Total Dissolved Solids: The Challenges Ahead US EPA Region 3 Freshwater Biology Team Wheeling, WV

Total Dissolved Solids: The Challenges Ahead

US EPA Region 3Freshwater Biology Team

Wheeling, WV

Freshwater Biology Team, EPA R3, EAID, OMA

• FBT Members– Amy Bergdale, Frank

Borsuk, Kelly Krock, Maggie Passmore, Greg Pond, Louis Reynolds

• Assist the states in methods development, bioassessment, biocriteria

• Assist EPA R3 in use of biological data– WQS, monitoring, TMDLs,

NPDES, superfund, etc.– Perform special studies

Background

• Many states have identified “ionic toxicity”, conductivity and/or total dissolved solids (TDS) as a stressor or pollutant in their integrated lists.

• EPA has also identified TDS (and component ions) as a stressor impairing aquatic life.

• EPA lacks aquatic life criteria for TDS mixtures.

• Some TMDLs have been deferred due to lack of criteria.

• We also need criteria for effluent limits for discharge permits.

What We Know

• Some component ions are toxic to aquatic life.

• Ex. Mount et al 1997 , acute endpointsK+ > HCO3

- =Mg2+ > Cl- > SO42-

• Laboratory fish are more tolerant than laboratory inverts.

• Test duration important.• Chronic endpoints important.• Resident fish are more tolerant than resident

inverts.

Mount et al1997.C. Dubia More Sensitive toTDS than D. magna or fatheads.

What We Know

• Ion mixtures have varying toxicity• Ion mixtures source specific

– Alkaline coal mine drainage (HCO3- ,

Mg2+, Ca2+, SO42- )

– Marcellus Shale Brine (Na+, Cl-,SO42-)

– Coal Bed Methane (Na+, HCO3- ,SO4

2-)

What We Know

• Effects synergistic, additive, or ameliorative

• Depends on the ions and their concentrations

• In some systems (e.g. Appalachian headwater streams) lab controlled toxicity tests are not a good predictor of instream aquatic life use impairment.

Two Webinars on TDS (2009)

• Toxicity testing approaches to develop criteria for individual ions– Surrogate organisms– Iowa: chloride and sulfate– Illinois: sulfate

• Empirical approaches– bioassessment and water quality data to

develop a criterion for an ion mixture:– Ex. Alkaline mine drainage in southern WV and

KY Appalachian streams.

The Case for Single Ion Criteria• Lab experiments are controlled• Other stressors are excluded• Toxicity testing data deemed more “defensible”• Pollutant specific criteria instead of integrative

parameters such as TDS or conductivity– Easier to implement than narrative criteria– Easier to check compliance– Permit writers understand it

• Can still incorporate site-specific conditions• Resources will focus on source reduction• Regulating TDS “futile”; Ion mixtures too

complex.

Chloride LC50 vs. HardnessC. dubia

LC 50 VS. Hardness

LC50 = 440.74*(Hardness)0.2144

R2 = 0.8246

100

1000

10000

10 100 1000

Hardness (Caco3 mg/l)

LC

50

(m

g/l

)

Chloride LC50 vs. SulfateC. dubia

LC 50 VS. Sulfate

LC50 = 1736.9*(Sulfate)-0.0588

R2 = 0.3153

100

1000

10000

10 100 1000

Sulfate (mg/l)

LC

50 (

mg

/l)

Iowa Cl Criteria

Iowa Sulfate Criteria

Illinois Sulfate Criterion Also Based on Acute Tests

Illinois Sulfate Criterion

Illinois states that “Sensitive organisms reside in receiving streams with sulfate concentrations of 2,000 mg/L.”

Illinois Sulfate Criterion

The Case for an Empirical Approach

• Context is important. • Aquatic life in small Appalachian streams is not

the same as in Iowa or Illinois! • We must protect the resident aquatic life uses.• Unlike Illinois, we routinely see aquatic life use

impairment downstream of alkaline mine drainage.

• Elevated TDS, hardness and alkalinity, in the absence of other stressors (e.g. habitat, low pH, metals violations).

• TDS and component ions are strongly correlated to this impairment.

OH

KY

WV

PA

VA

Context is Important. What aquatic life are we trying to protect? What is the natural

water quality? What is the effluent quality?

NPDES discharge

Bio-Monitoring

Effluent Dominated Streams

HeptageniidaeEpeorus

HeptageniidaeHeptagenia

Ephemerellidae

E. Fleek, NC DWQ

Ephemere

lla

Mayflies represent ~25-50% of Abundance; ~1/3rd biodiversityIn natural, undegraded Appalachian streams

KY AppalachianHeadwaters(sandstone)y = 0.7821x - 28.661

R2 = 0.9754

0

500

1000

1500

2000

2500

3000

3500

4000

4500

0 500 1000 1500 2000 2500 3000 3500 4000 4500

Conductivity

TD

SWe use conductivity as a surrogate for

TDS

We also use conductivity as a surrogate for sulfate (Kentucky Data)

y = 0.574x - 54.165

R2 = 0.93

0

500

1000

1500

2000

2500

0 500 1000 1500 2000 2500 3000 3500

Conductivity

SO

4

y = 1.2148x - 1.042R2 = 0.94

0

0.5

1

1.5

2

2.5

3

3.5

1.5 1.7 1.9 2.1 2.3 2.5 2.7 2.9 3.1 3.3 3.5

log Cond

log

SO

4West Virginia Data

Using Empirical Data

• Note – conductivity of 500-1000 uS/cm approximates

sulfate of 200-400 mg/l– Iowa sulfate criteria ranges 500-2000 mg/l– Illinois sulfate criteria in range of 1000-1500

mg/l

Reference

Mined

Mined/Residential

%E

ph

emer

op

tera

Conductivity

0

10

20

30

40

50

60

70

80

0 500 1000 1500 2000 2500

Resident Mayflies Very Sensitive

(Eastern Kentucky Coalfields)

Note: strong nonlinear “threshold” response

0 500 1000 1500 2000 2500 3000

Conductivity

0

10

20

30

40

50

60

70

80

90%

May

flie

s

Unmined

Mined

Independent Datasets Confirm Sensitivity (West Virginia southern coal fields)

EPA EIS data (WV)based on mean monthly WQ concentrations (n=13

months)Spearman's Correlation Coefficients

n=89 # Ephem Taxa % EphemTDS -0.88 -0.86Conductivity -0.87 -0.86SULFATE -0.87 -0.85CALCIUM -0.87 -0.85MAGNESIUM -0.86 -0.83POTASSIUM -0.85 -0.82SELENIUM -0.74 -0.72NITRATE/NITRITE NITROGEN -0.72 -0.69pH -0.64 -0.60SODIUM -0.60 -0.59IRON, DISSOLVED -0.57 -0.61CHLORIDE -0.39 -0.46MANGANESE -0.34 -0.35NICKEL -0.31 -0.31TOTAL ORGANIC CARBON -0.31 -0.35COPPER -0.05 -0.13TSS -0.03 0.03Temperature -0.02 -0.02D.O. 0.02 -0.02ALUMINUM 0.07 0.10BARIUM 0.10 0.05ZINC 0.19 0.16LEAD 0.25 0.23bold values = p<0.05

TDS andIons stronglyCorrelated To mayfliesAnd impairment

0-200

200-400

400-600

600-1000>1000

CONDUCTIVITY

0

10

20

30

40

% E

ph

emer

ella

% S

ensi

tive

May

flie

s

EpeorusEphemerellaAmeletusDrunellaCinygmulaParaleptophlebia

EpeorusEphemerellaAmeletusDrunellaCinygmulaParaleptophlebia

Is aquatic life in small Appalachian streams more sensitive to TDS pollution than that in midwestern

streams?

Sensitive Mayflies:

0-200

200-400

400-600

600-1000>1000

CONDUCTIVITY

0

10

20

30

40

50

60

70

0-200

200-400

400-600

600-1000>1000

CONDUCTIVITY

0

10

20

30

40

50

% Is

on

ych

ia

0

10

20

30

40

50

60

70

80

% T

ole

ran

t M

ayfl

ies

0-200

200-400

400-600

600-1000>1000

CONDUCTIVITY

Isonychia, Tricorythodes, Baetis, Caenis

What aquatic life is found in the midwest? Perhaps more TDS-tolerant

invertebrates?

Facultative/Tolerant Mayflies:

The Case for an Empirical Approach

• The concentrations of ions that are correlated with high probability of aquatic life use impairment are much lower than the toxicity testing data imply would be protective.– Suggests that common toxicity testing organisms

are not as sensitive as resident aquatic invertebrates.

– Many of the toxicity test results have been based on acute tests. The tests and endpoints should be chronic and the toxicity tests should test sensitive life stages.

• There may be seasonal issues due to insect life cycles.

• Empirical data may help us determine the more sensitive resident species.

• Bioassessment endpoints are the best tool to capture the total effect of a complex ion mixture.

Examples of ambient toxicity

C. dubia Chronic Effects

0

20

40

60

80

100

120

0 1000 2000 3000

Sp. Cond. Field (us/cm)

EC

25

Re

pro

du

cti

on

(%

)

Chronic effects were detected in samples with field conductivity >1800 µS/cm.There is NO dilution capacity in these streams.

Chronic Effects Levels

C. dubia Chronic Effects

0

20

40

60

80

100

120

0 500 1000 1500

Sp. Cond. Estimated @ EC25 (uS/cm)

EC

25

Re

pro

du

cti

on

(%

)

Estimated conductivity at EC25 % ranged from 448-1243 with an average of 820 µS/cm.

This range is slightly higher than where we see effects with resident biota.

C. dubia more tolerant than resident Aquatic Life

Stream Resident Biota More Sensitive Than WET Surrogate

0

20

40

60

80

100

0 500 1000 1500 2000 2500 3000

Sp. Cond. Field (uS/cm)

GLIMPSS

EC25

All sites were rated impaired using the genus level GLIMPSS (<66) , which directly measures aquatic life use impairment. The resident biota are more sensitive than the WET surrogate, C. dubia. Can’t use C. dubia alone to express “safe” thresholds, but it can be used as an indicator of the more toxic discharges.

Ref for GLIMPSSNot tox tested

Using Empirical Data

• Linear regression• Quantile regression• Conditional Probability Analysis• Regression Trees• Note

– conductivity of 500-1000 uS/cm approximates sulfate of 200-400 mg/l

– Iowa sulfate criteria ranges 500-2000 mg/l– Illinois sufate criteria in range of 1000-1500 mg/l

Regression of GLIMPSS by log COND (R²=0.476)

0

10

20

30

40

50

60

70

80

90

100

1 1.5 2 2.5 3 3.5

log COND

GL

IMP

SS

ActiveModelConf. interval (Mean 90%)Conf. interval (Obs. 90%)

125 uS/cm 880 uS/cm

Ex: Linear Regression

Ex: Quantile Regression (summer)

N=535

IMPAIRMENT THRESHOLD

Ex: Quantile Regression (spring)

N=276

IMPAIRMENT THRESHOLD

Ex. Conditional Probability Approach

Paul and McDonald (2005)• CPA relies on a large dataset to develop

criteria.– Simply asks “what is the probability of

impairment given conductivity value ≥ x”?• P(y|x) where y is impairment threshold (IBI),

and x is some TDS or conductivity value.

• J. Paul (EPA, RTP, in review) found – 100% chance of MAHA sites being impaired

when conductivity >575 and – 100% chance of Florida streams impaired when

conductivity >750

N=949RBP HAB>130

Ex: CPA: WV DEP data: Summer pH>6

Conductivity

Pro

bab

ility

of

impa

irmen

t Probability of Impairment Over 90% when Cond > 500

88.2% variance

All Ions, Metals, pH, Hardness

%EPHEMMean=20.45SD=18.236

N=64

Mean=4.04SD=5.945

N=30

Mean=34.94SD=11.947

N=34

SULFATE<350.66

Mean=1.45SD=2.040

N=23

Mean=12.5SD=6.720

N=7

Mn DISS.<0.0074

Mean=23.83SD=6.393

N=8

Mean=38.4SD=11.196

N=26

CONDUCTIVITY<433.1

Mean=34.0SD=9.799

N=14

SULFATE<15.6

Mean=44.1SD=10.179

N=12

Mean=29.66

SD=9.077N=9

ZINC<0.023

Mean=40.13

SD=7.688N=5

Mean=39.95

SD=11.966N=6

Mean=48.33

SD=6.533N=6

MAGNESIUM<6.9

Split Variable PRE Improvement 1 SULFATE 0.726 0.726 2 Mn DISS 0.758 0.032 3 CONDUCTIVITY 0.819 0.062 4 SULFATE 0.855 0.036 5 ZINCTOTAL 0.872 0.017 6 MAGNESIUM 0.882 0.010

Ex: Regression Tree (MTM/VF EIS)

How do these empirical results compare to Iowa’s Sulfate Criteria?

We have not reviewed any bioassessment data from Iowa.R3 Empirical examples suggest impairment at sulfate 200-400 mg/l

Water Quality Based Approachto Pollution Control

DetermineProtection Level

(EPA Criteria/State WQS)

Conduct WQAssessment

(Identify Impaired Waters)

Set Priorities(Rank/Target Waterbodies)

Evaluate Appropriatenessof WQS for Specific Waters

(Reaffirm WQS)

Define and AllocateControl Responsibilities

(TMDL/WLA/LA)

Establish SourceControls

(Point Source, NPS)

Monitor and EnforceCompliance

(including instream bioassessments)

Measure Progress

Recommendations•Do not rely solely on toxicity

testing to determine protective limits.

•Consider chronic toxicity testing endpoints.

•Consider dilution ratios.•Combine toxicity testing and

empirical data approaches when field data are available.

Recommendations• Prepare a technical support

document on TDS–reflects acute and chronic toxicity testing literature

–offers some examples of empirical datasets and how they would be used to characterize aquatic life, and develop, refine or evaluate criteria and permits.

Recommendations• Always use bioassessments to

assess aquatic life uses downstream of discharges with TDS.

• These data should feed back into the permit and possibly result in site specific criteria.–Reflect all toxicants in discharge–Protect actual aquatic life that should be residing in that stream type

Ongoing Research - Surrogates

• Toxicity of TDS to surrogate lab organisms– Review literature for

TDS– Develop empirical

datasets between TDS and aquatic life

– Acute and chronic tests with mining effluent and reconstituted salts and surrogate organisms (e.g. C. dubia)

• USGS Columbia Lab, Duluth EPA Lab

• Preliminary Data…Hassell et al 2006

Ongoing Research - Natives• Metal and osmotic

ecophysiology • Deploy insects in situ – sample

individuals in a time course– Measure growth, metal and

electrolyte content, subcellular compartmentalization of metals

– Explain any differences in metal tolerance, bioaccumulation and toxicity

• Laboratory Exposures– Monitor oxygen consumption,

osmoregulatory status and Adenosine triphosphate (ATP) levels

– Characterize “energetic costs” to living in high conductivity

• Outcome– Provide information on whether

metal uptake is contributing to impairment

– Provide information on mechanism for TDS impairment

• North Carolina State

Buckwalter et al, 2007

Discussion

• Where do we go from here?• Technical Barriers?• Non-Technical Barriers?• What do you need from EPA?• What can you expect from EPA?• How do we advance aquatic life criteria?• How do we advance TMDL development?