External Review Draft | EPA 910-R-12-004d | May 2012

An Assessment of Potential Mining Impacts on Ecosystems of ,

Executive Summary

U.S. Environmental Protection Agency, Seattle, WA www.epa.gov/bristolbay

External Review Draft – Do Not Cite or Quote DISCLAIMER This document is distributed solely for the purpose of pre-dissemination peer review under applicable information quality guidelines. It has not been formally disseminated by the U.S. Environmental Protection Agency (USEPA). It does not represent and should not be construed to represent any Agency determination or policy. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. Tributary of Napotoli Creek, near the Humble claim Photo: Michael Wiedmer

Executive Summary

he Bristol Bay watershed in southwestern Alaska supports the largest !shery in the world, is home to 25 Federally Recognized Tribal TGovernments, and contains large 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 !sheries, and the future of Alaska Native tribes in the watershed who have maintained a salmon-based culture and subsistence-based lifestyle for at least 4,000 years. The U.S. Environmental Protection Agency (USEPA) launched this assessment to determine the signi!cance of Bristol Bay’s ecological resources and evaluate the potential impacts of large- scale mining on these resources. The USEPA will use the results of this assessment to inform the consideration of options consistent with its role under the Clean Water Act. The assessment is intended to provide a scienti!c and technical foundation for future decision making; the USEPA will not address use of its regulatory authority until the assessment becomes !nal and has made no judgment about whether to use that authority at this time. In addition to informing future USEPA actions, this report is of potential use to other federal and state government entities with an interest in mining in the Bristol Bay region. It is also of interest to both proponents and opponents of mining. By providing an unbiased assessment of potential risks, this assessment informs an active debate concerning the risks of mining development to the sustainability of the Bristol Bay salmon !shery.

1 The Lower between Portage Creek and Ekwok Photo: Michael Wiedmer (ADFG)

Scope of the Assessment This assessment reviews, analyzes, and synthesizes available information on the potential impacts of large-scale mining development on Bristol Bay !sheries and subsequent e"ects on the wildlife and Alaska Native cultures of the region. The primary focus of the assessment is the quality, quantity, and genetic diversity r ive a R of salmonid !sh. Because wildlife and tn ha ulc M KVICHAK rk la Alaska Native cultures in Bristol Bay are C ke r La e ! v intimately connected and dependent i Port Alsworth R iver na R C hulit k NUSHAGAK a g K upon !sh, the quantity and diversity a o kt h ul ! s i R u iv Nondalton of wildlife and the culture and human N er otoli Creek Pedro Bay Nap Koliganek Iliamna ! welfare of indigenous peoples, as ! ! ! a"ected by changes in the !sheries are Newhalen

New Stuyahok Iliamna Lake additional endpoints of the assessment. ! ! Aleknagik ! Kokhanok ! er W v Ekwok Igiugig! i o er R iv o k R The geographic scope of the assessment d k ha ic R a v g K i v a e h r r Dillingham ! e is the Nushagak River and Kvichak River s Levelock iv ! u

Manokotak k R ! N A watersheds. These are the largest of lag n a ! Cook Inlet ! Clark's Point Portage Creek the Bristol Bay watershed’s six major !Ekuk river basins and compose about 50% of the total watershed area. These two watersheds are also identi!ed Bristol Bay as mineral development areas by the State of Alaska. The Pebble deposit, the most likely site for near-term large- scale mining development in the ± region, is located at the intersection of 0 25 50 Kilometers Watershed Boundary the Nushagak River and Kvichak River 0 25 50 Miles Approximate Pebble Deposit Location

2 The Nushagak River and Kvichak River Watersheds of Bristol Bay

r e

v r i e v R i a R n i t ul a kt

h Ko

c

l

u M

NORTH FORK KOKTULI

r e iv R

i l u TSF 1 t U k p o Mine Pit p

K e r NUSHAGAK Waste k T r a o Rock Area l F a h r i r t k o

N C

r

e

e k

SOUTH FORK KOKTULI UPPER TALARIK r Note: Sampling intensity is greatly reduced away from the Pebble deposit area. ive i R Streams without data may not have been surveyed; thus, it is unknown tul Sou ok whether or not they provide suitable habitat for these species. th Fork K

± Spawning Minimum Mine Size 0 2.5 5 Kilometers S Rearing Site Watershed tuy ah 0 o 2.5 5 k R ive Miles Present (Life Stage UnknoNwDnA)RY Watershed Boundary r OU SHED B TER A ek W re C K k A a CH n KVI ka KVICHAK K- as A K Reported Salmon (Sockeye, Chinook,G Coho, Pink, and Chum Combined) Distribution in the North and South Fork Koktuli River and Upper Talarik A H Creek. Designation of species spawning,S rearing, and presence is based on ADFG Draft 2012 Anadromous Waters Catalog (Johnson and Blanche pers. comm.). SpawningIliamn a Lake U = spawning adults observed, rearingN = juveniles observed, present = present, but life stage use not determined. Life stage-speci!c reach designations are likely underestimates, given the logistical constraints on the ability to accurately capture all streams that may support life stage use at various times of the year.

watersheds. The headwaters of three and failures remediated. Because mining biologically productive tributaries wastes would be altered by geologic originate in this region: the North Fork processes but would not degrade, this Koktuli River, located to the northwest of period would continue for centuries and r ive a R the Pebble deposit, which #ows into the potentially “in perpetuity.” tn ha ulc M KVICHAK rk la C Nushagak River via the ; ke r La e ! The assessment was conducted as an v i Port Alsworth the South Fork Koktuli River, which drains R River litna k C hu a NUSHAGAK ecological risk assessment. We started g Ko the Pebble deposit area and converges a kt h ul ! s i R u iv Nondalton with a thorough review of what is known N er with the North Fork west of the Pebble otoli Creek Pedro Bay Nap Koliganek Iliamna ! about the Bristol Bay watershed !shery ! ! deposit; and Upper Talarik Creek, which ! Newhalen drains the eastern portion of the Pebble and wildlife and the Alaska Native

New Stuyahok Iliamna Lake cultures. We also reviewed information ! deposit and #ows into the Kvichak River ! Aleknagik ! Kokhanok about copper mining and available ! er via Iliamna Lake, the largest undeveloped W v Ekwok Igiugig! i o er R iv o k R d k cha information outlining proposed mining a vi R K lake in the . g i v a e h r r Dillingham ! e s Levelock iv operations for the Pebble deposit that ! u Manokotak k R ! N Alag n a The assessment addresses two general has been the focus of much exploratory ! Cook Inlet ! Clark's Point Portage Creek !Ekuk time periods for mine activities. The study and has received much attention !rst is the development and operation from various groups in and outside phase, during which mine infrastructure of Alaska. Using that information, we is built and the mine is operated. This developed a set of conceptual models Bristol Bay phase may last from 25 to 100 years or to show potential associations between more. The second is the post-mining, the endpoints of interest—the salmon or post-closure, phase, during which !shery and salmon populations—and ± the site would be monitored and, as the various types of environmental 0 25 50 necessary, water treatment and other stressors that might reasonably be Kilometers Watershed Boundary 0 25 50 waste management activities continued expected as a result of large-scale Miles Approximate Pebble Deposit Location

3 Sockeye salmon near Gibraltar Lake Photo: Thomas Quinn (University of Washington)

mining. Those conceptual models were re!ned through interactions with regional stakeholders. The assessment was then developed based upon the

r e rk iv la background characterization studies and R C e

a k r KVICHAK La e NUSHAGAK n ! t iv a R e Port Alsworth v the conceptual models. h tna Ri r c li k u l C h a u g M a h s ! u Nondalton This is not an in-depth assessment of a N ek potoli Cre Pedro Bay Na Koliganek Iliamna ! speci!c mine, but rather an examination ! ! ! of the impacts of mining activities at Newhalen New Stuyahok Iliamna Lake the scale and with the characteristics ! ! ! Aleknagik W Ekwok ! r Igiugig Kokhanok realistically foreseeable in the Bristol o ! e o v d i ver Ri R R k ha i ic v k v Bay region, given the nature of mineral e a K r g Dillingham a ! ver ! h Ri Cook Inlet

s Levelock k u a deposits in the watershed and the A la gn N ! requirements for successful mining !Clark's Point ! Portage Creek development. Known information about Ekuk the Pebble deposit is very relevant, because it is likely representative of any Bristol Bay potential near-future mine development Approximate Pebble Deposit Location

in the area. Thus, the assessment largely Watershed Boundary

analyzes a mine scenario that re#ects Confirmed (66%)

the expected characteristics of mining Mapped, no field evidence, but use likely (4%) operations at the Pebble deposit. ± Potential/probable, but undocumented (7%) However, the analysis is intended to 0 25 50 Kilometers Mapped, no field evidence, but use unlikely or limited (1%) provide a baseline for understanding 0 25 50 Miles No evidence (22%) the potential impacts of mining development throughout the Nushagak River and Kvichak River watersheds. The potential mining of other existing Salmon-Producing Subwatersheds in the Nushagak River and Kvichak River 2 copper deposits in the region would Watersheds. A total of 568 subwatersheds (total area of 61,317 km ) were assessed in the likely re#ect the same type of mining Nushagak River and Kvichak River watersheds. The percentage of this area in each category is activities and facilities analyzed for shown in parentheses in the legend. Note that the southwestern portion of the Nushagak River watershed (i.e., the watershed) was not included in this analysis. Data from 4 Demory et al. (1964), Nelson (1967), Salomone et al. (2009), Johnson and Blanche (2011), and ADFG (2012). A. B.

Bristol Bay Kodiak Nushagak & Naknek-Kvichak Russia Mainland & Islands Cook Inlet Egegik West Kamchatka Prince William Sound Ugashik East Kamchatka Southeast Alaska Togiak Western Alaska (excluding Bristol Bay) North British Columbia South South British Columbia, Washington & Oregon

Average Annual Relative Abundance and Commercial Harvest of Wild Sockeye Salmon. A. Average annual relative abundance of wild sockeye salmon stocks in the North Paci!c, 1956 to 2005; with the exception of Bristol Bay, stocks are ordered from west to east across the North Paci!c, from Russia (Russia Mainland and Islands, West Kamchatka, East Kamchatka) to western North America (all other sites). B. Average annual relative commercial sockeye harvest in Bristol Bay watersheds, 1990 to 2009. Data from Ruggerone et al. (2010) and Salomone (pers. comm.).

the Pebble deposit scenario (open pit populations are entirely wild. These !sh mining, waste rock piles, tailing storage are anadromous—hatching and rearing facilities) and, therefore, would present in freshwater systems, migrating to the potential risks similar to those outlined in sea to grow to adult size, and returning to this assessment. freshwater systems to spawn and die.

The most abundant salmon species in the Ecological Resources watershed is sockeye salmon. The Bristol The Bristol Bay watershed provides Bay watershed supports the largest habitat for numerous animal species, sockeye salmon !shery in the world, with including 35 !shes, more than 190 birds, approximately 46% of the average global and more than 40 terrestrial mammals. abundance of wild sockeye salmon. Many of these species are essential Between 1990 and 2010, the annual to the structure and function of the average inshore run of sockeye salmon in region’s ecosystems and economies. Bristol Bay was approximately 37.5 million Chief among these resources is a world- !sh. Annual commercial harvest of class commercial and sport !shery for sockeye over this same period averaged Paci!c salmon and other important 27.5 million. Approximately half of the resident !shes. The watershed supports Bristol Bay sockeye salmon production production of all !ve species of Paci!c is from the Nushagak River and Kvichak salmon found in North America: sockeye River watersheds—the area of focus for (Oncorhynchus nerka), coho (O. kisutch), this assessment. Chinook or king (O. tshawytscha), chum (O. keta), and pink (O. gorbuscha). Because In addition to sockeye salmon, Chinook no hatchery !sh are raised or released salmon are also abundant. For example, in the watershed, Bristol Bay’s salmon Chinook returns to the Nushagak River

5 Sockeye salmon near Pedro Bay, Iliamna Lake Photo: Thomas Quinn (University of Washington)

are consistently greater than 100,000 The exceptional quality of the Bristol !sh per year and have exceeded 200,000 Bay watershed’s !sh populations can !sh in 11 years between 1966 and 2010, be attributed to several factors, the frequently placing Nushagak River most important of which is perhaps Chinook runs at or near the world’s the watershed’s high-quality, diverse largest. This is noteworthy given the aquatic habitats, which are untouched Nushagak River’s small watershed area by human-engineered structures and compared to other Chinook-producing #ow management controls. Surface and rivers such as the Yukon River, which subsurface waters are highly connected, spans Alaska, and the Kuskokwim River in enabling hydrologic and biochemical , just north of Bristol Bay. connectivity between wetlands, ponds, streams, and rivers, thus increasing the The Bristol Bay watershed also supports diversity and stability of habitats able populations of resident !shes that to support !sh. The high diversity of typically remain within the watershed’s habitats, high quality of surface and freshwater habitats throughout their subsurface waters, and relatively low life cycles. The region contains highly development pressures all contribute to productive waters for such sport and making Bristol Bay a highly productive subsistence !sh species as rainbow trout system. This high diversity of habitats (Oncorhynchus mykiss), Dolly Varden also has enabled the development of (Salvelinus malma), Arctic char (Salvelinus high genetic diversity of !sh populations. alpinus), Arctic grayling (Thymallus This genetic diversity acts to reduce year- arcticus), and lake trout (Salvelinus to-year variability in total production and namaycush). These !sh species occupy a increases the stability of the !shery. variety of habitats within the watershed, from headwater streams to wetlands The return of salmon from the Paci!c to large rivers and lakes. The Bristol Bay Ocean brings nutrients into the region is especially renowned for the watershed and fuels terrestrial and abundance and size of its rainbow trout: aquatic food webs. The condition of between 2003 and 2007 an estimated terrestrial ecosystems in Bristol Bay, 196,825 rainbow trout were caught in the therefore, is intimately linked to the Bristol Bay Sport Fish Management Area. condition of salmon populations. Unlike

6 Brown bear in the Kvichak River watershed Photo: USEPA most terrestrial ecosystems, the Bristol Indigenous Cultures Bay watershed has undergone little The Alaska Native cultures present in development and remains largely intact. the Nushagak River and Kvichak River Consequently, the watershed continues watersheds—the Yup’ik and Dena’ina— to support its historic complement of are two of the last intact, sustainable species, including large carnivores such salmon-based cultures in the world. In as brown bears (Ursus arctos), bald eagles contrast, other Paci!c Northwest salmon- (Haliaeetus leucocephalus), and gray based cultures are severely threatened wolves (Canis lupus); ungulates such as due to development, degraded natural moose (Alces alces gigas) and caribou resources, and declining salmon (Rangifer tarandus granti); and numerous resources. Paci!c salmon are no longer waterfowl species. found in 40% of their historical breeding Wildlife populations tend to be relatively ranges in the western United States, and large in the region, due to the increased where populations remain, they tend to biological productivity associated be signi!cantly reduced or dominated with Paci!c salmon runs. Brown bears by hatchery !sh. Salmon are integral to are abundant in the Nushagak River the entire way of life in these cultures as and Kvichak River watersheds. Moose subsistence food and as the foundation and caribou also are abundant, with for their language, spirituality, and social populations especially high in the structure. The cultures have a strong Nushagak River watershed where felt- connection to the landscape and its leaf willow, a preferred plant species, resources. In the Bristol Bay watershed, is abundant. The Nushagak River and this connection has been maintained Kvichak River watersheds are used by for at least the past 4,000 years and caribou, primarily the Mulchatna caribou is in part due to and responsible for herd. This herd ranges widely through the continued pristine condition of these watersheds, but also spends the region’s landscape and biological considerable time in other watersheds. resources. The respect and importance given salmon and other wildlife, along with the traditional knowledge of the environment, have produced a

7 Fishing boats at Naknek, Alaska Photo: USEPA

sustainable subsistence-based economy. this opportunity. All are concerned with This subsistence-based way of life is a the long-term sustainability of their key element of indigenous identity and it communities. serves a wide range of economic, social, and cultural functions in Yup’ik and Economics Of Ecological Dena’ina societies. Resources Fourteen of Bristol Bay’s 25 Alaska Native The Bristol Bay watershed supports villages and communities are within several economic sectors that are the Nushagak River and Kvichak River wilderness-compatible and sustainable: watersheds, with a total population commercial, sport and subsistence of 4,337 in 2010. Thirteen of the 14 !shing, sport and subsistence hunting, communities are Federally Recognized and non-consumptive recreation. Tribal Governments. In the Bristol Bay Considering all these sectors, the region, salmon constitute approximately ecological resources of the Bristol Bay 52% of the subsistence harvest. watershed generated nearly $480 million Subsistence from all sources (!sh, moose, (M) in direct economic expenditures and and other wildlife) accounts for an sales, in 2009, and provided employment average of 80% of protein consumed for over 14,000 full- and part-time workers. by area residents. The subsistence way of life in many Alaska Native villages is The Bristol Bay commercial salmon augmented with activities supporting !shery generates the largest component cash economy transactions. Alaska of economic activity and was valued Native villages, in partnership with Alaska at approximately $300 M in 2009 Native corporations and other business (!rst wholesale value) and provided interests, are considering a variety of employment for over 11,500 full- and economic development opportunities— part-time workers at the peak of the mining included. Some Alaska Native season. These estimates do not include villages have decided for themselves retail expenditures from national and that large-scale hard rock mining is not international sales. the direction they would like to go, while Based on 2009 data, the Bristol Bay sport- a few others are seriously considering !shing industry supports approximately

8 er iv Pebble deposit area Photo: Lorraine Edmond (USEPA) a R atn ch

29,000 sport-!shing trips, generates approximately $60 M per year, and Lake Clark directly employs over 850 full- and part- time workers. The vast majority of this !Port Alsworth revenue is spent in the Bristol Bay region. Sport hunting—mostly of caribou, moose, and brown bear—generates more than $8 M per year and employs over 130 full- and part-time workers. Nondalton ! r e The scenic value of the watershed, oktuli iv K R R th For k iv e o r e il N r P measured in terms of wildlife viewing and tourism, is estimated to generate an

r e additional $100 M per year and supports KVICHAK iv r R a Rive uli River n h Fork Kokt n m ut e lia So k l I e a Pedro Bay nearly 1,700 full- and part-time workers. re h NUSHAGAK C w Knutson Bay ! e k ri N la Iliamna The subsistence harvest of !sh also a T ! Pile Bay r Pedro Bay e p Newhalen contributes to the region’s economy p ! U

! Williamsport when Alaskan households spend money on subsistence-related supplies. These

Iliamna Bay contributions are estimated to be slightly Iliamna Lake over $6 M per year.

Geological Resources Cook Inlet ! Kokhanok In addition to signi!cant and valuable

Kamishak Bay ecological resources, the Nushagak River and Kvichak River watersheds contain considerable mineral resources. The potential for large-scale mining ± Transportation Corridor 0 5 10 development within the region is Kilometers Watershed Boundary 0 5 10 greatest for copper deposits and, to a Miles Approximate Pebble Deposit Location Kukaklek Lake lesser extent, for intrusion-related gold er Riv k deposits. Because these deposits are na l ag Potential 139-kilometer (86-mile) Transportation Corridor Connecting the 9 Pebble Deposit Area to Cook Inlet Mine pit at Fort Knox, Alaska Photo: Phil North (USEPA)

low-grade—meaning that they contain Mine Scenario relatively small amounts of metals A detailed and !nal mine plan has not relative to the amount of ore—mining been made available for any of the copper will be economic only if conducted over deposits identi!ed in the Bristol Bay a large area, and a large amount of waste watershed, nor is one strictly needed to material will be produced as a result of conduct this assessment. To examine the mining and processing. mining-related stressors that could a"ect The largest known deposit and the ecological resources in the watershed, we deposit most explored to assess future developed a hypothetical mine scenario, mining potential is the Pebble deposit. designed to be as realistic as possible. If fully mined, the Pebble deposit could The mine scenario is based on mining produce more than 11 billion metric tons of the Pebble deposit, because it is the (1 metric ton = 1,000 kg, approximately best-characterized mineral resource 2,200 pounds) of ore, which would make and the most likely to be developed in it the largest mine of its type in North the near term. Thus, the mine scenario America. In comparison, the largest draws on plans published by the Pebble existing copper mine in the United Limited Partnership (PLP) and baseline States is the Sa"ord Mine in Arizona with data developed by PLP to characterize 7.3 billion metric tons of ore. Although the likely mine site and surrounding the Pebble deposit represents the environment. Details of a mining plan for most imminent and likely site of mine the Pebble deposit or for other deposits development, other mineral deposits in the watershed may di"er from our with potentially signi!cant resources mine scenario; however, our scenario exist within the Nushagak River and re#ects the general characteristics of Kvichak River watersheds. Several mineral deposits in the watershed, speci!c claims have been !led, many contemporary mining technologies and near the Pebble deposit. Findings of this best practices, the scale of mining activity assessment concerning the potential required for economic development of impacts of large-scale mining are the resource, and necessary development generally applicable to these other sites. of infrastructure to support large-scale mining. Therefore, the USEPA concludes that the mine scenario represents the 10 sort of development plan that can be Landscape near the Pebble deposit location Photo: Joe Ebersole (USEPA) r e iv R r li Fo k K tu rth o k No Minimum Mine Size anticipated for a copper deposit in the Upp er T ala ri k Bristol Bay watershed. Uncertainties

C

r e e k associated with the mine scenario are TSF 1 Mine Pit discussed later in this executive summary.

Waste Rock Area The mine scenario includes minimum r e v i

R and maximum mine sizes, based on

i l

u

t

k

o the amount of ore processed (2 billion

K

k r

o metric tons vs. 6.5 billion metric tons),

F

h

t

u o

± S and approximate corresponding mine 0 1 2 life spans of 25 to 78 years, respectively. Kilometers 0 1 2 Components of the minimum mine Miles Freshwater Habitat would include a 5.5 km2 (1,358 acre) mine pit, a 14.9-km2 (3,686-acre) tailings r e iv R rk Kokt uli impoundment behind a 208 m-high Fo th or N Waste Rock (685-foot-high) earthen dam; a 13.3- Maximum Mine Size Area km2 (3,286-acre) waste rock pile; a 139-

Up km (86-mile) road with four pipelines p e r

T a for product concentrate, return water, l a r ik TSF 1 C re diesel, and natural gas; and facilities for Mine Pit e k ore processing and support services. The maximum size mine would include TSF 3 Waste Rock a much larger pit and waste rock pile, Area with a combined area of 38.4 km2 (9,486

acres), potentially an underground mine, r

e and three tailings impoundments, with a v TSF 2 i R li 2 u t combined area of 43.7 km (10,807 acres). k

o

K

k

r

o

F

h

t

u The !rst part of the assessment considers o S routine operation, which assumes that the mine would be designed using practices Minimum and Maximum Mine Footprints in the Assessment Scenario. to minimize environmental impacts and Individual mine components are the mine pit, waste rock piles, and one or more tailings that no signi!cant human or engineering storage facilities (TSFs). The dark bar at the north end of TSF 1 indicates the dam for 11 which tailings dam failure is modeled. Sockeye salmon near Pedro Bay, Iliamna Lake Photo: Thomas Quinn (University of Washington)

failures occur during or for centuries Overall Risks to Salmon and after operation. The second part of the Other Fish assessment considers various failures that have occurred during the operation Based on the mine scenario, the of other mines and have the potential to assessment de!nes potential mining- occur here. related stressors that could a"ect the Bristol Bay watershed’s !sh and would The assessment does not consider all consequently have impacts on wildlife mining-related development. Although and human welfare. the mine scenario assumes development of a deep-water port on Cook Inlet to No Failure ship concentrated product elsewhere No failure, or routine operation, is a for smelting and re!ning, impacts of the mode of operation de!ned as using development and operation of a deep- the highest design standards and day- water port are not assessed. Additionally, to-day practices, with all equipment the assessment does not evaluate the and management systems operated in potential environmental impacts of one accordance with applicable speci!cations or more electricity-generating power and requirements. In the no failure plants that would need to be constructed mode of operation, we assume that best to provide power at the mine site and the practical engineering and mitigation deep-water port facility. This assessment practices are in place and in optimal also does not consider potential impacts operating condition. We do not specify resulting from secondary development all of those mitigation practices, but that is likely to accompany a large- rather, we assume that they would scale mine development. Secondary be in place and properly functioning. development includes, but is not limited Analyzing routine operations is not to, additional support services for mine meant to imply that a failure-free mining employees and their families, increased operation is likely; rather, it is meant to recreational development due to increased isolate the inevitable and foreseeable access, development of vacation homes, e"ects of mining from those that are and increased transportation infrastructure unintended and thus more di$cult to (i.e., airports, docks, and roads).

12 Rainbow trout caught in American Creek, in the Kvichak River watershed Photo: USEPA

r Koktuli R ive ork h F rt o N Minimum Mine Size predict. With no failures, adverse e"ects

Upper Ta lari outside the mine footprint are minimized k Cr e e k by complete containment of waste rock

TSF 1 Mine Pit and mine tailings, reliable collection of all water from the site, and e"ective

Waste Rock Area treatment of e%uents. Nonetheless,

r e v impacts on !sh resulting from habitat i

R

i

l

u

t loss and modi!cation within and beyond

k

o K

Stream Eliminated

k the area of mining activity would r

o

F

h

Stream Blocked t

u result from six key direct and indirect

o S ± Wetland Eliminated mechanisms. 0 1 2 Kilometers Wetland Blocked 0 1 2 (1) Eliminated or blocked streams Miles Freshwater Habitat under the minimum and maximum mine footprints (i.e., the mine

r e pit, waste rock piles, and tailings iv R rk Kokt uli Fo rth storage facilities) would result in No Waste Rock Maximum Mine Size Area the loss of 87.5 to 141.4 km (55 to 87 miles), respectively, of possible Up pe r spawning or rearing habitats for T a l a r i k , Chinook salmon, C r TSF 1 e e Mine Pit k sockeye salmon, rainbow trout, and Dolly Varden.

TSF 3 Waste Rock (2) Reduced !ow resulting from Area water retention for use in mine

operations, ore processing,

r e transport, and other processes

v TSF 2 i R li u would reduce the amount and t k

o

K

k

r quality of !sh habitat. Reductions

o

F

h

t

u o in stream#ow exceeding 20% S would adversely a"ect habitat in

Streams and Wetlands Lost (Eliminated or Blocked) Under the Minimum and 13 Maximum Mine Footprints Lower Twin Lake in the Mulchatna River watershed Photo: Michael Wiedmer (USGS)

an additional 2 to 10 km (1.2 to t Reduced food resources 6.2 miles) of streams, reducing would result from the loss of production of coho salmon, organic material and drifting sockeye salmon, Chinook salmon, invertebrates from the 87.5 rainbow trout, and Dolly Varden. to 141.4 km (55 to 87 miles) An unquanti!able area of riparian of streams and streamside #oodplain wetland habitat would wetlands lost to the mine either be lost or su"er substantial footprint. changes in hydrologic connectivity with streams due to reduced #ow t The balance of surface water from the mine footprint. and groundwater inputs to downstream reaches would (3) Removal of 10.2 to 17.3 km2 shift, potentially reducing (2,512 to 4,286 acres) of wetlands winter !sh habitat and making in the footprint of the mine would the streams less suitable for eliminate o"-channel habitat for spawning and rearing. salmon and other !shes. Wetland loss would reduce availability t Water treatment and reduced and access to hydraulically and passage through groundwater thermally diverse habitats that #owpaths could increase can provide enhanced foraging summer water temperatures opportunities and important and decrease winter water rearing habitats for juvenile salmon temperatures, making streams less suitable for salmon, trout, (4) Indirect e"ects of stream and and char. wetland removal would include reductions in the quality of These indirect e"ects cannot be downstream habitat for the same quanti!ed but likely would diminish !sh species listed above in the three production downstream of the mine site. headwater streams draining the (5) Diminished habitat quality in mine site. Sources of these indirect streams below road crossings e"ects would include the following: would result primarily from altered #ow, runo" of road salts, and 14 Groundwater upwelling near Kaskanak Creek in the Lower Talarik Basin Photo: Joe Ebersole (USEPA)

siltation of spawning habitat and scenario would be stored in tailings reduced invertebrate prey. The storage facilities (TSFs) consisting of road is adjacent to Iliamna Lake and tailings dams and impoundments. The crosses multiple tributary streams. annual probability of failure for each These habitats are important tailings dam would be in the range spawning areas for sockeye salmon, of one-in-ten-thousand to one-in-a- putting sockeye particularly at risk million. The probability of one of several to impacts from the road. tailings dams failing increases with the number of dams. The minimum mine size (6) Inhibition of salmonid outlined in the mine scenario includes movement at road crossings one TSF with three dams; the maximum could result from culverts that may, mine size includes three TSFs, with a over time, block or diminish use of total of eight dams. The TSFs and their the full stream length. component dams are likely to be in place for hundreds to thousands of years, long Failure beyond the life of the mine. Although The assessment evaluates four failures details for the actual design of mining that have occurred at other large-scale operations at the Pebble deposit are mining and related infrastructure unknown, available reports from the projects and that could occur during PLP suggest tailings dams as high as 208 mine operations or after mine closure: m (685 feet) at TSF 1. At this height, the tailings dam failure, product concentrate tailings dam would be higher than the St. or return water pipeline failure, water Louis Gateway Arch and the Washington collection and treatment failures, and Monument. We evaluated two dam failures of roads and culverts. Risks failures in this assessment: one when associated with each of these failures are the TSF was partially full (partial-volume summarized in the following table. failure) and one when it was completely full (full-volume failure). In both cases we Tailings Dam Failure assumed a release of 20% of the tailings, a conservative estimate that is well Tailings are the waste materials produced within the range of historical tailings dam during ore processing, which in our failures.

15 Summary of Probability and Consequences of Potential Failures under the Mine Scenario

Failure Probabilitya Consequences

Tailings dam 10-4 to 10-6 per dam-year More than 30 km of salmonid stream = recurrence frequency of would be destroyed and more streams 10,000 to 1 million yearsb and rivers would have greatly degraded habitat for decades.

Product concentrate pipeline 10-3 per km-year = 98% Most failures would occur between stream chance per pipeline in 25 or wetland crossings and might have little years e"ect on !sh.

Concentrate spill into a stream 2x10-2 per year = 1.5 stream- Fish and invertebrates would experience contaminating spills in 78 acute exposure to toxic water and chronic years exposure to toxic sediment in a stream and potentially extending to Iliamna Lake.

Concentrate spill into a wetland 3 x 10-2 per year = 2 Invertebrates and potentially !sh would wetland-contaminating experience acute exposure to toxic water spills in 78 years and chronic exposure to toxic sediment in a pond or other wetland.

Return water pipeline Same as product concentrate Fish and invertebrates would experience pipeline acute exposure to toxic water.

Culvert, operation Low Frequent inspections and regular maintenance would result in few impassable culverts.

Culvert, post-operation 3 x 10-1 to 6 x 10-1 per In surveys of road culverts, roughly one- culvert-instantaneous = 4 to third to two-thirds are impassable to !sh 10 culverts at any one time. This would result in 4 to 10 salmonid streams blocked.

Water collection and treatment, High Collection and treatment failures are operation highly likely to result in release of untreated leachates for hours to months.

Water collection and treatment, High Collection and treatment failures are planned post-closure highly likely to result in release of untreated leachates for days to months.

Water collection and treatment, Certain When water is no longer managed, premature post-closure or perpetuity untreated leachates would #ow to the streams.

a Because of di"erences in derivation, the probabilities are not directly comparable. b Based on expected state safety requirements. Observed failure rates for earthen dams are higher (about 5 x 10-4 per year or a recurrence frequency of 2,000 years).

16 Transamerica Building – 260 Meters 260 240 220 Tailings Dam TSF 1 – 208 Meters 200 Gateway Arch – 192 Meters 180 Washington Monument – 169 Meters 160 Full Volume 140 Tailings Reservoir

Meters 120 100 80 60 Tailings Dam Partial Volume 40 20 0

Height of the Partial-Volume and Full-Volume Dam at TSF 1, Relative to Common Landmarks

The range of estimated probabilities assessment of the correlation of dam of dam failure is wide, re#ecting the failure probabilities with safety factors great uncertainty concerning such against slope instability suggests an failures. The most straightforward annual probability of failure of 1 in method of estimating the annual 1,000,000 for Category I Facilities (those probability of failure of a tailings dam designed, built, and operated with is to use the historical failure rate of state-of-the-practice engineering) and 1 similar dams. Three reviews of tailings in 10,000 for Category II Facilities (those dam failures produced an average designed, built, and operated using rate of approximately 1 failure per standard engineering practice). This 2,000 dam years, or 5 x 10-4 failures spans the failure frequency used in our per dam year. The argument against failure assessment. The advantage of this approach is that it does not fully this approach is that it addresses current re#ect current engineering practice. regulatory guidelines and engineering Some studies suggest that improved practices. The disadvantage is that we do design, construction, and monitoring not know whether standard practice or practices can reduce the failure rate by state-of-the practice dams will perform an order of magnitude or more, resulting as expected, particularly given the large in an estimated failure probability size of potential dams. In addition, slope within our assumed range. The State instability is only one type of failure; of Alaska’s guidelines suggest that an other failure modes, such as overtopping applicant follow accepted industry during a #ood, would increase overall design practices such as those provided failure rates. by the U.S. Army Corps of Engineers (USACE), Federal Energy Regulatory Failure of the dam at TSF 1 would result Commission (FERC), and other agencies. in the release of a #ood of tailings Both USACE and FERC require a minimum slurry into the North Fork Koktuli River, factor of safety of 1.5 against slope scouring the valley and depositing instability, for the loading condition tailings several meters (yards) in depth corresponding to steady seepage over the entire #oodplain of the river. from the maximum storage facility. An The complete loss of suitable salmon habitat in the North Fork Koktuli River

17 Tributary of Napotoli Creek, near the Humble claim Photo: Michael Wiedmer

along at least 30 km (18.6 miles) of stream both !sh and the invertebrates habitat—the spatial limit of the modeling they eat. Based largely on their conducted for this assessment—in the copper content, deposited short term (fewer than 10 years) and tailings would be toxic to the high likelihood of very low-quality benthic macroinvertebrates, spawning and rearing habitat in the long although existing data term (decades) would result in near- concerning toxicity to !sh is complete loss of mainstem North Fork less clear. Koktuli River !sh populations. The North Fork Koktuli River currently supports t Deposited tailings would spawning and rearing populations of continue to erode from the sockeye, Chinook, and coho salmon; North Fork Koktuli and Koktuli spawning populations of ; River valleys. and rearing populations of Dolly Varden t Suspension and redeposition and rainbow trout. The slurry #ood of tailings would likely cause would continue down the Koktuli River serious habitat degradation with similar e"ects, the extent of which in the Koktuli River and cannot be estimated at this time due to downstream into the model and data limitations. Mulchatna River. The tailings dam failures evaluated here (2) Those waters would provide very are predicted to have the following severe low-quality spawning and rearing direct and indirect e"ects on aquatic habitat for a period of decades. resources, particularly salmonid !sh. t Recovery of suitable substrates (1) It is likely the the North Fork Koktuli via mobilization and transport River below the TSF 1 dam, and of tailings would take years much of the Koktuli River, would to decades, and would a"ect not support salmonid !sh in the much of the watershed short term (fewer than 10 years). downstream of a failed dam. t Deposited tailings would degrade habitat quality for

18 Knutson Creek draining into the Knutson Bay area of Iliamna Lake Photo: Keith Denton

t Ultimately, spring #oods and to loss of the Koktuli River storm#ows would carry some watershed population; an proportion of the tailings into additional 10 to 20% could be the Nushagak River. lost due to tailings deposited in the Mulchatna River and its t For some years, periods of tributaries. high #ow would be expected to suspend su$cient t Sockeye are the most abundant concentrations of tailings to salmon returning to the cause avoidance, reduced Nushagak River watershed, growth and fecundity, and with annual runs averaging even death of !sh. more than 1.3 million !sh. The proportion of sockeye (3) Near-complete loss of North Fork and other salmon species of Koktuli River !sh populations Koktuli-Mulchatna origin is would likely result from these unknown. habitat losses. t Similarly, populations of t The Koktuli River watershed rainbow trout and Dolly is an important producer of Varden would be lost for Chinook salmon. The Nushagak years to decades. Quantitative River watershed, of which the estimates of the impacts Koktuli River watershed is a on population sizes are not part, is the largest producer of possible. Chinook salmon in the Bristol Bay region, with annual runs E"ects would be qualitatively the same averaging over 160,000 !sh. for both the partial-volume and full- volume dam failures, although e"ects t The tailings spill would be from the full-volume failure would expected to eliminate 28% extend further and last longer. Failure of of the Chinook salmon run dams at the two additional TSFs under in the Nushagak River due the maximum mine size (TSF 2 and

19 TSF 3) were not modeled, but would design or operation, or the failure to have similar e"ects in the South Fork maintain and operate these systems in Koktuli River and downstream. However, perpetuity, could result in contamination because their volumes would be smaller, of one or more streams draining the e"ects would be less extensive. site. Based on a review of historical and currently operating mines, some failure Pipeline Failures of the collection and treatment systems Under the mine scenario, the primary is likely during operation or post-closure product of the mine would be a periods. These failures could range from concentrate of copper and other metals operational failures resulting in short- that would be pumped in a pipeline to term releases of untreated leachates, a shipping facility on Cook Inlet. Water to long-term failures to operate the carrying the sand-like concentrate would collection and treatment system in be returned to the mine site in a second perpetuity. Our evaluation looked at the pipeline. Based on the record of pipelines realistic possibility of leachate escaping in general, and the world’s largest metal at the base of TSF 1. We also considered a concentrate pipeline in particular, one to failure to collect and treat leachate from two near-stream failures of each of these waste rock piles around the mine pit. pipelines would be expected to occur Test leachates from the tailings and over the life of the maximum mine (78 non-ore-bearing Tertiary waste rocks— years). Failure of either the product or those formed between approximately the return water pipelines would release 65 million to 2.5 million years ago—are water that is expected to be highly toxic, mildly toxic; they would require an potentially killing !sh and invertebrates approximately two-fold dilution to in the a"ected stream over a relatively achieve water quality criteria for copper, brief period. If concentrate spilled into but they are not expected to be toxic a stream, it would settle and form bed to salmonids. If Tertiary rock were to sediment predicted to be highly toxic be used as planned for construction based on its high copper content and of mining infrastructure, leachate acidity. Unless the receiving stream was from these areas would need to be dredged, causing additional long-term collected and treated to avoid toxic damage, this sediment would persist e"ects on benthic invertebrates. Our for decades before ultimately being risk assessment did not evaluate this washed into Iliamna Lake. Potential potential pathway in detail. concentrations in the lake could not be predicted, but near the pipeline route Pre-Tertiary waste rocks, which would Iliamna Lake contains important beach be excavated to expose the ore body, spawning areas for sockeye salmon are acid-forming with high copper that could be exposed to a toxic spill. concentrations in test leachates and Sockeye also spawn in the lower reaches would require 2,900 to 52,000-fold of streams which could be directly dilution to achieve water quality contaminated by a spill. criteria. If leachate from a waste rock pile surrounding the mine pit was Water Collection and Treatment not collected, the 10.6 million m3 Failures (approximately 2.8 billion gallons) of There is a long history of unplanned leachate per year from the waste rock discharges of contaminated waters pile could constitute source water for from mine sites into surface and ground Upper Talarik Creek, which #ows to waters. Water in contact with tailings Iliamna Lake. The total #ow of Upper or waste rock would leach copper and Talarik Creek would provide only 18-fold other metals. The failure of collection dilution, so failure to prevent leachate and treatment systems due to imperfect releases could cause the entire creek and

20 Washed out culvert on Kenai Peninsula Photo: Robert Ru"ner (Kenai Watershed Forum)

a potentially large mixing zone in the lake outmigration and were not cleared for to become toxic to !sh and the sensitive several days, production of a year-class invertebrates upon which they feed. The (i.e., !sh spawned in the same year) could signi!cance of such an event to salmon is be lost or diminished. illustrated by the abundance of spawning salmon in Upper Talarik Creek. As many Culverts can also fail to convey water as as 33,000 sockeye and 6,300 coho a result of landslides or, more commonly, spawners have been counted in the creek #oods that wash out the culvert. In such on a single day; in 2008, 82,000 sockeye failures, the stream could be temporarily were counted in Upper Talarik Creek and impassible to !sh until the culvert is one of its tributaries in a single day. The repaired or until erosion reestablishes toxic event described could kill adult !sh the channel. If the failure occurs during or the millions of eggs, larvae, and fry a critical period in salmon migration, that they generate. the e"ects would be the same as with a debris blockage (i.e., a lost or diminished Road and Culvert Failures year-class). Within the Kvichak River watershed, the Culvert failures also would result in the transportation corridor would cross 34 downstream transport and deposition of streams and rivers supporting migrating silt, which could cause returning salmon and/or resident salmonids, including to avoid a stream if they arrived during 17 streams designated as anadromous or immediately following the failure. waters at the location of the crossing. More likely, deposition of silt would The most likely serious failure associated smother salmon eggs and larvae, if they with the transportation corridor would were present, and would degrade the be blockage or failure of culverts. downstream habitat for salmonid !sh Culverts commonly become blocked and the invertebrates that they eat. by debris that may not stop water #ow but would block !sh passage. If Extended blockage of !sh passage these blockages occurred during adult at road crossings is unlikely during salmon immigration or juvenile salmon operation assuming best-case scenario daily inspection and maintenance.

21 Homes near Newhalen Photo: David Allnut (USEPA)

However, after mine operations cease, pipeline failures that would release the road may be maintained less carefully product slurry, return water, or diesel or be transferred to a governmental fuel. The e"ects of each of these accidents entity. In that case, the proportion of individually would be the same as discussed culverts that are impassable would be previously, but their co-occurrence would expected to revert to levels found in cause cumulative e"ects on salmonid published surveys of public roads (30 to populations and make any mitigative 66%). Of the approximately 50 culverts response more di$cult. that would be required, 17 would be on streams that are believed to support Over the perpetual timeframe that salmonids. Hence, over the long term, tailings, mine pit, and waste rock 4 to 10 streams would be expected to would be in place, the likelihood of lose passage of salmon, rainbow trout, multiple extreme precipitation events, or Dolly Varden, and some proportion earthquakes, or combinations of these of those streams would have degraded events becomes much greater. Multiple downstream habitat resulting from the events further increase the chances sedimentation from washout of the road. of weakening and eventual failure of facilities that are still in place. Common Mode Failures Overall Loss of Wetlands Multiple, simultaneous failures could occur as a result of a common event, Wetlands are a dominant feature of such as the occurrence of a severe storm the landscape in the Pebble deposit with heavy precipitation (particularly one area and throughout the Bristol Bay that fell on spring snow cover) or a major watershed, and are important habitats earthquake. Such an event could cause for salmon and other !sh. Ponds and one to three tailings dam failures that riparian wetlands provide spawning, would spill tailings slurry into streams rearing, and refuge habitat for both and rivers, road culvert washouts that anadromous salmonids and resident !sh would send sediments downstream species. Other wetlands moderate #ows and potentially block !sh passage, and and water quality, and can in#uence downstream delivery of dissolved

22 organic matter, particulate organic spawning migration. Those nutrients are matter, and aquatic macroinvertebrates released into streams when the salmon that supply food sources to !sh. Under die, enhancing the production of other the mine scenario, wetlands would be aquatic species that feed wildlife. Salmon !lled or excavated in 10.2 km2 (2,512 predators deposit these nutrients on acres) and 17.3 km2 (4,286 acres) of the landscape, thereby fertilizing the the minimum and maximum mine vegetation and increasing the abundance footprints, respectively. An additional and production of moose, caribou, and 1.9 km2 (481 acres) and 1.1 km2 (267 other wildlife that depend on vegetation acres) of riparian wetlands would be for food. blocked by the minimum and maximum footprints, respectively, and would be Fish-Mediated Risk to lost or su"er substantial changes in Indigenous Culture hydrologic connectivity with streams as a result of reduced #ow from the mine Under routine operations with no major footprint. Another 0.18 km2 (44 acres) of accidents or failures, the predicted loss wetlands would be !lled in the Kvichak and degradation of salmon, char, and River watershed by the roadbed of the trout habitat in North Fork Koktuli and transportation corridor. By interrupting South Fork Koktuli Rivers and Upper #ow and adding silt and salts, the Talarik Creek is expected to have some roadbed would also a"ect approximately impact on Alaska Native cultures of 2.4 to 4.9 km2 (593 to 1,211 acres) of the Bristol Bay watershed. Fishing and wetlands. Finally, a tailings or product hunting practices are expected to concentrate spill could damage wetlands change in direct response to the stream, and eliminate or degrade their capacity wetland, and terrestrial habitats lost due to support !sh. to the footprints of the mine site and the transportation corridor. Additionally, it Fish-Mediated Risk to Wildlife is also possible that subsistence use of salmon resources could decrease based Although the e"ects of reduced on the perception of reduced !sh or salmon, trout, and char production water quality resulting from mining. on wildlife—the !sh-mediated risk to wildlife—cannot be quanti!ed given The potential for signi!cant e"ects on available data, some reduction in wildlife indigenous cultures is much greater would be expected under the mine from a mine failure than from routine scenario. Changes in the occurrence operations. As described above, and abundance of salmon have the failures could reduce or eliminate potential to change animal behavior and !sh populations in a"ected areas, reduce wildlife population abundances. including areas signi!cant distances Assuming no failures, routine operations downstream from the mine. Any loss would be expected to have local e"ects of !sh production from these potential on brown bears, wolves, bald eagles, and failures would reduce the availability other wildlife that consume salmon as of those subsistence resources to local a result of reduced salmon abundance Alaska Native villages, and the reduction from the loss and degradation of habitat of food supply potentially would have in or immediately downstream of the negative consequences on human mine footprint. Any of the accidents or health if alternative food resources are failures evaluated would increase e"ects not available. Salmon-based subsistence on salmon, which would proportionately is integral to Alaska Native cultures. If reduce the abundance of their predators. salmon quality or quantity is adversely a"ected, the nutritional, social, and The abundance and production of spiritual health of Alaska Natives and their wildlife also is enhanced by the marine culture will potentially decline. nutrients that salmon carry on their

23 Cumulative Risks Summary Of Uncertainties In This assessment has focused on the Mine Design And Operation potential e"ects of a single, hypothetical This assessment of a hypothetical mine on salmon and other resources mine scenario is generally applicable in the Nushagak and Kvichak River to the copper deposits in the Bristol watersheds, including the cumulative Bay watershed and is based on speci!c e"ects of multiple stressors associated characteristics of the Pebble deposit. with that mine. However, the potential The mine scenario does not represent exists for development of multiple mines the plans of any mining company; if and associated infrastructure in these the resource is mined in the future, watersheds. Each potential mine poses actual events will undoubtedly deviate risks similar to those identi!ed for the from this scenario. This is not a source mine scenario. Estimates of the loss of of uncertainty, but rather an inherent stream and wetland habitats would di"er aspect of a predictive assessment. across di"erent deposits based on the Even an environmental assessment of size and location of mining operations a proposed plan by a mining company within the watersheds. Individually, each would be an assessment of a scenario mine footprint would eliminate some that undoubtedly would di"er from the amount of !sh-supporting habitat and, ultimate development. should human or engineering failures occur, a"ect !sh habitats beyond the Multiple uncertainties are inherent mine footprint. Cumulatively, multiple in planning, designing, constructing, mines have the potential to decrease the operating, and closing a mine. abundance and genetic diversity of !sh t Mines are complex systems requiring populations and thereby increase their skilled engineered design and annual variability. operation. The uncertainties facing We considered development of mines mining and geotechnical engineers at several sites in the Nushagak River include unknown geologic defects, watershed, including Big Chunk, uncertain values in geological Groundhog Mountain, and Humble properties, limited knowledge claims. These sites were chosen, because of mechanisms and processes, all contain copper deposits that have and human error in design and generated exploratory interest. If construction. Vick (2002) notes all four mine sites were developed, that models used to predict the the cumulative area covered by TSFs behavior of an engineered system alone would be close to 73 km2 (19,038 are “idealizations of the processes acres). Loss of stream habitats as a they are taken to represent, and it is result of eliminated or blocked streams well recognized that the necessary could reach 233 km (144 miles). The simpli!cations and approximations combined facilities would eliminate can introduce error in the model.” an estimated 34.6 km (21.5 miles) of Engineers use professional judgment documented salmon streams. The length in addressing uncertainty (Vick 2002). of salmon stream a"ected is likely an t Accidents are inherently underestimate, because most streams unpredictable. Though systems can have not been sampled for the presence be put into place to protect against of salmon. Loss of these distinct streams system failures, seemingly logical would likely result in the loss of their decisions about how to respond to a associated salmon populations, reducing given situation can have unexpected the genetic and life-history diversity consequences resulting from human generated through the existence of error (e.g., the January 2012 over#ow numerous distinct populations. of the tailings dam at the Nixon Fork

24 Salmon art on a building in Dillingham Photo: Alan Boraas (Kenai Peninsula College)

Mine near McGrath, Alaska). Further, Such uncertainties are inherent in unforeseen events or events that are any complex enterprise, particularly more extreme than anticipated can when they involve an incompletely negate the apparent wisdom of prior characterized natural system. However, decisions (Caldwell and Charlebois the large scales and long durations 2010). implied by the e"ort required to exploit this resource make these inherent t The ore deposit would be mined for uncertainties more prominent. decades and the waste would require management for centuries or even in Summary of Uncertainties and perpetuity. Engineered waste storage systems of mines have only been in Limitations in the Assessment existence for about 50 years. Their Signi!cant uncertainties about and long-term behavior is not known. limitations of the estimated potential The response of our best technology e"ects of the mine scenario, as judged in the construction of tailings dams is by the assessment authors, include the untested and unknown in the face of following. centuries of extreme events such as earthquakes and weather. t Any mine plan submitted by a mining company may not exactly re#ect the t Mine management or ownership location and sizes of the mine pit, may change over time. Over the waste rock pile, and tailings storage long timespan (centuries) of mining facilities, and the location and length and post-mining care, generations of the transportation corridor used of mine operators must exercise in the scenario for this assessment. due diligence. Priorities are likely An actual mine plan may be smaller, to change in the face of !nancial larger, or laid out di"erently than the circumstances, changing markets for mine scenario considered here. metals, new information about the resource, political priorities, or any t The estimated annual probability number of currently unforeseeable of tailings dam failure is uncertain changes in circumstance. and based on both design goals and

25 Kvichak River below Iliamna Lake and Igiugig Photo: Joe Ebersole (USEPA)

historical experience. Actual failure not available. Estimating changes rates could be higher or lower than in populations would require the estimated probability. population modeling, which requires knowledge of life stage-speci!c t The proportion of the tailings that survival and production as well as would spill in the event of a dam knowledge of limiting factors and failure could be larger than the processes that are not available. largest value modeled (20%). Further, it requires knowledge of how t The long-term fate of the spilled temperature, habitat structure, prey tailings in the event of a dam failure availability, density dependence, could not be quanti!ed. Analogous and sublethal toxicity in#uence to other cases, it is likely that tailings life stage-speci!c survival and would erode from the areas of initial production, which is not available. deposition and move downstream Obtaining that information would over a period of more than a decade. require more detailed monitoring However, the data needed to model and experimentation. Further, that process and the resources salmon populations naturally vary in needed to develop that model were size because of a great many factors not available. that vary among locations and years. Collecting su$cient data to establish t Consequences of the loss and reliable salmon population estimates degradation of habitat on !sh takes many years. Estimated e"ects populations could not be quanti!ed of mining on habitat become the because of the lack of quantitative available surrogate for estimated information concerning salmon, e"ects on !sh populations. char, and trout populations. The occurrence of salmonid species in t Standard leaching test data are rivers and major streams is known, available for test tailings and waste but information on abundances, rocks from the Pebble deposit, productivities, and limiting factors but these results are uncertain within each of the watersheds is predictors of the actual composition of leachates from tailings

26 New Stuyahok Photo: David Allnut (USEPA)

impoundments, tailings deposited in streams and on their #oodplains, and waste rocks in piles.

t The e"ects of tailings and product concentrate deposited in spawning and rearing habitat are uncertain. It is clear that they would have harmful physical and toxicological e"ects on salmonid larvae or sheltering juveniles, but the concentration in spawning gravels required to reduce salmonid reproductive success is unknown.

t The actual response of Alaska Native cultures to any impacts of the mine scenario 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 likely include some degree of cultural disruption. It is not possible to predict speci!c changes in demographics, cultural practices, or physical and mental health.

27 References

ADFG (Alaska Department of Fish and Game). 2012. Alaska Freshwater Fish Inventory. Alaska Department of Fish and Game, Division of Sport Fish. http://www.adfg.alaska.gov/index.cfm?adfg=$nventory.main.

Blight, G. E. 2010. Geotechnical Engineering for Mine Waste Storage Facilities. CRC Press, Boca Raton.

Caldwell, J. A., and L. Charlebois. 2010. Tailings Impoundment Failures, Black Swans, Incident Avoidance and Checklists. Pages 3–39 in Tailings and Mine Waste 2010: Proceedings of the 14th International Conference on Tailings and Mine Waste, Vail, Colorado, USA, October 17–20, 2010. CRC Press, Boca Raton, FL.

Demory, R. L., R. F. Orrell, and D. R. Heinle. 1964. Spawning ground catalog of the Kvichak River system, Bristol Bay, Alaska, Special Scienti!c Report-Fisheries No. 488. U.S. Dept. of the Interior, Bureau of Commercial Fisheries, Washington.

Johnson, J., and P. Blanche. 2011. Catalog of waters important for spawning, rearing, or migration of anadromous !shes – Southwestern Region, E"ective June 1, 2011, Special Publication No. 11-08. Alaska Department of Fish and Game, Anchorage, AK. http://www.adfg.alaska.gov/sf/SARR/AWC/index.cfm?ADFG=main.overview.

Johnson, J. and P. Blanche. In Press. Catalog of waters important for spawning, rearing, or migration of anadromous !shes – Southwestern Region, E"ective June 1, 2012. Alaska Department of Fish and Game, Special Publication No. 12-08, Anchorage.

Nelson, M. L. 1967. Red salmon spawning ground surveys in the Nushagak and Togiak districts, Bristol Bay, 1966; Informational Lea#et 96. Alaska Department of Fish and Game, Division of Commercial Fisheries, Juneau, AK.

Ruggerone, G. T., R. M. Peterman, and B. Dorner. 2010. Magnitude and trends in abundance of hatchery and wild , chum salmon, and sockeye salmon in the North Paci!c Ocean. Marine and Coastal Fisheries: Dynamics, Management, and Ecosystem Science 2:306-328.

Salomone, P., S. Morstad, T. Sands, and M. Jones. 2009. Salmon spawning ground surveys in the Bristol Bay Area, Alaska, 2008; Fishery Management Report No. 09-42. Alaska Department of Fish and Game, Division of Commercial Fisheries, Anchorage, AK.

Salomone, P. Area Management Biologist, Alaska Department of Fish and Game. Unpublished data.

Vick, S. G. 2002. Degrees of Belief: Subjective Probability and Engineering Judgement. American Society of Civil Engineers, Reston, VA.

Photo Credits Front Cover, Main Photo ...... Upper Talarik Creek, Joe Ebersole (USEPA) Front Cover, Thumbnail 1 ...... Brown bear, Steve Hillebrand (USFWS) Front Cover, Thumbnail 2 ...... Fishing boats at Naknek, Alaska, USEPA Front Cover, Thumbnail 3 ...... Iliamna Lake, Lorraine Edmond (USEPA) Front Cover, Thumbnail 4 ...... Sockeye salmon in the Wood River, Thomas Quinn (University of Washington) Inside Front Cover ...... Sockeye salmon in the Wood River, Thomas Quinn (University of Washington) Inside Back Cover ...... Sockeye salmon in the Wood River, Thomas Quinn (University of Washington) Back Cover...... Tributary of Napotoli Creek, near the Humble claim, Michael Wiedmer

External Review Draft – Do Not Cite or Quote

Region 10, Seattle, WA