Hydrogeological Modeling of the
Pullman-Moscow Basin Basalt
Aquifer System, WA and ID
Joan Wu, Farida Leek, Kent Keller
Washington State University
John Bush
University of Idaho
OUTLINE
 Introduction
 Hydrogeologic Setting
 Methodology
 GIS database development
 Ground-water flow modeling
 Results and Discussions
 Summary
 Position Announcement
2
INTRODUCTION
 The aquifer system in the CRBG is the sole water
supply source for the Palouse Basin
 The continuous water-level decline and the
projected future development have led to serious
public concerns
 PBAC: a multi-stakeholder, multi-agency (city,
county, university) organization promoting
conservation and sound ground-water
management
 The 2003 MOA with PBAC: GIS database
3
INTRODUCTION (cont’d)
Past Studies on Hydrogeological Characterization
 Crosby and Cavin (1960)
 Foxworthy and Washburn (1963); Jones and Ross
(1972)
 Bush and colleagues (1998, 2000, 2001, 2003)
Past Studies on Groundwater Modeling
 Barker (1979), overly conservative
 Lum et al. (1990), overly optimistic
 Both models proved inadequate by year 2000
4
INTRODUCTION (cont’d)
Goal
 To develop a foundation for improved and informed
Palouse Basin groundwater resources assessment
and management
Objectives
 To develop a hydrogeology GIS database for the
Palouse Basin to improve data accessibility and data
processing and analysis efficiency
 To develop a groundwater flow model for the
basaltic aquifer system of the Pullman-Moscow
area based on new spatial and temporal data
5
HYDROGEOLOGIC SETTING






Palouse loess
Saddle Mts.
Wanapum basalt
Grande Ronde basalt
Imnaha basalt
Pre-basalt
CRBG
6
HYDROGEOLOGIC SETTING (cont’d)
 Palouse loess: rural domestic use
 Wanapum basalt: major aquifer for Moscow
till 1960’s
 Grande Ronde basalt: source for more than
90% of water supply, with a recent
construction of WSU #8
7
 Occurred during late Miocene and
early Pliocene (17–6 mya BP)
 Engulfing ~ 1.6×105 km2 of the
Pacific Northwest between
Cascade Range and Rocky Mt.,
covering parts of ID, WA, and OR
 Over 300 high-volume individual
lava flows identified, along with
countless smaller flows, with vents
up to 150 km long
 Eventually accumulating to more
than 1,800 m thick
 Tectonic origin (Hooper, 1997)
 Yellowstone hot spot
 Thinning of continental
lithosphere due to spreading
behind Cascade arc
 Proximity of fissure vents to
tectonic boundary between
accreted terranes and lithospheres
of old N. Am. Plate
Source: USGS, http://vulcan.wr.usgs.gov/
Source: ND Space Grant Consortium, http://volcano.und.edu/
METHODOLOGY:
I. GIS DATABASE DEVELOPMENT
 Data Collection





Well log
Groundwater level
Pumpage
Precipitation
Geochemistry
 Data Compilation
 Digitizing into ArcGIS
 Processing existing and new coverages:
• Topography
• Township and range to UTM conversion of well coordinates
• Stream network
• Land use
• Soil
12
• Watershed boundary
Digitizing & Processing Well Data
Digitizing & Processing Well Data cont’d
R 45E
R 46E
R
6 W
R
5
W
T 15N
IDAHO
T 40N
WASHINGTON
T 16N
T 39N
T 14N
T 38N
R5W
R 46E
T 15N
6
5
4
7
8
9
18
17
16
19
20
21
30
29
28
31
32
33
D
C
B
A
E
F
G
H
M
L
K
J
N
P
Q
R
Well 15/46-31J1
6
5
4
3
2
1
7
8
9
10
11
12
18
27
16
15
14
13
19
20
21
22
23
24
30
29
28
27
26
25
31
32
33
34
35
36
b
c
a
Aa
d
Well 39N/5W-7ad2
METHODOLOGY:
I. GIS DATABASE DEVELOPMENT
 Data Analysis
 Plot long-term hydrographs
• Separate vs composite
• Their relations with precipitation and pumpage
 Build structural contour maps
•
To depict the shape of stratigraphic horizons
 Construct aquifer contour maps
• Wanapum
• Grande Ronde
 Develop hydrogeological cross-sections
• Across most of the basin
• In various directions
17
RESULTS AND DISSCUSSION:
I. GEOSPATIAL DATA ANALYSIS
Composite Hydrograph of Wells in the Palouse Basin
Composite Hydrograph for Palouse Basin Aquifer
Water Level Elevation, a.m.s.l., ft
2700
2650
2700
Private well (Carson)
Jones
Eveland
USGS
2650
Palouse Loess
2600
2550
Wanapum
2500
2450
2400
2350
Moscow # 1
Moscow # 2
Private well (Freight)
UI # 2
Moscow # 3
Moscow-Arden
UI-Irrigation
UI # 1
Moscow # 7
Cemet. well
Grande Ronde
2300
Pullman # 1
Pullman # 2
Pullman # 3
Moscow # 6
Moscow # 9
UI # 4
Pullman # 4
Pullman # 6
WSU # 3
WSU # 4
WSU # 5
WSU # 6
WSU # 7
UI # 3
Pullman # 5
WTEST
Pullman # 7
Moscow # 8
2250
2200
1923
2600
2550
2500
2450
2400
2350
2300
2250
1931
1939
1947
1955
1963
Year
1971
1979
1987
1995
2200
2003
Water Level, Elevation, a.m.s.l., ft
Long-term Hydrograph for Pullman and WSU Grande Ronde Wells
2310
2310
2305
2305
2300
2300
2295
2295
2290
2290
2285
2285
2280
2280
2275
2275
2270
2265
2260
2255
2250
2245
2240
2235
2230
1974
2270
Pullman 3
Pullman 4
Pullman 6
Pullman 5
Pullman 1
Pullman 2
Pullman 7
DOE
WSU 3
WSU 4
WSU 5
WSU 6
WSU 7
WTEST
1977
1980
2265
2260
2255
2250
2245
2240
2235
2230
1983
1986
1989
Year
1992
1995
1998
2001
2004
Long-term Hydrograph for Moscow and UI Grande Ronde Wells
2300
2300
2295
2290
Water Level Elevation, a.m.s.l., ft
2295
Moscow 6
Moscow 8
Moscow 9
UI 3
UI 4
2285
2280
2290
2285
2280
2275
2275
2270
2270
2265
2265
2260
2260
2255
2255
2250
2250
2245
2245
2240
2240
2235
2235
2230
2230
2225
2225
2220
1974
2220
1977
1980
1983
1986
1989
Year
1992
1995
1998
2001
2004
3500
0
3000
500
2500
2000
1000
Pullman pumpage
Moscow pumpage
Total pumpage
Pullman precipitation
Moscow precipitation
Annual precipitation, mm
Annual Pumpage, MGY
Long-term Groundwater Pumpage from Two Aquifers
1500
1000
500
(a)
0
1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
Year
Long-term Hydrographs
 Each aquifer has a distinct pattern of water-level
fluctuations in relation to pumping, climate, recharge
 Wanapum saw its groundwater level recovery since
1960’s when pumping shifted to the Grande Ronde
 Relatively more consistent pattern of fluctuation in
Grande Ronde wells in Pullman than in Moscow
 0.3–0.6 m/yr groundwater level decline observed at both
pumping centers
28
Contour Map of Top Altitude of Wanapum Formation
Contour Map of Top Altitude of Grande Ronde Formation
Structural Contour Maps
Wanapum
 Wanapum basalt is to the NW controlled by NW
trending folds, and dips and thickens E and W away
from Pullman
Grande Ronde
 The top of GR drops in elevation E towards Moscow
and W and NW away from Pullman
 Substantial lateral changes in the occurrence and nature
of sediments exist between Pullman and Moscow
32
Potentiometric surface contour map of the Wanapum aquifer (1960s)
Potentiometric surface contour map of the G. Ronde aquifer (1990s)
Potentiometric Surface Contour Maps
Wanapum
 Hydraulic connection between Pullman and Moscow is
weak
 General groundwater movement is to W and NW
Grande Ronde
 Piezometric surface shows two cones of depression as
a result of heavy pumping
 The open shape of cones of depression to the W and
NW is possibly controlled by structural features
35
METHODOLOGY:
II. DEVELOPING A NEW MODEL
Water Release from a Confined Aquifer:
Water Expansion + Aquifer Compression
41
Source: http://www.bae.uky.edu/sworkman/AEN438G/theiseq/theiseq.html
Unsteady-State Flow in “Ideal”
Aquifer: Theis (1935) Equation
“The flow of ground water has many analogies to the flow of heat by
conduction. We have exact analogies … for thermal gradient, thermal
conductivity, and specific heat…solution of some of our problems
is probably already worked out in the theory of heat conduction…”
Source: http://www.olemiss.edu/sciencenet/saltnet/theisbio.html
42
Unsteady-State Flow in “Ideal”
Aquifer: The Solution
“Actually derived by a mathematician friend of Theis, C.I. Lubin.
Reportedly, Lubin declined co-authorship of the paper because
he regarded his contribution as mathematically trivial.” [Fetter, 1994]
43
Groundwater Flow Model Development
 Industry standard MODFLOW
 MODular 3-d finite-difference
groundwater FLOW model
 Free source codes from the USGS
and GUI versions available
 PEST (nonlinear parameter
estimator) can be used with
MODFLOW for optimal
parameterization
Source: http://water.usgs.gov/nrp/gwsoftware/modflow2000/modflow2000.html
44
Comparison of Model Domain and
Structure
Barker (1979)
Western BC at Union Flat Cr.;
One lumped basalt aquifer; “single-layer-cake”
Lum et al.
(1990)
Western BC at Snake R.;
Palouse Loess + two separate basalt aquifers, layers
horizontal
New Model
Western BC as in Barker (1979);
Three model layers with actual top/bottom altitudes
45
Comparison of Western Boundary
Condition
Barker (1979)
Dirichlet (head) at Union Flat Cr. for lumped aquifer
Lum et al. (1990)
Cauchy (weighted head and flux) at Snake R. for all
three aquifers
New Model
Same as in Barker (1979) but for three distinct aquifers
46
Comparison of Hydraulic
Parameterization
Comparison of Hydraulic Parameterization
Barker (1979)
Uniform hydraulic properties within zones:
Kh = Kv = 0.03–7.9 m/d, S = 0.005
Lum et al. (1990)
Uniform hydraulic properties within zones of each aquifer:
Loess: Kh = 1.5 m/d, Kv = 0.02 m/d
Wanapum: Kh = 0.1–0.2 m/d, Kv = 2.4–3.6×10−4 m/d
Grande Ronde: Kh = 0.1–3.7 m/d, Kv = 3.1–76×10−5 m/d
S = 0.001
New Model
Apply inverse modeling to a wealth of historical head data
for greatly improved parameterization
47
Comparison of Recharge Distribution
Barker (1979)
17 mm yr−1 uniform across model domain
Lum et al. (1990) 71 mm yr−1 uniform across model domain
New Model
Spatially varying following O’Green (2005):
3 mm yr−1 in 33% (near Moscow Mt.),
10 mm yr−1 in 37% (Pullman area),
actual infiltration in 10% (valleys) of the basin area
48
Management Alternatives
Given: pumpage needs 2,400 MGY = 9.1×106 m3, basin area 660 km2
Aerial Recharge
Recharge needs: 14 mm
Winter wheat consumes up to 90% annual precipitation of 550 mm
Winter runoff loss unavoidable from conventionally farmed fields
Low permeability across Bovill sediment–Wanapum basalt contact in places
Transporting Surface Water from Snake R.
Economic feasibility low but of potential
Artificial Recharge
Of greatest potential when using streams incised into Wanapum
Ground-water modeling imperative in determining the effectiveness
49
SUMMARY AND CONCLUSIONS
 GIS database has in the first time brought together the
various scattered data pertinent to PBA hydrogeology
and placed it in uniform and easily accessible form
 Such database facilitates efficient data retrieval and
analysis and allows continuous updating and
refinement, forming a solid foundation for future transboundary hydrogeolocial investigation
 A great deal has been learned from this newly
available digital temporal and spatial data
 Development of an improve basin-scale groundwater
flow model is underway
50
THANK YOU !
Pullman─Moscow Cross-section
Pullman─Moscow Cross-section
Pullman side
 Less sedimentary interbedding
 Loess is in direct contact with the basalt
 Wanapum is unproductive
Moscow side
 More sedimentary interbeds
 Wanapum is highly productive
 Current hydraulic gradient and ground-water flow in
Grande Ronde between Pullman and Moscow is
minimal, reflecting good hydraulic connection and lack
of dike barrier as suggested by some scientists
52
Long-term Hydrographs Revisit
 Relatively consistent pattern of fluctuation in
Grande Ronde wells in Pullman
 Aquifer is shown to have been depressurized!
 Greater fluctuation in Grande Ronde wells in
Moscow due to
 Multi-layered sediment system
 Proximity to low-permeability boundaries created by
non-basaltic rocks
 Confined nature of aquifer
 All these factors tend to cause longer recovery period
for the wells to reach equilibrium
53
Pullman─Albion─Colfax Cross-section
 Fracture patterns and degree of weathering dominantly
control the productivity of wells
 Grande Ronde dips eastward towards Colfax with a
hydraulic head drop of 150 m
 Intrusion of low-permeability pre-Tertiary rocks are
considered to form barriers between Pullman and Colfax
and cause the drastic change in hydraulic head
 Certain previous pump test results may be questionable;
substantial ground-water flow from Pullman to Colfax
appears unlikely
54
Pullman–Union Flat Creek–Snake River
 Significant difference (~460 m) exists in hydraulic heads
of the Wanapum and Grande Ronde near the Snake R.;
this sudden change in head may be related to the dip of
the basalt flows to the NW away from the Snake R.
 Cross-sections and potentiometric surface maps suggest a
major flow direction of NW along the Snake R.;
significant seepage along the canyon walls of the Snake
R. from the Grande Ronde aquifer is unlikely
 Geochemistry data from previous studies (Larson et al.,
2000) also indicates a lack of Grande Ronde discharge to
the Snake R.
55
SUMMARY AND CONCLUSIONS (cont’d)
 Long-term trends of the hydrographs indicate weak
vertical hydraulic connection between the two basalt
aquifers, consistent with pervious isotope geochemistry
studies
 Each aquifer exhibits a distinct pattern of water-level
fluctuation as affected by pumping, climate and
recharge, with the top basalt aquifer seemingly receiving
Holocene precipitation recharge and the bottom aquifer
pre-Holocene recharge
56
SUMMARY AND CONCLUSIONS (cont’d)
 Potentiometric surface contour maps of the basalt
aquifers display a general pattern with the groundwater level dipping S–NW along the ancient basalt
flow
 Existing structural features (monoclines, anticlines and
synclines) tended to create local areas with rapid
changes in water levels in the approximate direction of
their major axis
 Previous modeling studies using Snake R. as a Cauchy
boundary and forced high recharge may have been the
key causes of the model failures
57
SUMMARY AND CONCLUSIONS
 Geologic and hydrogeologic conditions at the two
cities of Pullman, WA and Moscow, ID in the Palouse
Basin are rather different; yet the hydraulic connection
appears strong
 The nature and position of stratigraphic units and their
inherent spatial heterogeneity together with geologic
structures have significant effects on the ground-water
flow regime in a fractured complex basalt system,
which should be carefully taken into account in future
modeling efforts
58