Fish Mercury Impairment in California Reservoirs:

Historic Mines and Other FactorsEPA Region 9 State-of-the-Science Workshop on

Mercury Remediation in Aquatic EnvironmentsSeptember 26, 2013

California Water BoardsMulti-Region Team

Michelle Wood, Carrie Austin, Steve Louie & many others

•WANTED: Your Feedback!

OutlineIntroduction to Statewide Mercury Control Program for

ReservoirsQuick overview of:

California reservoirs fish MeHg impairment

Linkage between fish MeHg bioaccumulation, sources, and other factors

Mercury sources and where they occur

Where might mine remediation enable measurable and timely fish mercury reductions?

•2

total mercury (Hg)inorganic mercuryOutline

Introduction to Statewide Mercury Control Program for Reservoirs

Quick overview of:California reservoirs fish MeHg impairment

Linkage between fish MeHg bioaccumulation, sources, and other factors

Mercury sources and where they occur

Where might mine remediation enable measurable and timely fish mercury reductions?

•3

monomethylmercuryMeHg

Statewide Mercury Control Program for Reservoirs

Goal: Quickly, measurably reduce fish MeHg

•4

Website with fact sheets & updates www.waterboards.ca.gov/

water_issues/programs/mercury

Sign up for email notices at: www.waterboards.ca.gov/resources/

email_subscriptions/swrcb_subscribe.shtml#quality

Fish MeHg Levels

74 CWA 303(d) listed reservoirs

another ~70+ soon to be listed

Rainbow trout have low MeHg

But so do black bass! [green boxes]

•5

Red and orangeorange indicatefish MeHg exceeds

draft target of0.2 mg/kg

Pacific Ocean

California

• High elevationSierra Nevada

Fish MeHg Levels

74 CWA 303(d) listed reservoirs

another ~70+

soon to be listed

•6

Control program might need to address ~500 reservoirs

Red and orangeorange indicatefish MeHg exceeds

draft target of0.2 mg/kg

Pacific Ocean

California

• Fish were collected from about 350 lakes and reservoirs

• ~50% impaired<typical for U.S.>

• There are >1,000 reservoirs in CA

It’s a complicated story…

•7

It’s a complicated story…

•8

MultipleFactors

low sedHg,high fish MeHg variability

upstream mines & high sedHg, low fish MeHg

3 factors are equally important!

Model Equation

•9

LN [Fish methylmercury] = 0.563 * [aqueous total Hg]

+ 0.338 * [aqueous MeHg] / [chlorophyll-a]+ 0.394 * (annual water level fluctuation) – 0.912

R2 = 0.83Adjusted R2 = 0.81Predicted R2 = 0.72

n = 26 reservoirs, P < 0.001

MeHg

Today’s focus: mine waste remediation

Multiple factors → Multiple possible tools

•10

(a)Source control: Reduce Hg sources to reduce MeHg production in reservoirs

Some reservoirs are in naturally Hg-enriched areas

•11

Mercury & Gold Mines

•12

74 303(d)-Listed reservoir watershed boundaries indicated by black outline

Hg Mines Au MinesAu Mines

•13

Reservoir fish MeHg compared to modeled 2001 atmospheric Hg deposition rate

Reservoirs with no record of upstream gold or mercury mines;

60 have fish MeHg > target

At least 1 recorded upstream gold or mercury mine

Can have high fish MeHg but low atm dep and no minesCan have low fish MeHg but very high atm Hg depVery highest fish MeHg associated with extensive Hg mining

MultipleFactors

Key Question: Where can mine waste remediation make quick reductions in reservoir fish MeHg?

•14

Program Goal: Quickly & measurably reduce fish MeHg

(a) Source control: Reduce Hg sources to reduce MeHg production in reservoirs

Desk-top [GIS-based] analysis: Mine factors consideredHigh reservoir sediment Hg compared to background

indicates substantial mine contribution

Mine sites localized to a relatively small watershed area indicates highly contaminated soils likely not dispersed

throughout watershed

Mines near reservoirs (e.g., within 10 to 20 km) likely do not have many miles of creek channels filled with waste

that can be difficult to remediate

•15

Initial desk-top analysis results:53 of the 74 Hg-impaired reservoirs have at least

one recorded upstream mine or prospect

Of these 53:only 3 “probably” and 2 “maybe” reservoirs where mine waste remediation expected

to make timely and measurable improvements

•16

A comparison of two neighbors…Part 1

•17

350 mm Bass MeHg

(mg/kg)

Reservoir sediment

Hg (mg/kg)

Localized mines?

Mine proximity

(km)

REMSAD atm dep

rate (g/km2/yr)

San Antonio 0.24 0.07 1 prospect na 8.0 (low)

Nacimiento 1.1 0.39 Yes <10-20 8.2 (low)

Halfwayto target

A comparison of two neighbors…Part 2

•18

Likelyfish MeHg reduction

from source control

Watershed soil Hg (mg/kg)

Annual reservoir

water level fluctuation

(feet)

Aqueous MeHg

Geomean [peak] (ng/L)

Chlor-a

(μg/L)

San Antonio minimal 0.04 25 0.04 [0.9] 6.2

Nacimiento ~40% 0.08 48 0.08 [3.7] 2.2

MeHgLinkage Model

= TotHg sources

+ aqMeHg / Chlor-a

+ water level fluctuation

A comparison of two neighbors…Part 2

•19

Likelyfish MeHg reduction

from source control

Watershed soil Hg (mg/kg)

Reservoir sediment

Hg (mg/kg)

Aqueous MeHg

Geomean [peak] (ng/L)

Chlor-a

(μg/L)

San Antonio minimal 0.04 0.07 0.04 [0.9] 6.2

Nacimiento ~40% 0.08 0.39 0.08 [3.7] 2.2

Halfwayto target

Not controllable

A comparison of two neighbors…Part 2

•20

Likelyfish MeHg reduction

from source control

Watershed soil Hg (mg/kg)

Annual reservoir

water level fluctuation

(feet)

Aqueous MeHg

Geomean [peak] (ng/L)

Chlor-a

(μg/L)

San Antonio minimal 0.04 25 0.04 [0.9] 6.2

Nacimiento ~40% 0.08 48 0.08 [3.7] 2.2

MeHgLinkage Model

= TotHg sources

+ aqMeHg / Chlor-a

+ water level fluctuation

Halfwayto target

Not controllable

Nacimiento ≈ MeHg machine!

A comparison of two neighbors…Part 2

•21

Likelyfish MeHg reduction

from source control

Watershed soil Hg (mg/kg)

Annual reservoir

water level fluctuation

(feet)

Aqueous MeHg

Geomean [peak] (ng/L)

Chlor-a

(μg/L)

San Antonio minimal 0.04 25 0.04 [0.9] 6.2

Nacimiento ~40% 0.08 48 0.08 [3.7] 2.2

Halfwayto target

aqMeHg/Chlor-a ratio >5x higher inNacimiento

Not controllable

Likely not controllable Likely controllable

We evaluated:High reservoir sediment Hg compared to backgroundLocalized mine sites Mines near reservoirs

•22

What other factors can we consider?mine processesmine productivityothers???

Key Question: Where can mine waste remediation make timely reductions in reservoir fish MeHg?

We need realistic expectations

of where quick improvements are possible from

mine waste remediation

Let there be no doubt…We are still advocating mine waste remediation as a tool to reduce fish MeHg…And we are looking forward to coordinating with stakeholders to explore ways to prioritize specific sites within a watershed, e.g.…

Proximity and erosion of waste to surface water: High threat - visual evidence or high potential of wastes eroding into surface waters

Medium threat - wastes near waters but no visual evidence of erosion

Low threat - wastes located far from waters and no visible evidence of erosion

Level of Hg contamination: Historical mine processes and productivity

Waste pile and portal discharges: Hg concentrations and volumes

Hg concentrations in downstream water and sediment

Site accessibility

•23

Find Out More, Stay in Touch Website with fact sheets & updates www.waterboards.ca.gov/

water_issues/programs/mercury

Sign up for email notices at: www.waterboards.ca.gov/resources/

email_subscriptions/swrcb_subscribe.shtml#quality

Hand out Project goals & contact info Discussion questions from

this presentation

•24

We evaluated:High reservoir sediment Hg compared to backgroundLocalized mine sites Mines near reservoirs

•25

What other factors can we consider?mine processesmine productivityothers???

Key Question: Where can mine waste remediation make timely reductions in reservoir fish MeHg?

We need realistic expectations

of where quick improvements are possible from

mine waste remediation

•26

•27

Extra Slides for Possible Questions

•28

We have reasons to be hopeful…

•29

Gambonini Mercury Mine•Largest Hg pollution source to Walker Creek & Tomales Bay•Erosion control alone – no capping! >90% Hg load reductions & >50% sediment load reductions•But we don’t yet have fish MeHg data

Fish

MeH

g C

onc.

(mg/

kg)

Lake Pinchi Mercury Mine, BC•Some waste capping & erosion control since 1975 mine closure; additional remediation planned•Initial dramatic fish MeHg reduction, then modest reductions•Coring indicate slow burial process – no large tributaries to provide significantly cleaner sediment

Source: Teck Cominco-Azimuth 2008 (Figure 4.6-1) Source: Kirchner et al. 2011 (Figure 3a)

•Logscale

Prioritizing where to start additional mine remediation within a watershedProximity and erosion of waste to surface water:

High threat - visual evidence or high potential of wastes eroding into surface waters

Medium threat - wastes near waters but no visual evidence of erosion Low threat - wastes located far from waters and no visible evidence of erosion

Level of Hg contamination: Historical mine processes and productivity Waste pile and portal discharges: Hg concentrations and volumes Hg concentrations in downstream water and sediment

Site accessibility •30

Next stepsLearn from what we hear today and enhance our

evaluation!

Have follow-up meetings with workshop participants and other interested stakeholders during the next couple months to continue to enhance the evaluation

Scientific peer review of draft technical report in 2014

•31

Nacimiento Reservoir Sediment THg

•32

Hubbert 1991 Water Boards 2013

Area General Location[range ci ted in

Rice et a l . 1994]

Site Result

Area Average Site Result

Pebblestone Shut-In 0.01Oak Shores area 0.04Las Tablas Creek Arm 1.5Las Tablas Creek Arm 1.6

East of LTC Arm

Mile 19 (Las Tablas Creek Arm) 0.05-1.2 0.08 0.08 na

Dip Creek shore 0.02Mile 17 (Dip Creek) 0.18Snake Creek Arm; Heritage Ranch 0.08Dip Creek mouth 0.07Snake Creek mouth 0.13

3.09Las Tablas Creek Arm

Rice et al. 1994

na

0.05

Lake Nacimiento Bottom Sediment Total Mercury Data (mg/kg)

0.04-0.16

0.05 to 0.62

0.17-0.99 1.55

0.03

0.10

Upper

Near Dam

•33

Nacimiento Reservoir Fish THg(ATSDR 2007, Figure 5)

•34

Sources and Methylation of Mercury

•35

Before After

LinkageAnalysis

>30 reservoir and watershed variables

This is just one analysis example…

•36

Pearson'sr

Spearman'sRho

[aq MeHg] Geomean / [Chl-a] Geomean 0 0.67 0.70Reservoir Sediment [THg] Geomean 0 0.50 0.47Watershed Soil [THg] Geomean 0 0.40 0.44Reservoir Longitude 5 0.39 0.40Reservoir [Chl-a] Geomean -0.22 0.34 0.27Average Water Level Fluctuation 0 0.33 0.35Watershed Percent Vegetation 3 0.32 0.29[aq MeHg] Geomean -0.5 -0.31 -0.38[aq THg] Geomean 0 0.30 0.25Watershed Percent Open Water 0 -0.27 -0.30Reservoir Dam Height 0.5 0.25 0.34Reservoir Elevation 0.21 -0.22 -0.27Watershed Percent Forests 2 0.22 0.12CA Hg Atm Dep Rate to the Watershed 0 0.19 0.17Watershed Productive Mines per Mile -3.77 -0.17 -0.05Number of Mines in Watershed (PAMP) -0.5 -0.15 -0.17Year Dam Built 5 0.15 0.19Watershed Mines per Mile -2 -0.14 -0.01Number of Dams Upstream of Reservoir -0.22 -0.13 -0.06Reservoir Maximum Capacity 0 0.10 0.17Watershed Area/Reservoir Surface Area -0.11 -0.09 -0.19CA Hg Atm Dep Rate to the Reservoir Surface 0 0.08 0.12Reservoir Latitude 5 0.08 0.04Watershed Surface Area 0 -0.05 0.13All Hg Atm Dep Rate to the Watershed -1 -0.03 -0.02All Hg Wet Atm Dep Rate to the Reservoir Surface 0 -0.03 0.03Number of Productive Mines in Watershed -0.13 -0.03 -0.002Watershed Percent Wetlands -5 0.02 0.002All Hg Atm Dep Rate to the Reservoir Surface -1 0.02 -0.05All Hg Wet Atm Dep Rate to the Watershed 0 0.01 -0.04Watershed Percent Agriculture -5 0.01 0.08Reservoir Surface Area 0 0.01 0.05Number of Mines in Watershed (MRDS) 0 -0.002 -0.03* Highlighted environmental factors indicate statstically signficant correlations with fish tissue mercury concentrations for the parametric, non-parametric, or both analyses (using their respective two-sided tests of signficance, P < 0.05).

Table 1: Correlation coefficients for 350 mm standardized predatory fish [MeHg]

versus reservoir and watershed factors

Environmental Factors*Lambda Trans-

formation Correlation Coefficient

Reservoirs: Multiple factorscontribute to impairment

Comes down to 3 factors:Total Hg = (a) sourcesMeHg/Chl-a = (b) methylation / (c) bioaccumulationReservoir fluctuation = (b) methylation / (c) bioaccumulation

•37

Several tools: Select right tool(s) for each reservoir

one size fits all

Source control

Water chemistry(decrease methylation)

Fisheries management

•38

Potential Solutions toReduce Fish Mercury

Multiple factors → Multiple possible tools

•39

(b) Water chemistry (reduce methylation):Studies & pilot tests to reduce in-reservoir MeHg production:

- Reduce anoxia with artificial circulation or oxygenation, or adjust redox potential by adding nitrate

- If low pH, increase pH by adding lime or reducing air emissions (NOx, SOx, etc.) that produce acid rain

Multiple factors → Multiple possible tools

•40

(c) Fisheries & food web management: Studies & pilot tests to manage food web to reduce MeHg bioaccumulation

- Add nutrients to oligotrophic reservoirs to increase productivity at food web base

- Manipulate food web, e.g., intensive fishing, to increase growth rate of remaining fish

- Restore native anadromous fisheries (e.g., salmon & steelhead trout)- Change stocking to increase numbers of less predatory fish

•41

Reservoir fish MeHg compared to modeled 2001 atmospheric Hg deposition rate

No record of upstream gold or mercury mines

At least 1 recorded upstream gold or mercury mine

Can have high fish MeHg but low atm dep and no minesCan have low fish MeHg but very high atm Hg depVery highest fish MeHg associated with extensive mining

MultipleFactors

USEPA’s REMSAD 2001 model output for atmospheric Hg deposition throughout the U.S.

•42

Overall, atm dep rates low in CA

Some limited areas rival atm dep rates in eastern U.S.

Source: USEPA 2008a (Figure 6-3c)

USEPA’s REMSAD 2001: % of atmospheric Hg deposition from sources CA regulators cannot regulate

•43

Several reservoirs in areas with 30 to 80% of atm dep from CA industrial sources (blue arrows)

YELLOW-RED:

>90%natural + international industrial

DARKEST GREEN:>60% CA industrial

Statewide:~10% :

CA industrial sources

~60% :International

industrial

sources

~30% :Natural sources

Sources of Hg in atmospheric deposition per USEPA’s REMSAD model output for 2001 & literature review

•44

Reservoir fish MeHg and

upstream mines

•45

Reservoir or lake with no upstream mines

Almost half of reservoirs (~150) exceed draft target

More than a third of those (~60) do not have upstream mine sites

No mines, high fish

MeHg

No mines, high fish

MeHg

Reservoirs with very high fish MeHg

•46

~60%~40%

Mine waste remediation is a useful tool, but benefits may take a while to observe in many downstream reservoirs

•47

Many reservoirs have a high density of mine sites over much of their watersheds, some with watershed sizes bigger than many countries…

Need to consider mine waste already in creek and river channels that may be difficult to remediate

MRDS gold mine features upstream of 303(d) Listed

Hg-impaired reservoirsin the Feather, Yuba, & Bear

Rivers’ watersheds

Mine waste remediation is a useful tool, but benefits may take a while to observe in many downstream reservoirs

•48

Many reservoirs have a high density of mine sites over much of their watersheds,some with watershed sizes bigger than many countries…

Need to consider mine waste already in creek and river channels

Need to consider potential for widespread elevated soil THg from past mining activities’ air emissions

Profile of Hg concentration in soil impacted by release of particulate Hg and vapor from Hg mines in the New Idria Mercury District, California [Source: Rytuba 2002, Figure 2]