Monty Porter - Streams/Rivers Monitoring CoordinatorJason Childress – Biological Team Leader

Oklahoma Water Resources Board (OWRB) Water Quality Programs Division - Monitoring Section

Methods Comparability in Oklahoma’s Low Gradient Streams

Roadmap

• Oklahoma’s part in the National Wadeable Streams Assessment (NWSA) and Methods Comparability

• What does an Oklahoma low gradient stream look like?

• A look at the data and some answers—maybe??

• How can we do it better?

Questions???

• Do different sampling methods produce similar or different answers?

• Can the results of different sampling methods be combined to produce a national assessment?

• Is it realistic to expect states to change methods?

Oklahoma’s Part in the NWSA• Target Sites

– 20 sites visited with comparability on 18 (Red River site excluded and West Buffalo Creek had sample issues)

– 15 of these sites are low gradient

• Reference Site Selection– 10 stations visited with comparability

on 8 (Trader’s creek had sample issues)– 4 of these sites are low gradient– Station's were ecoregion-based– Used OCC reference site work and best

professional judgment

NWSA Parametric Coverage• Typical in-situ parameters such as pH, dissolved

oxygen, temperature, specific conductance• Nutrients• Major cations and anions• Variety of metals • Benthic Macroinvertebrates—single habitat• Periphyton• Physical Habitat• Stream Flow

Cimmaron River, Beaver County, Southwest Tablelands Ecoregion

Trail Creek, Dewey County, Central Great Plains Ecoregion

Red River, Tillman County, Central Great Plains Ecoregion

Grey Horse Creek, Osage County, Flint Hills Ecoregion

Hybarger Creek, McClain County, Central Great Plains Ecoregion

Unknown Creek in the Arbuckle Uplift, Johnston County, Central Oklahoma/Texas Plains Ecoregion

Methods Comparability Study (MCS)Methods Comparability Study (MCS)

MCS Design

• Spatially consists of Fourteen WSA Cooperators

• Incorporates both target and reference sites

• Side-by-side collections using multiple methods (Oklahoma compared W-EMAP to State RBP)

Objectives of the MCS

• Compare state to federal methods to determine the extent of the difference between the methods.

• Is needed in future studies to assess the condition of the Nation’s waters.

• Allays the need for standardized protocols across states

• Looked at both physical habitat and benthic macroinvertebrate sampling methodologies

Oklahoma’s Benthic Macroinvertebrate Collections• Source of Method

– Rapid Bioassessment Protocol adopted from Plafkin, et al (1999)– Method adopted by Oklahoma state agencies in OWRB Technical Document 99-3 (1999)– Use is codified into Oklahoma Administrative Code (OAC) through the Oklahoma Water Quality

standards (OAC 785:45) and the Use Support Assessment Protocols (OAC 785:46)• Oklahoma RBP Method

– Multi-habitat method targeting richest habitats in flowing water over a 400-800 meter reach– 500 uM nets and sieves are used– Composite Riffle—3 kicks in a fast, medium, and slow riffle– Streamside Vegetation—reachwide 3-minute collection of composited jabs– Woody Debris—reachwide 5-minute collection of composited scrapes/picks– 100-150 organism subsample with a large and rare scan– Identified to lowest practical taxonomic level

NWSA Methods—Benthic Macroinvertebrate Collections

• Sample taken reach wide

• Kick sample with modified D-frame net at 11 equidistant transects

• Work L, R, C

• Composite Sample

• Processed at any of a number of EPA contract labs

• 300-500 organism subsample

• Identified to lowest taxonomic level

MCS Results

• Meetings– National meeting of cooperators in Baltimore, MD– Regional meeting of cooperators in Lawrence, KS

• Have results to date on only 6 states• Focused comparability on macroinvertebrates

– Evaluate relationships of Indices of Biotic Integrity (IBIs)

– Evaluate relationships of condition class assessment– Evaluate relationships of pass-fail assessment– Investigate effects of natural slope gradient– Investigate effects of stressor gradient– Investigate relationships with biological condition

gradient

Data Analysis

• Data Sets– WSA_WSA IBI—is the WSA dataset processed through the

WSA IBI (Courtesy of Versar, Inc. and USEPA, OWOW)– WSA_Ok IBI—is the WSA dataset processed through the

Oklahoma IBI – OK_Ok IBI—is the OK dataset processed through the Oklahoma

IBI

• Oklahoma IBI– Metrics are Taxa Richness, EPT Taxa Richness, EPT

Abundance, Modified Hilsenhoff Biotic Index, EPT/Chironomidae Proportion, Percent Dominant Taxa, and Shannon-Weaver Diversity

– 5 condition classes including Reference (Good), Non-impaired (Good), Slightly Impaired (Fair), Moderately Impaired (Poor), and Severely Impaired (Poor)

Data Analysis (continued)

• Data Analysis

– Evaluate relationships of Indices of Biotic Integrity (IBIs)

– Evaluate relationships of condition class assessment

– Evaluate relationships of pass-fail assessment

– Make comparisons of IBI scores to some habitat and land use metrics

– Look at some simple boxplots of data to see where some variation may exist

• Data Issues

– Subsample counts not standardized between datasets

– Taxonomic resolution

– Did not compare Oklahoma Data to NSA IBI

– Sample size is small

– Data collection issues including sample handling and weather

– Oklahoma IBI is well used but application of reference was a first run

– Do no look at biological condition gradient in this analysis

Draft Interim Comparability Study. 2006. Versar, Inc. and USEPA OW

Relationships of Indices of Biotic Integrity (IBIs)

• Value of r2 for Oklahoma Low Gradient Stream Data– WSA_Ok IBI vs. Ok_Ok IBI

0.20

– WSA_WSA IBI vs. WSA_Ok IBI 0.68

– WSA_WSA IBI vs. OK_Ok IBI 0.28

Draft Interim Comparability Study. 2006. Versar, Inc. and USEPA OW

WSA_OK IBI Scores vs. OK_OK IBI Scores

R2 = 0.2796

0

20

40

60

80

100

120

140

160

0 20 40 60 80 100 120 140

WSA_OK IBI

OK

OK

_IB

I

WSA_OK IBI Scores vs. WSA_WSA IBI Scores

R2 = 0.667

0

10

20

30

40

50

60

70

0 20 40 60 80 100 120 140

WSA_OK IBI

WS

A_W

SA

IB

I

OK_OK IBI Scores vs. WSA_WSA IBI Scores

R2 = 0.1964

0

10

20

30

40

50

60

70

0 20 40 60 80 100 120 140 160

OK_OK IBI

WS

A_W

SA

IBI

WSA_WSA IBI WSA_OK IBI OK_OK IBI

mean 34.7 89.1 88.9median 36.1 100.0 89.0p25 23.8 79.0 72.4p75 47.3 111.0 106.1minimum 23.8 79.0 72.4maximum 65.1 118.0 135.7

Comparison of WSA_OK IBI and OK_OK IBI Scores to Various Habitat Metrics

IBI scores vs. % Fines

R2 = 0.0647

R2 = 0.0932

0

20

40

60

80

100

120

140

160

0.00 20.00 40.00 60.00 80.00 100.00 120.00

% Fines

IBI S

core

WSA_OK IBI OK_OK IBILinear (OK_OK IBI) Linear (WSA_OK IBI)

IBI scores vs. % Sand

R2 = 0.0046

R2 = 0.1381

0

20

40

60

80

100

120

140

160

0.00 20.00 40.00 60.00 80.00 100.00

% Sand

IBI S

core

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)Linear (WSA_OK IBI)

IBI scores vs. % Pool

R2 = 0.0127

R2 = 0.3308

0

20

40

60

80

100

120

140

160

0.00 20.00 40.00 60.00 80.00 100.00 120.00

% Pool

IBI S

core

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

IBI scores vs. % Rock Substrate

R2 = 0.0001

R2 = 0.2465

0

20

40

60

80

100

120

140

160

0.00 20.00 40.00 60.00 80.00 100.00 120.00

% Rock Substrate

IBI S

core

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

Comparison of WSA_OK IBI and OK_OK IBI Scores to Several Land Use Metrics

• RHUM0—% of human land use at the site

• RAGT0—% of total agricultural land use at the site

• POPDENS—population density

IBI scores vs. RHUM0

R2 = 0.2765

R2 = 0.0928

0

20

40

60

80

100

120

140

160

0 10 20 30 40 50 60 70 80

RHUM0

IBI S

core

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

IBI scores vs. RAGT0

R2 = 0.2437

R2 = 0.0681

0

20

40

60

80

100

120

140

160

0 10 20 30 40 50 60 70 80

RAGT0

IBI S

core

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

IBI scores vs. POPDENS

R2 = 0.1327

R2 = 0.2476

0

20

40

60

80

100

120

140

160

0 5 10 15 20 25 30 35 40

POPDENS

IBI S

co

re

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

Relationship of Condition Class Assessment

• WSA Method/WSA IBI vs. OK Method/OK IBI– Not well compared – 8 of 18, or 44% have different class

• WSA Method/WSA IBI vs. WSA Method/OK IBI

– Not well compared – 9 of 18, or 50% have different class

• WSA Method/OK IBI vs. OK Method/OK IBI– Not well compared – 9 of 19, or 42% have different class– However, 50% of the divergence is at

the Fair/Good classifications

Draft Interim Comparability Study. 2006. Versar, Inc. and USEPA OW

WSA_OKIBI Poor Fair Good Total

Poor 1 1 1 3Fair 1 0 1 2Good 0 4 10 14

Total 2 5 12 19

OK_OK IBI

WSA_WSA IBI Poor Fair Good Total

Poor 3 1 4 8Fair 0 1 4 5Good 0 0 5 5

Total 3 2 13 18

WSA OK IBI

WSA_WSA IBI Poor Fair Good Total

Poor 2 3 3 8Fair 0 5 0 5Good 0 2 3 5

Total 2 10 6 18

OK_OK IBI

Comparison of WSA_OK IBI and OK_OK IBI Condition Classifications to Various Habitat Metrics

Condition Class vs. % Rock Substrate

R2 = 0.3289

R2 = 0.0055

0

1

2

3

4

0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00

% Rock Substrate

Con

ditio

n C

lass

WSA_OK IBI OK_OK IBILinear (OK_OK IBI) Linear (WSA_OK IBI)

Condition Class vs. % Pool

R2 = 5E-05 R2 = 0.3883

0

1

2

3

4

0.00 20.00 40.00 60.00 80.00 100.00 120.00

% Pool

Co

nd

itio

n C

lass

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

Condition Class vs. % Sand

R2 = 0.002

R2 = 0.2305

0

1

2

3

4

0.00 20.00 40.00 60.00 80.00 100.00

% Sand

Co

nd

itio

n C

lass

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

Condition Class vs. % Fines

R2 = 0.1313

R2 = 0.1329

0

1

2

3

4

0.00 20.00 40.00 60.00 80.00 100.00 120.00

% Fines

Co

nd

itio

n C

lass

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

Comparison of WSA_OK IBI and OK_OK IBI Condition Classifications to Several Land Use Metrics

• RHUM0—% of human land use at the site

• RAGT0—% of total agricultural land use at the site

• POPDENS—population density

Condition Class vs. RAGT0

R2 = 0.3456

R2 = 0.095

0

1

2

3

4

0 10 20 30 40 50 60 70 80

RAGT0

Co

nd

itio

n C

lass

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

Condition Class vs. RHUM0

R2 = 0.3812

R2 = 0.132

0

1

2

3

4

0 10 20 30 40 50 60 70 80

RHUM0

Co

nd

itio

n C

lass

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

Condition Class vs. POPDENS

R2 = 0.1735

R2 = 0.2904

0

1

2

3

4

0 5 10 15 20 25 30 35 40

POPDENS

Co

nd

itio

n C

lass

WSA_OK IBI OK_OK IBILinear (WSA_OK IBI) Linear (OK_OK IBI)

Relationship of Pass/Fail Assessment

• WSA Method/WSA IBI vs. OK Method/OK IBI– Not well compared– 6 of 18, or 33% have different status– More sites impaired per the WSA IBI

• WSA Method/WSA IBI vs. WSA Method/OK IBI

– Not well compared– 5 of 18, or 28% have different status– More sites impaired per the WSA IBI

• WSA Method/OK IBI vs. OK Method/OK IBI– Well compared– 3 of 19, or 16% have different status

Draft Interim Comparability Study. 2006. Versar, Inc. and USEPA OW

WSA_WSA IBI Fail Pass Total

Fail 2 6 8Pass 0 10 10

Total 2 16 18

OK_OK IBI

WSA_WSA IBI Fail Pass Total

Fail 3 5 8Pass 0 10 10

Total 3 15 18

WSA OK IBI

WSA_OKIBI Fail Pass Total

Fail 1 2 3Pass 1 15 16

Total 2 17 19

OK_OK IBI

OKNSA

35

25

15

5

Dataset

Num

ber

of ta

xa

OKNSA

7

6

5

4

Dataset

Mod

Hils

enho

ff B

iotic

Ind

OKNSA

4

3

2

Dataset

Sha

nnon

-Wea

ver

dive

rsity

inde

x

OKNSA

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

Dataset

EP

T/to

tal

OKNSA

15

10

5

0

Dataset

EP

T ta

xa

OKNSA

1.0

0.5

0.0

Dataset

EP

T/E

PT

+ C

hiro

nom

idae

OKNSA

1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

Dataset

Chi

rono

mid

ae/to

tal

Answers???• Do different sampling methods produce similar or different answers?

• Can the results of different sampling methods be combined to produce a national assessment?

• The r2 values between IBI scores show little to some relationship

• When use different IBI’s see poor comparability between both condition classes and pass/fail designations

• Using same IBI

– See poor comparability at the condition class level but 50% comes at the fair/good classifications

– However, when use same IBI see a little better comparability between pass/fail designations

– Relationship of datasets to habitat and land use metrics is a mixed bag

Answer

• Is it realistic to expect states to change methods?– No– Not economically feasible– Not politically practical– Most states have developed, adopted and

revised methods that fit that state– Are following RBP’s adopted and pushed

long ago by the EPA – RBP’s have been used effectively by the

states to both screen waterbodies and determine impairment status of streams

So where can we go from here?

• Need to invest money in comparability work

• Investigate comparability at all levels from sampling techniques to metric and index development

• In the end need to decide if the same final answer can be obtained

• Studies need to be more tightly designed and controlled

• Need to continue developing the good working relationship that came out of the national study and continues into the National Lakes Study

Questions?Questions?