativity is then entered into the pumping well data analysis as a constant (S is much less sensitive than T) which reduces the number of variables to two (T and skin). For a single well pump test, a storativity value must be estimated from other tests performed in the area. This is a major weakness of a single
skin ~( :. -l}n(:J well pump test. The skin factor (skin), is mathematically defined as (Earlougher, 1977): (3)

where k ks rs rw

is the permeability of the formation, units in area [L2], is the permeability of the skin zone, units in area [U], is the radial thickness of the skin zone, units in length [L], is the radius of the well, units in length [L].

The above equation is based on the following assumptions:

1. If a radial assymetric orientation of the skin zone exists, the effects of

differential clogging or improvement will be averaged into a single,

symmetric radial representation of the zone.

2. The skin zone and surrounding aquifer are homogenous and isotropic.

I

l

1 1

The

total

head

loss

or

drawdown

(s)

at

a

pumping

well

is

accounted

for

by

adding

the

effects of the aquifer and the skin zone (due to different permeabilities), which is
I expressed in the following equation (Van Everdingen, 1953):
I

t
'

\

I '
I j_-

Flow Rate Vs. Time for the St. Marys Pump Test

73.00

72.50

I.
I

-eEcn. 72.00
-Q) 71.50
....c.a.

~ 71.00

~

70.50

70.00 +,---+--+-------,__-----+----+--
10/1/97 0:00 10/2/97 0:00 10/3/97 0:00 10/4/97 0:00 10/5/97 0:00
date and time

'
Figure B-2. Graph of flow rate versus time over the duration of the pumping phase of the St. Marys pump test, showing flow rate steps used in the analysis.
.t:.
OQ

.> .-r-~.c:.-"'1 ..... ~

, .
o -.,.... ..~ ~ ',..

49

s = _g__ Ins__+ skinQ

(4)

21tT rw 21tT

where s Q T rc rw

is the total drawdown or head loss, units in length [L]; is the volumetric flow rate, units in volume per time [L3ff], . is the transmissivity of the aquifer, units in area per time [L2ff],
is the radius of influence of the pumping well, units in length [L],
is the radius of the pumping well, units in length [L],

_g__,ns__ is the head loss at the pumping well due to the aquifer, units in length [L], 21tT rw

skinQ is the head loss at the pumping well due to the skin zone, units in length [L]. 21tT

The Theis equation which is modified to take into account the skin factor for a fully penetrating pumping well then becomes (Van Everdingen, 1953):

s = _g__[w( u) + 2skin].

(5)

41tT

(c) From equations (2) and (5), the modified Theis equation used in the spreadsheet,

which takes into account both multiple flow rates and skin factor is (Snipes et al., 1996):

I S= AP= Qn -Qn-l [w(un)+2skin].

(6)

n=l 41tT

(d) The W(u) for each flow step is calculated by a Fortran program within the spreadsheet based on the previously mentioned approximations.

(e) The step drawdowns are then summed to give the total calculated drawdown (Table B-1, column D).

(4) The theoretical "calculated" data is plotted on the same chart as the measured data. As discussed in step (3), each theoretical data point is calculated based on guesses forT ' and S for observation well data analysis and guesses for T and skin for pumping well data ! analysis.

, (5) The guesses forT and S (observation well data a.nalysis) or T and skin (pumping well , data analysis) are varied by the interpreter and entered into the spreadsheet whereby the
plots of the measured and calculated drawdown are compared. By varying the T, the interpeter can change both the slope (T appears in both the equation for u and the , equation for drawdown) and the vertical displacement of the calculated curve. Varying the S or skin changes only the vertical displacement of the calculated curve. A time weighted residual value (Table B-1, cell E 10) enables the interpreter to produce the best possible curve match. The lowest residual value results in the best curve match. The equation for the residual value in the spreadsheet is:

50

n-1

residual=L:(calculatedn

-measured 0

)

2

*(to

-tn+l)

l

where

n

is the number of data points (time steps) in the pumping phase,

starting at t-1 000 seconds (early time measured data does not

follow the theoretical drawdown curve due to well

bore storage effects),

calculatedn is the calculated drawdown at each time step,

measuredn is the measured drawdown at each time step,

t

is the elapsed time in relation to "pump on" at each time step (n).

In the St. Marys test, data from the observation wells was not used to calculate storativity for the pumped aquifer (aquifer X) because observation well St. Marys 2 was not screened in the same aquifer and St. Marys 1 was unscreened with an unknown open interval. Therefore only the punping well data was analyzed to determine estimates for T and the pumping well skin factor using a typical value for storativity for a confined aquifer of 0.0001. To calculate a more realistic estimate of the storativity for the aquifer would require a step drawdown pump test. Figure B-3 shows the curve match using initial guesses forT and skin, with an estimated storativity of 0.0001, to begin the analysis for the pumping well St. Marys 3. Note that the calculated drawdown curve is well above the measured drawdown curve and therefore is not a match. By increasing the guess for the skin factor, the calculated drawdown curve will increase and move toward the measured drawdown curve as shown in Figure B-4. By increasing the skin factor from 0.00 to 10.00, the vertical position of the theoretical drawdown approximately matches the vertical position of the measured drawdown (Figure B-4). However, the calculated drawdown slope is not steep enough to match the measured drawdown. Decreasing the guess for T will increase the slope as well as the vertical displacement of the theoretical drawdown curve. InFigure 8-5, the T is decreased to 50 m2/day which results in a good match for the slopes of the two drawdown curves, although the theoretical drawdown curve is significantly lower than the measured drawdown curve. By decreasing the skin factor from 10 to 5, as shown in Figure 8-6, the theoretical drawdown is increased which moves the theoretical drawdown curve closer to the measured drawdown curve. Decreasing the skin factor from 5 to 3.44 results in the best curve match according to the lowest possible residual value (Figure B-7). The further manipulation of T and/or skin by small increments does not result in a decrease in the residual value. Early time data (0 to 1000 seconds) is not used in the curve match analysis because the measured data is skewed due to well bore storage effects. Late time measured data can also be skewed due to certain hydrologic conditions. The measured drawdown can be less than expected due to a water source (i.e., leakage, constant head source) or it can be more than expected (i.e., impermeable boundary). ~ecovery analysis is possible if a check valve is installed oil the pump which would not allow the water contained in the. discharge apparatus to rush back into the well bore. In the St. Marys test, a check valve was installed on the pump, making recovery analysis possible.

- . - . :.,,..... ,,.....r..::.:., ...,-7.'..,. ;...f..~i.~-
51

St. Marys 3 Pumping Well Theis Curve Match

5.000

g

c

~ 0

11.000

"0

~

cI! 15.000

100

S=0.0001 T=BO m112/day skin=O.OO

1000

10000

Time (sec)

100000

1000000 --calculated drawdown --observed draw down

------------------------------
Figure 8-3. Semi-log curve match of measured drawdown (all correction factors applied) and Theis equation based calculated drawdown for pumping well St. Marys 3, using a storativity of 0.0001, and intiat guesses for T (80m2/day) and skin (0.00). The value for S (0.0001) was not able to be estimated from the observation well data analysis and is therefore assumed, based on storativity values from other confined aquifers.
----------------------
St. Marys 3 Pumping Well Theis Curve Match

5.000.

E 7.000.
c 9.000 ..

~ 11.000. ,'

0

"0 ~

13.000 ..

cI! 15.000.

17.000

100

5=0.0001 T=BO mii2Jday . skin=10.00

1000

10000

100000

Time (sec)

F - 1000000

calculated draw down

. 1=--observed draw down

Figure 8-4. Semi-log plot of pumping well St. Marys 3 theoretical calculated drawdown
and measured drawdown (all correction factors applied), using a value for S of 0.0001 and guesses forT of 80m2/day and a skin of 10.00.

52

St. Marys 3 Pumping Well Theis Curve Match
5=0.0001 T=50 m"2/day skin=10.00

1000

10000

100000

Time (sec)

~ 1000000 ---ca-lc-u-la-te-d-d-ra-w -do-wnJ --observed draw down ------------

Figure 8-5. Semi-log plot of pumping well St. Marys 3 theoretical calculated drawdown . and measured drawdown (all correction factors applied), using a value for S of 0.0001
and guesses forT or' 50m2/day and a skin of, 10.00. Notice the slope of the theoretical drawdown approximately matches the slope of the measured drawdown.
---'------------------------------------------------------St. Marys 3 Pumping Well Theis Curve Match

5.000.

! 7.000. 9.000.

c
~
1
l!

11.000. 13.000. 15.000.

Q 17.000.

19.000.

100

5=0.0001 T=50 m"2/day skin=5

1000

10000

Time (sec)

100000

------------1 1000000

-calculated draw down



--observed draw down

- ----------------------

Figure 8-6. Plot of pumping well St. Marys 3 theoretical calculated drawdown and
measured drawdown (all correction factors applied), using a value for S of 0.0001 and guesses forT of 50m2/day and a skin of 5.00.

53

St. Marys 3 Pumping Well Theis Curve Match

5.000.

g 7.000. 9.000.

, c
3:

11.000

0 13.000.

3: 15.000 .
c1.'!! 17.000 ..

19.000 ..

100

5=0.0001
T=SO m112/day
skin=3.44

1000

10000 100000

Time (sec)

r=-- 1000000

cal~~lated draw down

--observed draw down

-----------------
Figure 8-7. Plot of pumping well St. Marys 3 theoretical calculated drawdown and measured drawdown (all correction factors applied) using a value for S of 0.0001 and guesses for T of 50 m2/day and a skin of 3.44. This is the best curve match possible according to the lowest residual value.

(6) The well efficiency of a pumping well is calculated by determining the ratio of the theoretical calculated drawdown with a skin factor of 0 to the actual measured drawdown. A perfectly efficient well would have a well efficiency equal to one. In the case of the St. Marys pump test, a storativity value for the aquifer had to be assumed because it could not be estimated from observation well data analysis. The skin factor in , the pumping well data analysis depends on the accuracy of storativity estimate and therefore makes the computation of the well efficiency invalid. To get a more accurate estimate of the skin factor (and therefore storativity) in the pumping well without the application of observation well data analysis, a step drawdown pump test on the pumping. well would be neccesary. Assuming the value for storativity is accurate, the well efficiency of the pumping well St. Marys 3 is 73.8% as shown in Figure 8-8.

:

.... - .

54

9.000. 10.000

Drawdown vs. Time for St. Marys 3 Showing Measured

Drawdown and Theoretical Drawdown

. T"F';'I

T=50 m112/day
S=0.0001

11.000 .
E 12.000. i0 13.000
I 14.000 .
a
15.000- .

..

---calculated drawdown with a

skin factor of 0

--observed drawdown

16.000 .

17.000. 1000

10000

100000

1000000

Time (sec)

Figure B-8. Plots of pumping well St. Marys 3 measured drawdown (all correction
factors applied) and theoretical drawdown with skin = 0, T = 2000 m2/day and S =
0.0001, .showing well efficiency calculations. A well efficiency of73.8% was calculated but may not be realistic due to the storativity (and therefore the pumping well skin factor) not being able to be estimated from the data analysis.

Editor: William M. Steele
Copies: 250 Cost: $1900.00
The Department of Natural Resources (DNR) is an equal opportunity employer and offers all persons the opportunity to compete and participate in each area of DNR employment regardless of race, color, religion, national
origin, age, handicap, or other non-merit factors.