comparison of sampling methods in low-gradient streams...comparison of sampling methods in...

Comparison of sampling methods in low-gradient streams

Raphael Mazor1 Andy Rehn2 Ken Schiff1

1Southern California Coastal Water Research Project, Westminster, CA2Aquatic Bioassessment Laboratory, Rancho Cordova, CA

November 29, 2006

Intro

Initiation of regional biomonitoring program: Integrating data from SWAMP and NPDES programs.

Sampling methods and assessment tools (SoCal-IBI) have been proposed.

Do these tools work in low-gradient streams?

Low-gradient streams are common in southern California, and their health is of great public interest.

Questions

1. Does the So-Cal IBI function well in low-gradient streams?

2. Which sampling methods are the most precise?

3. Do different sampling methods give similar results?

Background

Sampling methods

CSBP: Targets richest habitats (riffles, margins)

San Gabriel Arroyo Seco Arroyo Seco

Background

Sampling methods

CSBP: Targets richest habitats (riffles, margins)

?

San Gabriel Arroyo Seco Arroyo Seco

? ?

Background

Sampling methods

MH: Multi-habitat (25%, 50%, and 75% of channel width)

San Gabriel Arroyo Seco Arroyo Seco

Background

Sampling methods

MH: Multi-habitat (25%, 50%, and 75% of channel width)

San Gabriel Arroyo Seco Arroyo Seco

Background

Sampling methods

MCM: Margin-Center-Margin (also gets richest habitats)

San Gabriel Arroyo Seco Arroyo Seco

Background

Sampling methods

MCM: Margin-Center-Margin (also gets richest habitats)

San Gabriel Arroyo Seco Arroyo Seco

Methods

Low-gradient streams sampled in southern California:

-Santa Clara River (4 sites)

-Rio Hondo

-Santa Margarita River (2 sites)

-Santa Ana River

-Las Virgenes Creek

-Agua Hedionda

Methods

Each method tested in each river, often sampled in triplicate.

500-count samples were sorted and identified.

Metrics and IBI scores were calculated for each sample.

Results

Number of samples:

River CSBP MCM MHSanta Clara 5 5 6Agua Hedionda 2 3 3Rio Hondo 3 3 3Santa Margarita C 2 3 3Santa Margarita D 2 3 3Santa Ana 3 3 3Las Virgenes 2 3 3TOTAL 19 23 24

Results

CSBP MCM MHRichness 18.7 19.9 16.3Individuals per sample* 453 481 377

*p < 0.05

Of 66 samples total, 16 had < 450 organisms, of which 10 were MHsamples

Sampling method does NOT affect richness, but it may result in small samples.

Results

Very good

Good

Fair

Poor

Very poor

Sant

a C

lara

Riv

er

Agu

a H

edio

nda

Cre

ek

Rio

Hon

do

Sant

a M

arga

rita

C

Sant

a M

arga

rita

D

Sant

a A

na R

iver

Las

Virg

enes

Cre

ek

SC-IB

I

0

10

20

30

40

50

60

70

80

90

100

Results

Very good

Good

Fair

Poor

Very poor

Sant

a C

lara

Riv

er

Agu

a H

edio

nda

Cre

ek

Rio

Hon

do

Sant

a M

arga

rita

C

Sant

a M

arga

rita

D

Sant

a A

na R

iver

Las

Virg

enes

Cre

ek

SC-IB

I

0

10

20

30

40

50

60

70

80

90

100CSBP MCM MH

Results

SS d.f. F p

Method 39.9 2 0.6 0.575River 2641.3 5 14.8 <0.001Interaction 380.2 10 1.1 0.408Residuals 1426.1 40

Exclude Santa Margarita C

Two-way ANOVA on IBI Score

Method does NOT affect IBI score at most sites

CSBP

0 5 10 15 20 25 30 35

MC

M

0

5

10

15

20

25

30

35

CSBP

0 5 10 15 20 25 30 35M

H0

5

10

15

20

25

30

35

SoCal-IBI

R2 = 0.52 R2 = 0.58Slope = 0.62 Slope = 0.77

Results

Good relationship between all methods.

Results

CSBP: High variability at low-scoring sites.

Other methods: High variability at all scores.

CSBP

IBI

0 5 10 15 20 25 30 35

SD

0

2

4

6

8

10

MCM

IBI

0 5 10 15 20 25 30 350

2

4

6

8

10

MH

IBI

0 5 10 15 20 25 30 350

2

4

6

8

10

CSBP

0 2 4 6 8 10

MC

M

0

2

4

6

8

10

CSBP

0 2 4 6 8 10

MH

0

2

4

6

8

10

EPT Taxa

R2 = 0.76

Slope = 0.56

R2 = 0.43

Slope = 0.38

Results

Better relationship between CSBP and MCM than CSBP and MH.

CSBP

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

MC

M

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

CSBP

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7M

H0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

% Non-insect

R2 = 0.73

Slope = 0.60

R2 = 0.96

Slope = 0.85

Results

Better relationship between CSBP and MH than CSBP and MCM.

Results - Precision

Comparisons among streamsTypical questions:

Are streams in San Diego of fair or better condition?

Are streams draining urban areas worse than streams draining open space?

Among-stream variability (SD of site averages):

CSBP 6.6

MCM 6.1

MH 4.2

MH << MCM < CSBP

Results - Precision

Comparisons within streamsTypical questions:

Is this site in better condition following restoration?

Is this site above a biocriterion threshold?

Within-stream variability (average within-site SD):

CSBP 3.8

MCM 3.9

MH 4.1

All methods more-or-less the same.

NMDS 1 (20%)-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

NM

DS

2 (2

9%)

-1.0

-0.5

0.0

0.5

1.0

1.5

Stress = 10.2

Santa Clara

Santa Margarita D

Rio Hondo

AguaHedionda

Las Virgenes

SantaMargarita C

Santa Ana

Geography strongly influences community structure.

NMDS 1 (20%)-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

NM

DS

2 (2

9%)

-1.0

-0.5

0.0

0.5

1.0

1.5

CSBP MCM MH

Stress = 10.2

But sampling method does not.

Conclusions

1. Does the So-Cal IBI function well in low-gradient streams?

All streams are in poor condition.

True status of low-gradient streams?

What about “reference” streams?

Conclusions

2. Which sampling methods are the most precise?

All methods similar for within-stream comparisons.

MH best for among-stream comparisons.

But: low power for most applications.

Conclusions

3. Do different sampling methods give similar results?

Geography, not sampling method, has the strongest influence on community structure and IBI scores.

Correlations between methods are good.

Conclusions

Next steps:

“Better” reference sites (Central Coast).

Test other assessment techniques (e.g., RIVPACS).

Examine physical habitat data. What drives between-site differences?

End