EPA910-R-14-001ES | January 2014
&EPA
United States
Environmental Protection
Agency
An Assessment of Potential Mining Impacts
on Salmon Ecosystems of Bristol Bay, Alaska
Executive Summary
Region 10, Seattle, WA
www.epa.gov/bristolbay
-------
-------
The Bristol Bay watershed in southwestern Alaska supports the largest
sockeye salmon fishery in the world, is home to 25 federally recognized
tribal governments, and contains significant mineral resources. The potential
for large-scale mining activities in the watershed has raised concerns
about the impact of mining on the sustainability of Bristol Bay's world-class
commercial, recreational, and subsistence fisheries and the future of Alaska
Native tribes in the watershed, who have maintained a salmon-based culture
and subsistence-based way of life for at least 4,000 years.
The U.S. Environmental Protection Agency (USEPA) launched this assessment
to determine the significance of Bristol Bay's ecological resources and
evaluate the potential impacts of large-scale mining on these resources. It
uses the well-established methodology of an ecological risk assessment,
which is a type of scientific investigation that provides technical information
and analyses to foster public understanding and inform future decision
making. As a scientific assessment, it does not discuss or recommend policy,
legal, or regulatory decisions, nor does it outline or analyze options for
future decisions.
-------
This assessment characterizes the biological
and mineral resources of the Bristol Bay
watershed. It is intended to increase
understanding of potential impacts of large-
scale mining on the region's fish resources
and serve as a technical resource for the
public and for federal, state, and tribal
governments as they consider how best to
address the challenges posed by mining
and ecological protection in the Bristol Bay
watershed. It will inform ongoing discussions
of the risks of mine development to the
sustainability of the Bristol Bay salmon
fisheries and thus will be of value to the many
stakeholders in this debate.
The assessment also will inform the
consideration of options for future
government action, including, possibly,
by USEPA, which has been petitioned by
multiple groups to address mining activity
in the Bristol Bay watershed using its
authority under the Clean Water Act (CWA).
Should specific mine projects reach the
permitting stage, the assessment will enable
state and federal permitting authorities to
make informed decisions to grant, deny, or
condition permits and/or conduct additional
Cook Inlet
Bristol Bay
N
A
25 50
] Kilometers
25
] Miles
Approximate Pebble Deposit Location
Towns and Villages
Parks, Refuges, or Preserves
Watershed Boundary
The Nushagak River and Kvichak River Watersheds of Bristol Bay
-------
Present
Spawning
Rearing
Pebble 6.5 Components
Mine Scenario Watersheds
Watershed Boundary
Reported Salmon (Sockeye, Chinook, Coho, Pink, and Chum Combined) Distribution in the South and North Fork Koktuli River and Upper Talarik
Creek Watersheds. Designation of species spawning, rearing, and presence is based on the Anadromous Waters Catalog (Johnson and Blanche 2012). Life-stage-
specific reach designations are likely underestimates, given the challenges inherent in surveying all streams that may support life-stage use throughout the year.
research or assessment as a basis for
such decisions. USEPA conducted this
assessment consistent with its authority
under the CWA Section 104(a) and (b).
Scope of the Assessment
This assessment reviews, analyzes, and
synthesizes information relevant to
potential impacts of large-scale mine
development on Bristol Bay fisheries and
consequent effects on wildlife and Alaska
Native cultures in the region. Given
the economic, ecological, and cultural
importance of the region's key salmonids
(sockeye, Chinook, coho, chum, and
pink salmon, as well as rainbow trout
and Dolly Varden) and stakeholder
and public concern that a mine could
affect those species, the primary focus
of the assessment is the abundance,
productivity, and diversity of these
fishes. Because wildlife in Bristol Bay are
intimately connected to and dependent
on these and other fishes, changes in
these fisheries are expected to affect
the abundance and health of wildlife
populations. Alaska Native cultures have
strong nutritional, cultural, social, and
spiritual dependence on salmon, so
changes in salmon fisheries are expected
to affect the health and welfare of Alaska
Native populations. Therefore, wildlife
and Alaska Native cultures are also
considered as assessment endpoints, but
only as they are affected by changes in
salmonid fisheries.
The assessment considers multiple
geographic scales. The largest scale is the
Bristol Bay watershed, which is a largely
undisturbed region with outstanding
natural, cultural, and mineral resources.
Within the larger Bristol Bay watershed,
the assessment focuses on the Nushagak
and Kvichak River watersheds. These are
the largest of the Bristol Bay watershed's
six major river basins, containing about
50% of the total watershed area, and are
identified as mineral development areas
by the State of Alaska. Given its size and
-------
Sockeye salmon near Gibraltar Lake Photo: Thomas Quinn (University of Washington)
extent of characterization, the Pebble
deposit is the most likely site for near-
term, large-scale mine development in
the region. Because the Pebble deposit is
located in the headwaters of tributaries
to both the Nushagak and Kvichak Rivers,
both of these watersheds are subject to
potential risks from mining. The third
geographic scale is the watersheds of
the three tributaries that originate within
the potential footprint of a mine on the
Pebble deposit: the South Fork Koktuli
River, which drains the Pebble deposit
area and converges with the North
Fork west of the Pebble deposit; the
North Fork Koktuli River, located to the
northwest of the Pebble deposit, which
flows into the Nushagak River via the
Koktuli and Mulchatna Rivers; and Upper
Talarik Creek, which drains the eastern
portion of the Pebble deposit and
flows into the Kvichak River via Iliamna
Lake, the largest undeveloped lake in
the United States. The mine footprints
in the three realistic mine scenarios
evaluated in the assessment make up
the fourth geographic scale. These
scenarios—Pebble 0.25, Pebble 2.0,
and Pebble 6.5—define three potential
mine sizes, representing different stages
in the potential mining of the Pebble
deposit. The final geographic scale is the
combined area of the subwatersheds
between the mine footprints and the
Kvichak River watershed's eastern
boundary that would be crossed by a
transportation corridor linking the mine
site to Cook Inlet.
The assessment also addresses two
periods for mine activities. The first is
the development and operation phase,
during which mine infrastructure would
be built and the mine would be operated.
This phase may last from 20 to 100 years
or more. The second is the post-mining
phase, during which the site would
be monitored and maintained. Water
treatment and other waste management
activities would continue as necessary
and any failures would be remediated.
Because mine wastes would be persistent,
this period could continue for centuries
and potentially in perpetuity.
We began the assessment with a
thorough review of what is known about
the Bristol Bay watershed, its fisheries and
wildlife populations, and its Alaska Native
cultures. We also reviewed information
about copper mining and publicly
available information outlining proposed
-------
Bristol Bay
Russia Mainland & Islands
West Kamchatka
East Kamchatka
Western Alaska (excluding Bristol Bay)
- South Alaska Peninsula
Cook Inlet
Prince William Sound
Southeast Alaska
North British Columbia
South British Columbia, Washington & Oregon
Togiak
Nushagak
Kvichak
Naknek
Egegik
Ugashik
Proportion of Total Sockeye Salmon Run Sizes by (A) Region and (B) Watershed in the Bristol Bay Region. Values are averages from (A) 1956 to 2005
from Ruggerone et al. 2010 and (B) 1956 to 2010 from Baker pers. comm.
mine operations for the Pebble deposit.
The Pebble deposit has been the focus of
much exploratory study and has received
significant attention from groups in
and outside of Alaska. With the help of
regional stakeholders, we developed
a set of conceptual models to show
potential associations between salmon
populations and the environmental
stressors that might reasonably result
from large-scale mining. Then, following
the USEPA's ecological risk assessment
framework, we analyzed the sources and
exposures that would occur and potential
responses to those exposures. Finally, we
characterized the risks to fish habitats,
salmon, and other fish populations, as
well as the implications of those risks for
the wildlife and Alaska Native cultures
that use them.
This is not an in-depth assessment of a
specific mine, but rather an examination
of potential impacts of reasonably
foreseeable mining activities in the
Bristol Bay region, given the nature of
the watershed's mineral deposits and
the requirements for successful mine
development. The assessment analyzes
mine scenarios that reflect the expected
characteristics of mine operation at the
Pebble deposit. It is intended to provide
a baseline for understanding potential
impacts of mine development, not just at
the Pebble deposit but throughout the
Nushagak and Kvichak River watersheds.
The mining of other existing porphyry
copper deposits in the region would be
expected to include the same types of
activities and facilities evaluated in this
assessment for the Pebble deposit (open
pit mining and the creation of waste
rock piles and tailings storage facilities
[TSFs]), and therefore would present
potential risks similar to those outlined
in this assessment. However, because the
region's other ore bodies are believed to
be much smaller than the Pebble deposit,
those mines would likely be most similar
to the smallest mine scenario analyzed in
this assessment (Pebble 0.25).
-------
t--
^mm ^m*
Sockeye salmon near Pedro Bay, Iliamna Lake
Photo: Thomas Quinn (University of Washington)
This assessment considers many but
not all potential impacts associated
with future large-scale mining in the
Bristol Bay watershed. Although the
mine scenarios assume development
of a deep-water port on Cook Inlet to
ship product concentrate elsewhere
for smelting and refining, impacts
of port development and operation
are not assessed. The assessment
does not evaluate impacts of the one
or more large-capacity electricity-
generating power plants that would
be required to power the mine and
the port. We recognize that large-scale
mine development would induce the
development of additional support
services for mine employees and their
families, vacation homes and other
recreational facilities, and transportation
infrastructure beyond the main
corridor (i.e., airports, docks, and roads).
The assessment describes but does
not evaluate the effects of induced
development resulting from large-scale
mining in the region. Direct effects of
mining on Alaska Natives and wildlife are
not assessed. The assessment also does
not include a cost-benefit analysis and does
not compare mining to other ongoing
activities such as commercial fishing.
Ecological Resources
The Bristol Bay watershed provides
habitat for numerous animal species,
including at least 29 fish species, more
than 40 terrestrial mammal species,
and more than 190 bird species. Many
of these species are essential to the
structure and function of the region's
ecosystems and current economies. The
Bristol Bay watershed supports several
wilderness compatible and sustainable
economic sectors, such as commercial,
sport, and subsistence fishing; sport
and subsistence hunting; and non-
consumptive recreation. Considering all
these sectors, the Bristol Bay watershed's
ecological resources generated nearly
$480 million in direct economic
expenditures and sales in 2009 and
provided employment for over 14,000
full- and part-time workers.
Chief among these ecological resources
are world-class commercial and sport
fisheries for Pacific salmon and other
salmonids.The region's commercial
salmon fishery generates the largest
component of economic activity. The
watershed supports production of all
five species of Pacific salmon found in
North America: sockeye (Oncorhynchus
-------
Bald eagle Photo: Conrad Gowell (Oregon State University)
nerka), coho (O. kisutch), Chinook (O.
tshawytscha), chum (O. keta), and
pink (O. gorbuscha). These fishes are
anadromous, meaning that they hatch
and rear in freshwater systems, migrate
to sea to grow to adult size, and return
to freshwater systems to spawn and die.
Because no hatchery fish are raised or
released in the watershed, Bristol Bay's
salmon populations are entirely wild.
The most abundant salmon species in the
Bristol Bay watershed is sockeye salmon.
The watershed supports the largest
sockeye salmon fishery in the world, with
approximately 46% of the average global
abundance of wild sockeye salmon.
Between 1990 and 2009, the annual
average inshore run of sockeye salmon in
Bristol Bay was approximately 37.5 million
fish. Annual commercial harvest of
sockeye over this same period averaged
25.7 million fish. Approximately half of
Bristol Bay's sockeye salmon production
is from the Nushagak and Kvichak River
watersheds, the main area of focus for
this assessment.
Chinooksalmon are also abundant in the
region. Chinook returns to the Nushagak
River are consistently greater than
100,000 fish per year and have exceeded
200,000 fish in 11 years between 1966
and 2010, frequently placing Nushagak
River Chinook runs at or near the world's
largest. This is noteworthy given the
Nushagak River's small watershed area
compared to other Chinook-producing
rivers such as the Yukon River, which spans
Alaska and much of northwestern Canada,
and the Kuskokwim River in southwestern
Alaska, just north of Bristol Bay.
The Bristol Bay watershed also supports
populations of non-salmon fishes that
typically (but not always) remain in
the watershed's freshwater habitats
throughout their life cycles. The region
contains highly productive waters for
sport and subsistence fish species,
including rainbow trout (O. mykiss), Dolly
Varden (Salvelinus malma), Arctic char (5.
alpinus), lake trout (S. namaycush), Arctic
grayling (Thymallus arcticus), northern
pike (Esoxludus), and humpback
whitefish (Coregonus pidschian). These
fishes occupy a variety of habitats in the
watershed, from headwater streams to
wetlands to large rivers and lakes.
The Bristol Bay region is especially
renowned for the size and abundance
of its rainbow trout: between 2003 and
2007, an estimated 183,000 rainbow
trout were caught in the Bristol Bay
Management Area.
-------
The exceptional quality of the Bristol
Bay watershed's fish populations can be
attributed to several factors, the most
important of which is the watershed's
high-quality, diverse aquatic habitats
unaltered by human-engineered
structures and flow management
controls. Surface and subsurface
waters are highly connected, enabling
hydrologicand biochemical connectivity
between wetlands, ponds, streams,
and rivers and thereby increasing the
diversity and stability of habitats able to
support fish. These factors all contribute
to making the Bristol Bay watershed a
highly productive system. High aquatic
habitat diversity also supports the high
genetic diversity offish populations.
This diversity in genetics, life history,
and habitat acts to reduce year-to-
year variability in total production and
increase overall stability of the fishery.
The return of spawning salmon from
the Pacific Ocean brings marine-derived
nutrients into the watershed and fuels
both aquatic and terrestrial foodwebs.
Thus, the condition of Bristol Bay's
terrestrial ecosystems is intimately linked
to the condition of salmon populations,
as well as to almost totally undisturbed
terrestrial habitats. The watershed
continues to support large carnivores
such as brown bears (Ursusarctos), bald
eagles (Haliaeetusleucocephalus), and
gray wolves (Canis lupus); ungulates
such as moose (Alces alces gigas) and
caribou (Rangifer tarandus gmnti); and
numerous waterfowl and small mammal
species. Brown bears are abundant in the
Nushagak and Kvichak River watersheds.
Moose also are abundant, particularly
in the Nushagak River watershed where
felt-leaf willow, a preferred forage
species, is plentiful. The Nushagak and
Kvichak River watersheds are used by
caribou, primarily the Mulchatna caribou
herd. This herd ranges widely through
these watersheds, but also spends
considerable time in other watersheds.
Alaska Native Cultures
The predominant Alaska Native cultures
present in the Nushagak and Kvichak
River watersheds—the Yup'ik and
Dena'ina—are two of the last intact,
sustainable, salmon-based cultures
in the world. In contrast, other Pacific
Northwest salmon-based cultures are
severely threatened by development,
degraded natural resources, and
declining salmon resources.
-------
Salmon are integral to these cultures'
entire way of life via the provision of
subsistence food and subsistence-
based livelihoods, and are an important
foundation for their language, spirituality,
and social structure. The cultures have a
strong connection to the landscape and
its resources. In the Bristol Bay watershed,
this connection has been maintained for
at least 4,000 years and is in part both
due to and responsible for the continued
undisturbed condition of the region's
landscape and biological resources. The
respect and importance given salmon
and other wildlife, along with traditional
knowledge of the environment, have
produced a sustainable subsistence-
based economy. This subsistence-based
way of life is a key element of Alaska
Native identity and serves a wide range of
economic, social, and cultural functions in
Yup'ikand Dena'ina societies.
There are 31 Alaska Native villages in
the wider Bristol Bay region, 25 of which
are located in the Bristol Bay watershed.
Fourteen of these communities are
within the Nushagak and Kvichak River
watersheds, with a total population
of 4,337 in 2010. Thirteen of these 14
communities have federally recognized
tribal governments and a majority Alaska
Native population. Many of the non-
Alaska Native residents in the watersheds
have developed cultural ties to the
region and they also practice subsistence.
Virtually every household in the
watersheds uses subsistence resources. In
the Bristol Bay region, salmon constitute
approximately 52% of the subsistence
harvest; for some communities this
proportion is substantially higher.
The subsistence-based way of life in many
Alaska Native villages is augmented with
activities that support cash economy
transactions, including commercial fishing.
Alaska Native villages, in partnership
with Alaska Native corporations and
other business interests, are considering
a variety of economic development
opportunities. Some Alaska Native
villages have decided that large-scale
mining is not the course they would
like to pursue, whereas a few others are
seriously considering this opportunity.
All are concerned with the long-term
sustainability of their communities.
Geological Resources
In addition to significant and valuable
ecological resources, the Nushagak
and Kvichak River watersheds contain
considerable mineral resources.
-------
The potential for large-scale mine
development in the region is greatest for
copper deposits and, to a lesser extent,
for intrusion-related gold deposits.
Because these deposits are low-grade—
meaning that they contain relatively
small amounts of metals relative to the
amount of ore—mining will be economic
only if conducted over large areas and
will necessarily produce large amounts of
waste material.
The largest known and most explored
deposit is the Pebble deposit. If fully
mined, the claim holder estimates that
the Pebble deposit would produce
more than 11 billion tons of ore, which
would make it the largest mine of its
type in North America. A mine at the
Pebble deposit could ultimately generate
revenues between $300 billion to $500
billion over the life of the mine, as well
as provide more than 2,000 jobs during
mine construction and more than 1,000
jobs during mine operation.
Although the Pebble deposit represents
the most imminent site of mine
development, other mineral deposits
with potentially significant resources
exist in the Nushagak and Kvichak River
watersheds. Ten specific claims with
more than minimal recent exploration
(in addition to the Pebble deposit claim)
have been filed for copper deposits.
Most of these claims are near the
Pebble deposit. The potential impacts
of large-scale mining considered in this
assessment are generally applicable to
these other sites.
Mine Scenarios
Like all risk assessments, this assessment
is based on scenarios that define a set of
possible future activities and outcomes.
To assess mining-related stressors that
would affect ecological resources in
the watershed, we developed realistic
mine scenarios that include a range of
mine sizes and operating conditions.
These mine scenarios are based on the
Pebble deposit because it is the best-
characterized mineral resource and
the most likely to be developed in the
near term. The mine scenarios draw on
preliminary plans developed for Northern
Dynasty Minerals, consultation with
experts, and baseline data collected
by the Pebble Limited Partnership to
characterize the mine site, mine activities,
and the surrounding environment.
The exact details of any future mine
plan for the Pebble deposit or for other
deposits in the watershed will differ
from our mine scenarios. However,
our scenarios reflect the general
characteristics of mineral deposits in
10
-------
the watershed, modern conventional mining
technologies and practices, the scale of mining
activity required for economic development of
the resource, and the infrastructure needed to
support large-scale mining. Therefore, the mine
scenarios evaluated in this assessment realistically
represent the type of development plan that would
be anticipated for a porphyry copper deposit in
the Bristol Bay watershed. Uncertainties associated
with the mine scenarios are discussed later in this
executive summary.
The three mine scenarios evaluated in the
assessment represent different stages of mining
at the Pebble deposit, based on the amount of ore
processed: Pebble 0.25 (approximately 0.25 billion
tons [0.23 billion metric tons] of ore over 20 years),
Pebble 2.0 (approximately 2.0 billion tons [1.8 billion
metric tons] of ore over 25 years), and Pebble 6.5
(approximately 6.5 billion tons [5.9 billion metric
tons] of ore over 78 years). The major components
of each mine would be an open mine pit, waste
rock piles, and one or more TSFs. Other significant
features include plant and ancillary facilities (e.g.,
a water collection and treatment system, an ore-
processing facility, and other facilities associated
with mine operations) and the groundwater
drawdown zone (the area over which the water
table is lowered due to dewatering of the mine pit).
An underground extension of the mine, which could
increase the size of the mine to 11 billion tons of ore,
is not included in this assessment.
Mine Scenario Parameters
Parameter
Amount of ore mined (billion metric tons)
Approximate duration of mining (years)
Ore processing rate (metric tons/day)
Mine Pit
Surface area (km2)
Pebble 0.25
Pebble 2.0
198,000
Pebble 6.5
208,000
Waste Rock Pile
Surface area (km2)
PAG waste rock (million metric tons)
NAG waste rock (million metric tons)
Capacity, dry weight (billion metric tons)
Surface area, exterior (km2)
Maximum dam height (m)
TSF2a
Capacity, dry weight (billion metric
•face area, exterior (kir
Capacity, dry weight (billion metric
face area, exterior (kn
Total TSF surface area, exterior (I
Notes:" Final value, whenTSF is full; PAG = potentially acid-generating; NAG = non-acid-generating; TSF = tailings storage facility; NA = not applicable
-------
.
Pebble deposit area Photo: Lorraine Edmond (USEPA)
Each of these mine scenarios includes a
138-km (86-mile) transportation corridor;
113 km (70 miles) of the corridor would fall
within the Kvichak River watershed. This
corridor would include a gravel-surfaced
road and four pipelines (one each for
product concentrate, return water, diesel
fuel, and natural gas).
The assessment considers risks from
routine operation of a mine designed
using modern conventional design,
practices, and mitigation technologies,
assuming no significant human or
engineering failures. The assessment also
considers various types of failures that
have occurred during the operation of
other mines and that could occur in this
case, including failures of a wastewater
treatment plant, a tailings dam, pipelines,
and culverts.
Risks to Salmon and
Other Fishes
Based on the mine scenarios, the
assessment defines mining-related
stressors that would affect the Bristol Bay
watershed's fish and consequently affect
wildlife and human welfare. The scenarios
include both routine operations and
several potential failure scenarios.
A -
A .
1 2 ^| Pebble 0.25 Components
•=H Kilometers _ Pebble 2.0 Components
^™— — ' Miles Pebble 6.5 Components
1 *-'
12
Major Mine Components for the Three Scenarios Evaluated in the Assessment.
Pebble 0.25 represents 0.25 billion tons of ore; Pebble 2.0 represents 2.0 billion tons of
ore; Pebble 6.5 represents 6.5 billion tons of ore. Each mine footprint includes the mine
components shown here, as well as the drawdown zone and the area covered by plant and
ancillary facilities. Light blue areas indicate streams and rivers from the National Hydrography
Dataset (USGS 2012) and lakes and ponds from the National Wetlands Inventory (USFWS
2012); dark blue areas indicate wetlands from the National Wetlands Inventory (USFWS 2012).
-------
m
Lake Clark
»o
c>V Nc
if Approximate Pebble Deposit Location
™ — Transportation Corridor (Outside Assessment Area)
^™ Transportation Corridor
— Existing Roads
I I Transportation Corridor Area
1 _ I Subwatersheds within Area
Cook Inlet
The Transportation Corridor Area, Comprising 32 Subwatersheds in the Kvichak River Watershed that Drain to Iliamna Lake. Subwatersheds are
defined by 12-digit hydrologic unit codes according to the National Hydrography Dataset (USGS 2012).
Mine Footprint
Effects on fish resulting from habitat
loss and modification would occur
directly in the area of mine activity and
indirectly downstream because of habitat
destruction. These habitat loss estimates
are believed to be low due to incomplete
delineation of streams, wetlands, and
salmon distribution across the region.
However, it is possible that careful siting of
mine facilities could reduce habitat losses
to some degree.
Due to the mine footprint (the area covered
by the mine pit, waste rock piles, TSFs,
groundwater drawdown zone, and plant
and ancillary facilities), 38,89, and 151 km
(24,55, and 94 miles) of streams would
be lost—that is, eliminated, blocked, or
dewatered—in the Pebble 0.25,2.0, and
6.5 scenarios, respectively. This translates
to losses of 8,22, and 36 km (5,14, and
22 miles) of streams known to provide
spawning or rearing habitats for coho
salmon, sockeye salmon, Chinook salmon,
and Dolly Varden.
Altered streamflow due to retention
and discharge of water used in mine
operations, ore processing and transport,
and other mine activities would reduce
the amount and quality offish habitat.
Streamflow alterations exceeding 20%
would adversely affect habitat in an
additional 15,27, and 53 km (9.3,17,
and 33 miles) of streams in the
Pebble 0.25,2.0, and 6.5 scenarios,
respectively, reducing production of
sockeye salmon, coho salmon, Chinook
salmon, rainbow trout, and Dolly
Varden. Reduced streamflows would
also result in the loss or alteration of an
unquantifiable area of riparian floodplain
wetland habitat due to loss of hydrologic
connectivity with streams.
Off-channel habitats for salmon and
other fishes would be reduced due to
losses of 4.5,12, and 18 km2 (1,200,
3,000 and 4,900 acres) of wetlands
and 0.41,0.93, and 1.8 km2 (100,230,
and 450 acres) of ponds and lakes to
the mine footprints in the Pebble 0.25,
2.0, and 6.5 scenarios, respectively.
13
-------
These losses would reduce availability of
and access to hydraulically and thermally
diverse habitats that provide enhanced
foraging opportunities and important
rearing habitats for juvenile salmon.
Indirect effects of stream and wetland
losses would include reductions in
the quality of downstream habitat for
coho salmon, sockeye salmon, Chinook
salmon, rainbow trout, and Dolly Varden.
Although these indirect effects cannot be
quantified, such effects would be expected
to diminish fish production downstream of
the mine site because fish depend on these
habitats. Indirect effects would be caused
by the following alterations.
• Reduced food resources would result
from the loss of organic material and
drifting invertebrates from streams
and streamside wetlands lost to the
mine footprint.
• The balance of surface water and
groundwater inputs to downstream
reaches would shift, potentially
reducing winter fish habitat and
making streams less suitable for
spawning and rearing.
A
0 1 2
0 1 2
] Miles
Pebble 6.5 Components __
Groundwater Drawdown Zone * -^ -- <
A ^
Eliminated, Blocked, or Dewatered •\-^S^' *N
Streams, Lakes, and Ponds
Eliminated, Blocked, or Dewatered
Wetlands
14
Streams and Wetlands Lost (Eliminated, Blocked, or Dewatered) in the
Pebble 6.5 Scenario. Light blue areas indicate streams and rivers from the
National Hydrography Dataset (USGS 2012) and lakes and ponds from the National
Wetlands Inventory (USFWS 2012); dark blue areas indicate wetlands from the
National Wetlands Inventory (USFWS 2012).
-------
Summary of Estimated Stream Lengths Potentially Affected in the Three Mine Scenarios, Assuming Routine Operations
Pebble 0.25
Stream Length Affected (km)
Pebble 2.0
Pebble 6.5
Eliminated, blocked, or dewatered
Eliminated, blocked, or dewatered—anadromous
>20% streamflow alteration3
Direct toxicity to fisha
Direct toxicity to invertebrates3
n"wnstream of transportation corrid™
Summary of Estimated Wetland, Pond, and Lake Area Potentially Affected in the Three Mine Scenarios, Assuming Routine Operations
Pebble 0.25
Lost to the mine footprint
Lost to reduced streamflow below mine footprint
Filled by roadbed
Influenced by the road (within 200 m)
Wetland, Pond, and Lake Area Affected (km2)
Pebble 2.0
Unqualified
Notes:" Stream reaches with streamflow alterations partially overlap those with toxicity.
Pebble 6.5
• Seasonal temperatures could be altered by
water treatment and reduced groundwater
flowpaths, making streams less suitable
forsalmonids.
Water Quality
Leakage during Routine Operations
Water from the mine site would enter streams
through wastewater treatment plant discharges
and in uncollected runoff and leakage of
leachates from the waste rock piles and TSFs.
Wastewater treatment is assumed to meet
all state standards and national criteria, or
equivalent benchmarks for chemicals that have
no criteria. However, water quality would be
diminished by uncollected leakage of tailings
and waste rock leachates from the containment
system, which would occur during routine
operations. Test leachates from the tailings and
non-acid-generating waste rocks are mildly
toxic. They would require an approximately
two-fold dilution to achieve water quality criteria
for copper, but are not estimated to be toxic
to salmonids. Waste rocks associated with the
ore body are acid-forming with high copper
concentrations in test leachates, and would
require 2,900- to 52,000-fold dilution to achieve
water quality criteria. Several metals could be
sufficiently elevated to contribute to toxicity, but
copper is the dominant toxicant.
Uncollected leachates from waste rock
piles and TSFs would elevate instream
copper levels and cause direct effects
on salmonids ranging from aversion and
avoidance of the contaminated habitat
to rapidly induced death of many or all
fish. Avoidance of streams by salmonids
would occur in 24 and 34 to 57 km (15
and 21 to 35 miles) of streams in the
Pebble 2.0 and Pebble 6.5 scenarios,
respectively. Rapidly induced death of
many or all fish would occur in 12 km
(7.4 miles) of streams in the Pebble 6.5
scenario. Copper would cause death
or reduced reproduction of aquatic
invertebrates in 21,40 to 62, and 60 to
82 km (13,25 to 38, and 37 to 51 miles)
of streams in the Pebble 0.25,2.0,
and 6.5 scenarios, respectively. These
invertebrates are the primary food source
for juvenile salmon and all life stages of
other salmonids, so reduced invertebrate
productivity would be expected to
reduce fish productivity. These results
are sensitive to the assumed efficiency
of the leachate capture system, and a
more efficient system could be devised.
However, greater than 99% capture
efficiency would be required to prevent
exceedance of the copper criteria for
15
-------
the South Fork Koktuli River in the
Pebble 6.5 scenario, which would require
technologies beyond those specified in
our scenarios or identified in the most
recent preliminary mine plan.
Wastewater Treatment Plant Failure
Based on a review of historical and
currently operating mines, some failure of
water collection and treatment systems
would be expected to occur during
operation or post-closure periods. A
variety of water collection and treatment
failures are possible, ranging from
operational failures that result in short-
term releases of untreated or partially
treated leachates to long-term failures to
operate water collection and treatment
systems in perpetuity. A reasonable but
severe failure scenario would involve a
complete loss of water treatment and
release of average untreated wastewater
flows into average dilution flows. In that
failure scenario, copper concentrations
would be sufficient to cause direct
effects on salmonids in 27,64 to 87, and
74 to 97 km (17,40 to 54, and 46 to 60
miles) of streams in the Pebble 0.25,
2.0, and 6.5 scenarios, respectively.
Aquatic invertebrates would be killed
or their reproduction reduced in 78 to
100 km (48 to 62 miles) of streams in all
three scenarios. In the Pebble 2.0 and
6.5 scenarios, a fish kill would occur
rapidly in 3.8 and 31 km (2.4 and 19
miles) of streams, respectively, following
treatment failure.
Spillway Release
In the event of TSF overfilling,
supernatant water would be released via
a spillway. If the water was equivalent to
the test tailings supernatant, 2.6 km (1.6
miles) of streams would be avoided by
fish and 3.4 to 23 km (2.1 to 14 miles) of
streams would be toxic to invertebrates,
independent of other sources.
Transportation Corridor
Construction and Routine Operation
In the Kvichak River watershed, the
transportation corridor would cross
approximately 64 streams and rivers. Of
those, 55 are known or likely to support
migrating and resident salmonids,
including 20 streams designated as
anadromous waters at the location of
the crossing. The corridor would run
near Iliamna Lake and cross multiple
tributary streams near their confluences
with the lake. These habitats are
important spawning areas for sockeye
salmon, putting sockeye particularly at
16
-------
Probabilities and Consequences of Potential Failures in the Mine Scenarios
Probability3
Consequences
Product concentrate pipeline
1(r per km-year = 95% chance per pipeline in 25 years
Most failures would occur between stream or wetland crossings and
might have little effect on fish.
Concentrate spill into a stream
1.5x 102 per year = 1 stream-contaminating spill in 78
Fish and invertebrates would experience acute exposure to toxic water
and chronic exposure to toxic sediment in a stream and potentially
extending to Iliamna Lake.
Concentrate spill into a wetland
2.6 x 102 per year = 2 wetland-contaminating spills in
78 years
Invertebrates and potentially fish would experience acute exposure to
toxic water and chronic exposure to toxic sediment in a pond or other
wetland.
Return water pipeline spill
Same as product concentrate pipeline
Fish and invertebrates would experience acute exposure to toxic water
if return water spilled to a stream or wetland.
Diesel pipeline spill
Same as product concentrate pipeline
Acute toxicity would reduce the abundance and diversity of
invertebrates and possibly cause a fish kill if diesel spilled to a stream
or wetland.
Culvert, operation
Frequent inspections and regular maintenance would result in few
impassable culverts, but for those few, blockage of migration could
persist for a migration period, particularly for juvenile fish.
Culvert, post-operation
3 x 10"1 to ~6 x 10"1 per culvert; instantaneous = 11 to
22 culverts
In surveys of road culverts, 30 to 61% are impassable to fish at any one
time. This would result in 11 to 22 salmonid streams blocked at any one
time. In 10 to 19 of the 32 culverted streams with restricted upstream
habitat, salmon spawning may fail or be reduced and the streams would
likely not be able to support long-term populations of resident species.
Truck accidents
1.9 x 107 spills per mile of travel = 4 accidents in 25
years and 2 near-stream spills in 78 years
Accidents that spill processing chemicals into a stream or wetland could
cause a fish kill. A spill of molybdenum concentrate may also be toxic.
Water collection and treatment,
operation
0.93 = proportion of recent U.S. porphyry copper mines
with reportable water collection and treatment failures.
Water collection and treatment failures could result in exceedance
of standards potentially including death offish and invertebrates.
However, these failures would not necessarily be as severe or extensive
as estimated in the failure scenario, which would result in toxic effects
from copper in more than 60 km of stream habitat.
Tailings storage facility spillway
release
No data, but spills are known to occur and are sufficiently
frequent to justify routine spillway construction
Spilled supernatant from the tailings storage facility could result in
toxicity to invertebrates and fish avoidance for the duration of the
event.
Water collection and treatment,
managed post-closure
Somewhat higher than operation
Post-closure collection and treatment failures are very likely to result
in release of untreated or incompletely treated leachates for days
to months, but the water would be less toxic due to elimination of
potentially acid-generating waste rock.
Water collection and treatment, Certain, by definition
When water is no longer managed, untreated leachates would flow to
' Because of differences in derivation, the probabilities are not directly comparable.
b Based on expected state safety requirements. Observed failure rates for earthen dams are higher (about 5x104 per year or a recurrence frequency of 2,000 years).
-------
Note: Sampling intensity is greatly reduced away from the Pebble
deposit area. Streams without data points may not have been
surveyed; thus, it is unknown whether or not they contain these species.
•^T Approximate Pebble Deposit Location
• •.. Transportation Corridor (Outside Assessment Area)
^™ Transportation Corridor
I I Transportation Corridor Area
1 1 Subwatersheds within Area
i^™ Salmon
A DollyVarden
O Rainbow Trout
\\ f*yj
"na-i ^
Bay
Reported Salmon, Dolly Varden, and Rainbow Trout Distribution Along the Transportation Corridor. Salmon presence data are from the
Anadromous Waters Catalog (Johnson and Blanche 2012); Dolly Varden and rainbow trout presence data are from the Alaska Freshwater Fish Inventory (ADF&G
2012). Note that rainbow trout have also been documented in the Iliamna River and Chinkelyes Creek, although these points are not indicated on this map.
risk from the road. Diminished habitat
quality in streams and wetlands below
road crossings would result primarily
from altered streamflow, runoff of road
salts, and siltation of habitat for salmon
spawning and rearing and invertebrate
prey production.
Culvert Failure
Culverts commonly fail to allow free
passage of fish. They can become
blocked by debris or ice that may not
stop water flow but that create a barrier
to fish movement. Fish passage also may
be blocked or inhibited by erosion below
a culvert that "perches" the culvert and
creates a waterfall, by shallow water
caused by a wide culvert and periodic
low streamflows, or by excessively high
gradients. If blockages occurred during
adult salmon immigration or juvenile
salmon emigration and were not cleared
for several days, production of a year-
class (i.e., fish spawned in the same year)
would be lost from or diminished in the
stream above the culvert.
Culverts can also fail to convey water due
to landslides or, more commonly, floods
that wash out undersized or improperly
installed culverts. In such failures, the
stream would be temporarily impassible
to fish until the culvert is repaired or until
erosion re-establishes the channel. If the
failure occurs during a critical period in
salmon migration, effects would be the
same as with a debris blockage (i.e., a lost
or diminished year-class).
Culvert failures also would result in the
downstream transport and deposition of
silt, which could cause returning salmon
to avoid a stream if they arrived during
or immediately following the failure.
Deposition of silt would smother salmon
eggs and alevins if they were present,
and would degrade downstream habitat
for salmonids and the invertebrates that
they eat.
Blockages of culverts could persist
for as long as the intervals between
culvert inspections. We assume that
the transportation corridor would be
inspected daily and maintained during
18
-------
- '^j&£;^ ;
-••
mine operation. The level of surveillance
along the corridor can be expected to
affect the frequency of culvert failure
detection. Driving inspections would
likely identify a single erosional failure
of a culvert that damaged the road, or
a debris blockage sufficient to cause
water to pool above the road. However,
long-term fixes may not be possible
until conditions are suitable for culvert
replacement, and these fixes may not
fully address fish passage, which may be
reduced or blocked for longer periods.
Extended blockage of migration would
be less likely if daily road inspections
included stops to inspect each end of
each culvert.
After mine operations cease, the road
would likely be maintained less carefully
by the operator or may be transferred
to a government entity that would be
expected to employ a more conventional
inspection and maintenance schedule. In
either case, the proportion of impassable
culverts at any one time would be
expected to revert to levels found in
published surveys of public roads (mean
of 48% [range of 30 to 61%] of culverts
that had failed and not been repaired
when surveyed). Of the approximately 45
culverts that would be required, 36 would
be on streams that are believed to support
salmonids. Hence, 11 to 22 streams would
be expected to have impeded passage
of salmon, rainbow trout, or Dolly Varden
for an indefinite period of time, and some
proportion of those streams would have
degraded downstream habitat resulting
from sedimentation following washout of
the road.
Truck Accidents
Trucks would carry ore processing
chemicals to the mine site and
molybdenum product concentrate to the
port. Truck accident records indicate that
truck accidents near streams are likely
over the long period of mine operation.
These accidents could release sodium ethyl
xanthate, cyanide, other process chemicals,
or molybdenum product concentrate
to streams or wetlands, resulting in toxic
effects on invertebrates and fish. However,
the risk of spills could be mitigated by using
impact-resistant containers.
Tailings Dam Failure
Tailings are the waste materials
produced during ore processing. In
our scenarios, these wastes would be
stored in TSFs consisting of tailings dams
and impoundments. The probability
of a tailings dam failure increases with
the number of dams. The Pebble 0.25
19
-------
Transamerica Building - 260 Meters
lings Dam TSF1- 209 f
Gateway Arch - 1 92 Meters
Washington Monument -169 Meters
Tailings Reservoir
Height of the Dam at TSF1 in the Pebble 2.0 and Pebble 6.5 Scenarios, Relative to U.S. Landmarks
scenario would include one TSF with
a single dam, the Pebble 2.0 scenario
would include one TSF with three dams,
and the Pebble 6.5 scenario would
include three TSFs with a total of eight
dams. Because their removal is not
feasible, the TSFs and their component
dams would be in place for hundreds to
thousands of years, long beyond the life
of the mine. Available reports from the
Pebble Limited Partnership suggest a
tailings dam as high as 209-m (685 feet)
at TSF 1. We evaluated two potential dam
failures at TSF 1 in this assessment: one
at a volume approximating the complete
Pebble 0.25 scenario (92-m dam height)
and one at a volume approximating the
complete Pebble 2.0 scenario (209-m
dam height). In both cases we assumed
20% of the tailings would be released, a
conservative estimate that is well within
the range of historical tailings dam
failures. Failures of the TSF 2 and TSF
3 tailings dams were not analyzed but
would be expected to be similar in terms
of types of effects.
The range of estimated dam failure
probabilities is wide, reflecting the great
uncertainty concerning such failures.
The most straightforward method of
estimating the annual probability of a
tailings dam failure is to use the historical
failure rate of similar dams. Three reviews
of tailings dam failures produced an
average rate of approximately 1 failure
per 2,000 dam-years, or 5 x 10~4 failures
per dam-year. Strictly speaking, these
frequencies are properties that apply to
a group of dams. However, by extension,
if there is one dam and it is typical of
the population, it would be expected
to fail, on average, within a 2,000-year
period. This does not mean it is expected
to fail 2,000 years after it is built. Rather,
it indicates that, after 2,000 years have
passed, it is more likely than not that
the dam would have failed and that
expected failure could occur any year in
that 2,000-year window with an average
annual probability of 0.0005.
The argument against this method is
that the record of past failures does not
fully reflect current engineering practice.
Some studies suggest that improved
design, construction, and monitoring
practices can reduce the failure rate by
an order of magnitude or more, resulting
in an estimated failure probability
within the range assumed here. The
State of Alaska's guidelines suggest that
an applicant follow accepted industry
design practices such as those provided
20
-------
r : • •
by the U.S. Army Corps of Engineers
(USAGE), the Federal Energy Regulatory
Commission (FERC), and other agencies.
Based on safety factors in USAGE and
FERC guidance, we estimate that the
probability of failure for all causes
requires a minimum factor of safety
of 1.5 against slope instability for the
loading condition corresponding to
steady seepage from the filled storage
facility. An assessment of the correlation
of dam failure probabilities with slope
instability safety factors suggests an
annual probability of failure of 1 in
250,000 per year for facilities designed,
built, and operated with state-of-
the-practice engineering (Category
I facilities) and 1 in 2,500 per year for
facilities designed, built, and operated
using standard engineering practice
(Category II facilities). The advantage of
this approach is that it addresses current
regulatory guidelines and engineering
practices. The disadvantage is that we do
not know whether standard practice or
state-of-the practice dams will perform as
expected, particularly given the potential
dam heights and subarctic conditions in
these scenarios.
Failure of the dam at TSF 1 (the TSF
included in all three mine scenarios)
would result in the release of a flood of
tailings slurry into the North Fork Koktuli
River. This flood would scour the valley
and deposit many meters of tailings fines
in a sediment wedge across the entire
valley near the TSF dam, with lesser
quantities of fines deposited as far as the
North Fork's confluence with the South
Fork Koktuli River. The North Fork Koktuli
River currently supports spawning and
rearing populations of sockeye, coho, and
Chinook salmon; spawning populations
of chum salmon; and rearing populations
of Dolly Varden and rainbow trout. The
tailings slurry flood would continue
down the mainstem Koktuli River with
similar effects, the extent of which
cannot be estimated at this time due to
model and data limitations.
The tailings dam failures evaluated in
the assessment would be expected to
have the following severe direct and
indirect effects on aquatic resources,
particularly salmonids.
It is expected that the North Fork
Koktuli River below the TSF 1 dam and
much of the mainstem Koktuli River
would not support salmonids in the
short term (less than 10 years).
21
-------
• In the tailings dam failure scenarios,
spilled tailings would bury salmon
habitat under meters of fines along
nearly the entire length of the North
Fork Koktuli River valley downstream
of the dam (over 29 km or 18 miles in
the Pebble 0.25 dam failure scenario),
and beyond (in the Pebble 2.0 dam
failure scenario).
• Deposited tailings would degrade
habitat quality for both fish and the
invertebrates they eat. Based largely
on their copper content, deposited
tailings would be toxic to benthic
macroinvertebrates, but existing data
concerning toxicity to fish are less clear.
• Deposited tailings would continue to
erode from the North Fork Koktuli River
and mainstem Koktuli River valleys.
• Suspension and redeposition of tailings
would be expected to cause serious
habitat degradation in the mainstem
Koktuli River and downstream into
the Mulchatna River; however, the
extent of these effects cannot be
estimated at this time due to model
and data limitations.
The affected streams would provide
low-quality spawning and rearing
habitat for a period of decades.
• Recovery of suitable substrates via
mobilization and transport of tailings
would take years to decades, and
would affect much of the watershed
downstream of the failed dam.
• Ultimately, spring floods and
stormflows would carry some of the
tailings into the Nushagak River.
• For some years, periods of high
flow would be expected to suspend
sufficient concentrations of tailings to
cause avoidance, reduced growth and
fecundity, and even death offish.
Near-complete loss of North Fork
Koktuli River fish populations
downstream of the TSF and additional
fish population losses in the mainstem
Koktuli, Nushagak, and Mulchatna
Rivers would be expected to result
from these habitat losses.
• The Koktuli River watershed is an
important producer of Chinook salmon.
The Nushagak River watershed, of which
the Koktuli River watershed is a part, is
the largest producer of Chinook salmon
22
-------
The Koktuli River Photo: Jeff Frithsen (USEPA)
in the Bristol Bay region, with annual
runs averaging over 190,000 fish.
• A tailings spill could eliminate 29% or
more of the Chinook salmon run in
the Nushagak River due to loss of the
Koktuli River watershed population. An
additional 10 to 20% could be lost due
to tailings deposited in the Mulchatna
River and its tributaries.
• Sockeye are the most abundant
salmon returning to the Nushagak
River watershed, with annual runs
averaging more than 1.9 million fish.
The proportion of sockeye and other
salmon species of Koktuli-Mulchatna
origin is unknown.
• Similarly, the North Fork Koktuli River
populations of rainbow trout and
Dolly Varden would be lost for years to
decades if they could not successfully
be maintained entirely in headwater
networks upstream of the affected zone.
Quantitative estimates of these losses are
not possible given available information.
Effects would be qualitatively similar
for both the Pebble 0.25 and Pebble 2.0
dam failures, although effects from the
Pebble 2.0 dam failure would extend
farther and last longer. Failure of dams
at the two additional TSFs in the Pebble
6.5 scenario (TSF 2 and TSF 3) were not
modeled, but would have similar types
of effects in the South Fork Koktuli River
and downstream rivers.
Pipeline Failure
In the mine scenarios, the primary mine
product would be a sand-like copper
concentrate with traces of other metals,
which would be pumped via pipeline to
a port on Cook Inlet. Water that carried
the concentrate would be returned
to the mine site in a second pipeline.
Based on the general record of pipelines
and further supported by the record of
metal concentrate pipelines at existing
mines, one near-stream failure and two
near-wetland failures of each of these
pipelines would be expected to occur
over the life of the Pebble 6.5 scenario
(approximately 78 years).
Failure of either the product or the
return water pipeline would release
water that is expected to be highly toxic
due to dissolved copper and possibly
processing chemicals. Invertebrates and
potentially early fish life stages would
be killed in the affected stream over a
relatively brief period. If concentrate
23
-------
spilled into a stream, it would settle
and form highly toxic bed sediment
based on its high copper content and
acid generation. The mean velocities of
many streams crossed by the pipelines
are sufficient to carry the concentrate
downstream to Iliamna Lake, but some
would collect in low-velocity areas of
the receiving stream. If the spill occurred
during low streamflows, dredging could
recover some concentrate but would
cause physical damage to the stream.
Concentrations in Iliamna Lake could not
be predicted, but near the pipeline route
Iliamna Lake contains important sockeye
salmon beach spawning areas that would
be exposed to a spill. Sockeye also spawn
in the lower reaches of streams that
could be directly contaminated by a spill.
Based on petroleum pipeline failure
rates, the diesel fuel pipeline also would
be expected to spill near a stream over
the life of the Pebble 6.5 mine. Evidence
from modeling the dissolved and
dispersed oil concentrations in
streams, laboratory tests of diesel
toxicity, and studies of actual spills in
streams indicates that a diesel spill at
a stream crossing would be expected
to immediately kill invertebrates and
likely fish as well. Remediation would be
difficult but recovery would be expected
to occur within 3 years. Failure of the
natural gas pipeline would also be
expected, but significant effects on fish
would not be expected.
Spills into wetlands that support fish
would be expected to have greater toxic
effects because contaminants would
be washed out slowly, if at all. However,
retention of contaminants within the
wetland would make remediation by
removal more practical.
Common Mode Failures
Multiple, simultaneous failures could
occur due to a common event, such as
a severe storm with heavy precipitation
(particularly precipitation that fell on
spring snow cover) or a major earthquake.
Over the long period that tailings
impoundments, a mine pit, and waste
rock piles would be in place, the likelihood
of multiple extreme precipitation events,
earthquakes, or combinations of these
events becomes much greater. Multiple
events further increase the chances that
facilities remaining in place will weaken and
eventually fail.
24
-------
?.=i ;•*; •-»-•• •* f. >
A/. 'J \
v- %'./'-
Such an event could cause multiple tailings
dam failures that would spill tailings
slurry into both the South and North Fork
Koktuli Rivers; road culvert washouts that
would send sediments downstream and
potentially block fish passage; and pipeline
failures that would release product slurry,
return water, or diesel fuel. The effects
of each of these accidents individually
would be the same as discussed
previously, but their co-occurrence would
cause cumulative effects on salmonid
populations and make any remedial
responses more difficult.
Fish-Mediated Risks
to Wildlife
Although the effects of salmonid
reductions on wildlife—that is, fish-
mediated risks to wildlife—cannot be
quantified given available data, some
reduction in wildlife would be expected
in the mine scenarios. Changes in the
occurrence and abundance of salmon
have the potential to change animal
behavior and reduce wildlife population
abundances. The mine footprints would
be expected to have local effects on
brown bears, wolves, bald eagles, and
other wildlife that consume salmon,
due to reduced salmon abundance
from habitat loss and degradation in or
immediately downstream of the mine
footprint. Any of the accidents or failures
evaluated would increase effects on
salmon, which would further reduce the
abundance of their predators.
The abundance and production of wildlife
also is enhanced by the marine-derived
nutrients that salmon carry upstream on
their spawning migration. These nutrients
are released into streams when the
salmon die, enhancing the production of
other aquatic species that feed wildlife.
Salmon predators deposit these nutrients
on the landscape, thereby fertilizing
terrestrial vegetation that, in turn,
provides food for moose, caribou, and
other wildlife. The loss of these nutrients
due to a reduction in salmon would be
expected to reduce the production of
riparian and upland species.
Fish-Mediated Risks to
Alaska Native Cultures
Under routine operations with no major
accidents or failures, the predicted loss
and degradation of salmonid habitat
in the South and North Fork Koktuli
Rivers and Upper Talarik Creek would be
expected to have some impact on Alaska
25
-------
Kvichak River below Iliamna Lake and Igiugig Photo: Joe Ebersole (USEPA)
Native cultures of the Nushagak and
Kvichak River watersheds. Fishing and
hunting practices would be expected
to change in direct response to the
stream, wetland, and terrestrial habitats
lost to the mine footprints and the
transportation corridor. It is also possible
that subsistence use of salmon resources
would decline based on perceptions of
reduced fish or water quality resulting
from mining.
The potential for significant effects on
Alaska Native cultures is much greater
from mine failures that reduced or
eliminated fish populations in affected
areas, including areas significant
distances downstream from the
mine. In the case of the tailings dam
failures described in the assessment,
the significant loss of Chinook salmon
populations would have severe
consequences, especially for villages in
the Nushagak River watershed.
Any loss offish production from these
failures would reduce the availability
of these subsistence resources to local
Alaska Native villages, and the reduction
of this highly nutritious food supply
could have negative consequences on
human health. Because salmon-based
subsistence is integral to Alaska Native
cultures, the effects of salmon losses
go beyond the loss of food resources. If
salmon quality or quantity was (or was
perceived to be) adversely affected, the
nutritional, social, and spiritual health of
Alaska Natives would decline.
Cumulative Risks of
Multiple Mines
This assessment has focused on the
effects that a single large mine at the
Pebble deposit would have on salmon
and other resources in the Nushagak and
Kvichak River watersheds, including the
cumulative effects of multiple stressors
associated with that mine. However,
multiple mines and their associated
infrastructure may be developed in these
watersheds. Each mine would pose risks
similar to those identified in the mine
scenarios. Estimates of the stream and
wetland habitats lost would differ across
different deposits, based on the size
and location of mine operations within
the watersheds. Individually, each mine
footprint would eliminate some amount
of fish-supporting habitat and, should
operator or engineering failures occur,
affect fish habitats well beyond the
mine footprint.
26
-------
We considered development of mines
at the Pebble South/PEB, Big Chunk
South, Big Chunk North, Groundhog,
AUDN/lliamna, and Humble claims
in the Nushagak and Kvichak River
watersheds. These sites were chosen
because all contain copper deposits that
have generated exploratory interest. If
all six mine sites were developed, the
cumulative area covered by these six
mine footprints could be 37 to 57 km2
(9,100 to 14,000 acres). Stream habitats
eliminated or blocked could be 43 to 70
km (27 to 43 miles). Cumulative wetland
losses could be 7.9 to 27 km2 (2,000 to
6,700 acres).
These are conservative estimates
of habitat loss, because we did not
estimate the hydrologic drawdown
zones around each mine pit as was done
for the Pebble scenarios. Inclusion of
the drawdown area in the Pebble 0.25
scenario increased the area of stream and
wetlands losses by roughly 50%. A similar
increase might be expected at the other
mine sites, depending on local geology.
These mines also would be expected to
modify streamflows and diminish water
quality to approximately the same extent
as the Pebble 0.25 scenario. Waters on
these claim blocks include the Chulitna
River and Rock, Jensen, Yellow, Napotoli,
Klutuk, and Kenakuchuk Creeks, as well
as over 250 unnamed tributaries and over
50 unnamed lakes and ponds. Although
not all support salmon, many do. Loss of
substantial habitat across the watersheds
could contribute to diminishing the
genetic diversity of salmon stocks and
thereby increasing annual variability in
salmon returns.
Mitigation and Remediation
The mine scenarios assessed here
include modern conventional mitigation
practices as reflected in Northern
Dynasty Mineral's published plan for
the Pebble deposit, plus practices
suggested in the mining literature
and consultations with experts. These
practices include, but are not limited
to, processing all potentially acid-
generating waste rock before closure,
managing effluent water temperatures,
inspecting and maintaining roads daily,
and providing automatic monitoring
and remote shut-off for the pipelines.
However, we recognize that risks could
be further reduced by unconventional
or even novel mitigation measures, such
as dry stack tailings disposal or the use
of armored containers on the trucks
carrying process chemicals to the site.
These practices may be unconventional
because they are expensive, unproven,
or impractical. However, these obstacles
to implementation might be overcome
and justified by the large mineral resource
and the highly valued natural and cultural
resources of the Bristol Bay watershed.
Although remediation would be
considered if spills contaminated
streams, features of the Pebble deposit
area would make remediation difficult.
Spilled tailings from a dam failure would
flow into streams, rivers, and floodplains
that are in roadless areas and that are not
large enough to float a barge-mounted
dredge. Recovery, transport, and disposal
of hundreds of millions of metric tons of
tailings under those conditions would
be extremely difficult and would result
in additional environmental damage.
Compensatory mitigation measures
could offset some of the stream and
wetland losses, although there are
substantial challenges regarding the
efficacy of these measures to offset
adverse impacts. Pipeline crossings of
streams would be near Iliamna Lake,
so the time available to block or collect
spilled material before it reached the
lake would be short. Spilled return
water and the aqueous phase of the
product concentrate slurry would be
unrecoverable. The product concentrate
itself would resemble fine sand, and
mean velocities in many receiving
streams would be sufficient to suspend
and transport it. Hence, concentrate
27
-------
—
UpperTalarik Creek floodplain Photo: Joe Ebersole (USEPA)
spilled or washed into streams could
be recovered only where it collected in
low-velocity locations. Diesel spills would
dissolve, vaporize, and flow as a slick to
Iliamna Lake. Booms and absorbents are
not very effective in moderate- to high-
velocity streams.
Summary of Uncertainties in
Mine Design and Operation
This assessment considers realistic mine
scenarios that are based on specific
characteristics of the Pebble deposit and
preliminary plans proposed by Northern
Dynasty Minerals. These scenarios
are generally applicable to copper
deposits in the Bristol Bay watershed.
If the Pebble deposit is mined, actual
events will undoubtedly deviate from
these scenarios. This is not a source of
uncertainty, but rather an inherent aspect
of a predictive assessment. Even an
environmental assessment of a specific
plan proposed for permitting by a mining
company would be an assessment of a
scenario that undoubtedly would differ
from actual mine development.
Multiple uncertainties are inherent
in planning, designing, constructing,
operating, and closing a mine. These
uncertainties, summarized below, are
inherent in any complex enterprise,
particularly when that enterprise involves
an incompletely characterized natural
system. However, the large spatial scales
and long durations required to mine
the Pebble deposit make these inherent
uncertainties more prominent.
• Mines are complex systems requiring
skilled engineering, design, and
operation. The uncertainties facing
mining and geotechnical engineers
include unknown geological features,
uncertain values of geological
properties, limited knowledge
of mechanisms and processes,
and human error in design and
construction. Models used to predict
the behavior of engineered systems
represent idealized processes and
by necessity contain simplifications
and approximations that potentially
introduce errors.
• Accidents are unplanned and
inherently unpredictable. Although
systems can be put into place to
protect against system failures,
seemingly logical decisions about how
to respond to a given situation can
have unexpected consequences due
28
-------
II
Homes in Nondalton Photo: Alan Boraas (Kenai Peninsula College)
to human error (e.g., the January 2012
overflow of the tailings dam at the
Nixon Fork Mine near McGrath, Alaska).
Further, unforeseen events or events
that are more extreme than anticipated
can negate apparently reasonable
operation and mitigation plans.
Climate change will likely exacerbate
this uncertainty. In the Bristol Bay
region, climate change is expected
to lead to changes in snowpack and
the timing of snowmelt, an increased
chance of rain-on-snow precipitation,
and increased flooding. All of these
changes are likely to affect multiple
aspects of any large-scale mining in
the area, including mine infrastructure,
the transportation corridor, water
treatment and discharge, and post-
closure management, in unknown and
potentially unpredictable ways.
• The ore deposit would be mined for
decades and wastes would require
management for centuries or even in
perpetuity. Engineered mine waste
storage systems have been in existence
for only about 50 years, and their
long-term behavior is not known. The
response of current technology in
tailings dam construction is untested
and unknown in the face of centuries of
unpredictable events such as extreme
weather and earthquakes.
• Over the long time span (centuries)
of mining and post-mining care,
generations of mine operators must
exercise due diligence. Priorities
could change in the face of financial
circumstances, changing markets for
metals, new information about the
resource, political priorities, or any
number of currently unforeseeable
changes in circumstance.
Summary of Uncertainties and
Limitations in the Assessment
The most important uncertainties
concerning estimated effects of the mine
scenarios, as judged by the assessment
authors, are identified below.
• Consequences of habitat loss and
degradation for fish populations
could not be quantified because of
the lack of quantitative information
concerning salmonid populations in
freshwater habitats. The occurrence
of salmonid species in rivers and
major streams is known, but detailed
and comprehensive information
on abundances, productivities,
29
-------
Salmon art on a building in Dillingham Photo: Alan Boraas (Kenai Peninsula College)
and limiting factors in each of the
watersheds is not available. Estimating
fish population changes would
require population modeling, which
requires knowledge of life-stage-
specific survival and production
and limiting factors and processes.
Further, it requires knowledge of how
temperature, habitat structure, prey
availability, density dependence, and
sublethal toxicity influence life-stage-
specific survival and production.
Obtaining this information would
require more detailed monitoring and
experimentation. Salmon populations
naturally vary in size due to many
factors that vary among locations and
years. At present, data are insufficient
to establish reliable salmon population
estimates, and obtaining such data
would take many years. Estimated
effects of mining on fish habitat thus
become the best available surrogate for
estimated effects on fish populations.
• Standard leaching test data are
available for test tailings and waste
rocks from the Pebble deposit, but
these results are uncertain predictors
of the actual composition of leachates
from waste rock piles, tailings
impoundments, or tailings deposited in
streams and on their floodplains.
• Leachate capture efficiencies are
uncertain. We assume 50% capture for
waste rock leachates outside of the
mine pit drawdown zone. In the Pebble
2.0 scenario, for example, this would
result in capture of 84% of the leachate
by the pit drawdown zone and the
wells combined. To avoid exceeding
water quality criteria for copper, more
than 99% capture would be required.
• The quantitative effects of tailings
and product concentrate deposited
in spawning and rearing habitat are
uncertain. It is clear that they would
have harmful physical and toxicological
effects on salmonid larvae or sheltering
juveniles, but the concentration in
spawning gravels required to reduce
salmonid reproductive success
is unknown.
• The estimated annual probability of
tailings dam failure is uncertain because
it is based on design goals. Historical
experience is presumed to provide
an upper bound of failure probability.
Features that should reduce failure
frequencies have not been tested for
the thousands of years that they must
function properly. Hence, actual failure
rates could be higher or lower than the
estimated probability.
30
-------
Frying Pan Lake Photo: R. Halford
• The proportion of tailings that would
spill in the event of a dam failure
could be larger than the largest value
modeled (20%).
• The long-term fate of spilled tailings in
the event of a dam failure could not be
quantified. It is expected that tailings
would erode from areas of initial
deposition and move downstream
over more than a decade. However, the
data needed to model that process and
the resources needed to develop that
model are not available.
• The actual response of Alaska Native
cultures to any impacts of the mine
scenarios is uncertain. Interviews with
village Elders and culture bearers and
other evidence suggest that responses
would involve more than the need
to compensate for lost food, and
would be expected to include some
degree of cultural disruption. It is not
possible to predict specific changes in
demographics, cultural practices, or
physical and mental health.
• Because we mention but do not
evaluate potential direct effects
of mining on wildlife or on Alaska
Natives, this assessment represents a
conservative estimate of how these
endpoints would be affected by mine
development and operation.
Uses of the Assessment
This assessment is a scientific
investigation. It does not reflect any
conclusions or judgments about
the need for or scope of potential
government action, nor does it offer
or analyze options for future decisions.
Rather, it is intended to provide a
characterization of the biological and
mineral resources of the Bristol Bay
watershed, increase understanding of
the risks from large-scale mining to the
region's fish resources, and inform future
government decisions. The assessment
will also better inform dialogues among
interested stakeholders concerning the
resources in the Bristol Bay watershed
and the potential impacts of large-scale
mining on those resources.
31
-------
References
ADF&G (Alaska Department of Fish and Game). 2012. Alaska Freshwater Fish Inventory Database.
Available: http://www.adfg.alaska.gov/index.cfm?adfg=ffinventory.main.
Baker, T. Area Fishery Research Biologist, Bristol Bay Salmon Program. Alaska Department of Fish and Game,
Anchorage, AK. July 8,2011 —email to Rebecca Shaftel.
Johnson, J., and P. Blanche. 2012. Catalog of Waters Important for Spawning, Rearing, or Migration ofAnadromous
Fishes—Southwestern Region, Effective June 1,2012. Special Publication No. 12-08. Anchorage, AK: Alaska
Department of Fish and Game.
Ruggerone, G.T., R. M. Peterman, and B. Dorner. 2010. Magnitude and trends in abundance of hatchery and wild pink
salmon, chum salmon, and sockeye salmon in the North Pacific Ocean. Marine and Coastal Fisheries: Dynamics,
Management, and Ecosystem Science 2:306-328.
USFWS (U.S. Fish and Wildlife Service). 2012. National Wetlands Inventory.
Available: http://www.fws.gov/wetlands/Wetlands-Mapper.html. Accessed: December 11, 2012.
USGS (U.S. Geological Survey). 2012. National Hydrography Dataset, High Resolution, Alaska.
Available: ftp://nhdftp.usgs.gov/DataSets/Staged/States/FileGDB/HighResolution. Accessed: October 16, 2012.
Photo Credits
Front Cover, Main Photo....
Front Cover, Thumbnail 1.
Front Cover, Thumbnail 2.
Front Cover, Thumbnail 3.
Front Cover, Thumbnail 4.
Inside Front Cover
Inside Back Cover
Back Cover
. Upper Talarik Creek, Joe Ebersole (USEPA)
. Brown bear, Steve Hillebrand (USFWS)
. Fishing boats at Naknek, USEPA
. IliamnaLake, Lorraine Edmond (USEPA)
. Sockeye salmon in the Wood River, Thomas Quinn (University of Washington)
. Sockeye salmon in the Wood River, Thomas Quinn (University of Washington)
. Sockeye salmon in the Wood River, Thomas Quinn (University of Washington)
. Tributary ofNapotoli Creek, near the Humble claim, Michael Wiedmer
-------
-------
United States
Environmental Protection
Agency
Region 10
1200 Sixth Avenue, Suite 900
Seattle, WA 98101
Recycled/Recyclable
Printed with vegetable-based ink on paper that
contains a minimum of 50% post-consumer
fiber and is processed chlorine free.
------- |