WNH THREE CANYONS (WNH3C) GROUNDWATER FLOW …...WNH THREE CANYONS (WNH3C) GROUNDWATER FLOW MODEL...

ASSESSMENT OF WESTERN NAVAJO AND HOPI (WNH)

WATER SUPPLY NEEDS, ALTERNATIVES AND IMPACTS

TECHNICAL MEMORANDUM

WNH THREE CANYONS (WNH3C)

GROUNDWATER FLOW MODEL

This memorandum describes the evaluation and revision of an existing groundwater

flow model for the Three Canyons (Clear Creek, Chevelon Canyon, and Jack’s Canyon)

area between the southwest corner of the Navajo Nation and the Mogollon Rim. The

existing model was developed HydroGeoChem, Inc. (1997) for the Bureau of Indian

Affairs and the reader is referred to that document for supporting information.

HydroGeoChem(1997) used a proprietary model code, called ABEL, to develop the

existing model, now informally called the “Hopi Ranches Model”. The revised model is

called the Western Navajo-Hopi Three Canyons (WNH3C) model and is based on the

freely-available MODFLOW model code. The purpose of the WNH3C model is to

estimate the impacts from pumping on baseflows in Clear Creek and Chevelon Canyon.

This memorandum only addresses the development and calibration of a steady state

simulation using the WNH3C model.

1.0 Initial Model Preparation

The WNH3C model was developed from basic data, rather than conversion of the

input files developed for the ABEL-based Hopi Ranches Model. The first step was to

plot the basic data in the form of GIS files in ArcGIS ArcView version 8.2. The GIS files

are in UTM Zone 12, NAD 1927 coordinates. The WNH3C model uses feet and days as

units. The offset from the origin of UTM Zone 12 to model coordinates is: 1,570,505

of 7

Revision 0 December 24, 2002 HDR Review Draft Western Navajo Hopi Three Canyons Model Technical Memorandum

feet in the X direction, 12,521,635 feet in the y direction, and no tilt from an easterly

orientation. The steady state model was developed from GIS files using Groundwater

Vistas Version 3.3.

The model grid was designed, following the HydroGeoChem (1997) report as a grid

of 211 columns and 183 rows. Each grid cell is a square has the same size: 1320 feet on

a side.

The next step was to plot and contour the available water-level measurements

indicated in HydroGeoChem (1999) and provided as an EXCEL file by John Ward, one

of the authors of the HydroGeoChem (1997) report. The resulting contours were similar

to those presented in HydroGeoChem (1997) and were used to assign prescribed head

boundary conditions around the model periphery.

The top of the C-Aquifer was delineated from surface elevations where the C-Aquifer

is exposed at land surface (determined from the Arizona Geological Survey Geologic

Map of Arizona). Where the Moenkopi and some younger units overlie the C-Aquifer,

modifications were made to the land surface-based estimate for the top of the C-Aquifer.

In the north and northeast areas of the model, where the C-Aquifer is deeply buried,

contours on the base of the overlying Moenkopi Formation were transcribed from Ulrich

& others (1984). The remaining areas covered by Moenkopi Formation were addressed

manually by continuing land surface elevation contours across from opposing sides of

limited areas of Moenkopi Formation occurrence.

The base of the C-Aquifer was derived by subtracting the saturated thickness

presented by HydroGeoChem (1997) from the estimated water-level elevation in the C-

Aquifer for areas of exposed C-Aquifer. For the north-northeast area, a thickness of

approximately 650 feet was subtracted from the elevation of the base of the Moenkopi

Formation.

of 7


The available hydraulic conductivity estimates were plotted by HydroGeoChem

(1997). The supporting aquifer test data were not obtained or re-analyzed for this effort.

Based on inspection of the plot of the available hydraulic conductivity estimates, a range

of values of 0.1 to 20 feet per day (ft/d) was considered reasonable. No systematic trend

or grouping is apparent for the WNH3C model area from inspection of the

HydroGeoChem (1997) plot. The three-part distribution used by HydroGeoChem (1997)

in their model (one zone each with 1, 6, or 6.5 ft/d) was used as an initial distribution.

The locations of channels potentially interacting with groundwater were identified by

comparing the water-level elevation contours with land surface elevation contours shown

on USGS 7.5 minute quadrangle sheets along Clear Creek and Chevelon Canyon. Three

primary areas were identified as having potential groundwater interaction:

• Clear Creek in the vicinity of Blue Ridge Reservoir,

• Clear Creek near its confluence with the Little Colorado River, and

• Chevelon Canyon near its confluence with the Little Colorado River.

The lower Clear Creek and lower Chevelon Canyon reaches were documented by

HydroGeoChem (1997) as perennial, but no discussion was provided of the perennial

nature of Clear Creek and its tributaries near Blue Ridge Reservoir in the far

southwestern corner of the Hopi Ranches Model. Nevertheless, comparison of water-

level elevations and the land surface confirms the potential for groundwater-surface water

interaction in the vicinity of Blue Ridge Reservoir. The model grid was overlain on the

7.5-minute quadrangle sheets in ArcView 8.2 and parameters for the MODFLOW Stream

Package were estimated fro each cell representing Clear Creek, Chevelon Canyon, and a

few tributaries of Clear Creek near Blue Ridge Reservoir.

Although HydroGeoChem (1997) did not incorporate recharge into their simulation,

review of a recent white paper on recharge developed by HDR (2002) indicated that the

exposed C-Aquifer should receive significant recharge, particularly at the elevations

included in the WNH3C model. Simplified regions of equal recharge based on the

of 7


elevations, precipitation, and percent of precipitation becoming recharge were adopted as

initial recharge inputs for the WNH3C model for areas of exposed C-Aquifer.

An area in the southeastern portion of the WNH3C model, noted as dry by

HydroGeoChem (1999) was set as inactive in the WNH3C model.

The depths to water across the model were calculated and reviewed. Based on this

review, the potential or evapotranspiration from the groundwater was considered to be

negligible at this time. Future evaluation may provide a basis for adding the simulation

of evapotranspiration in limited areas.

The pre-conditioned conjugate gradient solver package (PCG2 – Hill, 1990) was

selected for the WNH3C model. The solver was set to achieve a head residual of 1 x 10-4

feet and a mass balance residual of 1 cubic foot per day (ft3/d) using 1,000 outer

iterations and 10 inner iterations.

3.0 Water-Level Targets and Calibration

The water-level data provided by John Ward were input to Groundwater Vistas

Version 3.3 as targets for calibration the model was operated with MODFLOW-2000 and

the match of model simulation to measurements was observed. Overall match was

relatively poor so calibration was pursued. Calibration consists of adjusting input

parameters to improve the match of model-simulated water-level elevations to measured

water-level elevations. The hydraulic conductivity and recharge distributions described

above were modified to achieve an acceptable calibration. The required modifications

were well within the range of reasonable values.

of 7


After acceptable calibration was achieved, a band of cells surrounding the inactive

area in the southeast portion of the model were set to inactive as their bottom elevations

were above the simulated water level elevations.

4.0 Results

The inflow portion of the WNH3C model mass balance for steady conditions consists

of 16,915 acre-feet per year (ac-ft/yr) of inflow from prescribed head boundaries, joined

by 35,450 ac-ft/yr of diffuse recharge, and 2,840 ac-ft/yr of recharge from streams. The

outflow portion of the WNH3C model mass balance for steady conditions consists of

50,330 ac-ft/yr of outflow to prescribed head boundaries and 4,875 ac-ft/yr of baseflow

out to streams. Total inflow or outflow under steady conditions is approximately 53,000

ac-ft/yr.

The base flow gains for Clear Creek and Chevelon Canyon simulated by the WNH3C

model are as follows:

• 4 cubic feet per second (cfs) for Upper Clear Creek in the vicinity of Blue

Ridge Reservoir

• 2.5 cfs for Lower Clear Creek

• 0.4 cfs for Lower Chevelon Canyon

These values are different than those derived from model simulations by

HydroGeoChem (1997), who found:

• No flow for Upper Clear Creek in the vicinity of Blue Ridge Reservoir

• 4 cfs for Lower Clear Creek

• 5.5 cfs for Lower Chevelon Canyon.

A useful assessment of the meaning of the differences between the HydroGeoChem

(1999) and WNH3C models would require detailed analysis of available stream flow

of 7


records and observations. However, given the acceptable simulation of the water-level

elevations by the WNH3C model, particularly in the areas of the groundwater-surface

water interactions, such an analysis is not considered necessary at this time.

Figure 1 shows the extents of the WNH3C model, modified from the existing

HydroGeoChem model (1997) as discussed in this memorandum. Also shown on Figure

1 are the inactive cells, prescribed head (constant water-level elevations in this case)

cells, stream package cells, the Little Colorado River and lesser surface water channels,

township and range boundaries, and the locations of major roads.

Figures 2 and 3 show the distributions of hydraulic conductivity and recharge in the

calibrated WNH3C model. Figure 4 shows the measured water-level elevations at points,

interpreted contours of the measurements, and the water-level elevations simulated by the

WNH3C model.

Figure 5 shows the cross plot of simulated and measured water level elevations after

calibration. The current standard deviation of the water-level elevation simulation

residuals is 39 feet, which is approximately two percent of the overall range in measured

water-level elevations (1758 feet). Figure 6 shows the distribution of WNH3C

calibration residuals (measured water-level elevation minus simulated water-level

elevation). The WNH3C model is ready for simulations of the response of base flows in

Clear Creek and Chevelon Canyon to future pumping from the C-Aquifer.

4.0 REFERENCES

HDR Engineering, 2002. White Paper – Recharge Assessment of Western Navajo Hopi Water Supply N-Aquifer Model (WNHN), prepared for and presented to the Technical Support Group.

Hill, M. C., 1990, Preconditioned Conjugate-Gradient 2 (PCG2), a Computer Program

for Solving Ground-Water Flow Equations: U.S. Geological Survey Water-Resources Investigations Report 90-4048, 43 pages.

of 7


HydroGeoChem, Inc. 1997. Results of Numerical Simulations of Impacts on Clear and

Chevelon Creeks from Proposed Groundwater Developments on Clear Creek and Aja Ranches (Hopi Ranches), prepared for the U. S. Bureau of Indian Affairs, dated, December 4, 1997, 24 pages plus tables and figures.

Ulrich, G.E., Billingsley, G.H., Hereford, R., Wolfe, E.W., Nealey, L.D., and R.L.

Sutton, 1984. Map Showing Geology, Structure, and Uranium Deposits of the Flagstaff 1 Degree by 2 Degree Quadrangle, Arizona, U.S. Geological Survey, Miscellaneous Investigation Series, Map I-1446.

0 6 12 18 24 303Miles

®

Jacks Canyon

Clear Creek

Chevelon Canyon

Blue Ridge Reservoir

Chevelon Canyon Lake

Little Colorado River

Figure 1 WHN Three Canyons Model Extents, Surface Water Features and Boundary Conditions

ExplanationModel Extents

Roads

Township/Range

Moenkopi Extents

Unique Model Cell Type

CH

Noflow

Stream


Roads

Township & Range

Kx0.6

7

12

0 10 20 30 402.5 5 7.5Miles

Figure 2 Distribution of Calibrated Hydraulic Conductivity WNH Three Canyons Model

Kx = horizontal hydraulic conductivity in feet per day

Area where C-Aquiferis assumed to be dry.


Roads

Township & Range

Recharge R

0

0.00009

0.0001

0.00015

0.0003

Recharge Rate is a flux density in feet per day

Figure 3 Distribution of Recharge Applied to WNH Three Canyons Model

®

0 6 12 18 24 302 4Miles


5600

5100

5000

5700

5200

4900

5800

5500

4800

5300

5400

5900

6000

61006200

6300

4700

6400

65006600

6700

68006900

5000

5700

4900

5800

5100

4800

5600 5200

5500

5300

5400

5900

6000

6100

6200

4700

6300

6400

6500

660067006800

5100

5200

5600

5500

4707 46774925

469646794750465649045005 4719

47844810 4725 47354710482848314724

5008 4770 4723 47124744

47844819 4754 4789 47754884

482648804922 4803 48335007 4930 484448434848

485648505090

48754878 4908

4871 496949004883 49864969

48994899 4951 50124987

49865109 5003499049144940 5000

501349405109

4997500549665135 50065140

501251335131

5017 5060

5193 50985058

50875272 5048 50555235 5036

53265077 5084

5300 515560705812 5200

56675654 527651805825

58936340

6375 634964026258 5405

637355236372 5463

6414

57615525

552555836280

58225618559359206020 56576726

5833 559755946907

®

0 10 20 30 402.5 5 7.5Miles

Figure 4 Water-Level Elevations WNH Three Canyons Model Area

Interpreted Contour of Equal MeasuredWater-Level Elevation(thick solid blue line)

Contour of Equal Simulated Water-Level Elevation(thinner - dashed blue line)


Measured Water-Level Elevation

Road

Township-Range Boundary

4500 4750 5000 5250 5500 5750 6000 6250 6500Measured Water-Level Elevation (feet above mean sea level)

4500

4750

5000

5250

5500

5750

6000

6250

6500Si

mul

ated

Wat

er-L

evel

Ele

vatio

n (fe

et a

bove

mea

n se

a le

vel)

Water-level ElevationsC-Aquifer Water Levels1:1 (exact correspondence)+/- 1 SD (38.9 feet)

Figure 5 Cross Plot and Residuals Plots for Measured and Simulated Water-Levels, WNH Three Canyons Model

4500 4750 5000 5250 5500 5750 6000 6250 6500

-150

-100

-50

0

50

100

150

Resid

ual (

mea

sure

d - s

imul

ated

)

$+$+ $+ $+ $+$+

$+$+

$+$+

$+

$+ $+$+ $+$+ $+$+

$+$+

$+

$+

$+ $+ $+$+$+

$+$+$+

$+$+

$+$+

$+$+

$+

$+$+$+ $+$+

$+$+

$+$+

$+

$+ $+

$+

$+$+ $+ $+$+

$+ $+$+

$+$+

$+$+ $+$+$+$+$+ $+ $+$+$+ $+$+

$+

$+$+

$+$+

$+$+$+$+$+$+

$+ $+$+$+ $+$+

$+ $+$+$+$+ $+$+$+$+$+$+$+

$+$+$+ $+$+$+ $+

$+$+ $+

$+ $+$+

65

47

-4

-9

27

71

5.50.2

-336.8

1.3

6.33.1

3.79.6

1.32.8

3.4 0.21.3

-32

5.2

1.9

-22

-24

128

-3.1

-9.912.914.8

76.7-0.7-0.532.7

39.619.734.1 12.6

-0.621.7-8.6

12.7

28.3

18.4-3.9-4.4

-3.822.4

28.530.4

76.823.1

84.2

76.1

65.8-6.4

21.7

81.5-0.8

-4.7

-8.435.758.6

19.568.6-1.3

17.233.732.3

24.6-1.3 83.213.4

20.2-7.7

-34.5-11.9-23.3

-12.6-45.7 -17.9 -14.6

-35.1 -48.5-17.8 -14.9 -20.8

-11.7

-17.6

-11.2

-14.8-15.2

-15.3

-19.5-64.4

-23.3

-32.9 -23.2 -10.6

-90.2

-46.7

-31.2

-79.5

-31.2-41.1 -30.4-20.8

-13.8

-98.9-20.1

-22.3-95.9

0 6 12 18 24 303Miles

®

Figure 6 WHN Three Canyons Model Calibration Residuals

Explanation$+ Residual (feet)

Model Extents

Roads

Township/Range

Unique Model Cell Type

CH

Noflow

Stream

WNH THREE CANYONS (WNH3C) GROUNDWATER FLOW …...WNH THREE CANYONS (WNH3C) GROUNDWATER FLOW MODEL...

Documents

Transcript of WNH THREE CANYONS (WNH3C) GROUNDWATER FLOW …...WNH THREE CANYONS (WNH3C) GROUNDWATER FLOW MODEL...