Date post: | 03-Mar-2015 |
Category: |
Documents |
Upload: | marsha-monestersky |
View: | 69 times |
Download: | 2 times |
1
Review of Peabody Western Coal Company’s (1984–2004)
Determination of Probable Hydrologic Consequences for the
Black Mesa‐Kayenta Coal Mine
Daniel Higgins, PhD
June 2011
2
INTRODUCTION
This report summarizes my evaluation of the predictive reliability and methodological adequacy of
Peabody Coal Company’s 1985 determination of Probable Hydrologic Consequences (PCC‐PHC 1985) that
was included in its Mining and Reclamation Plan: Black Mesa and Kayenta Mines1 (PCC 1985). This
report includes an evaluation of two related impact assessments — performed by the Office of Surface
Mining Reclamation and Enforcement (OSM) — that were based upon Peabody’s PHC.
Probable Hydrologic Consequences (PHC)
On 31 October 1984, Peabody submitted a permit application to OSM seeking approval of its mining
plan for the Black Mesa‐Kayenta Coal Mine (PCC 1985). This application required Peabody to determine
the probable hydrologic consequences that would result from the proposed activities (OSM 2004). To
make this determination, Peabody used the simulation results of a groundwater model developed by
the U.S. Geological Survey (see Eychaner 1983, and Brown and Eychaner 1988).
Cumulative Hydrologic Impact Assessment (CHIA)
Peabody’s 1985 PHC served as “the main source of input for the development of the cumulative
hydrologic impact assessment” (OSM 2002). In 1989, OSM released its Cumulative Hydrologic Impact
Assessment of the Peabody Coal Company Black Mesa / Kayenta Mine (referenced as OSM‐CHIA 1989).
Environmental Impact Statement (EIS)
The information in OSM’s 1989 CHIA was used to provide the technical hydrologic data needed in its
Draft and Final Environmental Impact Statement (referenced as OSM‐EIS 1990, 1989).
Because the USGS groundwater model (Eychaner 1983; Brown and Eychaner 1988) provided the
mine‐related impact predictions for Peabody’s 1985 PHC (PCC 1985) and OSM’s 1989 CHIA (OSM‐CHIA
1989) and 1990 EIS (OSM‐EIS 1990, 1989), an evaluation of their predictive‐reliability is warranted.
Collectively, the predictions in these documents provided the foundation for all regulatory decisions
related to the protection of Black Mesa’s groundwater resources since the mid‐1980s.
The following report summarizes my evaluation of the predictive‐reliability and methodological
adequacy of these interrelated documents.
1 Peabody’s permit application (PCC 1985) was made public in 2011 through the Freedom of Information Act. OSM’s CHIA (OSM‐CHIA 1989) and EIS (OSM‐EIS 1990) were acquired via OSM’s website: http://www.osmre.gov/.
3
Study contents and summary of findings
OVERVIEW
Literature review on the regulatory application of groundwater models Peabody’s PHC and the evolution of the USGS groundwater model Evaluation method of Peabody’s PHC
1) PEABODY WITHDRAWALS
a) Peabody withdrawals in relation to N‐aquifer recharge
The PHC, CHIA, and EIS underestimated Peabody withdrawals by approximately 6,735 acre‐feet during the fifteen year period the mine was fully operational (after completion of the EIS).
Assuming that Peabody’s safe yield, water‐budget methodology is an appropriate method for determining sustainable groundwater exploitation, and assuming that the USGS’s most recent recharge estimate is correct, Peabody’s total withdrawals have exceeded the rate of natural recharge by approximately 21,000 to 53,000 acre‐feet.
b) The Water‐Budget Myth Since 1940, USGS hydrogeologists have continuously demonstrated that the safe yield water‐budget methodology is a fallacy and is not an appropriate method for determining sustainable rates of groundwater development.
c) Future pumping projections The 1983 USGS model predicted that, by 1990, the rate of municipal withdrawals from the tribal communities would overtake the rate of Peabody withdrawals. This never occurred.
2) GROUNDWATER QUANTITY
a) KAYENTA
Disagreement regarding hydrologic characteristics at Kayenta precedes mining operations and continues to be the source of much contention. The PHC, CHIA, and EIS assume that 85% of the water level decline at Kayenta would be caused by Kayenta’s withdrawals. The EIS explains that the closest that potentiometric surface comes to the top of the N‐aquifer is 366 feet and would not occur until the year 2052.
In reality, however, the water level at Kayenta had fallen below the top of the N‐aquifer in 2005. Thus, this EIS prediction is off by nearly 50 years and approximately 370 feet. i) Water level in BM3 and Kayenta withdrawals
There is no statistically significant relationship between the rate of Kayenta withdrawals and water level decline at Kayenta well BM3. In fact, this analysis demonstrates that as Kayenta’s withdrawals increase, the water level does not fall, but it actually rises.
4
ii) Water level in BM3 and Peabody withdrawals There is a strong, statistically significant relationship between the rate of Peabody’s increasing withdrawals and water level decline in Kayenta well BM3. It is hypothesized here that the magnitude of Peabody withdrawals is so great that it simply skews any indicator of Kayenta’s withdrawals on the water level in BM3.
iii) Indicators using box‐plots with outliers 1985—the year that the mine ceased withdrawals due to maintenance at the Mohave Generating Station—is a statistical outlier for Peabody withdrawals. This coincides with an outlier for water level in Kayenta well BM3: 1985 is the only year that water levels fell below 2 standard deviations of the mean water level.
Moreover, Kayenta’s total withdrawals in 1985 are near their mean, thus having no effect on the BM3’s conspicuous recovery that year.
iv) OSM’s conclusion on the source of Kayenta’s water level decline
OSM attributes drawdown at Kayenta to municipal withdrawals.
b) TUBA CITY / MOENKOPI Water level decline at Tuba City and Moenkopi (TC/M) is attributed entirely to local, municipal withdrawals. The closest that any measureable mining‐related impact is expected to occur is fifteen miles away from TC/M and will not occur until 2052.
The USGS monitors three NTUA and three BIA wells in the vicinity of Tuba City and Moenkopi. Water level decline—attributed to municipal pumpage—was overestimated by a mean 44% in the three NTUA wells, by a mean 103% in the three BIA wells, or by a mean 73% in the six combined wells. i) No correlation between Tuba City withdrawals and Tuba City water levels
The model overestimated the rate of TC/M’s municipal withdrawals by 11%. The disparity between overestimated decline and overestimated withdrawals indicate model error.
ii) Discharge from Moenkopi School Spring and to Moenkopi Wash iii) The USGS model and N‐aquifer discharge
Spring discharge near TC/M was predicted to decline by 1‐2% but this would be caused entirely by municipal withdrawals.
iv) The relationship between springs and water level v) N‐aquifer springs vi) CHIA criterion for material damage to spring discharge
OSM determined the material damage criterion for mine‐related impacts: discharge from springs must not decline by more than 10% in response to mine‐related groundwater withdrawals.
5
vii) Evaluation of the criterion The USGS monitoring program has measured discharge from four springs since approximately 1987. However, through water year 2005, the USGS reports inadvertently concealed declining trends in spring discharge through the use of logarithmic charts. Since 1987, the decline at Moenkopi School Spring exceeds 26.2%; if this is attributable to Peabody withdrawals rather than municipal withdrawals, then the threshold for material damage to spring discharge has been far exceeded.
viii) Moenkopi School Spring and Peabody Withdrawals There is a remarkably strong, statistically significant relationship between the rate of Peabody’s increasing withdrawals and the rate of declining discharge from Moenkopi School Spring. (r = ‐0.84; R2 = 0.71; p < 0.0001)
ix) Moenkopi School Spring and Tuba City withdrawals There is no statistically significant relationship between the rate of Tuba City withdrawals and discharge from Moenkopi School Spring. (r = ‐0.31; R2 = 0.09; p = 0.28)
x) Moenkopi School Spring and Tuba City Precipitation There is no statistically significant relationship between local precipitation at Tuba City and the rate of discharge from Moenkopi School Spring. (R2 = 0.11; p = 0.17)
xi) OSM on precipitation and spring discharge
OSM’s 1989 CHIA also reported that there is no apparent relationship between local precipitation at Tuba City and Moenkopi and N‐aquifer discharge in the area, but attributes declining spring discharge to municipal withdrawals.
c) FOREST LAKE
The well at Forest Lake represents the closest community well to the mining area. Maximum water level decline was predicted to occur in 2007; it would recover to its 1985‐level by 2009. Total decline since the pre‐mining period was predicted to be 148 feet. Impacts from mining would be minor. i) Recovery at Forest Lake
Due to the cessation of pumping at the Black Mesa mine at the end of 2005 rather than 2006 as the model simulated, maximum decline should have occurred sooner than 2007 as predicted, should be less than predicted, and should recover sooner than predicted. However, Maximum water level decline did not occur in 2007. Between April 2008 and May 2010, the water level rose 8.9 ft. At the time of this study (June 2011), no current data are available and it is uncertain if the water level is recovering. Thus, in 2011—having no clear recovery, much less recovery to its 1985 level, and despite its extra full‐year for recovery—the groundwater prediction demonstrates the model’s significant conceptual and/or input data error.
6
d) PINON
Maximum water level decline was predicted to occur in 2007. Total decline from the pre‐mining period will be 107 ft. i) Recovery at Pinon For the period 1985 to 2007, drawdown at Pinon was underestimated by 97%. From the pre‐mining period through 2007, drawdown was underestimated by 52%. At the time this study was performed, no appreciable recovery was apparent at the community of Pinon.
3) GROUNDWATER QUALITY Change in groundwater quality in the N‐aquifer would occur of withdrawals induced leakage from the overlying and poorer quality D‐aquifer. The PHC, CHIA, and EIS concluded that there is no evidence of significant vertical leakage from the D‐ into the N‐aquifer. In fact, the model demonstrated that only 4 acre feet of leakage per year would be induced by Peabody withdrawals. Considering that volume of groundwater in storage (estimated at 180 million acre feet), the effect on water quality would be negligible due to the 2 million to 1 dilution. a) Discussion
The D‐aquifer has never been monitored and the USGS groundwater model was not developed in consideration of any groundwater quality parameters: it has no hydrgeochemical simulation capacity. The model cannot determine change in groundwater quality, no matter the source.
Monitoring data express two N‐aquifer wells with TDS levels in excess of EPA’s Maximum Contaminant Level. The concentration of arsenic at Keams Canyon exceeds EPA’s Maximum Contaminant Level by a factor of four. Discharge from Moenkopi School Spring expresses significant increasing trends in TDS, chloride, and sulfate.
4) REACH OF PEABODY’S WITHDRAWALS
Numerous indicators express that the impact from Peabody’s withdrawals, and the areal extent of the confined N‐aquifer, far exceeds previous assumptions. These indicators are illustrated in USGS water level charts from wells distributed across the (assumed) confined and unconfined portions of the aquifer.
5) SUMMARY
6) REFERENCES
7
Figure 1. Tribal communities in the Black Mesa hydrologic basin area and the location of the Black Mesa‐Kayenta coal mine lease‐area (illustration from Macy 2010)
8
PROBABLE HYDROLOGIC CONSEQUENCES (PHC)
Literature review on the regulatory application of groundwater models
Konikow and Bredehoeft (1992) explain that, like hypotheses, groundwater models are generated as
a means of suggesting explanations for observed phenomena and predicting causal relationships
between phenomena: understanding of a groundwater basin increases when the model is iteratively
tested, falsified, and refined over time to develop a more accurate representation of the system.
The underlying philosophy of process‐simulating deterministic‐modeling… is that, given a
comprehensive understanding of the processes by which stresses on a system produce
subsequent responses in that system, a system’s response to any set of stresses can be
defined or predetermined through that understanding of the governing (or controlling)
processes, even if the magnitude of the new stresses fall outside of the range of historically
observed stresses. Predictions made this way assume an understanding of cause‐and‐effect
relations. The accuracy of such deterministic forecasts thus depends, in part, upon how
closely our concepts of the governing processes reflect the processes that actually control
the system’s behavior. (Konikow 1986)
Rather than demonstrating the accuracy and completeness of our hydrologic knowledge,
groundwater models expose “uncertainties and facilitate discussion of possible responses, which may
include various precautionary actions, steps to increase or maintain social flexibility and ecological
resilience, and/or research and monitoring schemes to reduce uncertainty” (Carpenter et al. 2002).
The accuracy of deterministic groundwater models comes into question when regulatory agencies
and the public require assurance that potential impacts from a proposed project have not been
underestimated; because decision‐makers rely upon impact‐assessments to approve or disapprove
projects that could adversely affect social and ecological systems, concern regarding the predictive
reliability or “correctness” of groundwater models is warranted (Hassan 2004; Woessner and Anderson
1996; Oreskes et al. 1994; Sargent 1990).
Consequently, modelers have pursued methods for testing the veracity of their models: “the notion
has emerged that numerical models can be “verified” or “validated”... Claims about verification and
validation of model results are now routinely found in the published literature” (Oreskes et al. 1994).
However, procedural inconsistency, semantic confusion, and disagreement regarding model capabilities
continue to hinder the modeling process, problematize policy‐decisions, and foster public skepticism
(NRC 1990, 2000; Leijnse and Hassanizadeh 1994; Konikow and Bredehoeft 1992; Anderson and
Woessner 1992a, 1992b).
9
“Acceptability of ground water models should be determined by using confirming observations to
support subjective judgment. The judgment is made in the context of the stated purpose of the model
and the nature of the supporting observations associated with each component of the modeling
process” (Woessner and Anderson 1996). The National Research Council2 (1990) acknowledges the
inherent limitations of deterministic groundwater models, adding:
Modelers must contend with the practical reality that model results, more than other
expressions of professional judgment, have the capacity to appear more certain, more
precise, and more authoritative than they really are. Many people who are using or relying
upon the results of contaminant transport models are not fully aware of the assumptions
and idealizations that are incorporated into them or of the limitations of the state of the art.
There is a danger that some may infer from the smoothness of the computer graphics or the
number of decimal places that appear on the tabulation of the calculations a level of
accuracy that far exceeds that of the model. There are inherent inaccuracies in the
theoretical equations, the boundary conditions, and other conditions and in the codes.
Special care therefore must be taken in the presentation of modeling results. Modelers
must understand the legal framework within which their work is used. Similarly, decision‐
makers, whether they operate agencies or in courts, must understand the limitations of
models.
There is no standard protocol for groundwater system‐modeling, ascertaining model validity, or
reporting simulation‐results. Consequently, agreement upon standards for accepting the conceptual
accuracy and predictive‐reliability of a deterministic groundwater model continues to be the source of
much contention (Anderson and Woessner 1992a; Hassan 2004). Because there are numerous
interpretations of the terms verification and validation, there are numerous approaches for conducting
these processes and disparate standards for gauging their attainment: “Both words imply authentication
of both the truth and accuracy of the model” (Konikow and Bredehoeft 1992). When the terms are used
interchangeably “to indicate that model predictions are consistent with observational data… modelers
misleadingly imply that validation and verification are synonymous, and that validation establishes the
veracity of the model” (Oreskes et al. 1994). It is often mistakenly assumed that once a model has been
calibrated, it has also been validated, and thus the model is perceived as an acceptable tool for
predicting the future conditions of a hydrological basin (Freyberg 1988; Konikow and Bredehoeft 1992).
However, Verification, calibration, and validation are three distinct processes, and all three are required
2 See Ground Water Models: Scientific and Regulatory Applications (1990). NRC members “are drawn from the councils of the National Academy of Sciences, the National Academy of Engineering, and the Institute of Medicine. The Members of the committee responsible for the report were chosen for their special competence and with regard to appropriate balance” (NRC 1990).
10
for determining the biophysical‐accuracy and predictive‐reliability of a groundwater model. Given the
inherent complexity of groundwater systems and the subjectivity required in the modeling process,
Woessner and Anderson (1996) have proposed three underlying principles for modelers and model‐
users to keep in mind: (1) groundwater modeling is inherently uncertain; (2) model acceptability should
be based upon the strength and number of actual observations that confirm the model’s predictions;
and (3) subjective judgment determines if a model appropriately represents the system.
Calibration of a groundwater model determines the level of accuracy that a model can reproduce
historical conditions within a predetermined range of acceptability. Because determining the actual
distribution of aquifer parameters is both technically and economically unfeasible, calibration is a
method of subjectively selecting a set of parameter values through manual trial and error (or automated
programs) for the system. Referred to as history matching, the process forces the model to reproduce
historical conditions by changing parameter values until an “acceptable” range of accuracy is achieved;
“There are no rules other than one’s judgment” (Konikow 1986; Konikow and Bredehoeft 1992, 1993).
Calibrated models are often presented, either implicitly or explicitly, as empirically adequate
representations of the system—that is, as valid representation of the system—but this is misleading.
Konikow (1986) acknowledges one consistent source of model error when calibration is equated with
validation:
It should be recognized that when model parameters have been adjusted during calibration
to obtain “best fit” to historical data, there is a bias towards extrapolating existing trends
when predicting future conditions, in part because predictions of future stresses are often
based on existing trends…. Concepts inherent in a given model may be adequate over the
observed range of stresses, but may prove to be oversimplified of invalid approximations
under a new and previously inexperienced type or magnitude of stresses.
Oreskes et al. (1994) explain that the necessity to refine a calibrated model “suggests that the
empirical adequacy of numerical models is forced… Consider the difference between stating that a
model is “verified” and stating that it has “forced empirical accuracy””.
It is axiomatic that models of complex systems cannot be validated: they can only be invalidated and
refined over time by testing the extent to which they diverge from reality (Holdgate 1978; Holling 1978;
Konikow and Bredehoeft 1992). In deterministic groundwater models, calibration is commonly used as
the basis for sizing sustainable rates of exploitation. However, because the parameter‐solution is non‐
unique, successful comparisons can result from an erroneous model: the iterative process of comparing
11
model‐predictions to field observations only reveals errors—it does not signify the absence of errors
(Konikow and Bredehoeft 1992; Greenberg et al. 1976).
“There are cases in hydrology… where our understanding of processes may be great, but predictive
ability is low, and other cases where understanding is minimal, but predictive accuracy is very high. In
any event, the accuracy of the prediction cannot be assessed until after the predicted period of time has
passed” (Konikow and Bredehoeft 1992; emphasis added).
If validation is interpreted to mean that a model can reliably predict the system’s future behavior
(i.e. predictive validation), then it can only be achieved by performing a postaudit (Anderson and
Woessner 1992b). However, validation is unlikely: “The issue of validation is mainly a regulatory one,
not a scientific one… Because our understanding of a system will always be incomplete, a model can
never be proven valid from a scientific standpoint” (Anderson and Woessner 1992b).
Peabody’s PHC and the evolution of the USGS groundwater model The U.S. Geological Survey (Eychaner 1983) has modeled the effects of the pumping in a fashion which permits the distinction of the two separate drawdown amounts at common locations… As can be seen, by the year 2001, the magnitude of the Peabody pumpage impact, in terms of the areal extent of drawdowns, is substantial. However, when comparing the magnitude of the Peabody drawdowns in terms of total drawdown at a location, the impact takes on less significance… The significance of Peabody caused drawdowns on total available community well water heights is not significant. (PCC‐PHC 1985: 37, 39)
The US Geological Survey first developed its two‐dimensional numerical model of the N‐aquifer in
1981 (Eychaner 1981) using MODFLOW (Trescott et al. 1976). The model was routinely evaluated and
updated throughout its period of use. By 1982, total pumpage from the N‐aquifer had increased enough
to cause significant water level decline; the USGS updated its model “to improve the estimates in areas
of uncertainty, and the resulting estimates collectively are considered to be more reliable” (Eychaner
1983; this version of the USGS groundwater model, referred to as “Eychaner 1983”, provided the
simulations that were used in Peabody’s determination of Probable hydrologic Consequences).
Although the 1983 model was judged to reproduce N‐Aquifer behavior “reasonably” well, its limitations
were explicitly outlined: inflow and outflow were uncertain, vertical leakage was ignored, future
precipitation and withdrawal rates were unknowable, further studies were needed to verify
assumptions and increase understanding of aquifer dynamics and response to pumping stress, and the
water‐budget could not be calculated given insufficient data (Eychaner 1983, 1981).
12
The unavailable aquifer parameter values were estimated during the model’s calibration procedures
because “data are not available for most nodes. Some data may be incorrect or be subject to multiple
interpretations” (Eychaner 1983: 14). Parameter (aquifer thickness, transmissivity, etc.) were estimated
and the model calculated water levels for each node; disparity between computed and actual water
levels was expected (Eychaner 1983). It is notable that the model predicted that most of the
groundwater lost to industrial pumpage “would be recovered within a few years if withdrawals at the
mine ceased”, and that, by 1990, “municipal supply pumpage is expected to exceed pumpage at the
mine” (Eychaner 1983).
By 1985, divergence between simulations and observations was evident, so the model was rerun to
include pumpage for the five years from 1980 to 1984 and parameter adjustments (recalibration) were
made to reflect the measured changes (Hill and Whetten 1986).
Through 1986, groundwater‐levels in non‐industrial wells in the confined portion of the N‐aquifer
continued to decline steadily. That same year, a reversal of the trend occurred in response to Peabody
shutting down their wells for 6 months in 1985 due to maintenance at the Mohave Generating Station3
(Hill and Whetten 1986). When pumping resumed in 1987, water‐levels in the non‐industrial wells
resumed their declining trends and reported their lowest levels since mining operations commenced.
Further, “most of the observation and non‐industrial wells in the northeastern section of the confined
area of the N‐aquifer showed record declines” (Hill and Sottilare 1987).
USGS evaluated the model’s performance again in 1987 and concluded that it was a reasonable
representation of N‐aquifer conditions (Hill and Sottilare 1987). However, because the software
program did not run properly on the new computer system when USGS attempted the model‐update, it
was converted to a new modeling program (McDonald and Harbaugh 1984), was recalibrated using
revised aquifer‐parameter estimates4, and used a finer spatial grid.
The 1988 update (Brown and Eychaner 1988) ran five pumping scenarios including both industrial
and municipal pumping for the period 1985–2051 (twenty years after groundwater pumping would
cease under the newly proposed permit application). Each simulation was based upon the proposed
3 This “reversal” appears to be seen throughout many N‐aquifer wells monitored by the USGS. Figure 19, at the end of this report, illustrates numerous wells with a conspicuous water level recovery “spike” around this period. 4 Re‐parameterization of the model was significant: conductivity “yielded unsatisfactory results so they were completely recalibrated”; recharge increased 4% to maintain a better balance between inflow and outflow; evapotranspiration increased from 6,000 af/y to 6,600 af/y; this was “balanced”; outflow to rivers and springs was reduced from 9,640 af/y to 7,030 af/y; river and drain conductances to achieve a reasonable match of both inflow and head; water‐table depth affecting evapotranspiration rates were changed; transmissivity throughout the model area changed; windmill pumpage was included (Brown and Eychaner 1988).
13
development‐regime and community growth projections. However, the USGS explained: “the accuracy
of projected water‐levels is uncertain. Future withdrawals that greatly exceed the amounts withdrawn
through 1984 may cause unanticipated responses in the aquifer. Long‐term projections that include
large increases in pumpage should be used with caution” (Brown and Eychaner 1988).
Consequently, the USGS explained, “These regional models cannot adequately represent the local
geology and simulate hydrologic processes in detail”, and that the model simulations “are better used to
compare effects of different development plans rather than estimate actual future water levels and
water‐budget components” (Brown and Eychaner 1988).
In 1993, Papadopolus and Associates (1993) evaluated the model’s input data and technical
correctness, and Waterstone Environmental Hydrology and Engineering (1995) evaluated its input data,
statistical uncertainty, and predictive reliability; both studies determined that the model performed
reasonably well. According to OSM (2006: 6), the USGS updated the model in 1994 with new
monitoring data and ran simulations (Littin and Monroe 1995; this report makes no reference to this
update, however). According to Peabody (HSIGeoTrans and WEHE 1999: 7‐1), the USGS updated the
model in 1996 (Littin and Monroe 1997).
In 1998, a Supervisory Hydrologist with the USGS Water Resources Division articulated his concern
regarding the incongruity between N‐aquifer oversight and monitoring, explaining that the USGS
monitoring program was “at best an early warning system in that it is indicative rather than
deterministic and is not set up to specifically address many of the criterion [sic]… The bottom line is
that… we need to tailor the current monitoring program in such a way as to more specifically address
the above criteria and in a deterministic fashion” (Hart 1998, quoted in NRDC 2000).
In 2001, the USGS suggested that (1) a better understanding of recharge, discharge, and leakage
should be ascertained before investing more time in refining the model; and (2) a new model using
current software may provide a better representation of aquifer dynamics (Thomas 2002).
In 2008, the USGS explained that the model was a poor representation of the N‐aquifer and had not
performed well (Leake 2008).
Evaluation method of Peabody’s PHC
To evaluate the accuracy of the PHC predictions (and concurrently, those of the USGS groundwater
model and OSM’s 1989 CHIA and 1990 EIS) this study reviews and summarizes the prediction data in
Peabody’s PHC and compares them to actual measured observations after the prediction period has
passed. In so doing, this study demonstrates the accuracy and success of regulatory procedures.
14
Regression analysis provides the methodological basis of any postaudit; it has numerous
advantageous: (1) it is widely used throughout numerous scientific and technological disciplines; (2) it is
relatively easy to understand by those without scientific and technical training; (3) it provides a strong
empirical evaluation of prediction accuracy; and finally (4), because modelers use regression to calibrate
and determine a model’s fitness for regulatory purposes, it is both a methodologically appropriate and
fair means for evaluating a model’s prediction accuracy after the prediction period has passed (Anderson
1995; Oreskes et al. 1994; Flavelle 1992; Anderson and Woessner 1992; Flavelle et al. 1990; Konikow
1986).
15
PEABODY WITHDRAWALS (1988‐2006)
The 1985 PHC, 1989 CHIA, and 1990 EIS projected that industrial withdrawals would remain
constant at 4,400 af/y through 2006, reduced to 1,100 af/y from 2007 through 2011, and then cease in
2012. However, according to the Final EIS, the model overestimated industrial withdrawals:
Since 1988, PCC [Peabody Coal Company] has used water from selected impoundments for dust suppression instead of water from the N‐aquifer. This equates to approximately 400 acre‐feet of water per year that is not removed from the N‐aquifer… For this reason, the 4,400 acre‐feet per year used in the N‐aquifer simulation overestimates the withdrawal from the N‐aquifer system by 400 acre‐feet or approximately 9 percent. (OSM‐EIS 1990: IV‐24)
However, industrial withdrawals were not overestimated (Figure 2). For the period 1990‐2005 (the
first water‐year following the completion of the Final EIS through the final year that both mines were
fully operational), Peabody withdrawals averaged 4,449 af/y. Assuming that 400 af/y of surface‐water
was used for dust suppression (as explained in the EIS), the mine’s groundwater‐budget of 4,000 af/y
was underestimated by approximately 449 acre‐feet per year, totaling 6,735 acre‐feet for the period5.
Figure 2. Peabody withdrawals, 1990‐2005. Withdrawal data from Macy (2010).
5 Peabody and OSM consistently report is that Navajo residents freely take water from the Peabody wells. Recognizing that this usage occurs in the context of residents driving automobiles to the lease‐area, filling water‐containers at the Peabody wells, and returning to their homes where the water is consumed, it is unreasonable to argue that this component accounts for anything more than a miniscule fraction of Peabody’s withdrawals (see, for example, OSM‐EIS 2008, 2006, 1990; OSM‐CHIA 2008, 1989; OSM 2006, 2004, 2000, 1998’ Peabody 2002).
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Peab
ody Withdrawals (af/y)
Peabody Withdrawals, 1990‐2005(acre‐feet/year)
16
Peabody Withdrawals and N‐Aquifer Recharge
OSM explained that, because “at no time does the total withdrawal from the system exceed the
recharge to the system (13,000 to 16,000 acre‐feet per year),” all impacts resulting from the mine’s
withdrawals “are overestimated” (OSM‐EIS 1990: IV‐24). Assuming that the conventional safe‐yield
water budget methodology is appropriate for determining sustainable rates of groundwater
exploitation, consideration of the USGS’s most recent and most accurate estimate for recharge to the N‐
aquifer is warranted.
In 1997, the USGS used geochemical analysis to re‐estimate N‐aquifer recharge and concluded that
it ranged from 2,500 to 3,500 af/y (Lopes and Hoffman 1997). Assuming this estimate is correct, for the
thirty‐two water‐years during the period 1972‐2005 (the period that both Black Mesa and Kayenta
Mines were fully operational, not including water year 1984‐1985 when withdrawals ceased for 6
months, Peabody withdrawals averaged 4,150 af/y), under the best case recharge scenario (3,500 af/y),
the mine depleted N‐aquifer storage by a total of nearly 21,000 acre‐feet. Under the worst case
recharge scenario (2,500 af/y), the mine depleted total storage in the N‐aquifer by nearly 53,000 acre‐
feet (Figure 3).
Figure 3. Comparison of Peabody withdrawals to the most current USGS recharge estimate. Withdrawal data from Macy (2010); recharge data from Lopes and Hoffman (1997).
‐
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
5,000
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
Acre‐feet / year
Peabody Withdrawals Compared to the 1997 USGS Recharge Estimate
min. recharge max. recharge Peabody withdrawals
17
The Water‐Budget Myth6
Methodologically, the safe yield water‐budget methodology is not an appropriate means for
determining sustainable rates of groundwater exploitation. In fact, the “Water‐Budget Myth” is
commonly referred to as an attractive fallacy and a perceptual convenience throughout the scientific
literature in hydrogeology, groundwater engineering and modeling, and natural resource management.
The method assumes that if the volume of groundwater withdrawn from an aquifer is less than the rate
of natural recharge, then development will be sustainable; if withdrawals equal recharge, then discharge
will be stopped but storage will remain intact, and is thus sustainable; and if withdrawals exceed
recharge, then discharge will be stopped, storage will be depleted, and development is not sustainable
(i.e. groundwater mining). However, despite the apparent logic of the water‐budget paradigm, the
method flouts the fundamental principles of hydrogeological engineering:
Water‐resource scientists are concerned that some basic principles are being overlooked by water managers…. Perhaps the most common misconception in groundwater hydrology is that a water budget of an area determines the magnitude of possible groundwater development. Several well‐known hydrologists have addressed this misconception and attempted to dispel it. Somehow, though, it persists and continues to color decisions by the water‐management community. (Bredehoeft et al. 1982)
According to the U.S. Geological Survey, the water‐budget paradigm “is an oversimplification of the
information that is needed to understand the effects of developing a groundwater system… A pre‐
development water‐budget by itself is of limited value in determining the amount of groundwater that
can be withdrawn on a sustained basis” (Alley et al. 1999).
Balleau and Mayer (1988) explain that “It would be hydrologically inaccurate and economically
inefficient to ignore the transition period and to assume that ground water is only of two types: 100
percent mined or 100 percent recharged”. Nonetheless, policy decision‐makers prefer the method
because its mathematical precision implies expert‐consensus regarding sustainable groundwater
exploitation (Ludwig et al. 1993).
While empirical precision is helpful in winning political buy‐in, public‐interests are not well served
“by adopting an attractive fallacy that the natural recharge rate represents a safe rate of yield” (Balleau
and Mayer 1988).
6 Numerous hydrogeologists have continuously demonstrated the problems with the water budget methodology, dating back to Theis (1940). Others include Lohman 1972; Bredehoeft et al. 1982; Bredehoeft 1997, 2002, 2003, 2006; Bredehoeft and Durbin 2009; Konikow and Bredehoeft 1992; Sophocleous 1997, 1998, 2000; Devlin and Sophocleous 2004; Balleau 1988; Balleau and Mayer 1988; Brown 1996; Alley and Leake 2004; Alley et al. 1999; Alley 2007; Alley and Emery 1986; Konikow 1986.
18
Sophocleous (1997) agreed, “policy‐makers are primarily concerned about aquifer drawdown and
surface‐water depletion, both unrelated to the natural recharge rate. Despite its irrelevance, natural
recharge is often used in groundwater policy to balance groundwater use under the banner of safe yield.
Adopting such an attractive fallacy does not provide scientific credibility.”
Bredehoeft (1997) decried: “Sustainable groundwater development has almost nothing to do with
recharge... However, I continue to hear my colleagues say they are studying the recharge in order to size
a development… The water‐budget as it is usually applied to scale development is a myth—Theis said
this in 1940. Yet the profession continues to perpetuate this wrong paradigm.”
Five years later, Bredehoeft (2002) implored “...the myth goes on; it is so ingrained in the
community’s collective thinking that nothing seems to derail it.”
Devlin and Sophocleous (2005) explained that despite the “conclusive theoretical proof” that the
water‐budget paradigm has no scientific merit, “it still persists”.
In 2008, Milly et al. (2008) explained that climate variability further exacerbates conventional
methods of water management:
Systems for management of water throughout the developed world have been designed and operated under the assumption of stationarity. Stationarity—the idea that natural systems fluctuate within an unchanging envelope of variability—is a foundational concept that permeates training and practice in water‐resource engineering…. In view of the magnitude and ubiquity of the hydroclimatic change apparently now under way… we assert that stationarity is dead and should no longer serve as a central, default assumption in water‐resource risk assessment and planning.
If stationarity is dead—that is, if the rate of precipitation cannot be assumed to fall within a
predictable range for a specified area—then the paramount variable in all of groundwater management
is a fixed uncertainty: “…if the climate is changing, as recent evidence suggests, then the assumption of
equilibrium should be questioned” (Milly et al. 2008).
Bredehoeft and Durbin (2009) explain that if equilibrium of a hydrologic basin is uncertain, then the
predetermination of sustainable rates of development is precluded by the time to full capture problem:
“large systems pose a challenge to the water manager, especially when the water manager is committed
to attempting to reach a new equilibrium state in which water levels will stabilize and the system can be
maintained indefinitely.” These attempts are mired by two realities: “(1) a large groundwater system
creates a delayed response between the observation of an impact and its maximum effect and (2) there
is a long time lag between changing the stress and observing an impact at a distant boundary.”
19
Future pumping stress projections
One particularly conspicuous assumption in the 1983 version of the USGS model—and thus,
Peabody’s 1985 PHC—illustrates the model’s conceptual uncertainty and the modeler’s inaccurate
pumping projections. Eychaner (1983: 1) explained: “By 1990… municipal supply pumpage is expected
to exceed pumpage at the mine”. In reality, however, municipal withdrawals would never overtake
Peabody withdrawals while the Black Mesa‐Kayenta mine was fully operational.
Further, it is imperative to recognize that municipal withdrawals are distributed throughout Black
Mesa’s 5,400 mi2 area (in both confined and unconfined conditions) whereas Peabody withdrawals are
centralized at a single area within confined conditions.
Figure 4. Total industrial and municipal withdrawals from the N‐aquifer, 1968‐2008. Withdrawal data are from Macy (2010).
WATER QUANTITY
Peabody’s PHC concludes that the magnitude of the mine’s groundwater withdrawals would have a
“substantial” impact on the areal extent of drawdown and the cone of depression would spread “over
considerable distance” (PCC‐PHC 1985: 37, 39). However, the impacts from the proposed extension of
the mining period at the tribal communities would be negligible and confined to the mining area only:
‐
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
5,000
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
Comparison of Peabody and Municipal Withdrawals, 1968‐2008
Municipal (Confined & Unconfined) Peabody (Confined)
20
The USGS model assumed a PWCC pumping rate of 3,700 AF/y… and that PWCC pumpage would cease in 2001; whereas the current coal supply agreements run through 2007 and 2011 for the Black Mesa and Kayenta mines, respectively. Even considering this difference, it is not anticipated that the Peabody induced drawdowns would increase much… the amount of additional drawdown from the longer pumpage would only be on the order of 10 to 20 feet near the leasehold. This would cause negligible increases in the percent of available water height losses. (PCC‐PHC 1985: 39)
The EIS, CHIA, and EIS provide testable water‐level predictions for the communities of Kayenta, Tuba
City/Moenkopi, Forest Lake, and Pinon (Table 1). However, many of the predictions are not presented
in a uniform manner; comparisons are made where possible.
Table 1. Predicted and Actual Rates of N‐Aquifer Drawdown in 2001
Prediction data for 2001 from PCC‐PHC (1985). Actual water level data for 2001 is from Thomas (2002).
21
Kayenta
The significance of tribal community pumpage should not be understressed. Eychaner states that in areas of large community pumpage, little of the simulated decline is caused by mining. At Kayenta, over 85% of the total water level declines will be caused by community pumpage.
Peabody’s 1985 determination of Probably Hydrologic Consequences
(PCC‐PHC 1985: 46)
If a confined aquifer’s potentiometric surface is lowered to or below the aquifer’s top, it is possible that he aquifer matrix will compress (collapse) causing general surface subsidence and irreversibly changing the aquifer’s water producing characteristics.
OSM’s 1989 Cumulative Hydrologic Impact Assessment
(OSM‐CHIA 1989: 5‐5)
For there to be a reduction in well production or for structural damage to occur… the potentiometric surface would have to be drawn down to below the top of the confined portion of the aquifer… It can be seen that at no time does the potentiometric surface drop to this level anywhere within the affected area for any scenario. The closest the potentiometric surface gets to the top of the confined aquifer for scenario B is 366 feet at Keams Canyon in the year 2052.
OSM’s 1990 Final Environmental Impact Statement
(OSM‐EIS 1990: IV‐28)
Since mining operations commenced in 1968, there has been no consensus regarding the
hydrogeological dynamics of the N‐aquifer near the Diné community of Kayenta, north of the
mine. In fact, in 1971 the Bureau of Reclamation explained: “Hydrologists do not agree whether
these domestic wells [in Kayenta] are in the same pressure zones as the Peabody wells, but a
monitoring program has been devised to ascertain those facts” (USBR 1971: 39).
The EIS only provides the predicted water level at Kayenta in 2007; no other prediction data exist.
Page IV‐30 (OSM‐EIS 1990) states that drawdown from the pre‐stress period to 2007 would be 58 feet7.
Between 1965 and 2007, the water‐level at two wells in the immediate Kayenta area have been
tracked by the USGS monitoring program. By 2007, the well BM3 had fallen 106.1 feet (Truini and Macy
2008; BMMP 2010). Drawdown was underestimated by 48.1 ft, or 82.9%. Decline at USGS observation
well BM2, which is 7.5 miles southeast of Kayenta, was underestimated by 56%.
7 A significant incongruity exists here: Peabody’s 1985 PHC predicted that, by 2001, water level at Kayenta would decline by 110 ft (PCC‐PHC 1985: 40). Five years later, OSM’s 1990 EIS—which includes six additional years of Peabody withdrawals—states that, in 2007, water level decline at Kayenta would be 58 ft (OSM‐EIS 1990: IV‐30). Because the PHC provided the information for the 1989 CHIA which in turn provided the technical information for the 1990 EIS, the nearly 50% change in the drawdown prediction is a significant and unexplained inconsistency, and demonstrates the uncertainties and errors in the groundwater model.
22
Figure 5. Kayenta groundwater‐levels, BM3 (top), and BM2, (bottom; BMMP 2010)
No measurable recovery has occurred at Kayenta.
OSM’s material damage threshold for structural stability of the N‐aquifer is potentiometric surface
(i.e. head) no lower than 100 ft above the top of the aquifer (OSM‐CHIA 1989). In 2005, the USGS
reported the water level at Kayenta well BM3 was approximately 3.6 feet below the top of the aquifer
itself (Truini and Macy 2007), thus the damage threshold has been exceeded by 103.6 ft. Because the N‐
aquifer is no longer saturated at Kayenta, it is, theoretically, vulnerable to compaction.
Water Level in BM3 and Kayenta Withdrawals
The regression of BM3 water levels to Kayenta withdrawals (1984‐2008) demonstrates a weak linear
relationship. However, as the volume of Kayenta’s withdrawals increase, the water‐level in BM3 does
not fall, it rises: r = 0.‐44; R2 = 0.19; p = 0.047 (Figure 6, top chart).
It is hypothesized that the magnitude of Peabody’s withdrawals simply obscures Kayenta’s effect on
BM3 (Figure 7).
It is notable that the data point representing Kayenta’s smallest volume of groundwater withdrawn
in any single year (381 acre‐feet in 2008; Figure 6, top chart, circled in blue) correlates with the BM3’s
largest water level decline (161.9 ft below land surface).
Conversely, Kayenta’s largest withdrawals (708 af in 1987 and 690 af in1988) correlate with the
some of the well’s highest water levels (131 and 135 ft below land surface, respectively).
23
Figure 6. Comparison of Water Level in BM3 and Kayenta withdrawals8, 1984‐2008
8 In the top chart, the value for water level in BM3 (y‐axis) is shown as feet below land surface, so as the value increases, the water level is falling. Kayenta withdrawal data are collected from the USGS Black Mesa Monitoring Program’s annual reports, 1984‐2010; withdrawal data are not available for water‐years 1992, 1993, 1997, 2006. Water level data collected from BMMP (2011).
2008
24
Figure 7. Peabody and Kayenta withdrawals, 1984‐2005. Withdrawal data from Macy (2010).
Water Level in BM3 and Peabody Withdrawals 9
Because the record for withdrawals from the Kayenta well system begins in 1984, the regression of
BM3 water levels to Peabody withdrawals is performed for the same period (Figure 8, top chart). The
regression conflicts with OSM’s model‐based conclusion: a relatively strong, linear relationship exists: as
the Peabody’s withdrawals increase, the water‐level in BM3 falls: r = 0.75; R2 = 0.56; p < 0.0001.
Figure 8 (bottom chart) demonstrates this relationship temporally for the period (1984‐2005).
Data for water‐year 1985 (circled in red) is a strong indicator of the mine’s influence on Kayenta’s
water level. Recall that, in 1985, Peabody ceased its groundwater withdrawals pipeline operations for
six months due to maintenance at the Mohave Generating Station, withdrawing its smallest volume of in
any single year (2,520 acre‐feet). For the period that the Black Mesa‐Kayenta Mine was fully
operational, 1985 is a statistical outlier for Peabody withdrawals – the only year between 1972 and 2005
that the mine ceased groundwater pumping (water years 2006‐2008 cannot be included in the analysis
due cessation of groundwater withdrawals at the Black Mesa Mine. See Appendix A for an explanation
of these outliers. It is likely that Peabody’s decrease in withdrawals beginning in 2006 do not yet appear
as recovery in BM3 given the principle of superposition).
For the period 1984‐2005, water‐year 1985 expresses the highest water level in BM3 for any single
year (120 ft. below land surface) correlating with Peabody’s lowest volume of groundwater withdrawals.
9 The value for water level in BM3 is shown as feet below land surface, so as the value increases, the water level is falling. Withdrawal data collected from the USGS Black Mesa Monitoring Program annual reports, 1984‐2010. Peabody withdrawals for water years 2006‐2008 are not included in the analysis because they are statistical outliers.
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
1984
1985
1986
1987
1988
1989
1990
1991
1994
1995
1996
1998
1999
2000
2001
2002
2003
2004
2005
Comparison of Peabody and Kayenta Withdrawals, 1984‐2005
Peabody withdrawals avg. 4,085 af/y
Kayenta withdrawals avg. 569 af/y
25
Figure 8. Comparison of Water Level in BM3 and Peabody Withdrawals, 1984‐2005
1985
26
Indicators using box‐plots with outliers10
The box‐plot of Peabody withdrawals (Figure 9, top box‐plot) demonstrates that water year 1985
(highlighted in red) is a statistical outlier for the period 1972‐2005.
Concurrently, the box‐plot of water levels in BM3 (middle box‐plot) demonstrates that 1985 is the
only year outside 2 standard deviations of the mean water level (1984‐2008).
Finally, the box‐plot for Kayenta withdrawals (bottom box‐plot) shows that 1985 is near the mean
for withdrawals, and moreover, in 2008, Kayenta withdrew 381 acre‐feet (a statistical outlier) yet BM3
fell to its lowest level on record, which is inexplicable if Kayenta is causing 87% of the drawdown.
Figure 9. Box‐plots: Peabody withdrawals; BM3 water level; and Kayenta withdrawals
10 Peabody withdrawal data from Macy (2010); BM3 water level data from BMMP (2010); Kayenta withdrawal data collected from the USGS monitoring reports for the period 1984‐2010.
1985
1985
2008 1985
2008
27
OSM’s conclusion on the source of Kayenta’s water level decline
In 1999, Peabody’s hydrology consultants argued that “The drawdown at BM3 appears to be almost
entirely related to local pumping at Kayenta and is similar in all [simulated pumping] scenarios”
(HSIGeoTrans and WEHE 1999: 6‐16).
In 2005, actual conditions diverged from the simulations, OSM then explained:
Approximately 423 acre‐feet were withdrawn from the N‐aquifer at this location by the BIA and Navajo Nation wells [i.e. the Kayenta well system] in 2004. The Kayenta municipal pumping currently accounts for 34% of all the non‐industrial pumpage from the confined portion of the aquifer and approximately 15% of all municipal production from the N‐aquifer (confined and unconfined). Therefore, lowering of potentiometric surface in the area near BM3 can be attributed to municipal use as well as PWCC pumping. (OSM 2006: 5)
OSM seems to be implying that, in 2004, the magnitude of Kayenta’s withdrawals (423 acre‐feet) is
profound because it accounts for 34% of 1,240 acre‐feet comprising total tribal withdrawals (it is notable
these withdrawals are distributed throughout the entire confined portion of the N‐aquifer).
This implication lacks empirical reasoning; arguably, it lacks understanding of basic hydrogeological
principles, for in 2004, Peabody withdrew 4,370 acre‐feet from one centralized location in the confined
N‐aquifer – more than ten times the volume of Kayenta’s 423 acre‐feet.
In 2005, OSM also concluded that the material damage criterion for structural stability “is subject to
review and modification”, but that it would continue to be used for OSM’s 2005 evaluation (OSM 2006:
4). That year, BM3 was 3.6 ft below the top of the N‐aquifer (Truini and Macy 2007).
To clarify, in 1989, Peabody’s PHC concluded that 85% of the drawdown at Kayenta would be caused
by municipal withdrawals (PCC‐PHC 1985), and in 1990, OSM concluded that “For there to be a
reduction in well production or for structural damage to occur… the potentiometric surface would have
to be drawn down to below the top of the confined portion of the aquifer… It can be seen that at no
time does the potentiometric surface drop to this level anywhere within the affected area for any
scenario. The closest that the potentiometric surface gets to the top of the confined aquifer for scenario
B [i.e. the proposed pumping rate] is 366 feet at Keams Canyon in the year 2052” (OSM‐EIS 1990: IV‐28).
By 2005, OSM’s threshold for material damage to the N‐aquifer’s structural stability had been
exceeded by 103.6 ft., and the potentiometric surface had fallen below the top of the aquifer.
Nonetheless, OSM concluded: “material damage to the hydrologic balance of the N‐aquifer, caused by
mining, with respect to maintaining the potentiometric head above the top of the N‐aquifer, has not
occurred” (OSM 2006: 7).
28
Tuba City / Moenkopi Drawdown at Tuba City/Moenkopi is due entirely to community withdrawals. Total decline from the pre‐stress period to 2007 will be 51 feet: “OSM concludes that under all alternatives there would be negligible short and long‐term impacts in the Tuba City/Moenkopi area due to mine‐related pumping” (OSM‐EIS 1990: IV‐29).
USGS monitors six wells within a five mile radius of Tuba City and Moenkopi. However, municipal
withdrawals occur at three clustered Navajo Tribal Utility Authority (NTUA) wells. From the pre‐stress
period through 2008, these wells declined by 31 ft, 27.7 ft, and 27.4 ft (Figure 10). Mean decline is 28.7
ft; drawdown was overestimated by 44%.
Three other wells monitored by USGS are located east and south of Moenkopi. The water level in
BIA wells 3K‐325 and Tuba City Rare Metals increased 6.2 ft and 4.9 ft for the period, while BIA 3T‐333
declined 6.3 ft (Truini and Macy 2008). The mean water‐level for the three wells is +1.6 ft. The
predicted drawdown of 51 ft was overestimated by 103% (Figure 11). Thus, at Tuba City / Moenkopi,
the model overestimated the impact on municipal withdrawals on water level by 44% to 105%; mean
drawdown of the six wells is 13.6 ft, a 73% overestimation.
29
Figure 10. Tuba City NTUA 1 (top), 3, (middle), and 4 (bottom. BMMP 2010)
30
Figure 11. Tuba City BIA 3K‐325 (top), 3T‐333 (middle), and Rare Metals (BMMP 2010)
31
No correlation between Tuba City withdrawals and Tuba City water levels
Because the actual water‐level decline was overestimated (44% in three NTUA wells, 103% in three
BIA wells, and 73% in six area wells), it is reasonable to assume that the model’s projection for municipal
withdrawals was overestimated by an approximately equivalent amount.
Municipal withdrawals from the Tuba City well‐system were recorded in 23 of 24 years for the
period 1985 – 2008 (withdrawal data were unavailable in 2006). For this period, projected withdrawals
totaled 27,787 acre‐feet, for an average of 1,208 af/y 11 (Figure 12).
Actual withdrawals for the same period totaled 24,730 acre‐feet, averaging 1,075 af/y. Actual
withdrawals were overestimated by a total of 3,057 acre‐feet, or 133 acre‐feet per year (11%).
Thus, assuming a linear relationship between withdrawals and water‐level, the water‐level decline
at Tuba City (overestimated by 44%‐105%, or mean 73%) does not correlate with an equivalent
overestimation of municipal withdrawals (11%).
Figure 12. Municipal Withdrawals from the Tuba City Well‐System, 1985‐2008.
11 In 1985, the Tuba City well‐system withdrew 906 acre‐feet, and the model assumed that municipal withdrawals
would increase annually by 2.5%.
32
Discharge from Moenkopi School Spring and to Moenkopi Wash
Because changes in water‐level correlate with change in spring‐discharge, the overestimated water‐
level decline should correlate with an overestimated decline in spring‐discharge (the actual decline at
Moenkopi School Spring should be less than the predicted 1‐2% decline).
Peabody’s 1985 PHC addresses mine‐related impacts to N‐aquifer springs in a single paragraph:
Outflows to springs and streams from the N‐aquifer are projected by the USGS model to decline 150 acre‐feet due to Peabody pumpage and 330 acre‐feet because of Tribal community pumpage. Overall, this will amount to about a six percent reduction in stream and spring flow. Although the projected declines are small, it is unlikely that they will ever return to pre‐1965 flow levels following cessation of the Peabody pumpage because Tribal community pumpage will continue and is presently causing a majority of the flow reductions. (PCC‐PHC 1985: 46)
Subsequently, OSM’s 1989 CHIA concluded:
No material damage to spring discharge to the hydrologic balance [sic] is projected to
occur for N‐aquifer spring discharge. (OSM‐CHIA 1989: 7‐5)
As well, OSM’s 1990 Final EIS concluded:
Spring discharges in Pasture Canyon would not change as a result of mine‐related withdrawals. Simulated outflow from the N‐aquifer to Moenkopi Wash, west of Black Mesa through Blue Canyon to Moenkopi, would decrease by 1 to 2 percent under all pumping scenarios. Community pumping would have a slightly greater effect on the outflow to Moenkopi Wash than would varying the duration of pumping at the mine; therefore, the short‐ and long‐term impact of the mine on Moenkopi Wash baseflow discharge from the N‐aquifer would be negligible. (OSM‐EIS 1990: IV‐28)
The USGS model and N‐aquifer discharge
It is imperative to recognize that, when the USGS model was developed, the basic components of
the N‐aquifer’s water budget were (and continue to be) highly uncertain: the water‐budget could not be
calculated because inflow and outflow were uncertain; vertical leakage was ignored; further studies
were needed to verify assumptions; and as with all groundwater models, future precipitation and
withdrawal rates were unknowable. Most data were “not available for most nodes. Some data may be
incorrect or be subject to multiple interpretations” (Eychaner 1983). Divergence between simulated
and actual water levels was expected to occur, and the USGS explained that “the accuracy of projected
water‐levels is uncertain” (Brown and Eychaner 1988). Finally, the USGS explained that the model
simulations “are better used to compare effects of different development plans rather than estimate
actual future water levels and water‐budget components” (Brown and Eychaner 1988).
33
To clarify, the USGS groundwater model is not capable of accurately predicting water level at any
particular well, it is incapable of predicting spring discharge from any particular spring, and it is
incapable of predicting outflow to streams (Brown and Eychaner 1988; Eychaner 1983, 1981).
This recognition reflects the uncertainties that continue to characterize current knowledge of the N‐
aquifer, for in 1999 Peabody’s own hydrology consultants explained that, because spring and stream
discharge is poorly understood, “a regional scale model cannot currently be developed for the basin that
will accurately predict the impacts of pumping on individual springs” (HSIGeoTrans and WEHE 1999: 5‐
23). Thus, they concluded that “developing a qualitative estimate of the discharge (e.g., 0.1 cfs), or
absolute change in discharge, is not feasible” (HSIGeoTrans and WEHE 1999: 5‐24). In fact, no
groundwater model ever developed for the N‐aquifer can accurately simulate spring or stream discharge
from the N‐aquifer (HSIGeoTrans and WEHE 1999; Brown and Eychaner 1988; Eychaner 1983, 1981).
The relationship between springs and water level
Spring discharge is reduced when the potentiometric surface (i.e. head) of an aquifer declines.
Recall that any groundwater withdrawn from an aquifer alters its equilibrium and, initially, 100% of the
withdrawals will come from storage. Over time, however, given the Principle of Superposition, the
groundwater withdrawn from the aquifer will reduce its potentiometric surface; aquifer discharge is
affected as the cone of depression radiates out and away from the pumping source and reaches
discharge locations. As withdrawals slowly become balanced by the (1) reduced discharge, (2) induced
recharge, or (3) some combination of both (this is known as “capture”). New equilibrium is reached
when the capture is equal to the amount of groundwater withdrawn from the system (Bredehoeft 2002,
1997; Balleau and Mayer 1988; Bredehoeft et al. 1982; Theis 1940).
“Discharge from springs is proportional to potentiometric head. An analysis of groundwater
quantity is basically a determination of where and how much the potentiometric head would change as
a result of imposed activities… and the corresponding changes in spring discharge rates resulting from
the head changes” (OSM‐EIS 1990: IV‐20).
N‐aquifer Springs
Four springs are monitored by the USGS monitoring program: (1) Moenkopi School Spring, (2) Burro
Spring, (3) Pasture Canyon Spring, and (4) Unnamed Spring near Dennehotso (Figure 13).
34
Figure 13. Map of spring monitored by the USGS (in Truini and Macy 2006)
Peabody’s PHC, and OSM’s 1989 CHIA and 1990 EIS explain that spring discharge in the area of
Moenkopi may be reduced by 1‐2%; however, this would be caused entirely by municipal withdrawals
from the Tuba City well system. In all five of the groundwater model’s pumping scenarios, mine‐related
withdrawals would have no effect on N‐aquifer springs (OSM‐EIS 1990: IV‐28).
Furthermore, the USGS groundwater model projects that the nearest measureable impact related
to Peabody’s withdrawals (1 foot of water level decline) comes no closer than fifteen from Tuba
City/Moenkopi and does not occur until the year 2052.
35
CHIA criterion for material damage to spring discharge
Based on the simulation results of the groundwater model, OSM developed the material damage
criterion for discharge from springs: discharge from springs must not be reduced by more than 10% in
response to mine‐related withdrawals; potentiometric head must not fall below 100 ft. from the top of
the N‐aquifer so that the average rate of discharge from springs, minus the measurement error of 10%,
will be maintained (OSM‐CHIA 1989: 5‐5).
Evaluation of the criterion
According to OSM, the evaluation of the material damage criterion for springs “is based on
computer simulation of changes in areas of hydraulic head near the confined/unconfined boundary of
the N‐aquifer, not specific spring discharge points” (OSM 2006: 8). To clarify, OSM does not use actual
spring discharge monitoring data in its evaluation. Rather, it relies upon simulations of the USGS
groundwater model—which is incapable of simulating discharge from springs—to make its conclusions.
Further, it is important to recognize that the USGS groundwater model itself is not updated with
each year’s monitoring data, it is not run annually to evaluate each year’s simulated impacts, and the
model’s performance is not evaluated on a regular schedule (the groundwater model is not a
responsibility, per se, of the hydrologists monitoring the N‐aquifer). Rather, the model is updated and
tested by USGS modelers when the funding is available to do so (recall that USGS applies to the Bureau
of Indian Affairs every year to fund the monitoring program itself). Thus, OSM’s annual evaluation of
material damage criteria is based upon the most recent simulation, whenever it occurred12.
To clarify, because the model cannot simulate spring discharge, the damage threshold for springs is
based upon by the models most recently run simulations of change in hydraulic head (this study has
already demonstrated the models poor prediction‐accuracy in regard to head).
12 OSM’s CHIA criterion for evaluating N‐aquifer discharge to streams (OSM‐CHIA 1989) has never been evaluated using the method original designed by OSM. Recall that the evaluation of this criterion is based upon the results of the simulation model; moreover, if the indicator‐threshold is crossed, OSM requires further studies to discern the cause of the decline before corrective actions are taken (OSM‐CHIA 1989; OSM‐CHIA Addendum 1990). OSM’s most recent evaluation of the N‐aquifer was performed in 2006, and the agency explained that the model’s most recent simulation‐run occurred in 1994. Thus, in 2006, OSM’s evaluation of, and conclusions regarding groundwater discharge to streams were based upon a simulation of conditions in 1994 (OSM 2006; of course, the evaluation also assumes that, technically, the 1994 model‐simulation was accurate and, theoretically, that the uncertainties in natural systems can be replaced by the “certainty” of model‐simulations).
36
Through 2006, the USGS monitoring reports used logarithmic charts (Figure 14) to illustrate spring‐
discharge from the N‐aquifer, and no appreciable trends were apparent through that time.
Concurrently, the 2006 OSM report concluded that material damage to the N‐aquifer springs had not
occurred in relation to mine‐related withdrawals (OSM 2006; also see OSM‐EIS 2004, 2006).
Figure 14. USGS’s logarithmic chart of spring discharge (in Truini and Macy 2006)
It was later revealed to USGS hydrologists working with Black Mesa Monitoring Program, however,
that the logarithmic charts unintentionally concealed negative trends (In logarithmic charts, the unit of
measure, circled in red (Figure 14) increases exponentially rather than incrementally. As such, if the
measured value has small changes within the higher range of values, as in the case of Moenkopi School
Spring, trends may be inadvertently concealed, which is evident when comparing discharge data for
Moenkopi School Spring in an incremental chart, Figure 15, below). Subsequently, USGS began
illustrating spring‐discharge using incremental charts (Truini and Macy 2008, 2007; Macy 2010, 2009).
Moenkopi School Spring
Recall that Moenkopi School Spring is located near the Hopi village of Lower Moenkopi, along the
bank of Moenkopi Wash (Figure 15), approximately 60 miles downstream of the mine, and it is in the
western‐most portion of the unconfined N‐aquifer. According to Peabody’s consultants, “Drawdown at
Tuba City is solely caused by pumping from local community wells… There are no impacts at Tuba City
from PWCC pumpage in either the 2D or 3D model” (HSIGeoTrans and WEHE 1999: 8‐7 and 8‐8).
Figure 17 demonstrates that the rate of discharge (in gallons per minute) varied throughout the
period of record (1987‐2009): the highest discharge (16 gpm) occurred the first year of record and
lowest discharge (8.0 gpm) occurred during the most recent year of record). Mean discharge for the
period is 11.81 gpm, which is 26.2% lower than the 1987 rate (exceeding the CHIA criterion by 16.2%).
37
Figure 15. Discharge from Moenkopi School Spring, 1987‐2008 (data from Macy 2010).
The regression of discharge from Moenkopi School Spring to Peabody withdrawals13 (Figure 16)
demonstrates a strong, indirect, linear relationship between decreasing discharge and increasing
Peabody withdrawals: r = ‐0.84; R2 = 0.71; p < 0.0001.
Figure 16. Regression of Moenkopi School Spring discharge and Peabody Withdrawals
13 Analysis includes the period 1987‐2005. Spring discharge measurements began in 1987. Peabody withdrawals for water years 2006, 2007, 2008 are statistical outliers and are not included in the regression (Peabody reduced its withdrawals 70% beginning in 2006 due to closure of MGS, Black Mesa Pipeline, and Black Mesa Mine). See Appendix A for the statistical explanation for the outliers.
R² = 0.5475
6
8
10
12
14
16
18
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Moenkopi School Spring‐Discharge, 1987‐2009
n = 15r = ‐0.84R² = 0.71p < 0.0001
3,600
3,800
4,000
4,200
4,400
4,600
4,800
9 10 11 12 13 14 15 16 17
Peabody Wihtdrawals (af/y)
Moenkopi School Spring Discharge (gpm)
Regression of Spring Discharge and Peabody Withdrawals
38
To clarify, with the data that are available, 71% of the variability in discharge from Moenkopi School
Spring (a declining trend) can be accounted for on the basis of the variability in Peabody withdrawals (an
increasing trend). The null hypothesis (Peabody withdrawals do not affect discharge from Moenkopi
School Spring) is rejected because the chance that the demonstrated relationship (R2) does not actually
exist is less than 0.01% (p).
Although OSM has not evaluated the material damage criteria since 2006 (water‐year 2005), its
conclusion that mine‐related groundwater withdrawals have had no effect on N‐aquifer springs has not
changed (OSM‐CHIA 2008).
Moenkopi School Spring and Tuba City Withdrawals
Moenkopi School Spring is within five miles of the wells that comprise the Tuba City well system.
The regression of spring discharge to Tuba City withdrawals (Figure 17) demonstrate that, for the period
of record, there is no statistically significant relationship between spring discharge and local
withdrawals: r = ‐0.30; R2 = 0.09; p = 0.28.
Figure 17. Regression of Moenkopi School Spring discharge & Tuba City Withdrawals
To clarify, with the data that are available, 9% of the variability in discharge from Moenkopi School
Spring (a declining trend) can be accounted for on the basis of the variability in Tuba City withdrawals
(an increasing trend). However, the null hypothesis (Tuba City withdrawals do not affect discharge from
Moenkopi School Spring) is not rejected because there is a 28% chance (p) that the demonstrated
relationship (R2) occurred by chance.
R² = 0.0893
80085090095010001050110011501200125013001350
8 9 10 11 12 13 14 15 16
Tuba City Withdrawals (af/y)
Moenkopi School Spring Discharge (gpm)
Regression of Spring Discharge and Tuba City Withdrawals
39
Moenkopi School Spring and Tuba City Precipitation
The regression of discharge from Moenkopi School Spring and local precipitation at Tuba City (Figure
18) demonstrate that there is no statistically significant relationship between discharge and rainfall in
the area: the null hypothesis (Precipitation at Tuba City does not affect discharge from Moenkopi School
Spring) is not rejected because there is a 17% chance (p) that the demonstrated relationship (R2)
occurred by chance.
Figure 18. Regression of Moenkopi School Spring discharge & Tuba City Precipitation (precipitation data for the Tuba City weather station from the Desert Research Institute: http://www.wrcc.dri.edu/).
OSM on Precipitation and Spring Discharge
OSM also found no discernable relationship between precipitation and discharge in the Tuba
City/Moenkopi area (OSM‐CHIA 1989). In fact, OSM found an inverse relationship: the (pre‐mining)
period of higher precipitation correlated with low‐flow characteristics while the (post‐mining) period of
low precipitation correlated with high‐flow characteristics:
The inconsistencies noted above indicate that other factors which influence run‐off within the basin effectively mask any consistent relation between precipitation on Black Mesa and discharge at Moenkopi Wash at Moenkopi, at least for the relatively short period of record analyzed. (OSM‐CHIA 1989: 3‐13)
R² = 0.1148
0
2
4
6
8
10
12
7 8 9 10 11 12 13 14 15 16 17
Precipitation (in)
Moenkopi School Spring Discharge (gpm)
Regression of Spring Discharge and Tuba City Precip
40
Forest Lake The maximum water‐level decline at Forest Lake will occur in 2007; it will be 40 ft. below the 1985
water‐level. Water levels will recover to the 1985‐level in 2009. Total decline from the pre‐stress period
to 2007 will be 148 feet. “OSM concludes that the impacts in the area due to mine related pumping
would be minor over both the short and the long‐term” (OSM‐EIS 1990: IV‐29).
Forest Lake (BIA well 4T‐523, completed in 1980) is the closest community to the Peabody lease‐
area. Directly south of the mine (see Figure 1), it was expected to incur the greatest mine‐related water
level decline of all the communities on Black Mesa. Given proximity to the mine and similar
hydrogeological characteristics, it is intuitive that the model‐predictions for Forest Lake would be more
accurate than distant communities with different hydrogeological characteristics; large discrepancies
would be indicative of significant model errors.
The USGS model estimated the pre‐stress water‐level at the Forest Lake well at 1,096 ft. below land
surface (Macy 2010); by 1985, the water level had declined 28 ft. (1,124 ft. below land surface).
However, due to the closure of the Mohave Generating Station, the withdrawal reduction in the model
simulation occurred at the end of 2005 rather than 2006: the aquifer had one full year to recover in
comparison to the simulation. With the extra full‐year of recovery, maximum water‐level decline should
have occurred prior to 2007 (as predicted), should be less than the model predicted, and recovery to the
1985‐water‐level should have occurred earlier than 2009 (as predicted). A significant lack of recovery by
2009 would indicate error in the conceptual model.
In 2007, the water‐level at Forest Lake was 1,189 ft. below land surface (65 ft. lower than the 1985
water‐level; see BMMP 2010; Truini and Macy 2008). While the predicted 40 ft. decline between 1985
and 2007 was off by merely 25 ft, this is an underestimation of mine‐related impacts at Forest Lake by
62.5%. Because the Forest Lake well (NTUA1) was only completed in 1980, water‐level decline from the
pre‐stress period to 1984 was estimated by the model at 108 ft (Brown and Eychaner 1988: 41).
However, USGS recorded total drawdown for the period 1965 to 2007 at 93.8 ft.
Because 65 ft of the total decline occurred between 1984 and 2007, only 28.8 ft of drawdown at
Forest Lake occurred prior to 1984. Thus, the model estimate of 108 ft decline occurring prior to 1985
was an overestimation of the actual drawdown that occurred prior to 1985 of 73.3%, expressing
significant conceptual or input data error.
41
Figure 19. Groundwater‐levels at Forest Lake NTUA1 (BMMP 2010)
Recovery at Forest Lake
Water‐level at Forest Lake did not begin to recover following the early reduction in 2006. In fact,
water‐levels remained approximately 1,190 ft below land surface through 2008. In 2009—the year
water‐level was predicted to recovery to the 1985‐level—head at Forest Lake had recovered only 3 of the
predicted 40 ft (or actual 66 ft) decline. Given the variation of water‐level at Forest Lake over time, at
the time of this study, it is not yet apparent of maximum decline has been reached or if any recovery has
occurred.
42
Pinon Drawdown at Pinon will be similar to Forest Lake, but community withdrawals will have a greater effect.
Maximum water level will decline will be in 2007, being 34 feet below the 1985 levels; total decline from
the pre‐stress period to 2007 will be 106 feet. “OSM concludes that impacts in the Pinon area from
mine‐related pumping would be minor over both the short and long‐term” (OSM‐EIS 1990: IV‐29).
The pre‐stress water level at Pinon was 743 ft below land surface; in 1985 it was 803.3 ft below land
surface (a decline of 60.3 ft). In 2007, the water‐level was measured at 904.3 ft below land surface, a
decline of 101 ft from the 1985‐level: drawdown for this period was underestimated by 67 ft (97%).
Between 1965 and 2007, water‐level declined 161.3 ft, an underestimation of 55.3 ft (52%).
Figure 20. Groundwater‐levels at Pinon well PM6 (BMMP 2010)
Recovery at Pinon
Water‐level continued to decline after the reduction of industrial withdrawals. In 2008 and 2009,
water‐levels were measured at 904.9 ft below land surface (BMMP 2010; Macy 2009). At the time this
study was conducted, it was unclear whether or not the water‐level at Pinon had yet reached its
maximum decline or if recovery had initiated.
43
GROUNDWATER QUALITY
The sandstone aquifers supplying water to the Peabody wells contain a vast quantity in
storage, which is isolated from water in the relatively shallow local wells on the mesa by
impermeable formations overlying the Navajo, Kayenta, and other deep formations.
United States Bureau of Reclamation (USBR 1971)
Environmental Statement for the Black Mesa‐Kayenta Mine
Effects of induced leakage of poorer quality water from the overlying D‐aquifer system on N‐aquifer water quality. The water quality from Peabody and Tribal wells completed in the N‐aquifer has been periodically monitored by the USGS since 1967. The thrust of the N‐aquifer water quality monitoring effort has been towards assessing if vertical leakage from the overlying D‐aquifer system is significant. The concentration of dissolved solids, chloride, and sulfate ions in the D‐aquifer is about 7 times, 11 times, and 30 times greater, respectively, than in the N‐aquifer. If the N‐aquifer water level declines are inducing large amounts of vertical leakage from the D‐aquifer system, there should be marked changes with time in these parameter concentrations.
Peabody Western Coal Company (PCC‐PHC 1985: 46) Determination of Probably Hydrologic Consequences
Peabody’s PHC concludes that “there is no evidence to suggest that significant vertical leakage is
occurring from the D‐aquifer system into the N‐aquifer system” (PCC‐PHC 1985: 46). OSM subsequently
generated the material damage criterion for water quality: a value of leakage from the overlying D‐
aquifer, caused by mine‐related withdrawals, is not to exceed 10%.
Like the criterion for structural stability and spring discharge, the USGS groundwater model
simulations provide the basis for the evaluation of water‐quality: the estimated pre‐stress leakage from
the D‐aquifer was estimated at 180 acre‐feet per year. The simulation of municipal pumping only
(Scenario E) shows leakage of 239 acre‐feet per year, and the simulation of municipal and proposed
industrial pumping (Scenario B) shows leakage increasing to 243 af/y. Thus, because mine‐related
groundwater pumping would increase leakage from the overlying D‐aquifer by merely 4 af/y, “any effect
on water quality would be negligible due to a 2 million to 1 dilution and no material damage to water
quality would occur” (OSM‐CHIA 1989: 7‐4).
OSM reiterates their findings in the Final‐EIS: “This is further substantiated by the 1989 USGS
monitoring program progress report (Hart and Sottilare 1989) which concludes that no impacts are
observable from leakage of the D‐aquifer to the N‐aquifer. OSM concludes that the potential for
degradation… is considered to be minor over both the short and the long‐term” (OSM‐EIS 1990: IV‐34).
44
OSM upheld this conclusion in 2006: “material damage to the N‐aquifer, caused by mining, with respect
to leakage from the overlying D‐aquifer, has not occurred” (OSM 2006).
However, OSM offers no explanation regarding exactly how the criterion for chemical quality will be
evaluated. The USGS model is a mathematical flow model; it was not developed in consideration of any
chemical quality parameters, and it has no hydrogeochemical capacity.
Rather, OSM predetermined that mining will not diminish the water‐quality because all five of the
model’s pumping scenarios showed negligible volumes of leakage from the overlying D‐aquifer. In
short, because the model predetermined that mining will not induce leakage, the mine will not harm the
N‐aquifer’s chemical quality. Thus, any future evaluation of the criterion is unnecessary.
Discussion
The quantification of leakage between aquifers that are thousands of feet deep requires intensive
investigation into vertical hydraulic conductivities between the aquifers, water measurements from the
aquifers, and other hydrogeological information that are not available. Moreover, the D‐aquifer is not
monitored and investigation of its leakage characteristics did not begin until 2003 (Truini and
Longsworth 2003; Truini and Macy 2005).
OSM acknowledges, “Neither the USGS nor PWCC monitors water‐levels in the confined portion of
the D‐aquifer as part of its monitoring effort. D‐aquifer water‐level information would be needed to
directly evaluate the change in leakage from the D‐aquifer to the N‐aquifer” (OSM 2006: 7).
In review of USGS monitoring data, two wells (Rough Rock PM5 and Keams Canyon PM2) have TDS‐
levels exceeding the EPA’s recommended drinking water limit, and “appreciably higher levels of
chloride” at 97 mg/L and 113 mg/L, respectively (OSM 2006).
In 2007, the concentration of arsenic in Keams Canyon PM2 was measured at 40.3 μg/L, exceeding
the EPA’s standard for Maximum Contaminant Level (MCL: 10 μg/L) (Truini and Macy 2007).
The USGS monitoring report for 2007‐2008 (Macy 2009) illustrates significant increasing trends TDS,
Chloride, and Sulfate at Moenkopi School Spring (Figure 21).
However, OSM concluded that mining‐activities generate only an additional 4 acre‐feet of leakage
per year to the N‐aquifer and, as a consequence, material damage to the hydrologic balance of the N‐
aquifer, in response to mining, has not and will not occur (OSM‐CHIA 1989).
45
Figure 21. Water chemistry data for select springs (in Macy 2009)
46
REACH OF PEABODY’S WITHDRAWALS
Figure 22 provides the water‐level record for six NTUA and BIA wells throughout the confined and
(assumed) unconfined portion of the N‐aquifer (water levels are shown as feet below land surface).
The red circles may highlight water levels following Peabody’s 6 month cessation of withdrawals; the
red‐dashed line indicates Peabody’s 70% reduction in withdrawals beginning in 2006.
It is hypothesized that the confined portion of the N‐aquifer reaches farther south than has
previously been recognized, and the water level recovery in these wells has not registered, as it did in
1985 (Hill and Whetten 1986), due to the principle of superposition (Bredehoeft et al. 1982; Theis 1940).
Indeed, in the subsequent 1999 model by Peabody consultants, the confined boundary of the N‐
aquifer “extends further to the southwest towards Goldtooth and southeast towards jadito [sic].
Differences in boundary locations range from 10 to 25 miles” (HSIGeoTrans and WEHE 1999: 7‐10).
Figure 19, below, illustrates that the conspicuous groundwater recovery “spike” as far south as
Goldtooth.
47
Longhouse Valley (north) Shonto Southeast (west)
White Mesa Arch (west) Rough Rock 9Y‐92 (northeast)
Tuba City 3K‐325 (southwest) Goldtooth (southwest)
Figure 22. Water Level Data for Five NTUA/BIA Wells14 (BMMP 2010)
14 Longhouse Valley and Shonto Southeast, north and west of the well‐field, are continuing declining trends.
Though water levels are variable in wells White Mesa Arch and Rough Rock (9Y‐92) both express the conspicuous
single year recovery “spike” in water levels following water year 1985 (highlighted in red), the year that Peabody’s
groundwater withdrawals ceased for 6 months due to maintenance at the Mohave Generating Station. Tuba City
3K‐325 remains variable ±5 ft., and Goldtooth expresses a slight decreasing trend, yet the post‐1985 recovery spike
is evident as far south as these two wells. Many other wells throughout the N‐aquifer also demonstrate the 1985
48
SUMMARY
This study found that the extraordinary range of hydrogeological uncertainties upon which
Peabody’s PHC is based undermines the conclusiveness of its determination of mining‐related impacts.
This finding is supported via the convergence of the following evidence:
(1) The USGS groundwater model is characterized by a high level of hydrogeological uncertainties (recharge, leakage, and discharge) and limited simulation capacities (Brown and Eychaner 1988; Eychaner 1983). These uncertainties explain why:
(2) Peabody’s PHC and OSM’s CHIA and EIS did not predict any groundwater parameters with any appreciable level of accuracy. This should not be surprising given the number of times the model had to be recalibrated during its first decade of use (i.e. it was first calibrated to match actual observations 1981 but divergence was apparent by 1982, in 1985, and in 1987, requiring recalibration each time. It is unknown if recalibration also occurred when the model was updated in 1994 and 1996).
(3) The USGS groundwater model, Peabody’s PHC, and OSM’s CHIA and EIS failed to capture the essentially linear relationship between water level decline and spring discharge. Recharge, leakage, and discharge continue to be poorly understood. The impact of Peabody withdrawals were generally underestimated and the effect of municipal withdrawals were generally overestimated. The boundaries of the confined and unconfined aquifer boundary areas are in question and there is significant reason to suspect that the confined portion of the aquifer reaches south of Goldtooth. There are robust indicators that mining‐impacts have had a far greater impact on the N‐aquifer’s hydrologic balance than has heretofore been recognized or acknowledged.
(4) In 2001, the USGS concluded that new studies investigating recharge, leakage, and discharge were needed before investing more time in refining the existing USGS groundwater model (Thomas 2002). In 2008, a USGS hydrologist explained that the existing USGS groundwater model was not a good representation of the N‐aquifer and, overall, had not performed well (Leake 2008).
(5) In the most current groundwater model developed by Peabody consultants (HSIGeoTrans and WEHE 1999), a discharge estimate was not attempted because discharge remains poorly understood. No new studies were performed to estimate either recharge or leakage.
“reversal” recorded by the USGS (Hill and Whetten 1986) These include Rough Rock 10T‐258, Rough Rock 10R‐
111, Pinon PM6, and possibly Keams Canyon PM2 and USGS observation well BM3.
49
APPENDIX A: Descriptive Statistics and Outliers for Peabody Withdrawals
Data for Kayenta withdrawals are available from 1984‐2008; data for spring discharge are
available for the period 1987‐2008; and Peabody withdrawal data are also available for this period
(Chart 1, below). However, because withdrawals declined by approximately 70% after 2005, the
regression of Peabody Withdrawals to discharge from Moenkopi School Spring, for example, is skewed
by the three years of reduced withdrawals (Chart 2).
As might be expected, descriptive statistics for Peabody withdrawals (the period 1987‐2008) express
three outliers representing withdrawals for 2006, 2006, and 2008 (Chart 3). Due to these three years,
the data are not distributed evenly under the normal fit; the box‐plot expresses three outliers; and the
three outliers skew the distribution.
It is not acceptable to remove outliers simply because they seem anomalous and do not trend with
other data. However, that is not the case here; these three years are not anomalous data points within
the data set. Rather, Peabody made a management decision to change its withdrawal regime by
decreasing withdrawals by 70%. In this case, the outliers are removed because they do not represent
conditions that are normal to the data set.
With these three years removed from the data set (Chart 4), descriptive statistics express the small
number of data points within the data set as distributed more evenly.
CHART 1
‐
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
5,000
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
With
draw
als
in a
cre-
feet
per
yea
r
50
CHART 2
51
CHART 3
52
CHART 4
53
Key references for Black Mesa and the Black Mesa‐Kayenta Coal Mine
Alley, W.M., and P.A. Emery. 1986. Groundwater model of the Blue River Basin, Nebraska – twenty years later. Journal of Hydrology, 85, pp. 225-250. Alley, W.M., O.L. Franke, and T.E. Reilly. 1999. Sustainability of Ground-Water Resources, U.S. Geological Survey Circular 1186. Denver, Colorado. 79 p. Alley, W.M., and S.A. Leake. 2004. The Journey from Safe Yield to Sustainability. Groundwater, Vol.42, No. 1, January-February 2004, pp. 12-16. American Society of Civil Engineers. 1972. Groundwater Management, Manual of Engineering Practice 40, New York. Anderson, M.P. 1995. Groundwater Modeling in the 21st Century. In Al-Kadi (ed.) Groundwater Models for Resources Analysis and Management. Lewis Publishers, Boca Raton. Anderson, P.F., and S. Lu. 2005. A Post Audit of a Model-Designed Ground Water Extraction System. Ground Water. Vol. 41, No.2, pp. 212-218. Anderson M.P., and W.W. Woessner. 1992a. Applied Groundwater Modeling: Simulation of Flow and Advective Transport. Academic Press, New York.
Anderson M.P., and W.W. Woessner. 1992b. The role of the postaudit in model validation. Adv. Water Res., 15, 167, 1992. Anderson M.P., D.S. Ward, E.G. Lappala, and T.A. Prickett. 1993. Computer Models for Subsurface Water. In D.R. Maidment (ed.) 1993, Handbook of Hydrology, McGraw-Hill, Inc. New York. Arizona Department of Water Resources, 2006. Arizona Water Atlas, Volume 2: Eastern Plateau Planning Area. AZ Department of Water Resources, Draft, June 2006. Arizona Department of Water Resources, 2008. Power Plants in Arizona – an Emerging Industry, a New Water User. Arizona Water Resources Research Center, University of Arizona, last accessed 1 April 2010: http://ag.arizona.edu/AZWATER/awr/janfeb01/feature1.htm.
54
Balleau W.P., and A.B. Mayer. 1988. The transition from groundwater mining to induced recharge in generalized hydrogeologic systems. Proceedings of FOCUS conference on Southwestern Ground Water Issues, National Water Well Association, Dublin Ohio, pp. 81-103. Barnett, T.P., D.W. Pierce. 2008. When will Lake Mead go dry? Water Resources Research, Vol. 44, W03201, doi:10, 1029/2007WR006704. http://meteora.ucsd.edu/~pierce/papers/Barnett_Pierce_2008_JWRR_Lake_Mead.pdf Barnett, T.P., and D.W. Pierce. 2009. Sustainable water deliveries from the Colorado River in a changing climate. Proceedings of the National Academy of Sciences. Vol. 106, no. 18. http://meteora.ucsd.edu/~pierce/papers/Barnett_and_Pierce_2009_PNAS_Colo_Sustainability.pdf Black Mesa Monitoring Program (BMMP). 2010. U.S. Geological Survey; Jamie Macy, Project Director. Cooperating Agencies include the Bureau of Indian Affairs and the Arizona Department of Water Resources. See the link to the “Interactive Data Map” for groundwater level and chemical data: http://az.water.usgs.gov/projects/9671-9E9/index.html. Last accessed 18 March 2010. Bredehoeft, J.D. 1997. Safe Yield and the Water Budget Myth. Editorial in Groundwater, Vol.35, No.6, November-December 1997. Bredehoeft, J.D. 2002. The Water Budget Myth Revisited: Why Hydrogeologists Model. Groundwater, Vol. 40, No. 4, July-August 2002, pp. 340-345. Bredehoeft, J.D. 2006. On Modeling Philosophies. Ground Water, Vol. 44, No.4. July-August, pp. 496-499. Bredehoeft, J.D., S. S. Papadopulos, and H. H. Cooper. 1982. Groundwater: The Water-Budget Myth. In Scientific Basis of Water-Resource Management, Studies in Geophysics, Washington DC: National Academy Press, pp. 51-57. Bredehoeft, J.D., and L.F. Konikow. 1993. Groundwater models: validate or invalidate. Ground Water, 31(2), 178, 1993. Brown, R.H. 1963(a). The cone of depression and the area of diversion around a discharging well in an infinite strip aquifer subject to uniform recharge. U.S Geological Survey Water Supply Paper 1545C.
55
Brown, S.C. 1963(b). Potential Ground Water Supply of the South Shonto-Tsegi Area, Navajo County, Arizona; Sentry Royalty Company, St. Louis, Missouri, 42 pages. Brown, J.G., and J.H. Eychaner. 1988. Simulation of five ground-water withdrawal projections for the Black Mesa area, Navajo and Hopi Indian Reservations, Arizona. U.S. Geological Survey Water-Resources Investigations Report 88–4000, 51 p. Carpenter, M.C. 1999. Part 1: South-Central Arizona. In D. Galloway, D.R. Jones, and S.E. Ingebritsen (Eds.) Land Subsidence in the United States (Circular No. 1182, pp. 65-78). Reston, VA: U.S. Geological Surgey. Carpenter, S. R., W. A. Brock, and D. Ludwig. 2002. Collapse, Learning, and Renewal. In L.H. Gunderson & C.S. Holling (Eds.), Panarchy: Understanding Transformations in Human and Natural Systems. Island Press, Washington D.C. Cooley, M.E., J.W. Harshbarger, J.P Akers, and W.F. Hardt. 1964. Selected Lithology Logs, Drillers’ Logs, and Stratigraphic Sections Part 3 of Geohydrologic Data in the Navajo and Hopi Indian Reservations, Arizona, New Mexico, and Utah; Arizona State Land Department, Water Resources Report 12-C, 157 pages (this study was subsequently integrated into Cooley et al. 1969). Cooley, M.E., J.W. Harshbarger, J.P. Akers, and H.F. Hardt. 1969. Regional Hydrogeology of Navajo and Hopi Indian Reservations, Arizona, New Mexico, and Utah. With a section on vegetation by O.N. Hicks. United States Department of the Interior, Geological Survey Professional Paper 521-A, prepared in cooperation with the Bureau of Indian Affairs and the Navajo Tribe. Davis, G.E., W.F. Hardt, L.K. Thompson, and M.E. Cooley. 1963. Records of ground-water supplies, part 1 of geohydrologic data in the Navajo and Hopi Indian Reservations, Arizona, New Mexico, and Utah: Arizona State Land Department Water-Resources Report 12–A, 159 p. Desert Research Institute, 2010. Precipitation and other climate data throughout Black Mesa’s communities is available at: http://www.wrcc.dri.edu/. Tuba City precip. data: http://www.wrcc.dri.edu/cgi-bin/cliMAIN.pl?az8792 . Devlin, J.F., and M.A. Sophocleous. 2005. The persistence of the water budget myth and its relationship to sustainability. Hydrogeology Journal, 13(4):549-554. El-Kadi, A. 1995. Needs for the next Generation of Models. In El-Kadi (ed.), Groundwater Models for Resources Analysis and Management. Lewis Publishers, Boca Raton.
56
Eychaner, J.H., 1981. Geohydrology and Effects of Water Use in the Black Mesa Arizona, Navajo and Hope Indian Reservations, Arizona. U.S. Geological Survey Open-File Report 81-911, Prepared in cooperation with the Arizona Department of Water Resources. Eychaner, J.H., 1983. Geohydrology and effects of water use in the Black Mesa area, Navajo and Hopi Indian Reservations, Arizona: U.S. Geological Survey Water-Supply Paper 2201, 26 p. Fairhead, J., and M. Leach. 1998. Reframing Deforestation: Global Analyses and Local Realities: Studies in West Africa, Routledge: London. Flavelle, P. 1992. A Quantitative Measure of Model Validation and its Potential Use for Regulatory Purposes. Advances in Water Resources, 15, 5-13. Flavelle, P., S. Nguyen, W. Napier, and D. Lafleur. 1990. Lessons learned from model validation – a regulatory perspective. In GEOVAL ’90 (Ref. 21) pp.441-8. Freyberg, D.L. 1988. An exercise in groundwater model calibration and prediction. Ground Water 26(3), 350-360. GeoTrans, Inc. 1987. A two-dimensional finite-difference flow model simulating the effects of withdrawals to the N aquifer, Black Mesa area, Arizona: Boulder, Colorado, GeoTrans, Inc., report prepared for Peabody Western Coal Company. GeoTrans, Inc. 2005. A three-dimensional flow model of the D and N aquifers, Black Mesa Basin, Arizona: Supplement 1. Louisville, Colorado 80027. Report prepared for Peabody Western Coal Company, 50 p. Glennon, R.G. 2002. Water Follies: Groundwater Pumping and the Fate of America’s Fresh Waters. Island Press, Washington D.C. Greenberger, M., M. Crenson, and B. Crissey. 1976. Models in the Policy Process. New York: Russell Sage Foundation. Harshbarger, J.W., D.D. Lewis, H.E. Skibitzke, W.L. Heckler, and L.R. Kister. 1966. Arizona Water. U.S. Geological Survey Water-Supply Paper 1648, 85 p. Hart, R.J. 1998. Draft memorandum from Bob Hart, Supervisory Hydrologist in the USGS Water Resources Division, to Greg Littin, Hydrologist in the USGS Water Resources Division (19 August 1998); pp. 1, 3 (Black Mesa Monitoring Program – a New Focus). Quoted in the NRDC’s report:
57
Drawdown: Groundwater Mining on Black Mesa. The updated version of the report is at: http://www.nrdc.org/water/conservation/draw/drawinx.asp Hart, R.J., and J.P. Sottilare. 1988. Progress report on the ground-water, surface-water, and quality-of-water monitoring program, Black Mesa area, northeastern Arizona—1987–88. U.S. Geological Survey Open-File Report 88–467, 27 p. Hart, R.J., and J.P. Sottilare. 1989. Progress report on the ground-water, surface-water, and quality-of-water monitoring program, Black Mesa area, northeastern Arizona—1988–89. U.S. Geological Survey Open-File Report 89–383, 33 p. Hassan, A. 2004. Validation of Numerical Ground Water Models Used to Guide Decision Making. Ground Water, Vol. 42, No. 2, March-April 2004, pp. 277-290. Hill, G.W., 1985, Progress report on Black Mesa monitoring program—1984. U.S. Geological Survey Open-File Report 85–483, 24 p. Hill, G.W., and J.P. Sottilare. 1987. Progress report on the ground-water, surface-water, and quality-of-water monitoring program, Black Mesa area, northeastern Arizona—1987. U.S. Geological Survey Open-File Report 87–458, 29 p. Hill, G.W., and M.I. Whetten. 1986. Progress report on Black Mesa monitoring program—1985–86. U.S. Geological Survey Open-File Report 86–414, 23 p. Hill, M.C. 1992. Methods and Guidelines for effective model calibration with application to UCODE, a computer code for universal inverse modeling. U.S. Geological Survey Water Resources investigations Report 98-4005. Holling, C.S., 1978. Adaptive Environmental Assessment and Management. The Blackburn Press, Caldwell, NJ. Holling, C.S. 1973. Resilience and stability of ecological systems. Ann. Rev. of Ecol. and Syst. 4: 1-23. HSIGeoTrans, Inc., 1993. Investigation of the N- and D-aquifer geochemistry and flow characteristics using major ion and isotopic chemistry, petrography, rock stress analyses, and dendrochronology in the Black Mesa area, Arizona. Boulder, Colorado, HSIGeoTrans, Inc., prepared for Peabody Coal Company, 400 v.p.
58
HSIGeoTrans, Inc., and Waterstone Environmental Hydrology and Engineering, Inc., 1999, A three-dimensional flow model of the D and N aquifers, Black Mesa Basin, Arizona. Prepared for Peabody Western Coal Company, 75 p. Kashef, A-A.I. 1986. Groundwater Engineering. McGraw-Hill Book Company, New York. Kister, L.R., and J.L. Hatchett. 1963. Selected chemical analyses of the ground water, part 2 of Geohydrologic data in the Navajo and Hopi Indian Reservations, Arizona, New Mexico, and Utah. Arizona State Land Department Water-Resources Report 12–B, 58 p. Konikow L.F. 1986. Predictive Accuracy of a Ground-Water Model – Lessons from a Postaudit. Groundwater, Vol. 24, No. 2, March-April 1986. Konikow, L.F. 1995. The Value of Postaudits in Groundwater Model Applications. In Al-Kadi (ed.) Groundwater Models for Resources Analysis and Management. Lewis Publishers, Boca Raton.
Konikow L.F., and J.D. Bredehoeft. 1974. Modeling flow and chemical quality changes in an irrigated stream aquifer-system. Water Resources Research, 10, pp. 546-562. Konikow L.F., and J.D. Bredehoeft. 1992. Groundwater models cannot be validated. Adv. Water Res., 15, 75. Konikow L.F., and J.D. Bredehoeft. 1993. The Myth of Validation in Ground Water Modeling. In Proceedings 1993 Groundwater Modeling Conference, A-4. Golden, Colorado, International Groundwater Modeling Center.
Konikow L.F., and E.P. Patten. 1985. Groundwater Forecasting. In Hydrological Forecasting, ed. M.G. Anderson and T.P. Burt. John Wiley and Sons Ltd. New York. Konikow, L.F., and M.A. Person. 1985. Assessment of Long-term salinity changes in an irrigated stream aquifer system. Water Resources Research, Vol. 21, 1611-1624.
Konikow, L.F., and L.A. Swain. 1990. Assessment of predictive accuracy of a model of artificial recharge effects in the Upper Coachella Valley, California. In Simpson, E.S., and Sharp, J.M. (eds.) Selected Papers on Hydrogeology, Vol. 1. International Association of Hydrogeologists, Proc. 1989 IGC Meeting, Washington, D.C., pp. 433-449.
59
Leake, S. 2008. Personal conversation with Stanley Leake, Hydrogeologist, USGS office in Tucson, AZ, May 2009. Leijnse, A., and S.M. Hassanizadeh. 1994. Model Definition and Model Validation. Advances in Water Resources, 17 (1994) 197-200. Levitt, N., and P. Gross. 1994. The Perils of Democratizing Science. The Chronicle of Higher Education, 5 October, B1, B2. Lewis, B.D., and F.J. Goldstein. 1982. Evaluation of a Predictive Ground-Water Solute-Transport model at the Idaho National Engineering Laboratory, Idaho. U.S. Geological Survey Water Resour. Inv. 82-25, 1982. Littin, G.R. 1992. Results of ground-water, surface-water, and water-quality monitoring, Black Mesa area, northeastern Arizona—1990–91. U.S. Geological Survey Water-Resources Investigations Report 92–4045, 32 p. Littin, G.R. 1993. Results of ground-water, surface-water, and water-quality monitoring, Black Mesa area, northeastern Arizona—1991–92. U.S. Geological Survey Water-Resources Investigations Report 93–4111, 23 p. Littin, G.R., B.M. Baum, and M. Truini. 1999. Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—1999. U.S. Geological Survey Open-File Report 98–653, 27 p. Littin, G.R., and S.A. Monroe. 1995a. Results of groundwater, surface-water, and water-quality monitoring, Black Mesa area, northeastern Arizona—1992–93. U.S. Geological Survey Water-Resources Investigations Report 95–4156, 37 p. Littin, G.R., and S.A. Monroe. 1995b. Results of ground-water, surface-water, and water-chemistry monitoring, Black Mesa area, northeastern Arizona—1994. U.S. Geological Survey Water-Resources Investigations Report 95–4238, 25 p. Littin, G.R., and S.A. Monroe. 1996. Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—1995. U.S. Geological Survey Open-File Report 96–616, 22 p. Littin, G.R., and S.A. Monroe. 1997. Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—1996. U.S. Geological Survey Open-File Report 97–566, 27 p.
Lohman, S.W. 1972. Definitions of selected groundwater terms – revisions and conceptual refinements. U.S. Geological Survey, Water Supply Paper 1988, 21 pp.
60
Lopes, T.J., and J.P. Hoffmann. 1997. Geochemical analyses of ground-water ages, recharge rates, and hydraulic conductivity of the N Aquifer, Black Mesa area, Arizona. U.S. Geological Survey Water-Resources Investigations Report 96–4190, 42 p. Macy, J.P. 2009. Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—2007–08: U.S. Geological Survey Open-File Report 2009-1148, 43 p. Macy, J.P. 2010. Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—2008–09: U.S. Geological Survey Open-File Report 2010-1038, 43 p. McDonald, M.G., and A.W. Harbaugh. 1988. A Modular Three-Dimensional Finite Difference Groundwater Flow Model: U.S. Geological Survey Open File Report 83-875, p. 528. Milly, P.C.D., J. Betancourt, M. Falkenmark, R.M. Hirsch, Z.W. Kundewicz, D.P. Lettenmaier, and R.J. Stouffer. 2008. Stationarity is Dead: Whither Water Management. Science Vol. 319. Mills, S. 1997. Discourse. Routledge, London. Moench, M. 2004. Groundwater: the challenge of monitoring and management. In: Gleick, P.H. (ed.), The World’s Water: 2004-2005, Washington DC: Island Press, pp. 79-100. National Research Council, 1990. Ground Water Models: Scientific and Regulatory Applications. Water Science and Technology Board Committee on Ground Water Modeling Assessment, Commission on Physical Sciences, Mathematics, and Resources. National Academy Press, Washington DC. National Research Council, 2000. Research Needs in Subsurface Science. National Academy Press, Washington D.C. Nichols, W.D. 1993 (Western Region Groundwater Specialist, USGS) Letter to William M. Alley, Chief of the Office of Groundwater, Water Resources Division, USGS, 10/28/1993. Office of Surface Mining Reclamation and Enforcement. 1985. Guidelines for Preparation of a Cumulative Hydrologic Impact Assessment (CHIA) [Draft]. December 1985. Office of Surface Mining Reclamation and Enforcement (OSM-CHIA). 1989. Cumulative Hydrologic Impact Assessment of the Peabody Coal Company Black Mesa / Kayenta Mine. Prepared by the OSMRE, Western Field Operations, April 1989.
61
Office of Surface Mining Reclamation and Enforcement (OSM-EIS). 1989. Proposed Permit Application, Black Mesa / Kayenta Mine, Navajo and Hopi Indian Reservations, Arizona. Draft Environmental Impact Statement, OSM-EIS-25. June 1989. Office of Surface Mining Reclamation and Enforcement. 1990. Interior Department Issues Permit for Kayenta Coal Mine, Defers Decision on Black Mesa. Press release, 6 July 1990. Office of Surface Mining Reclamation and Enforcement (OSM-EIS). 1990. Proposed Permit Application, Black Mesa / Kayenta Mine, Navajo and Hopi Indian Reservations, Arizona. Volume 1: Report. Final Environmental Impact Statement, OSM-EIS-25. May 1990. Office of Surface Mining Reclamation and Enforcement (OSM-CHIA Addendum). 1990. Addendum to the CHIA; distributed by Peter Rutledge, Chief, Federal Programs Division, 21 June 1990. The addendum addresses and corrects an error in the CHIA criterion for the baseflow of Moenkopi Wash. Office of Surface Mining Reclamation and Enforcement. 1998. Report on its Review and Analysis of Peabody Coal Company’s 1996 “Annual Hydrologic Data Report” and the U.S. Geological Survey’s “Groundwater, Surface-Water, and Water-Chemistry Data, Black Mesa Area, Northeastern Arizona – 1996, 1995, 1994, 1992-93.” June 1998. Office of Surface Mining Reclamation and Enforcement. 2000. Report on its Review and Analysis of Peabody Coal Company’s 1998 “Annual Hydrologic Data Report” and the U.S. Geological Survey’s “Groundwater, Surface-Water, and Water-Chemistry Data, Black Mesa Area, Northeastern Arizona – 1998.” July 5, Corrected , August 7. Office of Surface Mining Reclamation and Enforcement. 2002. Permitting Hydrology: A Technical Reference Document for Determination of Probable Hydrologic Consequences (PHC) and Cumulative Hydrologic Impact Assessments (CHIA), Baseline Data. May 2002. Office of Surface Mining Reclamation and Enforcement. 2004. Report on its Review and Analysis of Peabody Coal Company’s 2003 “Annual Hydrologic Data Report” and the U.S. Geological Survey’s “Groundwater, Surface-Water, and Water-Chemistry Data, Black Mesa Area, Northeastern Arizona – 2003-03.” September 2004. Office of Surface Mining Reclamation and Enforcement. 2005. Report on its Review and Analysis of Peabody Coal Company’s 2004 “Annual Hydrologic Data Report” and the U.S. Geological Survey’s “Groundwater, Surface-Water, and Water-Chemistry Data, Black Mesa Area, Northeastern Arizona – 2003-04.” September 2004.
62
Office of Surface Mining Reclamation and Enforcement. 2005(b). Correspondence from Paul Clark, OSM hydrologist, to Tim Grabiel, Natural Resources Defense Council, regarding the 2005 update of Peabody’s 1999 3D groundwater model. Office of Surface Mining Reclamation and Enforcement. 2006. Report on its Review and Analysis of Peabody Coal Company’s 2005 “Annual Hydrologic Data Report” and the U.S. Geological Survey’s “Groundwater, Surface-Water, and Water-Chemistry Data, Black Mesa Area, Northeastern Arizona – 2004-05.” November 2006. Office of Surface Mining Reclamation and Enforcement (OSM-EIS). 2006. Black Mesa Project: Draft Environmental Impact Statement. Prepared by the OSMRE, Western Region, November 2008. Office of Surface Mining Reclamation and Enforcement. 2007. Hydrologic Considerations for Permitting and Liability Release, Technical Reference Document for the Midcontinent Region, OSM, 2007. Office of Surface Mining Reclamation and Enforcement (OSM-EIS). 2008. Black Mesa Project: Final Environmental Impact Statement, Volume I–Report, and Volume II–Comments and Responses. Prepared by the OSMRE, Western Region, November 2008. Office of Surface Mining Reclamation and Enforcement (OSM-CHIA). 2008. Cumulative Hydrologic Impact Assessment of the Peabody Western Coal Company Black Mesa Complex. Prepared by the OSMRE, Western Region, December 2008. 98 p. Office of Surface Mining Reclamation and Enforcement. 2008 (OSM 2008). See OSM’s Indian Lands Program: http://www.wrcc.osmre.gov/OLD/Indlnds.html Office of Surface Mining Reclamation and Enforcement. 2008b (OSM 2008b). Significant Permit Revision Record of Decision, Black Mesa Complex. Peabody Western Coal Company, Navajo Country, Arizona. Permit No. AZ-0001D. OSM Project Number AZ-001-E-P-01. December 2008. Oreskes, N., K. Shrader-Frechette, and K. Belitz. 1994. Verification, Validation, and Confirmation of Numerical Models in the Earth Sciences. SCIENCE, Vol. 263, 4 Feb. 1994. Oreskes, N. 1998. Evaluation (Not Validation) of Quantitative Models. Environmental Health Perspectives, Vol. 106, Supplement 6, December 1998. Oreskes, N. 2003. The Role of Quantitative Models in Science. In Canham, C.D., J.J. Cole, and W.K Lauenroth (eds), Models in Ecosystem Science, Princeton University Press, Princeton, NJ.
63
Oreskes, N. 2007. The American Denial of Global Warming. Guest lecture for the series Perspectives on Ocean Science, University of California Television, first airing 17 December 2007. The lecture may be viewed at: http://www.uctv.tv/search-details.aspx?showID=13459. Last accessed 1 May 2010. Papadopulos, S.S. and Associates, 1993. Navajo aquifer water study, final report, prepared for the Hopi Tribe, Navajo Nation, and Peabody Coal Company. Bethesda Maryland, Papadopulos and Associates, Inc. 49. Peabody Coal Company (PCC). 1985. Mining and Reclamation Plan: Black Mesa and Kayenta Mines. Arizona Division, Peabody Coal Company. Peabody Coal Company (PCC-PHC). 1985. Probable Hydrologic Consequences. Chapter 18 of Peabody Coal Company, 1985, Mining and Reclamation Plan: Black Mesa and Kayenta Mines. , Arizona Division, Peabody Coal Company. Person, M., and L.F. Konikow. 1986. Recalibration and predictive reliability of a solute-transport model of an irrigated stream-aquifer system. Journal of Hydrology 87, 145-165. Sargent, R.G. 1990. Validation of Mathematical Models. In Proceedings of GEOVAL ’90 Symposium on Validation of Geosphere Performance Assessment Models, Stockholm, Sweden, May 14-17 pp. 571-579. Stockholm: Swedish Nuclear Power Inspectorate. Smith, C.F., D.W. Anning, N.R. Duet, G.G. Fisk, H.F. McCormack, G.L. Pope, P.D. Rigas, and B.L. Wallace. 1995. Water-resources data for Arizona, water year 1994: U.S. Geological Survey Water-Data Report AZ–94–1, 320 p. Smith, C.F., F.C. Boner, R.G. Davis, N.R. Duet, and P.D. Rigas. 1993. Water-resources data for Arizona, water year 1992. U.S. Geological Survey Water-Data Report AZ–92–1, 360 p. Smith, C.F., P.D. Rigas, L.K. Ham, N.R. Duet, and D.W. Anning. 1994. Water-resources data for Arizona, water year 1993. U.S. Geological Survey Water-Data Report AZ–93–1, 360 p. Smith, C.F., N.R. Duet, G.G. Fisk, H.F. McCormack, C.K. Partin, G.L. Pope, P.D. Rigas, and S. Tadayon. 1996. Water-resources data for Arizona, water year 1995: U.S. Geological Survey Water-Data Report AZ–95–1, 306 p. Smith, C.F., N.R. Duet, G.G. Fisk, H.F. McCormack, C.K. Partin, G.L. Pope, P.D. Rigas. 1997. Water-resources data for Arizona, water year 1996: U.S. Geological Survey Water-Data Report AZ–96–1, 328 p.
64
Sophocleous, M.A. 1997. Managing water resources systems: why “safe yield” is not sustainable. Groundwater Vol.35, No. 4, 561. Sophocleous, M.A. 1998. On the Elusive Concept of Safe Yield and the Response of Interconnected Stream-Aquifer Systems to Development. In: Sophocleous, M.A. (ed.) Perspectives on Sustainable Development of Water Resources in Kansas. Kansas Geological Survey, Bulletin 239, p. 86-95. Sophocleous, M.A., 2000. From safe yield to sustainable development of water resources, and the Kansas experience. Journal of Hydrology, 235(1-2):27-43. Sottilare, J.P., 1992, Results of ground-water, surface-water, and water-quality monitoring, Black Mesa area, northeastern Arizona—1989–90: U.S. Geological Survey Water-Resources Investigations Report 92–4008, 38 p. Surface Mining Control and Reclamation Act (SMCRA). 1977. Public Law 95-97. A copy of the Act and all of its revisions through December 1993 is accessible through the Office of Surface Mining Reclamation and Enforcement’s website: see http://osmre.gov Tadayon, S., N.R. Duet, G.G. Fisk, H.F. McCormack, C.K. Partin, G.L. Pope, and P.D. Rigas. 1999. Water-resources data for Arizona, water year 1998. U.S. Geological Survey Water-Data Report AZ–98–1, 454 p. Tadayon, S., N.R. Duet, G.G. Fisk, H.F. McCormack, C.K. Partin, G.L. Pope, and P.D. Rigas. 2000. Water-resources data for Arizona, water year 1999. U.S. Geological Survey Water-Data Report AZ–99–1, 389 p. Tadayon, S., N.R. Duet, G.G. Fisk, H.F. McCormack, C.K. Partin, G.L. Pope, and P.D. Rigas. 2001, Water-resources data for Arizona, water year 2000: U.S. Geological Survey Water-Data Report AZ–00–1, 390 p. Tadayon, S., N.R. Duet, G.G. Fisk, H.F. McCormack, G.L. Pope, and P.D. Rigas. 1998, Water-resources data for Arizona, water year 1997: U.S. Geological Survey Water-Data Report AZ–97–1, 416 p. Taylor, T.B., and C.C. Humpstone. 1973. The Restoration of the Earth. Harper and Row, New York. Theis, C.V. 1940. The Source of Water Derived from Wells: Essential Factors Controlling the Response of an Aquifer in Development. Civil Engineer 10, 277-280.
65
Thomas, B.E. 2002. Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—2000–2001, and performance and sensitivity of the 1988 USGS numerical model of the N aquifer. U.S. Geological Survey Water-Resources Investigations Report 02–4211, 75 p Thomas, B.E., and M. Truini. 2000. Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona–1999. U.S. Geological Survey Open-File Report 00–453, 42 p. TIME Magazine. 1964. WESTward Ho! Friday, 2 October 1964. Trescott, P.C., G.F. Pinder, and S.P. Larson. 1976. Finite Difference Model for Aquifer Simulation in Two-Dimensions, with Results of Numerical Experiments. U.S. Geological Survey Techniques of Water Resources Investigations, Book 7, Chapter C1, p. 116. Truini, Margot, Baum, B.M., Littin, G.R., and Shingoitewa-Honanie, Gayl, 2000, Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—1998: U.S. Geological Survey Open-File Report 00–66, 37 p. Truini, Margot, and Longsworth, S.A., 2003, Hydrogeology of the D aquifer and movement and ages of ground water determined from geochemical and isotopic analyses, Black Mesa area, northeastern Arizona: U.S. Geological Survey Water-Resources Investigations Report 03–4189, 38 p. Truini, Margot, and Thomas, B.E., 2004, Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—2002–03: U.S. Geological Survey Open-File Report 03–503, 43 p. Truini, Margot and Macy, J.P., 2005, Lithology and thickness of the Carmel Formation as related to leakage between the D and N aquifer, Black Mesa, Arizona: U.S. Geological Survey Scientific Investigation Report 2005-5187, 7 p. Truini, Margot, Macy, J.P., and Porter T.J., 2005, Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—2003–04: U.S. Geological Survey Open-File Report 2005–1080, 44 p. Truini, Margot, and Macy, J.P., 2006, Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—2004–05: U.S. Geological Survey Open-File Report 2006–1058, 42 p. Truini, Margot, and Macy, J.P., 2007, Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—2005-06: U.S. Geological Survey Open-File Report 2007–1041, 41 p.
66
Truini, Margot, and Macy, J.P., 2008, Ground-water, surface-water, and water-chemistry data, Black Mesa area, northeastern Arizona—2005-06: U.S. Geological Survey Open-File Report 2008–1324, 33 p. U.S. Department of the Interior, Bureau of Reclamation. 1971. Draft of Environmental Statement for Black Mesa Coal Mining Operation. The Final EIS for the Black Mesa Coal Mining Operation is incorporated into the U.S. DOI BOR’s 1972 Final Environmental Impact Statement for the Navajo Project, Arizona. U.S. Department of the Interior, Bureau of Reclamation. 1972. Navajo Project, Arizona: Final Environmental Statement. 4 February 1972 van der Heijde, P.K.M. 1995. Model Testing: A Functional Analysis, Performance Evaluation, and Applicability Assessment Protocol. In El-Kadi (ed.), Groundwater Models for Resources Analysis and Management. Lewis Publishers, Boca Raton. van der Heijde, P.K.M., and O.S Elnawaywy. 1992. Quality assurance and quality control in the development and application of groundwater models, Report EPA/600/R-93/011. U.S. Environmental Protection Agency, Ada. OK. van der Heijde, P.K.M., P.S. Huyakorn, and J.W. Mercer. 1985. Testing and validation of groundwater models, in Proc. Conf. on Practical Applications of Groundwater Models, National Water Well Association, Dublin OH. van der Heijde, P.K.M., and R.A. Park. 1986. U.S. EPA Groundwater modeling policy study group; report on findings and discussion of selected groundwater modeling issues, Report GWMI 86-13, International Ground Water Modeling Center, Indianapolis. van der Heijde, P.K.M., S.S. Paschke, and D.A. Kanzer. 1993. Groundwater flow and solute transport model functionality testing and performance evaluation. In Proc. Thirteenth AGU Hydrology Days, Morel-Syetoux, H.J. Ed., Fort Collins, CO. van Fraassen, B. 1980. The Scientific Image. Oxford University Press, New York. White, N.D., and W.B. Garrett. 1984. Water resources data for Arizona, water year 1982: U.S. Geological Survey Water-Data Report AZ–82–1, 440 p.
67
White, N.D., and W.B. Garrett. 1986, Water resources data for Arizona, water year 1983: U.S. Geological Survey Water-Data Report AZ–83–1, 387 p. White, N.D., and W.B. Garrett.1987, Water resources data for Arizona, water year 1984: U.S. Geological Survey Water-Data Report AZ–84–1, 381 p. White, N.D., and W.B. Garrett. 1988, Water resources data for Arizona, water year 1985: U.S. Geological Survey Water-Data Report AZ–85–1, 343 p. Wilson, L. 1998. A Practical Method for Environmental Impact Assessment Audits. Environmental Impact Assessment Review, 18: 59-71. Woessner, W.W., and M.P. Anderson. 1996. Good Model—Bad Model, Understanding the Flow Modeling Process. In Ritchey, J.D., Rumbaugh, J.O. (eds), Subsurface Fluid Flow (Ground-Water and Vadose Zone) Modeling. ASTM Conshohocken, PA. Wood, C. 2008. Dry Spring: The Coming Water Crisis of North America. Raincoast Books, Vancouver, B.C. Wood, L.A. 1976. US Geological Survey, written communication to J. Eychaner, noted in Eychaner, J.H., 1983. Geohydrology and effects of water use in the Black Mesa area, Navajo and Hopi Indian Reservations, Arizona: U.S. Geological Survey Water-Supply Paper 2201, 26 p. Zhu, Chen, Waddell, R.K., Star, I., and Ostrander, M., 1998, Responses of ground water in the Black Mesa Basin, northeastern Arizona, to paleoclimatic changes during the late Pleistocene and Holocene: Geology, v. 26, no. 2, p. 127–130. Zhu, Chen, 2000, Estimate of recharge from radiocarbon dating of groundwater and numerical flow and transport modeling: Water Resources Research, v. 36, no. 9, p. 2607–2620.