HomeMy WebLinkAboutNCD122263825_20150827_JFD Electronics - Channel Master_FRBCERCLA SPD_Draft Proposed Buffer Zone Around Area A for LUR-OCR-~---.--• •·
. . WArERS EDGE ENVIRONMENTAL, LLC
. , 4901 WATERS EDGE DRIVE, SUITE 201 • RALEIGH, NC 27606 • PHONE 919.859.9987 • FAX919.859.9930
Date
To
From
Subject
,
MEMORANDUM l\'b';)\0; lL. -~\
August 27, 2015 :f?f!>_,.-(J.C"'i'.,\¢'0 '
"' \'.>-' ,-,r;:,\S '.". st? '-Ms. Mary Siedlecki-NCDENR -" \\S ; \ .. ~oo . \\ . \\"'L <;,ec\\O . /
Mr. Phil Rahn-Waters Edge {b~ ,~-.s\C /,
t\2~. _,,,:J/ . l..{(,.r, -~·,-: l _,.,
Draft Proposed Buffer Zone Around Area A for LUR Submittal~·
Daimler Facility
Cleveland, North Carolina
I have been tl)'ing to research a method of calculating a "Buffer Zone" around Area A for
use as an Area B for the land use restrictions (LUR) at the Daimler facility in Cleveland,
North Carolina You had referred me to another facility (Schneider Electric site in
Knightdale, NC) which had undergone a similar simulation as an example of something
that would be acceptable to NCDENR. We were able to download the Schneider Facility
Remedial Action Plan supplement (Appendix F), and did determine that the consultant
had undergone integrating a MODFLOW calibrated groundwater model which we
believe was primarily used to predict the downgradient extent of contaminant impact but
then was also used for this capture zone simulation. The corresponding difficulty is that
this MODFLOW calibrated groundwater model is a considerable costly undertaking and
we again believe that the primary use of this model was for contaminant plume extent
simulations.
Our belief is that there may be other less expensive methods of calculating this capture ~= zone that would not involve a calibrated groundwater model. Based on some additional
research, we did find an article titled Domestic Well Capture Zone and InJluence of_Jf)f, ••
Gravel Pack Length (see Apriendix A). On Page 3, Paragrarih 2, it states "where regional
grounClwat#flowisdbminant and-local rechargeTnegligib'te,_the cariture zone ofa
domestic well can also be easily comriuted if the well fulliJ:>eootralae 119ui"fer system_
ordoes not stronglY. affect regional groundwater flow. The widih, w, ofihe capture zone .... =,=.;,~ ' -,
of a domestic well is then obtained by simple mass balance (Todd 1980):
w = Ql(I'*i)
where Q is the pumping rate, T is the aquifer transmissivity, ?P.d~Liscthe::regional::::i
• V hydraulic gradient.
Pumping Rate (0)-We did find a reference for the pumping rate (yield) in an article titled
Statistical Analysis Relating Well Yield to Construction Practices and Siting of Wells in
Domestic Well Capture Zone
Daimler Facility
Cleveland, North Carolina
August 27, 2015
Page 2 of2
the Piedmont and Blue Ridge Provinces of North Carolina (see Appendix B). This
article cross referenced well yields to both hydrogeologic conditions and topography and
we were able to find that the average well yield for the hydrogeologic conditions (only
past well log that referenced stratigraphy was from MW-I (boring log included as
Appendix C) which described the bedrock type as a meta-diorite [Mil]-see Appendix B)
and topography (we believe that this site would be described as being situated on a hill or
ridge) identified at the Daimler site was an average of 13.3 gallons per minutes (GPM-
Appendix 8-Page A23-Table 7). Converting 13.3 GPM to cubic meters per minute is as
follows:
13.3 GPM * 0.003785 m3/gallon = 0.05034 m3/minute
=72.49 m3/day
=26,458.97 m3/year
Transmissivity m-We found a reference titled Classification of Transmissivity
Magnitude and Variation (Appendix D-Page 231 Table 1-Classification of
Transmissivity Magnitude) which provides an "intermediate" coefficient of
transmissivity at I 00 m2/day. Converting I 00 m2/day to m2/year is as follows:
100 m2/day x 365 days/year= 36,500 m2/year
Hydraulic Gradient (i). The most recently calculated hydraulic gradient from the May
20 I 5 groundwater report is 0. 022 f1/ft.
Using these variables, the derived formula would be as follows:
w = 26,458.97 m3/year/(36,500 m2/year * 0.022)
= 32.95 m * 3.28 feet/meter
= 108.08-foot capture zone width
We have depicted a 110' capture zone width for Area Bon Figure 1 for review purposes.
We would like to use this as our Area B Zone for the LUR submittal and would ask that
you review this submittal to see if this is acceptable.
FIGURES
APPENDIX A
Domestic Well Capture Zone an :, Influence of the Gravel Pack Length
2 Judith E. Hom and Thomas Harter'
3 Department of Land, Air, and Water Resources
4 University of California
5 Davis, CA 95616-8629
6 • (corresponding author; ThHarter@ucdavis.edu; 530-752-2709)
7
8 Accepted for publication in the journal "Ground Water", September 2008
9
10 Abstract
11 Domestic wells in North America and elsewhere are typically constructed at relatively shallow
12 depths and with the sand or gravel pack extending far above the intake screen of the well (shallow
13 well seal). The source areas of these domestic wells and the effect of an extended gravel pack on the
14 source area are typically unknown and few resources exist for estimating these. In this paper, we use
15 detailed, high-resolution groundwater modeling to estimate the capture zone (source area) of a
16 typical domestic well located in an alluvial aquifer. Results for a wide range of aquifer and gravel
17 pack hydraulic conductivities are compared to a simple analytical model. Correction factors for the
18 analytical model are computed based on statistical regression of the numerical results against the
19 analytical model. This tool can be applied to estimate the source area of a domestic well for a wide
20 range of conditions. We show that an extended gravel pack above the well screen maY. co11trib_ute
21 significantly to the overall inflow to a domestic well, especially in less permeable aquifers, where
22 that contribution may range from 20% to 50%; and that an extended gravel pack may lead to a
23 significantly elongated capture zone, in some instances nearly doubling the length of the capture
I Hom
24 zone. Extending the gravel pack much above the intake screen therefore significantly increases the
25 vulnerability of the water source.
26
27 Introduction
28 Most households in rural areas of the United States, outside the service area of incorporated cities,
29 rely on domestic wells for their water supply (McCray 2005, U.S. EPA 1997). And many of these
30 domestic wells are constructed with a well-screen at depth and a sand or gravel pack that extends
31 upward to the mandatory minimum depth of the well seal, which is dictated by local and state
32 guidelines. A question commonly asked by homeowners is: Where does our water come from? The
33 capture zones (also referred to as the source area or recharge area) of domestic wells are rarely
34 determined. Attention has instead focused on public supply wells and their capture zones as these are
35 regulated through U.S. EPA's source water protection program. Domestic wells, typically serving a
36 single family, are often constructed to relatively shallow depths when compared to public or
37 municipal water supply wells (Burow et al. 2004).
38
39
40
41
42
43
44
45
Methods for delineating well capture zones range from very simple to very complex. In general, the
various approaches fall into four categories (Harter, 2008):
1.c1aeometric or graphical methods involve the use of a pre-determined fixed radius without any
special consideration of the flow system, or the use of simplified shapes_ that have been pre-
calculated for a range of pumping and aquifer conditions.
2. ~alytical methods allow calculation of distances for protection zones using equations that can be
solved using a hand calculator or microcomputer spreadsheet program.
2 Hom
46 3. Hydrogeologic mapping involves identifying the recharge zone and the source zone based on
47 geomorphic, geologic, hydrologic, and hydrochemical characteristics of an aquifer.
48 4. Computer modeling methods involve devising, calibrating, and applying complex analytical or
49 numerical models that simulate groundwater flow and contaminant transport processes.
50 The long-term average pumping rate of domestic wells typically ranges from less than 4 L/min [ I =====
51 gallon/min) to 20 L/min (5 gallon/min]. Using the graphical method employed by California's
52 Drinking Water Source Assessment and Protection (DWSAP) Program (California OHS, 1999), for
53 example, the default source area of a domestic well pumping 1,233.5 m3/year (I acre-foot per year,
54 the typical annual consumption ofa U.S. single family household) is a circle with a radius of 15 m
55 (~50 ft) for areal recharge of 450 mm/year (typical for very humid areas or rural residences in semi-<:::e:=:...,.
56 arid areas surrounded by irrigated lawn and fields) or with a radius of 31 m ( ~ 100ft) at a recharge
57 rate of 100 mm/year (typical of many semi-arid regions). This simple geometric approach neglects
58 the effects of the regional groundwater flow on the capture zone of a domestic well.
59
60 On the other hand, where regional groundwater flow is dominant and local recharge is negligible, the
61 c~(Jture zone of a domestic well can also be easily computed if the well fully penetrates the aquifer
62 sxstem or does not strong!y affect regional groundwater flow. The width, w, of the capture zone of a
63 domestic well is then obtained by simple mass balance (Todd 1980):
64 w = QI (T* i)
65 where Q is the pumping rate, Tis the aquifer transmissivity, and i is the regional hydraulic gradient.
66 For example, at a relatively low transmissivity, T, of 10 m2/d, a regional groundwater gradient of
67 0.5% and a pumping rate of 1,233 m3/year, the width of the capture zone is approximately 60 m
3 Horn
68 (~200 feel). Al values of Ttypical for productive aquifers, the width of the capture zone is often on
69 the order of I m -IO m ( ~ 3 feet -~ 30 feet) or even less.
70
71 Both, the geometric approach and equation (I) above provide simple approximations for extremely
72 idealized conditions. Here, our objective is lo determine the capture zone of a domestic well with a
73 sand or gravel pack, completed in an unconfined aquifer,1where both, recharge and regional
74 groundwater flow are significant. We use high-resolution computer simulations to determine the
75 source area and to explicitly determine the influence of the gravel pack on the well capture zone. For
76 reference, we compare those to a simple analytical model of the capture zone for a low-producing
77 well in an unconfined aquifer with recharge. Our study's focus is on rural domestic wells in irrigated
78 agricultural regions, e.g., of the Southwestern United States, where significant recharge is due to
79 irrigation return flows and much of the groundwater production is for irrigation purposes. Our
80 findings have general implications that are independent of this particular climate scenario.
81
82 Conceptual Framework
83 Domestic wells in rural areas are assumed to be completed near the uppermost portion of a regional
84 aquifer system. Furthermore, we assume that Dsignificant downward gradientfexists in the regional
85 aquifer system due to recharge at the water table and due to significant groundwater production
86 (mostly for irrigation) from the deeper portions of the aquifer system (e.g., Belitz and Phillips,
87 1995). Burow et al. (2004), for example, report typical recharge rates in irrigated areas in the San
88 Joaquin Valley, California, lo be on the order of 550 -750 mm/a with the majority of recharge
89 originating from irrigation return flows. For simplicity, regional groundwater flow is considered lo
90 be uniform around the source area of the domestic well and at steady-state. The superposition of
91 regional groundwater flow with the downward gradient induced by water table recharge and deeper
4 Hom
92 gro1IDdwater production yields a gro1IDdwater flow field that is vertically inclined relative to the
93 slope of the water table (Figure 1).
94
95
96
97
98
99
100
IOI
102
103
104
105
106
107
108
109
110
Ill
plan view
cross soction
.-capture zone-1
width o/\ . · · \
1--lenglh ' --<
I ! I x.
C
Cl> e u UI
a) z, b) L--'--'--...V
t--whole capture zone ----t
! elongated part ! main part !
Figure 1: Conceptual framework of gro1IDdwater flow towards a partially penetrating domestic well
(a) without and (b) with a gravel pack that extends for several meters to several tens of meters above
the well screen. Top: plan view, bottom: cross-sectional view. Regional gro1IDdwater flow is from
right to left with a vertical flow component controlled by IIDiform recharge at the top and aquifer
pumping from large production wells dispersed in the deep part of the aquifer below. The aquifer
bottom is assumed to be much deeper than the typical depth of the (relatively shallow) domestic
well; I: length, w: width. All other symbols: see text for details.
A simple method to compute the approximate source area size and location is available, if we
neglect the effects of the gravel pack and the effects of domestic well pumping, Q, on the local
gro1IDdwater flow field. Then, the source area location is obtained from the length and depth of the
domestic well screen, and from the angle, w, of gro1IDdwater flow relative to the slope of the water
table (Figure la):
Zo-Zl
X1 = X w + --(3) tanw
,,,.. = x, -x. (4)
5 Horn
I 12
113
114
115
116
117
118
119
A,,_0 = QI R (5)
A w = --1!!!!!.. (6) ,1-0 I ,,..
where Xw is the location of the well (along the regional groundwater gradient), Xh is the location of
the downgradient edge of the recharge (source) area, x, is the location of the updgractient edge of the
recharge area, z0 is the elevation of the water table, Zh is the elevation of the top the well screen, z, is
the elevation of the bottom of the well screen (Figure la), l,1,eo, w,heo, A,heo are the theoretical length,
width, and area of the recharge zone, and:
tan w =RI (Kh • i) (7)
120 where R is the uniform recharge rate, Kh is aquifer hydraulic conductivity, and i is the regional
121 hydraulic gradient. Equations (2)-(7) provide a simple analytical model to determine the capture
122 zone of a domestic well in an unconfined aquifer with uniform flow, recharge, and deep production.
123
124 To account for the influence of domestic well pumping on the local groundwater flow system around
125 the well and to account specifically for the influence of the gravel pack on the recharge area (Figure
126 lb), we constructed a numerical model, described in the next section.
127
128 Modeling Methods
129 The capture zone of a domestic well with a gravel pack is computed for a fully three-dimensional
130 steady-state groundwater flow field. The steady-state head and flux distribution are computed using
131 the MOD FLOW groundwater flow model (McDonald and Harbaugh 1988). The capture zone
132 corresponding to a particular groundwater flow solution is delineated using the backward particle
133 tracking model MODPATH (Pollock 1994).
134
6 Hom
135 Briefly, MODFLOW solves the steady-state groundwater flow equation
136 VK"vh=O (8)
137 where his the hydraulic head, by using a fully three-dimensional block-centered finite difference
138 scheme for the user-specified boundary conditions, K is the hydraulic conductivity tensor. In the
139 following simulations pumping induces only a small drawdown of the piezometric surface, so the
140 linear flow model (8) is sufficiently accurate for our purposes. We effectively invoke the Dupuit
141 assumption equivalent to the MODFLOW "unconfined layer" algorithm. There, the unconfined layer
142 thicknesses are set constant and only updated iteratively. From the hydraulic head solution,
143 MODFLOW also computes the flux, q, across each of the six faces of each finite difference cell in
144 the modeling domain. The flux solution becomes input to MODPATI-1, which computes backward
145 particle travel paths given the linear groundwater velocity, v = qln, where n is the effective porosity,
146 across each finite difference cell face. Starting locations for backward particle paths are user-defined.
147 MODPA TI-I uses a semi-analytical linear interpolation scheme to compute a spatially continuous
148 particle path (Pollock 1994).
7 Hom
149
150
151
152
153
154
155
156
157
158
159
160
161
a . I
~ I .. I "C C: :::, I 0 .rJ I "C I .. ., I .c
. ... I C: B E '" I
91 C: I 0 u
' ' ' I
II I I
-~
b) ----x-direclion
T
E
i
l
1~ I
-~ I
1 I
I I '
387m
~· .. "C .
C: :::, . 0 · .
.c
."C ' ii .c
C·:
B, ,n . C: 8
\ I
_ _/ wall (888 Figura 4) conatant flux boundaiy
-
Figure 2: Model grid in (a) cross-sectional view aty = 0 (Vertical exaggeration= 4.2x) and (b) in
plan view. Due to the symmetiy of the flow field, the model domain simulates only half of a well
and half of the capture zone. The well and gravel pack are very finely discretized. A close-up view
of the model around the well screen is shown in Figure 4.
Our modeling domain is a finite difference grid with 141,750 cells of which 137,937 are active. The
modeling domain is 58 m high, 387.23 m long and 59.695 m wide and consists of 45 rows, 90
columns, and 35 layers. The modeling domain takes advantage of the symmetry in the well flow
field, which is symmetric across the x-axis (mean flow direction) centered on the domestic well (y =
0, see below). The model is therefore designed to model only one-half of the well capture zone
(Figure 2). The second half of the well-capture zone mirrors the first half Grid spacing is non-
8 Hom
162 uniform in both the vertical and horizontal direction. Vertical grid spacing varies from I mat the
163 elevation of the well screen to 4 m elsewhere (Figures 2, 3). Horizontal grid-spacing varies from
164 0.0 I m near the well and in the gravel pack to nearly 20 m near the model boundaries. The
165 horizontal increase in cell-size between adjacent rows or columns of the finite-difference grid is set
166 to not exceed 50 % of its width.
167
168 The hydraulic gradient along the x-axis is produced by defining a constant head boundary of 58.00
169 m to the exterior block of cells al the upgradienl vertical side of the model and a constant head of
170 57.61 mat the downgradient vertical side of the model (Figure 2). This is equivalent to a hydraulic
I 71 gradient of 0.0018, which is typical for the study area. The other two vertical planes of the model are
172 assigned no-flow boundary condition: the vertical plane adjacent to the well half is a symmetry
173 plane. The vertical plane opposite of the half well is at sufficient distance to the well that the local
174 effect of pumping on the groundwater flow field can be neglected and flow is parallel to regional
175 groundwater flow. The average (steady-state) recharge rate is set to 0.669 m/year, a value typical for
176 semi-arid, irrigated agricultural regions such as the Modesto Area, San Joaquin Valley, California.
177 The bottom of the model domain is considered permeable and open to the regional aquifer system
178 below. It is assigned a uniform constant (downward) flux boundary condition, with total outflow
179 across the bottom boundary set equal to the difference between the total recharge inflow at the top
180 and the well outflow rate. In this way we implicitly enhance our model to greater aquifer depths. In
181 the Modesto Area large irrigation wells up lo a depth of almost 370 m below land surface pump
182 large amounts of water and produce a vertical flow component, even through a confining clay unit
183 above the irrigation wells (Burow el al. 2004).
184
9 Hom
l 85 The well construction was chosen to be representative of domestic well construction in the San
l 86 Joaquin Valley, California (e.g., Burow et al., 2004). The model well has a total depth of 56 m below
187 the water table. A seal to 18 m below the water table overlies a 30 m long gravel pack around a
l 88 blank well casing. The casing has a diameter of 0.2 m. The perforated well screen is located at 48 m
189 to 55 m below the water table, followed by a conceptual well sump from 55 m -56 m. Casing and
190 screen are surrounded by a 0.09 m thick gravel pack. The total borehole diameter is 0.38 m. Due to
191 the relatively low pumping rate, the well-loss and skin effect are assumed to be negligible. Inflow
192 along the screen is computed by the model and non-uniformly distributed.
193
194 The grouted well seal above the gravel pack and the well casing are modeled as "no-flow" cells
195 (black cells in Figure 3). The pump is simulated by 74 "well" cells inside the casing. They are
196 located significantly above the top of the screen, opposite of the well seal bottom, which creates an
l 97 upward flow inside the screen and casing. The MODFLOW "well" package is used to simulate the
198 pump cells (light-grey cells in Figure 3). The total pumping rate of the domestic well is 3.5 m3/d,
I 99 half of which is uniformly distributed across the individual "well" cells at the top of the casing.
200 Flow inside the model well casing was modeled by approximating the flow with eq. (8) using very
201 high hydraulic conductivity. The gravel pack (grey cells in Figure 3) is modeled by choosing a
202 separate hydraulic conductivity that is higher than that of the surrounding aquifer and ranges
203 between 50 and 1000 [mid] (Table 1). Modeling the pump inside the well allows the model to
204 properly distribute the flow across the well screen, with screen inflow highest near the top of the
205 screen and lowest at the bottom of the screen.
Hom
E
18
•,
IDIDIIIIDlfflfflllllllfflllll
1111111 · H · IIHIIIIIIIIIIIII
WIii] ~ .0 calls
••••• • ■ •••
• • • ■ •••• . . . • • • • • • ii •• • • • •
gravel pack
.
I ■ ■ I I
■ I I I I ■ ■
I I I I I I I
I I I I I I
I I I I I
• •••••••••••••••••••• • •••• ■ •••••• ■ • • • • • ■ ■
I I ■ I I I ■ I I I I I I I ■ ■ I I I
I I ■ ■ ■ ■ ■ I I ■ I I I I ■ I I I
I I I I ■ I
■
I I I ■ I I I ..........
■ I ■ ■ I
..... 1111-rth-~g pcilnt of parlk:lall
.........
■ ■ ■ ■ I I I ■ I ■
■ I I I I ■ I I I
I I I I I I I I I I
I I I ■ I I I I I I
■ I I I I I I I I I . . . . . . . I I I I I ■ I ■ I ■
I I I I I I I I I I I ■ ■ I I I I ■
I ■ ■ ■ ■ I ■ ■ I I I ■ I I I ■ I ■ I I I I I
I I I I I Ill I I ■ ■ I I I I I I I I I I ■ I I I
I ■ ■ I I ■ ■ I I I I ■ I ■ I I I I I I
■
■ I ■ ■ I I
I I ■ I I I I ■ I I I I I I I I I I ■ ■ I
■
I I ■
I I I I ■ I I I I I I I ■ I ■ I I I ■ ■ ■ ■ I I . . .................... .
■ I I I ■ I ■ ■ I I I I I I ■ I I I II
I I I I I ■ I ■ I I I ■ I I I I I I
■ ■ I I I I I I I I I ■ ■
■ I ■ I I I I I I I I
I I ■ I
206 . 0.38 m 1-----0.38 m -------l
207 Figure 3: Model well configuration and grid discretization around the well. Left: Cross-section at
208 the model boundary (y = 0). Right: Plan view at the land surface (right top), at the top layer of the
209 casing containing the well cells (right center), and at the screen elevation (right bottom). Black cells:
210 casing and well seal (impermeable). Grey cells: gravel pack. Dark grey cells: constant flux boundary
211 cells at the model bottom. Grey dots in the lower left panel indicate the starting location for
212 backward particle tracking.
213
214
11 Hom
215 The hydraulic conductivity, Kh, is assumed to be isotropic in the horizontal plane, while the vertical
216 aquifer hydraulic conductivity, K.., is lower, as typically observed in alluvial aquifers (e.g., Phillips
217 et al., 2007). Two representative anisotropy ratios, Kh/K, = 5 and 2, were chosen to bracket a
218 representative range typically found in alluvial aquifers (ibid.). The gravel pack itself is assumed to
219 have a completely isotropic hydraulic conductivity, Kg, that is larger than Kh, For illustration and
220 application purposes, we modeled well capture zones for a wide range ofrepresentative values for
221 the horizontal hydraulic conductivity, Kh, and the gravel pack hydraulic conductivity, Kg, and for two
222 anisotropy ratios (Table 1 ).
Kh K, K
I 0.2 50,125,250,500,750, 1000
I 0.5 50,125,250,500,750 1000
3 0.6 50 125 250 500 750 1000
3· 1.5 50 125 250 500 750 1000
5 I 50 125 250 500 750 1000
5 2.5 so, 125,250,500,750, 1000
10 2 so, 125,250,500,750, 1000
10 5 50 125 250 500 750 1000
30 6 50 125 250 500 750 1000
30 15 50 125 250 500 750 1000
100 20 125 250 500 750 1000
100 50 125,250,500,750, 1000
300 60 500,750 1000
300 150 500 750, 1000
223
224 Table 1: Model configurations with various
225 combinations of the horizontal hydraulic
226 conductivity, Kh, the vertical hydraulic
227 conductivity, K,, and the gravel pack hydraulic
228 conductivity, Kg-All values are in units of
229 [mid].
230
231 Results
232 Head contour configurations in the aquifer around the domestic well are highly dependent on the
233 aquifer and gravel pack hydraulic conductivities. Cross-sectional head contour lines along the
12 Hom
234 regional flowpath are vertical under strictly_regional flow with no recharge and no pumping. As
235 expected from the analytical model above, the modeled contours deviate from the vertical due to the
236 vertical flow component imposed by the recharge at the top of the model area and the regional
237 pumping below the modeled zone. Contour lines increasingly deviate from the vertical alignment
238 with smaller and smaller ratios of Kh IR (Figure 4). In addition, in aquifers with relatively low
239 hydraulic conductivity, the domestic well creates a distinct zone oflocal influence in the aquifer
240 around the well screen, whereas the influence is minimal in the highly permeable aquifer. The
241 anisotropy of the aquifer hydraulic conductivity creates significant flow zonation: much of the
242 impact of domestic well pumping on the pressure field is seen at the elevation of the well screens,
243 especially for those cases with the higher aquifer anisotropy. Another distinct horizontal zone is
244 created by the top of the gravel pack. The higher the gravel pack hydraulic conductivity (relative to
245 Kh), and the higher the aquifer anisotropy ratio, Kh!K,, the more pronounced is the effect that the
246 transition between the top of the gravel pack and the annular seal has on the head contour lines (e.g.,
247 Figure 4). Inflow to the well varies non-uniformly along the screen. It is highest near the top of the
248 screen, which is nearest to the pump intake inside the well-casing. The difference between the screen
249 inflow at the top (layer 26) and the screen inflow near the bottom (usually in layer 31 just above the
250 bottom of the layer) varies from approximately 45% for highly permeable aquifers to more than
251 100% for veiy low permeable aquifers with veiy high gravel pack K8. This is consistent with
252 analytical models (Nahrgang, 1954; Garg and Lal, 1971) and with field observations on large
253 production wells (VonHofe and Helweg, 1998).
13 Hom
-
254
255 Figure 4: Head contour lines around the well for a conductivity of 10 mid, an anisotropy ratio of 2,
256 and hydraulic gravel pack conductivity of 750 mid. The heads depend on the conductivity of the
257 aquifer, the anisotropy and the relative difference in the conductivities between gravel pack and
258 aquifer. Horizontal dimension is 7 m, vertical dimension is 58 m. Due to the horizontal exaggeration
259 (12.4x) the inclination of the head contours in the regional flow field (near top of the cross-section
260 appears nearly horizontal although it is actually nearly vertical.
261
262 Corresponding to the head field, pathlines in low hydraulic conductivity aquifers are significantly
263 steeper and the capture zone is much closer to the well-head than in an aquifer with high hydraulic
264 conductivity (Figure 5). For Kh 2': 10 ml day, the modeled pathlines are in fact sufficiently flat that
265 the source area is outside the model area In those simulations, we computed the pathlines outside
266 the numerical modeling area by analytically calculating the extension of the pathlines to the water
267 table using equation (7). Also, for model scenarios with hydraulic conductivities of 1, 3, and 5 mid,
268 the pathlines in the top aquifer layer were computed from eq. (7), because MODP A lll computations
269 in the top layer were subject to numerical error.
14 Hom
270
271 The source area of the domestic well has a distinct shape composed of two features: the main
272 capture zone, a relatively large and wide oval area, corresponding to pathlines that enter the annulus
273 of the well below the top of the well-screen for horizontal delivery into the well. At the
274 downgradient (well-facing) side of this main capture zone, we observe a narrow elongated capture
275 subzone that represents those pathlines that enter the gravel-pack of the well at some distance above
276 the well screen. These pathlines capture domestic water through the high permeability field of the
277 gravel pack above the well screen (Figure lb, Figure 5). The greater the hydraulic conductivity
278 difference between gravel pack and aquifer, the higher is the relative downward flow in the upper
279 gravel pack, and the more MODPA TH virtual water particles enter the well flowing through the
280 upper gravel pack. Moreover, the steeper the particle path gradient, the higher is the highest point of
281 entry into the gravel pack of pathlines that ultimately will be captured by the well. Thus, the gravel
282 pack, where it extends to elevations much higher than the well-screen, significantly extends the
283 length of the source area towards the well, albeit within a very narrow transverse range (Figure 5).
-~---==
,.
" -' .,
~
~ ~
284
285 Figure 5: Pathlines in cross section (left) and plan view (half the well, right) with the elongated and
286 main capture zone parts for an aquifer conductivity of 10 mid, an anisotropy ratio of 2, and gravel
287 pack conductivity of 750 mid. Corresponding heads are shown in Figure 4.
15 Hom
'.'
288
289 For further analysis of the capture zone location and size, we separately refer to the width and length
290 of the narrow, "elongated" part of the capture zone nearer to the well and of the "main" part of the
291 capture zone (Figure 6) from where the majority of the water originates. The simulations show that
292 the length of the elongated part increases faster than the length of the main part as horizontal aquifer
293 conductivity increases, but the gravel pack conductivity has a significant influence only on the
294 length of the elongated part (Figure 6c, d). The same is true for the width of the two capture zone
295 parts: The gravel pack conductivity has a significant influence only on the width of the elongated
296 part but little, yet discemable influence on the main part. The width of the elongated part increases
297 several-fold with gravel pack hydraulic conductivity, K8, especially in less productive (low K)
298 aquifers. By the same token, the widths of the main and elongated parts (Figure 6a, b) decrease
299 with higher aquifer conductivities (more narrow, but longer source area). For low gravel pack
300 conductivities, the width of the elongated part of the capture zone remains nearly constant,
301 regardless of aquifer conductivity
16 Hom
a)
50
E 40 fil
~ 30 "i
i 20 a.
C -~ 10
■ 0
3 5 10 30 100 300
c) 1200
Kh [mid]
-1000 .E. II
.c 800 0)
C .9l 600
i::: ro 400 a. I!!
C -~ 200
0 ■ • • Ii Ii
-b)
12
E 10
,5
-0 8 "i
i 6
a.
-0 4 .9l ro Cl 2 C .Q
0) 0
■ • 6
0
B
e
3 5 10 30
o Kg: 50
□ Kg: 125
◊ Kg: 250
6 Kg: 500
• Kg: 750
■ Kg: 1000
100 300
K• [mid]
d) 100 ~---~-~---~
.E. 600
-g, 500
C
.9l 400 i::: ~ 300
]l 200
"' i? 100 .Q " ow:!...--m...t1.....tl------~~
1 3 5 10 30 100 300 3 5 10 30 100 300
3m ~~ ~~
303 Figure 6: Widths (top panels) and lengths (bottom panels) of the elongated part (right panels) and
304 the main part (left panels) of the capture zone for an anisotropy ratio ofKv: Ki,= 1 : 2. Behavior of
305 the models with an anisotropy ratio Kv : Ki, = 1 : 5 is similar.
306
307 The analytical model (eqs. 2,3) of the source area location provides good approximations of the
308 source area only in highly permeable aquifers. For aquifers with intermediate and low conductivity,
309 the gravel pack has significant influence on the distance of the downgradient edge of the capture
310 zone from the well (Figure 7a), where the source area can be as much as 90% closer to the well than
311 estimated from eq. 2. The analytical approximation of the distance from the well to the upgradient
312 edge of the source area (eq. 3) is relatively close to the numerical simulations if aquifer hydraulic
313 conductivities are above 5 mid. In those cases, the relative difference between analytical and
314 numerical model is on the order of 10% or less (Figure 7b), regardless of anisotropy ratio and gravel
315 pack hydraulic conductivity.
17 Hom
a)
1,0
b)
0,9 GI el " !'l o Kg: 50 C: ! ]j 0,8 " .'9 0 = o Kg: 125 i 0,8 .!ll ◊ Kg: 250 'O 0,7
E § t. Kg: 500
.; 0 6 \'I 2 E 0,6 • Kg: 750 'I< Bl Kg: 1000 ·-' ~ 0,5 E
~ □ .. ~ 0,4 e1 0,4 ◊ Q) • 5l 0 d j 0,3
:i □ 6l
] 0,2 0 ◊ ";' 0,2
□ 0 al Ill
ai 0 G ~ 0,1 I!! 'O &l 1l El ~ 0,0 ll E 0,0
3 5 10 30 100 300 3 5 10 30 100 300
316 Kh [mid] Kh [m/d]
317 Figure 7: Comparison of the distances of the source areas to the well provided by the numerical and
318 by the analytical model exemplarily for an anisotropy ratio of Kv: Ki,= 1 : 2. (a) Normalized
319 differences between the modeled and analytically calculated distances of the downgradient edges of
320 the source areas to the well. (b) Normalized differences of the distances of the upgradient edges of
321 the source areas to the well.
322
323 The simulation results show that water moves downward inside the gravel pack above the well-
324 screen from considerable distances: For Kh less than 10 mid and high gravel pack hydraulic
325 conductivities, water travels downward from as far as the top of the gravel pack, 30 m above the
326 well-screen (Figure 8). Again, the more permeable the gravel pack in the annulus, the larger the
327 above-screen capture of source water. The fraction of well pumpage that originates from capture in
328 the gravel pack above the well-screen increases as the aquifer hydraulic conductivity decreases
329 (Figure 9). In intermediate and low permeable aquifers, domestic wells with highly permeable
330 gravel packs receive from 20% to 50% of the total well flow from the extended gravel pack above
331 the screened aquifer horizon. This model result is qualitatively consistent with the field data of
332 Houben (2006), who found iron oxide incrustations in the gravel pack significantly above the top of
333 the well screen, where the incrustations were due to a significant amount of water flowing through
334 the upper part of the gravel pack. At high aquifer conductivities (Kh > 10 mid), less than 12 % of the
18 Hom
335 total domestic well flow originate from the gravel pack above the well screen. Aquifer anisotropy
336 has little influence on the height of the capture zone within the gravel pack.
337
338 The height of the gravel pack participating in flow to the well and the percent fraction of the
339 pumpage originating from the gravel pack above the screen can be expressed quantitatively: Table
340 2 provides the regression coefficients obtained by fitting data in Figures Sa and Sb to nonlinear
341 exponential regression equations of the form:
342 y = a*exp(-log(Ki,)/b) (9)
343 using the Leven berg-Marquardt algorithm for optimization. For application to a specific site, linear
344 interpolation of the values for a and bin Table 2 may be used to compute the height of capture in
345 the gravel pack and the proportion of flow originating from the gravel pack above the well screen for
346 values of the anisotropy ratio and of Kg other than those given in the Table. This modified analytical
347 tool provides a much more realistic source area than the much simpler graphical method employed
348 in many states as part of their source water assessment programs (e.g., California OHS 1999).
19 Hom
a) b)
E 30 0 !;I ~ 60 0 Kg:50 .,,_ 0 Kg: 125 0 25 50 m <> Kg: 250 gi_
.; 0 • A Kg: 500
~ 20 C • Kg: 750
◊ 40 A c Kg: 1000 en ~ a, • ~ C £ 15 ~ 30 0 • . 5 A m E 0 D '.§ A • -i 10 20 0
◊ A " .: 0 • E 0 g 0 ◊ A ::, D 0 E 5 A 10 0 ◊ 0 "ij ~ Q 0 0 0
E Iii 0 0 0
1 3 5 10 30 100 300 3 5 10 30 100 300
349 Kh [m/d] Kh [m/d]
350 Figure 8: (a) Maximum virtual water particle heights in the gravel pack above the well screen
351 serving to capture water (b) Percentage of inflow into the well screen flowing through the gravel
352 pack from above the screen. Both for an anisotropy ratio of 2.
353
354
355
356
Anisotrop Kg Parameter for Adjusted Parameter for inflow Adjusted
y maximum heights: r2 from above: r2
a b A b
2 50 33.31 1.05 1.00 12.69 0.60 0.99
2 125 77.01 0.88 0.99 21.44 0.69 0.99
2 250 87.95 1.02 1.00 30.59 0.76 0.99
2 500 163.35 0.93 1.00 40.57 0.86 0.99
2 750 172.04 1.00 1.00 47.16 0.92 0.99
2 1000 226.70 0.96 0.94 52.10 0.96 0.98
5 50 79.01 0.78 0.95 14.94 0.61 1.00
5 125 98.28 0.93 0.97 25.14 0.70 0.99
5 250 216.33 0.80 0.93 35.57 0.77 0.99
5 500 276.67 0.87 0.99 47.59 0.85 0.99
5 750 396.12 0.85 0.99 55.54 0.91 0.99
5 1000 401.91 0.88 1.00 61.39 0.96 0.98 _,. . . 357 Table 2. Coefficients and adjusted coefficients of detemunahon (r) for the equanons descnbmg the
358 maximum heights of the capture zone in the gravel pack, and the inflow of water entering the well
359 from the gravel pack above the screen.
360
361
362
20 Hom
363 Discussion
364 For application to specific sites, Figure 7 provides a tool to estimate the additional source area due to
365 the gravel pack, when compared to the simple approximation (eq. 2). These results can also be
366 applied for conditions with smaller or larger recharge rates, R ', than the rate R = 0.669 m/a used in
367 our computations. For R' not equal to R, results shown in Figures 7-10 and expressed in the above
368 equation are looked up for a scaled hydraulic conductivity K' rather than for the actual hydraulic
369 conductivity K, where K' = K · R 1/R. This scaling procedure is approximate because it does not
370 simultaneously scale other parameters controlling the observed results, e.g., screen length and
371 pumping rate. However, for applications in unconsolidated sedimentary aquifers, this scaling
372 approach works well as the drawdown created by domestic wells is relatively small. For depths to
373 the top of the screen different from that used here, the simple geometric conceptual model outlined
374 in Figure 1 and expressed in eq. 2 provides a framework for adjusting the distance of the source area
375 from the well head. Equation 9 (with Table 2) can be used to estimate the fraction of flow
376 originating from the elongated part of the source area.
377
378 The numerical modeling shows the significant influence of the gravel pack on the source area of a
379 domestic well, particularly for lower permeable aquifers (horizontal hydraulic conductivities of less
380 than 10 mid). In highly permeable aquifers (relative to the recharge rate of0.669 m/year used in this
381 study), the analytical model (eqs. 2, 3) provides a relatively good approximation of the upgradient
382 and downgradient edge of the source area. Lower hydraulic conductivities lead to significantly
383 longer capture zones than predicted by the analytical model (eqs. 2-3). In our configuration of screen
384 length and gravel pack length, which represents an average domestic well construction for Central
385 California, the elongation due to the presence of a gravel pack constitutes up to 70 % of the total
21 Hom
386 length of the capture zone. The elongation is relatively narrow but higher gravel pack conductivities
387 lead to significant increases in that width. The width of the main capture zone, in turn, slightly
388 decreases at higher gravel pack conductivities. The greater the difference between hydraulic
389 conductivity of the aquifer and that of the gravel pack, the greater is the elongated part relative to the
390 total length of the capture zone.
391
392 For many contaminants, chemical or microbial, aquifer attenuation is a dynamic, time-dependent
393 process. Travel times for potential contaminants decrease approximately linearly with increased
394 gravel pack length above the well screen. This is due to the strong influence of recharge on vertical
395 downward displacement of water (and contaminants) and the relatively small influence that the
396 domestic well pumping exerts on the overall groundwater flow field. A linear decrease in travel time
397 from the time of recharge until arrival at the gravel pack is associated with exponentially increased
398 contaminant concentrations. The gravel pack itself typically provides much less attenuation capacity
399 than the aquifer material. Hence, a short seal and vertically extended gravel pack constitute a
400 potential short-circuit for contaminants.
401
402 We also note that the fraction of flow captured by the gravel pack above the well screen may be
403 relatively small in a productive (high K) aquifer. But for some contaminants the resulting dilution
404 with (good) groundwater collected by the well at the depth of the screen may not be sufficient. This
405 includes contaminants that reach the water table at concentrations that are several orders of
406 magnitude above regulatory drinking water limits including solvents, pesticides, other organic
407 chemicals, and pathogens. A possibly common source of such contamination are septic tank leach
22 Hom
408 fields, which -in rural and semi-rural housing developments -are often located in the vicinity of
409 domestic wells.
410
411 Conclusions
412 Our work provides a tool to quickly estimate the size and location of the source area of domestic
413 wells in regions with significant recharge (for example, due to irrigation). The influence of the
414 gravel (or sand) pack in the well annulus above the well screen is explicitly accounted for. Results
415 allow for estimation of source area and gravel pack impact for a wide range of scenarios.
416 Importantly, we show that the gravel pack above the well screen poses a significantly increased risk
417 for domestic well contamination. A gravel pack that extends significantly above the well screen ( due
418 to short seal length), may significantly enhance the length of the source area, thus exposing the well
419 to a larger cross-section of potential contaminant sources. The extended gravel pack also decreases
420 travel time and distance for contaminants from the source area to the well allowing for contaminants
421 to partially circumvent natural aquifer attenuation. This is especially true in aquifers with low to
422 intermediate hydraulic conductivity (K :<::: 10 mid). We therefore strongly recommend that the gravel
423 (or sand) pack not be extended more than a few meters above the well screen of a domestic well.
424
425 Acknowledgments. We gratefully acknowledge the careful review and constructive comments of
426 Karen Burow, USGS, and two anonymous reviewers. Funding for this research was provided
427 through a fellowship of the German Academic Exchange Service (DAAD) to Judith Hom.
428
429
430
23 Hom
431 References
432 Belitz, K., and S. P. Phillips (I 995), Alternative to agriculture drains in California's San Joanquin
433 valley: results of a regional-scale hydrogeologic approach, Water Resour. Res., 31(8), 1845-1862.
434
435 Burow, K. R., J. L. Shelton, J. A. Hevesi, and G. S. Weissmann. 2004. Hydrogeologic
436 characterization of the Modesto area, San Joaquin Valley, California. USGS Scientific Investigations
437 Report 2004-5232.
438
439 California OHS. 1999. Drinking Water Source Assessment and Protection (DWSAP) Program.
440 Division of Drinking Water and Environmental Management, California Department of Health
441 Services. Sacramento, CA. http://www.dhs.ca.gov/ps/ddwem/dwsap/DWSAPindex.htm
442
443 Garg, S. P. and J. Lal, 1971. Rational design of well screens. J. Irrig. and Drain. Div., ASCE,
444 97(1):24-35.
445
446 Harter, T. 2008. Delineation of wellhead protection areas. In: T. Harter and L. Rollins (eds.), 2008.
447 Watersheds, Groundwater, and Drinking Water: A Practical Guide, University of California, UC
448 ANR Communications Services Publication 3497, Davis, CA 95616, 274p.
449
450 Houben, G. H. 2006. The influence of well hydraulics on the spatial distribution of well
451 incrustations. Ground Water 44, no. 4: 668-675.
452
453 McCray, J. E., S. L. Kirkland, R. L. Siegrist, and G. D. Thyne. 2005. Model parameters for
454 simulating fate and transport of on-site wastewater nutrients. Ground Water 43, no. 4: 628-629.
455
456 McDonald, M. G. and A. W. Harbaugh. 1988. Technics of water-resources investigation of the
457 United States Geological Survey. USGS Open-File Report 83-875.
458
459 Nahrgang, G., 1954. Zur Theorie des vollkommenen und unvollkommenen Brunnens. 43 p.
460
461 Phillips, S. P., C. T. Green, K. R. Burow, J. L. Shelton, and D. L. Rewis, 2007. Simulation of
462 ground-water flow in part of the Northeastern San Joaquin Valley, California, U.S. Geological
463 Survey, Scientific Investigations Report, SIR 2007-5009.
464
465 Pollock, D. W. 1994. User's guide for MODPATH/MODPATH-PLOT, Version 3: A particle
466 tracking post-processing package for MODFLOW, the U. S. Geological Survey finite-difference
467 ground-water flow model. USGS Open-File Report 94-464.
468
469 Todd, D. K. 1980. Ground water hydrology, 2nd ed. New York: John Wiley and Sons.
470
471 U.S. EPA. 1997. Response to Congress on use of decentralized wastewater treatment systems.
472 Washington, D.C.: Office of Water, U.S. EPA.
473
474 VonHofe, F. and 0. J. Helweg, 1998. Modeling well hydrodynamics. ASCE J. Hydr. Eng.
475 124(12):1198-1202.
24 Hom
APPENDIXB
Statistical Analysis Relating Well Yield
to Construction Practices and Siting
of Wells in the Piedmont and
Blue Ridge Provinces of North Carolina
United States
Geological
Survey
Water-Supply
Paper 2341-A
Prepared in cooperation
with the North Carolina
Department of
Natural Resource~ and
Community Development
SELECTED SERIES OF U.S. GEOLOGICAL SURVEY PUBLICATIONS
Period lea ls
Earthquakes & Volcanoes (issued bimonthly).
Preliminary DetermlnaUon or Eplcenters(usued monthly).
Technical Books and Reports
Proresslonal Papen arc mainly comprehensive scientific reports of
wide and lastins interest and importance to professional scientislS and en-
gineers. Included are reports on the resu!IS of resource •tudies and of
topographic, hydrologic, and geologic investigations. They also include
collections of relalod papen addrusing diffem,t aspe<:ISof a single scien-
tific topic.
BuDetlns contain •ignificant dBla and inte,prelations th8l are of last-
ing scientific interest but arc generally more limited in scope or
geographic coverage than Professional Pupers. They include the results
of resource studies and of geologic and topographic investigations: as well
as collections of short papers rclau:d to a specific topic.
Water-Supply Papers arc comprehensive reports thBI present sig-
nificmt interpretive results of hydrologic investigations of wide interest
to professional geologists, hydrologists, and engineers. The serics coven
investigations in all phases of hydrology, including hydrogeology,
availability of water, quality of water, and use of WIiler.
Cliculan present adminiJtralive information or important scientific
infonnarion of wide popular intttest in a format designed for dislribution
at no cost to the public. Information is usually of shon--tcnn interest.
Water-Resources Investigations Reports are papen of an inte,pre-
tive nature made available to the public outside the formal USGS publi-
cations .. rie,. Copies are n:pro<iu=I on reque.t unlil:e fonnal USGS
publications, and they ue also av ail able for public inspection at
depositories indicalod in USGS catalog,.
Open-FIie Reports include unpublished mmruscript reports, maps,
and other material tha1 arc made available for public consuliation at
depositories. They arc a nonpermancnl form of publication that may be.
cited in other publications as sources of information.
Maps
Geologlc Quadrangle Maps are multicolor geologic rnaps on
topographic bases in 7 Ill.-or I 5-minutequadrangle fonnBIS (scales main-
ly I :24,000 or I :62,500) showing bedrock, surficial, or engineering geol-
ogy. Mops generally include brief texlS; some mops include •tructure
and columnar sections only.
Geophysical Investigations Maps are on topographic or planimelric
bases at vmious scales; they show results of rurveys using geophysical
techniques. such as gravity, magnetic, seismic, or radioactivity, which
reflect subsurface structures that are of coonomic or ge.ologic significance.
Ma."ly maps include correlatiom with the geology.
Miscellaneous lnVl!Stlgallons Serles Maps arc on planimelric or
topographic bases of regular and irregular arcss at various scales; they
present a wide variety of fonnat and subject matter. The series also in-
cludes 71/2-minutequadrangle photogeologic maps on planimetric bases
which show geology u interpreted &om aerial photogrophs. Series also
includes maps of Man: and lhe Moon.
Coal Investlj:aUons Maps are geologic maps on topographic or
planimetric: bases at various scales showing bedrock or surficia1 geol-
ogy, stratigraphy. and structural relations in certain coal-resource areas.
OU and Gas Investigations Charts show sll'llligraphic information
for certain oil and gas fields and other areas having petroleum potential.
Mlscellnneous Field Studies Maps are multicolor or black-and-
white mops on topographic or planimenic bases on quadrangle or ir-
regular areas a1 various scales. Pre-1971 maps show bedrock geology
in relation to specific mining or mineral-deposit problems; post-1971
maps are primarily blade-and-white maps on various ,ubjectJl such as
environmental studies or wilderness mineral investigations.
Hydrologlc Investigations Atlases are multicolored or black-and-
whitemops on topographicorplanimelric bases presenting a wide range
of geohydrologic dalaofboth regular and irregular areas; principal scale
is 1:24,000 and regional studies are at 1:250,000 scale or smaller.
catalogs
Permanent catalogs. u well as some othen, giving cornpre.hc:n-
sive listing, of U.S. Geological Survey publications ore available under
theconditions indicalod below &om the U.S. Geological Survey, Boob
and Open-File Reports Section, Federal Center, Box 25425, Denver,
CO 80225. (See larat Price and Availability List)
"Publications of the Geologk:al Survey, 11179-1961" may be pur-
chased by mail and over the counu:r in paperback book form and as a
set of rniaofiche.
"Publications of the Geologlcal Survey, 1962-1970" may be pur-
chased by mail and ovu the counter in pape,back book form and as a
Bel of microfiche.
• PubUcaUons or the U.S. Geologic al Survey, 1971-1981" may be
purchased by mail and over the counu:r in paperback book form (two
volumes, publications listing and index) and as a set of microfiche.
Supplements for 1982, 1983, 1984, 1985, 1986, and for subsequent
years since the last pemwtc:nt catalog may be purchased by mail and
over the counter in paperback book form.
State catalogs, "List of U.S. Geological Survey Geologic and
W111er-Supply Reporu and Maps For (Stale)," may be pun:hased by mail
and over the counter in paperback booklet form only.
"Price and A nllabWty List of U.S. Geologlcal Survey Publlca-
tlons, • issued annually, is available free of charge in paperbeck book-
let form only.
Seltttod roplesor a monthly catalog "New Publications of the U.S.
Geological Survey" available free of charge by mail or may be obtained
over the counter in paperback book.let fonn only. Those wishing a free
,uhsaiption to the monthly catalog "New Publications of the U.S.
Geological Survey" should write to the U.S. Geological Survey, 582
National Center, Reston, VA 220!/2.
Note.-Price.5 of Government publications listed in older catalogs,
announcements, and publications may be incorrect Therefore, the
pri~ charged may differ from the prices in calalogs, announcemenlS,
and poblicatlons.
Chapter A
Statistical Analysis Relating Well Yield
to Construction Practices and Siting
of Wells in the Piedmont and
Blue Ridge Provinces of North Carolina
By CHARLES C. DANIEL Ill
Prepared in cooperation with the North Carolina
Department of Natural Resources and Community
Development
A statistical analysis of data from more than 6,200
water ~ells was made to identify geologic,
topographic, and construction factors associated
with high-yield wells
U.S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 2341
GROUND-WATER RESOURCES OF THE PIEDMONT-BLUE RIDGE PROVINCES OF
NORTH CAROLINA
DEPARTMENT OF THE INTERIOR
MANUEL LUJAN, Jr., Secretary
U.S. GEOLOGICAL SURVEY
Dallas L. Peck, Director
Any use of trade, product, or firm names in this publication is for descriptive
purposes only and does not imply endorsement by the U.S. Government
UNITED STATES GOVERNMENT PRINTING OFFICE : 1989
For sale by the Books and Open-File Reports Section,
U.S. Geological Survey, Federal Center,
Box 25425, Denver, CO 80225
Library of Congress Cataloging in Publication Data
Daniel, Charles C., Ill
Statistical analysis relating well yield to construction practices and siting of
wells in the Piedmont and Blue Ridge provinces of North Carolina.
(U.S. Geological Survey water•supply paper; 2341-A)
Supt. of Docs. no. : I 19.13:2341A)
"'Prepared in cooperation with the North Carolina Department of Natural
Resources and Community Development."'
Bibliography: p
1. Water, Underground-North Carolina-Statistical methods.
2. Wells-North Carolina-Statistical methods. 3. Water•supply-North
Carolina-Statistical methods. I. North Carolina. Dept. of Natural
Resources and Community Development. II. Title. Ill. Series.
TD224.N8D36 1989 628.1"14'09756 88-600118
CONTENTS
Abstract Al
Introduction A 1
Purpose and Scope A3
Previous Investigations A3
Description of the Study Area A3
Physiography A3
Geology A6
Hydrogeologic Units A6
Geologic Belts and Terranes AB
Compilation of the Data Base and Statistical Procedures AB
lnfomtation Categories in the Data Base AB
Statistical Procedures Al3
Relation of Well Yield to Construction Practices and Siting of Wells Al4
Results of the Analysis Al4
Well Yields by Hydrogeologic Unit A20
Well Yields by Geologic Belts and Terrones A22
Summary and Conclusions A23
References A26
FIGURES
I. Map of North Carolina showing area of investigation. counties, and physio-
graphic provinces Al
2. Index map of North Carolina showing study areas of reconnaissance ground-
water investigations that were the sources of well data for this study A4
3. Map and geologic section showing the physical setting of the ground-
water system in North Carolina AS
4. Index map of North Carolina showing areas of geologic maps used in compi-
lation of the hydrogeologic unit map of the Piedmont and Blue Ridge
provinces A9
5. Hydrogeologic unit map of Guilford and Alamance Counties and vicinity in
the north-central Piedmont of North Carolina AlO
6. Map showing geologic belts, terranes, and some major structural features
within the Piedmont and Blue Ridge provinces of North Carolina All
7. Graph showing variation of average yield. average depth. and average yield
per foot of well depth with well bore diameter Al 7
8-IO. Contour plots of trend surfaces showing:
8. Relation between well yield, total well depth, and well diameter for wells
that are located in draws and valleys Al8
9. Relation between well yield, total well depth. and well diameter for wells
that are located on slopes and flats Al9
JO. Relation between well yield, total well depth, and well diameter for wells
that are located on hills and ridges A20
11 . Graph showing variation of average yield and yield per foot of well depth
with depth for wells having diameters between 5.5 and 6.5 in. All
Content! Ill
12. Contour plot of trend surface showing relation between yield per foot of well
depth, total well depth, and well diameter A22
13--14. Graphs showing:
I 3. Average yield of wells of average construction in the hydrogeologic units
of the Piedmont and Blue Ridge provinces of North Carolina A24
14. Average yield of wells of average construction in the geologic belts and
tcrrancs of the Piedmont and Blue Ridge provinces of North
Carolina A2S
TABLES
1. Classification and lithologic description of hydrogcologic units in the Piedmont and
Blue Ridge provinces of North Carolina A 7
2. Geologic belts and terrancs of the Blue Ridge, Piedmont, and Coastal Plain provinces
of North Carolina A12
3. Total number of entries for each variable in the water-well data base A13
4. Average and median values of selected well characteristics according to topographic
sening compared to statistics for all wells A14
5. Summary statistics defining depth to water, casing depth, and saturated thickness of
regolith according to topographic group in the Blue Ridge and Piedmont physiographic
provinces A16
6. Relation of selected well characteristics to the use of the well A16
7. Relation of well yields to hydrogeologic unit and topography A23
METRIC CONVERSION FACTORS
For readers who wish to convert measurements from the inch-pound system of units to the metric
system of units. the conversion factors are listed below:
Multiply inch-pound unit By To obtain SI unit
l..tngth
inch (in.) 2.5.4 millimeter (mm)
foot (ft) .3048 meter (m)
mile (mi) 1.609 kilometer (km)
Area
square mile (mi2) 2.590 SQWU'C kilometer (km2)
Volume
gallon (gal) 3.785 liter (L)
.003785 cubic meter (m3)
Flow
gallon per minute (gaVmin) 3.785 liter per minute (Umin)
.003785 cubic meter per minute (m3/min)
Flow per Length
gallon per minute per foot 12.418 liter per minute per meter
[(gaVmin)lft] [(Umin)/m]
.01242 cubic meter per minute per meter
[(m3/min)lm]
ALTITUDE DATUM
Sea level: ln this report "sea level" refers to the National GeodeticVertical Datum of 1929
(NOYD of 1929)-a geodetic datum derived from a general adjustment of the first-order level nets
of both the United States and Canada, formerly called Sea Level Datum of 1929.
IV Contenb
Statistical Analysis Relating Well Yield
to Construction Practices and Siting
of Wells in the Piedmont and
Blue Ridge Provinces of North Carolina
By Charles C. Daniel Ill
Abstract
A statistical analysis was made of data from more
than 6,200 water wells drilled into the fractured.crystalline
rocks of the Piedmoiil:"Blue Ridge, and western ed~ of
'tlie C~astal Plain where crystalline rocks underlie sedi-
ments at shallow depths. The study area encompassed 6.5
counties in western North Carolina, an area of 30,544
square miles, which comprises nearly two-thirds of the
State. Additional water supplies will be needed in western
North Carolina as population and industrial development
continue to increase. Ground water is an attractive alter-
native to surface-water sources for moderate to large
supplies. The statistical analysis was made to identify
geologic, topograptii~d·construction·fa'ctors=thar'are
asrociated'wiikliigh:yield wells.
=tfls-generally believed-tliat the crystalline rocks of
the Piedmont and Blue Ridge provinces yield only small
amounts of water to wells, that water is obtained from
vertical fractures that pinch out at a depth of about 300
feet because of lithostatic pressure, and that the function
of a large diameter well is primarily for storage. These
concepts are reasonable when based upon the fact that
the average well drilled in these rocks is a domestic well,
125 feet deep, 6 inches or less in diameter, and located on
a hill or ridge. However, statistical analysis shows that
wells in draws or valleys have average yields three times
those of wells on hills and ridges. Wells in the most
productive hydrogeologic units have average yields twice
those of wells in the least productive units. Wells in draws
and valleys in the most productive units average five times
more yield than wells on hills and ridges in the least
productive units.
Well diameter can have a significant influence on
yield: for a given depth, yield is directly proportional to
well diameter. Maximum well yields are obtained from
much greater depths than previously believed. For exam-
ple, the average yield of &-inch diameter wells located in
draws and valleys can be expected to reach a maximum of
about 45 gallons per minute at depths of 500 to 525 feet;
for similarly located 12•inch diameter wells, the average
yield can be expected to reach a maximum of about 150
gallons per minute at depths of 700 to 800 feet.
INTRODUCTION
Additional water supplies will be needed in the
Piedmont and Blue Ridge provinces of North Carolina (fig.
I) as population and industrial development continue to
increase. Municipal and industrial water supplies are
derived almost exclusively from surface water sources.
However. the potential for further development of surface
water is limited, and ground water is an attractive alterna-
tive for moderate to large water supplies.
Ground water has many attractive features as a source
of supply. Ground water in the crystalline rocks of the
Piedmont and Blue Ridge provinces has a relatively low
cost of development (Cederntrom, 1972).Generally, ground
water in these areas is of good chemical quality and requires
linle treatment. Because of the large quantity of water in
storage, the ground-water system usually can sustain mod-
erate yields during seasonal dry periods. The use of ground
water generally pennits other land-use activities if they do
not impede the infiltration of recharge or diminish water
quality.
The crystalline rocks that underlie the Piedmont and
Blue Ridge are reputed to furnish only small quantities of
ground water. This impression is the outgrowth of drilling
large numbers of domestic wells that do not represent
efforts to obtain quantities of water beyond the minimum
requirement of 2 to 10 gallons per minute (gal/min). About
70 percent of all wells drilled in the Piedmont and Blue
Ridge are for domestic supply. and most were located and
drilled without regard to geology .topography, and optimal
constrUction. In spite of these shortcomings, a significant
number of wells yield a few tens to a few hundreds of
gallons per minute. Additional high-yield wells likely can
be developed at carefully selected sites throughout the area.
Analysis Relating Well Yield to ConstrucOon and Sit;ng, Piedmont•Blue Ridge Provinces, North Carolina A 1
••• . ,.
36°
...
North Carolina study area:
30,544 square miles
65 counties
, ..
1980 population 4.36 million
Figure 1. Area of investigation showing counties and physiographic provinces.
rs• 11•
0 25 50 100 MIL.ES •--~~--~~----~
0 •• 100 KILOMETERS
Results of studies in several areas of the Piedmont,
both within and outside North Carolina, show that the
ground-water system can suppon large well yields. For
example, Daniel and Sharpless (1983) reported finding
more than 300 wells in an eight-county area of central North
Carolina that produce 50 gaVmin or more. Cressler and
others (1983) found a substantial number of wells in the
Georgia Piedmont that yield more than 100 gal/min and-
some that yield nearly 500 gal/min. They also found 66
mainly industrial and municipal wells that had been in use
for periods of 12 to more than 30 years without experienc-
ing declining yields. Similarly, Cederstrom (1972) found
that yields of I 00 to 300 gal/min are not uncommon for
bedrock wells in the Piedmont and Blue Ridge provinces
from Maine to Virginia.
To evaluate the potential for large ground-water
supplies in the Piedmont and Blue Ridge provinces of North
Carolina, the U.S. Geological Survey-in cooperation with
the North Carolina Department of Natural Resources and
Community Development-conducted a 5-year study of
ground-water resources in the region. This report is part of
that study.
Purpose and Scope
This report describes a statistical analysis of data
from a large number of water wells in the Piedmont and
Blue Ridge provinces of North Carolina. The analysis was
undertaken to identify factors that are associated with
high-yield wells.
The statistical analysis was made by using hydro-
logic, geologic.topographic, and well-construction data that
were obtained from records of more than 6,200 water wells.
The wells are in an area including all of the Piedmont and
Blue Ridge provinces in the State and an adjoining narrow
strip at the western edge of the Coastal Plain province where
a number of wells draw water from Piedmont crystalline
rocks at shallow depth beneath the sedimentary cover. The
study area encompassed 65 counties in North Carolina, an
area of 30,544 square miles (mi2), which comprises nearly
two-thirds of the State (fig. I).
The records of water wells, obtained from published
sources. were used to compile information on well yields
and water levels; use of the water; well-construction vari-
ables such as total depth. diameter, and casing depth: and
the siting of wells in relation to topography and geology. A
total of 14 geologic terranes considered to be hydrologically
significant were identified in the study area. Within these
terranes are 21 major rock types of igneous, metaigneous,
metasedimentary, metavolcanic, and sedimentary origin
that are considered to have quantifiable hydrogeologic
properties. Because of their hydrogeologic properties, these
major rock types are designated herein as hydrogeologic
units.
The data on both geologic terranes and hydrogeologic
units were obtained largely from the work, both published
and unpublished, of other investigators. Field studies were
kept to a minimum.
Previous Investigations
Between 1946 and 1971, a total of 14 reconnaissance
ground-water investigations (fig. 2) were completed that
provided information on ground-water resources in all the
counties in the Piedmont and Blue Ridge provinces of North
Carolina. All but one of these reports (Peace and
Link,1971) were prepared by the U.S. Geological Survey in
cooperation with various Nonh Carolina State agencies.
Included in the 14 reports, which were the main sources of
data for this report. are maps showing well locations in each
county and tables of well records providing details of well
construction, yield, use, topographic setting, water-bearing
formation·. plus miscellaneous notes.
DESCRIPTION OF THE STUDY AREA
Physiography
North Carolina lies in three physiographic provinces
of the southeastern United States (fig. 3): the Blue Ridge.
the Piedmont, and the Coastal Plain (Fenneman, 1938).
The Blue Ridge province in western North Carolina
contains the greatest mountain masses, highest altitudes.
and the most rugged topography in eastern North America.
The province is marked by steep, forest-covered slopes that
are cut by numerous small stream valleys. More than 40
peaks are greater than 6.000 feet (fl) in altitude and another
82 peaks are between 5,000 and 6,000 fl in altitude (Conrad
and others, 1975). The province is bounded on the west in
Tennessee by the Ridge and Valley province. On the east.
the boundary of the Blue Ridge with the Piedmont province
is marked by the escarpment of the Blue Ridge front-a
prominent topographic feature thought in part to be associ-
ated with faulting. The Blue Ridge front rises more than
1,700 ft above the Piedmont surface Ill the North Carolina-
Virginia border and reaches a maximum relief of nearly
2,500 ft in central North Carolina.
The topography of the Piedmont consists of low,
well-rounded hills and long, rolling. northeast-trending
ridges. The tops of many ridges and interstream divides are
relatively !lat. They are thought lo be remnants of the
Piedmont peneplain, an ancient erosional surface of low
relief. More recent erosion and downculling by streams has
dissected the Piedmont peneplain and created a local topo-
graphic relief of 100 to 200 ft between interstream divides
and stream bottoms. The Piedmont surface is 300 to 600 ft
Analysis Relating Well Yield to Construdion and Siting, Piedmont•Blue Ridge Provinces, North Carolina A3
83'
35°
3 Trapp, 1970 10
4 Sumsion and Laney, 1967 11
5 Peace and Link, 1971 12
6 LeGrand and Mundorfl, 1952 13
7 LeGrand, 1954 14
81' so•
Bain, 1966
May and Thomas, 1988
Schipl, 1961
Pusey, 1980
Mundorff, 1946
_ _L
78°
0 ' -
0
Figure 2. Study areas of reconnaissance ground-water investigations that were the sources of well data for this study.
76°
" so 100 MILES
so 100 KILOMETER$
•••
, ..
,..
FEET
8,000
SEA LEVEL
8,000
Figure 3.
••• •••
A BLUE RIDGE
... -,·· • •• -r·
PIEDMONT
Blue Ridge Front
/
Gneiss and
schist
Piedmont
Charlotte
Belt
Plateau
790 1&• 11_• ______ c',•c•-----, ----~----·.--, .. _,.
Triassic
basin
I Fall Line
•~--""---'~'------•"OO MILES
0 ,. 100 KILOMETER'S
COASTAL PLAIN
Cape
Hatteras
Pamlico
Sound
A'
Physical setting of the ground-water system in North Carolina (modified from Heath, 1980).
in altitude along the eastern border and rises gradually to the
west to about 1,500 ft in altitude at the foot of the Blue
Ridge front.
Scattered across the rolling Piedmont surface are
remnants of once higher mountains that because of their
resistance to erosion stand as much as 500 to 1,600 ft above
the local land surface. Some form prominent lines of hills.
Others are isolated hills and mountains. called monadnocks,
that stand alone above the Piedmont surface and. although
more common in the western Piedmont. are found through-
out the province.
The Piedmont is bounded on the east by the Fall Line
where the hard crystalline rocks of the Piedmont give way
to the softer sedimentary rocks of the Coastal Plain prov-
ince. At the Fall Line. the swift-flowing streams of the
Piedmont enter the Coastal Plain over a zone of rapids and
low falls.
The Coastal Plain has little relief in contrast to the
adjoining Piedmont. It is marked by sluggish streams
flowing in broad valleys cut into predominantly sand and
clay units that thicken seaward from a feather edge at the
Fall Line. Along the western edge of the Coastal Plain, the
sediments are underlain at shallow depth by crystalline
Piedmont rocks (fig. 3).
Geology
The geology of the Piedmont and Blue Ridge is
extremely complex. All major classes of rocks-metamor-
phic, igneous, and sedimentary-are represented. although
metamorphic rocks are the most abundant. The metamor-
phic and igneous rocks range in composition from felsic to
ultrarnalic and range in age from Precambrian in the Blue
Ridge to Triassic and Jurassic in the Piedmont. The meta-
morphism of the rocks varies in grade from low rank to high
rank: that is, varying in degree of recrystallization and
destruction of the original texture: many have been folded
and refolded during multiple metamorphic and orogenic
events. The rocks are broken and displaced by numerous
faults and zones of shearing, some of which are many miles
in length. Nearly everywhere are rock fractures without
displacement called joints. The joints commonly cluster in
groups orientated about one or more preferred directions.
Within the crystalline rocks of the Piedmont are down-
faulted basins (grabens) filled with sedimentary rocks of
Triassic age.
Three or more periods of igneous intrusion (Fullagar,
1971) have resulted in the emplacement of plutonic bodies
that range in size from batholiths down to dikes. sills. and
veins. Most intrusions have been metamorphosed.
deformed, and fractured, but some are massive and have
little or no foliation. All rocks have been subjected to uplift,
weathering. and erosion, which resulted in the widening of
fractures and the formation of new openings such as
stress-relief fractures. These breaks in the otherwise solid
rock are the conduits for ground-water flow. All of the
events and processes that are part of the geologic history of
the area have given the hydrogeologic system properties
that control the present-day movement and circulation of
ground water.
Bedding and planes of metamorphic foliation gener-
ally are folded and tilted and can have almost any attitude
and orientation. Fractures, bedding. and foliation create in-
homogeneities in the rocks and result in permeability that is
usually greatest parallel to bedding. foliation, and zones of
fracture concentration: permeability is usually least at right
angles to the plane of these features.
Bedrock may be exposed at land surface on steep
slopes, rugged hilltops. or in stream valleys. but nearly
everywhere else it is overlain by unconsolidated material
that may reach depths greater than 100 ft. Collectively this
unconsolidated material, which is composed of saprolite,
alluvium, and soil, is referred to as regolith. Saprolite is
clay-rich, residual material derived from in-place weather-
ing of the bedrock. When the bedrock weathers to form
saprolite, the relict structures generally are retained, and the
directional properties of permeability are also retained. In
many valleys. the saprolite has been removed by erosion,
and bedrock is exposed or thinly covered by alluvial
deposits. Soil is present nearly everywhere as a thin mantle
on top of both the saprolite and alluvium. The water-storing
and transmitting characteristics of bedrock and regolith and
the hydrologic relation between them determines the water-
supply potential of the ground-water system in the Piedmont
and Blue Ridge provinces.
Hydrogeologic Units
Within the Piedmont and Blue Ridge of North Caro-
lina there are hundreds of rock units that have been defined
and named by various.conventions more in keeping with
classical geologic nomenclature than hydrologic terminolo-
gy. The geologic nomenclature does little to reflect the
water-bearing potential of the different units. To overcome
this shortcoming and to reduce the number of rock units to
the minimum necessary to reflect the differences in water-
bearing potential, a classification scheme based on origin,
composition, and texture was devised (table I). The ration-
ale behind the hydrogeologic units shown in table I is the
hypothesis that these factors would be linked not only to a
rock's primary porosity but also to its susceptibility to the
development of secondary porosity in the form of fractures
and solution openings. The composition and texture would
also determine, in part, the rate and depth of weathering of
these units and the water-bearing properties of the resulting
regolith.
The origin of the hydrogeologic units is indicated by
the rock class (igneous, metamorphic, or sedimentary) or
A6 Ground-Water Resources of the Piedmont-Blue Ridge Provinces of North Carolina
Table 1. Classification and lithologic description of hydrogeologic units in the Piedmont and Blue Ridge provinces of
North Carolina
Symbol Hydrogeologic unit
IPI . .. . . . Igneous, felsic intrusive .... , ..... , ..... .
III.. . . . . . Igneous, intermediate intrusive ......... .
IMl ..... Igneous, mafic intrusive ............... .
MIF.. .. . Metaigncous, felsic ................... ..
Lithologlc description
IGNEOUS INTRUSIVE ROCKS
Light-colored. m~dy granitic rocks, fine-to come-grained, some prophyritic.
usually massive. locally foliated; includes granite, granodioritc, quartz diorite.
quartz monzonite. a1askites.
Gray to greenish-gray, medium-to coarse-grained. massive rocks of dioritic
composition; includes assemblages of closely associated diorite and gabbro
where they are too closely associated to be mapped separately.
Dark-greenish-gray to black, medium-to coarse-grained intrusive bodies: prima-
rily gabbroic in composition. includes closely associated gabbro and diorite
where they an, too closely associated to be mapped separately. ulrramafic rocks.
diabase. dunite.
MET AMORPIIIC ROCKS
Metalpeous Rocks (lotrusln)
Light-colored, massive to foliated metamorphosed bodies of varying
assemblages of felsic intrusive rock types; local shearing and jointing arc
common.
MIi ..... Metaigneous, intermediate .•............ Gray to greenish.gray. medium-to coarse-grained, massive to foliated, well-
jointed, metamorphosed bodies of dioritic composition.
MIM .... Meta.igneous, mafic ..................... Massive to schistose greenstone, amphibolite, metagabbro and mctadiabase, may
be strongly sheared and recrystallized; metamorphosed ultramafic bodies are
often strongly foliated, altered to serpentine. talc. chlorite-trcmolite schist and
gneiss.
Metavol.canic Rocb (Extrusl~Eruptlvt)
MVF .... Metavolcanic, felsic .................... Chiefly dense. fine-grained, light-colored to greenish-gray felsic tuft's and fclsic
crystal tuffs. includes interbedded fclsic flows. Felsic lithic tuffs, tuff breccias,
and some epiclastic rocks: recrystallized fine-grained groundmass contains feld-
spar, sericite. chlorite. and quartz. Often with well-developed cleavage, may be
locally sheared: phyllitic zones are common throughout the Carolina slate belt.
MVI . . . . Metavolcanic, intennediate...... . . . . . . . . Gray to dark-grayish-green ruffs and crystaJ tuffs generally of andesitic composi-
tion; most with well-developed cleavagei also includes interbedded lithic tuffs
and flows of probable andcsitic and bmaltic composition and minor felsic vol-
canic rocks.
MVM . . . Metavolcanic, mafic...... . . . . . . . . . . . . . . Grayish-green to dark-green. fine-to medium-grained andcsitic to basaltic tuffs.
crystal tuffs. crystal-lithic tuffs. tuff breccias and flows; pyroclastic varieties
may contain lithic fragments; usually exhibits prominent cleavage; alteration
minerals include chlorite, epidote, calcite. and trcmolite-actinolite.
MVE.. . . Metavolcanic. cpiclastic......... . • . . . . . . Primarily coarse sediments including lnterbedded graywackes and arkoses and
minor conglomerates, interbedded argillites and felsic volcanic rocks; much of
the sequence is probably subaqueous in origin and most of the rocks were
derived from volcanic tcrranes.
MVU .... Mctavolcanic, undifferentiated ........... Volcanic rocks of all origins (extrusive and eruptive) and compositions (felsic to
mafic) interbedded in such a complex assemblage that mapping of individual
units is not practical.
Metasedlmentary Rocks
ARG . . . . Argillite .................... : . . . . . . . . . . Fine-grained, thinly laminated rock having prominent bedding plane and axial
plane cleavage; locally includes beds of mud.stone, shale, thinly Jaminated silt-
stone, conglomerate, and felsic volcanic rock.
GNF .... Gneiss. felsic ............. , . . . . . . . . . . . . Mainly granitic gneiss; light-colored to gray, fine-to coarse-grained rocks. usu-
ally with distinct layering and foliation, often interlayered with mafic gneisses
and schists.
GNM .... Gneiss, mafic .................•........ Mainly biotite hornblende gneiss; fine-to coarse-grained, dark-gray to green to
black rock, commonJy with distinct layering and foliation. often interlayered
with biotite and hornblende gneisses and schists. and occasional amphibolite
layers.
MBL. . . . Marble .......... , . . . . . . . . . . . . . . . . . . . . . Fine-to medium-grained, recrystallized limestone and dolostone: found prima-
rily in the Blue Ridge belt.
PHL.. . . . Phyllite.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Light-gray to greenish-gray to white, fine-grained rock having well-developed
cleavage; composed primarily of sericite but may contain chlorite: phyllitic
zones arc common throughout the Carolina slate belt and probably represent
zones of shearing alt.uough displacement of units is usually not recognizable.
Analysis Relating Well Yield to Construction and Siting, Piedmont-Blue Ridge Provinces, North Carolina A7
Table 1. Classification and lithologic description of hydrogeologic units in the Piedmont and Blue Ridge provinces of
North Carolina-Continued
Symbol Hydrogeologic unit lithologic description
QTZ .. .. Quartzile ............................. . Me1asands1one, often feldspalhic 10 highly feldspall>ic. ll>in-10 !hick-bedded
with occasional graded bedding, includes meta•arkose and metaconglomerate;
often interbedded with mica schist, phyllite, and slate.
SCH . . . . Schis1 ................................ . Schistose rocks containing primarily the micas muscovite or biotite or both.
occasional sericite and chlorite schists; locally interlayered with hornblende
gneiss and schist. commonly with distinct layering and foliation.
SLT ..... Stale ................................. . Fine.grained metamorphic rock formed from such rocks as shale and volcanic
ash. possesses the property to part along planes independent of the original bed·
din (slaiy cleavage).
MISCELLANEOUS
TRI . . . . . Triassic sedimentary rocks ............. . Mainly red beds. composed of shale, sandstone, arkose. and conglomerate (fan•
glomerate near basin margins).
CPL. . . . . Coastal Plain basement ................ . Undifferentiated crystalline ba.~ment rocks of igneous and metamorphic origin
overlain unconfonnably by sedimentary sands. gravels, clays, and marine
deposits.
subclass (melaigneous, melavolcanic. or metasedimenlary).
The composition of the igneous, metaigneous, and meta-
volcanic rocks is designated as felsic, intennediate.or mafic
excepl for lhe addition in the me1avolcanic group of
epiclastic rocks and compositionally undifferenliated rocks.
These las1 two groups were necessary because of lhe
significant areas of epiclas1ic rocks where reworking by
sedimenlary processes and admixlure of terrigenous sedi-
ment during deposition made lhe rocks lexturally distinc1
and the other areas where the complex and small-scale
stratigraphic changes made differentiation of separale units
impractical. Composition is also shown in lhe me1asedi-
mentary units of gneiss, marble, and quartzite. The other
melasediments are designated primarily on lhe basis of
texture (grain size, degree of metamorphism, and develop-
menl of foliation).
The two miscellaneous classifications account for the
sedimeniary rocks within 1he Triassic basins and 1he undif-
ferentialed crys1alline basement rocks east of 1he Fall Line
that are overlain unconfonnably by sediments of Crelaceous
age and younger.
By using the classification scheme in table I and the
most recent geologic maps available (fig. 4), a hydrogeo-
logic unit map was compiled for the study area. Part of this
map for Guilford and Alamance Counties in lhe north-
central Piedmont (fig. I) is shown in figure 5. Well-localion
maps were later superimposed on lhis hydrogeologic unit
map, and the units corresponding to lhe well locations were
coded and entered inlo a computerized dala file for analysis
10 detennine the well yields in each unit.
Geologic Belts and Terranes
The Piedmont and Blue Ridge have been divided inlo
a number of northeast-ttending geologic bells (fig. 6).
Within a belt, rocks are to some degree similar to each other
wi1h respect to general appearance, melamorphic rank,
structural history, and relative abundance of igneous,
metaigneous, metasedimeniary. and metavokanic rocks
(Butler and Ragland, 1969). Areally, the most significant
are lhe Blue Ridge, Inner Piedmonl, Charlolle. Carolina
sla1e, and Raleigh belts. Two geologic terranes importan1 10
Ibis study have been added to lhe generally recognized
belts. These are lhe Triassic basins and lhe Coaslal Plain
immedia1ely eaSI of lhe Fall Line. A brief summary of lhe
belts and lhe hydrogeologic units 1ha1 conSlitute the belts is
given in table 2. Wells lapping rocks wilhin 1hese bells and
terranes were analyzed to determine well yields within each
area.
COMPILATION OF THE DATA BASE AND
STATISTICAL PROCEDURES
Information on 6,224 wells was compiled from pub-
lished sources (fig. 2) and s1atis1ically analyzed lo identify
relations between well yield and various geologic, lopo-
graphic, and construction faclors. This compilation con-
lained well records from every coumy in lhe 65-counly
study area and included 4 I 9 wells that derive water from
crys1alline rocks buried benealh the 1hin sedimen1ary cover
along the western edge of lhe Coastal Plain (fig. 3 ).
Information Categories in the Data Base
Specific types of infonnation calegories (variables) in
the data base included (I) the county where the well is
located, (2) the published well number. (3) 1he 101al depth
of the well, (4) well diameter, (5) casing depth, (6) static
wa1er level below land surface, (7) yield, (8) inlended use
when drilled, (9) lhe 1opographic selling of lhe well sile,
(10) the hydrogeologic unil into which lhe well is drilled,
( 11) lhe geologic bell or terrane in which lhe hydro geologic
unit is found, and ( 12) 1he reference to 1he published report
AB Ground•Water Resources of the Piedmont-Blue Ridge Pro\linces oi North Carolina
...
..
a.
b.
C.
d.
e.
I.
g.
h.
... 83' ... .,. ... 10• 78' 77• 78°
-------.1----,------=----------
e
a
INDEX TO GEOLOGIC MAPS USED IN COMPILATION
Compilation by C.C. Daniel Ill and Robert A. Payne, 1984·85,
Based on sources indicated by letter as loUows:
Stuckey, 1958 L McDaniel, 1980
Hadley and Nelson, t 971 ~ Wilson and Spence, 1979
Rankin and others. 1972 k. wnson, 1979
Goldsmith and others, 1982 I. WllsOn, t 981
Espenshade and others, 1975 m. Stromquist and Sundelius, 1975
Carpenter, 1982 n. Stromquist and others, 1971
Burt, 1981 0. Selders, 1981
Wilson and others, 1981
'\\
/
~
0
0 50 100 MILES
0 50 100 KILOMETERS
f',gure 4. Areas of geologic maps used in compilation of the hydrogeologic unit map of the Piedmont and Blue Ridge provinces.
~ ..
Cl
i! ~ ~ I ~ .. ! C
ij
~
5!. ;. ..
[
3 g
7 .. i: .. .. ;;:
":i ..
i. ~ n i
0 ~ z 0 :,. ,,.
n .. a if ..
TN
... w "' -..
"' G F 0 R
0., "'
u.. ..,
, .. --i/i
,,, I 1'1
~-WI i ------
R A N D
""'
""
0 • ,,."r~t
0 • 10 KILOMETERS
- --
"" D .
<5f ~,I
I 1,111
~~, ...
~~; "" -, ... --~, , i --~ . , '
c:!fa --, M A N C
,..
1!.llPLANATION
HYDROGEOLOGIC UNITS
IFI Igneous, felalc tnlrU&IY8
MIF Melatgneous, fe!sic
MIi Melaigneous, intermediate
MIM Melalgneous, maflc
MVF Metavolcanic, lelaic
MVI Metevotcanic, intermediate
MVM Metavolcanic. mafic
MVE Metavolcank:, eplclastlc
ARG ArgDBta
GNF Gneiss, felaic
GNM Gneiss, mallc
PHL Phyllite
SCH Schist
TRI Trlassic sedimentary
rocks
MAP OF NORTH CAROLINA
G E
""
Figure 5. Hydrogeologic unit map of Guilford and Alamance Counties and vicinity in the north-central Piedmont of North Carolina.
,..
,..
... .,. ,,.
~--
" Cl C 011 0 I A Bt•vard lault
EXPLANATION
IIIU •WUrohY bait
SAUAATOWN MTN. ANT. -Saun11cwn
wo1,1rnaln1 arrtlc:1/no,Lum
SR • 8f!ll1h Rl'fltf dochthOl'I
T..OCB -Oa'rie C~ty Trianle basin
Tr-ORB • Dan River Trlaulc: ba1in
cs -CaroMa st.w b.it
CP • Cou1a• Plain
Boundary betw••n Btua Ridge and Piedmont
phy~aohlc orovlncea
..,....--r Thru11 faun with teeth on uplhrown bloek
... ... ,,. , .. ,,. , ..
VIR01NIA
PIEDMONT
COASTAL PLAIN ---------
0 50 100 11111._(g
50 I 00 KII.OMETERS
Figure 6. Geologic belts, terranes, and some major structural features within the Piedmont and Blue Ridge provinces of North Carolina (from Brown and Parker,
1985).
Table 2. Geologic belts and terranes of the Blue Ridge, Piedmont, and Coastal Plain provinces of North Carolina
(The hydrogcologic units an: described in lable I]
Beh or terrane
Murphy belt .......................... .
Blue Ridge belt ...................... ..
Qiauga belt. .......................... .
(includes Brevard fault zone).
Inner Piedmont belt. ................... .
Smith River .................. , .. , ..... .
allochthon.
Sauratown Mountains .................. .
anticlinorium.
Kings Mountain belt. .................. .
Qiarlotte belt ......................... .
Milton belt. ........................... .
letter
designation
MU ...... .
BR ....... .
CA ....... .
IP ........ .
SR ....... .
SA ...... ..
KM .......
CH ........
MI ........
Boundaries
SU1TOunded by metasedimentacy rocks of Blue
Ridge belt.
Sedimentacy rocks of Ridge and Valley on north-
west and Brevard fault zone on southeast.
Blue Ridge belt on northwest. Inner Piedmont on
southeast.
Chauga and Blue Ridge belts on nonhwcst. Kings
Mountain and Olarlotte belts on southeast.
Blue Ridge belt on northeast and Sauratown Moun-
tains anticlinorium on southeast.
Smith River allochthon on northwest, Inner Pied-
mont belt on southwest. and Dan River Triassic
ba-.in and Milton belt on southeast.
Inner Piedmont belt on northwest and Charlotte
belt on southeast.
Kings Mountain and Inner Piedmont belts on north-
west. Milton belt on north, Gold Hill shear zone
and Carolina slate belt on southwest.
Igneous and metaigneous rocks of Charlotte belt on
south, Carolina slate belt on southeast, Dan River
Triassic basin and Sauratown MountaiIL'i anticli-
norium on northwest.
Dominant hydrogeer
logic units
SCH. SLT. MBL.
ONF. ONM, SCH,
QTZ. PHL.
ONF. ONM.
ONM. MIF.
ONF.
ONM. ONF. QTZ.
SCH, MIF. ONF.
Mil, MIF, MIM, JFI,
MVU.
ONM. ONF.
Gold Hill shear zone................ . . . . GH... . . . . . Metavolcanic and metaigneous rocks of Charlotte PHL.
belt on nonhwest and metavolcanic rocks of Cartr
line slate belt on southeast.
Carolina slate belt...................... CS . . . . . . . . Gold Hill. Charlotte, and Milton belts on north• ARO. MVE, MVU in
southwestern half of
belt-MVF, ARO,
MVU, MIF. Mil in
northeastern half of
belt.
west, Coastal Plain on southeast.
Raleigh belt............................ RA........ Bordered by Carolina slate belt rocks on east and Mlf, ONF, SCH.
west. Coastal Plain sediments on the south.
Triassic basins . . . . . . . . . . . . . . . . . . . . . . . . . TR........ Several bodies of sedimentary rock downfaulted TRT.
into the metamorphic crystalline rocks of the Pied-
mont.
Coastal Plain........................... CP .. .. . . .. Western edge of Coastal Plain province. CPL.
from which the well record was obtained. The total number
of entries for each variable is shown in table 3.
For inclusion in the data base, a well had to satisfy
certain requirements. The well had to be drilled into
bedrock, and the yield and location had to be known. All
wells in the resulting compilation are cased to the top of
bedrock and have no screened or slotted intervals in the
regolith, and nearly all are finished as open holes drilled
into bedrock. A small number of wells included in the data
base have casing. slotted casing. or screen extending into ·
bedrock to prevent fragmental rock debris from entering the
well bore. An extreme example is a well that is 600 ft deep
and is cased to the bottom of the hole. No other well has
more than 300 ft of casing, and only 157 wells, or 2.5
percent, are cased to within the bottom 5 ft of the well.
A12 Ground-Water Resour<ft of the Piedmont-Blue Ridge l'l'OviltCOI of North Carolina
Table 3. Total number of entries for each variable in the
water-well data base
Variable
Total
number of
data entries
County . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.224
Well number.. . .. . . .. . . . . . .. . . . . . . . . .. . .. .. .. . .. 6.224
Total depth .. .. .. . . . .. . . . . . . .. . . .. .. .. .. .. .. . . .. 6.204
Well diameter.. .. .. .. .. . . . . . .. .. . . .. . . . . . .. . . .. . 6,0!iO
Casing depth..... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.038
Static water level. ........... ,.................... 3,130
Yield. .. .. .. .. .. .. .. . .. .. .. . .. .. . .. .. .. .. . .. .. .. 6,224
Use............................................ 6,205
Topographic setting.............................. 5,234
Hydrogcologic unit... .. .. . . . . .. .. .. .. . .. .. .. .. .. 6,224
Geologic belt .. .. .. .. .. .. .. . . . .. .. .. .. .. .. . .. .. . 6.224
Reference....................................... 6,224
The wells range in diameter from 1.25 to 15 inches
(in.), and most (69 percent) of the wells have diameters
between 5.5 and 6.5 in. Only two drilled wells were as large
as 15 in.
Large-diameter bored or dug wells were not included
in the compilation because these wells arc not typical of
modem well construction. Nearly all new wells in the
Piedmont and Blue Ridge arc drilled by air rotary methods.
Further, large-diameter wells arc rarely dug below the top
of bedrock and do not represent attempts to obtain quantities
of water beyond that necessary for domestic supplies.
Transparencies were made of well-location maps
given in the published sources (fig. 2) and overlaid on maps
of the hydrogeologic units and geologic belts to assign the
wells to the units and belts in which they occur. The
hydrogeologic units reported in these publications were not
entered into the data file because of the conflicting variety
of names and naming conventions that were used by the
many authors. The reported hydrogcologic units were not
ignored, however. If a well was located on or near a contact
between units used in this report, the published description
helped guide the choice in the assignment of the unit and in
some places pointed out the need for revisions to the
hydrogeologic unit map. The published reports also were
used to identify wells drilled into diabase dikes. Diabase
dikes arc common in the Piedmont (Reinemund, 1955;
Weigand and Ragland, 1970; Ragland and others, 1983),
but generally they arc too narrow to accurately correlate
with well locations at the scale of the maps being used.
Wells drilled into diabase dikes arc included in the igneous,
mafic intrusive (IM!) hydrogcologic unit. By using a
combination of the new maps and the published descrip-
tions, each well in the data base subsequently was assigned
to I of the 21 hydrogeologic units.
All data related to well construction, yield, topo-
graphic setting, and static water level were entered as
reported. The intended use of each well was inferred from
the listed owner and other information in the remarks
column of the well-record tables. Wells were placed in one
of three use categories: domestic, commercial-industrial.
and public supply. Domestic wells serve single family
residences or, at most, a small number of homes. The
commercial-industrial category includes wells that serve
businesses ranging in size from large mills and factories
down to service stations and small shops. Public-supply
wells serve municipalities. subdivisions, trailer parks, hos-
pitals, churches, campgrounds, and other facilities having a
relatively large number of users.
Every item of information was not available for every
well. The static water level had the fewest number of data
entries; water levels were reponcd for only slightly more
than one-half of the wells. The second smallest number of
entries was for casing depth; less than two-thirds of the well
records had this information. The other variables had much
more complete records. The effect of these incomplete
records will be seen in the statistical analyses that follow.
especially for computations that arc based on more than one
variable. For example, in a calculation of yield per foot of
well depth by topographic setting, the variables yield,
depth, and topographic setting had 6,224, 6,204, and 5,234
data entries, respectively. Yet the final computation was
based on the 5,221 wells for which all three items of
information were available. This was generally the pattern;
the final computation was based on no more than and
commonly fewer observations than the smallest number of
variable entries.
Statistical Procedures
The data were statistically analyzed by using pro-
grams developed by the SAS Institute (SAS lnstinne, Inc.,
1982a) that arc available on the U.S. Geological Survey
computer system in Reston, Va. The most commonly used
SAS procedures were SORT, UNIV ARIA TE, RSQUARE,
GLM, and ANOV A.
The SORT procedure (SAS Institute, Inc., 1982a) is
a SAS utility procedure that sorts observations in a data set
by one or more variables. In this study, the SORT proce-
dure was used to sort the well data by topographic position,
use, hydrogeologic unit, and geologic belt so that statistics
could be computed for the soncd groups of data.
The UNIVARIATE procedure (SAS Institute, Inc.,
1982a) produces simple descriptive statistics including the
mean, median, range, standard deviation, and quantiles for
numeric variables.
A SAS procedure called RSQUARE (SAS Institute,
Inc., I 982b) was used for regression analysis because it
allows many possible regressions to be fitted to the data and
systematically analyzed to identify those combinations of
variables that best explain the variation in the data. Those
variables that repeatedly appeared in the models offering the
highest r•squarc were further tested by using SAS procedure
GLM (General Linear Models) (SAS Institute, Inc.,
Analysis Relating Well Yield to Construction and Siting, Piedmont-Blue Ridge Provinces, North Carolina A 13
1982b), which uses the method of least squares to determine
regression coefficients, intercepts, and statistical properties
of the models being tested.
Analysis-of-variance tests using the procedure
ANOV A (SAS Institute,Inc., 1982b) were made of the data
in the topographic classifications, hydrogeologic unit clas-
sifications, and geologic belt classifications to determine if
any of the apparent differences, or Jack of differences, in
mean values are statistically valid. Because the sample cells
have unequal numbers of observations, Tukey's studentized
range test, honestly significant difference (HSD) procedure
(Steel and Torrie, 1960, p. 109-110), was used to make the
multiple comparisons and to test for significant differences
at the 0.95 confidence level. Unequal cell size was not the
only reason for using Tu key's procedure. It is also a
conservative test compared to other procedures such as
Duncan's multiple-range test (Steel and Torrie, 1960. p.
107-109), which is most effective with samples of equal
cell size, and controls for the experiment-wise error rate
rather than on a percomparison basis. As a result. there is
Jess chance that Tukey' s procedure will declare some
differences between means to be significant even when the
means are a homogeneous set.
Duncan's multiple range test and the Duncan-Waller
k-ratio t-test were also attempted on data sets that were
manipulated to generate equal cell sizes. Equal cell sizes
were generated by taking the percentile values of frequency
distributions of data within a sample cell: this produced
cells containing I 00 observations. This transformation
woticed well for sample cells having large numbers of
observations in a distribution that was not excessively
skewed (skewness Jess than 4.0) and with similar values of
skewness. When these two conditions were not met, the cell
mean from the frequency distribution was different from the
cell mean of the raw data. Because of this problem, the
analysis-of-variance tests using Duncan's method and the
Duncan-Waller method produced inconsistent results,
although a pattern usually emerged that was similar to the
results from Tukey's procedure. Because of the properties
ofTukey's procedure, the nature of the data that were being
tested, and for overall consistency, Tukey' s HSD procedure
was used for all analysis-of-variance tests described in this
report. Further discussion of analysis of variance, including
Tukey's HSD procedure, can be found in Steel and Torrie
(1960) and SAS Institute, lnc.(1982b).
RELATION OF WELL YIELD TO
CONSTRUCTION PRACTICES AND
SITING OF WELLS
Results of the Analysis
The first group of statistics, presented in table 4,
characterize the wells in the study area with regard to their
physical and hydrologic characteristics. In the left half of
the table, the average and median values of these character-
istics are shown for wells in each of six topographic
settings. The topographic settings are arranged in order of
Table 4. Average and median values of selected well characteristics according to topographic setting compared to
statistics for all wells
Topographic setting All wells
Well characteristic,
Draw Valley Slope Flat Hill R;dgc Average First Median Third Ninth Number
Average yield1 quartile guanlle dectHe of wells
(gallons per minute) ......... 33.J 25.7 17.1 16.8 to.8 9.7 17.2 5 10 20 36 5,234
Median yield
(gaUons per minute) ......... 20 IS 10 10 6 6 5,234
A vcrage yield per foot
(gallons per minute per foot) .. .220 .205 .128 .131 .093 .086 .131 .038 .080 .165 .300 5,221
Median yield per foot
(gallons per minute per foot) .. .154 .14.l .082 .083 .056 .058 5.221
A vcragc depth
(feet) ...................... 175.1 157.8 152.6 150.0 150.2 153.1 154.0 85 119 179.5 297.4 5,221
Median depth
(C..t) ...................... 134 104 118 119 117 112 5,221
Average casing
(feel) ...................... 52.4 49.0 53.6 55.0 51.2 57.2 52.9 28 45 70 97 3,375
Median casing
(f .. t) ...................... 46 40 47 50 43.5 42 3,375
Ave~ water level (feet low land surface) ..... 24.3 18.6 32 .. l 28.6 38.6 43.6 32.2 18 28 40 60 2,825
Median watcr level
(feet below land surface) ..... 20 ,s 28 25 34 40 2,825
Average saruratcd thickness
of rcgolith (feet) ............ 31.7 35.4 23.6 27.5 20.5 18.4 24.8 0 15 40 65 2,161
Median wuratcd thickness
of regolith (feet) ............ 25 29 14 19 9 10.5 2,161
'Unadjusted for differences in depth and diameter.
A 14 Ground-Waler Resources of the Piedmont-Blue Ridge Provinces of North Carolina
decreasing average (mean) yield. The statistics of well
characteristics in the six topographic settings can be com-
pared to statistics computed for all wells in the sample that
are given in the right half of the table, which defines the
frequency at which a given value of a well characteristic can
be expected to occw. At the first quanile, 25 percent of the
wells in the sample have values that fall below the given
value; at the median or second quartile. half the wells have
values below the given value; at the third quanile, 75
percent of the wells fall below the given value; and at the
ninth dectile. 90 percent of the wells are below the given
value.
The yield per foot of well depth and saturated
thickness of regolith are computed characteristics. The yield
per foot is the yield divided by the total depth of the well.
The saturated thickness of regolith is the difference between
the depth of casing and the depth of the static water level.
If the water level in a well was below the bottom of the
casing, the saturated regolith thickness of that well was
considered to be zero.
In the computation of the saturated thickness of
regolith, casing depth was used to estimate regolith thick-
ness. The depth of surface casing in a drilled well is a good
approximation of regolith thickness in the Piedmont and
Blue Ridge (Daniel and Sharpless, 1983; Snipes and others,
1983). Surface casing is usually set no more than I or 2 ft
into fresh bedrock, just below the interface between it and
the overlying regolith. Wells drilled in North Carolina since
passage of the North Carolina Well Construction Act of
1967 (Heath and Coffield. 1970), however, are required to
have a minimum of 20 ft of casing, regardless of how
shallow the bedrock may be. Casing data from these wells
can lead to overestimated regolith thickness. Fortunately,
from a statistical standpoint, many of the records used in
this study were for wells drilled prior to I 967. Records
of casing depths as shallow as I ft for wells on bare-rock
exposures are included in the data compilation. These data
better reflect the natural range of depths to bedrock and thus
provide for a more accurate approximation of regolith
thickness.
The data in table 4 show a general pattern of
decreasing yield, yield per foot, and saturated thickness of
regolith at higher topographic settings (ridges and hilltops).
The depth to the water table follows the opposite pattern.
The amount of casing and the well depth do not show any
apparent relation to topographic setting e>cept that wells in
draws average from 17 to 25 ft deeper than wells in other
topographic positions.
Analysis-of-variance tests of the data in the six
topographic settings of table 4 were made in two steps, first
on the data in the six settings and then on grouped data
where significant differences were not found. In the first
analysis, casing depth was not statistically different in any
of the six topographic settings. The average depths for wells
on slopes, flats, hills, and ridges were also statistically the
same. The yield and depth of wells located in draws was
statistically different (greater) from the yield and depth of
wells located in valleys and other topographic settings. The
remainder of the data tended to cluster in three topographic
groups made up of those wells in draws and valleys, on
slopes and flats, and on hills and ridges. It is important to
point out that analysis-of-variance tests on yield per foot
data indicate that wells in draws and valleys are statistically
one group, because of adjustment of the yield to account for
the differences in well depth in these two topographic
settings. This finding is also an indication of the relation
between well yield and well depth that will be described in
more detail.
In the second pan of the analysis. the data were
merged according _to the three principal topographic groups
identified in the first pan of the test. Analysis of variance on
the grouped data still found no difference in casing depth,
nor did well depths on slopes and flats differ from well
depths on hills and ridges. Because the statistical tests
showed that the yield per foot for wells in draws was the
same as for wells in valleys, the yield and depth data for
wells in the~-e settings were combined. 1ne remainder of the
data fell into one of the three topographic groups and were
statistically distinct from the other groupings for a given
variable. Yields of wells in draws and valleys average
nearly three times the yields of wells on hills and ridges.
The highest yielding wells also were the wells having the
greatest saturated thickness of regolith and the highest water
level.
Statistics showing the depth to the water table. casing
depth, and saturated thickness of regolith for various
topographic settings in the three physiographic provinces in
the study area are given in table 5. The influence of
topography on the depth to the water table is apparent. The
effect of the higher relief and more rugged topography in
the Blue Ridge is reflected by the greater depths to the water
table than in comparable topographic settings in the Pied-
mont. An unexpected finding is the similarity of the
satwated thickness of regolith in the Piedmont and Blue
Ridge. This may be due in pan to compensating conditions
created by differences in rainfall and relief in the two
provinces. Generally. there is more rainfall and more
ground-water recharge in the Blue Ridge than in the
Piedmont. But there also is greater relief, and presumably
steeper ground-water gradients, in the Blue Ridge that
results in greater ground-water discharge. Although there is
less rainfall in the Piedmont (Eder and others. 1983), the
lower relief results in lesser rates of ground-water dis-
charge. Thus, the amount of ground water in long-term
storage in the two provinces is roughly equal.
Although the data for casing depth in table 4 indicate
little difference between wells in different topographic
settings when the study area is considered as a whole, the
data in table 5 show that there is an increase in casing depth
at higher topographic settings in the Blue Ridge. For wells
in the Piedmont, there is no apparent relation between
Analysis Relating Well Yield to Construction and Siting, Piedmont-Blue Ridge Pro\'inces, North Carolina A 1 s
Table 5. Summary statistics defining depth to water, casing depth, and saturated thickness of regolith according to
topographic group in the Blue Ridge and Piedmont physiographic provinces
[Statistics for ~'Clls pcoctrating bedrock beneath the-wcs1cm ed~c of the Coastal Plain sediments arc gi\.·cn for comparison)
elu@ Ridge Piedmont Coastal f'taln,
Well characteriMic OraW!i Slopes Hills Number DraW!i Slopes HIiis All Number Number All All and and and wi!lls of wells and and and wells of wells welts of wells 'lalli::n """ ridges valle;r:s flats ridges
A vcmgc water level
(feet below land surface) ...... 23.4 37.5 6'.?.9 37.l 507 22.l 29.3 36.8 31.3 2.326 18.8 145
Median water level
(feel below land surface) ..... 18 35 50 30 507 20 25 32 27 2,326 15 145
Average cn.c;i.ng
(feet) ..........•...........• SO.I 57.7 66.6 56.8 698 52.7 53.2 50.0 52.0 2,684 71.7 293
Median casing
53.5 698 45 293 (feet) ....................... 43 55 60 46 41 44 2.685 63
A verasc saturated thickness
or rtgolith (fttll .. , ....... 32.2 27.6 20.8 :?.8.0 422 33.6 '.?4.6 20.4 24.0 1,749 47.7 112
Median saturated thickness of rcgolith (fccu .......... 28 20 lO 20 421 2R 15 9 13 1,749 44.5 112
1Topogruphy of bedrock swface cannot be determined. Influence of topography on well yield in Coast~) Plain is unknown.
Table 6. Relation of selected well characteristics to the use of the well
[gal/min, galloru. per minutt: (gal/minl/ft. gallons per minute per foot: ft, feet]
Percentages of wells according to use in selected
topographic settings
Statistical summary of well characteristics
according to use
Use of well
om, Valll!y Slope Flat Hut
C>omestic ......... , ...... 47.5 54.5 71.5 72.0 82.0
Commercial-industrial ...... 31.0 27.9 13.8 12.5
Public ................... 21.5 17.6 14.7 12.5
casing depth and topographic setting. This difference may
be due to the greater relief in the Blue Ridge.
In relation to use (table 6), more than one-half the
wells in draws were commercial-industrial or public supply,
and nearly one-half the wells in valleys were in the same
two use categories. At the other topographic extreme, more
than 80 percent of tbe wells on hills and ridges were
domestic supply wells. The yields of domestic wells aver-
age about one-third the yields of the commercial-industrial
and public-supply wells and are about 100 ft shallower.
lnfonnation on well diameter (not shown) also indicated
that domestic-supply wells had the. smallest average diam-
eters and public-supply wells had the largest. Fewer than 2
percent of domestic wells were 8 in. in diameter or larger.
whereas 20 percent of the commercial-industrial and 26
percent of the public-supply wells were 8 in. or larger. The
implication of the data in this table is that public-supply and
commercial-industrial wells are more likely to be sited and
constructed in an effort to obtain as much water as possible.
whereas many domestic wells are at sites on hills and ridges
selected for setting and view. Also, many secondary roads
tend to follow the low ridgelines and drainage divides
connecting the better drained agricultural land. and many
rural homesites are near these roads.
7.8
10.2
Rktge
Averle Average Average
Average Average water Number yield yleldtfoot dept casing level of wells (gallminl (gallmln)lh (fll (ft) (ft)
83.6 I 1.6 0.117 123.6 51.8 30.8 4,408
3.3 27.7 .161 216.5 60.9 31.2 872
13.1 33.9 .171 229.8 69.2 34.7 905
The summary statistics strongly suggest a relation
between well yield and well depth and diameter, a definite
relation between topographic group and several well char-
acteristics, including yield, and an apparent cultural bias in
the siting and construction of wells related to the intended
use of the well.
The relation of well .yield to rock type, which has
been described by many past authors, also may be distorted
by cultural bias in siting and construction. For example. in
the upper Cape Fear River basin, as described by Daniel and
Sharpless ( 1983), the most productive rock unit is the
mafic.volcanics unit. They showed a concentration of
high-yield wells in central and northwestern Alamance
County coinciding with the area underlain by the mafic-
volcanics. Historically. this area has been a major center of
textile manufacturing and has a number of factories and
mills. The smaller towns have public water systems fur-
nished by wells. and many of the mills have, or have had,
their own ground-water supply systems. Thus, the area
underlain by the mafic-volcanics unit may have appeared to
be the most productive simply because it contained more
large-diameter. deep wells than any other area in the basin.
The relation between well yield and well depth and
diameter is indicated in figure 7. where average yield,
A 16 Ground•Water Resources of the Piedmont•Blue Ridge Pro\'inces of North Carolina
(/) w ::c
()
~
~
12
1 1
10
9
569 0.235
::::s~ 482 0.197 ~l\:.
~Average depth In feet
~·i~~i~~ .... ••,•.·
8 321 0.206
a: o 148 '-------Average yield per loot of w 7 I-· well depth. In gallons w :; 6
per minute per foot :s 141 0.134
0
w 5 0.113
a: Average yield 17 .0 gallons per minute
0 4 [D 0.0112 Average yield per toot 0.131 gallons
.J per minute per loot .J 3 130 0.076 w Average depth 151.9 feet ;!:; 2 121 0.065 Analysis based on data from 6,074 wells
1 81 0.092
0 20 40 60 80 100 120 140 160 180
AVERAGE YIELD, IN GALLONS PER MINUTE
Figure 7. Variation of average yield, average depth, and average yield per foot of well depth with wellbore
diameter.
average depth, and average yield per foot of well depth are
shown for wells of different diameters. 'The diameters arc
subdivided into 1-in. intervals; the actual diameters of the
6,074 wells summarized in figure 7 range from 1.2 in. to 12
in. The significance of figure 7 is the systematic increase in
yield and yield per foot that coincides with an increase in
depth and diameter.
To better define the narure of the interactions that are
indicated in figure 7, least-squares regression analysis was
employed. Yield and yield per foot of well depth were
treated as dependent variables to be explained in terms of
well depth and well diameter with the additional factor of
topographic setting to be considered. Including depth and-
diameter and interaction terms based on depth and diameter,
a total of 20 potential variables were tested in model
combinations containing from two to six variables in any
one model. The models finally identified as having the best
properties and best predictive capabilities contained three
variables. Models containing additional variables were only
increasingly complex without offering much more in pre-
dictive capability. 'The variance in the model of yield versus
depth and diameter was reduced by subsctting the data
according to the three topographic groups identified earlier
and recomputing the regression coefficients to produce
three regression equations of the general fonn:
yield = a -b (depth) + c (depth X diameter)
-d (depth2 x diameter)
where a. b. c, and d are regression coefficients.
The regression equations and contour plots of the
trend surfaces defined by these equations are shown in
figures 8, 9. and 10. 'The contour plots are limited lo the
range of known data. There are no small-diameter wells in
the data set deeper than the no-data boundary. The deepest
well in the data set is a 6-in. diameter well that is 1,301 ft
deep. A number of larger diameter wells in the data set are
nearly as deep. 'The shallowest well is 20 ft deep and 6 in.
in diameter.
Information contained in figun,s 8, 9, and 10 repre-
sents several significant new findings regarding drilled
wells in the crystalline rocks of the Piedmont and Blue
Ridge. The surfaces shown in these illustrations represent
the best average fit through yield data that has considerable
Analysis Rei.tins Well Yield to Conslnlction and Sitil'l!, Piedmont-Blue Ridge Provinces, North Carolina A 17
10
8
6
4 154 LINE OF EQUAL WELL YIELD--
Intervals 18 and 19 gallons per minute
Analyala based on data from 806 wells
2
/
I
I
YIELD •11.1570 -0.11337 (DEPTH)+0.04190 (DEPTH· DIAMETER)
-2.09739• 10' (DEPTH 2• DIAMETER)
0 '----'-----'----'---'---'----'----'-----'---'--'----'-----L..---'
0 200 400 600 800 1,000 1,200
TOTAL WELL DEPTH, IN FEET
Figure 8. Contour plot of trend surface showing relation between well yield, total well depth, and well diameter for wells
that are located in draws and valleys.
variation at any given point. That is, for a point on any of
the three contour plots there may be several wells of the
same depth and diameter, all having different yields. This is
important in interpreting the significance of the axes of the
yield surfaces and why the average yield for wells of a given
diameter decreases to the right of the yield-surface axes.
Take for example, a point on the surface of figure 9 (wells
on slopes and flats) representing a well depth of 525 ft and
a diameter of6 in. The predicted average yield at this point,
which also is on the yield-surface axis, is 32 gal/min. If a
6-in. well were drilled to this depth and had no water, two
things could be done: stop or drill deeper. If drilling were
stopped, that zero yield would be averaged with the yields
of all other 6-in., 525-ft wells, which would average about
32 gal/min. If the well is drilled deeper and finally obtains
water, the yield of that well averaged with other wells of the
same depth will be less than at the yield-surface axis. Thus,
for a given diameter well, the yield-surface axis represents
the depth at which the maximum average yield will be
obtained. Beyond the depth indicated by the axis, the
chances of obtaining significant amounts, or additionala-
mounts, of water decrease rapidly.
This is perhaps better illustrated by figure 11 which is
in effect a cross section of figures 8, 9, and 10. The figure
is for a narrow range of well diameters. average 6 in., and
shows the average yield and yield per foot for wells in
intervals of well depth. The large data base of wells having
diameters between 5.5 and 6.5 in. provides a sufficient
number of wells in each depth interval to give a consistent
picture and reduce scatter. A maximum average yield is
reached in the interval between 500 and 550 ft (fig. 11 ).
which is the approximate location of the yield-surface axes
for 6-in. wells in figures 8, 9, and 10. The likelihood of
A18 Ground-Water Resources of the Piedmont-Blue Ridge Provinces of North Carolina
rn w
10
~ 8
~
i!:;
ffi 1-w ~ :;!;
C
..J
6
uJ 4 ~ 7 113 LOIE OF EQUAL WELL YIELD--
lntervala 15 and 1 e gallons par minute
Analyala baaed on data from 2,681 wells
2 YIELD•8.1948 -0.11329 (DEPTH)+0,03494 (DEPTH-DIAMETER)
-1.50931 • 1()5 (DEPTH~ DIAMETER)
0 '----'---'----'----''----'----'--...1....----'--...I....----'----'-----'------'
0 200 400 600 800 1.000 1.200
TOTAL WELL DEPTH, IN FEET
Figure 9. Con lour plot of trend surface showing relation between well yield, total well depth, and well diameter for wells
that are located on slopes and flats.
obtaining significant additional quantities of water from
6-in. diameter wells decreases rapidly below depths of 550
ft. However. the increase in yield with increasing depth (up
to the optimum depth) docs not occur in proportion to depth
but actually decreases as the ratio to depth.
By subsetting the well data by topographic groups,
the regression analysis has resulted in three graphs (figures
8. 9, and 10) that at any well depth and diameter retain the
relative magnitudes of yields identified in table 4. At any
position on lhe graphs, the average yield for wells in valleys
and draws is nearly three times the yield for wells on hills
and ridges. The yield for wells on slopes and flats falls in
between. Al1hough then: are differences in yield, the
yield-surface axes of the three contour plots are nearly
coincident, suggesting that topography may have little
effect on the depth at which the maximum average yield is
attained. 'The real significance lies in the position and shape
of the yield-surface axes, which indicate that (I) well yield
increases with depth to a much greater depth than previ-
ously thought and (2) well yield increases dramatically as
well diameter increases. The curvature of the yield-surface
axes shows that depth is still a limiting factor, especially at
depths greater than 500 to 600 ft as the axes of the yield
surfaces rapidly curve away from the depth axes. However,
the maximum average yield for I 2-in. wells is reached
between 700 and 800 ft. This is much deeper than previ-
ously thought. Cressler and others (1983) recently described
similar large-diameter, deep, high-yield wells from the
Piedmont of Georgia. Even the depth at which 6-in. wells
reach their maximum average yield (about 500 ft) is 200 ft
deeper than is usually recognized in the literature (LeGrand,
1967; Snipes and others, 1983).
Analytis Relating Well Yield lo Construction and Siting, Piedmonl-lllue Ridge Provinces, North Carolina A 19
2
I
I
/
/
/
LINE OF EQUAL WELL YIELD--
Interval ■ e ~nd 7 gallons per minute
Analyala baaed on data from 1,880 wella
YIELD• 8,882 -0.05303(DEPTH+O.O 1588)(DEPTH • DIAMETER)
-7. 1124 7•106 (DEPTH\ DIAMETER)
0 L.___.L,__....1,__-1-_.-1. __ .___.L,_ _ _._ _ _._ _ __,_ __ .__ _ _._ _ __._ _ __,
0 200 400 600 800 1.000 1200
TOTAL WELL DEPTH, IN FEET
Figure 10. Contour plot of trend surface showing relation between well yield, total well depth, and well diameter for
wells that are located on hills and ridges.
Although the regression analysis indicates that aver-
age well yields continue to increase at greater depths than
previously thought, perhaps the most interesting finding is
the dramatic increase in average yield with an increase in
well diameter. The effectiveness of increasing well diame-
ter as opposed to drilling to greater depths is illustrated in
figure 12, which is the result of a regression analysis of
yield per foot versus well depth and diameter. The equation
was derived in the same manner described earlier for the
yield versus well depth and diameter relations. For a well of
a given diameter, the yield per foot of hole is inversely
proportional to the depth of the well, indicating that the
amount of additional water obtained by drilling deeper is
continuously decreasing. For wells of the same depth,
however, increases in diameter are directly proportional to
increases in yield per foot of well.
Well Yields by Hydrogeologic Unit
Well yields were matched to rock types to determine
the relative yields of the different hydrogeologic units. The
yield data were simultaneously sorted by topographic group
to compare the relative importance of hydrogeologic unit
versus topography as a consideration in selecting sites for
wells. The results of these computations to compare yield,
hydrogeologic unit, and topography are presented in table
7. Because yield is strongly influenced by well depth and
diameter, which can lead to cultural bias favoring one
hydrogeologic unit over another, a series of calculations
was performed to remove the variation in well yield
attributed to differences in depth and diameter. By using the
equations (figs. 8, 9, and 10) relating well yield to depth
and diameter for the three major topographic groups, the
well yields were adjusted to an average 154-ft depth and
A20 Cround-Water Resources of the Piedmont-Blue Ridge Provinces of North Carolina
50 ,----....----.-----.------,----~--~----.
w
I-::> z :i
a: w CL
en z 0 ..I ..I < C,
;!;
c5
..I !!:! >-
w C,
< a: w ~
40
30
20
10
0 200 400
"' ....
0
ci
"'
600
Analysis based on data
from 4,298 wells
800 1.000 1.200 1.400
TOTAL WELL DEPTH, IN FEET
Figure 11. Variation of average yield and yield per foot of well depth with depth for wells having
diameters between 5.5 and 6.5 in.
6-in. diameter, the average of all wells in the data set.
Because the influence of topography on well sites in the
Coastal Plain is uncertain. the yields of wells in the Coastal
Plain category were adjusted by using a regression equation
that was computed for the entire data set and disregards
topographic setting. It is nearly the same as the equation for
wells on slopes and flats. The hydrogeologic units !IT
(intermediate composition igneous intrusives), MBL (mar-
ble), and SLT (slate) each had fewer than 15 observations
having the necessary data (depth, diameter, yield, topogra-
phy) to adjust the yields. Statistics for these hydrogeologic
units, therefore, are not given, although the yields were
included in the summary statistics.
A regression of adjusted yields on hydrogeologic
units is shown in figure 13. The average yields range from
23.6 gal/min for SCH (schist) to 11.6 gal/min for TRI
(sedimentary rocks of Triassic age). The average difference
in yield between adjacent hydrogeologic units in the regres-
sion is only 0.6 gal/min. However, owing to the effect of
the large number of wells in the analysis, the hydrogeologic
unit can be used as a statistically reliable estimator (0.99
confidence level) of average well yield.
Analysis-of•variance tests were also used to deter-
mine whether any hydrogeologic units were significantly
different from other hydrogeologic units in terms of yield.
Because the average yields of all hydrogeologic units are
not very different and the range of yields within units is very
large, only those units toward opposite ends of the distri-
bution are statistically different (0.95 confidence level) as
indicated by the inequalities in figure 13.
Three groups of hydrogeologic units stand out in
figure 13. The metavolcanic units and ARO (argillite) form
a group at the low end of the graph with only TRI
(sedimentary rocks of Triassic age) having a lower average
yield. Midway in the range of yields are the igneous units.
At average or slightly above average yields are the metaig-
neous units and QTZ (quartzite). The Piedmont crystalline
Analysis Relating Well Yield to Construction and Siting, Piedmont-Blue Ridge Provinces, North Carolina A21
10
2
0.02
---
LINE OF EOUAL WELL YIELD--
Interval la varlable, gallona per
minute per foot
Analyala baaed on data from 5,077 wells
YIELD PER FOOT• 0.002443 -S.8582 ><105 (DEPTH)
-t-0.012997 (DIAMETER)+1.2S813 (DIAMETER/DEPTH)
0 '---.L--.L--..l.--..L.---'----'----'----L---L---L--....l__....[_ _ _J
0 200 400 600 800 1.000 1.200
TOTAL WELL DEPTH, IN FEET
Figure 12. Contour plot of trend surface showing relation between yield per foot of well depth, total well depth, and
well diameter.
rocks underlying the Coastal Plain have the second highest
average yield regardless of differences in crystalline rock
composition. The high yield of these wells is attributed to
the greater saturated thickness of overburden, which at an
average 47. 7 ft is 1.8 times thicker than the 26.8-ft average
for the rest of the study area based on 2,391 observations.
including wells for which topographic information was not
available.
Well Yields by Geologic Belts and Terranes
Comparison of well yields from the various geologic
belts and terranes generally reflects the average yield of the
predominant hydrogeologic unit(s). The yield data that were
used for this comparison also were corrected to an average
154-ft depth and 6-in. diameter. A regression analysis of
well yields in the various belts is shown in figure 14. The
average difference in yield between belts is 0.9 gal/min.
Average yield varies from a low of about 11.5 gaVmin for
the Smith River allochthon (SR) and Triassic basins (TR) to
a high of about 25.5 gal/min for the Murphy (MU).
Analysis of variance tests found that the average yield of
belts at the upper and lower ends of the data are signifi-
cantly different. The inequalities significant at the 0.95
confidence level are also shown in figure 14.
The belts having the highest yields, the Murphy
(MU), Blue Ridge (BR), Chauga (CA), and Inner Piedmont
(IP). are dominated by medium to high rank metasedimen-
tary rocks, mafic gneisses, schists, and quartzitcs, and they
include smaller ateas of metaigneous rocks, all of which
have above average yields. The Charlotte belt (CH), which
is characterized by igneous rocks intruded into country
rocks of metavolcanic and metaigneous origin (Fullagar,
1971 ). and the Carolina slate belt (CS), which is dominated
A22 Crouncf.Water Resources of the Piedmont•Blue Ridge Provinces of North Carolina
Table 7. Relation of well yields to hydrogeologic unit and topography
[Yield data arc adjusted to account for differences in yield due to d.iffcrecces in well depth and diameter. 'The average well is 6 in. in diameter and I S4
ft deep. The hydrogcologic unit!. are described in table I; gaVrnin, gallons per minute}
Mean yield by toposraphic group Yield of all wells
Hydrogeologic
{gaVmin) ( al/min) Number of
unit Draws and Slopes and Hills and Average First Median Third Ninth wells
valler:s flats ridges guartile guartile dectile
ARO ............ 26.8 16.3 12.5 14.6 7.0 11.5 17.0 27.0 319
CPL' ........... 21.7 9.1 14.5 21.8 37.2 419
GNF ............ 28.3 16.6 11.5 17.4 6.4 12.3 22.3 35.9 741
GNM ........... 33.5 19.6 12.3 19.9 6.5 12.5 23.4 40.7 1.129 IF!.. ............ 24.8 17.8 12.6 17.7 8.1 15.8 23.4 34.4 412
111' ............. 7 IMl2 ............ 24.4 12.1 17.8 4.7 14.0 19.9 44.0 29
MBL' ........... 3 MIF ............ 27.6 20.5 12.4 19.1 7.8 14.0 22.5 35.6 791
MIi ............. 22.1 20.6 ~13)'.;ti 18.4 8.8 16.0 23.3 36.2 284 MIM ............ 26.0 21.6 12:s 19.7 10.2 16.9 28.9 36.7 85
MVE'·· 16.6 11.9 16.9 7.5 11.8 16.0 25.0 95
MVF, ........... 19.0 15.1 9.5 13.0 6.2 11.2 17.8 25.9 280
MVI 17.1 15.5 16.8 9.2 13.4 23.6 35.2 43 MVM;:::::::::: 17.8 7.2 11.9 4.6 7.9 17.4 24.6 63
MVU ........... 27.1 23.4 10.9 20.2 8.1 14.8 24.5 41.2 141
PHL ............ 22.9 21.5 13.6 20.3 9.9 14.5 25.4 44.2 127
QTZ' ........... 20.6 16.8 18.6 4.8 15.2 29.4 46.5 65
SCH ............ 43.3 20.8 11.4 33.6 7.8 15.3 27.5 43.6 199
SL'f" ........... 2
TRI. ............ 19.0 12.2 8.5 11.6 4.7 9.0 14.5 25.5 269
All types 28.7 19.0 11.8 18.2 7.9 13.1 22.0 35.5 5,503
1Topography of bed.rock surface cannot be determined. Influence of topography on well yield in CPL area is unknown.
zStatistics for categories having lc!is than 15 observations are not given.
by metavolcanic rocks (Butler and Ragland, 1969), both are
belts having low average yields.
The areas containing sedimentary rocks, the Triassic
basins (TR) and the western edge of the Coasrnl Plain (CP).
are far apart in average yield, with the Triassic basins
having the next-to-lowest yield and the Coastal Plain the
third highest.
SUMMARY AND CONCLUSIONS
A statistical analysis was made of data from more
than 6,200 wells drilled into the crystalline rocks of the
Blue Ridge, the Piedmont, and the western edge of the
Coastal Plain where crystalline rocks underlie sediments at
shallow depths. This analysis was made to identify factors
that are associated with high-yield wells. The data were
classified according to geologic belts, hydrogeologic units
composed of similar rock types.topographic selling, total
and saturated thickness of regolith, water level, casing
depth, yield, total depth, well diameter, and waler use.
Six topographic seltings were combined into three
groups based on well yields: hills and ridges, slopes and
flats, and draws and valleys. Wells on hills and ridges had
the lowest yields (averaging about 10 gal/min): wells in
draws and valleys, the greatest (averaging about 30
gaVmin). Regolith thickness was about the same regardless
of topographic group. but saturated thickness was least
(about 19 ft) under hills and ridges and greatest (about 34 ft)
under draws and valleys. Average yields in the geologic
belts and hydrogeologic units ranged from about 11 to 25
gaVmin. There was considerable scatter in yields in all
geologic belts and hydrogeologic units. Of 14 geologic
belts, 10 were statistically different on the basis of well
yield, as were 8 of 21 hydrogeologic units.
About 70 percent of the wells were drilled for
domestic use and, on the average, yielded about 11 gal/min;
80 percent of these wells were located on hills and ridges.
The 30 percent of the wells drilled for public supply and
commercial-industrial supply yielded about 30 gal/min on
the average; about 50 percent of these wells were located in
draws and valleys. The domestic wells had an average depth
of about 125 ft; the public-supply and commercial-
industrial wells, about 225 ft. Fewer than 2 percent of the
domestic wells were 8 in. in diameter or larger. whereas
nearly 25 percent of the public-supply and commercial-
industrial wells were 8 in. or larger.
Selecting the most favorable hydrogeologic unit or
geologic belt alone can improve the chance of increasing the
yield of the average 6-in. diameter, 154-ft deep well from
Analysis Relating Well Yield to Construction and Siting, Piedmont~Blue Ridge Provinces, North Carolina A23
I:; ..
C'I i! " :I: i ..
I a
0 -1
::!! a. ~ ~ i .. ,I
~ .. a
1·
2. z i ,..
n w a = ~
w I-:::, z
::E
a: w
11.
en z
0 ..J
..J < C,
z
c:i
..J !:!:! > w C, < a: w ~
25,--,---,.----,---,----,-----,------r----r-----r--~--,---,---,---,----,----,---,......--,--~
24
22
20
18
16
14
12
gallons per minute
lnequalltles below were Identified by analysls ol
variance at the 95 percent confidence level
SCH ;t:ARG, MVF, MVM, TRI
CPL, GNM ;i:ARG, MVF, TRI
MIF:;t:MVF, TRI
Figure does not Include the hydrogeologlc unlta marble
(MBL), Igneous intrusive intermediate (Ill)
and slate (SL T) owing to lack ol data
The hydrogeologic units are described in table 1
• • • •
AVERAGE YIELD•-0.588 (HYDROGEOLOGIC UNIT)+23.14
Analysis based on data from 5,489 wells
•
IGNEOUS • • METAIGNEOUS ROCKS ROCKS METAVOLCANIC ROCKS
~
CPL PHL MVU GNM MIM MIF QTZ MIi IMI IFI GNF MVE MVI ARG MVF MVM TRI 10'--.J__.J__.J__J_ _ _L.. _ _L.. _ _L.._...J...__....J...._....J...._....J.... _ _,_ _ _._ _ _._ _ _i _ _i _ _i _ __J _ __J SCH
6 7 8 9 10 11 12 13 14 15 16 17 18 1 2 3 4 5
HYDROGEOLOGIC UNIT
figure 13. Average yield of wells of average construction in the hydrogeologic units of the Piedmont and Blue Ridge provinces of North Carolina.
~ ii:
.;
I f
i
~ ~
f .. c ~
f a l
w ~ ~ ::!:
0:: w a.
rn z g
..J < C,
~
ci ..J !Y >-w C, < 0:: w ~
26i-----.--,-----,--,-----,--,-------r--.------r--,--------r--,----~--~-~
Inequalities below were identified by analysis of
variance at the 95 percent confidence level
24 TR, Ml ;t BR, CP, CA. IP, SA, RA, CH
CS ;t BR, CP, IP
22 The geologic belts and terranes are described • In table 2 and shown in figure 8
20
• •
18 MEAN 18.24
gallons per minute •
16
AVERAGE YIELD•25.23 -0.900 (BELn
14 Analysis based on data from 5,498 wells
•
12 • •
MU BR CP CA IP SA GH RA CH KM cs Ml TR SR
10 1 2 3 4 5 6 7 8 9 10 11 12 13 14
GEOLOGIC BELT OR TERRANE
! Figure 14. Average yield of wells of average construction in the geologic belts and terranes of the Piedmont and Blue Ridge provinces of North Carolina.
"' i
about 11 to 12 gal/min to about 24 to 25 gal/min. about a
twofold increase. Considering topography alone, the
average well on hills and ridges can be expected to average
less than 12 gal/min, whereas wells in draws and valleys
can be expected to average about 29 gaVmin, an increase of
2.4 times. When the factors of hydrogeologic unit or
geologic belt are considered in combination with topo-
graphic setting, the range in yields is even greater. Wells in
draws and valleys in the most productive units average five
times more yield than wells on hills and ridges in the least
productive units.
The statistical analysis supported some concepts and
criteria for well-site selection, such as the siting of a well
with regard to topography. More importantly, however, the
analysis indicates that some previously held concepts may
be in error. First and foremost is the generally held concept
that the crystalline rocks yield only small amounts of water
to wells. The analysis showed that this concept may be due
to cultural bias. Most wells drilled in these rocks are small
diameter. are located primarily on hills and ridges-the
poorest possible sites for wells-and are drilled only to
depths where sufficient water for a domestic supply is
obtained. In the same theme, well diameter has not been
considered to have much effect on yield-a large-diameter
well was considered a storage tank. Statistical analysis
shows, however. that for a given depth the yield of a well
is directly proportional to the well diameter. The larger the
diameter the greater the yield.
Well construction in crystalline rocks has long been
based on the concept of a well intersecting near vertical
open fractures and joints that, because of lithostatic pres-
sure, pinch out at depths of about 300 ft. As a result, the
drilling of many wells has been arbilrarily stopped when the
depth of 300 ft was reached. The average well, whether
domestic or commercial-industrial. is not even that deep.
The analysis indicates that very few wells have been drilled
deep enough to test the full potential of the sites. For
example, the average yield of 6-in. diameter wells located
in draws or valleys reaches a maximum of about 45 gaVmin
at depths of 500 to 525 ft; the average yield of 12-in.
diameter wells located in draws or valleys reaches a
maximum of about 150 gal/min at depths of 700 to 800 ft.
Whatever the hydrogeologic unit or topographic loca-
tion, the chances of obtaining high yields are enhanced by
increasing the depth and diameter of the well to a much
greater extent than previously thought.
REFERENCES
Bain, G.L., 1966, Geology und ground-water resources of the
Durham area, Nonh Carolina: Nonh Carolina Department of
Water Resources Ground-Water Buhetin 7, 147 p.
Brown. P.M .. and Pwlcer, J.M., Ill, compiler.;, 1985, Geologic
map of North Carolina: North Carolina Department of Natu-
ral Resources and Community Development. 1 sheet. scale
1:500,000.
Burt, E.R .. 1981. Geologic map of Region H. North Carolina:
North Carolina Geological Swvey. Open-File Map no.
NCGS 81-4, scale 1:125,000.
Butler. J.R., and Ragland. P.C .. 1969, A petrochemical survey of
plutonic intrusions in the Piedmont, southeastern Appala•
chians. U.S.A.: Contributions to Mineralogy and Petrology.
v. 24, p. 164-190.
Carpenter, P.A., Ill, 1982, Geologic map of Region G, Nonh
Carolina: Nonh carolina Department of Natural Resources
and Community Development, Geological Survey Section.
Regional Geology Series 2, scale 1:125,000.
Cedarstrom. D.J., 1972, Evaluation of yields of wells in consol-
idated rocks, Virginia to Maine: U.S. Geological Survey
Water-Supply Paper 2021. 38 p.
Conrad. S.G .. Carpenter, P.A., Ill. and Wilson, W.F .. 1975,
Physiography, geology and mineral resources. in Clay, J. W .,
Orr, D.M .. and Sruan. A.W., eds .. North Carolina atlas,
portrait of a changing southern State: Chapel Hill. North
Carolina. University of North Carolina Press, p. 112-127.
Cressler. C.W., Thurmond. C.J., and Hester. W.G .. 1983.
Ground water in the greater Atlanta region: Georgia Depart•
ment of Natural Resources Environmental Protection Divi-
sion, Georgia Geologic Survey Information Circular 63.
144 p.
Daniel. C.C., III. and Sharpless, N.B., 1983. Ground-water
supply potential and procedures for well-site selection in the
upper Cape Fear River basin. North Carolina: North Carolina
Department of Natural Resource$ and Community Develop-
ment, 73 p.
Dodson, C.L.. and Laney. R.L., 1968. Geology and ground-water
resources of the Murphy area, Nonh Carolina: North Carolina
Department of Water and Air Resources Ground-Water
Bulletin 13. I 13 p.
Eder. B.K .. Davis. J.M .. and Robinson, P.J., 1983. Variations in
momhly precipitation over North Carolina: University of
North Carolina Water Resources Research Institute Report
no. 185. 50 p.
fapenshadc, G.H., Rankin. D.W .. Shaw, K.W .. and Neuman,
R.B., 1975. Geologic map of the east half of the Winston-
Salem quadrangle. Nonh Carolina-Virginia: U.S. Geological
Survey Miscellaneous Investigations Series Map 1-709-B.
scale 1:250,000.
Fenneman. N.M., 1938, Physiography of Eastern United States:
New York, McGraw-Hill, 714 p.
Floyd. E.O., 1965. Geology and ground-water resources of the
Monroe area. North Carolina: North Carolina Department of
Water Resources Ground-Water Bulletin 5, 109 p.
Fullagar. P.O., 1971. Age and origin of plutonic intru5ions in the
Piedmont of the 50Utheastcm Appalachians: Geological Soci-
ety of America Bulletin, v. 82. p. 2845-2862.
Goldsmith. Richard. Milton. D.J .. and Horton. J.W .. Jr .. 1982,
Simplified preliminary geologic map of the Charlotte IO x 2°
quadrangle, North Carolina and South Carolina: U.S. Geo-
logical Survey Open-File Report 82-56. scale I :250.000.
Hadley, J.B .. and Nelson. A.E., 1971. Geologic map of the
Knoxville quadrangle, North Carolina. Tennessee, and South
A26 Ground-Water Resources of the Piedmont-Blue Ridge Provinces of North Carolina
Carolina: U.S. Geological Survey Miscellaneous Investiga-
tions Series Map 1-654, scale 1:250,000.
Heath, M.S .. and Coffield, H.J .• Ill, 1970, Cases and materials on
ground water law-with particular reference to the law of
North Carolina: Chapel Hill, North Carolina, University of
North Carolina Instilute of Government, 72 p.
Heath, R.C., 1980. Basic elements of ground-water hydrology
with reference to conditions in North Carolina: U.S. Geolog-
ical Survey Water-Resources Investigations Open-File Report
80-44. 86 p.
LeGrand. H.E., 1954, Geology and ground water in the Statesville
area. Nonh Carolina: North Carolina Department of Conser-
vation and Development Bulletin 68, 68 p.
--1967, Ground water of the Piedmont and Blue Ridge
provinces in the southeastern States: U.S. Geological Survey
Circular 538. 11 p.
LeGrand, H,E., and Mundorlf, M.J .. 1952. Geology and ground
water in the Charlotte area. North Carolina: North Carolina
Department of Conservation and Development Bulletin 63,
88 p,
Marsh, Q.T .. and Laney, R.L., 1966, Reconnaissance of the
ground-water resources in the Waynesville area. North Caro-
lina: North Carolina Department of Water Resources Ground-
Water Bulletin 8. 131 p.
May, V.J., and Thomas, J.D .. 1968. Geology and ground-water
resources in the Raleigh area, North Carolina: North Carolina
Department of Wat~r and Air Resources Ground-Water
Bulletin 15, 135 p.
McDaniel, R.D .. 1980, Geologic map of Region K, Nonh
Carolina: North Carolina Geological Survey. Open-File Map
no. NCGS 80-2. scale 1:125,000.
Mundorlf. M.J ., I 946, Ground water in the Halifax area, Nonh
Carolina: North Carolina Department of Conser.·ation and
Development Bulletin 51, 76 p.
--1948, Geology and ground water in the Greensboro area.
North Carolina: North Carolina Department of Conservation
and Development Bulletin 55. 108 p.
Peace, R.R .. Jr., and Link, D.R., 1971, Geology and ground-
water resources of northwestern North Carolina: North Caro•
lina Department of Water and Air Resources Ground-Water
Bulletin 19. 135 p.
Pusey, R.D., I 960. Geology and ground water in the Goldsboro
area, Nono Carolin,: Nonh Carolina Depanment of Water
Resources Ground-Water Bulletin 2. 77 p.
Ragland, P.C .. Hatcher, R.D .. Jr., and Whittington, David, 1983,
Juxtaposed Mesozoic diabase dike sets from the Carolinas: A
preliminary assessment: Geology, July 1983, v, 11, p.
394-399,
Rankin. D.W., Espenshade. G.H .. and Neuman, R.B .. 1972,
Geologic map of the west half of the Winston-Salem quad-
nmgle, North Carolina, Virginia. and Tennessee: U.S. Geo•
logical Survey Miscellaneous Investigations Series Map
1-709-A, scale 1:125,000.
Reinemund. J.A., 1955, Geology of the Deep River coal field,
Nonh Carolina: U.S. Geological Survey Professional Paper
no. 246. 159 p.
SAS Institute Inc., 1982a. SAS user's guide: basics. 1982 edition:
Cary, North Carolina. SAS Institute, Inc., 923 p.
--1982b, SAS user's guide: statistics. 1982 edition: Cary.
Nonh Carolina, SAS Institute. Inc .. 584 p.
Schipf, R.G., 1961, Geology and ground-water resources of the
Fayetteville area: Nonh Carolina Department of Water
Resources Ground-Water Bulletin 3, 99 p,
Seiders, V,M .. 1981, Geologic map of the Asheboro, Nonh
Carolina, and adjacent areas: U.S. Geological Survey Mis•
cellancous Investigations Series Map 1-1314, scale I :62,500,
Snipes. D.S .. Padgett, G.S .. Hughes, W.S., and Springston.
G.E., 1983, Ground•water quantity and quality in fracture
zones in Abbeville County. South Carolina: Clemson. South
Carolina, Clemson University Water Resources Research
Institute Technical Repon no. l02, 54 p.
Steel. R.G.D .. and Torrie, J,H .. 1960. Principles and procedures
of statistics, with special reference to the biological sciences:
New York, McGraw-Hill, 481 p.
Stromquist. A.A .. Choquette. P.W .. and Sundelius, H.W .. 1971.
Geologic map of the Denton quadrangle, central Nonh
Carolina: U.S. Geological Survey Geologic Quadrangle Map
GQ-872, scale I :62,500.
Stromquist, A.A .. and Sundelius, H.W .. 1975. Interpretive geo-
logic map of the bedrock showing radioactivity, and aero-
magnetic map of the Salisbury, Southmont, Rockwell. and
Gold Hill quadrangles, Rowan and Davidson Counties. North
Carolina: U.S. Geological Survey Miscellaneous Investiga-
tions Series Map 1---888 (sheet 1 of 2). scale I :48,000.
Stuckey. J.L .. 1958, Geologic map of Nonh Carolina: Raleigh,
North Carolina. North Carolina Department of Conservation
and Development, Division of Mineral Resources. scale
1:500,000.
Sumsion, C.T .. and Laney, R.L .. 1967, Geology and ground-
water resources of the Morganton area: North Carolina
Department of Water Resources Ground-Water Bulletin 12,
119 p,
Trapp. Henry, Jr .. 1970, Geology and ground-water resources of
the Asheville area, North Carolina: North Carolina Depart-
ment of Water and Air Resources Ground-Water Bulletin 16,
127 p.
Weigand. P.W., and Ragland, P.C., 1970, Geochemistry of
Mesozoic dolerite dikes from eastern North America: Contri-
butions to Minerology and Petrology, v. 29, p. 195-214.
Wilson. W.F .. 1979. Geology of Wilson County, Nonh Carclina:
Nonh Carolina Depanment of Natural Resources and Com-
munity Development. Division of Land Resources. Geolog-
ical Survey Section, Open-File Map NCGS 79-2, scale
1:125,000.
--1981, Geology of Halifax County, Nonh Carolina: Nonh
Carolina Geological Survey, Open-File Map NCGS 81-3,
scale 1:125,000.
Wilson. W.F .. Carpenter, P.A .. Ill, and Parker. J.M., III. 1981,
Geologic map of Region J, North Carolina, in A guide for
North Caro1ina mineral resource development and land use
planning: North Carolina Department of Natural Resources
and Community Development. Geological Survey Section,
Regional Geology Series I, scale I: 125,000.
Wilson. W.F .. and Spence, W.H., 1979, Geology of Nash
County. North Carolina: North Carolina Department of Nat•
ural Resources and Community Development, Division of
Land Resources, Geologica1 Survey Section, Open-File Map
NCGS 79-3, scale 1:125,000.
Analysis Relating Well Yield to ConstNction and Siting. Piedmont-Blue Ridge Provinces, North Carolina A27
APPENDIXC
II
PROJECT NAME YAN I!!UCK It !!US DATE STARTED 3/6/91 COMPLETED 3/7/91 PAOE...LOF_II_
JOB NO. 36 COUNT¥ ROWAN CITY/TOWN CU:Vli:LAIID · STATE NC
BORING/WELL lllf-1 TOTAL DEPTH 20,0 BORING LOCATION UP OIWllENT ELEV. 1nu, FT
GEOLOGIST /ENGINEER BRIAN TEMPLE MONITORING EQUIPMENT USED OVA
DRILLERS WELL DRIU.EB~ INC DRIWNG EQUIPMENT 8 112 RIO METHOD UOb!,Olt g)'.f;!! AUG!;I! .
REMARKS: ALL MEASUREIIENTS ARE DELOY GROUND SURFACE . . .
DEPTH TO BOTTOM OF SCREEN 12;0 FT TOP OF SCREEN 2.0 FT SCREEN SLOT SIZE .01 IN
DEPTH TO BOTTOM OF SAND . 20.9 n TOP OF SAND UI E1 DEPTH TO TOP OF BENTONITE 119 [[
WELL MATERIAL USED PVC DIAMETER 2.00 !!! STICK UP 2.0[!
WATER ZONES (DEPTH) B.O l'I' STATIC WATER LEVEL ft.2& l'I'
REMARKS: ALL MEASUREMENTS ARE BELOW GROUND SURFACE
DEPTH SAMPLE BLOW % OVA
LITHOLOGY (FT) NO. COU T REC. SAMPLE DESCRIPTION READING
6" 6" 6'. (PPM)
OIIASS FIRST 2 CII
SUOIITLY IIICACEOUS
-5 I I I 3 t.B7F'I IIIIITE, DARK GRBEII, BRO'IN 7 1101ST, CLAYEY, F1NE BANDY, SILT
(6-Mn)
WATER Ill 8 FT
-10 I 2 3 6 1.10 n IIIIITE, DARK GREEII, BRO'IN
1101ST, CLAYEY, F111E/IIEDIUII SAIIDY ,3
SILT
(to-11.11 rr)
-15 2 4 8 2.20n IIETA-DIORITE 3 . l 1101ST, CLAnY, COARSI SANDY, SILT
(16-18.6 FT)
~ 20 BOIUNO TERMINATED O 20,0 FT
-25 IIOTE, ALL S00.S IOOIIB1111D IGH&:OUS
TEXTURES AT Dl!!PTIIS Bl:LOY 4,0 FT
'-. 30
-35
'-40
-45
SAMPLES SUBMmED FOR LABORATORY TESTING .QUATERRA, INC. I
2 (HONE) . ALEJOH,OREENSBORO,CHARLOTTE
3 NORffi CAROIJNA
APPENDIXD
Classification of Transmissivity
Magnitude and Variation
by Jiri Knlsny"
Abstract
Until now no objective claallcation of transmis.,ivity has bun introduced, in spite of the quantitative nature of
transmissivity values and their ohl'ious importance for quantitative appraisals of aquifen or major ground-water systems.
The usual subjective expres.sion of transmls.sivity, for example as "high" or "low," prevents the objective comparison of
transmlssivlty values characterizing different areas and hydrogeological environments. A combined da'5iflcation of magni-
tude and variation of transmissivity is proposed, with the intention to standardize the expression, comparison, and
representation oftrall'lmlsslvlty. This classification also enables Its compact and unambisuous depiction in tables and maps.
Introduction
Transmissivity is an important hydraulic property of
aquifers and water-bearing materials. In common with
permeability, transmissivity affords a notion about the
water-bearing characteristics of hydrogeological bodies.
Transmissivity values enable us to estimate the possibility of
ground-water abstraction, in the first approximation.
Therefore, knowledge of transmissivity distribution helps us
to draw important conclusions from hydrogeological stu-
dies, and for this reason, prevailing transmissivity values arc
often represented in hydrogeological maps. They provide a
basis for future ground-water exploration, development,
abstraction, and protection.
Yet, in spite of the quantitative nature of transmissivity
and its importance for quantitative appraisals, no objective
classification of transmissivity has been introduoed. Quan-
titative or semiquantitative terms describing transmi~ivity
are often used, denominating different grades or classes as
large, small, etc., but without strictly stating limils between
them. This is the current case with hydrogeological maps,
where the inexactly defined term "productivity" is some-
times used (cf. e.g., !AH et al., 1983). Even iftransmissivityis
exprtSSCd numerically, the verbal designation of numerical
classes might involuntarily reflect the relation between
hydrogeological conditions and water demand; in areas
where yields of water wells are sufficient to cover limited
water consumption, transmissivity may be designated as
high; on the other hand, where well yields do not suffice for
• Department of Hydrogealo~ and Engineering Geology,
Faculty of Science, Charles University, Prague, Czech Republic,
Received January 19921 revised July 1992, accepted Sep-
tember 1992.
Discussion open until September I, 1993.
230
large requirements, transmissivity might be designated as
low. Such a subjective approach prevents the objective
comparison of transmissivity at both local and regional
scales, including values represented on hydrogeological
maps.
Just as with transmissivity magnitude, transmissivity
variation affords important information on hydrogeological
properties. In spite ofits usefulness, however, this character-
istic is very seldom used in hydrogeological studies, and no
objective basis for its comparison has been defined.
Therefore, a combined classification of transmissivity
magnitude and variation is proposed with the intention to
standardize transmissivity expression, representation, and
comparison in regional and local hydrogeological studies.
This classification aiso enables its compact and unambigu-
ous depiction in tables and r,naps.
Background
The statistical distribution and prevailing values of
permeability and transmissivity strongly depend on the rela-
tion of the magnitude of elements of rocks heterogeneity to
the extent of the studied area (Rats, 1967). The idea was
hydrogeologically interpreted by Kiraly (1975). Conse-
quently, because of this "scale effect," representative values
of hydraulic parameters are substantially influenced by
methods used for their determination. Pumping tests arc the
most frequent procedure used to determine transmissivity
values that are often available as extensive data populations
that may be used to characterize a hydrogeological envi-
ronment. Wells are also the most common device for
ground-water abstraction in general. Many times, however,
various less reliable historical data are available in arehives.
These data are not suitable for determining exact hydraulic
parameters such as the transmissivity and hydraulic conduc-
tivity; on the other hand, many of these data are good
enough to contribute to a general idea about the regional
Vol. 31, No. 2-GROUNO WATER-March-April 1993
Table I. Classllleatlon of Transmlsslvlty Magnitude
Comparati~ regional
pararntttrs approximaltly
corresponding to the coefficient
of transmissivlly
Nonlogarithmic
Coeflic~nt of Class of Designation of -----
tr01Umissivl1y tran.rmlsslvily transmisslvity ~ciflc capacity
(m 2/d) magni1ude magnitude q in I/rm
I Very high
-1,000 JO
II High
100
Ill lntenncdiate
JO 0.1
IV Low
0.01
V Very low
.0.1 0.001
VI Imperceptible
distribution of transmissivity and/ or permeability values
and could be treated statistically. Therefore, a "category" of
comparative regional parameters was introduced (Jetel,
1964; Jetel and Krasny, 1968), as described below. This gives
us a good chance of drawing regional conclusions, especially
when using simple statistical procedures for data analysis.
The classification of transmissivity magnitude and variation
is proposed with the intention of expressing, representing,
and comparing available transmissivity data in a more
objective manner.
Transmlsslvlty Magnitude
This classification of transmissivity magnitude was
published by Krasny ( 1970) and has been modified several
times afterwards (e.g. Knlsny, 1986c) to reach the present
form. On the basis of transmissivity studies both in
Czechoslovakia and abroad, the range of values generally
found is logarithmically divided into six classes from very
high transmissivity (/ class: coefficient of transmissivity T
more than 1,000 m2/d) to imperceptible transmissivity ( VI
class: T less than 0.1 m2/d) as shown in Table I. In the table,
the range of different classes is expressed in SI units and in
approximately corresponding comparative regional parame-
ters (i.e., specific capacity q and index of transmissivity Y) as
well. In Figure I, English units are also shown for compari-
son. The divisions between classes in English units follow
Very approximate
Logarithmic expected discharge
----in 1/s ofa
Index Ground-water single well at
y supply potential 5 mdrawdown
Withdrawals of great
regional importance >SO
7.0
Withdrawals of lesser
regional importance 5 -50
6.0
Withdrawals for local
water supply (small
communities, plants,
etc.) 0.5 -5
5.0
Smaller withdrawals
for local water
supply (private
consumption, etc.) 0.05 -0.5
4.0
Withdrawals for local
water supply with
limited consumption 0.005-0.05
3.0
Sources for locaJ
water supply arc
difficult (if possible)
to ensure <0.005
more or less the orders of magnitude of SI units only by
chance, as for example, in the case of transmissivity
expressed in m2/d and ft2/d.
The index of transmissivity, Y, has been introduced as a
comparative regional parameter, a logarithmic tramforma-
tion of specific capacity (Jetel and Kr6sny, 1968). It has been
used advantageously since the distributions of the majority
of sample populations of transmissivity values are log-
arithmic-normal; when using a logarithmic parameter (i.e.,
the index Y, for example), the statistical distribution
changes to a simpler normal one. To calculate the index
Y =log(l06 q), specific capacity q has to be expressed in 1/s
m.
Approximate estimates of probable ground-water
yield to wells are appended to all classes in Table I. Thus,
results of individual wells can be evaluated in the first
approximation, but, more significantly, extended areas may
be assessed in a preliminary way aocording to prevailing
transmissivity values with respect to this important hydro-
geological issue (Table 1).
Many hydrogcological media are characterized by so-
called chaotic heterogeneity; these are media where perme-
ability (transmissivity) distribution is a random function
without any apparent regional tendency, cf. Borevskiy ct al.
( 1979). Where there are sufficient available data to be pro-
cessed statistically, sample populations delimited by differ-
231
ent rock types, areas, hydrogeological positions, etc. can be
treated to determine the arithmetic mean x and the standard
deviation of sample s of each of them. As the factor by which
to determine the class of transmissivity magnitude for each
population, the interval x ± s was chosen; if 70 or more
percent of the interval belongs to one class, prevailing
transmissivity is designated by the name or symbol of this
class; if 30 to 70 percent of that interval belongs to either of
the two (exceptionally three) adjacent classes, names (sym-
bols) of both (all) classes are concentrated to designate the
prevailing transmissivity, in order of the magnitude of their
participation. In the case where 30 to 10 percent belongs to a
class, its name (symbol) is enclosed in parentheses. No name
(symbol) is used with a percentage less than 10. Some exam-
ples of the classification procedures are given later. Trans-
missivity values expressed by index of transmissivity Y or by
coefficient of transmissivity or specific capacity, the last two
in a logarithmic fonn, should be used when treating the
samples statistically and determining the interval i ± s.
Examples of graphical statistical treatment of different
sample populations of transmissivity values are shown in
Figure 1, where probability paper for plotting cumulative
relative frequencies of samples was used. Similar representa-
tion was used by Walton (1962, in Davis and De Wiest, 1966).
More about the application of probability paper can be
found in Spiegel (1972).
Values of transmissivity outside the interval x ± s are
considered anomalies, positive and negative (Figure I).
Both are of practical importance: the positive anomalies (the
interval between i + s and x + 2s, designated as + A in
Figure I) indicate the zones of better possibilities for
ground-water abstraction compared with the area of prevail-
ing transmissivity (hydrogeological background). The nega-
tive ones(values betweeni -sand x-2s, i.e. -A)showthc
zones to be avoided for water-supply purposes or where, on
the other hand, there is generally less danger of ground-
water contamination, and where, consequently, location of
landfills may be considered. The extreme anomalies, posi-
tive(++ A) and negative ones (--A) can be found outside
the interval i ± 2s (Figure I).
In cases where data cannot be treated statistically, the
class of transmissivity magnitude is estimated directly on the
basis of prevailing values.
Variation In Transmlaslvlty
No homogeneous environment exists under natural
conditions; all environments are heterogeneous to a differ~
ent extent often depending on the extent of the studied area
CLASSES OF TRANSMISSIVITY MAGNITUDE
VERY LOW LOW INTERMEOIATE HIGH VERY HIGH
++A V IV 111 II ¼
~ ... ,.,.,
" +A ..
--){+S
"' "' .,
+I ., ll
IX " ..
\t~
+a+. .. l!:
~ X/X "
701 ~ ~ e • • "' " N ...,
'5 00 ,,;
<D O> 1! ,0 8 I ~ ~ -'-X-s IS
s ' 10
-A i ±s s
X-25 -• I
ll 13
Index y
--A :i: 15 I .0 I ,s 5.0 ss 6.0 65 7.0 7.S 8.
00!)1~ .. ,m, ,nb;-~· I ,0/';I.~, ~°'' I -~·r1t bi I I i'J';,o (~ I 1~•11.lf(l/sml q I I 1 I i911 I 1 , hii , I I ill I I (gpd/ft l
1 tO SO :200 SOQ1 2 1'-2 15z S.
01 02 , : 0 Ill 1 ' 1 1 (m21dl
, "' s 1 !frldl T
Fl&, t. Cumulative relative frequencies of different samples of transmlsshrity values and their classification according to transmlsslvlty
m11111ltude and variation, wbere q = 1pecillc capacity In 1/s m and in IJICl/ft; T = coefficient of transmlssivlly in m'/d and in lt'/d;
i = arithmetic mean; s::::: standard deviation; and ++A1 +A, -A, --A~ fields of positive and neaative anomalies (extreme
anomalies), respectively, and lndlvldual anomalous water wells (outdde tbe lntenal i ± s or prevailing tnnsmissivlty values =
hydrogeoloalcal background).
232
Table 2. Claulflcatlon or TrammlssMty Variation
Standard deviation
of rransmwlvily
index y•
Class of .O.signation of Hydrogeological environment from
the point of view of its
hydraulic heterogeneity 0
transmissivity trarumissiviry
wuiation variation
a Insignificant Homogeneous --0.2---------------------------------
b Small Slightly heterogeneous
--0.4---------------------------------
C Moderate Fairly heterogeneous --0.6---------------------------------
d Large Considerably heterogeneous --0.8-------------------------------
e Vcry heterogeneous Very large --1.0------------------'-'----------------
f Extremely large Extremely heterogeneous
• Or logarithmic transformation of any parameter cxpre.<ising transml~ivity.
•• Usable especially for permeability evaluation but also when evaluating transmi~ivity.
and the method used for hydraulic parameters assessment
(scale effect).
Variation in transmissivity can be expressed by using
the standard deviation of the sample population of trans-
missivity values. Six classes denominated a to/ are distin-
guished, with all classes having the same range: 0.2 of the
standard deviation of the logarithm of transmissivity. These
classes are based on the results ofstudies in different hydro-
geological media (Table 2). It is well-known that insignifi-
cant or small variation (classes a, b) is typical of samples
representing media having intergranular porosity ( e.g., well-
sorted fluvial deposits) while large or very large variation of
transmissivity (classes d, e) are characteristic of media with
dominant fissure porosity. Some examples are shown in
Figure 1: the steeper the slope of the line representing the
sample, the smaller the transrnissivity variation of the
sample.
Moreover, knowledge of transmissivity variation and
of the interval i ± s enables us to express a range of probable
transmissivity values and/or intervals representing anoma-
lies, and makes possible the prediction-on the basis of
previous data--0f future water well results, assuming that
wells are drilled in the same area under the same conditions
as those used for the statistical analysis. Specific intervals of
Y, q, or T values can be predicted on the respective axes, as
shown in Figure I. The smaller the standard deviation (i.e.,
the narrower the interval i ± s), the more reliable is the
prediction.
The classification of variation (Table 2) may, of course,
be used for the evaluation of permeability (hydraulic con-
ductivity) as well, if this parameter is expressed in a loga-
rithmic form. Variation of permeability and transmissivity
expresses a character of hydrogeological environment from
the point of view of its hydraulic heterogeneity. In the case of
transmissivity it can also reflect changes in aquifer thickness.
Cl888lllcatlon of Transmlsslvlty
The classification of transmissivity magnitude ( six
classes/-VI) combined with the classification of transmissiv-
ity variation (six classes a-/) can be used for a quick and very
simple quantitative characterization of different hydrogeo-
logical environments and for their comparison. Jn addition,
knowledge of transmissivity variation and of the interval
i ± s enables us to predict future well yields and to indicate
the hydraulic character of a hydrogeological environment.
For example, an environment classified as la, i.e., very high
transmissivity with insignificant variation, represents the
optimum environment for ground-water development. Of
course, this characteristic does not express natural ( or other)
ground-water recharge. On the other hand, classification
Via (imperceptible transmissivity with its insignificant varia-
tion) indicates an environment almost without promise of
ground-water abstraction possibilities, while Vie or Vlf
(imperceptible transmissivity with very or extremely large
variation) leaves some hope of discovering more permeable
zones or areas where transmissivity may differ significantly
from the mean values due to the extreme heterogeneity of
the hydrogeological environment.
The classification of transmissivity can be used in
hydrogeological maps as reported in Krasny and L6pez
(1989). It was used in the 1:500,000 hydrogeological map of
the Czech republic and in a slightly modified version also in
the new edition of hydrogeological maps of Czech territory
at 1 :50,000 scale. In maps and in hydrogeological studies,
three degrees may be distinguished of the density and reli-
ability of available data:
I. In areas with sufficient data to process statistically,
the interval i ± s of prevailing transmissivity values (hydro-
geological background) can be estimated; then both trans-
missivity magnitude and variation can be expressed accord-
ing to classifications in Tables I and 2, as mentioned above.
2. In areas where prevailing transmissivity is estimated
on the basis of less numerous data so that no statistical
treatment can be undertaken, only the class of transmissivity .,
magnitude may be assessed and expressed, for example by
symbols, as follows: /, Ill-IV. etc. I
3. In areas with few or no transmissivity data, the
prevailing transmissivity is assessed by analogy or by con-
sidering the geological environment (especially lithology
and structure) of the area and expressed by a symbol in
brackets, e.g. {VJ, [I-II}.
233
Table 3. Basic Data for the Examples
Example:
Sample:
Geological
environment:
Values of
the index Y:
n:
x:
s:
Index Y: i±s
T (m2/d):
Classes of
transmissivity
magnitude and
variation
of samples:
I
A
Crystal/IN
rocks (esp.
gnt!isS)
3.32
3.43
3.76
3.80
3.82
4.37
4.40
4.43
4.56
4.78
4.85
S.15
S.18
13
4.30
0.62
3.68-4.92
0.45-7.9
IV(-V)d
]
B
Cretaceow
sandstones
4.05
4.72
5.27
5.32
5.38
S.48
5.17
S.89
5.93
6.01
6.08
6.15
6.30
6.38
6.43
6.47
6.51
6.54
6.65
19
5.86
0.68
5.18~.54
14-330
lll-lld
3
C
Quaternary
fluvial
deposits
6.39
6.51
6.53
6.76
6.80
6.82
6.83
6.84
6.90
6.90
6.92
7.05
12
6.77
0.19
6.58~.96
J(i()-860
Ila
n = sample size; i = arithmetic mean of a sample; s = standard
deviation of the sample; T = coefficient of transmissivity. All the
values of the index of transmissivity Y (and accordingly those of
the specific capacity, from which indices Y were derived) were
determined as results of pumpina tests from water wells.
Examples of TransmlBBlvtty ClaBBlllcatlon
As examples of the clll&'lification procedure, and also to
demonstrate actual differences in transrnissivity magnitude
and variation, three real statistical samples oftransmissivity
values from different hydrogeological environments within
the Bohemian Massif in Czechoslovakia are presented. The
basic data for the examples are given in Table 3. Forsimplic-
ity, samples are characterized directly by the values of the
index of transmissivity Y, i.e., by a logarithmic modification
of specific capacity q, as stated above. The approximate
relation between the index of transrnissivity Y, specific
capacity q, and coefficient oftransmissivity T follows from
Table I or Figure I.
Example I. In an area formed by crystalline rocks
(especially gneiss) in southern Bohernia(sample A in Figure
I), the available values of index of transmissivity Y (fable 3)
give the arithmetic mean x as 4.30, and the sample standard
deviation s as 0.62. The classification interval i ± s covers
the range 3.68-4.92, of which 74.2 percent belongs to the
234
class /Vof transmissivity magnitude and 25.8 percent to the
class V. According to its standard deviation, the sample
belongs to the class d of transmissivity variation. Then the
designation of the sample is IV (-VJ d. i.e. low (to very low)
transmissivity with large variation.
The approximate conversion of the index Y values to
the respective transmissivitycoefficient T may be performed
using the equation T(in m2/d)= tov-u, X86,400(Jetel and
Krllsny, 1968). Then the arithmetic mean of sample
expressed as transmissivity Tis 1.89 m2/d, and the interval x
± s lies between 0.45 and 7.9 m2/d.
Example 2. A sample of Y values from sandstones of
the Cretaceous basin in northern Bohemia (sample B in
Table 3 and Figure I) has the arithmetic mean 5.86 ( of index
Y), or expressed as T = 69 m'/d. Standard deviation of
samples =0.68; 60.3 percent of the interval x± s (5.18-6.54)
belongs to the class Ill, 39. 7 percent to the class fl Accord·
ing to the cla.~ification the sample may be designated a."
Ill-lid, i.e., intennediate to high transmissivity with large
variation. Approximately 68 percent of the sample (interval x ± s) can be expected to be between transmissivity values I 4
and 330 m2/d.
Example 3. A sample ofY values from fluvial deposits
of the River Labe in central Bohemia (sample C in Table 3
and Figure I) has an arithmetic mean x = 6. 77 (T = 650
m2/d), standard deviations =0.19, interval x ± s = 6.58-0.96
(f = 360-860 m2/d). One hundred percent of the interval
belongs to the class II; according to the standard deviation,
the class of transmissivity variation is a. The sample can be
classified as //a, i.e., high transrnissivity with insignificant
variation.
Closing DlscuBBlon and Conclusions
The proposed classification of transmissivity magni-
tude and variation aims to provide a basis for a quantitative
and objective expression and representation of prevailing
transmissivity in hydrogeological maps and studies.
Expressing transmissivity classes in a compact form enables
a quick and objective comparison of different areas and/or
hydrogeological environments in local and regional studies.
Hydrogeological background and anomalies (positive and
negative ones) can be determined if statistical treatment of
samples is possible. Future water well results can then be
antiticipatcd on the basis of a statistical approach. This
methodology, of course, cannot replace detailed and com•
plex mathematical models. It can, however, help to formu-
late general regularities and to prepare a way for definition
of prevailing natural conditions. It may be considered as the
fiist step on the way towards hydrogcological data quanti•
fication.
Some interesting conclusions have been drawn from
various studies that have applied the methods described in
this paper:
• Similarities or, on the contrary, significant differ-
ences in prevailing transmissivity between distinct types of
rocks or of whole hydrogeological basins [Carlsson and
Carlstcdt, 1977; Krasny, 1975, 1976, 1986a; Krasny and
L6pez, 1989; Michlicek, 1982; and many explanatory notes
to hydrogeological map of Czechoslovakia I : 200,000, e.g.,
Hazdrova (ed.), 1983; Jetel (ed.}, 1986; Krasny (ed.), 1982].
• Differences in transmissivity depending on hydro-
geological position of wells ( areas of recharge or discharge,
Kdsny, 1974, 1984).
• Vertical-depth permeability changes in different
types of rocks (Krasny, 1975, 1986b).
• On the basis of mean (prevailing) transmissivity
values, a new method for ground-water runoff estimation in
fissured "hard" rocks was proposed (Krasny and Knezek,
19n; Knezek and Krasny, 1990).
A summary of some of the above-mentioned conclu-
sions was published recently (K.rasny, 1990).
After the first step of evaluating the spatial distribution
of transmissivity and determining an arithmetic mean and a
standard deviation, a redistribution of values among differ-
ent statistical samples is sometimes useful; thus, more
homogeneous samples can be produced, characterized by a
narrower interval of the hydrogeological background x ± s
and a less ambiguous classification oftransmissivity magni-
tude with less transmissivity variation. Anomalies indicating
zones with relatively greater or smaller transmissivity (per-
meability) may give hints: revelation of buried alluvial fans
may serve as an example (Krasny, 1983). In similar cases,
hydrogeological conclusions may even influence geological
conclusions.
To maintain one of the principal advantages of the
proposed classification, i.e., quick and easy data analysis,
simplifying assumptions were accepted which usually do not
affect results but which, however, should be taken into
account. Some of them are as follows:
l. Transmissivity (permeability) is considered to be a
random function of a natural hydrogeological environment
Therefore, the samples of data from wells are considered to
be random samples from an infinite population. This is not
the case for many available data samples, as many times
water wells are purposefully located in more promising sites.
As a consequence, the statistical distribution of a sample
may be biased toward greater values; that is, the distribution
is skewed more to the left when represented by a histogram.
More realistic statistical models can be obtained when data
are pretreated so that a selection of data from more perme-
able zones is made by choosing only selected values as
representative, or by calculation of a weighted arithmetic
mean.
2. Unsuccessful water wells naturally should belong to
a sample; however, usually neither specific capacity nor
yield arc stated in archive reports. If the number of unsuc-
cessful wells is known, an estimation of basic statistical
characteristics is possible by graphical or numerical approx-
imation of unknown value., by using the presumed statisti-
cal, i.e. lognormal model distribution [Krasny (ed.), 1982].
In case of an unknown number offailures, no remedy exists
and as a result, we have to consider the possibility that a
sample where unsuccessful wells were probably omitted
may overestimate the actual transmissivity.
3. Different depth of wells, naturally, may cause addi-
tional variation of transmissivity values. That is why a pre-
liminary selection should be made so as not to accept too
great differences in depth of wells within the same statistical
sample, particularly in hydrogeological environments where
thickness of aquifer may influence transmissivity changes.
Generally, however, relatively great differences in well
depths (e.g., tens of meters when mean depth is about 100
meters) cannot significantly influence variation in transmis-
sivity, as its horizontal changes caused by the variability
within a geological environment (lithology, fracturing) are
usually much more significant than those ones caused by
different depth of wells.
4. Diameter of most drilled water wells ranges from 0.2
to 0.4 m. These differences do not usually influence statisti-
cal characteristics significantly, as can be seen by examining
equations for steady.,.tate radial flow toward a pumped well.
Only large numbers of band-dug or other large-diameter
wells such as arc bored in flu vial deposits or in the weathered
zone of hard rocks may significantly influence the basic
transmissivity characteristics of a sample.
It is evident that these influences may cause some
inaccuracies in transmissivity data generalization and
regionalization. However, the author's experience during
previous studies suggests that they do not affect the principal
aim of the combined classification of transmissivity magni-
tude and variation presented here: to quantify and standard-
ize transmissivity expression, comparison, and representa-
tion in bydrogeological maps and studies. Within samples of
greater size, both inaccuracies and distinct influences are
usually balanoed, and even relatively small differences in
transmissivity magnitude and/or variation repeatedly found
in some areas or environments might have logical hydrogco-
logical grounds.
Transmissivity classification has usually been used in
regional hydrogeological studies, i.e. with most transmissiv-
ity values measured at the scale of pumping tests from water
wells and the space around them. Logarithmic-normal dis-
tribution of statistical samples of permeability and transmis-
sivity values prevail there. Transmissivity classification,
however, could be used for studies of areas of different size
(even under laboratory conditions) as well. In these cases,
other types of statistical distribution (models) and also dif-
ferent values could be expected because of the scale effect.
Therefore, it should be explicitly stated, when using the
classification, the extent of the studied area and conse-
quently the method of transmissivity (permeability) assess-
ment that was used.
References
Borevskiy, V. V., B. G. Samsonov, and L. S. Yazvin. 1979. Metod-
ika opredcleniya paramctrov vodonosnykh gorizontov po
dannym otkachek. Nedra, Moskva. 326 pp.
Carlsson, L and A. Carlstedl 1977. Estimation of transmissibility
and permeability in Swedish bedrock. Nordic Hydrology,
Copenhlll!Cn, v. 8, pp. 103-116.
Davis, S. N. and R.J.M. DcWicst. 1966. Hydrogcology. J. Wiley,
New York-London-Sydney. 463 pp.
Hazdrow, M. (ed.). 1983. Vysvctlivky k zllkladnl hydrogcologicke
mape CSSR I :200 000, list 12 Praha (in Czech: Explana-
tory notes to basic bydrogcological map of CSSR
I :200,000 shcel 12 Praha). tlsuednl uslav gcologicky,
Praha. 163 pp.
!AH, !ASH, and UNESCO. 1983. International legend for
235
hydroseological maps. Revised version. 5 I pp.
Jetcl, J. 1964. Pouzitl hodnot spccificke vydatnosti a novych
odvozenych parametru vhydrogeologii (in Czech: Applica•
tion of specific capacity values and new derived parameters
in hydrogcology). Oeol. Prutlc., Praha. v. 6, no, 5, pp,
144-145,
Jetel,J, (ed.), 1986. Vysvetlivky k zAldadnl hydrogeologickc mape
CSSR l: 200 000, list 03 Liberec, 04 NAchod (cAst) [in
Czech: Explanatory notes to basic hydrogcological map of
CSSR l: 200,000, sheet 03 Liberec, 04 NAchod (part)].
tlstrednl ustav gcologicky, Praha. 157 pp.
Jetel, J, and J. Knl.ffly. 1968, Approximative aquifer characteris-
tics in rogional hydrogeological study, Vest. Ustr. Ust, gcol.
Praha. v, 43, no, 5, pp, 459-461.
Kiraly, L. 1975, Rappon surl-actucldes connaissanccsdans le
domainc des caractUCS physiques des rochcs karstiqucs. In:
Hydrogeology oflCarstic Terrains, !AH Paris. pp, 53~7.
Knezek, M. and J. Krlasn)'. 1990. Natural groundwater resources
mapping in mountainous areas of the Bohemian Massif.
Mcm. 22nd Congress IAH. v, 2, Lausanne, pp. 846-854,
Krasny, J. 1970, Zpracovanl a plosnc vyj!.dn:nl nektcrych hydro-
gcologickych prvku v mapllch (in Czech: Elaboration and
areal expression of some hydrogeological clements in
maps). Shor. 5, hydrogcol. konfer., Zlln, pp, 57~.
Krisn)', J. 1974. Les differences de la transmissivite, statistiquc-
ment signilicatives, dans les zones d1nfihration et du drain-
age, Mcm. AIH 10, I. Communic., Montpellier, pp,
204-211,
Kr6snS,, J. 1975. Variation in tre.nsmissivity of L-rystalline rocks in
southern Bohemia. Vest. tlstr, Ust. gcol. Praha. v, 50, no. 4,
pp. 207-215.
Krasny, J, 1976, Statisticka analS,,. hydrogcologickycb dat z
podkrkonossk6 permokarbonskc panve (in Czech: Statisti-
cal analysis of hydrogeological data from the Permo-
Carboniferous Krkonosc-picdmont basin). Sbor. gcol.
Vcd, Hydrogcol.-lnz, Geel. Praha. v, 13, pp, 113-1S2.
Krasny, J. 1983. Hydrogcology and hydrochemistry of the Foot-
hill Zone, MS 0.0. Geel. Survey and Miner, Investig,,
Baghdad.
Krasny, J, 1984. Vliv hydrogeologick6 po,ice homin na jejich
propustnost (in Czech: Influence of hydrogcological posi-
236
tion ol rocks on their permeability). Geel. Pruzk,, Praha,
v, 26, no, 12, pp, 342-34S.
Kr6sny, J. 1986a. RegionAlnf tcndence hodnot transmisivity v
jihomoravskcm fluviAlnlm kvan~ru (in Czech: Regional
tendencies of transmissivity values in Southmoravian flu-
vial Quaternary deposits), Zpr, gcol. Vyzk, 1984, tlstr. ust.
gcol. Praha. pp. 121-123.
Krasny, J, 1986b. Charakter propustnosti a transmisivita flyso-
vych scdimentu jiznl casti Zd!nick6ho lesa (in Czech:
Character of permeability and transmissivity of flysh
deposits of the southcm pan of the Zdanicky les Mts.). Zpr,
geol. Vyzk. 1984, tlstr. ust. gcol. Praha. pp, 119-121.
Krjsny,J, 1986c. Klasifikaa: transmisivity ajcjl pou,.itf (in Czech:
Transmissivity classification and its application). Gcol.
Pruzk,, Praha. v, 28, DO, 6, pp, 177-179,
KrAsnY. J. 1990. Regionali7.ation of transmissivity data: hard
rocks of the Bohemian Massi!, Mem. 22nd Congress IAH,
v. l, Lausanne, pp, 98-105.
Krasny, J. (ed.), 1982. Vysvctlivky k zWadnl hydrogeologickc
mape CSSR I: 200 000, list 13 Hradec Kralovc (in Czech:
Explanatory notes to basic hydrogcologkal map ofCSSR
l: 200,000, sheet 13 Hradcc KrAlovc). Ustrednl ustav gco-
logicky, Praha, 159 pp.
Krlt.snj, J. and M. Kncz.cl::. 19TI. Regional estimate of ground-
water run-off from fissured rocks on using trtmsmissivity
coefficient and geomorphologic characteristics. J. Hydro!.
Sci., Warszawa. v, 4, no. 2, pp. 149-159.
Kr6sny, J. and A, L6pcz, 1989, The general hydrogeological map
of Nicaragua I: 250,000: Principles of compilation Wld
content. Mem. Internal. Symp. on Hydrogcot maps as
tools for econ. and social development, Hannover. pp.
413-422,
Michllcek, E. 1982. Statistick! analyza transmisivity homin vy-
chodni cAsti Cesk.omoravsk~ vrchoviny (in Czech: Statisti-
ca1 analysis of transmissivity of rocks in the eastern pan of
the Ccstomornvsk! vrchovina Mts.), Shor. gcol. Vcd,
Hydrogcol.-Inz. Gcol. Praha. v. 16, pp, 91-120.
Rats, M. V, 1967, Neodnorodnost gornykh pored i ikh fiziches•
kith svoystv, Nauka, Moskva.
Spiegel, M. R. 1972. Theory and Problems of Statistics in SI
Units. McGraw-Hill Book Co, 3S9 pp.