CE 626

W ELL H YDRAULICS II
Riddhi Singh Lecture 7
Email: [email protected]
1
Today we will learn about…

• Flow in leaky confined aquifer

• Flow in an unconfined aquifer
• Determining aquifer parameters for non-
equilibrium conditions
• Aquifers in non-ideal conditions

2
UNSTEADY STATE WELL HYDRAULICS:
LEAKY CONFINED AQUIFER

3
 Flow in a leaky confined aquifer Potentiometric Original potentiometric
surface surface

Q
 h 1 h S h q
2
+ = − b '' K ''
r 2
r r T t T
Leaky confining layer b ' K ',S'
Hydraulic head, h [m]
Storativity, S [-]
r h0
Transmissivity, T [m2/day] b
Time, t [days] h K, S
Radial distance from the pumping well, r [m]
Recharge to aquifer, q [m/day] Fully penetrating well in an aquifer overlain by a
semipermeable confining layer

h0 − h
q=K' Vertical hydraulic conductivity of the leaky layer
b' K’ [m/day]
 2 h 1 h S h K ' ( h0 − h ) Thickness of the leaky layer, b [m]
→ 2+ = − Drawdown, h0-h [m]
r r r T t Tb '
4
Case 1. Hantush’s (1956) solution for the case
Potentiometric Original potentiometric
when no water drains from the confining layer surface surface
Assumptions:
1. The aquifer is confined on the top by an Q
aquitard
2. The aquitard is overlain by an unconfined b '' K ''
aquifer, known as the source bed
Leaky confining layer b ' K ',S'
3. The water table in the source bed is
initially horizontal r h0
4. The water table in the source bed does b
not fall during pumping of the aquifer h K, S
S ' ( b ')
2

t , or , b '' K ''  100bK

10bK ' Fully penetrating well in an aquifer overlain by a
5. Groundwater flow in the aquitard is semipermeable confining layer
vertical
6. The aquitard is incompressible, so that no t  0.036b ' S '/ K ', or ,
water is released from storage in the
r  0.04b ( KS s / K ' S 's ) 
1/2
aquitard when the aquifer is pumped
7. The aquifer is compressible and water
drains instantaneously with a decline in
t  ( 30rw2 S / T ) 1 − (10rw / b )  , and
2
head  
Any well diameter:
5
rw / (Tb '/ K' )  0.1
1/2
Hantush’s (1956) solution for the case when
Potentiometric Original potentiometric
no water drains from the confining layer surface surface

 2 h 1 h S h K ' ( h0 − h )
Q
+ = −
r 2
r r T t Tb ' b '' K ''
Leaky confining layer b ' K ',S'
Q
( 0 )
h − h = W ( u, r / B )
4 T r h0
1/2 b
 Tb ' 
2
r S h
u= ,B=  K, S
4Tt  K'
Fully penetrating well in an aquifer overlain by a
Leaky artesian well function, W(u,r/B) semipermeable confining layer

The rate at which water is being drawn from elastic storage in the confined
aquifer, qs [m3/day] at a specific time, t [days] since pumping began is
determined from:
s (
q = Q exp −Tt / SB 2 )
For a total discharge of Q at time t, the water coming from leakage across
the aquitard is: q =Q−q
L s
6
Hantush’s (1956) solution for the case when
Potentiometric Original potentiometric
no water drains from the confining layer surface surface

Q
( h0 − h ) = W ( u, r / B ) Q
4 T
1/2 b '' K ''
 Tb ' 
2
r S
u= ,B=  Leaky confining layer b ' K ',S'
4Tt  K ' 
r h0
b
If the well is pumped for a long time, all h K, S
the water may come from leakage across
the aquitard and none from the elastic Fully penetrating well in an aquifer overlain by a
storage (Hantush & Jacob, 1954), this semipermeable confining layer
occurs when: 8b ' S
t
K'
Q
Drawdown in this case: ( h0 − h ) = K0 ( r / B )
4 T
Where K0 is a zero-order modified Bessel function of the second kind (see
from table) 7
Case 2. Hantush’s (1960) solution for the case when some
water comes from elastic storage in the aquitard Original potentiometric
surface
2 solutions: Potentiometric
S 'b' surface Q
1. Early time condition: t 
10 K '
Q b '' K ''
Solution: ( h − h ) = H ( u,  )
4 T Leaky confining layer b ' K ',S'
0

1/2 r
r  Tb '  h0
= ( S '/ S ) , B =  
1/2
b
4B  K '  h K, S
r 2S
u= Fully penetrating well in an aquifer overlain by a
4Tt semipermeable confining layer

Rate of flow from the storage in the main

aquifer is given by:
qs = Q exp ( t ) erfc ( t )
Where, the erfc is the complementary  = ( K '/ b ') ( S '/ S 2 )
error function = 1-erf(x)

8
Case 2. Hantush’s (1960) solution for the case when some
water comes from elastic storage in the aquitard Original potentiometric
surface
2 solutions: Potentiometric
surface Q
2. Equilibrium state:
8  S + ( S '/ 3) + S ''
t b '' K ''
( K '/ b ') + ( K ''/ b '' ) 
1/2
Leaky confining layer b ' K ',S'
Solution:
r h0
Q
(0 )
h − h = K 0 ( r / B ) ; rw / B  0.01 b
4 T h K, S
Fully penetrating well in an aquifer overlain by a
semipermeable confining layer

In first case, all water comes from the elastic

storage in the aquifer and the aquitard. At
equilibrium, all water comes from drainage from
the overlying source bed.

9
 Various solutions for
confined aquifers
Case Solution
No leakage Q
( ho − h ) = W (u )
4 T
Leakage (no
Q
storage from ( h0 − h ) = W ( u, r / B )
aquifer) 4 T
Leakage, Q
storage from ( 0 )
h − h = H ( u,  )
aquitard 4 T
Leakage,
Q
storage from
aquitard, ( h0 − h ) = K0 ( r / B )
equilibrium
4 T

Hantush, M.S., 1960. Modification of the theory of leaky

aquifers. Journal of Geophysical Research, 65(11), pp.3713-3725. 10
UNSTEADY STATE WELL HYDRAULICS:
UNCONFINED AQUIFER

11
Flow in an unconfined aquifer (Neuman & Witherspoon, 1969)
Original
Q
h0 − h potentiometric surface
 h K r h
2
h 2
h
Kr + + K = S
r 2 r r z 2 t
v s
Potentiometric
surface at time ,t
Saturated thickness of the aquifer, h [m]
Elevation above the base of the aquifer, z [m]
h0 r
Radial hydraulic conductivity, Kr [m/day]
Vertical hydraulic conductivity, Kz [m/day] h
Specific storage, Ss [1/m]
Time, t [days]
Fully penetrating well pumping from a confined
Radial distance from pumping well, r [m] aquifer

Assumptions in Neuman’s solution to this equation:

1. The aquifer is unconfined
2. Th vadose zone has no influence on the drawdown
3. Water initially pumped comes from instantaneous release of water from elastic storage
4. Eventually water comes from storage due to gravity drainage of interconnected pores
5. The drawdown in negligible compared with the saturated aquifer thickness
6. The specific yield is at least 10 times the elastic storativity
7. The aquifer may be anisotropic with differing radial and vertical hydraulic conductivities 12
Flow in an unconfined aquifer: Neuman’s solution
Original
Q
Q h0 − h potentiometric surface
( h0 − h ) = W ( u A , uB ,  )
4 T
r 2S Potentiometric
uA = , for early drawdown surface at time ,t
4Tt
h0 r
r 2Sy
uB = , for later drawdown
4Tt h
r 2 Kv
= 2
b Kh Fully penetrating well pumping from a confined
aquifer

Drawdown, h0-h [m] Time, t [days]

Pumping rate, Q [m3/day] Horizontal hydraulic conductivity, Kh [m/day]
Transmissivity, T [m2/day] Vertical hydraulic conductivity, Kz [m/day]
Radial distance from pumping well, r [m] Initial saturated thickness of the aquifer [m]
Storativity, S [-]
Specific yield, Sy [-]
13
 Non-equilibrium/ transient flow conditions
→ cone of depression continues to grow with time

DETERMINING AQUIFER PARAMETERS:

NON-EQUILIBRIUM CONDITIONS

14
Recall: A pumping test is an experiment when water is pumped from a well
(stress) at a specified rate and change in water level (response) is measured at
one or more locations surrounding the wells

1. To determine the hydraulic characteristics of the aquifer: hydraulic

conductivity, transmissivity, and storativity
2. Identify aquifer boundaries
3. Evaluate performance of the well
4. Water quality of the ground water

Assumptions:
1. The pumping well is screened only in the aquifer being tested
2. All observation wells are screened only in the aquifer being tested
3. The pumping well and observation wells are screened throughout the entire
thickness of the aquifer

15
Theis graphical solution
Q r 2S
Theis’s nonequilibrium equation: ho − h = W (u ) , u =
4 T 4Tt
Q 4Tut
Rearranging: T= W (u ) , S = 2
4 ( ho − h ) r

Steps:
1. Plot W(u) vs. (1/u) on log log scale – call this Theis reverse type curve
2. Plot drawdown (s=ho-h) vs. t on log log scale – call this observation curve
3. Overlay the observation curve over the Theis reverse type curve
4. Keeping the axes on both plots parallel, move the observation curve over the Theis
curve until the data points on the observation curves overlap with the Theis curves.
5. When the overlap occurs, identify any point on the Theis curve, preferably pick the
point where W(u) = 1 and 1/u = 1. This point is called the ‘match’ point.
6. For this point, read the x and y axis values on the observation curves.
7. You have W(u), u, (s=ho-h), and t, these can be substituted in the above equations to find
T and S.

16
Theis reverse non equilibrium type curve
(for fully confined aquifer)

Typical data from a well test

17
Matching field data to Theis curve to obtain aquifer
parameters

18
Cooper-Jacob straight line time-drawdown method
r 2S
After long time: u=  0.05
4Tt
Q
Ignore higher order terms in well T=  −0.5772 − ln u 
function: 4 ( ho − h )
2.3Q  2.25Tt 
→T = log  2 
4 ( ho − h )  r S 
Steps:
1. Plot drawdown vs. time on semi-log scale (time on log, drawdown on natural scale)
2. Extrapolate the line joining observation points to intersect the drawdown axis at s = 0
3. Read the value of time for this, call this to
4. The slope of the line gives the drawdown per log cycle of time: Δh-ho
5. Estimate transmissivity and storativity as: 2.3Q
T=
4 ( ho − h )
2.25Tto
S=
r2
19
Recovery test:
Q r 2S r 2S
Theis’s residual drawdown s' = W ( u ) − W ( u ')  , u = &u'=
estimate: 4 T  4Tt 4Tt '

Applying the approximation:

2.3Q t
s' = log
4 T t'
Steps:
1. Plot residual drawdown, s’ vs.
logarithm of t/t’
2. Estimate the slope of the
line, or the residual
drawdown, Δs, per log cycle
of t/t’
3. Estimate transmissivity as:
2.3Q
T=
4s '
20
What if the pumping well does not go through the full
aquifer width?
• 3D flow, vertical flow components
• Both vertical and horizontal hydraulic
conductivity values important
• Effects can be neglected for confined
aquifers if:
1. Observation wells are fully
penetrating
2. Observation wells are located more
than the following distance from the
pumped well: 1.5b K / K
h v A partially penetrating well in a confined aquifer.
3D flow due to vertical flow components.

• Effects of partially penetrating well in unconfined aquifers can be minimized if observation

wells fully penetrates the saturated thickness of the aquifer. Then:
1. Time drawdown curve for observation wells with following distances and time criteria
can be used to approximate the late time Theis curve r  b K / K , t  S / r 2T
obs h v obs y

2. Time drawdown curve for observation wells with following distances and time criteria
can be used to approximate the early time Theis curve
robs  0.03b K h / K v , tobs  S / r 2T 21
AQUIFERS IN NON-IDEAL
CONDITIONS

22
Effect of well interference: wells placed close to each
other

Composite pumping cone for three wells, pumping at different rates, tapping
the same aquifer.

• For confined aquifers, Laplace equation holds, and is linear. Therefore:

total drawdown = sum of individual drawdowns from each well
• For unconfined aquifers, the above method will under-estimate the actual drawdown
due to the non-linearity of the general GW equations in the unconfined case
23
Effect of hydrogeologic boundaries: rivers/lakes

A well bounded on one side by a stream,

a recharge boundary.

Recharge boundaries can be simulated by a recharging image well located opposite to

the river, at the same distance from the river as the original well

24
Effect of hydrogeologic boundaries: barriers

A well bounded on one side by an

impermeable boundary, or a barrier.

Barriers are simulated by locating a discharging well at an equivalent distance away from
the boundary on the opposite side.

25
Effect of hydrogeologic
boundaries: comparison

Recharge boundary retards the rate of

drawdown, therefore, the slope of drawdown vs.
time reduces in its presence.

Presence of barriers increases drawdown rates

above the ideal conditions (infinite aquifer), so
slopes are greater than the ideal case.

The observation of drawdown vs. time for long

time periods can be used to understand
whether recharge or barrier boundaries are
present in the aquifer system

26
Aquifer test design:
1. Where should I drill the pumping well?
– Use information on geology, geophysical surveys, aerial photos, including presence
of nearby wells
– Engineering and economic factors may play a greater role in determining the
location
– If possible, start by drilling a test/production well, after which the permanent well
is installed.
2. Make a borewell log for the well being drilled: a borewell log lists all the geologic
formations found at various depths during drilling
3. If possible, also note down the water levels in the well as the well is drilled
4. Use borewell log to determine potential aquifer zones, select one for testing
5. Install pump in the well and a device to measure flow rates (water meter within the
pipe line for small rates, orifice weir on discharge pipe for larger rates)
6. Make adequate arrangements to take water away from the test site
7. Maintain constant discharge during pumping, allow variations only within 10% (well
development helps here)
8. Account for any other source of water level changes (tidal, recharge sources, etc.)

27