EPA/600/R-19/057
Linking Physical, Biological,
and Social Sciences in
Natural Resources: An
Ecosystem Service
Framework for Monetary
Valuation of Environmental
Impacts Related to M ning in
Central Colorado
Progress for o Stronger Future
Office of Research and Development
Center for Environmental Solutions and Emergency Response
Land Remediation and Technology Division

-------
Revised: October 4, 2021
By
Andrew L. Gulley, and Robert R. Seal,II
USGS Eastern Geographic Science Center
Reston, VA
Carol Russell, and Terry Lyons
U.S. EPA/Center for Environmental Solutions and
Emergency Response/ Land Remediation and
Technology Division,
Cincinnati, OH 45268

-------
Notice/Disclaimer
The U.S. Environmental Protection Agency, through its Office of Research and Development,
funded and conducted the research described herein under an approved Quality Assurance
Project Plan (Quality Assurance Identification Number K-LRTD-0017045-RT-1-0). It has been
subjected to the Agency's peer and administrative review and has been approved for publication
as an EPA document. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.

-------
Foreword
The U.S. Environmental Protection Agency (US EPA) is charged by Congress with protecting
the Nation's land, air, and water resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet this
mandate, US EPA's research program is providing data and technical support for solving
environmental problems today and building a science knowledge base necessary to manage our
ecological resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.
The Center for Environmental Solutions and Emergency Response (CESER) within the Office of
Research and Development (ORD) conducts applied, stakeholder-driven research and provides
responsive technical support to help solve the Nation's environmental challenges. The Center's
research focuses on innovative approaches to address environmental challenges associated with
the built environment. We develop technologies and decision-support tools to help safeguard
public water systems and groundwater, guide sustainable materials management, remediate sites
from traditional contamination sources and emerging environmental stressors, and address
potential threats from terrorism and natural disasters. CESER collaborates with both public and
private sector partners to foster technologies that improve the effectiveness and reduce the cost
of compliance, while anticipating emerging problems. We provide technical support to EPA
regions and programs, states, tribal nations, and federal partners, and serve as the interagency
liaison for EPA in homeland security research and technology. The Center is a leader in
providing scientific solutions to protect human health and the environment.
Gregory Sayles, Director
Center for Environmental Solutions and Emergency Response

-------
Abstract
Remediation of historical mine sites with a legacy of environmental impairment typically aims to
restore water, soil, and sediment quality to levels that comply with relevant regulatory guidelines.
The achievement of compliance goals improves surrounding ecosystems, which can then provide
services to humans such as improved water quality for drinking, improved water quality for fish
populations that are attractive to anglers, and desirable vistas, among others. Although the costs
of varying degrees of remediation are clearly estimated, the value of ecosystem service
improvements is not explicitly considered in remedial planning. The purpose of this study is to
evaluate approaches for valuing ecosystem services affected by mine remediation and predicting
the growth of those post-remediation benefits. The literature was surveyed to identify studies that
value ecosystem service endpoints that can be linked to geo-environmental models of mine
pollution remediation. Benefit valuation applicable to mine site pollution were identified for
aquatic habitat, drinking water, groundwater, water supply reliability, lead contamination, air
particulate matter, mercury emissions, residential views, natural land cover, and fish populations
suitable for recreational angling (fish large enough to catch). To demonstrate the valuation of
ecosystem services affected by mine remediation, this exercise is applied to catchable
populations of brown trout in ten sample locations at two legacy mine sites on the National
Priorities List—specifically the Leadville district (California Gulch Superfund site) in the
Arkansas River watershed and the Gilman district (Eagle Mine Superfund site) in the Eagle River
watershed. The impact of dissolved metals on water quality, aquatic macroinvertebrates, brown
trout populations and their growth are modeled to allow the valuation of catchable brown trout
populations. This application, combined with the identification of studies that value ecosystem
services affected by mine pollution, outlines a framework for valuing changes in mine site
pollution for future research where site data are more readily available. While this study
evaluates the benefits of remediating mine site pollution, this framework may also be applied to
an increase in mine site pollution if site data are sufficient to allow the linkage of geo-
environmental modeling and ecosystem service endpoints.
II 	 I "MM ¦ ¦ '' ¦ 111 ii i Mi i »l
Historical mining commonly left a legacy of environmental impairment that affected surface
water, groundwater, soil, sediment, and associated ecosystems. Remediation of legacy mine sites
typically aims to restore water, soil, and sediment compositions to concentrations that comply
with relevant regulatory guidelines. A consequence of meeting compliance goals around legacy
mine sites is the restoration of a variety of ecosystem services, such as improved water quality
for drinking, improved water quality for fish populations that are attractive to anglers, and
desirable vistas. Despite the positive effects that remediation can have on ecosystem services, the
value of those improvements is not an explicit goal of remedial planning. The purposes of this
study are to evaluate approaches for the valuation of ecosystem services affected by mine

-------
remediation, and to explore approaches for predicting the growth of those post-remediation
benefits.
Legacy mine sites on the National Priorities List were the target of this exercise because of the
comprehensive and prescribed approach required to characterize environmental risks and
ecological damages. Four different mineral deposit types, with two sites each, were initially
selected for investigation, but inability to access and acquire data and reports from these
Superfund sites resulted in limiting this exercise to the carbonate-hosted lead-zinc-silver deposits
in central Colorado, specifically the Leadville district (California Gulch Superfund site) in the
Arkansas River watershed, and the Gilman district (Eagle Mine Superfund site) in the Eagle
River watershed. The primary data used for both areas are from reports by state and federal
agencies. Within these two sites, data limitations constrained the ecosystem services scope of the
study to recreational angling.
The deposits and mines in both districts experienced similar production histories, which resulted
in similar environmental legacies. They were all mined by underground methods, the ore was
crushed, sulfide concentrates were produced by flotation techniques, and mill tailings were
disposed on the surface. The Leadville district has the additional feature of tunnels that were
driven into the mountain side to facilitate easy removal of ore from the mine workings and to
drain the mine workings. Drainage from the tunnels is currently being addressed by active water
treatment plants. The remaining ecological and water-quality issues have largely been addressed
through source control - removal and disposal of solid mine waste. The water-quality issues
resulted from acid-mine drainage mixing with slightly alkaline river water. The mixing
effectively removed iron by precipitation of hydrated ferric oxides. The low solubility of lead
sulfate (anglesite) and lead carbonate (cerussite) and sorption on to the hydrated ferric oxides
effectively limited dissolved concentrations of lead. Likewise, a significant amount of the copper
was removed by sorption. Downstream exceedances of water quality criteria for the protection of
aquatic life in both the Arkansas River and Eagle River watersheds have been dominated by zinc
followed by cadmium and copper. The contribution of each element to the toxicity of the water is
zinc: 68 to 80%, cadmium: 10 to 20%, and copper: 10%. Downstream reductions in the
concentrations of these elements are due to dilution by groundwater and surface water influxes.
The bulk of remedial activities was completed in the Arkansas River watershed in 2000 and in
the Eagle River watershed in 2001, although minor environmental problems remain to be
addressed.
The improvements in water quality in both watersheds due to remediation have resulted in steady
improvement in the populations of aquatic macroinvertebrates and fish, specifically brown trout.
The Upper Arkansas River fishery achieved "Gold Medal" status from Colorado Parks and
Wildlife in 2014. Gold Medal status is reserved for waters with a combination of high population
density and at least 12 trout 14-in. (or longer) per acre. This can be achieved either by natural
populations or fisheries management. The growth in trout populations can be modeled using the
logistic function that uses a starting population, a growth rate, and a carrying capacity as inputs.
These input parameters were all estimated based on published trout population data for these
watersheds. More sophisticated population dynamic models are available in the literature, but all

-------
require a larger set of input data that is not currently available for the Arkansas and Eagle River
watersheds. Nevertheless, published studies that assessed the temporal variations in the
coefficient of variation for abundance of brown trout in healthy streams in Colorado ranged
between 15 and 82%, compared to an average for all trout species in the United States of 49%.
This large range in the coefficient of variation for temporal variations suggests that the logistic
function is adequate for predicting trout population growth in this study. In general, the logistic
function adequately describes population growth in these two watersheds. Specific sites that fall
significantly below the model typically have residual environmental issues, such as fluvial
tailings, untreated tunnel drainage, or seepage of leachate from waste piles that represent
ongoing sources of contamination.
For an ecosystem service valuation of recreational angling related to brown trout population at
these two sites, the relevant ecosystem service endpoint (which links the natural and social
science models) is the portion of the brown trout population that is large enough to be caught
(hereafter referred to as "catchable"). The goal is to estimate the change in net economic value
resulting from the increase in catchable trout population for its initial level to its carrying
capacity. The ecosystem service valuation literature provides two approaches for estimating the
change in net economic value to recreational anglers. The first approach directly values the
catchable fish population by combining the model fish population estimates with the estimated
value per catchable trout. The second approach multiplies a value estimate (for the number of
fish caught by recreational anglers) by the ratio of fish caught and catchable fish - which also
estimates the change in net economic value per catchable fish.
Linking biological and valuation models to determine the ecosystem service valuation of
catchable trout population was only successful at 10 sampling sites at locations in the Upper
Arkansas River and Eagle River. To provide a broader context, for the change in net economic
value due to the California Gulch Superfund remediation, an ecosystem service valuation was
conducted (without the benefit of geo-environmental modeling) along the impacted stretch of the
Upper Arkansas River. Due to data limitations, this ecosystem service valuation focused on the
use value of fish, i.e., those caught by recreational anglers.
Finally, outside of the specific study sites, the environmental and ecosystem service valuation
literature was widely surveyed to assess the literature's ability to value impacts of mine site
pollution through an ecosystem service framework. Catchable target fish populations are
explicitly modeled while guidance for valuation of ecosystem services related to aquatic habitat,
drinking water, groundwater, water supply reliability, lead contamination, air particulate matter,
mercury emissions, residential views, and natural land cover are also provided to enable future
research where site data prove to be more forthcoming.

-------
Table of Contents
Notice/Disclaimer	3
Foreword	4
Abstract	5
Executive Summary	5
Introduction	11
Defining ecosystem services and their endpoints	11
Ecosystem services impacted by mine site pollution	12
Initial set of ecosystem services explored	13
The endpoint problem: Difficulties in linking natural and social science research	14
Selection of ecosystem service endpoints capable of linking geo-environmental and valuation models
	16
Central Colorado study area	16
Statement of Problem	17
Scope Of Study	18
Sources Of Data	18
Environmental data	19
Economic data	20
Economic Valuation Background	20
Measures of Economic Value	21
Non-market valuation techniques	24
Benefit transfer	25
The use of meta-analysis for benefit transfer	26
Summary of environmental valuation literature regarding mine sites	26
Benefit Transfer Model for Valuation of Selected Ecosystem Services	28
Catchable target-fish population for recreational angling	29
WTP for an Angling Day	29
WTP for Fish Caught	29
WTP for Catchable Fish	35
Additional ecosystem services	36
Summary of Valuation Model	36
Study Site Description	37
Physical setting	37

-------
Regional geologic setting	37
Economic geology and mining history	40
Hydrologic setting	42
California Gulch	42
Eagle River	44
Water-quality variations	46
California Gulch	46
Eagle River	51
Mining-environmental landscape	56
Leadville District (California Gulch)	56
Gilman District (Eagle Mine)	57
Environmental Setting	57
Water	57
Influences on water quality	57
Aquatic organisms	67
Arkansas River/California Gulch	67
Eagle River	69
Discussion	72
Geo-environmental setting	72
Aquatic Setting	77
History of aquatic ecological recovery	77
Uncertainty in inter-annual healthy populations	78
Modeling population growth	79
Effects of residual wastes	84
Link between fish population growth predictions and estimates of their value	84
Ecosystem service valuation	85
Four sampling sites in the Upper Arkansas River	85
Data and methods	88
Discussion and Conclusions	91
Six sampling sites in the Eagle River	91
Recommendatons	92
Conclusions	92
References	94

-------
Appendices	102
Appendix A: Review of Environmental (and Ecosystem Service) Valuation Literature Pertaining to Mine
Site Pollution	102
Social Cost-Benefit-Analyses of Mine Remediation Schemes	102
Social Cost-Benefit-Analyses of Proposed Mine Sites	104
Natural Resource Damage Assessments Related to Mine Sites	105
References for Appendix A	106
Appendix B: Starting Points for Benefit Transfer Modeling of Additional Ecosystem Services Impacted by
Mine Site Pollution	108
Soil Quality for Human Health: Lead	108
Air Quality for Human Health: Particulate Matter	109
Air Quality for Human Health: Mercury Emissions	109
View from a Residence	113
Wetland, Open Water, Shrubland, Grassland, and Terrestrial Habitat	114
Municipal and Household Intake Water Quality	114
Drinking Water and Groundwater: An Incremental Approach	118
Household and Municipal Water: Supply Reliability	119
Non-Use Value of Acquatic Habitat	119
References for Appendix B	121
Appendix C: A Valuation of the Upper Arkansas River Fishery Recovery from Leadville to Canon City -
Providing Context for the Selection of Ecosystem Service Endpoints	124
Purpose and Scope	125
Data and Method	126
Non-Use Value of Aquatic Habitat Improvement	129
Results	129
Discussion, Implications and Conclusion	132
References for Appendix C	134

-------
5111
Extractive industries, especially the mining of mineral resources, have historically played
important roles in the economic development of the United States. Many parts of the United
States owe their existence and growth to mining. The California Gold Rush in 1849 and the
Colorado Silver Boom of 1879 advanced the development of those regions. This development
led to the sustainable economies found today, even though metal mining has essentially ceased in
the areas of initial focus. The development included not only the mines themselves, but also the
physical, commercial, and social infrastructure to support the mines including, roads, railroads,
energy supplies, stores, schools, hospitals, and other emergency services. However, the
environmental legacy of the mines, particularly that of historical mines, can also affect the long-
term economic benefits derived from ecosystems, otherwise known as ecosystem services, in the
surrounding regions.
Defining ecosystem services and their endpoints
To evaluate the effect of the environmental legacy of mines on long-term economic benefits
derived from surrounding ecosystems, the links between the natural and social sciences must be
explored. The field of ecology breaks an ecosystem into processes and services. Ecosystem
processes are the complex physical and biological interactions that underlie the natural world,
such as nutrient cycling, regulation of water chemistry, and maintenance of biological diversity.
By contrast, an ecosystem service is the result of ecosystem processes. Ecosystem services
sustain or enhance human life.
To link natural and social science models, this analysis requires a definition of ecosystem
services that satisfies the underlying ecological science, as well as the requirements of economic
application. Ecologists historically define ecosystem services as the benefits that humans derive
from ecosystems (Millennium Ecosystem Assessment, 2005, Wallace, 2007). This definition
tends to "double count" the benefits that humans derive from the ecosystem because it
encompasses ecosystem processes and ecosystem services. In other words, the value of the
function and the value of the service are both counted.
In contrast, Boyd and Banzhaf (2007) defined ecosystem services as the final components of
ecosystem processes. Instead of being benefits, ecosystem services are viewed as "components of
nature, directly enjoyed, consumed, or used to yield human well-being." By this definition,
ecosystem services are not the benefits humans obtain from ecosystems, but rather the final
ecological components that flow from the ecosystem. As such, these endpoints are combined
with other inputs to create human benefit (Boyd and Banzhaf, 2007, Boyd, 2007, Fisher et al.,
2008). This definition allows economists to incorporate ecosystem services as inputs in utility
functions, which they use to model human benefits. This advancement helps social scientists to
measure how much people care about changes in ecosystem services.

-------
To illustrate this definition, Boyd and Banzhaf (2007) provided the following example of how
ecosystem services (surface waters, fish populations, and scenic surroundings) combine with
human-made inputs (equipment, time, and access) to produce benefits for recreational anglers:
Consider, for example, the benefits of recreational angling. Angling requires ecosystem
services, including surface waters and fish populations, and other goods and services
including tackle, boats, time allocation, and access. For this reason, angling itself-or 'fish
landed'-is not a valid measure of ecosystem services. More fish may be landed simply
because better tackle are [sic] used... The fish population, surroundings, and water body
are the 'ecosystem end products' directly used by anglers to produce recreational
benefits. Thus, they are the ecosystem services that should be counted. The case of
commercial fishing is similar, but here aesthetics are unimportant, so only the target fish
populations need to be counted as ecosystem services.
This example illustrates that ecosystem services are not human benefits. Rather, they are
measurable, physical endpoints of ecosystem processes that are combined with other inputs by
humans to create benefit. Ecosystem services provide an avenue for valuation of environmental
quality because they are quantifiable. Changes in the quantity of ecosystem services impact
human benefit and these impacts can be valued.
Ecosystem services impacted by mine site pollution
The legacy of historical mines can include water-quality degradation due to mine drainage,
sediment-quality degradation from solid mine waste erosion, soil-quality erosion from wind-
blown tailing dust, and degradation of vistas in mining areas. Water-quality issues result from the
interaction of precipitation, surface water, and groundwater with rock within the mine workings,
waste rock excavated to access ore, or waste from processed ore. Sediment-quality and soil-
quality issues commonly result, respectively, from the water and wind erosion of mine waste,
particularly mill tailings that have been inadequately disposed. Mill tailings (typically the size of
sand or silt) are the waste materials produced by finely crushing ore and removing the ore
minerals using a variety of physical and chemical processes. Mining also alters the visual
appearance of the landscape around mines. Open pits are rarely filled once mining has ceased.
Waste rock and tailings storage facilities are common features on mine landscapes that have
historically been left with minimal restoration after mining has stopped. In contrast, best
practices for modern mining seek to restore landscapes to useful purposes without impacting the
environment. Mine buildings and structures such as head frames for mine shafts may be a
significant part of the legacy of mining landscapes. In some cases, these historical structures may
be an important part of the appeal of these areas for tourism.
Mine drainage can affect several ecosystem services. Acid or dissolved metals can affect aquatic
organisms, such as aquatic invertebrates that form the base of lotic ecosystems, and fish. Fish
support terrestrial ecosystems and can support recreational or subsistence fisheries depending
upon location. Mine drainage can also contain one or more solutes that may exceed drinking
water standards and affect the potability of groundwater or surface water in the vicinity of mines.

-------
Sources of drinking water are treated prior to human consumption. In the case of a stream
impacted by mining activity, there may be high concentrations of several metals present, such as
copper, lead, or arsenic, leading to increased costs to treat the water to a level suitable for human
consumption. Impacts of mine drainage on water quality in mining districts, therefore, have
economic impacts on the ecosystem services supported by water quality.
Solid mine waste can also affect ecosystem services, leachate from waste piles can serve as a
source of mine drainage, fine-grained mill tailings can be a source of wind-blown contamination
for residential and agricultural soils, and ongoing erosion of poorly contained tailings (as well as
catastrophic failure of tailings storage facilities) can serve as a source of fluvial tailings that are
transported downstream and become deposited in quiescent settings. Such fluvial tailings can
serve as a chronic source of contaminants to surface water.
Ecosystem service impacts related to mines can also extend beyond direct impacts to water, soil,
or sediments. Many abandoned and existing mines are located in remote and often scenic
terrains. Streams originating in these locations may also be used for recreational activities, such
as fishing, rafting, hiking, and camping. Additionally, terrestrial and riparian vegetation may be
lost from changes to the soil and sediment composition, resulting in undesirable aesthetics, loss
of cover for camping, and/or change in stream temperature. While other types of activities may
influence these same services, such as logging, clear cutting for electrical lines or wind turbines,
or laying of pipelines, we chose locations that had mineral extraction as the dominant (or
preferably sole) source of influence on the ecosystem.
Initial set of ecosystem services explored
In this research, we use existing data from mined locations to evaluate any losses of ecosystem
services that may result from that mining activity. Specifically, we focus on sites from the U.S.
Environmental Protection Agency (USEPA) National Priorities (Superfund) list from their
Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) program.
Sites from the Superfund list were chosen because of the detail and depth of data gathered as part
of the prescribed remedial investigations and subsequent post-remediation monitoring.
Table 2.1 below presents the ecosystem services that were initially identified as potential foci for
this research. In many cases, water quality is a metric. However, within water quality, there are a
number of measurements that may serve as the basis for evaluation. Assessment based on a sole
parameter, or analyte, may not be possible because all parameters contribute to the evaluation.
Therefore, water quality is the metric, but parameters and analytes are how the metric will be
evaluated. As described in the following sections, it was decided to focus on target fish
populations as they related to sport fishing because this ecosystem service had the best data set
available for this study.

-------
Table 2.1. Benefits, Ecosystem Services, Metrics, and Valuations Considered
Benefits
Ecosystem
Services
Metrics
Valuations
Human Health (Drinking
water)
Quality of
stream water,
quality of
groundwater
Concentrations of
constituents in the water (for
example, copper, cadmium,
hardness), water quality
parameters (for example, pH,
dissolved oxygen, color,
turbidity)
Avoided costs for
treatment or
replacement
Recreation (fishing,
hiking, camping, rafting,
swimming)
Target fish
population
Numbers of fish
Travel cost
surveys
Recreation (fishing,
hiking, camping, rafting,
swimming)
Surface water
body existence
and quality of
water
Miles of streams
Quality of stream water
Travel cost
surveys, avoided
costs for
treatment or
replacement
Recreation (fishing,
hiking, camping, rafting,
swimming)
Natural land
cover
Acres of riparian regions, soil
types
Travel cost
surveys
Aesthetic values
Vistas
Acres of undisturbed land
Hedonic pricing
method,
contingent
valuation
Aesthetic values
Water quality
(clarity)
Turbidity and color
Hedonic pricing
method,
contingent
valuation
Existence values
Species
abundance
Variety/number of species
Contingent
valuation, choice
modeling
Existence values
Wilderness
Acres of undisturbed land
Contingent
valuation, choice
modeling
The endpoint problem: Difficulties in linking natural and social
science research
Despite the advancement that the Boyd and Banzhaf (2007) definition brings to the linkage of
natural and social science models, problems persist (Boyd, 2007, Kontogianni et al., 2010). Boyd

-------
(2007) explained that "if linked social and natural science is a relay race, endpoints are the
baton. The problem is that the baton never gets handed off smoothly." An example of the
endpoint problem is provided by the quotation above from Boyd and Banzhaf (2007) (Section
1.1) on the ecosystem service of fish population and the benefits derived from recreational
angling.
The endpoint problem is that anglers do not assign an economic value for the target fish
population itself (Ng, 2011). Instead, they can assign a value for the trip that they take to an
angling destination or the day that they spend fishing. As a result, economic valuation studies
have focused on the value of an angling trip or day (Ng, 2011). Boyd and Banzhaf (2007)
referred to the angling trip/day in their quote as "angling itself', which is "not a valid measure of
ecosystem services."
Therefore, for an ecosystem service valuation, natural scientists can successfully model the
ecosystem service (target fish population), while economists are only able to model the value of
a related - but different - environmental good. In fact, the angling trip/day is two degrees
removed from the ecosystem service (catchable fish). The number of catchable fish influences
the number of fish caught (first degree of separation), and the number of fish caught influences
the number, and value of angling days/trips (second degree of separation) (Johnston et al., 2006,
Loomis andNg, 2009, Mazzotta, 2015, Ng, 2011, USEPA, 2006a).
A body of economic research (USEPA, 2006a) and a journal publication (Johnston et al., 2006),
resolved the second degree of separation by analyzing the valuation literature to determine what
information it revealed about anglers' value for catching another fish. In other words, USEPA
(2006a) conducted an analysis of angling day valuation studies to estimate how changes in the
number of fish caught affected the value, or number, of angling days. However, Boyd and
Banzhaf (2007) referred to the number of fish caught in their quote as "fish landed", which is
also not a valid measure of ecosystem services. Nonetheless, this research moves economists one
step closer to be being able to estimate the value of catchable fish.
Out of the 405 recreational angling valuation studies surveyed by USEPA (2006a), only four
were identified by this analysis as providing a possible remedy for resolving the first degree of
separation. The first two studies (Johnson et al., 1995, Mazzota et al., 2015) addressed the gap by
making static, proportional assumptions about the impact of changes in target fish population on
changes in the number of fish caught by anglers. The second two studies (Loomis and Ng, 2009,
and Ng, 2011) originated from the same body of research and addressed the gap by
econometrically modeling the catch per day.
Johnson et al. (1995) assumed that 60% of stocked trout were caught by anglers. Mazzota et al.
(2015) assumed that the percentage change in the catch rate is equal to the percentage change in
the abundance of catchable fish. Loomis and Ng (2009) and Ng (2011), on the other hand,
modeled the catch per day as a function of the number of catchable fish, angler skill, target
species, and the number of hours fished each day. While these four studies help economists to
value the ecosystem service in question, future research should be dedicated to the dynamic
modeling of the relationship between fish population and fish catch, even in the presence of data
constraints (Ng, 2011).

-------
In conclusion, the "relay race" of linked natural and social science is still working to hand the
baton off smoothly. For the purpose of this ecosystem service valuation, the endpoint problem is
particularly prevalent for trout population. With the endpoint problem in mind, we describe
which of the ecosystem services from Table 2.1 were selected for economic valuation modeling.
It is important to state that these selections were made in light of serious data difficulties
described in detail below.
Selection of ecosystem service endpoints capable of linking geo-
environmental and valuation models
Button et al. (1999) provided guidance on the selection of ecosystem services for valuation of
mine site pollution: "In terms of acid mine drainage (AMD) remediation, the importance of
restoring water quality, the restoration of scenic beauty, and the reintroduction of fish stock
emerge as key issues". One ecosystem service that follows this guidance, captures the majority
of impact value, and can be valued by the existing valuation literature is catchable target fish
population.
Olander et al. (2015) suggested using 'catchable' target fish population as the endpoint for the
ecosystem service of target fish population. This term focuses ecosystem service modeling on the
segment of target fish population that is large enough to be caught by anglers. It also relegates
the target fish population itself to the status of an ecosystem process. Catchable target fish
population can increase in a stream segment as a result of mine-site pollution remediation.
Alternatively, mine site development can decrease catchable target fish population by altering
water flow, reducing habitat, or accidentally releasing toxic effluent. In conclusion, data issues
limited the successful linkage of geo-environmental and valuation modeling to the single
ecosystem service of catchable trout population.
Central Colorado study area
Central Colorado provides an excellent opportunity to investigate the economic impact of legacy
mining on ecosystem services using a geologically based (geo-environmental) approach. Central
Colorado has historically been mined for a number of commodities from a variety of mineral
deposit types. The most significant deposit types include carbonate-hosted lead-zinc-silver
deposits, porphyry molybdenum deposits, and epithermal gold deposits. This study will focus on
the carbonate-hosted lead-zinc-silver deposits of Central Colorado, which provided the impetus
for the early settlement of this region. The main historical mining districts include Gilman,
Leadville, Kokomo, Aspen, and Sherman. The Leadville district is the most economically
significant of these camps (Beaty et al., 1990). The Leadville (California Gulch) and Gilman
(Eagle mine) districts were selected for this study because of the amount of data available from
these areas. Both sites offer well documented case studies of environmental impacts caused by

-------
geologically similar abandoned mines, their remediation, and the subsequent and ongoing
recovery of downstream surface water ecosystems.
' '! ¦ I "hi. III si I I I 11;' 11 i
Abandoned mines are complex in terms of the source, transport, and fate of contaminants and the
effect of these contaminants on the surrounding ecosystems and the services that these
ecosystems provide. The source of contaminants is influenced by the geology of the deposit, the
mining methods used, the ore-processing used, the waste management practices used, the
hydrologic setting, and climate - among other factors. The transport and fate of contaminants are
influenced by the local climate and hydrologic setting, the geology of the watershed as it relates
to the chemistry of receiving water bodies, and the specific contaminants themselves. The impact
of mine drainage and solid mine waste on aquatic organisms is equally complex and depends on
climate, hydrologic setting, the physical characteristics of the in-stream habitat, the number of
contaminants, and a variety of processes that serve to either increase or decrease the
concentrations and/or bioavailability of contaminants in surface water.
The recovery of the habitat in these streams impacted by abandoned mines is complicated by the
success in addressing water-quality issues, the distribution of residual sources of contamination,
the state of the habitat available once remediation is complete, and the availability of food
sources for aquatic organisms. The linkage of remediation to the value of an ecosystem service
requires knowledge of specific pathways of contaminants to the ecosystem services that they
impact, how remediation will influence those pathways in both the short and long term, and how
those changes will affect the value of those ecosystem services in the present and future.
For the present study, the exercise is restricted to the ecosystem services related to recreational
fishing because reasonably clear links can be established among water quality, the abundance of
catchable fish - specifically brown trout, and the value of those fish to anglers. The Leadville
(California Gulch) and Gilman (Eagle Mine) districts have geologically similar carbonate-hosted
lead-zinc-silver deposits that were mined and processed using similar approaches. Both sites
were addressed through the USEPA CERCLA program for remediation, and both sites have site-
specific environmental data spanning the period from prior to remediation, through remediation,
to after most remedial activities have been completed. The data encompass the physical and
chemical hydrology of downstream habitat, including information about fish and
macroinvertebrate populations extending at least 10 km downstream of the abandoned mine
sites.
The overall goal of this project is the valuation of ecosystem services to estimate the benefits for
what will be gained from remediation of abandoned mine sites having similar characteristics to
these existing sites. Specifically, the main objectives of this project are:
1. To develop a geologically based (geo-environmental) mineral-deposit model for carbonate-
hosted lead-zinc-silver deposits in Central Colorado that identifies specific geochemical risks
in a context that is relevant for the hydrogeologic setting of these deposits,

-------
2.	To incorporate damages to ecosystem services, specifically for recreational fishing, into a
watershed scale context spanning the period from before remediation to after remediation,
3.	To develop a model linking changes in water chemistry and aquatic habitat to changes in fish
populations, and
4.	To link changes in fish populations to changes in their value as an ecosystem service, specific
sport fisheries.
Scope Of Study
The original plan for this project was to consider a number of mineral deposit types in a number
of hydrologic and climatic settings. The original list included porphyry copper deposits (for
example, Bingham Canyon, Utah, Morenci, Arizona), epithermal gold deposits (for example,
Cripple Creek, Colorado, Golden Sunlight, Montana), volcanic-associated massive sulfide
deposits (for example, Holden, Washington, Greens Creek, Alaska), carbonate-hosted lead-zinc-
silver (Pb-Zn-Ag) deposits (for example, Leadville and Colorado Eagle, Colorado), and
Mississippi Valley-type lead-zinc deposits (for example, southeastern Missouri). The list of
potential deposit types to include in the study was narrowed based on the availability of relevant
secondary environmental and ecosystem services data meeting the data-quality criteria of having
been acquired by documented and approved methods with defined data quality objectives, i.e.,
accuracy and precision, etc., that were either consistent with methods approved by the federal
government or are universally accepted. The data availability exercise required the study to be
focused on the carbonate-hosted Pb-Zn-Ag deposits of Central Colorado (Leadville and Gilman
mining districts) because those areas are the only sites for which adequate environmental data are
available. Furthermore, the ultimate adequacy of data varied between those two sites with the
Leadville district having the most suitable data set.
A similar exercise was conducted to select ecosystem services for this study. Initially, plans were
to include the value of drinking water, vistas, aquatic habitat, and recreational fishing, but
insufficient data were identified to allow quantitative evaluation of the first three ecosystem
services. This project focuses exclusively on the value of sport-fishing recreation.
Sources Of Data
Data fell into three broad categories: geologic, environmental, and economic. Data were sought
and evaluated based on the data-quality hierarchy described below, in order of descending
quality:
1.	U.S. EPA or U.S. Geological Service (USGS) data collected under the respective agency's
quality assurance/quality control (QA/QC) program,
2.	State or tribal data collected under respective approved QA/QC programs,
3.	Data collected by other federal agencies (and their contractors) having approved QA/QC
programs (for example, Fish and Wildlife Service, Bureau of Land Management),

-------
4.	Data collected, peer-reviewed and published by academic organizations having described
QA/QC protocols, and
5.	Published data with limited QA/QC oversight (for example, independent contractors that may
or may not have established protocols) or publications that may or may not be peer-reviewed
(for example, conference proceedings from sources other than items 1-3 above).
Data in categories 1-4 above were used quantitatively, and data from category 5 was used only
qualitatively.
Environmental data
A significant portion of the hydrologic, hydrochemical, and biologic data used in this study was
derived from USGS data sources or reports published by the USGS and state agencies - for
example, Clements et al. (2010) and Woodling et al. (2005). The USGS National Water
Information System (NWIS, http://nwis.waterdata.usgs.gov/usa/nwis/qwdata/) was used for
stream flow and some water-quality data (pH, specific conductance) in both the Arkansas River
and Eagle River watersheds. Sources of geologic, hydrologic, geochemical, and biologic data are
summarized in Table 5.1.
Table 5.1. Sources of Geologic, Hydrologic, Geochemical, and Biologic data
Source
Data Types
Comments
General
Beaty et al. (1990)
Beaty (1990)
Thompson et al. (1990)
Wallace (1993)
Geologic setting
Geologic setting
Geologic setting
Geologic setting
Descriptive paper for
geologic background that
includes grade and tonnage
details
Descriptive paper for
geologic background
Descriptive paper for
geologic background
Descriptive paper for
geologic background
California Gulch
Clements et al. (2010)
Water quality, sediment
quality, macroinvertebrate
populations, fish
populations
Study includes two sites
upstream and two sites
downstream of California
Gulch. Study spans 17
years.
Eagle River
Woodling et al. (2005)
Water quality, sediment
quality, macroinvertebrate
populations, fish
populations

-------
Economic data
Economic data were primarily obtained from peer-reviewed studies as detailed in Table 5.2
below. These studies were all meta-analyses of environmental valuation literature that proved
capable of valuing ecosystem services relevant to this project. Creel survey data - including
information about angling hours/days, the number of fish caught, their location - were obtained
from the Colorado Parks and Wildlife annual fisheries inventory reports concerning the Upper
Arkansas River from the years 2012 and 2013. Sources of economic data are summarized in
Table 5.2.
Table 5.2. Sources of Economic and Ecosystem Services Data
Source
General
Data Types
Comments
USEPA (2006a)
Johnston et al. (2006)
Value estimates for an angler's
willingness to pay to catch an
additional fish
Value estimates for an angler's
willingness to pay to catch an
additional fish
A meta-analysis of the non-
market valuation literature on
recreational angling, with the
goal of estimating the angler's
willingness to pay to catch
another fish of the angler's
target species
This data source is the journal
article publication of the
research conducted in USEPA
(2006a) report described above.
California Gulch
Policky (2012)
Policky (2013)
Creel census data (angling
hours/days, catch rate,
proportion of out-of-state
anglers) for the years 1995,
2008, 2012
Extrapolation of creel site data
to the broader river reach that
the creel site is within - for the
years 1995,2008, 2012
Colorado Parks and Wildlife
Fisheries Inventory report for
the year 2012
Colorado Parks and Wildlife
Fisheries Inventory report for
the year 2013 - which provided
angling estimates for each river
reach from its respective creel
study site
II " ll-' I II i I'Ill" villi lil ''I | I 111 ¦1
This section defines and describes various economic values that can inform benefits derived
from ecosystem services to illustrate the effect that mine site pollution can have on the economic

-------
benefits of surrounding ecosystem services. Methods of estimating these economic values, in the
absence of formal markets, are discussed with a particular focus on the use of existing economic
valuation literature - a practice known as benefit transfer. Finally, a summary discussion of the
existing mine site valuation literature is provided to lend context to the benefit transfer model
described in the following chapter.
Measures of Economic Value
An excerpted discussion from Boyle et al. (1998) is provided in Figures 6.1 and 6.2 to illustrate
several measures of economic value. This discussion is clear and is directly applicable to the
primary ecosystem service values estimated in the benefit transfer model for catchable trout
population and in the Upper Arkansas River ecosystem service valuation application.
Trips per Year
Figure 6.1: Illustrative demand curve for fishing trips by an individual angler. Modified from Boyle
et al. (1998).
Figure 6.1 shows a demand curve for a representative angler. "The downward sloping demand curve
represents marginal willingness to pay per trip and indicates that each additional trip is valued less by
the angler than the preceding trip. All other factors being equal, the lower the cost per trip (vertical
axis) the more trips the angler will take (horizontal axis). The cost of a fishing trip serves as an
implicit price for fishing since a market price generally does not exist for this activity.
At $60 per trip, the angler would choose not to fish, but if fishing were free, the angler would take 20
fishing trips. At a cost per trip of $25 the angler takes 10 trips, with a total willingness to pay of $375
(area acde in Figure [6.1]). Total willingness to pay is the total value the angler places on
participation. The angler will not take more than 10 trips because the cost per trip ($25) exceeds what

-------
he would pay for an additional trip. For each trip between zero and 10, however, the angler would
actually have been willing to pay more than $25 (the demand curve, showing marginal willingness to
pay, lies above $25).
The difference between what the angler is willing to pay and what is actually paid is net economic
value. In this simple example, therefore, net economic value is $125 (($50 - $25) 10 2) (triangle
bed in Figure [6.1]) and angler expenditures are $250 ($25 x 10) (rectangle abde in Figure [6.1]).
Thus, the angler's total willingness to pay is composed of net economic value and total expenditures.
Net economic value is simply total willingness to pay minus expenditures. The relationship between
net economic value and expenditures is the basis for asserting that net economic value is an
appropriate measure of the benefit an individual derives from participation in an activity and that
expenditures are not the appropriate benefit measure.
Expenditures are out-of-pocket expenses on items an angler purchases in order to fish. The remaining
value, net willingness to pay (net economic value), is the economic measure of an individual's
satisfaction after all costs of participation have been paid.
Summing the net economic values of all individuals who participate in an activity derives the value to
society. For our example let us assume that there are 100 anglers who fish and all have demand
curves identical to that of our typical angler presented in Figure [6.1], The total value of this sport
fishery to society is $12,500 ($125 x 100)" [emphasis added]

-------
Shift in demand curve due to
increase in catch rate
Change in net economic value
due to increase in catch rate
0 5 10 15 20
Number of Trips per Year
25
Figure 6.2: Impact of an increase in catch rate on the angler demand curve for fishing trips.
Modified from Boyle et al. (1998).
"In many instances, all or nothing values, as shown in Figure [6.1], are not appropriate. Rather, a
change in quality shifts the demand curve, thereby resulting in a change in net economic value
(Figure [6.2]). In these instances, the change in net economic value is the appropriate benefit measure.
For example, assume a management activity will increase catch rates for anglers by 10 percent. This
change in the resource results in a shift of the demand curve upward and to the right, as presented in
Figure [6.2], The benefit to the angler of this increase in catch rate is the area cfgd. Estimation of this
area is possible by including harvest rates as explanatory variables in the estimated [value] equations"
[emphasis added]
In other words, when the following analysis mentions benefits, it refers to the net economic
value (triangle bed) - which is also known as the net willingness to pay (WTP) and the consumer
surplus. Similarly, the term 'benefits of remediation' refers to the monetary value of the area
cfgd in Figure 6.2 above. The final important point is that the change in net economic value is
the appropriate benefit measure that we will focus on.

-------
Non-market valuation techniques
Economic research informs environmental decisions through non-market valuation, which allows
a more direct comparison of costs and benefits relating to environmental change, such as
environmental degradation related to abandoned mines or the subsequent remediation of those
mines (Carson et al., 1992, Costanza et al., 1998). The goal of this section is to evaluate the
capacity of the environmental valuation literature to quantify environmental damage from mine
site pollution (from all types of mine sites) in monetary terms.
Economists estimate the benefit from directly using an environmental service through
environmental valuation. Value derived from use is known as use value. Economists can also
estimate the intrinsic value that humans place on environmental services - known as non-use
value. Non-use values are received from the existence of a service that the individual would
never use. In some cases, non-use values can be quite large.
Economists use four different techniques to value environmental services: replacement cost,
revealed preference, stated preference, and benefit transfer. In the first technique, a replacement
cost for a lost environmental service is calculated. Replacement cost techniques measure the cost
of employing human capital or labor, in lieu of an environmental service. This is an intuitively
appealing method for examining the value of an environmental services, however, it is often the
least desirable method because cost does not necessarily equate with value (Loomis, 2000).
For example, a valuable environmental service (such as potable water) may be relatively
inexpensive to replace. In this case, the replacement cost represents a minimum value of benefits
that the environmental service provides. In contrast, the replacement of a fishery in an area not
inhabited by humans would be costly and provide no value. Therefore, using replacement costs
as a measure of value can be misleading for economic decision-making.
Revealed preference methods estimate environmental service values from consumer behavior
that is observed in real markets. Examples of revealed preference methods include travel cost
valuation, hedonic valuation, averting behavior valuation, and production function valuation.
Travel cost valuation employs travel time and additional expenses incurred by individuals to
value recreational sites. Hedonic valuation uses market data on property to isolate the value of a
particular environmental service - such as a view. Averting behavior valuation sums up expenses
imposed due to poor environmental services. Finally, the production function valuation
technique estimates the value of environmental services as inputs to the production process. The
observation of actual human behavior in real markets is an advantage of using the revealed
preference technique. However, a shortcoming of this approach is that only use values are
measured.
Stated preference methods gather environmental service values through surveys that detail
hypothetical changes in non-market services, for example air quality, and ask respondents what
they would be willing to pay for those hypothetical changes. Stated preference methods use
carefully crafted surveys to allow respondents to directly state their willingness to pay to avoid a

-------
loss of environmental service quality. Examples of stated preference methods are contingent
valuation and conjoint choice modeling. The stated preference technique is the only technique
that can capture non-use value, which is a major advantage of using stated preferences methods
(Haab et al., 2013). However, some economists see the fact that the methods rely on hypothetical
markets, instead of real transactions, as a drawback (Hausman, 2012). Over the years, there has
been a lively debate regarding the validity of stated preference methods (Carson et al., 2001,
Diamond and Hausman, 1994, Hanemann, 1994, Portney, 1994).
A detailed literature review of all of these methods would be voluminous. Instead, the
purpose of describing these methods is to demonstrate that a plethora of approaches are available
for valuing environmental services.
Benefit transfer
The valuation techniques described in the previous section all produce primary valuation studies.
Primary valuation studies are conducted for a specific site, time, and context. By their nature,
primary valuations are expensive. The large number of mine sites that could be studied precludes
primary valuation of many of them, let alone all.
Constrained financial resources and the high cost of primary valuation studies provided incentive
to find cheaper ways to determine the value of non-market services at new sites (Bingham et al.,
1992). The result came to be known as benefit transfer. This technique transfers a benefit
estimate across time and space from the primary study site (known as the study site) to another
site where a policy is being evaluated (known as the policy site). Wilson and Hoehn (2006)
explained that:
[B]eneft transfer uses economic information captured at one place and time
to make inferences about the economic value of environmental goods and
services at another place and time.
Hypothetically, if there are two identical populations, environmental services, and contexts, then
the valuation of the environmental service should be the same for both sites. The need for
environmental service valuation, coupled with the expense of conducting primary valuation
studies, has propelled benefit transfer forward as a widely employed method to approximate the
value of environmental services at different locations (Wilson and Hoehn, 2006). Since the early
1990's, benefit transfer has been used in federal regulatory impact analysis for non-market,
environmental goods (Boyle et al., 2010). The literature on mine site pollution valuation
(Appendix A) supports the use of benefit transfer:
Conducting very detailed case studies on any prospective location to ameliorate [acid mine
drainage] is costly. A hybrid approach is, therefore, advocated, using existing information
coupled with new, case-specific analysis. Although there is the need to gather some information
on individual sites and those directly affected by remediation schemes, the use of literature
reviews, meta-analysis, and other techniques (i.e., benefit transfer) facilitates value transfers
(Button et al., 1999).

-------
Initially, benefit transfer comprised: 1. an evaluation of the policy site, 2. selection of a
corresponding primary valuation study from the existing literature, and 3. direct transfer of the
primary study's results to the study site. This is known as unit value benefit transfer (Loomis,
1992). However, benefit transfer is not limited to a single primary study, but can be employed
using many primary studies. Using additional primary valuation studies can facilitate more
accurate benefit transfer estimates. To advance this improvement, valuation experts began to
isolate the effect of explanatory variables such as income, study site characteristics, and region
on the valuation result (willingness to pay) that primary studies had generated. Using this
information, a benefit transfer function is constructed with the intent of further improving the
accuracy of benefit transfer (Loomis, 1992).
The use of meta-analysis for benefit transfer
As more benefit value estimates emerged for the same environmental service, it became clear
that the resulting estimates were seldom of the same magnitude. Some even had conflicting
signs. Whereas literature reviews were useful in qualitatively evaluating valuation result
disparities, a more quantitative approach was required (Boyle et al., 1994). Economists began to
evaluate primary valuation studies statistically via meta-analysis (Boyle et al., 1994, Carson et
al., 1996, Smith and Huang, 1995, Smith and Kaoru, 1990, Woodward and Wui, 2001).
Meta-analysis is a quantitative analysis of valuation analyses that uses regression to determine
the factors that cause variation between primary study results. First, the meta-analyst identifies
determinants of variation between primary study estimates of willingness to pay (WTP) to avoid
a reduction in environmental quality (Nelson and Kennedy, 2009). Usually these determinants
include population income, population demographics, primary study site characteristics, study
method, pollutant type, and publication method (Bergstrom and Taylor, 2006, Navrud and
Ready, 2007).
Once a meta-analysis function is estimated, the explanatory variables are set to reflect the policy
site as closely as possible. The result is a meta-regression model benefit transfer (Kirchhoff,
1998, Rosenberger and Loomis, 2000, Shrestha and Loomis, 2001). Meta-regression model
benefit transfer reduces the error between benefit transfer estimates and site-specific estimates
when study and policy sites were not particularly similar (Kaul et al., 2013), which is the
majority of the time in benefit transfer.
Meta-regression models are currently the state-of-the-art instruments for synthesizing and
transferring benefit estimates from the environmental literature to unstudied ('policy') sites.
Therefore, in subsequent chapters, meta-regression models are relied upon heavily to construct a
mine site pollution benefit transfer model and to apply this model to mine site pollution.
Summary of environmental valuation literature regarding mine sites
To employ the benefit transfer valuation technique, primary valuation studies must be located
that correspond to the 'policy' site in question. Given that the goal of this analysis is to value
changes in ecosystem services related to mine sites, the following section summarizes a more

-------
detailed review of the environmental valuation literature that relates to mine site pollution. See
Appendix B for the full literature review. The number of environmental valuation studies of
abandoned, proposed, or operating mine sites is rather small. The purpose of this summary is to
illustrate why studies from this literature were not helpful in the construction of the mine site
pollution benefit transfer model discussed in the following sections.
The first set of studies in this literature comprises attempts at social cost-benefit analysis of mine
remediation schemes. Valuable information can be gleaned from these social cost-benefit
analyses of mine site remediation. However, a benefit transfer model for the purpose of this
analysis cannot be built upon these studies because of one or more of the following flaws. First,
some of these studies confuse the economic values of expenditures and net WTP (Randall et al.,
1978). Second, most of these studies conducted the valuation at the level of the whole site, rather
than focusing on specific ecosystem services (Damigos and Kaliampakos, 2003, Farber and
Griner, 2000, Michael and Pearce, 1989, Neelawala et al., 2013, Williamson et al., 2008). Third,
some studies were unable to produce a benefit estimate (Button et al., 1999, Mendes et al., 2007).
Finally, some study sites do not correspond to sites in the United States (Ahlheim et al., 2004,
Burton et al., 2012, Lienhoop and Messner, 2009).
The second set of studies in this literature comprises attempts at social cost-benefit analysis of
mine development schemes. Whereas these studies are also informative, they cannot be used for
benefit transfer because they are conducted at the site-level (Trigg and Dubourg, 1993) or they
are benefit transfer exercises themselves (Damigos and Kaliampakos, 2006, Unaldi et al., 2011).
The final source of valuation studies is natural resource damage assessments (NRDA). Estimates
of economic benefit values are sparse in the NRDA literature. However, the few economic
benefit estimates that are provided prove to be difficult to use. The contending NRDA valuations
of the Eagle Mine are an example (Rowe and Schulze, 1985, Ward et al., 1992). Instead of
valuing each environmental component separately, the plaintiffs valuation (Rowe and Schulze,
1985) used a contingent valuation method conducted at the site level. On the other hand, the
defendant's valuation (Ward et al., 1992) estimated the cost to limit exposure to pollution from
the Eagle Mine. Cost estimation misses the point of valuation by confusing the concepts of
expenditure and net WTP.
In contrast to the Eagle Mine NRDAs stands the valuation of the BP Oil Spill (Board et al.,
2013). Board et al. (2013) broke the impact down into the value of each ecosystem service that
enters human utility functions to create benefit. By assigning a value to the magnitude of each
ecosystem service's impact, that ecosystem service's value can be transferred to dissimilar sites
that share a common ecosystem service.
In other words, economists need to work out exactly which environmental services are being
valued and how the value of each service impacts the total value of damage or benefit. Focusing
on specific services is important to economists because of substitution, scale effects, adding up,
internal consistency, and external consistency (Carson et al., 1992, 2001, Diamond, 1996,
Diamond and Hausman, 1994).

-------
Ecologists agree with this sentiment and argue that economists do not understand what they are
valuing (Limburg et al., 2002). Ecologists argue that this misunderstanding leads to double
counting and confusion of the human subjects surveyed during non-market valuation (Limburg et
al., 2002). A remedy is provided by the ecosystem service framework, which the following
benefit transfer model follows closely.
II < 'h< 'in III vh i 'i hi : ¦1 'II i i 'In vii h si ¦ <|. ¦ i< ¦ 1 II „ • i< 'Hi
Services
The goal of this section is to apply environmental valuation methods to ecosystem services that
are affected by mine site pollution. The purpose of this goal is to construct a model that measures
how much people care about changes in environmental quality as a result of changes in mine site
pollution. This model relies on three basic concepts. The first concept is that economists can
estimate how important a change in environmental quality is to a sample of people by
ascertaining the amount they are willing to pay to avoid the change1. The second concept is that
ecosystems are made up of interactive processes among minerals, water, and biota, which create
ecosystem services that people combine with time, effort, and equipment to derive benefit. The
third concept is that if economists estimate an ecosystem service's value for a population sample,
then it is feasible to apply the results to similar ecosystem services for similar population
samples.
Combining these three concepts allows environmental valuation to communicate with ecosystem
modeling and produces a scientifically rigorous valuation model by linking natural and social
sciences. Such a model can transfer benefit estimates from relevant environmental valuation
papers to unstudied mine sites and elucidate environmental costs and benefits of changing mine
site pollution. Ecosystem service valuation helps to incorporate the value of ecosystem services
into decisions regarding abandoned mine lands, legacy sites, operating sites, proposed mines, and
closure of operating sites. For example, application of the ecosystem service valuation model
quantifies the benefits of the legacy site remediation at California Gulch.
The ecosystem service chosen for this project is catchable target fish populations. This
ecosystem service was chosen for two main reasons. First, clear links can be established
between the ecosystem service and the underlying geology, geochemistry, hydrology, and
mining method. Second, this ecosystem service encapsulates the majority of costs and benefits
related to mining projects (Button et al., 1999). This section reviews the literature and methods
used to construct the benefits transfer tool for catchable target fish population.
1 Or, conversely, the amount they would need to be compensated to accept the change.

-------
Catchable target-fish population for recreational angling
The example provided by Boyd and Banzhaf (2007) regarding ecosystem services related to the
benefits of recreational angling helps to clearly delineate what is being valued by this segment of
the benefit transfer model. For the purpose of ecosystem service valuation, the Boyd and
Banzhaf (2007) definition implies that the total value from angling is the sum of value from all
of the inputs to angling. For example, the angler's value per day represents the value of angling
itself. The angler's value for fish landed during the day is some portion of the angler's total value
per day. Further, some portion of the value for fish landed represents the angler's value for the
ecosystem service offish population.
Finally, other portions of the angler's total value per day represent the value of benefits from
inputs not related to fish, some may be ecosystem services (scenic surroundings, fresh air, clean
water, wildlife, sunshine) and some may not (time with a friend, a cigar, a new fishing rod). In
other words, if the angler's total value per day is a pie, the value of fish landed is one slice of that
pie, and the value of the catchable target fish population is a portion of that slice.
WTP for an Angling Day
Traditionally, recreational angling has been valued by estimating an angler's willingness to pay
(WTP) for a day of fishing (Boyle et al., 1998, Loomis and Ng, 2009, Loomis and Richardson,
2007, Ng, 2011, Vaughan and Russell, 1982). As discussed above, WTP estimates for an angling
day include the value of additional ecosystem services - such as scenic view sheds, surface water
for boating, bird watching, wildlife viewing, and fresh air. Therefore, WTP estimates for an
angling day are included here only to provide context for the ecosystem service valuation of
catchable target fish population.
Loomis and Richardson (2007) conducted a meta-analysis of this extensive valuation literature
and provided WTP estimates for four species groups in six regions of the United States. The
estimates that were applicable to this analysis were for cold water species in the Intermountain
range - median net WTP per angling day of $51.27 and average net WTP per angling day of
$67.91. The majority of trips included in the meta-analysis that generated these estimates were
only for a single day.
WTP for Fish Caught
To value changes in the number of target fish caught, this project uses a benefit transfer tool
created by USEPA (2006a)2 in a "Regional Benefits Analysis for the Final Section 316(b) Phase
III Existing Facilities Rule June 2006." The overarching goal of USEPA (2006a) was to value
reductions in fish-kill from new regulations on power plant cooling water intake systems for
streams, rivers, and lakes in the United States. Foregone fishery yield was modeled via a model
that required estimates of species-specific size-at-age, stage-specific schedules of natural
2 The results of this research were also disseminated in the journal article Johnston et al. (2006). However,
for the sake of brevity, this research will be referred to by USEPA (2006a) only.

-------
mortality, and fishing mortality (USEPA, 2006a). Estimates of foregone fishery yield were
matched with species-specific estimates of anglers' willingness to pay to catch a fish. The meta-
analysis of recreational angling valuation literature that generated these value estimates is
described below.
USEPA (2006a) conducted an extensive literature review. All relevant studies in the published
economic literature, academic dissertations, and conference presentations were evaluated. Forty-
eight studies that provided marginal values of catching an additional fish were selected as the
study sample. Each study contained multiple values calculated based on various sample
characteristics and various model specifications - 391 WTP estimates in all. The 48 studies
varied in aspects such as study methodology (for example, stated vs. revealed preference),
elicitation method (for example, phone interview, survey, or in person interview), fish species,
study location, study date, baseline catch rate, and human sample characteristics. How species in
the primary valuation studies were aggregated for the meta-analysis is depicted in Table 7.1.
USEPA (2006a) econometrically estimated a regression on the 391 WTP estimates to estimate
the marginal value of catching an additional fish. This regression estimated the influence of
primary study variables such as baseline catch rate, species, angler income, and study
methodology. Once the influences of these variables are estimated, USEPA (2006a) used the
resulting meta-regression function for benefits transfer. This estimation involved setting the
function variables to correspond to the new site in question and predicting anglers' WTP to catch
an additional fish. For more detail, refer to USEPA (2006a) and Johnston et al. (2006). The
constant marginal value per fish results of USEPA (2006a), updated to 2013 dollars, is reported
in Table 7.2. These estimates represent mean value estimates. Confidence bounds to the
estimates shown in Table 7.2 are provided in Table 7.3.

-------
Table 7.1. Aggregate Species Groups
Aggregate
Group
Number of Observations
Species Included3
Big game
30
Billfish family, dogfish, rays, shark, skate, sturgeon,
swordfish, tarpon family, tuna, other big game
Small game
74
Barracuda, bluefish, bonito, cobia, dolly varden,
dolphinfish, jacks, mackerel, red drum, sea trout,
striped bass, weakfish, other small game
Flatfish
46
Halibut, sand dab, summer flounder, other flatfish
Other saltwater
89
Banded drum, black drum, chubby, cod family,
croaker, grouper, grunion, grunt, high-hat, kingfish,
lingcod, other drum, perch, porgy, rockfish, sablefish,
sand drum, sculpin, sea bass, smelt, snapper, spot,
spotted drum, star drum, white sea bass, wreck fish,
other bottom species, other coastal pelagics, "no
target" saltwater species
Salmon
44
Atlantic salmon, chinook salmon, coho salmon, other
salmon
Steelhead
16
Steelhead trout, rainbow trout (in Great Lakes only)b
Muskellunge
1
Muskellunge
Walleye/pike
12
Walleye, northern pike
Bass
14
Largemouth bass, smallmouth bass
Panfish
11
Catfish, carp, yellow perch, other panfish, "general
and no target" freshwater species
Trout
54
Brown trout, lake trout, rainbow trout, other trout
a Studies evaluated WTP for groups of species that did not fit cleanly into one of the aggregate species
groups established by EPA. In those cases, the group of species from the study was assigned to the
aggregate speces group with which they shared the most species.
b Rainbow trout in the Great Lakes were clasified as steelhead trout because they share similar physical
characteristics and life cycles with true anadromous steelhead. Although they have different common
names, rainbow trout and steelhead both belong to the species Oncorhyrychus mykiss. Source: USEPA,
2006a.

-------
Table 7.2. Marginal Recreational Value per Fish (by Region and Speciesa)
Species
California
North
Atlantic
Mid-
Atlantic
South
Atlantic
Gulf of
Mexico
Great
Lakes
Inland
Small game
$7.54
$6.17
$6.13
$5.94
$5.85

$5.56
Flatfish
$10.14
$6.19
$5.83




Other
saltwater
$3.07
$3.10
$3.03
$2.96
$2.89


Salmon





$13.78

Walleye/pike





$4.27
$4.25
Bass





$8.89
$9.36
Panfish


$1.10


$1.38
$1.10
Trout





$9.79
$2.94
Unidentified
$3.22
$3.12
$3.37
$2.97
$3.80
$6.46
$2.32
a All values are
in 2013 Dollars (2013$). Source: USEPA, 2006a.



Table 7.3. Confidence Bounds on Marginal Recreational Value per Fish3


Species
California
North
Atlantic
Mid-
Atlantic
South
Atlantic
Gulf of
Mexico
Great
Lakes
Inland
5% Lower
Confidence
Bounds'3
Small game
$4.12
$1.95
$2.06
$2.42
$2.53

$1.47
Flatfish
$4.85
$3.59
$3.45
$3.59



Other
saltwater
$1.62
$1.62
$1.65
$1.83
$1.80


Salmon





$10.38

Walleye/pike





$2.61
$2.28
Bass





$6.04
$5.49
Panfish


$0.59


$0.91
$0.59
Trout
$7.24 $1.50

-------
Species	California North Mid-	South Gulf of Great Inland
Atlantic Atlantic Atlantic Mexico Lakes
Unidentified $1.69 $1.63 $1.71 $1.84 $2.02	$4.43 $1.29
95% Lower
Confidence
Bounds'3
Small game
$13.76
$19.14
$17.94
$14.31
$13.31

$20.74
Flatfish
$20.89
$10.73
$9.95
$9.47



Other
saltwater
$5.86
$5.94
$5.60
$4.82
$4.65


Salmon





$18.29

Walleye/pike





$7.02
$8.03
Bass





$13.12
$15.98
Panfish


$2.01


$2.12
$2.01
Trout





$13.31
$4.46
Unidentified
$6.17
$5.99
$6.88
$4.87
$7.33
$9.47
$4.14
a All values are in 2013$.
b Upper and lower confidences bounds based results of the Krinsky and Robb (1986) approach. Source:
USEPA, 2006a.
This project uses the Table 7.2 inland value of trout at $2.94 for central Colorado, updated to 2013
dollars. To provide context for this figure, Table 7.4 shows the sub-sample of valuation studies from
USEPA (2006a) that apply directly to trout in Colorado.
Table 7.4. Trout Specific Values Corresponding to Eagle and Leadville 'Policy' Sites
Author and Year
State/Region
Study Methodology
Type of
Trout
Marginal Value per
Fish (June 2013$)
Boyle et al (1998)
U.S. Fish and Wildlife
Service Mountain
Trout Region
Contingent Valuation
General
$4.16
Johnson (1989)
Colorado
Contingent Valuation
Rainbow
$3.27
Johnson (1989)
Colorado
Contingent Valuation
General
$1.10
Johnson (1989)
Colorado
Contingent Valuation
General
$1.44
Johnson (1989)
Colorado
Contingent Valuation
General
$2.04
Johnson (1989)
Colorado
Contingent Valuation
General
$2.19
Johnson et al (1995)
Colorado
Contingent Valuation
General
$3.72
Johnson et al (1995)
Colorado
Contingent Valuation
General
$2.03

-------
Author and Year
State/Region
Study Methodology
Type of
Trout
Marginal Value per
Fish (June 2013$)
Johnson et al
1995)
Colorado
Contingent Valuation
General
$1.85
Johnson et al
1995)
Colorado
Contingent Valuation
General
$1.71
Johnson et al
1995)
Colorado
Contingent Valuation
General
$1.55
Johnson et al
1995)
Colorado
Contingent Valuation
General
$1.42
Johnson et al
1995)
Colorado
Contingent Valuation
General
$1.28
Johnson et al
1995)
Colorado
Contingent Valuation
General
$1.14
Johnson et al
1995)
Colorado
Contingent Valuation
General
$2.18
Johnson et al
1995)
Colorado
Contingent Valuation
General
$0.90
Johnson et al
1995)
Colorado
Contingent Valuation
General
$0.69
Johnson et al
1995)
Colorado
Contingent Valuation
General
$2.30
Johnson et al
1995)
Colorado
Contingent Valuation
General
$1.71
Johnson et al
1995)
Colorado
Contingent Valuation
General
$1.38
Johnson et al
1995)
Colorado
Contingent Valuation
General
$1.14
Johnson et al
1995)
Colorado
Contingent Valuation
General
$0.98
Johnson et al
1995)
Colorado
Contingent Valuation
General
$0.87
Johnson et al
1995)
Colorado
Contingent Valuation
General
$0.77
Johnson et al
1995)
Colorado
Contingent Valuation
General
$1.02
Vaughan
and
USA
Travel Cost
General
$1.44
Russell (1982)	
Overall Average	$ 1.70
Weighted Average	$1.08
by Study	
The criteria used to judge the correspondence between the primary valuation study and the
central Colorado sites were species (trout) and geographical region (Rocky Mountain Region).
Table 7.4 indicates that the value of an additional trout is within the range of $0.77 - $4.16 with
a study weighted average of $2.28. This range indicates that the value for trout of $2.94, from
USEPA (2006a) shown in Table 7.2, is well supported.
Johnson et al. (1995) provided a simplistic method to derive the value of the catchable target fish
population from the value of fish caught, which has been provided by USEPA (2006a). The logic
of this method is that the estimate of the percentage of target fish caught out of the catchable
target fish population can be used to determine how many catchable fish must be added for an
angler to catch another fish. Therefore, this method first estimates the ratio of the fish caught to
the catchable fish population and then multiplies the value of fish caught by this ratio.
For example, if the value of catching an inland trout is $2.94 (from Table 7.2) and one trout is
caught out of a catchable population of three trout, then the value of each trout in the target
population is $2.94 multiplied by one-third. It must be pointed out that many problems arise
from this simple solution. First, the percentage of fish caught from the catchable population now
has a large influence on the value of the catchable population. Second, there are obvious
dynamic effects between the number of fish caught and the number of catchable fish in the

-------
population (Ng, 2011). Whereas the modeling of such dynamic fisheries is outside the scope of
this analysis, this solution provides a mechanism for transferring the results of USEPA (2006a)
to the ecosystem service of catchable target fish populations.
WTP for Catchable Fish
Few studies have been dedicated to estimating the value of a catchable fish, even though such a
value would prove exceptionally useful for management decisions regarding fish stocking (Ng,
2011). Loomis and Ng (2009), as well as Ng (2011), directly addressed this problem through
their valuation survey of anglers at trout-stocked reservoirs in Colorado (Loomis and Ng, 2009).
The goal of Loomis and Ng (2009)3 was to inform fishery management decisions at the
reservoirs by estimating net WTP for angling trips, angling days, fish caught, and fish stocked
for trout and non-trout species. Loomis and Ng (2009) used survey data from 265 anglers to
estimate the value of angling trips via contingent valuation and travel cost methods.
The travel cost model was selected to estimate net WTP per angler trip. The dependent variable
was the number of annual trips to the site. The independent variables were the total cost of gas
per person, the hourly catch rate, a dummy for motorboat use, the number of people in the party,
the number of working household members, and the highest level of formal education. Mean net
WTP was the inverse of the estimated coefficient for total cost of gas per person. Finally, and
most importantly, the catch rate was modeled based on the number of catchable fish per acre,
skill, and a dummy for whether trout was the target species.
The endpoint problem surfaced for Loomis and Ng (2009) when the survey results indicated that
surveyed anglers had not traveled to their angling destinations just for fishing. They also
participated in other outdoor recreation activities such as camping, Whitewater rafting, hiking,
horseback riding, photography, and scenery viewing (Loomis and Ng, 2009, Ng, 2011). The
result was that survey respondents' stated value for an 'angling trip' included the value of
recreational benefits additional to angling.
Once net WTP per angler trip was estimated, the travel cost model and the catch rate model were
used to estimate the net WTP per angler day, net WTP per trout caught, and net WTP per trout
stocked (Loomis and Ng, 2009). Net WTP per angler day was calculated by dividing the net
WTP per angler trip by the average number of days per trip. Net WTP per trout caught and per
trout stocked were calculated by first doubling the average number of catchable (stocked) trout
per acre in the catch rate model, then estimating the impact on average catch rate, and finally
estimating the resulting increase in number of fishing trips4. Finally, the resulting increase in net
WTP was divided by the increased number of fish caught and by the increased number of fish
3	Ng (2011) was the PhD dissertation that resulted, in part, from the research of Loomis and Ng
(2009).
4	Notice that this approach directly mirrors the discussion above from Boyle et al. (1998) regarding
proper measurement of changes in net economic benefit (aka: net WTP).

-------
stocked. The resulting WTP per trout caught was $25.91, and the resulting WTP per catchable
trout was $0.60 in 2013 dollars.
Net WTP per angler day, trout caught, and trout stocked were all calculated from the base of the
inflated trip value. This result is clear when the Loomis and Ng (2009) net WTP per angler day
($173.66) is compared to a meta-analysis of net WTP per angler day values from the recreational
angling valuation literature ($67.91) (Loomis and Richardson, 2007). The same result emerges
when the net WTP per trout caught ($25.91) is compared to a meta-analysis of net WTP per trout
caught values from the recreational angling valuation literature ($2.94) (USEPA, 2006a).
The estimation of net WTP per stocked (catchable) trout is a novel accomplishment because it
values the ecosystem service itself. It also raises an important ecosystem service valuation
question. If the base value of the angling trip is inflated by the value of additional recreational
benefits, then does the increase in net economic value, which results from the increased number
of trips from doubling the catchable trout population, represent an ecosystem service value for
catchable trout? To explore this question, the net WTP per catchable trout from Loomis and Ng
(2009) will be compared with the net WTP per catchable trout derived from scaling the estimates
from USEPA (2006a) by the percentage of target fish caught.
Additional ecosystem services
The economic valuation literature has the capacity to model the value of many more ecosystem
services than were selected for this analysis. These ecosystem services were not modeled
because USEPA data concerning biological, geological, hydrological, and spatial characteristics
of intended study sites proved to be elusive. This deficiency prevented the modeling of changes
in additional ecosystem services and, instead, encouraged the focus on the one ecosystem service
with the most impact on total value (Button et al., 1999). Others have noted similar data
difficulties when attempting to value ecosystem services related to mine site pollution (Burton et
al., 1999, Button et al., 2012, Williamson et al., 2008). Nonetheless, Appendix B provides an
extended discussion of the potential to model the value of additional ecosystem services
impacted by mine site pollution.
Summary of Valuation Model
The benefit transfer model values catchable target fish population. Relevant value estimates are
summarized n Table 7.9. This benefit transfer model can be combined with a geo-environmental
model because it corresponds directly to the ecosystem services in question. By translating
changes in ecosystem services to monetary values of cost and benefit, decision makers can
compare services on a consistent basis when making decisions relating to ecosystem services. To
complete this process for the central Colorado sites, this project applies the benefit transfer
model to the outputs provided by the environmental model in Section 10.
Table 7.9. Summary Table of Values Used by the Project ($2013)

-------
Net Economic Benefit
Value
Units
WTP per angling day
$67.91	$ per angler per day
WTP to catch another trout
$2.91	$ per trout caught
WTP per catchable trout
$0.60	$ per catchable trout
'Hi'" ¦ - II" I	 11| iion
Physical setting
The study areas are located in the Rocky Mountains of central Colorado. The Leadville and
Gilman districts are on opposite sides of the continental divide. California Gulch drains the
Leadville area. California Gulch is a tributary of the Arkansas River near its headwaters. The
Arkansas River joins the Mississippi River approximately 2,400 km downstream from Leadville.
The Eagle River is on the western side of the continental divide from Leadville. It is a tributary
of the Colorado River.
The study area is in the ecoregion described by Bailey et al. (1994) as the Southern Rocky
Mountains Steppe - Open Woodland - Coniferous Forest - Alpine Meadow Province. Average
low temperatures for the area range from -16 °C in December and January to 3 °C in July and
August. Average high temperatures range from -1 °C in December and January to 22 °C in July.
The area receives an average of 29.4 cm of precipitation per year with 9 cm (31%) arriving in
July and August.
Regional geologic setting
The geologic history of central Colorado is summarized from Wallace (1993). The history began
more than 1.8 billion years ago in the Early Proterozoic with the accretion of volcanic arc and
back-arc complexes to the southern margin of the Archean Wyoming craton. These rocks were
deformed and then intruded by large Early and Middle Proterozoic batholiths. During Paleozoic
and Mesozoic time, the Proterozoic basement complex was buried beneath several kilometers of
marine and continental sediments, and it was partially exhumed during Pennsylvanian uplift.
Subduction-related calc-alkalic magmatism and uplift affected the region during the Late
Cretaceous-Early Tertiary Laramide orogeny. Post-subduction Oligocene and younger extension
generated the north-trending Rio Grande rift zone, which was accompanied by magmatic
activity. Most of the mineral deposits in the central Colorado mineral belt are associated with
Oligocene sub ducti on-related magmatism or later rift-related activity. Laramide-aged deposits
are relatively small, and a few carbonate-hosted deposits may be of Mississippian age.

-------
106°Q0'
N
t
\
\	39°00'
Explanation \
Cretaceous &	\
¦	Tertiary Intrusuns	\
\
¦	Cambrian to Missssippian \
¦	Sedimentary Rocks	\
0 10 Miles \
-i	\
I	r
0 10 Klometers
\
Ftgurej.6
Figure 8.1. Generalized Geologic Map of Central Colorado (modified from Wallace, 1993). Blue
areas are Cambrian through Mississippian sedimentary rocks. Tertiary intrusive rocks are shown
in pink. Unshaded areas are undifferentiated, but span that age range from Precambrian to
Quaternary.
The Leadville and Gilman districts lie on the eastern flank of the Sawatch Range. Rifting
exposed Paleozoic sedimentary rocks that overlie Proterozoic granites and were intruded by Late
Cretaceous and younger igneous rocks. Orogenic sediments were deposited in the graben during
uplift and erosion of the adjacent Sawatch Range and Mosquito Range to the east. Quaternary
glaciation further modified the landscape and locally redistributed the sediments in the district.
The oldest, volumetrically most important rocks exposed in the area are granites of the Middle
Proterozoic (approximately 1.4 Ga [billion years old]) Saint Kevin Granite (Tweto et al., 1978).
These rocks are overlain by shallow marine Paleozoic limestones, dolomites, sandstones, and
quartzites. The lower part of the stratigraphic section is composed predominantly of quartzite
with subordinate amounts of carbonate rocks and shale, ranging in age from Late Cambrian to
Late Devonian (Figures 8.1 and 8.2). Overlying these units are roughly 150 m of principally
carbonate rocks, including the Late Devonian Dyer Dolomite, Early Mississippian or Late
Devonian Gilman Sandstone, and the Early Mississippian Leadville Dolomite.

-------
Central Colorado was subjected to major intrusive events in Late Cretaceous-Early Tertiary time
(approximately 72-64 Ma [million years ago]) and again in the Middle Tertiary (43 to 39 Ma).
The intrusive activity produced sills, dikes, and small stocks of granodioritic to monzogranitic
composition (Bookstrom, 1990). Magmas invaded many faults, including shallow-dipping
Laramide thrust faults and high-angle younger faults, forming structurally controlled dikes. The
Pando Porphyry was emplaced at about 72 Ma. The Gray Porphyry includes igneous rocks
formed during several early to middle Tertiary intrusive events, units include the Lincoln Gulch
(66 Ma) and Evans Gulch (range in age from approximately 72 Ma to 30 Ma).
The more recent geology of the area was dominated by uplift and glaciation. After Middle
Tertiary uplift of the Sawatch and Mosquito Ranges and formation of the Arkansas River valley,
erosion of the ranges deposited sediments into the graben. At Leadville, erosion exposed many
orebodies, which consequently became oxidized during prolonged surface exposure. As
sedimentation in the graben continued during the late Tertiary, the orebodies, and probably much
of the area of the modern Leadville district, were progressively covered by sediments. These
poorly consolidated sediments are composed of sandy silt and interbedded sand and gravel
layers.
Stratified rocks of the Leadville district dip moderately to the east, forming a homocline that,
prior to Neogene rifting, once formed the eastern flank of the Sawatch uplift. This homocline is
cut by a complex network of faults, most of which dip steeply, but a few of which, as noted by
Thompson and Arehart (1990), are low-angle thrusts that presumably formed during the
Laramide orogeny. Trends of the principal faults in the district are approximately N15°E and
N20°W, consistent with the trends of major regional faults in the district (Tweto, 1960, 1968).
The mineralized rocks of the district owe their exposure in large part to formation of the Rio
Grande rift, a major intracontinental rift that extends northward from west Texas into at least
central Colorado.

-------
Minturn Formation
Belden Formation
Molas Formation
Leadville Dolomite
Castle Butte Member
Red Cliff Member
Chaffee Group
Gilman Sandstone
Dyer Dolomite
Parting Sandstone
Manitou Dolomite
Peerless Shale
Sawatch Quartzite
Saint Kevin Granite
Figure 8.2: Pre-Tertiary Stratigraphic Section of the Leadville District. The replacement ore
deposits of the district occur principally in the Leadville Dolomite and the Dyer Dolomite. Modified
from Thompson and Arehart (1990).
Economic geology and mining history
The discovery of the carbonate-hosted sulfide deposits of the central Colorado Mineral Belt
prompted much of the economic development of this part of the United States. Placer gold
deposits were discovered in the Leadville area in 1860, but these deposits were essentially
depleted by 1868 (Beaty et al., 1990). The carbonate replacement ores in the area were not
recognized until 1874. This discovery led to a prospecting rush focused on silver from 1877 to
1879 that resulted in the identification of similar deposits at Aspen, Gilman, Red Cliff, Tincup,
Kokomo, and Alma (Figure 8.1). The Black Cloud Mine in the Leadville district, the last
operating carbonate-hosted mine in the region, closed in 1999. Collectively, the Leadville and
Gilman districts produced over 1.5 million metric tons of zinc (Zn), 1.1 million metric tons of
lead (Pb), 100 thousand metric tons of copper (Cu), 10 million kg of silver (Ag), and 100
thousand kg of gold (Au), which was recovered from over 35 million metric tons of ore (Tables
8.1 and 8.2). This study will focus on the Leadville and Gilman districts.

-------
The deposits of the Leadville and Gilman districts are predominantly hosted by the Leadville
Dolomite and the Dyer Dolomite. The Manitou Dolomite also locally hosts ores at Leadville, and
the Sawatch Quartzite locally hosts ores at Gilman. The ores formed during a major mineralizing
event at about 39 Ma by wholesale replacement of the Paleozoic carbonate rocks by silver-, lead-
, zinc-, and gold-rich sulfide minerals. The carbonate rocks are also silicified adjacent to the
orebodies. Pyrite (iron sulfide), galena (lead sulfide), and sphalerite (zinc sulfide) are the most
common sulfide minerals in the replacement deposits of the Leadville district, with relatively
minor amounts of chalcopyrite (copper sulfide), tennantite-tetrahedrite (copper-arsenic-
antimony-sulfosalt), and magnetite (iron oxide). Silver principally occurred as argentite (silver
sulfide) with some argentiferous tetrahedrite, and gold is in its native form (Tweto, 1968).
Manganosiderite (manganese-iron carbonate) and quartz (mostly as fine-grained jasperoid) are
the principal nonsulfide gangue minerals (Beaty, 1990, Beaty et al., 1990, Thompson and
Arehart, 1990, Wallace, 1993). At both Leadville and Gilman, mineralization formed as mantos,
veins, and chimneys. Mantos are tabular replacement deposits typically confined by the
surrounding sedimentary stratigraphy. Chimneys are funnel-shaped bodies that may represent
feeder zones for the manto deposits. Igneous rocks are inferred to have been the source of both
the metals and sulfur for the Leadville-type deposits, and acid neutralization through reaction of
saline hydrothermal fluids with the carbonate hosts rocks is thought to have been the primary
depositional process for these ores (Beaty et al., 1990).
In the Leadville district, the orebodies are developed around the Breece Hill stock and cover an
area of approximately 6 km by 5 km (Thompson and Arehart, 1990). Typical sulfide ore grades
range from 3 to 8% lead, 6 to 30% zinc, 68 to 204 g/metric ton silver, and 1.7 to 7 g/metric ton
gold. The Zn:Pb ratios ranged from 1:1 to 4:1. Copper was present, but not in sufficient
quantities to warrant recovery. The main orebody at Gilman (Eagle Mine) supported most of the
mine production and covered an area of approximately 1 km by 2 km east of the Eagle River.
The ore processed at the mill from 1929 to 1977 averaged 2.0% lead, 11.6% zinc, 0.2% copper,
and 37.5 g/metric ton Ag (Beaty, 1990) with aZn:Pb ratio of 5.8:1.
Table 8.1. Metal Produced from Selected Carbonate-Hosted Districts in Central Colorado through
1987 (from Beaty et al., 1990).	
District
Zinc
Lead
Copper
Silver
Gold

metric tons
metric tons
metric tons
kilograms
kilograms
Gilman (Eagle)
866,000
147,000
96,000
2,116,000
12,400
Leadville
714,000
1,000,000
48,000
8,087,000
97,600
Kokomo
34,000
14,000
0
56,000
800
Aspen
10,000
267,000
minor
3,140,000
0
Sherman
0
3,000
0
228,000
0
Tincup
4,000
75,000
500
1,302,000
0

-------
Table 8.2. Cumulative Grade and Tonnage of Carbonate-Hosted Ore Produced in the Central
Colorado Mineral Belt from Selected Districts (from Beaty et al., 1990).	
District
Data Interval
Ore
Zinc
Lead
Copper
Silver
Gold


metric tons
%
%
%
mg/kg
mg/kg
Gilman (Total)
1880 - 1987
11,400,000
00
00
1.5
0.9
220
1.4
Gilman manto ores
1880 - 1987
8,600,000
11.6
2.0
0.2
38
0.7
Gilman chimney ores
1914- 1987
2,800,000
0.0
0.0
3.0
777
3.4
Kokomo
1905 - 1965
466,029
7.3
3.0
0.05
120
1.8
Leadville
1873 - 1987
23,800,000
3.0
4.2
0.2
320
3.7
Aspen
1880 - 1987
4,000,000
2.0
8.0
0.0
1,000
0.0
Sherman
1973 - 1984
645,005
4.0
0.8
0.1
485
0.0
Hydrologic setting
California Gulch
California Gulch is a 17-km long tributary of the Arkansas River that joins the Arkansas River
less than 32 km from its headwaters (Figure 8.3). The discharge from upper Arkansas River is
dominated by snowmelt typically peaking between early to mid-June (Figures 8.4 and 8.5). The
peak typically wanes throughout the summer, reaching base-flow conditions in early fall (Figure
8.5). The annual variation in discharge in the Arkansas River spans nearly two orders of
magnitude. The streambed is predominantly medium to large cobbles underlain by pebbles and
coarse sand (Clements et al., 2010).

-------
Figure 8.3. Map of Upper Arkansas River Showing Sample Sites. California Gulch, LMDT: Leadville
Mine Drainage Tunnel, and DT: Dinero Tunnel (modified from Clements et al., 2010).
SUSGS
USGS 07081200 ARKANSAS RIVER NEAR LEADVILLE, CO
2000.0 	
-o
c
I 1000.0	=j=| =|=|
1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
— Daily nean discharge	Period of approved data
Estinated daily nean discharge
Figure 8.4. Variation of Daily Discharge in the Arkansas River Upstream from California Gulch
from 1990 - 2010 (https://waterdata.usgs.gov/nwis).

-------
2USGS
USGS 07081200 ARKANSAS RIVER NEAR LEADVILLE, CO
— Daily nean discharge	Period of approved data
	 Estinated daily nean discharge
Figure 8.5. Detail of the Variation of Daily Discharge in the Arkansas River Upstream from
California Gulch from October 1, 2003 - October 1, 2005 (https://waterdata.usgs.gov/nwis).
Eagle River
The Eagle Mine is located on the northeastern bank of the Eagle River approximately 33 km
downstream from its headwaters (Figure 8.6). Similar to the Arkansas River, the discharge from
Eagle River is dominated by snowmelt typically peaking between early to mid-June (Figures 8.7
and 8.8). The annual variation in discharge in the Eagle River spans nearly two orders of
magnitude.

-------
Figure 8.6. Map Showing Sampling Sites along the Eagle River used in this Study from Woodling
et al. (2005). Potential sources of contamination exist from just north of Site 1 to just south of Site
4. The Redcliff Mine workings lie along Turkey Creek (not shown), which flows southwest to the
Eagle River, entering near Site 1. The Eagle Mine workings are located on the northeast side of the
Eagle River between Sites 1 and 2. Mine waste piles were located on both sides of the river
between Sites 2.9 and 4. Modified from Woodling et al. (2005).
ZUSGS
USGS 09064600 EAGLE RIVER NEAR MINTURN, CO
2000.0
> 10.0
s
° 5.0 	
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
	 Daily nean discharge	Period of approved data
	 Estinated daily nean discharge
Figure 8.7. Variation of Daily Discharge in the Eagle River Downstream from the Eagle Mine from
1990 - 2010 (https://waterdata.usgs.gov/nwis).

-------
USGS
USGS 09064600 EAGLE RIVER NEAR MINTURN, CO
800
"O
c
o
o
U
M
L
0
CL
m
m
o
¦H 100
3
O
0
M
L
A3
-C
o
w
¦H
10
Oct
2003
Jan
2004
Apr
2004
Jul
2004
Oct
2004
Jan
2005
fipr
2005
Jul
2005
Oct
2005
	 Daily nean discharge
— Estinated daily nean discharge
Period of approved data
Figure 8.8. Detail of the Variation of Daily Discharge in the Eagle River Downstream from the Eagle
Mine from October 1, 2003 - October 1, 2005 (https://waterdata.usgs.gov/nwis).
Water-quality variations
California Gulch
The long-term water quality of the Arkansas River is characterized by near neutral pH and
moderate specific conductance. Daily measurements of pH and specific conductance were taken
from 1990 to 1997 (Figures 8.9 and 8.10). The pH value generally ranged between 7.5 and 8.5
with limited excursions above and below this range and little or no seasonal variation. In
contrast, a distinct seasonal variation in specific conductance is evident with the lowest values
(greatest dilution) occurring in the late spring during snow melt and the highest values occurring
during base-flow conditions in the late summer through winter (Figure 8.10). The specific
conductance generally ranged between 50 and 300 |iS/cm.

-------
USGS 07081200 HRKHNSRS RIVER NERR LEADVILLE, CO
9*0 	
s 6*0
CL
5 5*5
-------
specific conductance shown in Figures 8.9 and 8.10. The alkalinity, hardness, and trace element
concentrations do not show any significant correlations with pH (Figures 8.11, 8.12, and 8.13).
Neither alkalinity nor hardness are shown to have any systematic variation upstream (sites EF5
and AR1) and downstream (sites AR3 and AR5) from California Gulch (Figures 8.11 and 8.12).
Significant differences in trace metals are evident upstream vs. downstream from California
Gulch. The downstream sites (AR3 and AR5) have higher dissolved zinc concentrations than the
upstream sites (EF5 and AR1, Figure 8.13). Dissolved cadmium and zinc have a general
correlation at higher concentrations (Figure 8.14), presumably reflecting a common source - the
mineral sphalerite in mine waste (Seal and Hammarstrom, 2003). Zinc is typically approximately
10 to 1,000 times more abundant than cadmium, on a mass basis, in the watershed. Compared to
copper, dissolved zinc is generally 1 to 1,000 times more abundant (Figure 8.15).
120
5.0
6.0
7.0
pH
8.0
9.0
Figure 8.11. Graph Showing pH and Alkalinity at Four Sites in the Arkansas River. Site locations
are shown in Figure 8.3. Data from Clements et al. (2010).
-------
250
200
O
u
(0
u
i-
0)
- 150

a.
U)
E
ro
*_
ao
O
*_
u
E
100
10
5.0
5.5
6.0
6.5
7.0
PH
7.5
8.0
8.5
9.0
Figure 8.13. Graph Showing pH and Dissolved Zinc Concentration at Four Sites in the Arkansas
River. Site locations are shown in Figure 8.3. Data from Clements et al. (2010).
-------
100
a 10

E
(D
3-
W)
O
1
£
£
I o.i
0.01
4 EF5
¦ AR1
A AR3
XAR5

vo,-
&

>
&
y x

x
4
-X^^X x/
X-X'w~xi
i
o-'
: >c
a.
to
E
aj
3-
bjo
O
i_
u
'£
i_r
a>
a.
a.
o
u
10
1
0.1
4 EF5
¦ AR1
i AR3
XAR5
X
i "
x>
X
> x x" X X ~
4*XXXI#
WZmXXKX XX ML ~
X
J? ~
~
~

=1	1—I—I—I—I—I—I—I—
=1	1—I—I—I—I—I—I—I—
=1	1	1—I—I—I—I—I—I—
=1	1	1—I—I—I—l-l-
10
100
Zinc, micrograms per liter
1000
10000
Figure 8.15. Graph Showing Dissolved Zinc and Dissolved Copper Concentrations at Four Sites in
the Arkansas River. Values that are below the analytical detection limit were plotted as the
detection limit, as is evident in many of the copper values falling at 0.5 pg/L. Site locations are
shown in Figure 8.3. Data from Clements et al. (2010).
-------
Eagle River
This river's long-term water quality is characterized by near neutral pH and moderate specific
conductance. Daily measurements of pH and specific conductance were taken periodically from
1989 to 2015 (Figures 8.16 and 8.17). The pH values generally ranged between 7.5 and 8.5,
which is similar to the range observed in the Arkansas River near Leadville. The specific
conductance range in the Eagle River was similar to that observed in the Arkansas River near
Leadville.
Figure 8.16. The Variation of Specific Conductance from 1989 - 2015 in the Eagle River
Downstream from the Eagle Mine (USGS Site ID 09064600). From the USGS National Water
Information System (http://nwis.waterdata.usgs.gov/usa/nwis/qwdata/).
-------
9
x
Q.
8
<#


» • •
»• • • •
N NN I
••• •
• •
• • •••
<£>
rCv

A
,0*
JF &	^ ^ ^ rf£}	^ ^ .ASJ rj>
^ ^ ^ ^ ^ ^ ^ ^ ^ ^
No\V ^ no\N ^ no\N ^ no\N no\N ^ no\N ^ no\N no\N ^
Date
Figure 8.17. The Variation of pH from 1989 - 2015 in the Eagle River Downstream from the Eagle
Mine (USGS Site ID 09064600). From the USGS National Water Information System
(htti)://nwis.waterdata.usgs.gov/usa/nwis/qwdata/).
Water-quality data, including pH, alkalinity, hardness, and trace element (Zn, Cd, Cu)
concentrations combined with biological data for the Eagle River in the vicinity of the Eagle
Mine are provided by Woodling et al. (2005) for the time period from 1990 to 2005. The
relationships among major and minor water-quality parameters are somewhat different from
those observed in the Arkansas River near Leadville. The alkalinity, hardness, and trace elements
concentrations show moderate correlations with pH (Figures 8.18, 8.19, and 8.20). Neither
alkalinity nor hardness indicates any systematic variation upstream (Sites 1-2) and downstream
(Sites 2.9 - 6) from the Eagle Mine, although the upstream and farthest downstream sites tend to
show the highest pH, hardness, and alkalinity values and the sites from within the mine zone tend
to have the lowest pH and highest dissolved zinc values (Figures 8.18 and 8.19). In contrast, the
downstream sites (except for Site 6 where zinc concentrations are lower than Site 3- because of
the diluting effect and higher pH of Gore Creek) have higher dissolved zinc concentrations than
the upstream sites (Figure 8.20). Dissolved cadmium and zinc also have a general correlation
(Figure 8.21). Zinc is typically approximately 100 to 1,000 times more abundant than cadmium
on a mass basis in the watershed. A significant correlation is not apparent for zinc and copper
(Figure 8.22).
-------
140
5.0
6.0
7.0
PH
8.0
9.0
Figure 8.18. Graph Showing pH and Alkalinity at Eight Sites in the Eagle River. Site locations are
shown in Figure 8.6. Data from Woodling et al. (2005). Blue symbols are sites upstream from
significant mining activity, red symbols represent sites from the reach of mining impacts on the
land surface, and yellow symbols represent sites downstream from surface disturbance related to
mining.
250
5.0
6.0
7.0
PH
8.0
9.0
Figure 8.19. Graph Showing pH and Hardness at Eight Sites in the Eagle River. Site locations are
shown in Figure 8.6. Data are from Woodling et al. (2005). Blue symbols are sites upstream from
significant mining activity, red symbols represent sites from the reach of mining impacts on the
land surface, and yellow symbols represent sites downstream from surface disturbance related to
mining.
-------
10000
-------
100
a>
a.
U)
E
ro
*_
m
O
i_
u
£
E
T3
TO
U
10	100	1000
Zinc, micrograms per liter
10000
Figure 8.21. Graph Showing Dissolved Zinc and Dissolved Cadmium Concentrations at Eight Sites
in the Eagle River. Site locations are shown in Figure 8.6. Data are from Woodling et al. (2005).
Blue symbols are sites upstream from significant mining activity, red symbols represent sites
from the reach of mining impacts on the land surface, and yellow symbols represent sites
downstream from surface disturbance related to mining.
1000
a>
~ 100

a.
a.
o
u
1
0.1
•	Site 1
¦	Site 1.9
¦	Site 2
A Site 2.9
•	Site 3
¦	Site 4
Site 5
Site 6

10	100	1000
Zinc, micrograms per liter
10000
Figure 8.22. Graph Showing Dissolved Zinc and Dissolved Copper Concentrations at Eight Sites in the Eagle River.
Site locations are shown in Figure 8.6. Data are from Woodling et al. (2005). Blue symbols are sites upstream from
significant mining activity, red symbols represent sites from the reach of mining impacts on the land surface, and
yellow symbols represent sites downstream from surface disturbance related to mining.
-------
Mining-environmental landscape
Leadville District (California Gulch)
The mining-environmental landscape of the Leadville district is adequately described by the
Operable Units defined by EPA for the site. The California Gulch (Leadville) Superfund Site
includes 47 km2 (18 mi2) of affected land. EPA has divided the site into 12 operable units that
include mine drainage, surface water, groundwater quality, sediment, waste rock, mill tailings (in
situ and fluvially transported), smelter sites including pyrometallurgical slag, and residential
soils (Table 8.3). California Gulch includes many operable units where remediation activities
have been completed and the operable units have been "deleted" from the National Priorities
List.
Table 8.3. Summary of operable units (OUs) at the California Gulch (Leadville) Superfund site.
OU Name
Description
Status
1	Yak Tunnel
2	Malta Gulch
3	Denver and Rio Grande Railroad Slag Piles,
Railroad Easement, Railroad Yard, and Mineral
Belt Trail
4	Upper California Gulch
5	ASARCO Smelter/Colorado Zinc-Lead Site
6	Stray Horse Gulch
7	Apache Tailings
8	Lower California Gulch
9	Populated residential areas
10	Oregon Gulch
11	Arkansas River floodplain
12	Site-wide surface water and groundwater quality
Mine drainage
Mill tailings
Slag
Surface water
Smelter and mill site,
smelter waste, waste
rock, mill tailings
Waste rock, mill
tailings, surface water,
groundwater
Mill tailings
Mill tailings, soils,
waste rock, stream
sediments
Soils
Mill tailings
Fluvial tailings
Surface water,
groundwater
Ongoing treatment
Deleted 7/23/2001
Operation and
Maintenance
Deleted 10/24/2014
Deleted 10/24/2014
Remedial design
Deleted 10/24/2014
Deleted 1/12/2010
Deleted 9/21/2011
Deleted 4/16/2001
Field work completed
Remedial design
Remediation of California Gulch and adjacent areas has been implemented in stages. The
Leadville Mine Drainage Tunnel water treatment plant, operated by the U.S. Bureau of
Reclamation, was completed in 1992. The Yak Tunnel water treatment plant, operated by
ASARCO, was completed in 1992. Tailings remediation in California Gulch and restoration of
riparian areas was completed in 1999 (Clements et al., 2010).
-------
Gilman District (Eagle Mine)
The mining-environmental landscape of the Gilman district is adequately described by the
Operable Units defined by EPA for the site. The Eagle Mine Superfund site is much smaller than
the California Gulch Superfund site. The Eagle Mine Superfund site only covers 95 hectares
(235 acres). EPA has divided the site into three operable units that include mine drainage,
surface water, groundwater quality, soils, waste rock, mill tailings, and roaster piles (Table 8.4).
Table 8.4. Summary of Operable Units (OUs) at the Eagle Mine (Gilman) Superfund Site
ou
Name
Description
Status
1
Eagle Mine, Roaster Pile, Waste Rock Piles,
Rex Flats, Old Tailings Pile, Consolidate
Tailings Pile
Mine drainage, surface
water, groundwater,
waste rock, mill tailings
Ongoing treatment, site
remediation completed by
2001
2
Gilman
Soils

3
North Property
Mill tailings, roaster piles
Remedial investigation
Remediation of the Eagle Mine and adjacent areas has been implemented in stages. A permanent
water treatment plant was constructed in 1990. The water treatment plant treats water collected
from the mine, groundwater beneath the tailings pile, and contaminated surface and groundwater
collected from multiple locations across the site. Most solid mine waste remediation was
completed by 2001, but additional relocation of roaster waste near Gilman occurred in 2006.
Einviironmental Settiirig
Water
Influences on water quality
Mining influenced water (MIW) is water that has had its chemical composition affected by
mining or mineral processing activities (Wildeman and Schmiermund, 2004). The downstream
effects of MIW on aquatic organisms throughout the remediation histories of the sites are best
considered in terms of the exceedance of water quality guidelines for metals of concern (Zn, Cd,
and Cu) in these watersheds. The toxicity of Zn, Cd, and Cu to aquatic organisms can be
modeled as a function of water hardness (Figures 7.12 and 7.19, USEPA, 2006b). Thus, the
inferred toxicity of dissolved metals (Zn, Cd, and Cu) in both rivers can be modeled using the
hardness-dependent aquatic life criteria in conjunction with site-specific water quality data from
the Arkansas River and the Eagle River. The Arkansas River water-quality data are from
Clements et al. (2010) and the Eagle River data are from Woodling et al. (2005). The
concentration of a metal for acute toxicity is expressed in terms of a criteria maximum
concentration (CMC), which can be compared to metal concentrations at sample sites. Dissolved
metal concentrations can be compared, or normalized, to their respective CMCs through a hazard
quotient (HQ), as described in Equation 9.1 below:
-------
HQ = ml c (9.1)
where m is the measured concentration of the metal and c is the hardness-adjusted CMC for that
metal in that specific sample. The simultaneous toxicity of multiple metals, such as Zn, Cd, and
Cu, can be evaluating by summing hazard quotients for individual metals in a sample as a hazard
index (HI), also known as a cumulative criterion unit (CCU), expressed as shown in Equation 9.2
below:
HI = CCU = hm/d (9.2)
where m, is the measured concentration of the 7th metal and c, is the hardness-adjusted CMC for
the 7th metal. An HQ or HI less than 1 implies metal concentrations that should not be toxic to
aquatic organisms. An HQ or HI above 1 implies toxic conditions.
In the Arkansas River, the cumulative (Zn, Cd, and Cu) hazard indices for the upstream reference
sites (EF5, AR1) averaged 0.95 and 0.78, respectively, over the course of the study (1991 -
2006), and those for the downstream sites (AR3, AR5) averaged 4.76 and 1.68, respectively. In
the Eagle River, the hazard indices averaged below 1 for the upstream reference site (Site 1), and
the sites from the reach with mine disturbance and downstream from mine disturbance ranged
from 0.99 (in the most downstream site) to 10.35 over the course of the study (1990 - 2005). In
other words, dissolved concentrations of zinc, cadmium, and copper locally exceed the aquatic
life guidelines.
The predicted relative contributions of zinc, cadmium, and copper to the aquatic toxicity of MIW
in the Arkansas River are Zn 68%, Cd 21%, and Cu 11% (Figure 9.1), in the Eagle River, they
are Zn 80%, Cd 9%, and Cu 11% (Figure 9.2). Thus, in both watersheds, zinc is the dominant
aquatic stressor and will be the focus of the following discussion.
Arkansas River
amln
¦	Cd
¦	Cu
Figure 9.1. Pie Diagram Showing the Relative Contributions of Zinc, Cadmium, and Copper to the
Predicted Toxicity (Expressed as Hazard Index) of Water in the Arkansas River Downstream of
California Gulch.
-------
Eagle River
Figure 9.2. Pie Diagram Showing the Relative Contributions of Zinc, Cadmium, and Copper to the
Predicted Toxicity (Expressed as Hazard Index) of Water in the Eagle River Downstream of the
Eagle Mine.
The variations among sites with time in terms of predicted toxicity for the combination of Zn,
Cd, and Cu compared to Zn alone, the predominant contaminant of concern, are very similar
(Figures 9.3 and 9.4). The similarities include upstream and downstream sites, influxes of metals,
and responses to remediation. Thus, it can be concluded that zinc is a reasonable proxy for
predicted metal toxicity in the Eagle River watershed. Similar conclusions can be reached for the
Arkansas River near Leadville (Figure 9.5).
For the Eagle River, there has been a general, but erratic, decrease in predicted toxicity of Zn,
Cd, and Cu over the course of the remediation project, although none of the sites near the sources
are below predicted toxicity limits as of 2005, the last year of data in the report (Figures 8.3 and
8.4, data from Woodling et al. (2005)). In contrast for the Arkansas River, all of the sites have
achieved, with time, zinc concentrations below that predicted acute toxicity limit (Figure 9.5).
-------
a>
8.00
7.00
6.00
5.00
~P 4.00
3.00
2.00
/ 1
' *\\ 1
AW ¦
/ \\\ * ¦
/ \\\ A ¦
/ U / u1

-•-Site 1

-¦-Site 1.9

A Site 2

-¦-Site 2.9

~ Site 3

9 Site 4
Site 5

V M\\ / l\>L
\ \\ // A\j >
Site 6




csN
Date
Figure 9.3. Variations of the Cumulative Acute Hazard Index (Zn + Cd + Cu) with Time for all Study
Sites. Note: A hazard index or quotient below 1 is considered "not toxic" (shown as the red
horizontal dashed line). The Eagle Mine Superfund site transitioned from active remediation to
operation and maintenance (O&M) in 2001 (shown as the vertical dashed green line). The site was
placed on the National Priorities List in 1986. Upstream sites are shown in blue, potential source
region sites are shown in red, and sites downstream from all known sources are shown in yellow.
Data from Woodling et al. (2005).
-------
10.00
Date
Figure 9.4. Variations of the Acute Hazard Quotient for Zn with Time for all Study Sites. Note: A
hazard quotient below 1 is considered "not toxic" (shown as the red horizontal dashed line). The
Eagle Mine Superfund Site transitioned from active remediation to operation and maintenance
(O&M) in 2001 (shown as the vertical dashed green line). The site was placed on the National
Priorities List in 1986. Upstream sites are shown in blue, potential source region sites are shown
in red, and sites downstream from all known sources are shown in yellow. Data from Woodling et
al. (2005).
-------
30
Figure 9.5. Variations of the Acute Hazard Quotient for Zn with Time for all Study Sites in the
Arkansas River (California Gulch Superfund Site). Note: A hazard quotient below 1 is considered
"not toxic" (shown as the red horizontal dashed line). Significant remediation events are shown
as vertical dashed lines. These events include the start of the Leadville Mine Drainage Tunnel
treatment plant (LMDT), the start of the Yak Tunnel treatment plant, and remediation of the
California Gulch (CG) tailings.
The downstream reductions in Zn concentrations, as reflected by the lower hazard quotients, may
be due to several processes, the most important of which is dilution due to the influx of
groundwater or surface-water tributaries to the Eagle and Arkansas rivers. This process is
particularly evident at Site 6 in the Eagle River that is downstream from the confluence with
Gore Creek. Removal by sorption on to hydrated ferric oxides or clay minerals is a less likely
possibility because the sorption edge for zinc is at a fairly high pH (Section 10) and
neutralization of MIW by receiving water bodies should have effectively precipitated most of the
hydrated ferric oxides at the site of mixing, and thus would have removed downstream sources of
trace metal sorbents (Section 10).
To verify that zinc is behaving conservatively in the watershed below all sources, the dissolved
concentrations that would be expected simply from dilution using the discharge data found in
Woodling et al. (2005) were calculated. Because potential sources of contaminants exist from
Site 1 to Site 4 (Figure 8.6), these calculations were made starting with Site 3 and again starting
-------
with Site 4. The comparison of the estimated concentrations with the measured concentrations is
shown in Figure 9.6.
Both sets of calculations yield strong correlations between measured and estimated
concentrations. Linear regression of both simulations yields r2 values greater than 0.86 (Figure
9.6). This strong correlation confirms that zinc behaves conservatively downstream from
sources, that is sorption and precipitation are not important in-stream processes affecting zinc
concentrations. This type of approach is not possible upstream from Sites 4 or 3 because of the
potential for multiple sources of contamination from mine waste piles and mine workings,
although this approach may have value in identifying reaches that still have sources of zinc.
Given that zinc behaves conservatively in the Arkansas and Eagle rivers, its toxicity should be
linked to the amount of dilution that has occurred at any point along the river downstream from
the primary sources of zinc.
2500
T3
CL)
£ 2000
TO
E
+->
to
cu
^ 1500
cu
cu
Q- 1000
to
E
TO
i—
ClD
o 500
u
'E
u
.E 0
M	0	500 1000 1500 2000 2500
Zinc, micrograms per liter, measured
Figure 9.6. Scatter Plot Comparing Measured Concentrations of Zinc in the Eagle River with
Concentrations Estimated Assuming that Dilution is the Only Process Causing Decreases in
Concentration. The dashed line represents perfect agreement with measured and estimated
values.
4 Relative to Site 3
Relative to Site 4
/J/
0
\
~
i *
The variation of fish populations in the Eagle River by site with time is shown in Figure 9.7.
Most sites, including the most upstream site, Site 1, exhibited depressed fish populations for the
early part of the Woodling et al. (2005) study. After remediation was completed in 2001, most
-------
sites experienced a steady increase in fish populations. The relationship between the fish
population and the hazard quotient for zinc for the Eagle River is shown in Figure 9.8. In
general, the lower hazard quotient corresponds to the higher fish populations. However, the
greatest increase in fish population is at a hazard quotient below 3 or 4 rather than the theoretical
value of 1.
Date
Figure 9.7. Variations of the Abundance of Fish (Predominantly Brown Trout) with Time for all
Study Sites in the Eagle River. The Eagle Mine Superfund Site transitioned from active
remediation to O&M in 2001 (shown as the vertical dashed green line). The site was placed on the
National Priorities List in 1986. Upstream sites are shown in blue, potential source region sites are
shown in red, and sites downstream from all known sources are shown in yellow. Data are from
Woodling et al. (2005).
-------
1500
dj 1000
i_
u
03
-------
Date
Figure 9.9. Variations of the Abundance of Brown Trout with Time for all Study sites in the
Arkansas River. Significant remediation events are shown as vertical dashed lines. These events
include the start of the Leadville Mine Drainage Tunnel treatment plant (LMDT), the start of the Yak
Tunnel treatment plant, and remediation of the California Gulch (CG) tailings. Data are from
Clements et al. (2010).
-------
2500
cu
i—
TO
+->
U
(L>
(L)
Q.
2000 2)
m
1500 ^
~
1000
500 -
O
w&
~ EF5 Spring
AEF5 Fall
~	AR1 Spring
OAR1 Fall
•	AR3 Spring
OAR3 Fall
¦ AR5 Spring
~ AR5 Fall
O
-i	1	1	1	1	1	1	1	1	1	r
"I	1	1	1	1	1	r
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Hazard quotient
Figure 9.10. Scatter Plot Showing the Relationship Between Brown Trout Populations and the
Acute Hazard Quotient for Dissolved Zn in the Arkansas River. Both Spring and Fall fish surveys
are shown. Data are from Clements et al. (2010).
Aquatic organisms
Arkansas River/California Gulch
Macroinvertebrate and brown trout (Salmo Irulla) data from 1989 to 2006 for four sites in the
Arkansas River in the vicinity of Leadville are described and presented by Clements et al.
(2010). Macroinvertebrate data include total abundance per area and number of mayfly
species per area and were sampled in both spring and fall each year. Brown trout data include
abundance and biomass per area and were sampled in August. The macroinvertebrates
demonstrated steady recovery after restoration events reaching levels above reference
streams (Figure 9.11). The brown trout data also indicate steady recovery downstream after
restoration events (Figure 9.12). The reader is referred to Clements et al. (2010) for more
detailed discussion of their biologic data.
-------
1400
rsl
E
1200
T—I
1000
o

L	
QJ
800
Q.

!—
600
OI

-Q
E
400
=3
200
0
4000
EF5
T

>
I- Ik
k
—<1
It—1

i yJ
—J\L
w


3000-
o
CD
2000-
O)
E 1000
3
~l	1	1	1	1	T"
~~1	1	1	1	1	T"
3000
AR3
jf
T
11
^ i M i i i i i i i i i i
Op cO' ^ oP Op (§*
& ^ # & & rf rf	r$>
Year
n—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r
oy # dp s> c?	<^°
$ & & S ^ ^ ^ ^ <#
Year
Figure 9.11. Variations in Total Macroinvertebrate Abundance in the Arkansas River from Sites
Upstream (EF5, AR1) and Downstream (AR3, AR5) from California Gulch. Arrows indicate the
completion of major restoration projects, and the horizontal lines bracket the mean value for
reference streams with good water quality in Colorado. Restoration events upstream are shown
by solid arrows, and downstream events are dashed. Upstream events include the start of
operation of the Leadviile Mine Tunnel and Yak Tunnel treatment plans, and downstream events
are tailings removals. Modified from Clements et al. (2010).
-------


Figure 9.12. Variations in Brown Trout Abundance and Biomass in the Arkansas River from Sites
Upstream (EF5, AR1) and Downstream (AR3, AR5) from California Gulch. Arrows indicate the
completion of major restoration projects. Restoration events upstream are shown by solid arrows,
and downstream events are dashed. Upstream events include the start of operation of the
Leadville Mine Tunnel and Yak Tunnel treatment plans, and downstream events are tailings
removals. Modified from Clements et al. (2010).
Eagle River
Macroinvertebrate and fish data from 1990 to 2005 for six sites in the Eagle River in the vicinity
of the Eagle Mine are described and presented by Woodling et al. (2005). Macroinvertebrate data
include the abundance and number of taxa. Fish data include abundance and size.
Macroinvertebrates at Sites 3 and 4 showed the most severe and persistent impacts from MIW,
whereas Sites 5 and 6, downstream from the area of surface disturbance, exhibited steady
improvement after the start of water treatment in 1991 (Figures 9.13 and 9.14). The brown trout
data have similar trends (Figure 9.15). The reader is referred to Woodling et al. (2005) for more
detailed discussion of their biologic data.
-------
Site 1
Site 1.9
Site 2
Year
Site 2.9

Year
Site 4

Year
Site 6

Year
Year
Figure 9.13. Variations in Macroinvertebrate Abundance in the Eagle River. Significant influx of
MIW occurred downstream of Site 2, and surface disturbance related to the mine was noted
upstream from Site 5. Modified from Woodling et al. (2005).
-------
Site 1
Site 1.9
Site 2
cy5 c^ c>P op o?3 d?> c?> Cs* CS-> C?1 cv>
vvvvvvvv^^'v^nr'v^'v3
Year
vvvvvvvv^^^'y3nr'v5
Year
Year
Figure 9.14. Variations in Number of Macroinvertebrate Taxa in the Eagle River. Significant influx
of MIW occurred downstream from Site 2, and surface disturbance related to the mine was noted
upstream from Site 5. Modified from Woodling et al. (2005).
-------
Site 1
Site 1.9
jfr _0^	_0?° ^	J& .CV' _cQ- .£?> _£¦?* J^5
\ear
Site 2.9
OJ
u
<
CLI
CL 1000 ¦
D
O
Year
Site 5
Year
Site 2

V V V V V V V V V V V V V V V v
Year
Site 4
VV)'v,VVVVV,VV,TJ'tJTx'r'r'>-
Year
Year
Figure 9.15. Variations in Brown Trout Abundance in the Eagle River in the Vicinity of the Eagle
Mine with 95% Confidence Intervals. Significant influx of MIW occurred downstream from Site 2,
and surface disturbance related to the mine was noted upstream from Site 5. Modified from
Woodling et al. (2005).
Discussion
Geo-environmental setting
Geo-environmental models have value for understanding environmental risks shared by mineral
deposits of similar types, such as the carbonate replacement sulfide deposits of the Leadville and
Gilman districts in central Colorado. The shared features include the nature of the ores, the
nature of the host rocks, the chemical behavior of the elements found in the ores and solid mines
-------
wastes, the mining methods, the ore-processing methods, and the waste management practices.
The environmental characteristics of the abandoned mines in the Leadville and Gilman districts
are influenced by the regional geologic characteristics of the watershed, especially the carbonate
rocks. The Manitou Dolomite, the Leadville Dolomite, and the Dyer Dolomite are important rock
units within the watersheds of both mining districts. The latter two represent the most important
ore hosts in the area. The effect of these dolomite units is to support near neutral pH, high
alkalinity, and high hardness in the watersheds that historically have been affected by MIW. It is
used in the following discussion because it is more general than the term "acid mine drainage"
and avoids confusion surrounding acid drainage that becomes neutralized yet still carries
potentially toxic concentrations of metals.
The primary ores are massive sulfide accumulations, dominated by pyrite, iron-rich sphalerite,
and galena with lesser amounts of chalcopyrite, important accessory minerals include argentite,
electrum, and tetrahedrite-tennantite (Thompson and Arehart, 1990). The pyrite, which ended up
in waste rock and mill tailings, embodies the considerable bulk of the acid-generating potential
of the mine waste. Iron from remnant sphalerite in the mine waste also contributes to its acid-
generating potential. Ultimately, the acid-generating potential of the mine waste depends on the
balance between the acid-generating potential from the pyrite and the acid-neutralizing potential
from the carbonate minerals derived from the host rock. Acid generated from pyrite oxidation
enhances the liberation of metals, such as iron, zinc, cadmium, copper, and lead, from mine
wastes. The carbonate host rocks represent acid-neutralizing potential for their solid mine wastes
(unprocessed waste rock or mill tailings). However, the ubiquitous silicification associated with
mineralization served to dilute the acid-neutralizing potential of the mined rock and processed
mill tailings.
The approach used to develop a mine also influences its environmental attributes. The mining
methods, ore-processing methods, and waste management practices utilized significantly
influenced environmental risks associated with mining in the Leadville and Gilman districts. The
dipping tabular nature of the ore bodies in both districts was conducive to underground mining
for most deposits. Underground mining minimizes the volume of waste rock generated, making
either hand-sorted, low grade waste for historical mining or mill tailings for more recent mining
the predominant solid mine waste materials. The differences in physical properties of coarse-
grained hand sorted waste and fine-grained mill tailings make their environmental behavior
distinct from one another. Water that has interacted with mine workings, waste rock, or mill
tailings (MIW) can become acidic due to the oxidative weathering of pyrite and can leach
significant quantities of metals, other trace elements, and sulfate. The coarse-grain size of hand
sorted waste (or waste rock) results in a fairly oxygenated, unsaturated waste pile - an ideal
environment for sulfide oxidation (Amos et al., 2015). In contrast, the fine-grain size of mill
tailings can limit the access of oxygen, and the retaining structures of tailings storage facilities
can maintain a saturated condition within much of the pile, which additionally can limit the
access of oxygen (Lindsay et al., 2015). The fine-grained nature of mill tailings also makes them
more prone to erosion, transport, and redeposition at low-gradient sites downstream where they
can act as additional sources of contamination. Historical mining operations typically employed
-------
waste management practices that ignored environmental mitigation strategies, whereas mining
operations permitted after 1970 were required to have environmental mitigation incorporated
into mine plans. The only mine in the study to fall in the latter category is the Black Cloud Mine
(1971 - 1999) in the Leadville district.
An additional, somewhat unique, feature of mining in the Leadville district was the construction
of tunnels that consolidated access to numerous small mines on the mountainside. The tunnels
facilitated haulage of ore and also served to drain water from the mines. Today, the tunnels are
major sources of acid mine drainage. The Leadville Mine Drainage Tunnel extends
southeastward from the Arkansas River valley north of the town of Leadville. It surfaces in the
watershed of the East Fork of the Arkansas River. The Yak Tunnel extends from California
Gulch northeastward beneath Breece Hill and surfaces in the California Gulch watershed. Both
are sites of active water treatment systems today.
In historical mining districts where mining has occurred over the span of a century, such as
Leadville and Gilman, mining, ore processing, and waste management practices typically evolve
over time due to technological advancements and increased regulation. The nature of
environmental risks will vary accordingly because this evolution in practices will affect the
nature of mine waste. The geo-environmental landscape of a historical mining district will,
therefore, be an agglomeration of these effects.
The regional geologic setting influences the watershed scale chemistry of surface water receiving
MIW. As mentioned, carbonate rocks, such as those found in central Colorado, serve to elevate
the pH, alkalinity, and hardness of rivers and streams. Neutralization of metal-rich acid drainage
due to mixing with larger, higher alkalinity streams under oxygenated conditions should result in
the precipitation of most of the dissolved iron, but should leave appreciable amounts of zinc and
cadmium in solution (Figure 10.1). In contrast, lead has limited solubility in sulfate and
carbonate rich waters due to the low solubility of anglesite and cerussite, respectively (Figures
10.2 and 10.3). Therefore, lead is considered to be of limited concern in the aqueous phase, but
does pose significant risks to humans in particulate form. The primary particulate forms of
concern are waste rock, and particularly mill tailings. The fine-grained nature of mill tailings
makes them more amenable to transport away from the site of initial disposal by wind or water
erosion. Trace metals may also be attenuated in aqueous settings by sorption onto hydrous ferric
oxides (Figure 10.4). Lead and copper are effectively removed at low pH (< 5.5), but significant
amounts of zinc and cadmium may persist to higher pH values. Therefore, neutralization of acid
drainage may effectively remove dissolved lead and copper in addition to dissolved iron, but zinc
and cadmium may persistent downstream.
-------
pH
Figure 10.1. Solubility of Various Metals as a Function of pH. Note low solubility of ferrihydrite
(Fe[OH3]) compared to the hydroxides of zinc, copper, and cadmium.. Modified from Nordstrom
and Alpers (1999).
Temperature (°C)
Figure 10.2. Logarithm of the Solubility Product (Ksp) of Selected Sulfate Minerals as a Function of
Temperature. Note the low solubility of anglesite (PbS04>. Modified from Rimstidt (1997).
-------
Temperature (°C)
Figure 10.3. Logarithm of the Solubility Product (Ksp) of Selected Carbonate Minerals as a
Function of Temperature. Note the low solubility of cerussite (PbCC>3). Modified from Rimstidt
(1997).
pH
Figure 10.4. Model Sorption Curves for Selected Metals and Sulfate on Hydrous Ferric Oxide.
Modified from Smith (1999).
The fate of the dissolved metals and other trace elements depends on the chemical characteristics
of the MIW and the receiving water body. The neutralization that commonly accompanies
mixing of acidic water with higher alkalinity receiving waters can partially or totally remove
some metals from solution. The behavior of metals during neutralization varies on an element by
element basis and is also influenced by other elements, especially iron, in the MIW. Hydrated
ferric oxides that can precipitate as a result of neutralization can act as strong sorbents of trace
metals such as Cu, Pb, Zn, and Cd. The sorption of these metals varies as a function of pH with
the onset of removal with increasing pH starting in the order of Pb, Cu, Zn, and Cd.
-------
Aquatic Setting
History of aquatic ecological recovery
The toxicity of metals in MIW to aquatic organisms depends on a number of factors.
Precipitation and sorption during the mixing of MIW with a receiving water body can reduce
metal concentrations in water. Dilution both due to mixing with receiving water bodies and the
influx of clean tributaries downstream serves to decrease metal concentrations. Decreases in
metal concentrations due to precipitation, sorption, and dilution all serve to reduce toxicity.
Toxicity to aquatic organisms also varies as a function of water chemistry. Water hardness is
important in mitigating the toxic effects of metals.
The effect of metals in MIW on fish populations is complex. Not only can the dissolved metals
directly affect the health of fish, but they can also affect aquatic macroinvertebrates that serve as
important food sources for fish. The precipitation of hydrated ferric oxides at the mixing zone of
MIW with receiving streams can cement cobbles and gravels, resulting in a degraded habitat for
aquatic macroinvertebrates. Surface-water impacts in the vicinity of the Leadville and Gilman
mining districts have been extensive. One of the most extensive early reports was written by
LaBounty et al. (1975). Their study included investigations of surface water chemistry, sediment
chemistry, macroinvertebrates, and fish collected from April through November 1974 in the
upper Arkansas River. They documented extensive impairment of the aquatic ecosystem due to
MIW.
The recovery of a watershed during and after remediation can be equally complex. Sources of
contamination must be identified and properly addressed. Common sources of MIW include
drainage from mine workings and leachate from solid waste piles. Drainage from mine workings
represents a particularly intractable challenge because source control is often difficult or
impossible. Instead, treatment, either passively or actively, is required. Solid mine waste
typically acts as a long-term source of contaminants for surface waters and groundwater. Mill
tailing impoundments can be prone to large-scale erosion, which can lead to downstream
transport and deposition. Fluvial tailings depositions can act as long-term sources of
contamination to surface waters that must be identified and addressed during remediation.
The recovery of fish populations following remediation involves a number of factors.
Contaminant sources must be addressed, including solid materials, such as tailings, that may
have been transported downstream and redeposited on stream banks. Water quality must return
to tolerable conditions for both the fish and the macroinvertebrates that serve as food sources for
the fish. Suitable habitat must be available for both fish and macroinvertebrates. Finally,
adequate time must be provided for both the macroinvertebrate and fish populations to rebound.
The abundance and diversity of fish and aquatic macroinvertebrates have shown steady increases
since the start of remediation in both watersheds. Clements et al. (2010) documented increases in
macroinvertebrates and fish in the upper Arkansas River in the period from 1989 to 2006 once
remediation began. They reported rapid response of macroinvertebrates to improvements in
water quality, reflecting their resilience to chemical stressors. This observation suggests that a
food source was not a limiting factor in trout recovery. Policky (2016) reported increased
-------
numbers of trout over 14 in long in the upper Arkansas River near Leadville beginning around
2002. The upper Arkansas River is predominantly a wild brown trout fishery, although rainbow
trout have been historically stocked mostly in lakes (Policky, 2016). Brown trout constitutes over
75% of the trout community. Woodling et al. (2005) reported similar changes in the Eagle River
for macroinvertebrates and fish after remediation began in the period from 1990 to 2005.
Uncertainty in inter-annual healthy populations
In the Arkansas River watershed, significant correlations of biologic measures of
macroinvertebrate and fish populations were found with level of exceedance of water-quality
criteria for Zn, Cd, and Cu, location, and sampling date. Broader aspects of the water chemistry
(temperature, pH, specific conductance, and alkalinity) and physical hydrology (stream depth,
velocity, and discharge) did not improve correlations (Clements et al., 2010). For brown trout
density and biomass, the highest correlation coefficients (r2) achieved were 0.54 and 0.51,
respectively. In other words, approximately 50% of the variance in brown trout population can
be related to the factors outlined above.
In the Eagle River watershed, logarithmic biologic measures of fish populations correlated
significantly with level of exceedance of water-quality criteria for Zn, Cd, and Cu that varied by
sample site, but most strongly at the three sites immediately downstream from the Eagle mine
but before the confluence with Gore Creek. The coefficients of correlation (r2) varied by site and
ranged from 0.03 to 0.86 (Woodling et al., 2005). Correlations of the number of 1-year old
brown trout with maximum flow from the preceding year were weak at best (Woodling et al.,
2005). As with the Arkansas River, a significant amount of the variation cannot be described by
differences in water quality and discharge.
For both watersheds, correlations among fish metrics and water-quality and physical hydrology
parameters generally yielded correlation coefficients (r2) that typically ranged between 0.50 and
0.70, but locally reached lows of 0.03 and highs of 0.86 (Woodling et al., 2005, Clements et al.,
2010). In other words, a significant portion of the variance is not described by the parameters
considered by these studies. High variance in trout populations from year to year is not unique to
watersheds that have been impacted by mine drainage. Dauwalter et al. (2009) examined
temporal variations in trout populations in North America through literature review. They found
that coefficients of variation for trout populations in healthy streams averaged 49 ± 27% (range
15 to 108%>) over time for all ages of trout. For brown trout in Colorado, they reported
coefficients of variation for abundance that ranged from 15%> for year 2+ trout to 82%> for year-1
trout in Little Beaver Creek. The coefficients of variation for brown trout in South Saint Vrain
Creek had a smaller range for various age groups from 17 to 49%>. Thus, the high variance
observed in the Arkansas River and Eagle River watersheds is characteristic of trout populations
in general, and not necessarily attributable to the effects of mine drainage and its remediation.
The rapid recovery of macroinvertebrates in the Arkansas River watershed, described by
Clements et al. (2010), suggests that the availability of food sources was not the limiting factor.
The rapid neutralization of acid-mine drainage upon mixing with higher alkalinity receiving
waters and subsequent precipitation of hydrated ferric oxides near the sites of mixing should
-------
limit physical impairment of aquatic habitats to the immediate site of mixing with limited
downstream impacts. Therefore, the recovery of the brown trout population should be considered
in terms of population growth through "normal" breeding and reproduction starting at the
remediation of water quality.
Modeling population growth
The literature on ecological population dynamics was reviewed to evaluate the role of time in
ecological recovery. Several models to describe population growth, particularly for fisheries,
were identified. The models vary in their complexity and can incorporate a number of variables
and other factors including growth, cooperation and competition within a species, interactions
with other species, age-class structure, and limiting physical characteristics of the habitat
(Berryman, 2003). Available models, particularly for brown trout, fall into two main categories:
population ecology models that seek to describe population dynamics and genetics, and
population distribution models that seek to describe habitat and spatial characteristics (Frank et
al., 2011). Population dynamics models are most appropriate for our ecosystem services
valuation exercise.
The logistic function (Verhulst, 1838) is considered to be an overly general, simplistic model for
describing ecological population dynamics (Turchin, 2001, Frank et al., 2011). More complex
population dynamic models, such as those based on the Leslie matrix (Leslie, 1945), require a
complex set of input parameters, including age-specific data on survival, fecundity, population
structure, and physical characteristics of the hydrologic setting (Sabaton et al., 1997, Gouraud et
al., 2001), all of which are beyond the scope of currently available data for this study. Despite its
simplistic approach to predicting population growth, the logistic function should be adequate for
the purposes of the present study because of the large coefficients of variations found for healthy
trout populations as described by Dauwalter et al. (2009). Clearly, if more sophisticated
population growth models are available for a watershed, they should be used instead of or in
addition to a logistic model.
The logistic function requires minimal inputs: a starting population, a carrying capacity for the
site, and a growth rate (Table 9.1), as described by Equation 10.1 below:
JV(t) =	—- (10.1)
v J N0+(K-N0)e"rt v '
where N(t) is the population at time t, N0 is the initial population, K is the carrying capacity of
the system, r is the rate of population growth, and t is time.
Site AR3 was used to calibrate the model for the entire watershed because that site displayed the
greatest and simplest increase in population after the last remediation event upstream from the
site - the remediation of the California Gulch tailings (Figure 8.3). The carrying capacity for the
study area was defined as the combined average population at site AR3 from 2002 to 2006. For
the Arkansas River, the starting population for each site was chosen as the population in 1991
(Clements et al., 2010). The simulations began at the end of the last remediation event upstream
-------
from the site. For EF5 and AR1, this event was the commissioning of the Leadville mine
drainage tunnel treatment system in 1992. For AR3 and AR5, the event was the California Gulch
tailings remediation, which was completed in 1999. Therefore, the only differences among the
models for individual sites are the starting populations and the recovery time (Table 10.1).
Table 10.1. Input Parameters Used for Trout Population Modeling of the Arkansas River Watershed
Site Initial Population	Carrying Capacity	Growth Rate
(number of fish)	(number of fish)	(% growth per year)
EF5 900	1456	0.944
AR1 1360	1456	0.944
AR3 279	1456	0.944
AR5 261	1456	0.944
The results of these simulations are presented in Figure 10.2 shown with fish population data
from those sites. The heavy green line represents the prediction for each site. The thin green lines
and thin black lines represent 20 and 50% variance, respectively. The 20% value approximates
the "best case" for observed Colorado brown trout coefficients of variations from Dauwalter et
al. (2009), and the 50% value reflects the approximate coefficient of variation for all trout in
their study.
-------
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Years since 1991
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Years since 1991
~i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Years since 1991
CD
i_
ru
Years since 1991
CD
i_
TO
Figure 10.2. Comparison of Fish Population Data with Ecological Population Growth Models. The
heavy green lines represent the population predictions, the thin green lines represent a variance
of 20%, and the thin black lines represent 50% variance. Arkansas River: Site EF5, Site AR1, Site
AR3, and Site AR5.
In general, the data for Sites EF5, AR1, and AR3 appear to scatter randomly about the prediction
within the amount of variance reported by Dauwalter et al. (2009). (Note that site AR3 was used
to calibrate the model for use in making predictions at the other sites.) In contrast, site AR5
appears to fall short of the predictions. Walton-Day and Mills (2015) described ongoing drainage
from the Dinero Tunnel on Lake Fork, a tributary of the Arkansas River (Figure 8.3), which
appears to have continued to suppress recovery in this part of the river.
The simulations for the Eagle River watershed assumed that the carrying capacity and growth
rate used in the Arkansas River watershed (Site AR3) would be applicable to the Eagle River
watershed given similar stream and habitat conditions. The simulations began at the nominal end
of the remediation at the Eagle mine in 2001 (Table 8.4). Input for the Eagle River simulations
are summarized in Table 10.2.
Table 10.2. Input Parameters Used for Trout Population Modeling of the Eagle River Watershed
Site	Initial Population	Carrying Capacity	Growth Rate
(number of fish)	(number of fish)	(% growth per year)
Site 1
786
1456
0.944
-------
Site
Initial Population
(number of fish)
Carrying Capacity
(number of fish)
Growth Rate
(% growth per year)
Site 1.9
1340
1456
0.944
Site 2.9
418
1456
0.944
Site 3
508
1456
0.944
Site 4
582
1456
0.944
Site 5
379
1456
0.944
Site 6
803
1456
0.944
In general, the data for Sites 1, 1.9, 3, 4, 5, and 6 are within the plus or minus 50% prediction for
population growth, consistent with the amount of variance reported by Dauwalter et al. (2009)
shown in Figure 10.4. Site 2.9 is the sole site for which population measurements fell below this
predicted range. Site 1, the upstream site, has had zinc concentrations that have been consistently
below the acute toxicity guideline (Figure 9.4). Site 2.9 has exhibited decreases in zinc
concentrations over time, yet it is the site that has consistently exceeded the acute toxicity
guideline by the greatest amount. This consistently high exceedance suggest that a source of zinc
must remain in this reach because the proximal sites both upstream and downstream exceed the
guideline by lesser amounts. The reach between Site 2.9 and Site 3 is known as the North
Property area (OU1, Table 8.4). A remedial investigation, released in 2006, of the North
Property area documented remnants of solid mine waste and contaminated seepage from waste
piles in this area (ERM, 2006). Even though most of the sites fall with the uncertainty range in
the predicted population growth, the majority of the sites (Sites 1, 29, 5, and 6) appear to be
biased toward the lower half of the predicted range (Figure 10.4). This low bias may indicate that
the carrying capacity of this stream may be lower than that of the Arkansas River that was used
in the model. The discharge in the two watersheds is similar with base flow typically ranging
between 10 and 20 cfs and peak flow between 200 and 1000 cfs (Figures 8.4 and 8.7,
respectively). Thus, differences in carrying capacity may be due to differences in the physical
habitat or food sources.
-------
2500
03 2000-
1500 -
03
Q_
1000
500
Date
3500-
03 3000-
cc 2500-
03 2000'
Q_
x: 1500.
cn
il 1000.
500.
Site 1.9
LnLninr^r^cocTicri
oioiaioioicioiai
cnmcnoioioicrioi
(N ro ro
rM r\l rM
2500
Date
03 2000
¦.	
O
TO
1500
03
Q.
"1000
500
0
Site 2.9
2500
03 2000-
i_
2 1500
03
Q.
1000-
500
Site 3
Date
Figure 10.4. Comparison of Fish Population Data with Ecological Population Growth Models. The
-------
heavy green lines represent the population predictions, the thin green lines represent a variance
of 20%, and the thin black lines represent 50% variance. Eagle River: Site 1, Site 1.9, Site 2.9, Site
3, Site 4, Site 5, and Site 6.
Effects of residual wastes
Both the Arkansas River and the Eagle River watersheds have trout populations that locally
deviate from the predicted values. For both watersheds, the areas of suppressed fish populations
can be linked to areas that contain either residual mine waste that was not removed during
remedial, contained seepage from waste piles, or fluvial tailings that have not been remediated.
In the Arkansas River watershed, the Dinero Tunnel continues to discharge contaminated
drainage (Walton-Day and Mills, 2015). In other words, deviations from predicted populations
appear to be an indicator of the continued presence of solid or aqueous contamination.
Link between fish population growth predictions and estimates of
their value
The recreational angling valuation literature currently provides three methods to link changes in
catchable target fish population to estimates of the resulting increase in net economic value, as
discussed in the endpoint problem and benefit transfer sections (Sections 6 and 7). The first
method is to estimate (or assume) the proportion of increased catchable fish population that will
be caught by anglers and multiply this percentage by the value of increased catch (Johnson et al.,
1995, Mazzota et al., 2015). The second method is to econometrically estimate the impact that
increased fish population has on the number of fish caught, while controlling for other variables
such as skill, equipment, species, and angling time (Loomis and Ng, 2009, Ng, 2011). The final
method (which is an extension of the second method) estimates the impact of increased catchable
population on catch rate and then the impact of the new catch rate on the number of angling
days/trips taken (Loomis and Ng, 2009, Ng, 2011). The resulting increase in net economic value
is divided by the increase in the catchable fish population to achieve a net economic value for the
change in catchable fish population (Loomis and Ng, 2009, Ng, 2011).
The following ecosystem service valuations cannot employ the second method because there are
no data available to control for other explanatory variables. For the Upper Arkansas River, there
is only one point in time and space (sampling site AR-5 in River Reach 1 for the year 2008)
where predictions regarding fish population and fish catch coincide. This data point is applied to
each of the four sampling sites to allow use of the first (proportional) method above. The final
method (WTP per catchable fish) is used to evaluate the sensitivity of the valuation exercise to
the proportional linkage assumption made in the first method. For the Eagle River, only the final
method of applying WTP per catchable fish is employed.
-------
Ecosystem service valuation
Four sampling sites in the Upper Arkansas River
This section values the geo-environmental model's trout population growth predictions for the
Arkansas River watershed (Table 10.1). This valuation estimates the change in net economic
value resulting from the initial catchable population's increase to the catchable carrying capacity.
It is important to note that this section does not value the full extent of catchable trout population
recovery resulting from the California Gulch remediation. Instead, it focuses on the areas
surrounding the four sampling sites where the geo-environmental model had sufficient data to
model changes in fish populations. (For a broader ecosystem service valuation of the remediation
(without the aid of geo-environmental modeling), please see Appendix C. The relative scale of
the fishery that was impacted by remediation of the California Gulch Superfund site and the area
where sufficient data were available for evaluation are shown in Figure 10.5.
-------
Upper Arkansas River Watershed
Figure 10.6 illustrates the location of the four sampling sites. The river reaches labeled 1 through
11 represent reaches of the Upper Arkansas River where Colorado Parks and Wildlife (CPW) has
estimated angling hours, angling days, and fish caught by conducting creel surveys at particular
locations (Policky, 2013). CPW has also extrapolated the localized creel survey results to
broader river reaches (Policky, 2013). The angling information collected by the creel surveys is
useful for linking changes in fish population to changes in benefits derived from humans, namely
-------
anglers. The river reaches labeled EF (East Fork), -1, and 0 represent reaches that have not been
surveyed by CPW, but correspond to sampling sites EF-5, AR-1, and AR-5, respectively.
Upper Arkansas River Watershed
Leadville
%
\ ^
i
&
,?•
, "2.
, n
jena
Vista
Johnson
Village
Salida
Turquoise
Lake
Poncha
Springs
10 mi
Canon Citv
Parkdale
Florence
Pueblo
West
^ Pueblo
20 km
Figure 10.6: Upper Arkansas River: Colorado Parks and Wildlife River Segments and EPA
Sampling Sites
Table 10.3 provides detailed information on the size of creel census areas (fishing access points
where anglers were surveyed) and their respective river reaches. River reach 1 is the only reach
relevant for this analysis, but Table 10.3 describes the relationship between creel census areas
-------
and broader river reaches for the Upper Arkansas River all the way down to Parkdale to provide
a broader context for this study.
Table 10.3. Colorado Parks and Wildlife River Reaches and Corresponding Creel Census Areas
River
Reach
River Reach Name
Miles
in
River
Reach
Creel Census Area
Creel
Census
Miles
EF
EF-1 - Confluence
2.2

NA
-1
Confluence - California Gulch
3.8

NA
0
California Gulch - Crystal Lakes
2.8

NA
1
Crystal Lakes - Kobe Bridge
5.1
Highway 24 - Kobe
3.2
2
Kobe Bridge - Lake Creek
4.2
Kobe - Two Bit Gulch
2.2
3
Lake Creek - Otero Bridge
10.6
Ball Town - Granite
2.6
4
Otero Bridge - Highway 285 Bridge
9
Otero Bridge - Railroad Bridge
3.1
5
Highway 285 Bridge - Ruby Mountain
6
Big Bend - F Street
5.7
6
Ruby Mountain - Stone Bridge
11.2
Big Bend - F Street
5.7
7
Stone Bridge - Stockyard Bridge
10.9
Big Bend - F Street
5.7
8
Stockyard Bridge - Howard Bridge
11.4
Stockyard Bridge - Badger Creek
5.9
9
Howard Bridge - Lazy J
8.3
Big Cottonwood Creek - Lone
Pine
3
10
Lazy J - Texas Creek
12
Big Cottonwood Creek - Lone
Pine
3
11
Texas Creek - Parkdale
13.3
Big Cottonwood Creek - Lone
Pine
3
Data and methods
Two approaches are employed in this benefit transfer valuation. Both estimate the change in
annual net economic benefit as a result of the change in catchable brown trout population from
initial to carrying capacity.
The first approach is straightforward because the value estimate employed from the economic
literature directly values the endpoint in question - catchable fish population. Modeled fish
population estimates (see Figure 10.2) are multiplied by the estimated hectares in each reach5
and estimated WTP for catchable trout ($0.60 in 2013 dollars from Table 7.9) from Loomis and
Ng (2009). The results of this approach are presented in Table 10.4.
5 Estimated using Google Earth map software.
-------
Table 10.4. Value of Catchable Fish Population, estimated via the WTP for Catchable Fish
Approach (2013$)	
River
Hectares
Catchable Fish Population
Catchable Fish Population
Reach
in Reach
Estimate

Value Estimate (2013$)


1995
2008
1995
2008
EF
2.2
1,436
1,456
$1,900
$1,900
-1
4.7
1,454
1,456
$4,100
$4,100
0
4.1
279
1,455
$700
$3,600
1
8.7
261
1,455
$1,400
$7,600
Total
19.6
3,430
5,822
$8,100
$17,200
The second approach requires more complex calculations because the benefit transfer estimate
from the literature values fish caught, which is one step removed from the endpoint of catchable
fish population. This subtle difference requires the use of creel census data that have been
extrapolated by an expert from the creel survey site to the broader river reach. This set of data is
used to estimate the percentage of catchable brown trout population that gets caught. The
percentage estimate serves as the link between the WTP per fish caught and the ecosystem
service. Calculation of the catch percentage requires data on the number of days that anglers
spent fishing and their daily catch rate to estimate the number of fish caught for a relevant time
and location within the sampling area. Two sets of data points fulfilling this purpose are provided
by CPW creel censuses conducted along the Upper Arkansas River between Crystal Lakes and
Kobe Bridge for the years 1995 and 2008 (Policky, 2012, Policky, 2013).
Because creel census areas are not representative of their broader river reaches, expert
knowledge is required to extrapolate creel census area estimates to broader river reach estimates.
Policky (2013) provides an expert extrapolation of creel census angling days to river reach
angling days for each river segment in the year 2012. Equation 10.2 combined with Table 10.5
illustrate how the extrapolation factor was calculated for river reach 1 in 2012. This extrapolation
factor is then applied to creel census data for the relevant years of 1995 and 2008:
2Q12ReachDayEstimate — 2Q12CreelDayEstimate
ExtrapolationFactor — 1 H		———;	
Table 10.5. CPW Creel Census and River Reach Angling Day Estimates Used to Calculate
Extrapolation Factor	
River Reach Number Creel Census Area
CPW Estimate of
Angling Day
Angling Days
Angling Days per
Extrapolation Factor

River Reach

2012
2012

1
3,600
7,200
1.99
-------
The extrapolation factor from Table 10.5 is multiplied by creel census angling hours (Policky,
2012, Policky, 2013) to estimate river reach angling hours in Table 10.6, which are then
multiplied by the average hourly catch rate to estimate the number of fish caught in Table 10.6.


Average

River
River Reach
Hourly
Estimated
Reach
Angling Hours
Catch Rate
Fish Catch


1995 2008
1995 2008
1
4,700
NA 0.9
0 4,322
To estimate the percentage of the fish population caught by anglers in river reach 1, the number
of fish caught is divided by the product of the number of hectares in the reach and the fish
population estimate for AR-5 (Table 10.7). River reach 1 is the only location where trout
population estimates and creel data coincide.
River
Hectares
Catchable
Actual
Percentage of Catchable
Fish Actually
Reach
in Reach
Fish Population
Fish Catch
Caught


1995 2008
1995 2008
1995 2008
1
8.7
2,281 12,717
0 4,322
0% 34%
The percentages estimated in Table 10.7 for river reach 1 are applied to reaches 0, -1, and EF as
well. Table 9.8 depicts how the predicted catchable fish population estimates are combined with
these percentage-catch estimates and WTP per fish caught ($2.94) estimates from USEPA
(2006a) to estimate the value of the catchable fish population.








Catchable Fish
River
Hectares
Fish Population
Fish Catch
Fish Catch
Population
Reach
in Reach
Estimate
Estimate
Value Estimate
Value Estimate


1995
2008
1995
2008
1995
2008
1995
2008
EF
2.2
3,099
3,143
0
1,068
$0
$3,140
$0
$1,067
-1
4.7
6,785
6,785
0
2,306
$0
$6,780
$0
$2,304
0
4.1
1,113
5,995
0
2,038
$0
$5,991
$0
$2,036
1
8.7
2,281
12,717
0
4,322
$0
$12,707
$0
$4,319
Total
19.6
13,277
28,639
-
9,734
$0
$28,617
$0
$9,726
-------
Discussion and Conclusions
The changes in net economic value resulting from the WTP for catchable fish approach and from
the 'percentage fish caught' approach are almost identical - $9,217 ($17,184 - $7,966 in Table
10.4) and $9,726 ($9,726 - $0 in Table 9.8) per year, respectively. In comparison, the total
documented costs for clean up at Leadville exceed $138,000,000 (Table C.9). The first approach
relies on WTP for catchable fish estimates from one study (Loomis and Ng, 2009), whereas the
second approach relies on a meta-analysis of WTP for fish caught estimates (USEPA, 2006a).
The first approach is appealing because of its simplicity, but future research must work to
develop the economic valuation literature regarding WTP for catchable fish. The second
approach is appealing because the WTP estimate from USEPA (2006a) is grounded in a robust
valuation literature search. However, the results are sensitive to the accuracy of creel surveys,
the extrapolation of these survey results to wider river reaches, and, above all, the percentage of
catchable fish in the population that are actually caught.
The valuation exercise in this section successfully links the geo-environmental ecosystem service
outputs to estimates of their value. However, the limited nature of available data limited this
linked natural and social science exercise to a small portion of the whole fishery that was
improved as a result of the California Gulch Superfund remediation. Appendix C provides an
attempt to more broadly characterize fishery benefits of remediation in the absence of geo-
environmental model outputs.
Six sampling sites in the Eagle River
This brief section values the model's trout population growth predictions for the sampling sites
on the Eagle River (Table 10.2). It is important to note that this section focuses narrowly on the
seven sampling sites where the model had sufficient data to model changes in fish populations.
No connection could be made between creel survey data and relevant river reaches, so the
change in net economic value estimates below cannot be extrapolated beyond their units of
dollars per hectare. Therefore, the information presented in Table 10.9 is best viewed as an
untested hypothesis about the value of this recovering ecosystem service in the Eagle River
watershed. Future creel and economic surveys will be needed to evaluate these predictions.
Table 10.9. Trout Population Characteristics and Estimated Values for Eagle River Watershed
Site
Initial
Carrying
Initial Population
Carrying Capacity
Change in Net

Population
Capacity
Value Estimate
Population Value
Economic

(trout/ha)
(trout/ha)
($/ha)
Estimate ($/ha)
Value ($/ha)
Site 1
786
1456
$472
$874
$402
Site 1.9
1340
1456
$804
$874
$70
Site 2.9
418
1456
$251
$874
$623
Site 3
508
1456
$305
$874
$569
Site 4
582
1456
$349
$874
$524
Site 5
379
1456
$227
$874
$646
Site 6
803
1456
$482
$874
$392
-------
Recommendatons
This study has pioneered the integration of: 1. a geologically-based environmental assessment of
remediation of abandoned mines sites dominated by MIW, 2. the progress of associated in-
stream biologic recovery as a result of remediation, and 3. the potential economic impacts of that
remediation and biologic recovery due to sport fishing. The study has relied on data from
existing sources to conduct this exercise. The study has also utilized existing models in disparate
fields such as geochemistry, hydrology, population ecology, and economics to form this
integrated approach. As originally envisioned, this study planned to conduct similar assessments
for, 1. mining districts that developed a variety of mineral deposit types having different
environmental contaminants, and 2. several ecosystem services. Ultimately, the scope of the
study was limited by the availability of adequate data from abandoned mine sites and their
watersheds, and by the availability of suitable geochemical, ecological, and economic models.
In the course of the geochemical and hydrological modeling in this study, the greatest limitations
were found in the availability of data. For the sites used in the study, minimum requirements
included pH, hardness, and trace metals, specifically Zn, Cd, and Cu. A broader range of water
quality parameters would have enabled more comprehensive modeling of source, transport, and
fate processes than was afforded by the available data.
The modeling of the trout population ecology used an admittedly simplistic approach to model
trout population growth, the logistic function, yet that level of sophistication was compatible
with the level of data available (Frank et al., 2011). With the current model, the input parameters
included a starting population, a growth rate, and the carrying capacity of the stream. The
starting population was easily obtained from the available data. However, the growth rate and
carrying capacity required a number of assumptions. A better estimate of the carrying capacity
may be obtainable from investigations in nearby streams in adjacent watersheds unaffected by
MIW. The growth rate was estimated based on the fastest and most systematic increase in
population in the Arkansas River watershed. A more accurate growth rate for recovering natural
populations will likely require extensive literature review of numerous case studies. Despite the
simplistic approach embodied in the logistic function, the uncertainties seem to fit within the
uncertainties observed in healthy natural populations year to year (Dauwater et al., 2009). More
sophisticated population dynamics modeling requires information such as age class structure of
populations and spawning efficiency, among others, that are beyond the scope of the current
study (Frank et al., 2011).
" >-Pi" III" I li
Decisions regarding mine site remediation can be clouded by the fact that its costs are easily estimated
but its benefits are not. Remediation costs for labor, equipment rental, and materials can all be
estimated within a reasonable range. However, benefits such as water of higher quality, improved fish
habitat, or healthy soil generate no revenue. For proposed mines, the opposite is true. The benefits are
-------
well summarized in discounted cash flows, while the costs of impact to the environment and
communities are less clear. A goal of this project was to explore a possible framework to begin to
remedy this problem.
Recent mine spills and mine development controversies reveal the importance of three ecosystem
services modeled via benefit transfer valuation (catchable fish population in Section 6.1, drinking water
in Appendix B, and aquatic habitat in Appendix B). Concerns about the Mount Polley mine tailings spill in
August 2014 (British Columbia, Canada) focused on salmon killed (or diverted from spawning habitat),
household drinking water intake quality, and impacts to aquatic habitat. The ecosystem service
valuation models presented in Section 6.1 and Appendix B are capable of valuing much of the ecosystem
damage from this spill if ecologists are able to model resulting changes to ecosystem services. Similarly,
public concern about development of the contentious Pebble project in southwestern Alaska focuses on
possible impacts to salmon populations and salmon habitat. In short, mine site pollution commonly
affects water and fish. Estimation of the value of related ecosystem services captures major components
of the value of mine site pollution, while abiding by the restrictive ecosystem service framework.
In contrast, the August 2015 Gold King mine spill in southwestern Colorado demonstratred the
shortcoming of current scientific and economic understanding when it comes to modeling the long-term
ecosystem service values impacted by a mine spill. First, the Animas River was not a pristine watershed
before the Gold King mine briefly turned it yellow. Any attempt to study the ecosystem service impacts
of the spill would first have to disentangle the impacts of background levels of contamination from the
impacts of contaminants introduced by the spill. Assuming that this could be done, ecologists would
then have to trace the impact of the spill's contaminant through the ecosystem until they were able to
quantify changes to an ecosystem service that economists could value.
For example, one concern related to the Gold King spill was the long-term contamination of river bed
sediment. This contamination could conceivably make its way into plants/micro-organisms and work its
way up the food chain until it had an impact on an ecosystem service, such as fish population, bird
population, or wildlife population. If it were scientifically possible to trace the contaminants through the
ecosystem, then economists would have to figure out how to value these ecosystem service changes.
Huntable bird and wild life populations do not have meta-analyses that have valued WTP to shoot
another bird/deer, as USEPA (2006a) has done for recreational angling. Existing estimates of WTP for a
day of hunting do not relate directly enough to the ecosystem services in question.
As for the ecosystem services mentioned in Appendix A, provision of ecosystem service changes,
additional time, and additional funding could allow value estimation and incorporation into the benefit
transfer model. Groundwater contamination, water supply reliability, and irrigation water are three of
the most important ecosystem services that were not valued. Such research would also be useful for
decisions regarding unconventional oil and gas development.
Future research emanating from this analysis calls for a full-scale characterization of ecosystem service
impacts from acid mine drainage in the Animas watershed and predictive modeling of ecosystem service
impacts from large, open-pit tailings storage facility failures. The Animas is representative of many
-------
watersheds impacted by acid mine drainage in the western United States. Characterization of the
impacts could lead to estimation of benefits from remediation. While remediation is mainly driven by
regulatory concerns, it is possible that defensible remediation benefit estimates could channel more
resources towards acid mine drainage remediation.
Predictive modeling of ecosystem service impacts from worst-case scenario failures may help engineers
of future mines to more accurately assess financial risks during design trade-off studies. From a
regulatory perspective, it would also provide more defensible amounts for environmental liability
bonding. In addition, insurance markets are currently unwilling to cover low frequency, high
consequence events like the tailing spills at Mount Polley and Samarco. If they could forecast ecosystem
service impact values and spill frequency, maybe they would be more likely to insure such events.
Insurance would spread the risks more widely and could provide a market mechanism to help regulate
these catastrophic failures.
Finally, the Office of the President recently mandated that all federal agencies begin incorporating the
value of ecosystem services into federal decisions (EOP, 2015). This chapter represents a robust analysis
of ecosystem service valuation for federal decisions related to the extractive sector. Future research
ought to be conducted to incorporate this model into the decision-making frameworks of relevant
federal agencies.
References
Ahlheim, M., Fror, O., Lehr, U., Wagenhals, G., and Wolf, U., 2004, Contingent valuation of mining land
Reclamation in East Germany: Inst, fur Volkswirtschaftslehre, Univ.
Amos, R.T., Blowes, D.W., Bailey, B.L., Sego, D.C., Smith, L., and Ritchie, A.I.M., 2015, Waste-rock
hydrogeology and geochemistry: Applied Geochemistry, v. 57, p. 140-156.
Bailey, R.G., Avers, P.E., King, T., and McNab, W.H., eds, 1994, Ecoregions and subregions of the
United States (map). Washington, DC: USDA Forest Service. 1:7,500,000. With supplementary table
of map unit descriptions, compiled and edited by W. H. McNab and R. G. Bailey.
BBC Research, 2008, The Economic Impacts of Hunting, Fishing and Wildlife Watching in Colorado:
BBC Research and Consulting.
Beaty, D.W., ed., 1990, Origin of the ore deposits at Gilman, Colorado: Economic Geology Monograph
7, p. 193-265.
Beaty, D.W., Landis, G.P., and Thompson, T.B., 1990, Carbonate-hosted sulfide deposits of the Central
Colorado Mineral Belt: Introduction, general discussion, and summary: Economic Geology
Monograph 7, p. 1-18.
Behnke, R., 2010, Trout and salmon of North America: Simon and Schuster, 384 p. Bergstedt, L.C.,
Chadwick, J.W., Conklin Jr, D.J., and Canton, S.P., 2005, Improvements in brown trout and
invertebrate populations in the Arkansas River during reclamation efforts on California Gulch
Proceedings, 22nd Annual Meeting of America Society of Mining and Reclamation, v. 3134, p. 54-70.
-------
Bergstrom, J. C., and Taylor, L. O., 2006, Using meta-analysis for benefits transfer: Theory and practice:
Ecological Economics, v. 60, p. 351-360.
Berryman, A.A., 2003, On principles, laws and theory in population ecology: Oikos, v. 103, p. 695-701.
Bingham, T. H., David, T., Graham-Tomasi, T., Kealy, M.J., LeBlanc, M., and Leeworthy, R., 1992,
Benefits transfer: procedures, problems, and research needs: Association of Environmental and
Resource Economists Workshop, Snowbird, Utah USA, p. 3-5.
Bookstrom, A.A., 1990, Tectonic setting, igneous rocks, and ore deposits of the northeastern segment of
the Colorado mineral belt, in Beaty, D.W., Landis, G.P., and Thompson, T.B., eds., Carbonate-hosted
sulfide deposits of the central Colorado mineral belt: Economic Geology Monograph No. 7, p. 45-65.
Borisova-Kidder, A., 2006, Meta-analytical estimates of values of environmental services enhanced by
government agricultural conservation programs. Dissertation, Ohio State University, 190 p.
Boyd, J.W., 2007, The endpoint problem. Resources Washington DC, v. 165, p. 26-28.
Boyd, J., and Banzhaf, S., 2007, What are ecosystem services? The need for standardized environmental
accounting units: Ecological Economics, v. 63, p. 616-626.
Boyle, K.J., Poe, G.L., and Bergstrom, J.C., 1994, What do we know about groundwater values?
Preliminary implications from a meta-analysis of contingent-valuation studies: American Journal of
Agricultural Economics, v. 76, p. 1055-1061.
Boyle, K.J., Roach, B., and Waddington, D.G., 1998, 1996 Net economic values for bass, trout and
walleye fishing, deer, elk and moose hunting, and wildlife watching: addendum to the 1996 National
Survey of Fishing, Hunting and Wildlife-Associated Recreation, v. 92. US Fish and Wildlife Service.
Boyle, K.J., Kuminoff, N.V., Parmeter, C.F., and Pope, J.C., 2010. The benefit-transfer challenges.
Annual Reviews of Resource Economics, v. 2, p. 161-182.
Burton, M., Jasmine Z.S., and White, B., 2012, Public preferences for timeliness and quality of mine site
rehabilitation. The case of bauxite mining in Western Australia. Resources Policy: v. 37, p. 1-9.
Button, K., Stough, R., Arena, P., Comp, A., and Casper, M., 1999, Dealing with environmental legacy
effects: the economic and social benefits of acid mine drainage remediation: International Journal of
Environment and Pollution, v. 12, p. 459-475.
Carson, R.T., Mitchell, R.C., Hanemann, W.M., Kopp, R.J., Presser, S., and Ruud, P.A., 1992, A
contingent valuation study of lost passive use values resulting from the Exxon Valdez oil spill:
Technical Report No. 6984, University Library of Munich, Germany.
Carson, R.T., Flores, N.E., and Meade, N.F., 2001, Contingent valuation: controversies and evidence:
Environmental and Resource Economics, v. 19, p. 173-210.
Clements, W.H., Vieira, N.K.M., and Church, S.E., 2010, Quantifying restoration success and recovery in
a metal-polluted stream: a 17-year assessment of physicochemical and biological responses: Journal
of Applied Ecology, v. 47, p. 899-910.
Costanza, R., De Groot, R., Farber, S., Grasso, M., Hannon, B., Limburg, K., Naeem, S., O'Neill, R.V.,
Paruelo, J., Raskin, R.G., Sutton, P., and Van den Belt, M., 1998, The value of the world's ecosystem
services and natural capital: Ecological Economics, v. 25, p. 3-15.
Damigos, D., and Kaliampakos, D., 2003, Assessing the benefits of reclaiming urban quarries: a CVM
analysis: Landscape and Urban Planning, v. 64, p. 249-258.
-------
Damigos, D., and Kaliampakos, D., 2006, The "battle of gold" under the light of green economics: a case
study from Greece: Environmental Geology, v. 50, p. 202-218.
Dauwalter, D.C., Rahel, F.J., and Gerow, K.G., 2009, Temporal variation in trout populations:
Implications for monitoring and trend detection: Transactions of the American Fisheries Society, v.
138, p. 38-51.
Diamond, P.A., and Hausman, J.A., 1994, Contingent valuation: Is some number better than no number?:
Journal of Economic Perspectives, v. 8, p. 45-64.
EOP, 2015, Memorandum for executive departments and agencies: Incorporating ecosystem services into
Federal decision making: Technical Report. M-16-01. Executive Office of the President.
ERM, 2006, Remedial investigation report: Bolts Lake area and areas within OU-1 of Eagle Mine Site:
Environmental Resources Management, Greenwood Village, Colorado, September 15, 2006, 260 p.
Farber, S., and Griner, B., 2000, Valuing watershed quality improvements using conjoint analysis:
Ecological Economics, v. 34, p. 63-76.
Fisher, B., Turner, K., Zylstra. M., Brouwer, R., Groot. R., Farber. S., Ferraro, P., Green, R., Hadley, D.,
Harlow, J. and Jefferiss, P., 2008. Ecosystem services and economic theory: integration for policy-
relevant research: Ecological Applications, v. 18, p. 2050-2067.
Frank, B.M., Piccolo, J.J., and Baret, P.V., 2011, A review of ecological models for brown trout towards
a new demogenetic model: Ecology of Freshwater Fish, v. 20, p. 167-198.
Gouraud, V., Bagliniere, Baran, P., Sabaton, C., Lim, P., and Ombredane, D., 2001, Factors regulating
brown trout populations in two French rivers: application of dynamic population model: Regulated
Rivers: Research and Management, v. 17, p. 557-569.
Haab, T. C., Interis, M. G., Petrolia, D. R., and Whitehead, J. C., 2013, From hopeless to curious?
thoughts on Hausman's 'dubious to hopeless' critique of contingent valuation: Applied Economic
Perspectives and Policy, v. 35, p. 593-612.
Hanemann, W.M, 1994, Valuing the environment through contingent valuation: Journal of Economic
Perspectives, v. 8, p. 19-43.
Hausman, J., 2012, Contingent valuation: From dubious to hopeless: Journal of Economic Perspectives, v.
26, p. 43-56.
IEc (Industrial Economics), 2006, Upper Arkansas River basin natural resource damage assessment:
Preliminary estimate of damages: Technical Report GS-10F-0224J. USFWS.
Ingraham, M.W., and Foster, S.G., 2008, The value of ecosystem services provided by the US National
Wildlife Refuge System in the contiguous US: Ecological Economics, v. 67, p. 608-618.
Johnson, D.M., 1989, Economic benefits of alternative fishery management programs. Ph.D. Dissertation,
Colorado State University, Fort Collins.
Johnson, D. M., Behnke, R.J., Harpman, D.A., and Walsh, R.G., 1995, Economic benefits and costs of
stocking catchable rainbow trout: a synthesis of economic analysis in Colorado: North American
Journal of Fisheries Management, v. 15, p. 26-32.
Johnston, R.J., Ranson, M.H., Besedin, E.Y., and Helm, E.C., 2006, What determines willingness to pay
per fish? A meta-analysis of recreational fishing values: Marine Resource Economics, v. 21, p. 1-32.
-------
Kaul, S., Boyle, K.J., Kuminoff, N.V., Parmeter, C.F., and Pope, J.C., 2013, What can we learn from
benefit transfer errors? Evidence from 20 years of research on convergent validity: Journal of
Environmental Economics and Management, v. 66, p. 90-104.
Kirchhoff, S., 1998, Benefit function transfer vs. meta-analysis as policy-making tools: a comparison. In:
Proceeding of Workshop on Meta-Analysis and Benefit Transfer: State-of-the art and Prospects.
Tinbergen Institute, Amsterdam.
Klemetsen, A., Amundsen, P.A., Dempson, J.B., Jonsson, B., Jonsson, N., O'Connell, M.F., and
Mortensen, E., 2003, Atlantic salmon Salmo salar L., brown trout Salmo trutta L. and Arctic char
Salvelinus alpinus (L.): a review of aspects of their life histories: Ecology of Freshwater Fish, v. 12,
p. 1-59.
Klucas, G. 2004. Leadville: The struggle to revive an American town. Island Press, 303 p.
Kontogianni, A., Luck, G.W., and Skourtos, M., 2010, Valuing ecosystem services on the basis of
service-providing units: A potential approach to address the cndpoint problem' and improve stated
preference methods: Ecological Economics, v. 69, p. 1479 1487.
Kroeger, T., Loomis J.B., and Casey, F., 2008, Introduction to the wildlife habitat benefits estimation
toolkit.
LaBounty, J.F., Sartoris, J.J., Klein, L.D., Monk, E.F., and Salman, H.A., 1975, Assessment of heavy
metals pollution in the upper Arkansas Rover of Colorado: Applied Sciences Branch, Division of
General Research, Engineering and Research Center, US Department of the Interior, Bureau of
Reclamation, 126 p.
Leslie, P.H., 1945, On the use of matrices in certain population mathematics: Biometrika, v. 35, p. 213-
245.
Lienhoop, N., and Messner, F., 2009, The economic value of allocating water to postmining lakes in East
Germany: Water Resources Management, v. 23, p. 965-980.
Limburg, K.E., O'Neill, R.V., Costanza, R., and Farber, S., 2002, Complex systems and valuation:
Ecological Economics, v. 41, p. 409-420.
Lindsay, M.B.J., Moncur, M.C., Bain, J.G., Jambor, J.L., Ptacek, C.J., and Blowes, D.W., 2015,
Geochemical and mineralogical aspects of sulfide mine tailings: Applied Geochemistry, v. 57, p. 157-
177.
Loomis J.B., Kroeger, T., Richardson, L. and Casey, F., 2016, The benefit transfer and use estimating
model toolkit. CSU Agricultural and Resource Economics.
http://dare.aesci.colostate.edu/outreach/tools/ Accessed June 5th, 2016.
Loomis J.B. and Ng, K., 2009, 2009 Colorado angler summary report: Economic values and survey
results. Technical Report Colorado State University.
Loomis J.B. and L. Richardson, 2007, Benefit transfer and visitor use estimating models of wildlife
recreation, species and habitats: Department of Agricultural and Resource Economics, Colorado State
University.
Loomis J.B., 1992, The evolution of a more rigorous approach to benefit transfer: benefit
function transfer: Water Resources Research, v. 28, p. 701-705.
-------
Loomis J.B., 2000, Environmental valuation techniques in water resource decision making: Journal of
Water Resources Planning and Management, v. 126, p. 339-344.
Marr, J.C.A., Bergman, H.L., Lipton, J., and Hogstrand, C., 1995, Differences in relative sensitivity of
naive and metals-acclimated brown and rainbow trout exposed to metals representative of the Clark
Fork River, Montana: Canadian Journal of Fisheries and Aquatic Science, v. 52, p. 2016-2030.
Mazzotta. M., Wainger, L., Sifleet. S., Petty, J.T., and Rashleigh, B., 2015, Benefit transfer with limited
data: An application to recreational fishing losses from surface mining: Ecological Economics, v.
119, p. 384-398.
Mendes, I., Sardinha, I., and Milheiras, I. 2007, The social and economic value of the S. Domingos
abandoned mine rehabilitation projects: The case of S. Domingos mine: Contributions of Corporate
Social Responsibility to Sustainable Development.
Michael, N., and Pearce, D., 1989, Cost benefit analysis and land reclamation: a case study: LEEC Paper-
London Environmental Economics Centre, International Institute for Environment and Development.
Millennium Ecosystem Assessment, 2005, Ecosystems and human well-being: synthesis: Millenium
Ecosystem Assessment., Washington, DC, Island Press. 155 p.
Morey, E.R., Breffle. W.S., Rowe, R.D., and Waldman, D.M., 2002, Estimating recreational trout fishing
damages in Montana's Clark Fork River basin: summary of a natural resource damage assessment:
Journal of Environmental Management, v. 66, p. 159-170.
Navrud, S., and Ready, R.C., 2007, Environmental value transfer: issues and methods: Springer.
Neelawala, P., Wilson, C., and Athukorala, W., 2013, The impact of mining and smelting activities on
property values: a study of Mount Isa city, Queensland, Australia: Australian Journal of Agricultural
and Resource Economics, v. 57, p. 60-78.
Nehring, B.R., and Policky, G., 2003, A Synopsis of Seventeen Years Trout Population Biometrics in the
Upper Arkansas River: Colorado Division of Wildlife.
Nelson, J.P., and Kennedy, P.E., 2009, The use (and abuse) of meta-analysis in environmental and natural
resource economics: an assessment: Environmental and Resource Economics, v. 42, p. 345-377.
Ng, K., 2011, Valuing economic benefits of water's ecosystem services with non-market valuation
methods and regional input-output model., Ph.D. dissertation, Colorado State University
Nordstrom, D.K., and Alpers, C.N., 1999, Geochemistry of acid mine waters. Reviews in Economic
Geology, v. 6A, p. 133-160.
NRC (National Research Council), 2004, Valuing ecosystem services: Toward better environmental
decision-making. Washington, DC: The National Academies Press, 278 p.
Ofiara, D.D., 2002, Natural resource damage assessments in the United States: rules and procedures for
compensation from spills of hazardous substances and oil in waterways under US jurisdiction: Marine
Pollution Bulletin, v. 44, p. 96-110.
Olander, L., Tallis, H., Polasky, S., and Johnston, R.J., 2015, Best practices for integrating ecosystem
services into Federal decision making: Duke University, National Ecosystem Services Partnership.
Poe, GL., Boyle, K.J., and Bergstrom, J.C., 2001, A preliminary meta analysis of contingent values for
ground water quality revisited: The Economic Value of Water Quality, p. 137-162.
-------
Policky, G.A., 2012, Fisheries inventory report: Arkansas River 2012: Colorado Parks and Wildlife.
Policky, G.A., 2013, Fisheries inventory report: Arkansas River 2013: Colorado Parks and Wildlife.
Policky, G.A., 2016, Upper Arkansas River, fish survey and management data: Colorado Parks and
Wildlife, 5 p., accessed May 18, 2016, at
http://cpw.state.co.us/thingstodo/Fishery%20Survey%20Summaries/ArkansasRiverUpper.pdf
Portney, P.R., 1994, The contingent valuation debate: why economists should care: Journal of Economic
Perspectives, v. 8, p. 3-17.
Randall, A., Grunewald, O., Johnson, S., Ausness, R., and Pagoulatos, A., 1978, Reclaiming coal surface
mines in central Appalachia: a case study of the benefits and costs: Land Economics, v. 54, p. 472-
489.
Rimstidt, J.D., 1997, Gangue Mineral Transport and Deposition, in Barnes, H.L., ed., Geochemistry of
Hydrothermal Ore Deposits, 3rd Edition, Wiley, New York, p. 487-515.
Roline, R.A., and Boehmke, J.R., 1981, Heavy metals pollution of the upper Arkansas River, Colorado
and its effects on the distribution of the aquatic macrofauna: In: RECERC, v. 81. US Department of
the Interior.
Rosenberger, R.S., and Loomis, J.B., 2000, Using meta-analysis for benefit transfer: In-sample
convergent validity tests of an outdoor recreation database: Water Resources Research, v. 36, p.
1097-1107.
Rowe, R.D., and Schulze, W.D., 1985, Economic assessment of damage related to the Eagle Mine
facility: Energy and Resource Consultants.
Sabaton, C., Siegler, L., Gouraud, V., Bagliniere, J.L., and Manne, S., 1997, Presentation and first
applications of a dynamic population model for brown trout, Salmo trutta L:. Aid to river
management: Fisheries Management and Ecology, v. 4, p. 425-438.
Seal, R.R., II, and Hammarstrom, J.M., 2003, Geoenvironmental models of mineral deposits: examples
from massive sulfide and gold deposits: Environmental Aspects of Mine Wastes, J.L. Jambor, D.W.
Blowes, and A.I.M. Ritchie (eds.), Mineralogical Association of Canada Short Course Series, v. 31, p.
11-50.
Shrestha, R.K., and Loomis, J.B., 2001, Testing a meta-analysis model for benefit transfer in international
outdoor recreation: Ecological Economics, v. 39, p. 67-83.
Smith, K.S., 1999, Metal sorption on mineral surfaces: An overview with examples relating to mineral
deposits: in Plumlee, G.S., and Logsdon, M.J., eds., The environmental geochemistry of mineral
deposits, Part A—Processes, techniques, and health issues: Reviews in Economic Geology, v., 6A, p.
161-182.
Smith, V.K., 1993, Nonmarket valuation of environmental resources: an interpretive appraisal: Land
Economics, p. 1-26.
Smith, V.K., and Huang, J., 1995, Can markets value air quality? A meta-analysis of hedonic property
value models: Journal of Political Economy, v. 103, p. 209-227.
Smith, V.K., and Kaoru, Y., 1990, Signals or noise? Explaining the variation in recreation benefit
estimates: American Journal of Agricultural Economics, v. 72, p. 419-433.
-------
Thompson, T.B., and Arehart, G.B., 1990, Geology and the origin of ore deposits in the Leadville
District, Colorado: Part I. Geologic studies of orebodies and wall rocks: Economic Geology
Monograph 7, p. 130-155.
Trigg, A.B., and Dubourg, W.R., 1993, Valuing the environmental impacts of opencast coal mining in the
UK: The case of the Trent Valley in North Staffordshire: Energy Policy, v. 21, p. 1110-1122.
Turchin, P., 2001, Does population ecology have general laws?: Oikos, v. 94, p. 17-26.
Tweto, O., 1960, Pre-ore age of faults at Leadville, Colorado: U.S. Geological Survey Professional Paper
400-B, p. 10-11.
Tweto, O., 1968, Leadville district, Colorado, in Ridge, J.D., ed., Ore Deposits of the United States, 1933-
1967 (Graton-Sales Volume): New York, American Institute of Mining, Metallurgical, and Petroleum
Engineers, Inc., p. 681-705.
Tweto, O., Moench, R.H., and Reed, J.C., Jr., 1978, Geologic map of the Leadville 1° x 2° quadrangle,
northwestern Colorado: U.S. Geological Survey Miscellaneous Investigations Map 1-999, scale
1:250,000.
Unaldi, O., Aksoy, S., and Gullu, G., 2011, Cost-benefit analysis of gold mining with environmental
valuation methods: World Mining Congress and Expo.
USEPA, 2002, EPA Superfund record of decision: Bunker Hill mining and metallurgical complex OU3.
Technical Report EPA/ROD/R10-02/032.
USEPA, 2006a, Regional benefits analysis for the final section 316(b) Phase III existing facilities rule.
Office of Water (4303T). Report EPA-821-R-04-007. June 2006.
http://water.epa.gov/lawsregs/lawsguidance/cwa/316b/phase3/upload/2006_08_09_316b_phase3_ph3
docs_p3-rba-final-parti .pdf
USEPA, 2006b, National recommended water quality criteria: U.S. Environmental Protection Agency,
accessed April 29, 2016, at https://www.epa.gov/wqc/national-recommended-water-quality-criteria-
aquatic-life-criteria-table
Vaughan, W.J., and Russell, C.S., 1982, Valuing a fishing day: An application of a systematic varying
parameter model: Land Economics, p. 450-463.
Verhulst, P.F., 1838, Notice sur la loi que la population suit dans son accrossement: Corr. Math. Phys., v.
10, p. 113-121.
Wallace, A.R., 1993, Geologic setting of the Leadville Mining District, Lake County, Colorado: U.S.
Geological Survey Open-File Report 93-343, 24 p.
Wallace, K.J., 2007, Classification of ecosystem services: problems and solutions: Biological
Conservation, v. 139, p. 235-246.
Walton-Day, K., and Mills, T.J., 2015, Hydrogeochemical effects of a bulkhead in the Dinero mine
tunnel, Sugar Loaf mining district, near Leadville, Colorado: Applied Geochemistry, v. 62, p. 61-74.
Walton-Day, K., Rossi, F.J., Gerner, L.J., Evans, J.B., Yager, T.J., Ranville, J.F., and Smith, K.S., 2000,
Effects of fluvial tailings deposits on soils and surface and ground-water quality, and implications for
remediation - Upper Arkansas River, Colorado, 1992-96: U.S. Geological Survey Water-Resources
Investigations Report 99-4273, 108 p.
-------
Ward, K.M., and Dueld, J.W., 1992, Natural resource damages: Law and economics. John Wiley and
Sons, Inc.
Wildeman, T.R., and Schmiermund, R., (2004). Mining influenced waters: Their chemistry and methods
of treatment. In: Proceedings of the 2004 National Meeting of the American Society of Mining and
Reclamation and the 25th West Virginia Surface Mine Drainage Task Force. April 18-24, 2004.
Published by ASMR, Lexington, KY 40502. pp. 2001-2013.
Williams, J. S., and P. W. Bettoli, 2003, Net value of trout fishing opportunities in Tennessee tailwaters:
Fisheries Report 03-21.
Williamson, J.M., Thurston, H.W., and Heberling, M.T., 2008, Valuing acid mine drainage remediation
in West Virginia: ahedonic modeling approach: Annals of Regional Science, v. 42, p. 987-999.
Wilson, M.A., and Hoehn, J.P., 2006, Valuing environmental goods and services using benefit transfer:
the state-of-the art and science: Ecological Economics, v. 60, p. 335-342.
Whitehead, J.C., and Aiken, R., 2000, An analysis of trends in net economic values for bass fishing from
the National Survey of Fishing, Hunting, and Wildlife-Associated Recreation: East Carolina
University, Department of Economics, Greenville, NC, 27 p.
Woodling, J., Rollings, A., and Wilson, J., 2005, Biological monitoring report for the Eagle Mine
Superfund Site, Eagle County, Colorado for the period of April 1990 to April 2005: Report prepared
for the Hazardous Material and Waste Management Division, Department of Public Health and the
Environment, State of Colorado, 95 p.
Woodward, R.T., and Wui, Y., 2001, The economic value of wetland services: a meta-analysis.
Ecological Economics, v. 37, p. 257-270.
-------
Appendices
Appendix A: Review of Environmental (and Ecosystem Service)
Valuation Literature Pertaining to Mine Site Pollution
To employ the benefit transfer valuation technique, primary valuation studies must be located
that correspond to the 'policy' site in question. Given that the goal of this analysis is to value
changes in ecosystem services related to mine sites, the following section reviews the
environmental valuation literature that relates to mine site pollution. As described below, the
number of environmental valuation studies geared towards abandoned, proposed, or operating
mine sites is rather small. Each study from this narrow literature resource is highlighted and
summarized below. The purpose of this review is to uncover potential material for the
construction of a mine site pollution benefit transfer model.
Social Cost-Benefit-Analyses of Mine Remediation Schemes
A common theme of this small body of literature is social cost-benefit-analysis of various
remediation schemes. Randall et al. (1978) employed water treatment costs, fish restocking costs,
government established per-day recreation values, and visual disamenity valuation to estimate
the benefits of proposed tightening of state and federal regulations regarding the reclamation of
surface coal mines. While Randall et al. (1978) provided an instructive framework for (and
application of) environmental valuation of mine site pollution, it would be difficult to transfer the
benefit results to a policy site that did not have the same topography, geology, hydrology,
geochemistry, and ecology. This difficulty is because the authors accounted for specific details in
slope, sediment loading, host-rock generated acid mine drainage, recreational characteristics of
the area, and the visual impacts of surface mining of coal. A benefit transfer practitioner would
be hard pressed to demonstrate site correspondence between this study site and a Central
Colorado hard rock mining site. Information can be obtained from this study, but one would have
to delve deeply into the basic components such as mine waste management, water treatment, and
fish replacement practices to approximate transferable benefits.
Similarly, Michael and Pearce (1989) conducted a social cost-benefit-analysis of a remediation
project in northwestern England that reclaimed a large abandoned coal field. A residential area
was built around coal spoil heaps that caught fire, collapsed into gardens, blew dust, and which
also included mine shafts. The coal field was turned into an agricultural area with forested
footpaths and soccer fields. Michael and Pearce (1989) employed a contingent valuation to
ascertain the total economic value that the average household placed on the remediation. As with
Randall et al. (1978), this study is instructive. However, its results could only be transferred to an
equally flammable and dangerous policy site.
Button et al. (1999) proposed a framework for valuing remediation benefits, but did not produce
a benefit value estimate. Nonetheless, Button et al. (1999) is an important literature source
because it highlights the complexities, synergies, and difficulties associated with remediation of
mine site pollution. For example, Button et al. (1999) noted that the sum of the benefits from
-------
remediating many sources of pollution within a watershed will be larger than remediating the
parts. Also, Button et al. (1999) mentioned that remediation can be coupled with heritage
activities (such as highlighting a region's mining history) to augment the benefits of remediation.
Finally, Button et al. (1999) discussed the difficulty and cost of collecting site-specific economic
valuation information and suggested benefit transfer as the best alternative.
Farber and Griner (2000) employed a conjoint analysis, in conjunction with a random utility
model, to value various combinations of stream quality improvements. The two study sites were
in western Pennsylvania and could support fishing, boating, and hiking. Depending on policy site
correspondence (for stream quality improvement, recreational characteristics, and population),
Farber and Griner (2000) may be useful in benefit transfer for mine site pollution affecting
stream quality.
The studies of Damigos and Kaliampakos (2003a) and Damigos and Kaliampakos (2003b) were
derived from the same contingent valuation of the proposed remediation of an abandoned rock
quarry in Athens, Greece. This contingent valuation does not appear to be of sufficient quality to
transfer the benefit value. Also, it would appear to only correspond to other urban quarry sites.
Ahlheim et al. (2004) and Lienhoop and Messner (2009) both applied a rigorous contingent
valuation to a remediation scheme in East Germany that converted open pit coal mines into
recreational lake parks. It is easy to imagine using these studies as study sites for open-pit hard
rock mines in the United States that could economically be turned into lake parks. Such a
possibility would depend on site-specific hydrology and rock chemistry. However, a lake for
recreation would likely provide more economic benefits than a bare, abandoned open pit.
Mendes et al. (2007) created a framework to value the non-market economic benefits of
remediating an open pit copper-silver-gold mine and smelter site in Portugal. Mendes et al.
(2007) set the stage for a contingent valuation at the site, but like Button et al. (1999) they were
unable to achieve a benefit value result. Nonetheless, Mendes et al. (2007) is a good example of
how much preparation must be conducted for primary valuation and of how prediction of
physical impacts of the remediation is required to conduct a valuation.
Williamson et al. (2008) employed a hedonic study in the Cheat River watershed of West
Virginia and showed that being within a quarter mile of an acid mine drainage impaired stream
reduces home property value by $8,525 (2013$). Each of the studies above highlights the
difficulty of employing mine site valuations in an ecosystem service framework. The sites are
valued wholesale, and the value of a particular ecosystem service cannot be parsed from others.
The study by Randall et al. (1978) is the only exception to this rule. Site correspondence
requirements pose another major challenge because of the differences between coal sites, hard
rock sites, and remediation schemes. In other words, these studies are only useful for benefit
transfer at sites that correspond to the site and context in question.
Burton et al. (2012) used conjoint analysis to estimate what the public would be willing to accept
for reducing various bauxite mine remediation schemes around Perth, Australia. The focus was
-------
on timing and reductions in plant species, richness, wildlife habitat, and bird populations. Like
many of the previous studies, it is difficult to translate Burton et al. (2012) into a benefit transfer.
The benefits/losses being estimated were particularly site-specific, and the policy site would
have to have been an excellent match for a valid benefit transfer.
While not a remediation scheme, Neelawala et al. (2013) used a hedonic property value model to
estimate the marginal willingness to pay to be farther from mining and smelting operations in
Queensland, Australia. The result is that households were willing to pay $13,703 (2013$) to be
one km farther from the pollution source when they are within a 4-km radius (Neelawala et al.,
2013).
In summary, valuable information can be gleaned from these social cost-benefit analyses of mine
site remediation, but a benefit transfer model for the purpose of this analysis cannot be built on
these studies. The studies by Farber and Griner (2000) and Williamson et al. (2008) may prove
useful for benefit transfer of the value of various stream-quality improvements, although it is not
clear exactly which ecosystem/environmental services are being valued. The studies by Ahlheim
et al. (2004) and Lienhoop and Messner (2009) could provide benchmarks for the value of
turning open-pit mines into recreational lakes, but that is not common practice in the United
States.
Social Cost-Benefit-Analyses of Proposed Mine Sites
The second common theme within this literature is social cost-benefit analysis of proposed mine
sites. Trigg and Dubourg (1993) took a hypothetical and expert opinion approach to a social cost-
benefit analysis of a proposed coal strip mine in North Staffordshire, England. Trigg and
Dubourg (1993) surveyed real estate agents for their expert opinion on how much property
values would decrease if the proposed mine was developed. The average estimate was roughly
30 %. This approach is certainly not up to National Oceanic and Atmospheric Administration
(NOAA) contingent valuation standards, and it would be dubious to transfer the results.
The study by Damigos and Kaliampakos (2006) was a social cost-benefit analysis of a proposed
open pit gold mine in Greece, which was funded by the project's owner. Damigos and
Kaliampakos (2006) indiscriminantly used results from many of the previously mentioned
studies for benefit transfer (Damigos and Kaliampakos, 2003a, Randall et al., 1978, Trigg and
Dubourg, 1993). While Damigos and Kaliampakos (2006) provided a framework for social cost-
benefit analysis of proposed mining projects, it did not appear that the policy site (an open-pit
gold mine in Greece) corresponds to the study sites from which the benefits are transferred (coal
fields and urban quarries). For example, Damigos and Kaliampakos (2006) employed the 30%
property value reduction from Trigg and Dubourg (1993). Considering that Trigg and Dubourg
(1993) did not identify which environmental goods (for example, view, air quality, and
congestion) were responsible for the decrease in property value, it seems difficult for the
approach of Damigos and Kaliampakos (2006) to directly transfer these values. Each benefit
transfer within Damigos and Kaliampakos (2006) was of this nature. The study site was related,
-------
but it is unclear how similar it was to the policy site or how valid the benefit transfer was. The
research by Unaldi et al. (2011) was a study that essentially duplicated the efforts of Damigos
and Kaliampakos (2006), but it was for a generic gold deposit in Turkey. In short, the social cost-
benefit-analyses of proposed mine sites were of low quality and should not be used.
Natural Resource Damage Assessments Related to Mine Sites
The final sources of valuation studies related to mining are natural resource damage assessments
(NRDA). Estimates of economic benefit values are sparse in the NRDA literature. EPA's
CERCLA cleanup of the Eagle Mine site in Gilman, Colorado prompted litigation regarding
cleanup costs, liability, and monetary compensation. Competing expert witness valuation
analyses from the trustee (Rowe and Schulze, 1985) and the defendant (Ward and others, 1992)
yielded completely different estimates. The plaintiffs expert witness conducted a contingent
valuation at the local, county, and state level where the focus of the survey was the subject's
willingness to pay to clean up the entire site (Rowe and Schulze, 1985). Once the value was
elicited, the survey asked respondents what percentage of their willingness to pay value was
represented by use value, non-use value, and existence value (Rowe and Schulze, 1985). The
defendant's expert witness, on the other hand, estimated replacement cost values to cover
contaminated areas with topsoil and to redrill wells for residential water (Ward et al., 1992).
These replacement values had little to do with the benefits that society has foregone as a result of
pollution, or society's willingness to pay to clean up the site.
The Coeur D'Alene, Idaho NRDA provided many cost estimates, but no benefits were estimated
(USEPA, 2002). Similarly, a preliminary estimate of damages at Leadville (IEc., 2006)
employed a habitat equivalency analysis (HEA) that valued damage by the cost that would be
required to repair the habitat to its original condition. The main shortcoming of IEc. (2006) (as
with all habitat equivalency analyses) was that its economic assessment of damages failed to
incorporate how the changes in these ecosystem services affected human well-being. Instead,
IEc. (2006) calculated damages using abatement costs required to return the ecosystem services
to their baseline condition. As a result, HEA was unlikely to result in efficient outcomes because
there was no balancing of costs and benefits of remediation. HEA was also the preferred
approach for the Holden Mine site in Washington, the Southeast Missouri Lead Mining District,
and the Blackbird Mine in Idaho.
Like many of the valuations above, these natural resource damage assessments are not useful for
benefit transfer. The contingent valuation from Rowe and Schulze (1985) valued the whole site,
rather than ecosystem services. Ward et al. (1992) estimated replacement values, which are the
minimum value for the ecosystem service in question. The Coeur D'Alene, Idaho NRDA did not
estimate any benefits. The habitat equivalency assessment for Leadville was a replacement cost
approach, not a technique for the optimization of social resources. An ideal literature portfolio
would provide numerous environmental valuation studies of mine site pollution. This literature
review demonstrates the limited number of studies on the subject.
-------
References for Appendix A
Ahlheim, M., Fror, 0., Lehr, U., Wagenhals, G., and Wolf, U., 2004, Contingent valuation of mining land
Reclamation in East Germany: Inst, fur Volkswirtschaftslehre, Univ.
Burton, M., Jasmine Z.S., and White, B., 2012, Public preferences for timeliness and quality of mine site
rehabilitation. The case of bauxite mining in Western Australia. Resources Policy: v. 37, p. 1-9.
Button, K., Stough, R., Arena, P., Comp, A., and Casper, M., 1999, Dealing with environmental legacy
effects: the economic and social benefits of acid mine drainage remediation: International Journal of
Environment and Pollution, v. 12, p. 459-475.
Damigos, D., and Kaliampakos, D., 2003, Assessing the benefits of reclaiming urban quarries: a CVM
analysis: Landscape and Urban Planning, v. 64, p. 249-258.
Damigos, D., and Kaliampakos, D., 2006, The "battle of gold" under the light of green economics: a case
study from Greece: Environmental Geology, v. 50, p. 202-218.
Farber, S., and Griner, B., 2000, Valuing watershed quality improvements using conjoint analysis:
Ecological Economics, v. 34, p. 63-76.
IEc (Industrial Economics), 2006, Upper Arkansas River basin natural resource damage assessment:
Preliminary estimate of damages: Technical Report GS-10F-0224J. US Fish and Wildlife Service.
Lienhoop, N., and Messner, F., 2009, The economic value of allocating water to postmining lakes in East
Germany: Water Resources Management, v. 23, p. 965-980.
Mendes, I., Sardinha, I., and Milheiras, I. 2007, The social and economic value of the S. Domingos
abandoned mine rehabilitation projects: The case of S. Domingos mine: Contributions of Corporate
Social Responsibility to Sustainable Development.
Michael, N., and Pearce, D. 1989. Cost benefit analysis and land reclamation: a case study. LEEC Paper-
London Environmental Economics Centre, International Institute for Environment and Development.
Neelawala, P., Wilson, C., and Athukorala, W., 2013, The impact of mining and smelting activities on
property values: a study of Mount Isa city, Queensland, Australia: Australian Journal of Agricultural
and Resource Economics, v. 57, p. 60-78.
Randall, A., Grunewald, O., Johnson, S., Ausness, R., and Pagoulatos, A., 1978, Reclaiming coal surface
mines in central Appalachia: a case study of the benefits and costs: Land Economics, v. 54, p. 472-
489.
Rowe, R.D., and Schulze, W.D., 1985, Economic assessment of damage related to the Eagle Mine
facility: Energy and Resource Consultants.
Trigg, A.B., and Dubourg, W.R., 1993, Valuing the environmental impacts of opencast coal mining in the
UK: The case of the Trent Valley in North Staffordshire: Energy Policy, v. 21, p. 1110-1122.
Unaldi, O., Aksoy, S., and Gullu, G., 2011, Cost-benefit analysis of gold mining with environmental
valuation methods: World Mining Congress and Expo.
USEPA, 2002, EPA Superfund record of decision: Bunker Hill mining and metallurgical complex OU3.
Technical Report EPA/ROD/R10-02/032.
-------
Ward, K.M., and Dueld, J.W., 1992, Natural resource damages: Law and economics. John Wiley and
Sons, Inc.
Williamson, J.M., Thurston, H.W., and Heberling, M.T., 2008, Valuing acid mine drainage remediation
in West Virginia: ahedonic modeling approach: Annals of Regional Science, v. 42, p. 987-999.
-------
I || <'h""i II ¦ -I »i iin, II iin i i II <'h<'in 11 vh i 'i hi: -1 'Ini, „ il
Additional Ecosystem Services Impacted by Mine Site
Pollution
Soil Quality for Human Health: Lead
Soil quality at legacy mine sites has often been contaminated by pollutants in mine tailings and
from smelting operations. Historical mine tailings have been scattered by wind and water over
the decades. Many historical smelting operations did not capture pollutants from their smoke
stacks, so they dispersed onto the surrounding communities6. When soil is contaminated with a
pollutant such as lead, it can become a pathway to children via unintentional soil ingestion. For
example, EPA models concluded that children in Leadville, Colorado who had backyard soil
lead levels greater than 500 ppm were 8.4 times more likely to have blood lead levels at or above
10 |ig/dL (CDH, 1990, p. 33), which causes health defects and triggers intervention (Gould et al.,
2009). When tested, these models over-estimated the connection between backyard soil lead
level and blood levels. Nonetheless, one can imagine that high levels of lead in the surrounding
environment could conceivably raise blood lead levels. To deal with this problem, Leadville
initiated a voluntary program called Kid's First to monitor children's blood lead level.
Elevated blood lead levels in children have been linked to IQ loss, attention deficit hyperactivity
disorder, the need for special education, and criminal behavior (Gould et al., 2009). Gould et al.
(2009) have summarized the medical literature on the effects of elevated blood lead level and the
monetization of their effects. Pairing Gould et al. (2009) with population blood lead level data
provides a start in modeling the value of soil quality improvements at a legacy site. Table B. 1
uses Leadville as an example of this process.
Several problems with this approach are enumerated below. First, the lead contamination in the
children may have nothing to do with lead from the surrounding environment. Blood lead levels
for children in Leadville were often on par with blood lead levels for children in inner cities. The
argument was made by Leadville residents (and the Potentially Responsible Parties [PRPs]) that
high blood lead levels could have been the result of lead paint. Second, a molybdenum mine that
employed many Leadville residents closed at the same time that EPA designated the area as a
Superfund site. This mine closure resulted in significant emigration from the town, especially for
families employed by the mine. Many children from Leadville were never included in the
baseline population. Third, it is not possible to determine how much of the improvement in blood
lead level was due to remediation vs. behavioral changes. For example, if children washed their
hands after playing outside and before eating, that may have been just as effective as physically
removing the mine waste and tailings. In other words, there is no direct link between a change in
6 Soil quality contamination at prospective mine sites is expected to be less pronounced
due to modern operators active management of tailings and smelting operations.
-------
the ecosystem service and the impact to be valued. Finally, the medical literature appears
divided on the issue of whether lead in soil contributes to lead absorption by children
(Kimbrough and Krouskas, 2012).
Table B.l. Impact of Elevated Blood Lead Levels in Children from Leadville, Colorado (2013$)
Blood Lead
(l-ig/dL)
Medical
Cost3
IQ Loss per
lag/dL1
Lost
Earnings per
Child3
Cost of
Special
Education3
Number of
Children
(1991)b
1991 Cohort
Total
2-10
$0
0.513
$63,608
$0
285
$18,100,000
10-15
$86
0.19
$49,080
$0
25
$1,200,000
15-20
$86
0.19
$68,712
$0
4
$300,000
20-25
$1,400
0.11
$51,147
$49,823
0
$0
25-45
$1,400
0.11
$79,562
$49,823
0
$0
45-70
$1,549
0.11
$130,709
$49,823
0
$0
>70
$3,995
0.11
$181,856
$49,823
0
$0
Total




314
$19,600,000
Source:a Gould et al. (2009),b EPA (1996)
Air Quality for Human Health: Particulate Matter
Air quality can also be affected at mine sites by suspended particulate matter. Smith and Huang
(1995) used a hedonic property value model to estimate the value of changes in total suspended
particulate matter. The median value of $52.39 (2013$) per household to reduce total suspended
particulates by 1 mg/m3 was estimated (Smith and Huang 1995). Similarly, Vassanadumrongdee
et al. (2004) provided a starting place for valuing the short-term health effects of air pollution
such as coughing, congestion, and asthma attacks.
Air Quality for Human Health: Mercury Emissions
A wide review of the literature suggests that IQ loss due to the consumption of mercury-
contaminated fish is the only properly monetized damage estimate relating to mercury emissions
(Sundseth et al., 2010). In the early 2000's, regulations were proposed to require coal-fired power
plants in the United States to abate mercury emissions from burning coal. These regulations
spawned attempts to weigh the costs and benefits of mercury emission abatement. Several
studies were conducted to map the chain of mercury emission, mercury deposition, conversion to
methylmercury, methylmercury bioaccumulation, consumption of contaminated fish, ensuing
-------
impact on fetal cognitive functioning, and loss of IQ (Hylander and Goodsite, 2006, Mergler et
al., 2007, Rice and Hammitt, 2005, Seigneur et al., 2004, Spadaro and Rabl, 2008, Sundseth et
al., 2010, Swain et al., 2007, Trasande et al., 2005, UNEP, 2013). These research efforts linked
atmospheric, oceanic, chemical, biological, and economic models - a titanic task fraught with
complexity. In short, these analyses focused on the roughly 2 % of elemental mercury (Hg°) that
becomes methylated, is ingested from fish fillets7, and affects the fetal nervous system.
From this literature, Trasande et al. (2005), Rice and Hammitt (2005), and Spadaro and Rabl
(2008) provided estimates of the translation from a quantity of mercury emission to a dollar
amount of lost earnings due to fetal IQ loss. The goal of Trasande et al. (2005) was to estimate
the economic costs of fetal neurodevelopmental impacts attributable to mercury emissions from
American power plants. To achieve this, Trasande et al. (2005) combined an environmentally
attributable fraction (EAF) model with national blood mercury prevalence data from the Centers
for Disease Control and Prevention. They found that between 316,588 and 637,233 children each
year have cord blood levels greater than the 5.8 |ig/L level associated with loss of IQ. The 5.8
|ig/L of cord blood level serves as the neurotoxicity threshold for all estimates from that study.
Trasande et al. (2005) estimated damages to the American economy due to IQ loss in an annual
birth cohort from mercury deposited in the United States from three sources. First, global
anthropogenic emissions are assumed to deposit 87,000 kg of mercury in the United States.
Assumptions regarding cord/maternal Hg blood level ratios and linear/logarithmic IQ decrements
produced a range of estimated damages from $2.9 billion (B) to $59.2B (2013$). Within this
range, Trasande et al. (2005) recommended a cord/maternal ratio of 1.7 and a logarithmic model
that resulted in a recommended value of $11.8B (2013$) for damages from mercury deposited in
the United States from global anthropogenic sources. Second, American anthropogenic
emissions were assumed to deposit 52,200 kg of mercury in the United States. The range of
estimated damages is $0.5B-$21.4B (2013$), and the recommended value is $4.2B (2013$).
Finally, estimates from American anthropogenic emissions from coal-fired power plants
provided a range $0.1B-$8.8B and a recommended value of $1.8B. Averaging the low,
recommended, and high estimates of damages per kg results in estimates of $11,000, $60,000, and
$193,000, respectively.
In a similar study, Rice and Hammitt (2005) estimated the economic benefits of greater control
of mercury emissions from coal-fired power plants in the United States. Mercury emissions
reduction was assumed to have a linear and proportional decrease in methylmercury
7 While methylmercury primarily concentrates in fish organs, methylmercury concentrations in the
muscle tissue are approximately 50% of liver concentrations (Oliveira Ribeiro et al., 1999). The USGS
found that 27% of fish sampled in US streams had skinless-fillet methylmercury concentrations higher than
the EPA human-health criterion (Scudder, 2010). Additionally, the mean methylmercury concentration of
skinless-fillets from the 59 fish sampled in basins with gold mining exceeded the EPA human-health
criterion (Scudder, 2010, p. 12).
-------
concentrations in fish. Changes in deposition rates were based on regional deposition modeling
from the EPA's analysis of the Clear Skies Initiative, under which power plants reduced mercury
emissions from 49,000 kg/year to either 26,000 kg/year or 15,000 kg/year. Human exposure to
methylmercury was modeled through commercial and non-commercial harvest of fish. Rice and
Hammitt (2005) used dose-response functions from recent methylmercury epidemiological
studies and data on fish consumption from the FDA to estimate damages of mercury deposition.
The estimates provided by Rice and Hammitt (2005) that are useful for this analysis are two
estimates of damages to the American economy due to IQ loss in an annual birth cohort from
mercury deposited in the United States by all sources. The first estimate of $4.2B assumed a
neurotoxicity threshold of maternal methylmercury intake greater than 0.1 |ig/kg of fish per day.
The second estimate of $26.9B assumed that there was no neurotoxicity threshold. Dividing
these estimates by the assumed 124,300 kg deposited in the United States by all sources yields
damages estimates of $33,000 and $208,000, respectively.
Finally, Spadaro and Rabl (2008) used worldwide average methylmercury doses from fish to
calculate global damages from total (anthropogenic and non-anthropogenic) emissions. Spadaro
and Rabl (2008) defined a comprehensive transfer factor for ingestion of methylmercury as a
ratio of global average dose rate (2.4 |ig/day) and global emission rate (6,000t/year). The
immediate problem with this approach is that methylmercury damages primarily come from the
high doses ingested by those consuming large amounts of fish. Using a global average dose rate
smooths the high doses out across the population to the point where they appear to have no
effect. Additionally, Spadaro and Rabl (2008) scaled damages based on income. For these
reasons, the estimates from Spadaro and Rabl (2008) are not incorporated in the averages for the
recommended value in Table B.2.
The IQ loss estimates from Trasande et al. (2005) and Rice and Hammitt (2005) were normalized
to reflect the same value per IQ point of Spadaro and Rabl (2008). The estimates were then
inflated to 2013$ using the CPI. One of the most influential factors in the calculation of lost
earnings from methylmercury poisoning is the incorporation of a neurotoxicity threshold. The
neurotoxicity threshold reduces the estimate of lost earnings by ruling out the large contingent of
infants who have trace amounts of MeHg in their blood. Current scientific understanding
indicates that a neurotoxicity threshold exists. Therefore, the recommended value is an average
of the non-outlier estimates that have a neurotoxicity threshold. Table B.2 details the valuations
and their conversion into lost IQ per kilogram of mercury released. The recommended value for
lost lifetime earnings due to IQ loss from 1 kg of vaporized mercury is $53,000.
The main weakness of this approach is the assumption of constant marginal impacts from each
kilogram of mercury emitted into the air. By dividing the economy-wide damage estimates from
Trasande et al. (2005) and Rice and Hammitt (2005) by the kilograms of mercury deposited, this
analysis assumes that the effect of each kilogram of mercury deposited is a linear function.
Future research needs to be conducted in the same vein as Spadaro and Rabl (2008) to more
-------
accurately estimate the marginal impacts from each kilogram of mercury emitted into the air. In
absence of such research, this analysis provides a strong starting point for comparing the
environmental impacts of various forms of gold mining.
Table B.2: Valuation of Environmental Damage Due to 1 kg Release of Mercury into the
Atmosphere	
Trasande et al. (2005) Cost of American Anthropogenic Coal Power Plant Hg Emissions Deposited in
US	
48,000
$1.8 billion
$27,000*
Mercury deposited in the United States from anthropogenic sources (kg)
Damages to American economy due to IQ loss in annual birth cohort3 (2013$)
Lost lifetime earnings due to IQ loss from 1 kg Hg air release (2013$)	
Trasande et al. (2005) Cost of American Anthropogenic Hg Emissions Deposited in US
Mercury deposited in United States from anthrogenic sources (kg)
Damages to American economy due to IQ loss in annual birth cohort3 (2013$)
Lost lifetime earnings due to IQ loss from 1 kg Hg air release (2013$)	
52,200
$4.2 billion
$57,000*
Trasande et al. (2005) Cost of Global Anthropogenic Hg Emissions Deposited in the US
Mercury deposited in United States from anthrogenic sources (kg)
Damages to American economy due to IQ loss in annual birth cohort3 (2013$)
Lost lifetime earnings due to IQ loss from 1 kg Hg air release (2013$)	
87,000
$11.8 billion
$97,000*
Rice and Hammitt (2005) Cost of Global Anthropogenic Hg Emissions Deposited in the US	
Mercury deposited in United States from anthrogenic sources (kg) 124,300
Damages to American economy due to IQ loss in annual birth cohort3 (2013$) $4.2 billion
Lost lifetime earnings due to IQ loss from 1 kg Hg air release (2013$)	$33,000*
Rice and Hammitt (2005) Cost Estimate Without Neurotoxicity Threshold
Mercury deposited in United States from anthrogenic sources (kg)
Damages to American economy due to IQ loss in annual birth cohort3 (2013$)
Lost lifetime earnings due to IQ loss from 1 kg Hg air release (2013$)	
124,300
$26.9 billion
$208,000
Trasande et al. (2005) Cost Estimate with Alternative Linear Model
Mercury deposited in United States from anthrogenic sources (kg)
Damages to American economy due to IQ loss in annual birth cohort3 (2013$)
Lost lifetime earnings due to IQ loss from 1 kg Hg air release (2013$)	
87,000
$44.5 billion
$366,000**
Global Estimate from Spadaro and Rabl (2008)
Lost lifetime earnings to global economy due to IQ loss in annual cohort3 (2013$)
Lost lifetime earnings to global economy due to IQ loss in annual cohortb (2013$)
$1,818**
$4,056
Recommended Value: Average of non-outlier threshold estimates
Average of all estimates
Average of all threshold estimates
3 Indicates a neurotoxicity threshold is assumed
b Indicates no neurotoxicity threshold is assumed
*	Indicates the value is included in the average of non-outlier threshold estimates
*	* Indicates an outlier value
$53,000***
$127,000
$115,000
The following example provides a concrete context for the model above. Gold Quarry is an open
pit gold mine in Nevada. The primary impact to the environment resulted from airborne mercury
emissions due to smelting of Gold Quarry ore. In 2006, the EPA and the state of Nevada
instituted a mercury emissions control program. Before the program (in 2005) 329.4 kg were
-------
emitted from the Gold Quarry smelter. This number was reduced to 48.3 kg in 2013. Table B.3
shows a valuation of the environmental damage per troy ounce of gold produced. This value was
$35 and $5 for the years 2005 and 2013, respectively.
Table B.3. Valuation of Environmental Damage per Ounce from the Gold Quarry Mine in Nevada
Environmental Damage per Ounce Calculation	
Gold Quarry Mine 2005	
Loss in American lifetime earnings due to 1 kg of mercury emissions (2014$)
Gold Quarry's airborne mercury emissions in 2005 (kg)a
Gold Quarry's portion of damages due to IQ loss in 2005 (2014$)
Gold Quarry's Gold equivalent production in 2005 (Troy oz)b
Environmental damage per ounce (2005)	
Gold Quarry Mine 2013
Gold Quarry's airborne mercury emissions in 2013 (kg)a
48.3
Gold Quarry's portion of damages due to IQ loss in 2013 (2014$)
$2,500,000
Gold Quarry's Gold equivalent production in 2013 (Troy oz)b
500,000
Environmental damage per ounce (2005)
$5
Source: a NMCP Annual Emissions Reporting,b Nevada Division of Minerals
View from a Residence
$53,000
329.4
$17,500,000
500,000
$35
In addition to the economic cost to residential consumers, the loss of service for municipal water
also includes business losses. Aubuchon and Morley (2013) calculated business losses as the
forgone business value, measured by industry level Gross Domestic Value, due to service
disruption. To do this, they took into consideration the variation in operating capacity among
industries in the event of the loss of water. For example, if water services were disrupted health
care and social assistance would shut down, whereas transportation and warehousing are resilient
and could continue to operate at almost the same capacity. Aubuchon and Morley (2013) used
resilience factors for each industry from two studies, one from the Applied Technology Council
(1991) and the other from Chang et al. (2002). A resilience factor is the percent of capacity an
industry could operate in the absence of water.
Using these resilience factors, Aubuchon and Morley (2013) calculated the economic loss using
equation B.6.1:
^GDPi^l-n)	(B.6.1)
365 * population4
This formula calculates a per capita daily loss for industry i, with an industry specific resilience
factor (r) and then sums across all industry. Aubuchon and Morley (2013) went on and calculated
a state level and a population weighted per capita per day business economic loss (see Aubuchon
and Morley [2013 Viewsheds can be affected negatively by proposed mines or positively by the
remediation of legacy sites. These changes often affect the value of residential property that has a
view of them. A benefit transfer model could be constructed by beginning with the literature
-------
reviews in Bourassa et al. (2003) and Walls et al. (2015). General conclusions are that the effect
of access to the scenic area must be parsed from the effect of the view of the scenic area, that a
mountain view increases the value of a property by 6%, that a forest view increases the value of
a property by 5%, and that a view of roads, railways, and industrial parks have various negative
impacts on property value.
Wetland, Open Water, Shrubland, Grassland, and Terrestrial Habitat
The following examples of natural land cover valuation require careful attention to the issue of
double-counting. Loomis and Richardson (2008) employed a wetland valuation meta-analysis
(Brander and Florax, 2007) for use in benefit transfer. Brander and Florax (2007) analyzed
European and North American wetland valuation studies that focused on flood prevention, water
quality, water quantity, fishing, birdwatching, habitat, and storm drainage. Loomis and
Richardson (2008) built a meta-regression model benefit transfer function that was used to value
wetlands based on their size, location, and the ecosystem services that they provided.
Similarly, Ingraham and Foster (2008) conducted a meta-analysis of valuation studies on the
indirect uses for natural land cover. Examples of such uses are carbon sequestration, disturbance
prevention, freshwater regulation and supply, habitat provision, and nutrient removal and waste
assimilation. Ingraham and Foster (2008) conducted this evaluation for five separate land
classifications: open water, forest, shrubland, grasslands, and wetlands. Ingraham and Foster
(2008) appeared to estimate per acre valuations of the land classifications, but they did not
explicitly enumerate the values.
Finally, Borisova-Kidder (2006) provided a valuation study for terrestrial open space and habitat.
Borisova-Kidder (2006) used 11 studies with 23 observations to conduct a meta-analysis of the
literature valuing terrestrial open space and habitat. The primary studies evaluated in this meta-
analysis were too disparate and the sample size too small to be incorporated in the current
analysis. Nonetheless, many mining sites affect terrestrial habitat and this study should be
highlighted as a good place to start for a per acre value of terrestrial habitat.
Municipal and Household Intake Water Quality
Intake water quality refers to the quality of groundwater or surface water brought into a water
system. For municipalities, the ecosystem service is the quality of raw water that enters the
treatment system. For households, the ecosystem service is the quality of well-water being
drawn into the home. In this context, quality is considered in terms of acceptability for use
(NRC, 1997, pp. 31). Therefore, the ecosystem service has value if the water is of high enough
quality that it can be treated for use. When the quality is too low for the water to be treated, its
value falls to zero. Aubuchon and Morley (2013) supported this binomial endpoint with a
thorough ecosystem service valuation. In contrast, the valuation literature was mute regarding
other possible methods of ecosystem service valuation for drinking water, such as changes in
water treatment cost as a function of continuous changes in contaminant concentrations.
-------
Municipal and household intake water quality is valued to account for the impact of mine
pollution events that temporarily make intake water quality so poor that it cannot be treated.
Such events have occurred at mine sites. For example, in 2014 a chemical used to wash coal was
accidentally spilled into the Elk River in West Virginia. Water treatment plants were
overwhelmed and 300,000 residents in nine counties were left without potable water. Similarly,
an impoundment failure at the Mount Polley tailings storage facility in British Columbia released
25 million m3 of water and tailings into Hazeltine Creek, Quesnel Lake, and Quesnel River. A
drinking water ban was imposed on approximately 150 households for 9 days.
Valuing household/municipal water quality first requires determining whether the water is
suitable for household/municipal use and then values the benefits (costs) of having (losing) that
ecosystem service. Contamination of household/municipal water from mining often involves an
increased risk to high concentrations of toxic elements, for example, West Virginia's coal
separation chemical spill of January 2014 (Plumber, 2014). Because contamination makes water
unusable, as opposed to merely raising treatment costs, the value of lost service is the best
approach. Estimating the foregone value due to water service disruption involves three steps: 1)
estimate benefit loss for residential consumers, 2) estimate benefit loss from the affected
business and commercial consumers, and 3) add these values together for a total economic loss.
This method was used by Aubuchon and Morley (2013), which this project will follow for water
service valuation. The rest of this section describes the details of the method used by Aubuchon
and Morley (2013).
This valuation technique requires an initial estimation of household demand for water at various
prices. Once a demand curve has been estimated, the area under the curve provides the
willingness to pay for household water service for a given quantity of water.
In the past, FEMA used a meta-analysis of studies that estimated the price elasticity of demand
for water (Dalhuisen et al., 2003) to evaluate the benefits of creating more secure water supplies
for municipalities and households. Aubuchon and Morley (2013) built on FEMA's method to
create a benefit transfer tool that estimates the cost of losing water service for residential
customers and businesses. Their results for U.S. residential customers based on per capita per
day (PCPD) consumption are presented in Table B.6.5.
-------
Table B.6.5. Impact to Residential Consumers, PCPD (2013$a)
Per Capita Per Day Consumption (gal)
Elasticity of Demand

-0.41
-0.35
-0.26
172
$57
$147
$2,248
98
$24
$40
$266
Population Weighted
$26
$47
$402
Current Value
Recommended Value = $158

a Includes the FEMA cost for Basic Water Requirements (6.6 gal @ $1.85/gal); Source: Aubuchon and
Morley (2013)
The values in the left column of B.6.5 represent different assumptions about the amount of water
that is consumed PCPD. The first row, 172 gallons PCPD, is an estimate used by FEMA. The
second row, 98 gallons PCPD, comes from an estimate from the USGS, and the third row is a
state level population weighted PCPD. The fourth row is a recommended value. Three estimates
of the cost of losing service PCPD are given for each consumption level based on different
assumptions of demand price elasticity. These elasticity values are represented in the top row.
Higher per capita consumption rates result in higher losses. Additionally, when demand is more
elastic (flatter demand curve), the benefit loss that people suffer from losing service is reduced.
However, for the purposes of this study the author's recommended value of $158 (2013$) will be
used. This is the average value of the three population weighted estimates.
Their results for U.S. total, state level, and population weighted per capita per day business
economic loss are shown in Table B.6.6, converted to 2013$, for both sets of resilience factors.
Using this process, the authors' recommended value is $57 per person per day for loss of water
for business uses. This is the average of the two estimates in the population-weighted column.8
Table B.6.6. Impact to Business Economic Activity, PCPD (2013$)

U.S. Total
State Mean
State Mean,
Population
Weighted
ATC-25(1991) Resilience Factors
$43
$42
$43
Chang et al. (2002) Resilience Factors
$71
$70
$71
Current Value
Recommended Value = $57

Source: Aubuchon and Morley (2013)
Combining the PCPD business loss and the residential PCPD economic loss, the authors
recommended a total economic impact of $215 per person per day for loss of municipal water.
8 The population-weighted estimates were considered by Aubuchon and Morley (2013) to
be the most accurate approach. See their paper for explicit formulas.
-------
This value is simply the sum of the two prior recommended values. Table B.6.7 summarizes the
possible combinations of total economic impact for business and residential based on various
assumptions of water consumption, resilience factors, and demand elasticity. Additional
summary statistics are provided at the end of Table B.6.7.
Table B.6.7. Total Economic Impact, PCPD (2013$a)

PCPD Consumption (gal)
Elasticity of Demand
-0.41 -0.35
-0.26
U.S. Total, ATC-25
172
$100
$191
$2,258

98
$67
$84
$310

USGS, Population Weighted
$69
$90
$444
U.S. Total, Chang et al. (2002)
172
$128
$219
$2,287

98
$95
$112
$338

USGS, Population Weighted
$97
$118
$473
State Mean, ATC-25
172
$98
$190
$2,257

98
$66
$82
$309

USGS, Population Weighted
$68
$88
$442
State Mean, Chang et al. (2002)
172
$127
$217
$2,257

98
$94
$111
$337

USGS, Population Weighted
$96
$117
$472
State Mean, Population Weighted,
172
$99
$191
$2,258
ATC-25
98
$67
$83
$310

USGS, Population Weighted
$69
$89
$443
State Mean, Population Weighted,
172
$128
$218
$2,286
Chang et al. (2002)
98
$95
$112
$338

USGS, Population Weighted
$97
$118
$473
Current Value
Recommended Value = $215



Mean

$93
$135
$1,016
Median

$96
$114
$458
Standard Deviation

$21
$53
$912
Minimum

$66
$82
$309
Maximum

$128
$219
$2,287
a U.S. and State Level (GDP and Consumption) Totals
Source: Aubuchon and Morley (2013)
Additional components of household/municipal water use value are supply reliability and
probability of contamination. Supply reliability refers to the value tied to the consistency of
being able to turn on a sink and have running water. Examples of the valuation of supply
-------
reliability are Howe et al. (1994), Griffin and Mjelde (2000), Koss and Khawaja (2001), and
Thorvaldson et al. (2010). Probability of contamination refers to the value of reducing the chance
of spoiling the water beyond a potable quality. Work on the value of changes in the probability
of contamination is best encapsulated by Poe et al. (2001). Although these two issues are
included in the value of a clean water ecosystem service, they are a second order consideration
when compared to complete water disruption and the current literature does not support their use
in benefit transfer (Poe et al. 2001).
Drinking Water and Groundwater: An Incremental Approach
Drinking water can be affected by excessive runoff at legacy sites or spills from operating mines.
Gorlach and Interwies (2003) summarized the drinking water literature up to 2003. Much of this
literature comprised averting behavior studies that evaluate the costs associated with poor
drinking water quality (Abdalla, 1990, Abdalla et al., 1992, Collins and Steinback, 1993,
Harrington et al., 1989, Laughland et al., 1993). Although these studies are for pollutants
unrelated to mining, they applied to the broad issue of contamination. More recent studies
relating to drinking water have been conducted in Brazil, Pakistan, and Nicaragua (Casey et al.,
2006, Khan et al., 2010, Vasquez et al., 2012). Construction of a benefit transfer model to value
drinking water quality would separate the studies into a group that focuses on safe/unsafe
drinking water and a group that focuses on percentage changes in quality.
Metal contamination of household and municipal water by mine sites can often be treated. In
this case, the additional costs of treatment should be taken into account. However, these costs do
not equate to the value of intake water quality because costs depend on more than the ecosystem
service of water quality, for example, water treatment plant specifications. This analysis is
unable to connect increased metal contamination with increased treatment cost due to several
factors. First, water chemistry and metal contamination have complex interactions, so
straightforward economic analyses of treatments costs for various contaminants could not be
found. Second, operating mine sites must adhere to regulations concerning the quality of the
water that they discharge. This limitation prevents situations where a municipality or household
routinely incurs additional treatment costs as a result of mining operations. Conversely,
households and municipalities are unlikely to locate water intakes in streams polluted by
abandoned mine runoff. Therefore, cases where treatment costs are reduced as a result of an
abandoned mine cleanup have proven elusive. This dearth of information on treatment costs
encourages the approach employed above.
Groundwater can also be affected by legacy sites and operating mines. Gorlach and Interwies
(2003) summarized the valuation literature regarding ground drinking water. This body of
literature is composed of contingent valuation studies, avoided treatment cost studies, and
replacement cost studies. Boyle et al. (1994) and Poe et al. (2001) conducted meta-analyses of
the contingent valuations regarding groundwater. Both Boyle et al. (1994) and Poe et al. (2001)
concluded that the groundwater contingent valuation literature produces defensible values.
-------
However, both recommend against using the meta-analysis for benefit transfer. A groundwater
valuation benefit transfer model, which is desperately needed in the unconventional oil and gas
development debate, could be constructed by gathering groundwater contingent valuation studies
conducted after Poe et al. (2001) and then updating the analysis using updated techniques for
benefit transfer error reduction.
Household and Municipal Water: Supply Reliability
Additional components of household and municipal water use value are supply reliability and
probability of contamination. Supply reliability refers to the value of having a consistent supply
of water. Starting places for the valuation of supply reliability are Howe et al. (1994), Griffin and
Mjelde (2000), Koss and Khawaja (2001), and Thorvaldson et al. (2010). Probability of
contamination refers to the possibility of having a water supply contaminated to the point that the
water is unusable. Work on the value of changes in the probability of contamination is best
captured in Poe et al. (2001). Although these two issues are included in the value of a clean water
ecosystem service, they are a second-order consideration when compared to complete water
disruption and the current literature does not support their use in benefit transfer (Poe et al.,
2001).
Non-Use Value of Acquatic Habitat
Aquatic habitat refers to the form of natural land cover that serves as habitat for aquatic
organisms. Johnston et al. (2005), as well as Loomis and Richardson (2008), supported the
endpoint of aquatic habitat at the level of a watershed or lake. Incorporation of the value of
aquatic habitat in an ecosystem service framework must be dealt with carefully. The value that
anglers hold for aquatic habitat should not be added to the value they hold for fish caught. That
combination would be double counting of use-value. However, non-users hold value for fish and
their aquatic habitat as well. This value is significant and should be incorporated in regard to
fishery improvements (or degradations). To capture non-use value associated with aquatic
habitat, a study sponsored by the US Forest Service (USFS) by Loomis and Richardson (2008) is
employed. Loomis and Richardson (2008) relied on a meta-analysis of aquatic habitat valuations
by Johnston et al. (2005) to create a benefit transfer tool for the non-use valuation of aquatic
habitat.
Johnston et al. (2005) surveyed the aquatic habitat valuation literature searching for studies in the
United States that: 1) contained both use and non-use value, 2) valued changes in water quality
affecting aquatic habitat, 3) used academically accepted methodologies, and 4) provided
sufficient information on the resource, context, and study attributes. Of 300 relevant studies,
Johnston et al. (2005) selected 34 as meta-data, which provide a total of 81 observations.
Johnston et al. (2005) regressed these 81 observations to determine the influence of relevant
variables and illuminate the magnitudes of use and non-use value of aquatic habitat.
Loomis and Richardson (2008) focused on the non-use portion of Johnston et al. (2005) and
constructed a user-friendly benefit transfer model for the USFS. An example for this benefit
-------
transfer model is provided in Table B.6.8 for the Upper Arkansas River in Central Colorado.
These values were calculated using the median household income for Colorado in 2006, and
final values were updated to 2013 dollars. The baseline and increase in water quality figures in
Table 7.9 are derived from the Vaughan (1981) water quality ladder, more popularly known as
the Resources for the Future (RFF) water quality ladder.
Best possible
water quality
10
9
8
7
6
5
4
3
2
1
Acceptable for drinking

J Js
Acceptable for drinking
^ s-J J#
Acceptable for drinking
A J
Acceptable for drinking

Worst possible
water quality
Figure B.6.3. Illustrative Water Quality Ladder. Modified from Vaughan (1981).
Table B.6.8. Non-Use Values of Aquatic Habitat for Upper Arkansas River, per Household per Year,
(From Loomis and Richardson, 2008)	
Baseline Water Quality Increase in Water	Less than 50% Fish More than 50% Fish
Quality	Population Change	Population Change
4	2	$12.48	$27.68
a Recommended value because the Upper Arkansas River went from ""unfishable" to ""fishable".
All values are in 2013$
Source: Loomis and Richardson (2008)
-------
When the Loomis and Richardson (2008) benefit transfer model was calibrated to reflect the
changes in Colorado, the result was an aquatic habitat non-use value of $27.68 (2013$) per
household. Note that the endpoint problem is an issue for valuing aquatic habitat nonuse value.
Economists conduct the valuation at the scale of a single river, multiple rivers, a single lake, or
multiple lakes. Environmental scientists do not conduct aquatic habitat improvements at the scale
of an entire river or lake. Instead, it is more likely to be by the river mile, wetland acre, or acre of
surface water. Additionally, ecologists likely do not agree with the simplicity of the water quality
ladder used as the basis for the example. For the example above, the endpoint problem is
partially alleviated because remediation was conducted at the scale of a river and because there is
no double counting of non-use value.
References for Appendix IB
Abdalla, C.W., 1990, Measuring economic losses from ground water contamination: An investigation of
household avoidance costs: Journal of the American Water Resources Association, v. 26, p. 451-463.
Abdalla, C.W, Roach, B.A, and Epp, D.J., 1992, Valuing environmental quality changes using averting
expenditures: an application to groundwater contamination: Land Economics, p. 163-169.
Applied Technology Council (ATC), 1991, Seismic vulnerability and impact of disruption of lifelines in
the conterminous United States: Federal Emergency Management Agency Report, 224 p.
Aubuchon, C.P., and K.M. Morley, 2013, The economic value of water: Providing confidence and context
to FEMA's methodology: Journal of Homeland Security and Emergency Management, v. 10, p. 1-21.
Borisova-Kidder, A., 2006, Meta-analytical estimates of values of environmental services enhanced by
government agricultural conservation programs: Ph.D. thesis, Ohio State University.
Bourassa, S.C., Hoesli, M., and Sun, J., 2003, What's in a view?: FAME Research Paper.
Boyle, K.J., Poe, G.L., and Bergstrom, J.C., 1994, What do we know about groundwater values?
Preliminary implications from a meta-analysis of contingent-valuation studies: American Journal of
Agricultural Economics, v. 76, p. 1055-1061.
Brander, L.M., and Florax, J.G.M., 2007, The valuation of wetlands: primary versus meta-analysis based
value transfer: Ashgate, Aldershot.
Casey, J.F., Kahn, J.R., and Rivas, A., 2006, Willingness to pay for improved water service in Manaus,
Amazonas, Brazil: Ecological Economics, v. 58, p. 365-372.
CDH (Colorado Department of Health), 1990, Leadville metals exposure study: State of Colorado.
Chang, S.E., Svekla, W. D., and Shinozuka, M., 2002, Linking infrastructure and urban economy:
Simulation of water-disruption impacts in earthquakes: Environment and Planning B: Planning and
Design, v. 29, p. 281-301
Collins, AR and Steinback, S., 1993, Rural household response to water contamination in West Virginia:
Water Resources Bulletin, v. 29, p. 199-209.
-------
Dalhuisen, J.M., Raymond, J.G.M. Florax, de Groot, H.L.F, and Nijkamp, P., 2003, Price and income
elasticities of residential water demand: A meta-analysis: Land Economics, v.79, p. 292-308.
Gorlach, B., and Interwies, E., 2003, Economic assessment of groundwater protection: a survey of the
literature: Berlin: Ecologic.
Gould, E., 2009, Childhood lead poisoning: conservative estimates of the social and economic benefits of
lead hazard control: Environ Health Perspectives, v. 117, p. 1162-1167.
Griffin, R.C., and Mjelde, J.W., 2000, Valuing water supply reliability: American Journal of Agricultural
Economics, v. 82, p. 414-426.
Harrington, W., Krupnick, A.J., and Spofford Jr, W.O., 1989, The economic losses of a waterborne
disease outbreak: Journal of Urban Economics, v. 25, p. 116-137.
Howe, C.W., Smith, M.G., Bennett, L., Charles, M., Flack, J.E., Hamm, R.M., Mann, R., Rozaklis, L.,
and Wunderlich, K., 1994, The value of water supply reliability in urban water systems: Journal of
Environmental Economics and Management, v. 26, p. 19-30.
Hylander, L.D., and Goodsite, M.E., 2006, Environmental costs of mercury pollution. Science of the
Total Environment, v. 368, p. 352-370.
Ingraham, M.W, and Foster, S.G., 2008, The value of ecosystem services provided by the US National
Wildlife Refuge System in the contiguous US: Ecological Economics, v. 67, p. 608-618.
Johnston, R.J., Besedin, E.Y., Iovanna, R., Miller, C.J., Wardwell, R.F., and Ranson, M.H. 2005,
Systematic variation in willingness to pay for aquatic resource improvements and implications for
benefit transfer: A Meta-Analysis: Canadian Journal of Agricultural Economics/Revue Canadienne
dAgroeconomie, v. 53, p. 221-248.
Khan, Himayatullah, Iqbal, Faiza, Saeed, Imranullah, and Khan, Inayatullah, 2010, Estimating
willingness to pay for improvements in drinking water quality: evidence from Peshawar, Northern
Pakistan: Journal of Environmental Economics, v. 2, p. 38-43.
Kimbrough, Renate D, and Krouskas, Constantine A., 2012, Contribution of lead in soil to children's lead
burden, an update: Chemical Speciation and Bioavailability, v. 24, p. 105-112.
Koss, Patricia, and Khawaja, M Sami, 2001, The value of water supply reliability in California: a
contingent valuation study: Water Policy, v. 3, p. 165-174.
Laughland, A.S., Musser, L.M., Musser, W.N., and Shortle, J.S., 1993, The opportunity cost of time and
averting expenditures for safe drinking water: Journal of the American Water Resources Association,
v. 29, p. 291-299.
Loomis, J.B., and Richardson, L., 2008, Technical documentation of benefit transfer and visitor use
estimating models of wildlife recreation, species and habitats: National Council for Science and the
Environment.
Mergler, D., Anderson, H.A., Chan, L.H.M., Mahaffey, K.R., Murray, M., Sakamoto, M., and Stern,
A.H., 2007, Methylmercury exposure and health effects in humans: a worldwide concern: AMBIO: A
Journal of the Human Environment, v. 36, p. 3-11.
-------
NRC (National Research Council), 1997, Valuing ground water: Economic concepts and approaches.
Washington, DC: The National Academies Press.
Plumber, B., 2014, Five big questions about the massive chemical spill in West Virginia: Washington
Post, January 21, 2014
Poe, G.L., Boyle, K.J., and Bergstrom, J.C., 2001, A preliminary meta analysis of contingent values for
ground water quality revisited: Economic Value of Water Quality, p. 137-162.
Rice, G., and Hammitt, J.K., 2005, Economic valuation of human health benefits of controlling mercury
emissions from US coal-red power plants: Report for NESCAUM, Northeast States for Coordinated
Air Use Management.
Seigneur, C., Vijayaraghavan, K., Lohman, K., Karamchandani, P., and Scott, C., 2004, Global source
attribution for mercury deposition in the United States: Environmental Science and Technology, v.
38, p. 555-569.
Spadaro, J.V., and Rabl, A., 2008, Global health impacts and costs due to mercury emissions: Risk
Analysis, v. 28, p. 603-613.
Sundseth, K., Pacyna, J.M., Pacyna, E.G., Munthe, J., Belhaj, M., and Astrom, S., 2010, Economic
benefits from decreased mercury emissions: Projections for 2020: Journal of Cleaner Production, v.
18, p. 386-394.
Swain, E.B., Jakus, P.M., Rice, G., Lupi, F., Maxson, P.A., Pacyna, J.M., Penn, A., Spiegel, S.J., and
Veiga, M.M., 2007, Socioeconomic consequences of mercury use and pollution: AMBIO: A Journal
of the Human Environment, v. 36, p. 45-61.
Thorvaldson, J., Pritchett, J.S, and Goemans, C., 2010, Western households' water knowledge,
preferences, and willingness to pay: Canadian Journal of Agricultural Economics/Revue Canadienne
dAgroeconomie, v. 58, p. 497-514.
Trasande, L., Landrigan, P.J., and Schechter, C., 2005, Public health and economic consequences of
methyl mercury toxicity to the developing brain: Environmental Health Perspectives, v. 113, p. 590-
596.
UNEP, 2013, Global mercury assessment: Sources, emissions, releases and environmental transport:
UNEP Chemicals Branch: Geneva, Switzerland.
Vassanadumrongdee, S., Matsuoka, S., and Shirakawa, H., 2004, Meta-analysis of contingent valuation
studies on air pollution-related morbidity risks: Environmental Economics and Policy Studies, v. 6, p.
11-47.
Vasquez, W.F, Franceschi, D., and Van Hecken, G.T., 2012, Household preferences for municipal water
services in Nicaragua: Environment and Development Economics, v. 17, p. 105-126.
Vaughan, W.J. (1981). The water quality ladder, Appendix II in Mitchell, Robert Cameron, and Richard
T. Carson. An Experiment in Determining Willingness to Pay for National Water Quality
Improvements. Washington, DC: Resources for the Future, 1981.
Walls, M., Kousky, C., and Chu, Z., 2015, Is what you see what you get? The value of natural landscape
views: Land Economics, v. 91, p. 1-19.
-------
Appendix C: A Valuation of the Upper Arkansas River Fishery
Recovery from Leadville to Canon City - Providing Context
for the Selection of Ecosystem Service Endpoints
A century's worth of legacy mine sites around Leadville created mine site pollution sources that
contaminated the Arkansas River headwaters with metals as far downstream as Canon City.
Before remediation, mine pollution reduced the fish population between Leadville and Lake
Creek to a level that was unfishable (Clements et al., 2010, Policky, 2012, Policky, 2013).
Farther downstream, between Lake Creek and Canon City, fish did not live beyond their third
year due to chronic toxicity of metals (Clements et al., 2010, Policky, 2012, Policky, 2013). The
California Gulch Superfund site remediation has been credited with a transformation in this
fishery that culminated in 2014 when it achieved the highest fishery designation possible - Gold
Medal Water.
This appendix applies the benefit transfer model components of WTP for fish caught, WTP for
angling days, and non-use value for aquatic habitat to the Upper Arkansas River Fishery in
Colorado. This exercise does not represent linked natural and social science because the data are
solely derived from economic studies, creel surveys, and demographic information. Instead, the
purpose is to circumvent data availability issues, which prevented geo-environmental modeling
of aquatic habitat and trout populations on the full scale of impacts, to quantify a significant
portion of the benefits produced by remediation. The relative scale of the fishery that was
impacted by remediation of the California Gulch Superfund site and the area where sufficient
data were available to conduct geo-environmental modeling of contaminant source processes is
shown in Figure C. 1.
-------
Upper Arkansas River Watershed
Purpose and Scope
The purpose of this analysis is to improve the understanding of the value of ecosystem
improvements due to the Superfund remediation. The scope covers improvements in fish
population on the Upper Arkansas River from the East Fork of the Arkansas River down to
Parkdale (Figure C.2). Because of the lack of comprehensive information and the resultant
inability to model the links among remedial action, water quality improvement, and fish
population in the framework of climate variability, an important assumption required by this
analysis is that the increase in trout catch is entirely due to the remediation. Although other
-------
factors surely contribute to the increase in trout catch, such as increasing numbers of anglers in
Colorado, this assumption is supported by previous analyses of the impact of trace metals on the
Upper Arkansas River (Clements et al., 2010, Policky, 2012, Policky, 2013).
Upper Arkansas River Watershed
Turquoise
Lake
Tunnel
Iowa Gulch
EmpireGulch
~^Big Union
Johnson
Village
Twin Lakes
Salida
Poncha
Springs
Parkdale
Florence
Leadville
AK-l
^ Leadville
AR-3
AR-5
AR-6
AR-7
AR-8
AR-9
10 mi
20 km
Pueblo
West
Pueblo
Figure C.2. Upper Arkansas River: Colorado Parks and Wildlife River Segments and EPA Sampling
Sites
Data and Method
The main data relied on by this analysis were collected from Colorado Parks and Wildlife (CPW)
creel censuses conducted along the river from Crystal Lakes to Parkdale for the years 1995,
2008, and 2012. CPW collects information on creel census area angling hours, angling days,
-------
hourly catch rate, and proportion of anglers who are from out-of-state. The specific sites
surveyed by the creel censuses are small and rarely represent the larger river segments that they
are located within. Table C.l provides detailed information on the size of creel census areas and
their respective river reaches.
Table C.l. Colorado Parks and Wildlife River Reaches and Corresponding Creel Census Areas
River
Reach
River Reach Name
Miles
in
River
Reach
Creel Census Area
Creel
Census
Miles
EF
EF-1 - Confluence
2.2

NA
-1
Confluence - California Gulch
3.8

NA
0
California Gulch - Chrystal Lakes
2.8

NA
1
Chrystal Lakes - Kobe Bridge
5.1
Highway 24 - Kobe
3.2
2
Kobe Bridge - Lake Creek
4.2
Kobe - Two Bit Gulch
2.2
3
Lake Creek - Otero Bridge
10.6
Ball Town - Granite
2.6
4
Otero Bridge - Highway 285 Bridge
9
Otero Bridge - Railroad Bridge
3.1
5
Highway 285 Bridge - Ruby Mountain
6
Big Bend - F Street
5.7
6
Ruby Mountain - Stone Bridge
11.2
Big Bend - F Street
5.7
7
Stone Bridge - Stockyard Bridge
10.9
Big Bend - F Street
5.7
8
Stockyard Bridge - Howard Bridge
11.4
Stockyard Bridge - Badger Creek
5.9
9
Howard Bridge - Lazy J
8.3
Big Cottonwood Creek - Lone
Pine
3
10
Lazy J - Texas Creek
12
Big Cottonwood Creek - Lone
Pine
3
11
Texas Creek - Parkdale
13.3
Big Cottonwood Creek - Lone
Pine
3
Because creel census areas are not representative of river reach, expert knowledge is required to
extrapolate creel census area estimates to broader river reach estimates. Policky (2013) provided
an expert extrapolation of creel census angling days to river reach angling days for each river
segment in the year 2012. How these factors were calculated is illustrated in Equation C. 1 and
Table C.2.
2Q12RmchDayEsHmate — 2012CreelDayEstimate
ExtrupolatianFactw = 1 +	WUCreelDayEstMe. 	
The extrapolation factors from Table C.2 are important because they allow extrapolation of creel
census area estimates to a larger and more useful scale. The extrapolation factors are multiplied
-------
by creel census estimates to estimate river reach angling hours (for 1995, 2008, and 2012) and
river reach angling days (for 1995 and 2008, Table C.3).
Table C.2: CPW Creel Census and River Reach Angling Day Estimates Used to Calculate
Extrapolation Factor	
River Reach
Creel Census Area Angling
CPW Estimate of Angling Angling Day Extrapolation
Number

Days

Days per River Reach
Factor



2012

2012


1

3,600

7,200
1.99

2

1,600

2,400
1.53

3

2,700

10,800
4.02

4

1,600

4,300
2.78

5

5,600

3,500
0.63

6

5,600

12,700
2.29

7

5,600

10,400
1.88

8

8,100

11,600
1.43

9

5,100

9,300
1.83

10

5,100

14,000
2.75

11

5,100

14,400
2.82

Total

49,700

100,600


Table C.3: Resulting Angling Hour and Day Estimates per River Reach
River Reach
Estimated River Reach Angling Hours Estimated River Reach Angling Days
Number







1995

2008
2012 1995
2008
2012
1


4,750
16,600
2,000
7,200
2
2,400

3,700
5,500 1,100
1,500
2,400
3
5,000

7,200
19,900 5,000
3,500
2,400
4
2,700

4,100
5,600 2,000
3,200
4,300
5
900

5,500
7,800 900
2,300
3,500
6
3,200

20,300
28,500 3,200
8,400
12,700
7
2,700

16,600
23,400 2,600
6,900
10,400
8
9,500

21,300
31,300 7,200
6,700
11,600
9
10,400

12,400
20,000 8,500
5,560
9,300
10
15,600

18,700
29,900 12,800
8,300
14,000
11
16,000

19,200
30,800 13,100
8,600
14,300
Total
68,400

133,800
219,200 56,600
56,900
100,600
Estimates of the number of hours fished for each river segment in 1995, 2008, and 2012 are
paired with average hourly catch rates from Policky (2012) to estimate the total catch for each
river segment (Table C.4). Next, the estimated catch is multiplied by the WTP to catch an
additional fish - $2.94 from Table C.5.
-------
To estimate the number of fish caught in the river segments above the highest CPW creel sites
(Segments 0, -1, and EF), this analysis applies an estimate of the percentage of the fish
population caught by anglers in Segment 1 (34%) to the fish population estimates from sample
sites AR-3 (Segment 0), AR-1 (Segment -1), and EF-5 (Segment EF). The results of this
application can be seen in Table C.6.
The use of an ecosystem service approach to value recreational fishing is rather novel.
Typically, it is the 'fishing day' that is valued to estimate benefits of recreational fisheries.
Therefore, to put the fish-centric valuation approach in full context, it is compared to a similar
valuation using the value of an angling day estimated by Loomis and Richardson (2008). This
'angler-centric' approach combines estimates of the number of angling days with the value of an
angling day ($67.91) to achieve a valuation of the change in angling over the study period. Table
C.6 in the following section provides the results of this analysis.
Non-Use Value of Aquatic Habitat Improvement
Finally, the value of fishery improvements to non-fisherman is estimated through non-use value
that households place on healthy aquatic habitat. People who do not fish still accrue benefits
from the remediation of the Upper Arkansas River fishery. This may be as simple as the pride a
resident feels from knowing that the Upper Arkansas River is now a Gold Medal trout fishery,
rather than a periodic conduit for acid-mine drainage. As mentioned previously, this sentiment is
referred to by economists as non-use value. It often has great weight in non-market valuation and
should be captured when possible.
Loomis and Richardson (2008) provided a straight-forward benefit transfer of non-use value
associated with aquatic habitat. To achieve this benefit transfer, the annual non-use value of
$27.68 per year is combined with the number of households living in the counties that straddle
this section of the Arkansas River. As reflected in Table C.7, the water quality in the Upper
Arkansas River moved up two rungs on the Resources for the Future (RFF) water quality ladder
from the fourth rung. The resulting improvement in fish population was greater than 50%
because fish could not live beyond their third year due to chronic metal toxicity.
Multiplying the recommended value of $27.68 from Table 6.9 by the number of households in
the three counties that encompass this watershed provides an estimate of the non-use value
created by the improvement in the Upper Arkansas River's aquatic habitat. County household
data were provided for 2010 by the U.S. Census Bureau. Table 9.8 details the population and
annual non-use aquatic habitat values.
Results
The fish-centric valuation approach resulted in a trebling of the annual value of fishing from
1995 to 2012 - $252,000 to $773,000 respectively. Remediation of the California Gulch site was
ongoing between 1995 and 2008, but was largely completed by 2009. Policky (2012) and
Policky (2013) indicated that the fishery was slow to improve until 2008. However, between
2008 and 2014, the fishery improved rapidly. The results of the fish-centric valuation reflect this
-------
sentiment. The original annual value ($252,000) took 13 years to double, but doubled again just
4 years afterward.
Table C.4. Fish-Centric Valuation Results (2013$)
River
Average Hourly Catch
Estimated Number of Fish
Estimated Value of Fish Caught
Reach

Rate

Caught per Year

per Year

Number










1995
2008
2012
1995
2008
2012
1995
2008
2012
EF
NA
NA
NA
1,100
1,100

$2,700
$2,700

-1
NA
NA
NA
2,306
2,300

$5,800
$5,700

0
NA
NA
NA
400
2,000

$1,000
$5,000

1
1
0.91
1.2
0
4,300
19,900
$0
$12,700
$58,400
2
1
0.91
1.2
2,400
3,300
6,600
$7,200
$9,800
$19,300
3
1
0.91
1.2
5,000
6,600
23,900
$14,800
$19,400
$70,200
4
1
0.91
1.2
2,700
3,800
6,800
$7,800
$11,100
$19,900
5
1.3
1.4
1.2
1,200
7,700
9,300
$3,400
$22,800
$27,400
6
1.3
1.4
1.2
4,200
28,400
34,200
$12,400
$83,400
$100,600
7
1.3
1.4
1.2
3,500
23,300
28,100
$10,200
$68,500
$82,500
8
1.3
1.4
1.2
12,300
29,900
37,600
$36,200
$87,800
$110,400
9
1.3
1.4
1.2
13,500
17,400
23,900
$39,700
$51,200
$70,400
10
1.3
1.4
1.2
20,300
26,100
35,900
$59,600
$76,800
$105,600
11
1.3
1.4
1.2
20,800
26,800
36,900
$61,200
$78,900
$108,500
Total



85,900
177,700
263,000
$252,500
$522,300
$773,300
The angler-centric valuation approach resulted in a less pronounced change in the annual value
of fishing. The annual value remained essentially unchanged between 1995 and 2008 - $3.83M
to $3.871M. The bump in annual value between 2008 and 2012 is also reflected in this analysis -
$3.871M to $6.825M. One important result of this comparison is that the fish-centric approach
results in annual values that are approximately one-tenth the annual value of the angler-centric
approach. This is an interesting result given the discussion above about the portion of total
angling value that is represented by the number of fish caught. Note that the angler-centric
valuation does not include Segments 0, -1, or EF because no estimates are available for the
number of angling days in these stretches and because no proxy could be found for these
estimates.
Table C.5. Angler-Centric Valuation Results (2013$)
River Reach Estimated River Reach	Estimated Value of Angling Days per
Number	Angling Days per Year	Year
	1995 2008 2012 1995	2008	2012
1	0 2,000 7,200	$0 $135,800 $489,000
2	1,100 1,500 2,400	$74,700 $101,900 $163,000
3	5,000 3,500 10,800	$339,600 $237,700 $733,400
4	2,000 3,200 4,300	$135,800 $217,300 $292,000
-------
River Reach
Estimated River Reach
Estimated Value of Angling Days per
Number
Angling Days per Year

Year

5
900
2,300
3,500
$61,100
$156,200
$237,700
6
3,200
8,400
12,700
$217,300
$570,400
$862,500
7
2,600
6,900
10,400
$176,600
$468,600
$706,300
8
7,200
6,700
11,600
$489,000
$455,000
$787,800
9
8,500
5,600
9,300
$577,200
$380,300
$631,600
10
12,800
8,300
14,000
$869,200
$563,700
$950,700
11
13,100
8,600
14,300
$889,600
$584,000
$971,100
Total
56,400 57,000 100,500 $3,830,100 $3,870,900 $6,825,100
Table C.6. Comparison of Valuation Results (2013$)
Increase in
Annual
Value
from 1995 to
2012
1995
2008
2012
Fish-centric $252,000 $522,000 $773,000 $521,000
Angler-centric $3,830,000 $3,871,000 $6,825,000 $2,995,000
Between 1995 and 2012, non-use value of aquatic habitat increased by $834,000 per year. This
dollar figure can be thought of as the value that residents place on the transformation of the
Arkansas River into a Gold Medal Fishery.
Table C.7. Valuation of Aquatic Habitat for Upper Arkansas River
County
Number of Households
Annual Non-Use Aquatic Habitat Value
Lake
3,100
$92,000
Chaffee
7,800
$235,000
Fremont
16,900
$507,000
Total	27,800	$834,000
The geo-environmental models of trout population improvement from AR3 and AR5 suggest that
remediation improved the fish populations in the first and second river stretches. Policky (2012)
and Policky (2013) suggest that the remediation also improved the other river stretches as well.
Extending this assumption to the fish-centric valuation, the benefit of the remediation pertains to
-------
all river segments, increasing from $252,000 per year in 1995 to $773,000 in 2012—resulting in an
additional benefit of $521,000 per year (see Table C.6). Extending this assumption to the angler-
centric valuation, the benefit of the remediation increases from $3,830,000 per year in 1995 to
$6,825,000 in 2012—resulting in an additional benefit of $2,995,000 per year (see Table C.6).
From a net present value standpoint, the fish-centric value of fishery improvement from 1995 to
2012 (using a 3% discount rate) is $13,401,000 for a 50-year time frame and $16,458,000 for a
100-year time frame. The angler-centric net present value of fishery improvement from 1995 to
2012 (using the same discount rate) is $77,100,000 for a 50-year time frame and $94,600,000 for
a 100-year time frame. Finally, the non-use aquatic habitat net present value is $21,500,000 for a
50-year time frame and $26,400,000 for a 100-year time frame.
Table C.8. Net Present Value of Fishery Improvements at 3% over 50 and 100 Years (2013$)


NPV
NPV

Annual Value
50 Years
100 Years

of



Improvement
3% Discount
3% Discount
Fish-Centric
$520,000
$13,400,000
$16,500,000
Angler-Centric
$2,995,000
$77,100,000
$94,600,000
Aquatic Habitat
$834,000
$21,500,000
$26,400,000
These benefits can be compared to the expenditures made during the remediation. Table C.9
details the expenditures, purpose, and funding source for expenditures found in the public record
for the California Gulch Superfund cleanup.
Discussion, Implications and Conclusion
The results of this analysis indicate that the net present value of benefits (at 3% over 100 years)
from the improvement of the fishery are, at most, $94.6 million plus $26.4 million equals $121
million (2013$). Although additional benefits accrued to society from the remediation, such as
reduction of blood lead level in children, increased quality of irrigation water, and greater
municipal water supply reliability, the causal link is more dubious and valuation of the benefits is
not possible due to poor availability of usable data. To compare the benefits that have been
estimated for the remediation of this fishery, Table C.9 below details the expenditures that could
be found for the California Gulch Superfund remediation.
Table C.9. Expenditures Located for the California Gulch Superfund Site	
Year Purpose
Funding Source
Expenditure
(2013$)
1988	Yak Tunnel Plug/Treatment Plant
1988	Annual Yak Tunnel O&M costs (1988-1992)
1999	1-year field demonstration, biosolids/lime in soil
2001	OU1 23 years ofYWTP Costs
2012	OU1 Costs associated with Black Cloud Mine
1994	OU2 Malta Gulch removal actions (1995-1996)
2001	OU2 15 years of monitoring
OU3 Denver and Rio Grande slag piles
ASARCO/Resurrection
ASARCO/Resurrection
ASARCO/Resurrection
ASARCO/Resurrection
Resurrection Mining
Hecla Mining
Hecla Mining
Union Pacific
$29,490,000
$4,530,000
$6,850,000
$21,180,000
$5,070,000
$1,070,000
$790,000
???
-------
Year
Purpose
Funding Source
Expenditure
(2013$)
1998
OU4 NPV of removal costs
Resurrection Mining
$5,830,000
2001
OU4 Erosion control/inspection
Resurrection Mining
$580,000
2001
OU5 AV/CZL and EGWA remediation costs
ASARCO
$4,280,000
2001
OU5 5 years of monitoring costs for AV/CZL site
ASARCO
$120,000
2001
OU5 Institutional control costs for EGWA
ASARCO
$40,000
2001
OU5 5 years of monitoring costs for EGWA site
ASARCO
$20,000

OU6 Removal action costs (1995-2001, '05, '08, '11
EPA
???
2010
OU6 Stray Horse Gulch Waste Rock Repository
EPA
$19,230,000
2010
OU6 100 years NPV of costs
EPA
$490,000

OU7 Remedial costs
ASARCO/Resurrection
???
2001
OU7 14 years of monitoring costs
ASARCO/Resurrection
$1,570,000

OU8 1995/1998 Oregon Gulch tailing removal
Resurrection Mining
???
2001
OU8 Fluvial tailings removal
Resurrection Mining
$1,300,000
2001
OU8 Stream sediment remediation costs
Resurrection Mining
$940,000
2001
OU8 14 years of monitoring costs
Resurrection Mining
$100,000
2001
OU9 Lead program costs over 12 years
ASARCO/Resurrection
$6,370,000
2012
OU9 Annual costs for Phase 2 of lead program
ASARCO/Resurrection
$760,000

OU10 Cost of remedial actions
Resurrection Mining
???
1997
OU10 30 years NPV of costs
Resurrection Mining
$3,690,000
2001
OU10 16 years of additional monitoring costs
Resurrection Mining
$250,000
2005
OU11 Combined capital and operational costs
ASARCO/Resurrection
$6,220,000
2012
OU11 Combined capital and operational costs
ASARCO/Resurrection
$15,870,000
2009
OU12 Institutional control monitoring costs
EPA
$1,370,000
2012
OU12 3 years of monitoring costs
EPA
$640,000
2012
OU12 3 years of enforcement costs
EPA
$150,000
Total


$138,810,000
These expenditures do not represent the full tally for this remediation effort because several large
expenditures could not be estimated. Nevertheless, the largest estimate of benefits that were
estimated (NPV over 100 years at 3%) do not quite cover this cost estimate. Future research
could address this question by locating the missing expenditures and estimating additional
benefits related to the remediation.
The fish-centric recreational angling valuation approach [using USEPA (2006)] is valuable
because of its explicit focus on fish caught - the closest possible endpoint to the fish population.
However, the marginal value of $2.94 per fish does not include the economic benefits generated
by anglers as a result of their fishing. On the other hand, the angler-centric approach transfers
values from Loomis and Richardson (2008), which used the economic benefits generated by the
angler to estimate the willingness to pay a value of $67.91 per angling day. While Loomis and
Richardson (2008)'s values may paint a clearer picture of the economic benefits of fishing, they
have more to do with the joy of a family fishing trip than with an increase in fish population.
However, the desirability of a specific reach of a stream as a destination for a family fishing trip
relies on its reputation as a source of abundant and large game fish.
-------
This discussion encompasses the end-point problem that natural and social scientists will
continue to work out in relation to ecosystem service valuation (Boyd, 2007). Future research on
this issue from the fisheries management side ought to isolate the impact of increasing fish
population on the number of fish caught. Data would be required for fish population, fish caught,
fishing capital, fishing skill, angler hours, and angler days, among other factors. Future research
from ecological economists should isolate the portion of angling-day value that comes from
catching each marginal fish.
References for Appendix C
Boyd, J.W., 2007, The endpoint problem. Resources, Washington DC, v. 165, p. 2-6.
Cantrell, R.N., Garcia, M., Leung, P. and Ziemann, D., 2004, Recreational anglers' willingness to pay for
increased catch rates of Pacific threadfin (Polydactylus sexfilis) in Hawaii. Fisheries Research, v. 68,
p. .1149-158.
Clements, W.H., Vieira, N.K.M., and Church, S.E., 2010, Quantifying restoration success and recovery in
a metal-polluted stream: a 17-year assessment of physicochemical and biological responses: Journal
of Applied Ecology, v. 47, p. 899-910.
Deisenroth, D.B., Loomis, J.B. and Bond, C.A., 2013, Using Revealed Preference Behavioral Models to
Correctly Account for Substitution Effects in Economic Impact Analysis. Journal of Regional
Analysis and Policy, v. 43, p. 157.
Johnson, D. M., Behnke, R.J., Harpman, D.A., and Walsh, R.G., 1995, Economic benefits and costs of
stocking catchable rainbow trout: a synthesis of economic analysis in Colorado: North American
Journal of Fisheries Management, v. 15, p. 26-32.
Loomis, J.B., and Richardson, L., 2008, Technical documentation of benefit transfer and visitor use
estimating models of wildlife recreation, species and habitats: National Council for Science and the
Environment.
Mazzotta. M., Wainger, L., Sifleet. S., Petty, J.T., and Rashleigh, B., 2015, Benefit transfer with limited
data: An application to recreational fishing losses from surface mining: Ecological Economics, v.
119, p. 384-398.
Morey, E.R., Breffle. W.S., Rowe, R.D., and Waldman, D M., 2002, Estimating recreational trout fishing
damages in Montana's Clark Fork River basin: summary of a natural resource damage assessment:
Journal of Environmental Management, v. 66, p. 159-170.
Patterson, W.B., and Frederick, 2011, Do hatchery trucks make happy anglers?. Dissertation, Royal
Roads University.
Policky, G., 2012, Fisheries inventory report: Arkansas River: Colorado Parks and Wildlife.
Policky, G., 2013, Fisheries inventory report: Arkansas River: Colorado Parks and Wildlife.
USEPA, 2006, Regional benefits analysis for the final section 316(b) Phase III existing facilities rule.
Office of Water (4303T). Report EPA-821-R-04-007. June 2006.
-------
-------