ECOLOGICAL INTEGRITY ASSESSMENT OF OZARK ...biscweb.uark.edu/Coop/Publications/andreasthesis.pdfreaches of 10 Ozark rivers, five with Wild and Scenic status. Thirty -four variables

ECOLOGICAL INTEGRITY ASSESSMENT OF OZARK RIVERS

TO DETERMINE SUITABILITY FOR PROTECTIVE STATUS

by

Andrea Radwell

Arkansas Cooperative Fish and Wildlife Research Unit U.S. Geological Survey, Biological Resources Division

Department of Biological Sciences University of Arkansas Fayetteville, Arkansas

2000

COOP UNIT PUBLICATION NO. 36

ECOLOGICAL INTEGRITY ASSESSMENT OF OZARK RIVERS TO DETERMINE SUITABILITY FOR PROTECTIVE STATUS

ECOLOGICAL INTEGRITY ASSESSMENT OF OZARK RIVERS TO DETERMINE SUITABILITY FOR PROTECTIVE STATUS

A thesis submitted in partial fulfillment

of the requirements for the degree of

Master of Science

by

Andrea Radwell, B.S., M.A.

Northwestern University, Evanston, Illinois 1971, 1972

May 2000

University of Arkansas

THESIS DUPLICATION RELEASE

I hereby authorize the University of Arkansas Libraries to duplicate this thesis when needed for research and/or scholarship. Agreed _______________________________________ Refused ______________________________________

iv

ACKNOWLEDGEMENTS

I wish to thank the many people who have expressed their confidence in me,

shared their expertise, and provided the encouragement and guidance needed to

complete the research presented in this thesis.

Without Dr. Tom Kwak’s initial confidence in my abilities to become a

researcher, this project would never have been undertaken. He provided the important

step in helping me develop my ideas for this project into a coherent, well-defined

research agenda. He has always encouraged me to proceed, never doubting my ability

to carry on. He has provided guidance from start to finish. I will always value both

his friendship and his contribution to my growth as a graduate student and a

researcher.

Dr. Art Brown is deserving of acknowledgement for sharing his wealth of

knowledge of stream ecology and his enthusiasm for studying and protecting the

natural environment. Kip Heth deserves special recognition for invaluable assistance

with invertebrate identification. I thank Dr. Paul Vendrell for providing insight into

water quality issues. And I would also like to acknowledge Dwayne Rambo for his

cooperative spirit in sharing data that he collected in 1996 that was incorporated into

this study.

Dr. James Dunn and Lynnette Duncan of the Mathematical Sciences

Department made a major contribution to my understanding of statistical analyses.

They offered great patience, genuine interest, and creative options for analysis of my

data. I am most grateful for their contribution to this research.

v

The Arkansas Cooperative Fish and Wildlife Research Unit provided financial

support, office support, and equipment, as well as field technician assistance. I thank

Barbara Parker, Dennis Lichtenberg, and Casey Pevey for being there for me. Many

individuals spent time in the rivers with me, and I extend my appreciation for the help

of Sam Allen, Rebecca Dukes, Jacque Hill, Shane Jackson, April King, Mike Mason,

Danielle Pender, Scott Quinton, Rhonda Rimer, and Jennifer Robbins. I offer special

thanks to Rhonda Rimer for the many hours she gave and for her spirited appreciation

of this project, and to Jacque Hill for the generous hospitality he offered to us at his

home during field work. The landowners along the rivers were gracious enough to

allow us access to sampling sites, as well as cordial and willing to share information

about the unique river environment they know so well.

Malcolm Williamson and Anne Gisiger from the Center for Advanced Spatial

Technologies were most helpful with the Geographic Information Systems (GIS)

component of this study. Conclusions surrounding the data that they assisted me in

obtaining were among the most important findings in this study. Also, Terri and

Bruce Gorham have been great friends and colleagues – providing further assistance

with GIS analyses, as well as sharing an enthusiasm for my project.

For much of my life and this study, I had the loyal support of my sister, Karen.

I only wish she were still with me to see the completion of this work. I know she

would have expressed pride in my achievements, as she often did over the course of

my life. I have worked hard to live up to her expectations. Margie and Nick, my

remaining sister and brother, have taken a sincere interest in my work, which has been

positive for all of us.

vi

Sharing the “trials and tribulations” as well as the joys of this research with my

son, Brent, has served to motivate me to do my very best. We are role models for one

another. His expressed support and excitement over my accomplishments is truly

rewarding.

To all these people, I extend a most sincere thank-you – it could not have been

done without you.

vii

TABLE OF CONTENTS

Page ACKNOWLEDGEMENTS

iv

LIST OF TABLES

ix

LIST OF FIGURES

xii

ABSTRACT

1

INTRODUCTION

3

OBJECTIVES AND RELEVANCE

8

BACKGROUND

10

Ecological Integrity

10

History of Aquatic Bioassessment

11

Recent Bioassessment Protocols

13

Ecoregion Perspective

14

Fish Assemblages

15

Macroinvertebrate Assemblages

16

Instream Habitat and Riparian Vegetation

18

METHODS

19

Study Rivers and Sampling Sites

19

Field and Laboratory Procedures

21

Fish Assemblages

21


23


24

viii

TABLE OF CONTENTS (Continued)

Page

Water Quality

25

Geographic Information System Analysis

25

Statistical Methods

26

Cluster and Discriminant Function Analyses

26

Multidimensional Scaling

28

RESULTS

32

Biological Attributes

32

Fish Assemblages

32


53

Physical and Chemical Attributes

53


53

Water Quality

75

Watershed Attributes

75

River Comparisons

81

Biotic and Physical Variable Comparisons

81

River Grouping based on Similarities

87

Ranking Rivers Relative to Ideal Conditions

90

DISCUSSION

94

LITERATURE CITED

101

ix

LIST OF TABLES

Page

Table 1. Geographic location and description of watersheds and sampling sites.

22

Table 2. Original data sources and categories used in GIS analysis of watershed attributes. Data were derived from the digital data archives of the Center for Advanced Spatial Technologies (CAST), University of Arkansas.

27

Table 3. Values assigned to metrics for a conceptually ideal headwater river reach in the Boston Mountain ecoregion.

29

Table 4. Fish species present, metric classification, and number of sites where each was present.

33

Table 5. Fish assemblage characteristics for headwater reaches of 10 Boston Mountain rivers.

37

Table 6. Density and biomass estimates for fishes sampled in Big Piney Creek 15 October 1998.

39

Table 7. Density and biomass estimates for fishes sampled in Hurricane Creek 28 June 1996 (Rambo 1998).

40

Table 8. Density and biomass estimates for fishes sampled in Kings River 29 July 1998.

41

Table 9. Density and biomass estimates for fishes sampled in Middle Fork Illinois Bayou 9 July 1996 (Rambo 1998).

42

Table 10. Density and biomass estimates for fishes sampled in Mulberry River 12 September 1998.

44

Table 11. Density and biomass estimates for fishes sampled in North Fork Illinois Bayou 2 July 1996 (Rambo 1998).

46

Table 12. Density and biomass estimates for fishes sampled in Richland Creek 20 August 1998.

47

Table 13. Density and biomass estimates for fishes sampled in Upper Buffalo River 20 July 1998.

48

x

LIST OF TABLES (Continued)

Page

Table 14. Density and biomass estimates for fishes sampled in War Eagle Creek 23 July 1998.

49

Table 15. Density and biomass estimates for fishes sampled in White

River 14 July 1998. 51

Table 16. Macroinvertebrate assemblage characteristics for headwater reaches of 10 Boston Mountain rivers.

54

Table 17. Identity and number of macroinvertebrates sampled in Big Piney Creek 16 October 1998.

55

Table 18. Identity and number of macroinvertebrates sampled in Hurricane Creek 1 August 1998.

56

Table 19. Identity and number of macroinvertebrates sampled in Kings River 30 July 1998.

58

Table 20. Identity and number of macroinvertebrates sampled in Middle Fork Illinois Bayou 24 October 1998.

60

Table 21. Identity and number of macroinvertebrates sampled in Mulberry River 12 August 1998.

61

Table 22. Identity and number of macroinvertebrates sampled in North Fork Illinois Bayou 26 September 1998.

63

Table 23. Identity and number of macroinvertebrates sampled in Richland Creek 10 October 1998.

65

Table 24. Identity and number of macroinvertebrates sampled in Upper Buffalo River 20 July 1998.

67

Table 25. Identity and number of macroinvertebrates sampled in War Eagle Creek 27 July 1998.

69

Table 26. Identity and number of macroinvertebrates sampled in White River 14 July 1998.

71

Table 27. Instream habitat and riparian vegetation characteristics for

headwater reaches of 10 Boston Mountain rivers. 73

xi

LIST OF TABLES (Continued)

Page

Table 28. Water quality characteristics for headwater reaches of 10 Boston Mountain rivers sampled in summer.

76

Table 29. Water quality characteristics for headwater reaches of 10 Boston Mountain rivers sampled in winter.

78

Table 30. Watershed attributes for headwater reaches of 10 Boston Mountain rivers.

80

Table 31. Statistical characteristics of metrics used in cluster analysis and Guttman’s scaling.

82

Table 32. Pearson correlation coefficient (r) between variables with significance level (probability) of 0.05 or lower.

85

Table 33. Significant variables (p < 0.07) that distinguished between river groupings based on stepwise discriminant analysis.

89

Table 34. Variables that characterized North Fork Illinois Bayou ranked as closest to ideal and War Eagle Creek and White River ranked as furthest from ideal.

93

xii

LIST OF FIGURES

Page

Figure 1. Map of Boston Mountain ecoregion showing study rivers and sites.

20

Figure 2. Grouping of study rivers based on cluster analysis. 88

Figure 3. Guttman’s scale ranking rivers relative to conceptually ideal conditions using 34 variables describing biotic, physical, chemical, and watershed characteristics.

91

1

ABSTRACT The Wild and Scenic Rivers Act of 1968 was passed to protect free-flowing

rivers with outstanding ecological and social values and requires suitability studies as

part of the designation process. An objective, quantitative method to determine

suitability based on ecological integrity was developed and tested using headwater

reaches of 10 Ozark rivers, five with Wild and Scenic status. Thirty-four variables

representing macroinvertebrate and fish assemblage characteristics, instream habitat,

riparian vegetation, water quality, and watershed attributes were quantified for each

river and analyzed using two multivariate approaches. Two groups of rivers were

identified based on similarities using cluster analysis, and discriminant analysis

identified only one variable (% forested watershed) that reliably distinguished groups.

The second approach compared variables for each river to conceptually ideal

conditions that were developed as a composite of the optimal conditions among the 10

rivers, which may serve as least-disturbed ecoregion reference conditions. The

composite distance of each river from ideal was then calculated using

multidimensional scaling. Two rivers without Wild and Scenic designation ranked

highest relative to ideal (highest ecological integrity), and two others, also without

designation, ranked most distant from ideal (lowest ecological integrity). Fish density,

number of intolerant fish species, and invertebrate density were influential biotic

variables for scaling. Contributing physical variables included riparian forest cover,

nitrate concentration, turbidity, percentage of forested watershed, percentage of

private land ownership, and road density. These methods provide a framework for

refinement and application in other regions to facilitate the process of identifying

2

rivers that have retained high ecological integrity for protection under the Wild and

Scenic Rivers Act or other mechanism, use as least-disturbed reference streams in

biomonitoring, and providing benchmarks for restoration efforts.

3

INTRODUCTION Management of aquatic ecosystems to provide water supplies, hydropower,

flood control, transportation, recreation, and other benefits has altered the flow and

biological processes of a substantial portion of United States rivers and streams

(National Research Council 1992). Estimates of the number of high quality, free-

flowing rivers in the country range from less than 2% to 9% of the total river

kilometers (Stanford and Ward 1979, Echevarria and Fosburgh 1988, Benke 1990).

The pressure of increasing human population accompanied by intensified municipal,

industrial, and agricultural development has led to the degradation of water quality and

habitat leaving few river reaches unaffected by human activity (Karr 1995). In an

effort to fulfill national conservation goals, the United States Congress passed the

Wild and Scenic Rivers Act of 1968 (PL 90-542) to preserve free-flowing rivers with

outstanding natural and cultural values, and to protect the water quality of such rivers.

The Wild and Scenic Rivers system currently includes over 16,000 km of 157 rivers.

The Wild and Scenic Rivers system remains distinctly smaller and less well-

known than other federal protection programs, such as the National Park system and

the Wilderness program (Palmer 1993). In addition to the fact that few river

kilometers are sufficiently unaltered by human activity to be eligible for consideration,

the Wild and Scenic Rivers Act is underused in comparison to other federal

conservation efforts (Palmer 1993, Cassidy 1994). Benke (1990) concluded that

utilization of the act has been slow and uneven across the nation. Progress in

expanding river protection has been accomplished by state agencies and private

organizations, but these programs do not confer the level of protection afforded by

4

federal status (Palmer 1993). Accelerated effort toward preservation and restoration is

needed, regardless of how protection is to be afforded to the least disturbed rivers and

streams of the Nation (National Research Council 1992, Allan and Flecker 1993,

Naiman et al. 1995).

In view of the limited number of high quality, free-flowing river kilometers

that presently exist, identification, assessment, and protection of these reaches are

critical to meeting the Nation’s commitment to preservation of extraordinary natural

resources, protection of water quality, and river restoration. A primary objective of

the Wild and Scenic Rivers Act is to preserve rivers with exceptional scenic beauty,

recreational potential, and intangible social values. In addition, the preservation of

these rivers is important because they represent historic ecological conditions as

closely as can be expected given the degree of anthropogenic activity that has

occurred. The National Research Council (1992) recommended that relatively

undisturbed reference reaches be designated and protected in each ecoregion of the

United States. Bioassessment using comparisons among relatively undisturbed

reference streams and disturbed reaches within ecoregions has become the

recommended method of water quality monitoring (Ohio EPA 1987a, 1987b, 1989,

Plafkin et al. 1989, Karr 1991, Rosenberg and Resh 1993, Davis and Simon 1995,

Barbour et al. 1999). Protected reference reaches can also be used for comparison to

disturbed sites to establish restoration goals and monitor progress in achieving those

goals (Hughes et al. 1986, Hughes 1995).

The original Wild and Scenic Rivers Act of 1968 (PL 90-542) stated that

“certain selected rivers of the Nation which with their environments, possess

5

outstandingly remarkable scenic, recreational, geologic, fish and wildlife, historic,

cultural, or other similar values, shall be preserved in free-flowing condition, and that

they and their immediate environments shall be protected for the benefit and

enjoyment of present and future generations.” The Revised Guidelines for Eligibility,

Classification and Management of Rivers under the Wild and Scenic Rivers Act of

1982 (PL 96-487) clarified the process by which rivers are to be designated for

protection. The minimum reach length requirement was eliminated, and ecological

value was added as a value that qualifies a river for consideration. In addition,

eligibility and suitability requirements were clarified. The minimum eligibility

requirements are that the river segment be completely free-flowing and possess at least

one outstandingly remarkable value. It also mandated that all river segments

designated as potential additions to the system be studied as to their suitability for

inclusion. Suitability studies to be completed by the U.S. Forest Service, Bureau of

Land Management, National Park Service, or a state agency are required to determine

if it is in the public interest to confer Wild and Scenic River status to rivers that meet

eligibility requirements.

River segments that meet eligibility requirements are listed in the Nationwide

Rivers Inventory (U.S. National Park Service 1982). The recommended procedure for

conducting suitability studies is to compare a group of eligible rivers in a region,

consider a number of alternatives that have been made available for public review, and

finally render a decision as to which rivers should be recommended for Wild and

Scenic status (U.S. Forest Service 1987.)

6

In 1986, eligible rivers within the Ozark National Forest boundaries or

surrounded by adequate U.S. Forest Service land to facilitate management were

identified in the Ozark-St. Francis National Forests Final Environmental Impact

Statement for the Land and Resources Plan (U.S. Forest Service undated). After

modification and addition of river segments, 13 segments were selected and studied by

the U.S. Forest Service to determine suitability for inclusion in the Wild and Scenic

Rivers system (U.S. Forest Service 1988). The preferred alternative selected by the

U.S. Forest Service was recommendation of 6 of the 13 river segments that possess the

most outstandingly remarkable values and which respond to public issues (U.S. Forest

Service 1991). Based on this recommendation, an amendment to the Wild and Scenic

Rivers Act, known as the Arkansas Wild and Scenic Rivers Act (PL 102-275), was

passed in 1992, and river segments of Big Piney Creek, Buffalo River, Hurricane

Creek, Mulberry River, North Sylamore Creek, and Richland Creek became part of the

Wild and Scenic Rivers system.

River assessment based on the values originally described in the Wild and

Scenic Rivers Act (e.g. scenic, recreational, cultural) is a necessarily subjective

process involving social, political, and aesthetic considerations. Consideration of the

ecological value of rivers adds a more objective component to the process. Physical,

chemical, and biological attributes can be assessed to compare ecological integrity of

rivers using quantitative, repeatable, and objective standards. The development of a

protocol to assess suitability based on ecological integrity may serve to reduce the

degree of subjectivity involved in suitability studies and facilitate the process of

selecting rivers for protective status. In addition, information from such an approach

7

may be useful in identifying reference stream conditions for monitoring water quality

and establishing benchmarks for progress in stream restoration. The goal of this

research is to contribute an objective, quantitative method for achieving those ends.

8

OBJECTIVES AND RELEVANCE

This study evaluated the ecological integrity of 10 free-flowing rivers that

originate in the Boston Mountain ecoregion of northwestern Arkansas that were

originally identified in the Nationwide Rivers Inventory (U.S. National Park Service

1982) as eligible for study of suitability for Wild and Scenic River status. Seven of

these river segments were part of a suitability study completed by the Ozark-St.

Francis Unit of the U.S. Forest Service, including five designated Wild and Scenic,

two that were not recommended for inclusion in the system, as well as three others

that have not been studied for suitability.

Data on fish and macroinvertebrate assemblages, habitat quality,

physicochemical properties of water, and watershed attributes were incorporated into a

multimetric model designed to determine a suite of metrics that are most useful in

assessing ecological integrity of rivers in the Boston Mountain ecoregion of Arkansas.

The objectives of this research were to (1) develop and test an objective, quantitative

method of assessing rivers for suitability based on ecological integrity, (2) determine

what physical or biotic variables distinguish among rivers in the Boston Mountain

ecoregion, and (3) identify least-disturbed reference conditions for headwater reaches

of Boston Mountain ecoregion rivers.

This evaluation method may serve as a model for refinement and application in

other regions and facilitate development of more efficient evaluation protocols and

management strategies. On a national scale, information from this research is

intended to be useful in assessment of rivers considered for protection under the Wild

and Scenic River Act or any other protective mechanism available. Furthermore, the

9

identification of rivers with the highest degree of ecological integrity serves the

purpose of providing a standard for ecoregion least-disturbed reference conditions that

may be useful in monitoring or restoring surface water quality.

10

BACKGROUND

Environmental degradation from exploitation of water resources has led to the

development of a large body of information on aquatic ecosystem assessment. The

methods developed in my research to assess ecological integrity are a composite of

various protocols that have been shown to be effective, repeatable, and practical for

use by management agencies. Below, I define ecological integrity, provide a historical

perspective of aquatic bioassessment, discuss assessment protocols in present use, and

address issues considered in experimental design and interpretation of results of this

research.

Ecological Integrity

The term “integrity” is frequently used in water management legislation and

water quality assessment, but it is rarely explicitly defined. The stated goal of the

Federal Pollution Control Act of 1972 (PL 92-500) is to “restore and maintain the

chemical, physical and biological integrity of the nation’s waters.” Karr and Dudley

(1981) defined biotic integrity as the “ability to support and maintain a balanced,

integrated, adaptive community of organisms having a species composition, diversity,

and functional organization comparable to that of natural habitat of the region.” They

further defined ecological integrity as the “summation of chemical, physical and

biological integrity.” Determination of ecological integrity, then, is a holistic

approach to assessment using a wide range of physical, chemical, and biological

variables.

11

History of Aquatic Bioassessment

Bioassessment, the measurement of living resources to assess aquatic

ecosystems, dates back to the mid 1800s with the recognition of the relationship

between disease and poor water quality. The Saprobien system, introduced in Europe

in the early 1900s, associated benthic and planktonic plants and animals with the

degree of decomposition of sewage in slow-moving water. While the Saprobien

system and general concept of indicator organisms had support from such early

scientists as Forbes and Richardson (1913), the idea was not widely accepted in the

United States well into the twentieth century (Cairns and Pratt 1993). Modified

versions of the Saprobien system continue to be employed outside the United States,

and it is important historically in that it focused attention on measuring water resource

quality by the presence or absence of a range of indicator biota providing the

foundation for further development of bioassessment (Davis 1995).

Several numeric biotic indices have been developed including the Shannon-

Wiener Diversity Index, Gammon’s Index of Well-Being, Beck’s Biotic Index, and

Hilsenhoff’s Biotic Index (Davis 1995). They represent an effort to numerically

characterize biological data in meaningful and understandable ways. While these

indices remain in use, there is considerable debate about their validity and usefulness

(Cairns and Pratt 1993, Norris and Georges 1993, Merritt and Cummins 1996).

Specifically, diversity indices have been criticized for lack of a sound or well-

understood foundation in biological or ecological theory.

The introduction of the Index of Biotic Integrity (IBI), (Karr 1981) marked a

major advancement in bioassessment. This multiple-metric index, designed to

12

evaluate stream biotic integrity using fishes, combines several assemblage attributes

(metrics) to provide a single numeric score. Development of a number of multiple-

metric indices followed, including the Invertebrate Community Index (Ohio EPA,

1989), a multiple benthic index for Arkansas developed by Shackleford (1988), the

North Carolina Biotic Index (Lenat 1993), and a set of composite indices called Rapid

Bioassessment Protocols (RBPs) for benthic macroinvertebrates and fish (Plafkin et al.

1989). The various indices in use to assess fish and macroinvertebrate assemblages

include metrics that reflect community structure, taxonomic composition, individual

condition, and biological processes. In addition, multimetric habitat indices, such as

the Qualitative Habitat Evaluation Index (Rankin, 1989, 1995) and RBP Habitat

Quality Index (Plafkin et al. 1989), have been developed to evaluate the physical

structure of streams.

Water quality assessment in the United States accelerated in response to

passage of the Clean Water Act of 1977 (PL 95-217). However, emphasis on

monitoring point source pollution stressed the use of chemical criteria, a legacy that

continues today in some water resource management agencies. Numerous arguments

have been made and continue for the use of biological criteria (Karr 1995).

Recognition of problems associated with diffuse pollution (also known as nonpoint

source pollution) has drawn attention to the need for an ecological approach to water

quality monitoring (Novotny and Olem 1994). An accelerated effort to increase

acceptance of using an ecological rather than a chemical endpoint as the most

appropriate and effective method of managing aquatic resources is reflected in the

revision of the U.S. Environmental Protection Agency (EPA) Rapid Bioassessment

13

Protocols (Barbour et al. 1999). This document provides guidelines for a wide range

of bioassessment using benthic macroinvertebrates, fish, periphyton, and habitat

quality parameters. Multi-species and assemblage approaches are encouraged, and

measurement of physicochemical parameters of water receives attention only within

the context of broader habitat assessment.

Changes in land use and vegetative cover are becoming recognized as perhaps

the most important anthropogenic activities influencing ecosystems worldwide

(Schlosser 1991, Vitousek 1994). The effect of these changes on aquatic ecosystems

has focused attention on landscape scale assessment (Roth et al. 1996, Wang et al.

1997, Lammert and Allan 1999). The development of geographic information

systems (GIS) has provided the technical advance needed to extend our ability to

address the influence of watershed disturbance on stream ecological integrity (Allan

and Johnson 1997, Johnson and Gage 1997).

Recent Bioassassment Protocols

The standard approach to aquatic bioassessment is to compare biological,

physical, and chemical attributes of a stream reflected in a set of metrics to those of a

least-disturbed reference stream, i.e., a stream that is judged to be maintaining natural

structure and function within the same ecoregion (Hughes et al. 1986, Hughes 1995).

Streams are assigned a composite score that is the sum of the scores for each metric

based on proximity to that of the reference stream. An alternative approach is to make

comparisons to least-disturbed reference conditions using metrics as response

variables in multivariate statistical analyses. Norris and Georges (1993) suggest that

14

multivariate approaches hold promise for community level biomonitoring in the

future. Statistical approaches are widely accepted in stream ecology research, but

have not been adopted by most water resource management agencies.

Ecoregion Perspective

Comparison of stream attributes to least-disturbed reference conditions in the

same ecoregion is central to contemporary bioassessment protocols. Because potential

natural vegetation, geology, soils, and other environmental factors vary across a

landscape, efforts have been made to identify regional patterns. Abiotic and biotic

factors that characterize ecoregions are important in assessment because they define

the biological expectations of regions (Karr, 1991, Omernik 1995). Bailey (1983) and

McNab and Avers (1994) developed stratified classifications of ecoregions based

primarily on differences in vegetation. To classify streams for more effective water

quality management, Omernik (1987) identified homogeneous areas analyzing land

surface form, soils, land use, as well as potential natural vegetation. Because

Omernik’s work is related to stream classification, his delineation of ecoregions is

widely used by aquatic ecologists and water resource managers (Ohio EPA 1987a,

1987b, 1989, Hughes et al. 1990, Karr 1991, ADPCE 1998, Barbour et al. 1999).

Rohm et al. (1987) surveyed 22 rivers in the six ecoregions in Arkansas defined by

Omernik (1987), and concluded that streams of the same ecoregion were more similar

than those of other ecoregions, based on fish assemblages, physical habitat, and water

chemistry variables. Ecological integrity comparisons among streams within an

15

ecoregion minimizes the need to compensate for biogeographic variation that occurs

among ecoregions.

Fish Assemblages

There are many approaches to assessing biotic integrity using fish

assemblages. One of the most widely used protocols is the Index of Biotic Integrity

(IBI), (Karr et al. 1986). Twelve metrics are used to represent fish assemblages in

warmwater Midwestern streams: (1) species richness, (2) number and identity of

darter species (Etheostomatinae), (3) number and identity of sunfish species

(Centrarchidae), (4) number and identity of sucker species (Catostomidae), (5) number

and identity of intolerant species, (6) proportion of individuals as green sunfish

(Lepomis cyanellus ), (7) proportion of individuals as omnivores, (8) proportion of

individuals as insectivorous cyprinids, (9) proportion of individuals as piscivores, (10)

number of individuals in sample, (11) proportion of individuals as hybrids, and (12)

proportion of individuals with disease, tumors, fin damage, and skeletal anomalies. In

the presence of disturbance, species richness and number of species or proportions of

darters, sunfishes, suckers, intolerant species, piscivores, and insectivorous cyprinids

are expected to decline, while the proportions of green sunfish and omnivores are

expected to increase. The number of individuals in the sample, which may be

expressed as fish density, may decline as a result of some forms of chemical pollution,

physical disturbance, or introduction of non-native species. However, nutrient loading

associated with agriculture or municipal sewage inflows may result in increased fish

density and changes in community structure associated with a greater proportion of

16

tolerant species (Yoder and Smith, 1999). Metric 11, presence of hybrids in the

sample, is relevant in situations involving non-native introductions. Conditions

described in metric 12, disease, tumors, fin damage, and skeletal anomalies, are most

often associated with sites below point source pollution or where toxic chemicals are

concentrated (Barbour et al. 1999). Selection and calibration of IBI metrics requires

consideration of regional differences in fish distribution and assemblage structure and

function. Alternative metrics appropriate for use in various parts of the Nation are

listed in the EPA Rapid Bioassessment Protocols (Barbour 1999).

Suter (1993) presents several criticisms of biotic indices including lack of

diagnostic power, effects of one component being eclipsed by other components, and

failure to provide a reason for high or low index values. Despite such criticisms, the

IBI remains the most widely accepted method of bioassessment using fish

assemblages, and refinements based on regional differences continue to expand its

usefulness.


While there is some degree of consensus on bioassessment of fish

assemblages, there are many different approaches to assessment of benthic

macroinvertebrates. Resh and Jackson (1993) reviewed 30 stream invertebrate

protocols presently in use, differing in sampling devices, habitat sampled, proportion

of the sample examined, level of taxon identification, or criteria used to determine

degree of impairment. Qualitative (kick nets) and quantitative (Surber or Hess

samplers) methods have been employed. Shackleford (1988) and Barbour and

17

Gerritsen (1996) advocated the use of fixed-sample counts. On the other hand,

Courtemanch (1996) argued that fixed counts may not adequately reflect species

richness. A wide range of metrics to represent macroinvertebrate assemblages has

been adopted; Resh and McElray (1993) listed 17 different indices used in lotic

studies of benthic invertebrates. Rosenberg and Resh (1993) suggested that the

number of indices in use may reflect a dissatisfaction with the results they provide. In

summary, there is presently a lack of consensus on many issues related to

macroinvertebrate collection and analysis.

Merritt and Cummins (1996) reviewed the rationale for a number of metrics

that are currently used in macroinvertebrate assessment. Taxa richness with resolution

to genus remains the most commonly used measure based on the premise that it

decreases in response to a decline in water quality. EPT (Ephemeroptera, Plecoptera,

Trichoptera) richness is also commonly applied because most taxa in these orders are

believed to be pollution sensitive, and because identification of these groups may be

easier than other taxa. Total number of individuals, which may be expressed as

invertebrate density, is presumed to decrease with pollution, but as is the case with

fish density, increased abundance may result from nutrient loading. The ratio of EPT

abundance to Chironomidae abundance may be used to reflect changes in community

structure associated with the ratio of intolerant to tolerant species. As previously

discussed, diversity indices, which are intended to reflect taxa richness as well as

evenness, remain widely used despite criticism of their validity. Karr (1998)

recommended percent dominant taxa (ratio of individuals in numerically dominant

taxa to total number of individuals) as an alternative to diversity indices.

18


Habitat indices such as the Qualitative Habitat Evaluation Index (Rankin 1989,

1995) and RBP Habitat Quality Index (Plafkin et al. 1989) are designed to assess the

physical integrity of streams by assessing characteristics known to be associated with

support of stream biota. Characteristics that are commonly evaluated include water

velocity and depth, substrate composition, bank condition, and riparian vegetation.

Since these characteristics are known to vary among landscapes, evaluation of their

ability to sustain healthy biota requires comparison to conditions known to be

minimally disturbed by human activity within the same ecoregion.

19

METHODS

Study Rivers and Sampling Sites

Selection of study rivers was based on two criteria – eligibility for Wild and

Scenic River status and origination in the Boston Mountain ecoregion of Arkansas.

The Nationwide Rivers Inventory (U.S. National Park Service 1982) and the Ozark-St.

Francis National Forests Final Environmental Impact Statement for the Land and

Resources Management Plan (U.S. Forest Service undated) were used to identify

rivers that met those criteria. The 10 rivers selected include Big Piney Creek,

Hurricane Creek, Kings River, Middle Fork Illinois Bayou, Mulberry River, North

Fork Illinois Bayou, Richland Creek, Upper Buffalo River, War Eagle Creek, and

White River (Figure 1).

Big Piney Creek, Hurricane Creek, Mulberry River, Richland Creek, and

Upper Buffalo River were conferred Wild and Scenic status based on the suitability

study described in the Wild and Scenic River Study Report and Final Environmental

Impact Statement of Thirteen Rivers in the Ozark National Forest (U.S. Forest Service

1991). It should be noted that the Buffalo River is a National River under

management by the National Park Service with the exception of the headwaters. The

Wild and Scenic designated portion of the river encompasses the headwaters and is

referred to as the Upper Buffalo River. Middle and North Forks of the Illinois Bayou

were included in the suitability study, but were not recommended for designation, and

the Kings, War Eagle, and White rivers were not studied because they are not within

Ozark National Forest boundaries or surrounded by adequate U.S. Forest Service land

to facilitate management.

20

21

A sampling site was selected on each river based on similar watershed sizes

(Table 1). The watersheds, encompassing the area from the headwaters to the sampling

site, are fully contained within the Boston Mountain ecoregion which consists of low

mountains with dense oak/hickory tree cover as described by Omernik (1987).

Field and Laboratory Procedures

Fish Assemblages

Fishes were sampled between July and October, 1998, from Big Piney Creek,

Kings River, Mulberry River, Richland Creek, Upper Buffalo River, War Eagle Creek,

and the White River. Because identical fish sampling techniques were used in 1996 on

Hurricane Creek and the Middle and North Forks of the Illinois Bayou (Rambo 1998),

those data were incorporated into my research. A 1,500-W two-electrode pulsed D.C.

barge electrofisher was used to collect fish in representative pool-riffle sequences ranging

from 22 m to 80 m. I followed procedures for a three-pass removal method described by

Bohlin et al. (1989) to estimate fish populations. Fish were collected for an equal period

of time for each pass, with time per pass ranging from 35 to 56 minutes depending on

sampling reach length. Fish from each pass were held separately; those that could be

identified stream-side were measured for length (TL + 1mm) and weight (+ 0.01g) and

released. Small or unidentified fish were preserved in 10% buffered formalin, returned to

the laboratory where they were rinsed, transferred to 70% ethanol, identified, and

measured for length and weight.

22

Table 1. Geographic location and description of watersheds and sampling sites. River

County

Digital elevation

models*

Digital line graphs

30x60-minute series*

Watershed size (ha)

UTM coordinates of

sampling site Big Piney Creek

Newton

Fallsville, Ozone, Rosetta,

Swain

Fly Gap Mountain

4,358

3958636N 464605E

Hurricane Creek

Newton Deer, Lurton Fly Gap Mountain 5,012 3957401N 487138E

Kings River

Madison, Newton Boston, Weathers Fly Gap Mountain 6,836 3980162N 450386E

Middle Fork Illinois Bayou

Pope, Searcy Moore, Smyrna, Tilly Witts Spring

Mountain View

6,775 3946730N 508431E

Mulberry River

Johnson, Newton Fallsville, Oark, Ozone

Fly Gap Mountain 7,488 3949266N 452240E

North Fork Illinois Bayou

Newton, Pope Sand Gap, Smyrna Fly Gap Mountain, Mountain View

4,352 3947715N 498201E

Richland Creek

Newton Lurton, Moore, Sand Gap, Smyrna,

Fly Gap Mountain, Mountain View

6,646 3958598N 501321E

Upper Buffalo River

Newton Boston, Boxley, Fallsville, Weathers


War Eagle Creek

Madison Boston, Japton, Pettigrew, St. Paul, Weathers, Witter


White River

Franklin, Johnson, Madison

Boston, Pettigrew, St. Paul Fly Gap Mountain 9,591 3964129N 430886E

*See Table 2 for data sou

23

Estimates of density and biomass were computed from catch, length, and

weight for each species using the three-pass removal, maximum likelihood method

computed with Pop/Pro software (Seber 1982, Bohlin et al. 1989, Kwak 1992). Fish

density, biomass, and metrics based on the IBI (Karr et al. 1986) were used to

characterize fish assemblages including species richness; numbers of darter, sunfish,

sucker, and intolerant species; and proportions of individuals as green sunfish,

omnivores, insectivorous cyprinids, and piscivores. Fishes were grouped into these

categories based on life history information found in Robinson and Buchanan (1988).

Other IBI metrics, such as the proportion of individuals as hybrids, were not used

because non-native species were absent, making hybridization unlikely. The

proportion of individuals with disease, tumors, fin damage, and skeletal anomalies was

not included because these conditions are usually associated with high levels of

pollution, a situation that does apply to these streams.


Three consecutive riffles in each stream, including the riffle that was

electrofished, were sampled for macroinvertebrates using a Hess sampler with a 600-

µm mesh net. Samples were preserved in 70% ethanol. Individuals were identified to

the genus level (with the exception of Chironomidae and Oligochaeta) according to

Merritt and Cummins (1996) and Poulton and Stewart (1991). Metrics used to

characterize macroinvertebrate assemblages included taxa richness; total abundance

expressed as density; Ephemeroptera, Plecoptera, Tricoptera (EPT) richness;

proportion of individuals as Chironomidae; ratio of EPT number to Chironomidae

24

number, and percent dominant taxa, calculated as the ratio of the number of

individuals from the three most abundant taxa to total abundance.


Instream habitat surveys were conducted within a 250-m river reach that

included the electrofishing and invertebrate sampling sites. Ten cross-sectional

transects were selected perpendicular to stream flow. Location of the first transect

was selected randomly, and all subsequent transects were spaced at 25-m intervals.

Within each 25-m section, the following measurements were made as a percent of the

total surface area or stream bank length: pool, riffle, run, eroded bank, undercut bank,

fish cover (including submerged and emerged vegetation, root wads, fine and coarse

woody debris, boulders, and rock-ledges), and shade over the water surface between

1100 and 1300 hours. Points at 1.0-m intervals were measured to characterize depth,

velocity, substrate composition, and percent embeddedness. Depth was measured

using a top-set wading rod, and a Marsh-McBirney Model 2000 flow meter was used

to measure mean water column velocity. Substrate composition was based on the

modified Wentworth particle size scale (Bovee and Milhous 1978); the two most

abundant substrates categories were used to characterize each square meter along the

transect. Data were condensed into percent bedrock, boulder, cobble, gravel, and sand

and silt with the most abundant category considered twice as abundant as the second

most abundant category. Bank angles of the water-land interface were measured with

a clinometer, and bank-full widths were visually determined and measured at each

transect with a range-finder. Riparian vegetation was assessed by measuring the

25

proportion of undisturbed forest, shrub, pasture, and road within a 50-m lateral buffer

at each transect.

Water Quality

Water samples were collected from each sampling reach in summer at base

flow (21 July to 26 October 1998) and in winter at higher flow (but not within seven

days of rain) (22-26 February 1999). Water temperature, dissolved oxygen, and

specific conductance were measured in the field. A 1-L sample was collected for

laboratory analysis of alkalinity; hardness; pH; specific conductance; total dissolved

solids; turbidity; and sulfate, chloride, and nitrate-nitrogen concentrations. A 40-ml

sample was collected for analysis for total reactive orthophosphate. A 0.5-L sample

was collected and preserved with H2SO4 to reduce pH to less than 2 and was analyzed

for concentrations of total Kjeldahl nitrogen, ammonia-nitrogen, and total phosphorus

concentrations. Laboratory analyses were completed by the Arkansas Water

Resources Center – Water Quality Laboratory, University of Arkansas.

Geographic Information Analysis

Geographical Resource Analysis Support System (GRASS), (USACERL

1993), a raster-based GIS, was used to map watersheds of the rivers at each study site.

The Digital Elevation Maps (DEMs) based on 7.5 minute 30-m U.S. Geological

Survey quadrangles that encompass the watershed of each river were patched together

using the routine, r.patch, to generate watershed DEMs. The watershed basin analysis

program, r.watershed 4.0, was then used to delineate the watershed boundary of each

river by inputting the UTM coordinates of the sampling site as the outlet point at a

26

resolution of 30 m. The result was a watershed basin map from sampling site to

headwaters. This base map provided the template upon which map layers were then

created for analysis of watershed land cover and ownership and road density in the

basin and within a 100-m buffer of the river channel. Specific data sources and

categories used in analysis are reported in Table 2.

Statistical Methods

Two multivariate approaches were used to compare biotic and physical

characteristics of the rivers studied. The first method involved clustering rivers on the

basis of their similarities, followed by stepwise discriminant analysis and discriminant

function analysis. A multidimensional scaling method based on paired comparisons

and rank order was also used (Guttman 1946). For both methods, 34 variables were

incorporated representing biota, instream habitat, riparian vegetation, water quality,

and watershed attributes. A Pearson correlation matrix was created to examine

correlation among these variables.

Cluster and Discriminant Function Analyses

Variables on a percent scale were transformed into log ratios or logits as

appropriate, and percents, log ratios, and logits were used in analysis. Hierarchical

cluster analysis was performed on the data using PROC CLUSTER (SAS Institute Inc.

1990) standardized using the AVERAGE method. Stepwise discriminant analysis

(PROC STEPDISC) (SAS Institute Inc. 1990) was then used to identify the

combination of variables responsible for the grouping. Finally, discriminant function

analysis (PROC DISCRIM) (SAS Institute Inc. 1990) was used to determine the linear

27

Table 2. Original data sources and categories used in GIS analysis of watershed attributes. Data were derived from the digital data archives of the Center for Advanced Spatial Technologies (CAST), University of Arkansas. Mapping

Data source

Categories

Watershed delineation

U.S. Geological Survey 7.5-minute 30-m Digital Elevation Models (DEMs) Level 2.

Streams Topologically Integrated Geographic Encoding and Referencing System (TIGER) HYDRO, 1992. U.S. Department of Commerce, Bureau of the Census, 1:100,000-scale topographic maps.

Land cover CAST GAP Analysis Project veg_state_urban30, 1992.

Forest, riverine vegetation, riverine bare, water, agricultural wet and dry, pasture.

Land ownership CAST GAP Analysis Project land management and ownership, 1992.

Private, U.S. Forest Service National Forest, U.S. Forest Service Wilderness, State of Arkansas, U.S. National Park Service, water, other.

Roads Digital Line Graphs (DLGs), 1983. 1:100,000-scale 30x60-minute series.

Class 1 - interstate and U.S. numbered highways, Class 2 - state and county highways, Class 3 - road or street without assignment of administrative responsibility, Class 4 - residential streets, Class 5 - four-wheel drive trails.

28

relationship among significant variables. Statistical analyses were performed using

SAS version 8 (SAS Institute Inc. 1990).

Multidimensional Scaling

Guttman’s scaling is a multidimensional scaling method of making paired

comparisons and assigning rank order (Guttman 1946). To utilize this approach, a set

of values is established, against which all observed values are compared. For my

study, this set of values represents the characteristics of a conceptually ideal headwater

river reach in the Boston Mountain ecoregion. The ideal value for each of the 34

variables was set at the upper or lower end of the range (maximum or minimum) or

median for the ten rivers based on its theoretical relationship to ecological integrity

(Table 3). For example, because species richness is considered to be directly

correlated with ecological integrity, the value was set at the maximum found for the

ten rivers. For attributes considered to be inversely correlated with ecological

integrity, such as the proportion of sand and silt as substrate, the minimum was used.

An intermediate value, the median, was used for attributes that are believed to show a

curvilinear relationship with disturbance; that is, in situations where either high or low

values are indicative of disturbance.

Guttman’s scaling uses a matrix method to compare the value for each variable

for each river to the corresponding ideal value. The absolute deviation from the

maximum, minimum, or median ideal value was calculated. The relative values

obtained were then used to rank the rivers for each variable. The ideal river was

always assigned rank 1, and no ties were allowed with the ideal river. Finally, ranks

29

Table 3. Values assigned to metrics for a conceptually ideal headwater river reach in the Boston Mountain ecoregion. Metric

Ideal value

Fish assemblage characteristics

Density (fish/ha)

Median

Biomass (kg/ha)

Median

Species richness

Maximum

Number of darter species

Maximum

Number of sunfish species

Maximum

Number of sucker species

Maximum

Number of intolerant species

Maximum

Proportion of individuals as green sunfish (%)

Minimum

Proportion of individuals as omnivores(%)

Minimum

Proportion of individuals as insectivorous cyprinids (%)

Maximum

Proportion of individuals as piscivores (%)

Maximum

Invertebrate assemblage characteristics

Taxa richness

Maximum

Density (invertebrates/m2 )

Median

Proportion of individuals as EPT (%)

Maximum

Proportion of individuals as Chironomidae (%)

Median

Ratio of EPT number to Chironomidae number

Maximum

Dominance - proportion of 3 most abundant taxa (%)

Minimum

30

Table 3. Continued. Metric

Ideal value

Water quality

Nitrate concentration summer (mg/L)

Minimum

Nitrate concentration winter (mg/L)

Minimum

Alkalinity summer (mg/L)

Median

Alkalinity winter (mg/L)

Median

Turbidity summer (NTU)

Minimum

Turbidity winter (NTU)

Minimum

Riparian vegetation and instream habitat

Proportion of 50-m riparian buffer as forest (%)

Maximum

Mean depth (cm)

Median

Mean velocity (m/s)

Median

Proportion of substrate as sand and silt (%)

Minimum

Pool to riffle ratio

Median

Proportion of bank eroded (%)

Median

Proportion of stream area with fish cover (%)

Maximum

Watershed attributes

Proportion of watershed as forest (%)

Maximum

Proportion of watershed as private land (%)

Minimum

Density of roads (km/ha)

Minimum

Density of roads in 100-m stream buffer (km/ha )

Minimum

31

were used to compute Guttman’s scale, a one-dimensional scale that describes a

composite distance between rivers, with the ideal receiving the highest score, and the

river farthest away from the ideal receiving a score of zero.

32

RESULTS

Biological Attributes

Fish Assemblages

Thirty-seven fish species from nine families were collected from the 10 rivers

(Table 4). Slender madtom (Noturus exilis) and longear sunfish (Lepomis megalotis)

were the only species common to all rivers. Eleven species were found at only one

site. Species richness ranged from 10 species in Richland Creek and the Upper

Buffalo River to 19 in the Mulberry and White rivers (Table 5). The North Fork

Illinois Bayou had the highest number of intolerant species. The percentage of

individuals as omnivores varied greatly from a low of 0.3% in the Kings River where

no central stonerollers (Campostoma anomalum) were caught to 51.1% in Big Piney

Creek where both central stonerollers and bluntnose minnows (Pimephales

notatus) contributed to the omnivore category. Lampreys were collected in the

Mulberry River, but because the specimens were juvenile, species identification was

not possible.

Density and biomass estimates according to species and river are presented in

Tables 6-15. Fish density among rivers ranged from 8,676 fish/ha to 46,150 fish/ha

with a mean of 22,328 fish/ha. The Upper Buffalo River had the highest fish density,

dominated by central stonerollers and rainbow darters (Etheostoma caeruleum)

accounting for 73% of the total. The White River, with the third highest density,

showed a similar pattern of dominance with 62% of the total consisting of the same

two species. Fish biomass ranged from 26.8 kg/ha to 202.9 kg/ha with a mean of

117.87 kg/ha. War Eagle Creek had the highest fish biomass.

33

Table 4. Fish species present, metric classification, and number of sites where each was present. ___________________________________________________________________ Family and species Classification Number of sites ___________________________________________________________________ Petromyzonitidae

Lamprey (Ichthyomyzon spp.)

none 1

Cyprinidae

Central stoneroller (Campostoma anomalum)

omnivore 9

Whitetail shiner (Cyprinella galactura)

insectivorous 1

Bigeye shiner (Notropis boops)

insectivorous, intolerant 5

Bluntnose minnow (Pimephales notatus)

omnivore 5

Duskystripe shiner (Luxilus pilsbryi)

insectivorous 4

Hornyhead chub (Nocomis biguttatus)

insectivorous 1

Ozark minnow (Notropis nubilus)

intolerant 3

Ozark shiner (Notropis ozarcanus)

none 1

Rosyface shiner (Notropis rubellus)

none 1

Telescope shiner (Notropis telescopus)

insectivorous 1

34

Table 4. Continued. ___________________________________________________________________ Family and species Classification Number of sites ___________________________________________________________________

Creek chub (Semotilus atromaculatus)

none 4

Catostomidae

Northern hog sucker (Hypentelium nigricans)

omnivore, intolerant 8

Black redhorse (Moxostoma duquesnei)

intolerant 2

Golden redhorse (Moxostoma erythrurum)

omnivore 2

Ictaluridae

Yellow bullhead (Ictalurus natalis)

none 1

Slender madtom (Noturus exilis )

none 10

Fundulidae

Northern studfish (Fundulus catenatus)

none 2

Blackspotted topminnow (Fundulus olivaceous)

none 4

Atherinidae

Brook silverside (Labidesthes sicculus)

none 2

35

Table 4. Continued. ___________________________________________________________________ Family and species Classification Number of sites ___________________________________________________________________ Centrarchidae

Shadow bass (Ambloplites ariommus)

piscivore, intolerant 1

Ozark bass (Ambloplites constellatus)

piscivore 3

Green sunfish (Lepomis cyanellus)

piscivore 9

Bluegill (Lepomis macrochirus)

none 1

Longear sunfish (Lepomis megalotis)

none 10

Smallmouth bass (Micropterus dolomieu)

piscivore, intolerant 8

Spotted bass (Micropterus punctulatus)

piscivore 2

Percidae

Greenside darter (Etheostoma blennoides)

none 9

Rainbow darter (Etheostoma caeruleum)

none 5

Fantail darter (Etheostoma flabellare)

none 4

Yoke darter (Etheostoma juliae)

none 1

36

Table 4. Continued. ___________________________________________________________________ Family and species Classification Number of sites ___________________________________________________________________

Stippled darter (Etheostoma punctulatum)

intolerant 7

Orangethroat darter (Etheostoma spectabile)

intolerant 9

Redfin darter (Etheostoma whipplei)

intolerant 5

Banded darter (Etheostoma zonale)

none 5

Logperch (Percina caprodes)

none 1

Cottidae

Banded sculpin (Cottus carolinae)

none 2

37

Table 5. Fish assemblage characteristics for headwater reaches of 10 Boston Mountain rivers. River

Total density

(fish/ha)

Total biomass

(kg/ha)

Species richness

No. of

darter species

No. of sunfish species

No. of sucker species

Big Piney Creek

17,185

86.13

12

3

2

2

Hurricane Creek

8,676 26.82 14 6 2 0

Kings River

17,480 150.85 15 4 3 1


17,965 154.77 17 6 3 1

Mulberry River

10,983 74.79 19 5 3 2


18,140 89.93 15 5 2 2

Richland Creek

21,309 146.24 10 3 2 0

Upper Buffalo River

46,150 95.28 10 3 1 1

War Eagle Creek

33,885 202.85 18 5 3 2

White River

31,509 151.03 19 4 3 1

38

Table 5. Continued. River

No. of

intolerant species

Green sunfish

(%)

Omnivores

(%)

Insectivorous

cyprinids (%)

Piscivores

(%)

Big Piney Creek

4

5.75

51.09

16.69

6.30

Hurricane Creek

3 0.37 7.70 7.60 1.92

Kings River

3 4.02 0.34 29.63 13.09


6 0.67 5.08 9.35 2.43

Mulberry River

6 1.29 46.26 7.17 4.27


7 0.26 16.32 5.46 1.26

Richland Creek

1 5.51 15.75 6.12 5.51

Upper Buffalo River

3 0 42.85 0.38 0

War Eagle Creek

4 0.79 14.77 5.55 9.81

White River

2 2.65 30.31 8.20 6.56

39

Table 6. Density and biomass estimates for fishes sampled in Big Piney Creek 15 October 1998. Values in parentheses are ± 2 standard error (SE). Species

Density (fish/ha)

Biomass (kg/ha)

Central stoneroller

5,713 (± 830)

29.21 (± 10.82)

Bigeye shiner

2,868 (± 168)

1.34 (± 0.13)

Bluntnose minnow

2,993 (± 4,530)

2.10 (± 3.20)

Northern hog sucker

51 (± 35)

1.07 (± 1.97)

Golden redhorse

23* 0.04*

Slender madtom

421*

0.82*

Green sunfish

988 (± 3,376)

33.90 (± 80.63)

Longear sunfish

1,903 (± 5,143)

11.23 (± 53.50)

Smallmouth bass

95 (± 11)

5.16 (± 4.00)

Greenside darter

433 (± 1,510)

0.46 (± 1.61)

Orangethroat darter

1,537 (± 110)

0.57 (± 0.09)

Redfin darter

160 (± 97)

0.23 (± 0.19)

Total

17,185 (± 7,071)

86.13 (± 97.54)

* population not depleted; minimum summing 3 passes.

40

Table 7. Density and biomass estimates for fishes sampled in Hurricane Creek 28 June 1996 (Rambo 1998). Values in parentheses are ± 2 SE. Species

Density (fish/ha)

Biomass (kg/ha)

Central stoneroller

651 (± 200)

1.88 (± 0.37)

Bigeye shiner

659 (± 84)

1.26 (± 0.20)

Bluntnose minnow

17*

0.04*

Creek chub

35 (± 12)

0.11 (± 0)

Slender madtom

712 (± 367)

1.38 (± 0.69)

Green sunfish

32 (± 42)

0.97 (± 1.74)

Longear sunfish

1,100 (± 261)

14.25 (± 6.98)

Smallmouth bass

135*

3.23*

Greenside darter

477 (± 93)

0.80 (± 0.16)

Fantail darter

1,046 (± 509)

0.65 (± 0.32)

Stippled darter

17*

0.04*

Orangethroat darter

3,369 (± 449)

1.81 (± 0.24)

Redfin darter

96 (± 72)

0.11 (± 0.09)

Banded darter

330 (± 768)

0.29 (± 0.70)

Total

8,676

(± 1,147) 26.82

(± 7.29)

*population not depleted; minimum summing 3 passes.

41

Table 8. Density and biomass estimates for fishes sampled in Kings River 29 July 1998. Values in parentheses are ± 2 SE. Species

Density (fish/ha)

Biomass (kg/ha)

Duskystripe shiner

4,707 (± 596)

1.30 (± 0.27)

Northern hog sucker

60*

1.56*

Slender madtom

1,924*

5.63*

Northern studfish

60*

0.08*

Blackspotted topminnow

721*

1.11*

Ozark bass

1,466 (± 1298)

54.33 (± 52.01)

Green sunfish

703 (± 918)

29.07 (± 39.62)

Longear sunfish

5,293 (± 725)

52.85 (± 8.74)

Smallmouth bass

60*

1.02*

Spotted bass

60*

0.19*

Greenside darter

131 (± 89)

0.69 (± 1.10)

Rainbow darter

1,995 (± 487)

1.91 (± 0.77)

Stippled darter

60*

0.49*

Orangethroat darter

120*

0.07*

Banded sculpin

120*

0.55*

Total

17,480 (± 1,911)

150.85 (± 65.98)


42

Table 9. Density and biomass estimates for fishes sampled in Middle Fork Illinois Bayou 9 July 1996 (Rambo 1998). Values in parentheses are ± 2 SE. Species

Density (fish/ha)

Biomass (kg/ha)

Central stoneroller

753 (± 164)

4.70 (± 1.29)

Bigeye shiner

1,677 (± 66)

2.92 (± 0.17)

Bluntnose minnow

142*

0.18*

Northern hog sucker

16*

0.01*

Yellow bullhead

16*

0.76*

Slender madtom

1,261*

2.51*

Brook silverside 32*

0.08*

Shadow bass

32*

0.21*

Green sunfish

120 (± 110)

3.51 (± 3.45)

Longear sunfish

9,108 (± 73,477)

119.75 (± 1,172.41)

Smallmouth bass

285 (± 58)

16.01 (± 3.40)

Greenside darter

260 (± 32)

0.89 (± 0.22)

Fantail darter

681 (± 100)

0.53 (± 0.11)

Stippled darter

63*

0.18*

Orangethroat darter

3,425 (± 1,497)

2.38 (± 1.06)

43

Table 9. Continued.

Species

Density (fish/ha)

Biomass (kg/ha)

Redfin darter

47*

0.09*

Banded darter

47*

0.07*

Total

17,965 (± 73,492)

154.78 (± 1,172.42)


44

Table 10. Density and biomass estimates for fishes sampled in Mulberry River 12 September 1998. Values in parenthesis are ± 2 SE. Species

Density (fish/ha)

Biomass (kg/ha)

Lamprey species

102*

0.20*

Central stoneroller

3,906 (± 57)

15.39 (± 1.22)

Bigeye shiner

788 (± 47)

1.60 (± 0.28)

Bluntnose minnow

1,135 (± 707)

0.50 (± 0.30)

Northern hog sucker

20*

7.55*

Golden redhorse

20*

0.40*

Slender madtom

305*

0.84*


228 (± 32)

0.82 (± 0.19)

Brook silverside

41*

0.06*

Green sunfish

142*

4.27*

Bluegill

20*

0.10*

Longear sunfish

2,317 (± 824)

27.27 (± 14.67)

Smallmouth bass

286 (± 423)

13.93 (± 23.58)

Spotted bass

41*

0.44*

Greenside darter

41*

0.05*

Fantail darter

20*

0.01*

Stippled darter

298 (± 933)

0.61 (± 1.91)

45

Table 10. Continued.

Species

Density (fish/ha)

Biomass (kg/ha)

Orangethroat darter

1,148 (± 179)

0.50 (±0.09)

Redfin darter

125 (± 21)

0.25 (± 0.09)

Total

10,983 (± 1,505)

74.79 (± 27.86)


46

Table 11. Density and biomass estimates for fishes sampled in North Fork Illinois Bayou 2 July 1996 (Rambo 1998). Values in parentheses are ± 2 SE. Species

Density (fish/ha)

Biomass (kg/ha)

Central stoneroller

1,321 (± 516)

8.08 (± 4.80)

Bigeye shiner

991 (± 60)

1.16 (± 0.14)

Bluntnose minnow

1,000 (± 186)

1.76 (± 0.29)

Creek chub

536 (± 311)

0.59 (± 0.13)

Northern hog sucker

640*

3.22*

Black redhorse

12*

1.79*

Slender madtom

350*

1.01*

Green sunfish

48*

3.14*

Longear sunfish

7,103 (± 16,135)

54.39 (± 128.42)

Smallmouth bass

181 (± 34)

8.85 (± 2.00)

Greenside darter

3,039 (± 300)

4.42 (± 0.42)

Fantail darter

12*

0.01*

Stippled darter

24*

0.04*

Orangethroat darter

2,804 (± 608)

1.30 (± 0.28)

Redfin darter

79 (± 31)

0.18 (± 0.13)

Total

18,140 (± 16,162)

89.94 (± 128.53)


47

Table 12. Density and biomass estimates for fishes sampled in Richland Creek 20 August 1998. Values in parentheses are ± 2 SE. Species

Density (fish/ha)

Biomass (kg/ha)

Central stoneroller

3,357 (± 326)

12.00 (± 3.08)

Duskystripe shiner

1,305 (± 280)

0.58 (± 0.27)

Ozark minnow

1,332 (± 315)

1.60 (± 0.30)

Creek chub

60*

0.03*

Slender madtom

5,246 (± 12,789)

35.54 (± 87.24)

Green sunfish

1,174 (± 520)

36.61 (± 14.83)

Longear sunfish

6,700 (± 20,463)

58.15 (± 158.36)

Greenside darter

60*

0.29*

Rainbow darter

1,356 (± 260)

1.04 (± 0.23)

Orangethroat darter

719*

0.40*

Total

21,309 (± 24,144)

146.24 (± 181.43)


48

Table 13. Density and biomass estimates for fishes sampled in Upper Buffalo River 20 July 1998. Values in parentheses are ± 2 SE. Species

Density (fish/ha)

Biomass (kg/ha)

Central stoneroller

19,428 (± 3,755)

65.35 (± 14.34)

Ozark minnow

8,758*

9.07*

Telescope shiner

174*

0.30*

Northern hog sucker

253 (± 121)

0.94 (± 1.52)

Slender madtom

2,079 (± 2,313)

2.09 (± 2.02)

Longear sunfish

58*

1.22*

Rainbow darter

14,285 (± 1,669)

11.89 (± 1.81)

Stippled darter

660 (± 497)

2.59 (± 2.18)

Orangethroat darter

339 (± 626)

0.17 (± 0.31)

Banded sculpin

116*

1.66*

Total

46,150 (± 4,784)

95.28 (± 14.84)


49

Table 14. Density and biomass estimates for fishes sampled in War Eagle Creek 23 July 1998. Values in parentheses are ± 2 SE. Species

Density (fish/ha)

Biomass (kg/ha)

Central stoneroller

4,951 (± 346)

4.88 (± 1.11)

Whitetail shiner

963*

2.57*

Duskystripe shiner

917 (± 49)

0.50 (± 0.06)

Creek chub

53*

0.02*

Northern hog sucker

53*

3.25*

Black redhorse

107*

15.83*

Slender madtom

1,177*

2.21


160*

0.48*

Ozark bass

1,621 (± 734)

47.79 (± 42.43)

Green sunfish

267 (± 371)

10.54 (± 17.92)

Longear sunfish

4,694 (± 892)

80.67 (± 16.03)

Smallmouth bass

1,436 (± 138

16.11 (± 3.93)

Greenside darter

2,863 (± 256)

5.68 (± 1.2)

Rainbow darter

5,591 (± 1,790)

6.18 (± 2.38)

Stippled darter

53*

0.21*

Orangethroat darter

8,506 (± 20,234)

5.24 (± 12.49)

50

Table 14. Continued. Species

Density (fish/ha)

Biomass (kg/ha)

Banded darter

160*

0.21*

Banded sculpin

313 (± 223)

0.31 (± 0.24)

Total

33,885 (± 20,355)

202.68 (± 50.58)


51

Table 15. Density and biomass estimates for fishes sampled in White River 14 July 1998. Values in parentheses are ± 2 SE. Species

Density (fish/ha)

Biomass (kg/ha)

Central stoneroller

9,215 (± 315)

41.83 (± 2.43)

Duskystripe shiner

2,585 (± 1,006)

5.24 (± 0.73)

Hornyhead chub

618 (± 817)

5.09 (± 4.25)

Ozark minnow

1,056 (± 80)

0.98 (± 0.13)

Ozark shiner

60*

0.08*

Rosyface shiner

30*

0.06*

Northern hog sucker

337 (± 924)

6.84 (± 19.16)

Slender madtom

1,580*

5.66*

Northern studfish

268*

0.62*


328*

0.75*

Ozark bass

686 (± 421)

20.28 (± 30.80)

Green sunfish

836 (± 1,000)

8.54 (± 8.67)

Longear sunfish

1,348 (± 1,311)

15.23 (± 6.79)

Smallmouth bass

545 (± 924)

25.76 (± 6.42)

Greenside darter

1,111 (± 302)

3.61 (± 1.25)

52

Table 15. Continued. Species

Density (fish/ha)

Biomass (kg/ha)

Rainbow darter

10,174 (± 2,087)

9.35 (± 17.13)

Yoke darter

60*

0.18*

Banded darter

612 (± 1,376)

0.61 (± 1.56)

Logperch

60*

0.32*

Total

31,509 (± 12,287)

151.03 (± 42.42)


53

Macroinvertebrate assemblages Aquatic macroinvertebrates were represented by 14 orders, 32 families, and 47

genera in the 10 rivers (Tables 16-26). Taxa richness ranged from 15 genera in Big

Piney Creek to 28 genera in Hurricane Creek. Density varied widely from 523

invertebrates/m2 in Middle Fork Illinois Bayou to 8,019 invertebrates/m2 in War Eagle

Creek with a mean of 2,230 invertebrates/m2. War Eagle Creek also had the highest

EPT richness at 75%. The proportion of individuals as Chironomidae was generally

low (less than 10%) with the exception of the Middle Fork Illinois Bayou and the

Upper Buffalo River. All rivers exhibited a high degree of dominance (measured as

the proportion of the three most abundant taxa). Dominance was 46% in the Kings

River, and at least 60% in all others. Psephenus was the genera that most often

contributed to dominance. Isonychia, Chimarra, and Cheumatopsyche made up 62%

of the War Eagle Creek sample, with nets of the two caddisfly species capturing a

dense mass of other species.

Physical and Chemical Attributes


All instream habitat surveys were done in the summer of 1998, a year which

was characterized by higher than average temperatures and drought. Instream and

riparian characteristics of the ten rivers are presented in Table 27. Mean depth ranged

from 11.7 cm in the Upper Buffalo River to 47.4 cm in War Eagle Creek. Dry riffles

in small Boston Mountain streams are common in the summer, but their occurrence

was more pronounced than usual in 1998. Differences in mean depth were due largely

54

Table 16. Macroinvertebrate assemblage characteristics for headwater reaches of 10 Boston Mountain rivers. River

Taxa

richness

Density

(invertebrates/m2)

EPT (%)

Chironomidae

(%)

EPT/Chironomidae

Dominance

(%) Big Piney Creek

15

748

65

7

8.3

77

Hurricane Creek

28 3,911 33 1 30.2 75

Kings River

26 919 49 0 117.0 46


18 523 58 27 2.1 60

Mulberry River

26 2,535 24 1 22.3 71


25 1,341 22 8 2.6 63

Richland Creek

22 1,066 22 9 2.4 60

Upper Buffalo River

23 1,663 35 22 1.5 74

War Eagle Creek

26 8,019 75 3 22.4 62

White River

20 1,574 40 2 16.2 72

55

Table 17. Identity and number of macroinvertebrates sampled in Big Piney Creek 16 October 1998. Order

Family

Genus

Number

Ephemeroptera

Isonychiidae

Isonychia

1

Heptageniidae Leucrocuta 1

Stenonema 12

Caenidae Caenis 101

Plecoptera Leuctridae Zealeuctra 2

Capniidae Allocapnia 7

Trichoptera Philopotamidae Chimarra 1

Odonata Gomphidae Stylogomphus 3

Coenagrionidae Argia 1

Megaloptera Corydalidae Corydalus 1

Coleoptera Psephenidae Psephenus 33

Elmidae Stenelmis 6

Diptera Chironomidae 14

Tipulidae Hexatoma 1

Decapoda Astacidae Orconectes 9

Total

193

56

Table 18. Identity and number of macroinvertebrates sampled in Hurricane Creek 1 August 1998. Order

Family

Genus

Number

Ephemeroptera Baetidae Baetis 1

Isonychiidae Isonychia 28


Stenacron 101

Stenonema 11

Caenidae Caenis 119

Leptophlebiidae Choroterpes 9


Perlidae Perlesta 5


Polycentropodidae Polycentropus 6

Hydropsychidae Cheumatopsyche 26

Helicopsychidae Helicopsyche 8

Leptoceridae 2


Aeshnidae Boyeria 1


Megaloptera Sialidae Sialus 2

57

Table 18. Continued. Order

Family

Genus

Number

Corydalidae Corydalus

1

Nigronia 10


Elmidae Stenelmis 35


Hemiptera Veliidae Microvelia 1

Rhagovelia 1

Gastropoda Pleuroceridae Elimia 16


Class Oligochaeta

8

Total

1,009

58

Table 19. Identity and number of macroinvertebrates sampled in Kings River 30 July 1998. Order

Family

Genus

Number

Ephemeroptera

Baetidae

Acentrella

1

Baetis 6


Heptageniidae Stenacron 14

Stenonema 31

Caenidae Caenis 35


Plecoptera Perlidae Acroneuria 2








Nigronia 15

Coleoptera Staphylinidae 1

Psephenidae Psephenus 42

59


Family

Genus

Number

Elmidae

Helicus

6

Stenelmis 9

Diptera Tipulidae Hexatoma 1

Hemiptera Veliidae Rhagovelia 5

Lepidoptera Pyralidae Petrophila 4

Gastropoda Pleuroceridae Elimia 2


Class Oligochaeta 12

Total

237

60

Table 20. Identity and number of macroinvertebrates sampled in Middle Fork Illinois Bayou 24 October 1998. Order

Family

Genus

Number

Ephemeroptera

Baetidae

Acentrella

9

Acerpenna 2

Baetis 1


Heptageniidae Stenonema 21

Caenidae Caenis 3

Leptophlebiidae Paraleptophlebia 1


Capniidae Allocapnia 23




Coleoptera Elmidae Stenelmis 3


Simuliidae Simulium 7



Isopoda Asellidae Lirceus 6

Total 135

61

Table 21. Identity and number of macroinvertebrates sampled in Mulberry River 12 August 1998. Order

Family

Genus

Number

Ephemeroptera

Baetidae

Acentrella

3

Baetis 2


Heptageniidae Stenacron 4

Stenonema 28

Caenidae Caenis 17

Plecoptera Leuctridae Leuctra 1




Leptoceridae Oecetis 2




Corydalidae Nigronia 7


Elmidae Helicus 1

Stenelmis 16

62


Family

Genus

Number

Diptera

Chironomidae

6

Ceratopogonidae Bezzia 1

Tipulidae Tipula 2


Rhagovelia 8


Acarina

Hydrachnoidea Hydracarina 2

Class Oligochaeta 2

Total

654

63

Table 22. Identity and number of macroinvertebrates sampled in North Fork Illinois Bayou 26 September 1998. Order

Family

Genus

Number

Ephemeroptera

Isonychiidae

Isonychia

1


Stenacron 6

Stenonema 20

Caenidae Caenis 21

Baetiscidae Baetisca 2


Plecoptera Capniidae Allocapnia 1

Perlidae Acroneuria 1





Odonata Gomphidae Stylogompus 82



Corydalidae Nigronia 3


64


Family

Genus

Number

Elmidae Dubiraphia 1

Stenelmis 37



Isopoda Asellidae Lirceus 2


Class Oligochaeta 1

Total

346

65

Table 23. Identity and number of macroinvertebrates sampled in Richland Creek 10 October 1998. Order

Family

Genus

Number

Ephemeroptera

Baetidae

Baetis

1



Stenacron 3

Stenonema 14

Caenidae Caenis 15









Elmidae Helicus 1

Stenelmis 66


Simuliidae Simulium 1

66


Family

Genus

Number

Tipulidae Hexatoma

8

Tabanidae Tabanus 4


Class Oligochaeta 8

Total

275

67

Table 24. Identity and number of macroinvertebrates sampled in Upper Buffalo River 20 July 1998. Order

Family

Genus

Number

Ephemeroptera

Baetidae

Acentrella

3

Baetis 67



Stenonema 1

Caenidae Caenis 3



Perlidae Neoperla 16



Hydropsychidae Ceratopsyche 1





Elmidae Helicus 1

Stenelmis 8

68


Family

Genus

Number

Diptera

Chironomidae

96

Ceratopogonidae Bezzia 1




Total

429

69

Table 25. Identity and number of macroinvertebrates in War Eagle Creek sampled 27 July 1998. Order

Family

Genus

Number

Ephemeroptera

Baetidae

Acentrella

5

Baetis 59



Stenonema 109

Caenidae Caenis 72


Plecoptera Perlidae Acroneuria 1








Elmidae Helicus 12

Stenelmis 10

Diptera Tipulidae Hexatoma 2

70


Family

Genus

Number

Empididae 1

Chironomidae 68

Hemiptera Veliidae Rhagovelia 1


Gastropoda Ancylidae Ferrissia 3

Pleuroceridae Elimia 21


Class Oligochaeta 29

Total

2,069

71

Table 26. Identity and number of macroinvertebrates sampled in White River 14 July 1998. Order

Family

Genus

Number

Ephemeroptera

Baetidae

Baetis

9



Stenacron 17

Stenonema 65

Tricorithidae Tricorythodes 6

Caenidae Caenis 20






Brachycentridae Micrasema 1


Leptoceridae Oecetis 3




72


Family

Genus

Number


Class Oligochaeta

5

Total

406

73

Table 27. Instream habitat and riparian vegetation characteristics for headwater reaches of 10 Boston Mountain rivers. River

Mean depth

(cm)

Mean velocity

(m/s)

Sand and silt

(%)

Pool/riffle

ratio

Eroded bank

(%)

Fish cover

(%)

Forest in 50-m

buffer (%) Big Piney Creek

24.4

0.082

1.0 0.6

13.5

38.0

97.0

Hurricane Creek

40.6 0.004 9.0 0.9 24.5 16.0 93.0

Kings River

30.4 0.005 8.1 1.5 48.8 42.7 67.6


38.6 0.071 1.6 2.5 65.5 38.5 95.0

Mulberry River

34.5 0 8.4 1.4 28.5 20.0 87.0


34.1 0.039 5.6 0.6 0 54.6 97.0

Richland Creek

25.9 0 1.1 0.3 23.5 32.7 70.1

Upper Buffalo River

11.7 0.016 1.8 0.7 32.3 38.5 100.0

War Eagle Creek

47.4 0.004 13.3 0.5 30.8 15.2 90.3

White River

14.1 0.018 7.5 0.1 15.3 30.5 57.5

74

to the presence of deep pools in some of the rivers. Mean velocity was very low in all

rivers, in some cases undetectable.

There was considerable substrate size diversity with the number of substrate

categories present ranging from 8 to 13 based on the modified Wentworth particle size

scale (Bovee and Milhous 1978). Substrates of the Kings and White rivers were

largely gravel and cobble with no bedrock, in contrast to the Mulberry River, North

Fork Illinois Bayou, Upper Buffalo River and War Eagle Creek that had from 34 to

51% bedrock. Mean bank angles were similar among rivers ranging from 140 to 155

degrees, but the percentage of eroded bank varied greatly from zero along North Fork

Illinois Bayou to 65.5% along Middle Fork Illinois Bayou. There were numerous

types of fish cover in the rivers including overhanging banks, coarse woody debris,

root wads, submerged and emerged vegetation, boulders, and rock ledges, Fish cover

ranged from 15.2% in War Eagle Creek to 54.6% in North Fork Ill Bayou.

Riparian vegetation in a 50-m buffer was most disturbed at the White River

reach with 57.5% as forest and the remainder as shrub and pasture. In contrast, the

Upper Buffalo River reach, located on the edge of the Upper Buffalo Wilderness Area,

was 100% forested. However, there was little shade over the water at the

electrofishing site on the Upper Buffalo River because there was a wide gravel and

cobble floodplain producing a mean bank full width of over 30 m. In contrast, the

Kings River site was densely vegetated along the water-land interface and was densely

shaded.

75

Water Quality

Physicochemical properties of water at summer and winter flow are presented

in Tables 28 and 29. The geology of the Boston Mountain ecoregion is primarily

sandstone and shale, resulting in streams with relatively low alkalinity (usually less

than 18 mg/L). Only Hurricane Creek, Kings River and War Eagle Creek had higher

alkalinity (38, 26, and 38 mg/L, respectively, measured at summer flow). This was

reflected in higher specific conductance for these three rivers as well. Phosphate

concentrations were low in all rivers. The Kings River, War Eagle Creek, and White

River had the highest concentrations of nitrates in both summer and winter, but

concentrations in all rivers were low, relative to the drinking water standard of 10.0

mg/L (Novotny and Olem 1994). Relative to other regions, all ion and nutrient

concentrations were low in these rivers. All water quality parameters were within

standards for the Boston Mountain ecoregion set by the Arkansas Department of

Environmental Quality (ADPCE 1998).

Watershed Attributes

All watersheds were primarily forested, ranging from 84% forest in the Kings

River watershed to 98% in the North Fork Illinois Bayou watershed (Table 30). The

remainder of land cover consisted of pasture or clear-cut forest which could not be

differentiated with the data available for this GIS analysis. Private land ownership

varied greatly from 9% in North Fork Illinois Bayou watershed to 99% in War Eagle

Creek watershed. Land not in private ownership was held by the U.S. Forest Service,

with 20% of Hurricane Creek watershed designated Wilderness. The highest

76

Table 28. Water quality characteristics for headwater reaches of 10 Boston Mountain rivers sampled in summer. River

Specific

conductance (µS/cm)

Chloride (mg/L)

Sulfate (mg/L)

Nitrogen NO3-N (mg/L)

Nitrogen NH4-N (mg/L)

Total Kjeldahl

Nitrogen (mg/L)

Phosphate

PO4-P (mg/L)

Big Piney Creek

30

1.08

2.26

0.028

0.079

0.14

0.005

Hurricane Creek

99 1.50 2.61 0.081 0.035 0.11 0.003

Kings River

85 2.62 1.75 0.735 0.035 0.17 0.007


37 2.12 2.40 0.017 0.016 0.10 0.003

Mulberry River

75 0.94 1.90 0.002 0.044 0.34 0.003


34 0.87 2.31 0.002 0.005 0.10 0.003

Richland Creek

43 1.28 1.82 0.007 0.038 0.78 0.003

Upper Buffalo River

41 0.88 2.08 0.084 0.017 0.10 0.006

War Eagle Creek

100 1.60 3.49 0.100 0.013 0.10 0.003

White River

43 1.48 2.54 0.339 0.016 0.21 0.006

77


Total

phosphorus (mg/L)

pH

Alkalinity

CaCO3 (mg/L)

Hardness CaCO3

(mg/L)

Turbidity

NTU

Total dissolved

solids (mg/L)

Big Piney Creek

0.020

6.81

12

10

4.5

29.0

Hurricane Creek

0.033 7.59 38 38 0.7 52.0

Kings River

0.011 7.12 26 30 3.3 47.0


0.009 6.81 12 14 2.0 29.5

Mulberry River

0.060 6.90 12 10 7.5 37.3


0.020 7.14 14 12 1.5 30.8

Richland Creek

0.060 7.18 18 18 2.5 34.0

Upper Buffalo River

0.011 6.51 18 14 4.7 34.3

War Eagle Creek

0.037 7.57 38 38 2.6 54.8

White River

0.011 6.81 14 13 4.8 32.0

78

Table 29. Water quality characteristics for headwater reaches of 10 Boston Mountain rivers sampled in winter. River

Specific

conductance (µS/cm)

Chloride (mg/L)

Sulfate (mg/L)

Nitrogen NO3-N (mg/L)

Nitrogen NH4-N (mg/L)

Total Kjeldahl

Nitrogen (mg/L)

Phosphate

PO4-P (mg/L)

Big Piney Creek

22

0.84

2.03

0.010

0.016

0.08

0.007

Hurricane Creek

40 1.00 2.79 0.009 0.015 0.02 0.002

Kings River

33 1.83 2.31 0.422 0.015 0.07 0.001


27 1.51 2.60 0.210 0.005 0.05 0.005

Mulberry River

19 1.24 1.95 0.131 0.005 0.02 0.007


21 0.87 2.26 0.003 0.006 0.02 0.001

Richland Creek

18 1.05 1.86 0.047 0.029 0.11 0.001

Upper Buffalo River

16 1.07 2.18 0.062 0.019 0.08 0.001

War Eagle Creek

38 1.61 2.75 0.479 0.005 0.07 0.002

White River

39 1.73 2.32 1.203 0.017 0.06 0.003

79


Total

phosphorus (mg/L)

pH

Alkalinity

CaCO3 (mg/L)

Hardness CaCO3

(mg/L)

Turbidity

NTU

Total dissolved

solids (mg/L)

Big Piney Creek

0.035

6.55

8

8

5.7

31.5

Hurricane Creek

0.025 6.85 16 18 3.2 41.5

Kings River

0.665 6.94 12 16 2.8 41.8


0.110 6.82 8 10 3.4 36.0

Mulberry River

0.030 6.40 6 6 3.8 27.5


0.030 6.65 6 8 2.6 41.0

Richland Creek

0.030 6.53 6 8 2.1 35.8

Upper Buffalo River

0.015 6.58 4 6 4.2 28.0

War Eagle Creek

0.055 7.19 16 20 5.0 52.3

White River

0.010 6.54 6 10 4.8 38.5

80

Table 30. Watershed attributes for headwater reaches of 10 Boston Mountain rivers. River

Forest (%)

Pasture *

(%)

Privately owned

(%)

U.S. Forest Service

(%)

Road density

(km/ha)

100-m buffer road

density (km/ha)

Big Piney Creek

97

3

17

83

0.0112

0.0016

Hurricane Creek

96 4 11 89** 0.0129 0.0007

Kings River

84 16 69 31 0.0130 0.0015


91 9 31 70 0.0128 0.0014

Mulberry River

96 4 35 65 0.0149 0.0015


98 2 9 91 0.0140 0.0015

Richland Creek

96 4 14 86 0.0123 0.0027

Upper Buffalo River

97 3 16 84 0.0108 0.0017

War Eagle Creek

87 13 99 1 0.0167 0.0033

White River

90 10 65 35 0.0165 0.0050

* agricultural pasture and clear-cut forest included in this category. ** includes 20% designated Wilderness.

81

watershed and riparian road densities were found in War Eagle Creek and White River

watersheds. Upper Buffalo River watershed had the lowest density of roads, but

ranked third behind War Eagle Creek and the White River in riparian road density.

River Comparisons

Biotic and Physical Variable Comparisons

The mean coefficient of variation was 63.55 for the biotic variables and 64.97

for the physical variables, indicating that the variation in data representing animal

biota was similar to that for the composite of physical, chemical, and watershed

variables. Table 31 summarizes standard statistics for 17 biotic variables describing

fish and macroinvertebrate assemblage characteristics and 17 variables describing

instream habitat, riparian vegetation, water quality, and watershed attributes.

Because there were 34 variables, a large number of significant correlations

were detected among them (Table 32). Some of these statistical relationships were

presumably important biologically or ecologically, while others were merely statistical

artifacts. Most of the relationships between biotic variables supported underlying

biological theory applied to bioassessment. Relationships among physical variables

reflected the influence of riparian vegetation and watershed attributes on water quality

and the presence of sand or silt. Few relationships between biotic and physical

variables provided insights into biological or ecological processes occurring in these

rivers.

82

Table 31. Statistical characteristics of metrics used in cluster analysis and Guttman’s scaling. Metric

Mean

Median

Standard deviation

Coefficient of

variation*

Range

Fish assemblage characteristics

Density (fish/ha)

22,328.20 18,052.50 11,489.11 51.46 8,676 - 46,150

Biomass (kg/ha)

117.87 120.76 51.66 43.83 26.82 - 202.85

Species richness

14.90 15.00 3.41 22.91 10 - 19

Darter species

4.40 4.50 1.17 26.68 3 - 6

Sunfish species

2.40 2.50 0.70 29.13 1 - 3

Sucker species

1.20 1.00 0.79 65.73 0 - 2

Intolerant species

3.90 3.50 1.91 49.02 1 - 7

% Green sunfish

2.13 1.04 2.21 103.92 0 - 5.75

% Omnivores

23.05 16.04 18.29 79.36 0.34 - 51.09

% Insectivorous cyprinids

9.35 7.39 7.40 79.16 0.38 - 26.93

% Piscivores

5.12 4.89 4.06 79.40 0 - 13.09

83


Mean

Median

Standard deviation

Coefficient of

variation

Range

Invertebrate assemblage characteristics

Taxa richness 22.90

24.00 4.15 18.12 15 - 28

Density (invertebrates/m2 ) 2,229.90

1457.50 2,264.64 101.56 523 - 8,019

% EPT 42.30

37.50 18.80 44.44 22 - 75

% Chironomidae 8.00

5.00 9.32 116.52 0 - 27

EPT/Chironomidae 22.49

12.27 34.79 154.69 1.52 - 117

Dominance 66.00 67.00 9.57 14.50 46 - 77 Water quality

Nitrate summer (mg/L) 0.14 0.05 0.23

166.46 0.002 - 0.735

Nitrate winter (mg/L) 0.26 0.10 0.37

145.17 0.003 - 1.203

Alkalinity summer (mg/L CaCO3) 20.20 16.00 10.30

51.01 12 - 38

Alkalinity winter (mg/L CaCO3) 8.80 7.00 4.34

49.33 4 - 16

Turbidity summer (NTU) 3.41 2.95 2.00

58.64 0.7 - 7.5

Turbidity winter (NTU) 3.76 3.60 1.16 30.77 2.1 - 5.7

84


Mean

Median

Standard deviation

Coefficient of

Variation*

Range

Riparian vegetation and instream habitat

% Riparian forest 85.45

91.65 14.86 17.39 57.5 - 100

Mean depth (cm) 30.16

32.24 11.36 37.65 11.7 - 47.36

Mean velocity (m/s) 0.02

0.01 0.03 126.38 0 - 0.082

% Sand and silt 5.74

6.54 4.23 73.65 0.98 - 13.3

Pool /riffle ratio 0.90

0.63 0.72 79.99 0.12 - 2.5

% Eroded bank 28.25

26.50 18.41 65.18 0 - 65.5

% Fish cover 32.67

35.35 12.59 38.53 15.2 - 54.65

Watershed attributes

% Forested 93.20

96.00 4.87 5.23 84 - 98

% Private land 36.60

24.00 30.77 84.07 9 - 99

Road density (km/ha) 0.0135

0.0129 0.0020 14.98 0.0108 - 0.0167

Buffer road density (km/ha) 0.0021 0.0016 0.0012 60.19 0.0007 - 0.0050

* Coefficient of variation calculated as standard deviation divided by the mean times 100.

85

Table 32. Pearson correlation coefficients (r) between variables with significance level (probability) of 0.05 or lower. Correlated variables

Probability

r

Correlated variables

Probability

r

Biotic vs biotic variables

Alkalinity winter vs mean depth

0.0188

0.720

Fish species richness vs sunfish species

0.0016 0.856 Alkalinity winter vs % sand and silt

0.0291 0.684

Intolerant fish species vs sucker species

0.0313 0.678 % Private land vs % sand and silt

0.0242 0.700

% Green sunfish vs darter species

0.0391 -0.657 Road density vs % sand and silt

0.0077 0.781

% Insect. cyprinids vs piscivore species

0.0173 0.726 Buffer road density vs % riparian forest

0.0379 -0.660

EPT/Chironomid vs % insect. cyprinids

0.0042 0.813 Alkalinity summer vs alkalinity winter

0.0004 0.900

EPT/Chironomid vs piscivore species

0.0155 0.735 % Forested watershed vs nitrate summer 0.0095 -0.768

Invert. dominance vs omnivore species

0.0169 0.728 % Private land vs nitrate winter

0.0215 0.710

Physical vs physical variables

Road density vs nitrate winter

0.0252 0.697

Pool/riffle vs % eroded bank

0.0030 0.829 Buffer road density vs nitrate winter

0.0018 0.852

Nitrate summer vs % riparian forest

0.0464 -0.640 % Forested watershed vs % private land

0.0008 -0.881

Alkalinity summer vs % sand and silt

0.0301 0.681 % Private land vs road density

0.0193 0.718

Nitrate winter vs % riparian forest

0.0179 -0.724 Road density vs buffer road density

0.0418 0.650

86

Table 32. Continued. Correlated variables

Probability

r

Correlated variables

Probability

r

Biotic vs physical variables

Darter species vs mean depth

0.0102 0.763 Piscivore species vs % private land

0.0086 0.774

Fish biomass vs % forested watershed

0.0165 -0.730 Invert. density vs % sand and silt

0.0069 0.787

Sunfish species vs % forested watershed

0.0211 -0.711 % Chironomids vs % sand and silt

0.0298 -0.682

Fish biomass vs % private land

0.0126 0.750 Invert. density vs fish cover

0.0182 -0.723

Sunfish species vs % private land

0.0226 0.706 EPT/Chironomid vs nitrate summer

0.0004 0.898

Invert. taxa richness vs mean velocity

0.0043 -0.812 Invert. density vs alkalinity summer

0.0087 0.773

Invert. taxa richness vs % sand and silt

0.0220 0.708 Invert. density vs alkalinity winter

0.0253 0.696

% Insect. cyprinids vs nitrate summer

0.0145 0.740 % EPT vs turbidity winter

0.0389 0.657

Piscivore species vs nitrate summer

0.0188 0.720 EPT/Chironomid vs % forested watershed

0.0241 -0.700

Invert. taxa richness vs alkalinity summer

0.0418 0.650 Invert. dominance vs % forested watershed

0.0324 0.675

Omnivore species vs turbidity summer

0.0078 0.780 Fish species richness vs road density

0.0018 0.850

Piscivore species vs % forested watershed 0.0042 -0.813

Sunfish species vs road density 0.0203 0.714

87

River Grouping based on Similarities

Kings River and War Eagle Creek, neither of which has Wild and Scenic River

designation, were distinguished from the remaining eight rivers using cluster analysis

(Figure 2). Using a significance level (alpha) of 0.05, the percentage of forested

watershed was the only variable identified that could distinguish between the groups

using stepwise discriminant analysis. However, a grouping based only on this variable

yielded one of the 10 rivers classified incorrectly. At an alpha level of 0.07, seven

variables were significant (Table 33), and using all seven yielded correct

classification. In an effort to develop a more parsimonious classification function,

percentage of forested watershed was tested with each of the other six significant

variables individually to find if using only percentage of forested watershed and one

other variable would yield a correct classification. Using the percentage of forested

watershed and the logit of the percentage of forested watershed gave an error rate of

zero and a cross validation error rate of zero. None of the other five variables when

paired with percentage of forested watershed produced zero error rates. Thus, the final

discriminant function contained the percentage of forested watershed and its logit.

The correlation between percentage of forested watershed and canonical

variable one from the discriminant function analysis was 0.70, and the correlation

between the logit of the percentage of forested watershed and canonical variable one

was 0.51. Because the number of canonical variables that can be identified is limited

to one less than the number of members of the smallest group, only the first canonical

variable could be examined. The linear discriminant functions generated were as

follows:

88

89

Table 33. Significant variables (p < 0.07) that distinguished between river groupings based on stepwise discriminant analysis. Variable

Degrees of freedom 1

Degrees of freedom 2

F statistic

Probability

% Forested watershed

1

8

18.14

0.0028

logit % Forested watershed

1 7 5.02 0.0601

% Fish cover

1 6 11.61 0.0144

Nitrate concentration summer

1 5 7.92 0.0481

Mean velocity

1 4 13.15 0.0361

% Sand and silt

1 3 23.19 0.0405

Fish species richness

1 2 22.57 0.0416

90

For membership in Group 1 (Kings River and War Eagle Creek),

G1 = 158.95 (percentage forested watershed) - 800.53 (logit percentage

forested watershed) - 6083.

For membership in Group 2 (8 remaining rivers),

G2 = 164.92 (percentage forested watershed) - 826.33 (logit percentage

forested watershed) - 6558.

The equation that generates the higher value (G1 or G2) determines group

membership.

Because a single watershed characteristic dominated the cluster analysis when

variables representing biotic, physical, chemical, and watershed attributes were used,

cluster analysis was also performed using only biotic variables (fish and

macroinvertebrate assemblage characteristics) to determine if any of these variables

could distinguish among rivers. No clear groupings were generated by this analysis.

Ranking Rivers Relative to Ideal Conditions

The ranking of the 10 rivers relative to conceptually ideal conditions is shown

in Figure 3. North Fork Illinois Bayou was closest to ideal followed by the Middle

Fork Illinois Bayou. Neither of these rivers has Wild and Scenic River designation,

although both were included in the U.S. Forest Service suitability study concluded in

1991 (U.S. Forest Service 1991). These two rivers were followed in ranking by the

five Wild and Scenic rivers and the Kings River, which as a group were similar in

their relationship to ideal. The White River and War Eagle Creek ranked farthest from

ideal and shared similar rankings for a number of variables. The variables responsible

91

92

for the contrast between North Fork Illinois Bayou (closest to ideal) and both the

White River and War Eagle Creek (distant from ideal) are shown in Table 34. Fish

density, number of intolerant fish species, and invertebrate density were important

biotic variables responsible for the rankings. Contributing physical variables included

riparian forest cover, nitrate concentration, turbidity, percentage of forested watershed,

percentage of private land ownership, and road density both in the watershed and in a

100-m buffer.

93

Table 34. Variables that characterized North Fork Illinois Bayou ranked as closest to ideal and War Eagle Creek and White River ranked as farthest from ideal. Blanks indicate minimal influence of the variable. Variable

North Fork

Illinois Bayou

War Eagle

Creek

White River

Fish density

Moderate

High

High

Fish biomass High

Number of intolerant fish species

High

Low

% Green sunfish Low

Invertebrate density Moderate

High

Invertebrate taxa richness Low

% Sand and silt

High

% Forest in 50-m riparian zone High

Low

% Fish cover High Low

Mean depth Moderate

High

Pool/riffle ratio Moderate

Nitrate concentration summer

Low High

Nitrate concentration winter Low High High

Alkalinity summer Moderate High

Turbidity summer Low

Turbidity winter Low High High

% Forest in watershed High Low Low

% Private land in watershed Low High High

Density of roads in watershed High High

Density of roads in 100-m buffer

High High

94

DISCUSSION

The two statistical approaches followed in this study, cluster with discriminant

function analyses, and multidimensional scaling, shared the common goal of identifying

rivers with the greatest degree of ecological integrity and the attributes that best

describe these rivers. The two methods, however, differed in the manner in which they

addressed the problem and ultimately generated different kinds of information. In

applying these procedures to my data set, the strengths and limitations of the two

approaches became apparent.

Cluster analysis followed by stepwise discriminant analysis grouped rivers

based on their similarities and then identified the attributes that best discriminated

between groups. This approach is based on the premise that rivers in a specific

ecoregion naturally share many common characteristics, and that certain biological,

chemical, or physical attributes change as a result of human activity. There are certain

limitations associated with this approach. A large number of variables is required to

adequately assess ecological integrity relative to the number of observations (rivers

assessed). However, the number of discriminating variables that can be identified by

stepwise discriminant analysis is limited to two less than the number of rivers. In this

case, 34 variables were used, but the number of discriminating variables that could be

identified was limited to eight. Also, because each variable selected is dependent on the

presence of all variables selected ahead of it, stepwise discriminant analysis stops when

a variable is detected that does not meet the set significance level (alpha). In this study,

raising the alpha was necessary to obtain more ecological information. Finally, only the

95

first canonical variable can be examined if there are only two members in a group, as

was the case in this study.

Regardless of these constraints, useful information was generated about the

rivers using cluster analysis. A watershed variable, percentage of forest, rather than any

biotic variable, distinguished among groups. No clear grouping resulted when biotic

variables alone were used, suggesting that changes in land cover may not have been

reflected in the biota. The rivers in this study were chosen from the Nationwide Rivers

Inventory (U.S. National Park Service 1982), a list that represents the Nation’s least-

disturbed rivers, and Boston Mountain rivers are recognized as being minimally

disturbed relative to many others (Omernik 1995). Cluster analysis supported this

perception in that it did not identify strong differences among the rivers. It may be

concluded that while there were differences in watershed land cover, the headwater

reaches of the rivers were not strongly affected by these differences. While the

circumstances surrounding analysis of this particular data set using cluster and

discriminant function analyses resulted in a limited result, the approach is sound, and

may provide useful information when applied to a data set in which at least some of the

rivers have sustained a greater degree of disturbance than that encountered in this study.

Guttman’s scaling involves an approach which is frequently used in

bioassessment protocols, employing comparison to reference conditions. Because at

least some of the rivers in this study are recognized as minimally disturbed, establishing

ideal conditions as a composite of the best conditions among the ten rivers was justified.

This approach may be useful in suitability studies because all rivers in these studies

must meet eligibility requirements, but in studies involving more disturbed rivers, this

96

approach may not be defensible. In this study, the process of establishing conceptually

ideal conditions served an important purpose beyond providing a point of reference for

ranking rivers; it determined least-disturbed reference conditions for upper headwater

reaches in the Boston Mountain ecoregion, a primary objective of this study.

Physical factors, including many watershed variables, played a greater role than

biotic factors in ranking rivers relative to ideal conditions. The percentage of forested

watershed played an important role in distinguishing among rivers, with the river

ranking first (highest integrity) having the highest percentage of forested watershed, and

the two rivers ranked farthest from ideal (lowest integrity) ranking eighth and ninth in

percentage of forested watershed. Roth et al. (1996) reported that land use was a useful

predictor of fish IBI scores in Michigan streams with 36 to 84% of their watersheds in

agricultural use, and Wang et al. (1997) concluded that fish IBI scores decreased when

at least 50% of the watershed was converted to agriculture in Wisconsin. In contrast,

the watersheds in this study were much less disturbed, with forest cover ranging from

84 to 98%. Some fish assemblage characteristics influenced river ranking, but six of the

rivers were very similar to one another, relative to ideal. The level of watershed

disturbance in some of these upper headwater reaches may not be sufficient to be

reflected in changes in biota.

There were indications, however, that changes in land use may have influenced

the biota of at least two of the rivers, the White River and War Eagle Creek. Relative to

the other rivers in this study, these two had low percentages of forested watershed, high

percentages of private land, high road densities, high levels of nitrate, and high fish

densities. In addition, War Eagle Creek had very high fish biomass and invertebrate

97

density, and the White River had the lowest number of intolerant fish speices and

lowest invertebrate taxa richness. My results confirm that Boston Mountain ecoregion

streams are nutrient poor, and total fish density, biomass , and production estimates

from these streams have been shown to be low, compared to other areas (Rambo 1998).

Steedman (1988) found fish abundance to be higher at moderate levels of degradation

(i.e. nutrient enrichment), and Yoder and Smith (1999) reported a pattern of increased

fish density and biomass with moderate species richness in disturbed streams. High fish

and invertebrate productivity, changes in fish assemblage structure, and lower

invertebrate taxa richness in the White River and War Eagle Creek may be a reflection

of the influence of nutrient enrichment associated with conversion from forest to pasture

in these watersheds.

Some metrics proved more useful than others in determining the relative ranking

of rivers in the Boston Mountain ecoregion. Fish density and biomass estimates are not

conventionally used in fish IBI assessments, but provided important information in this

study. The number of intolerant fish species, invertebrate density, and invertebrate taxa

richness were also important biological variables. The percentage of fish as omnivores

proved somewhat problematic because of a linkage with the distribution of central

stonerollers, the species that contributes the most to this category. This species is very

abundant at sites without shade cover, which was the case at the Upper Buffalo River

site. On the other hand, it was not collected at the Kings River site which was densely

shaded, but it was observed to be present during habitat assessment on a longer stream

reach. Central stonerollers have been recommended as an indicator of disturbance in

Ozark plateau streams (Petersen 1998). However, in this study, there was no evidence

98

that variation in estimates of this species was related to disturbance. Estimates of

central stonerollers influenced not only the percentage of fishes as omnivores, but also

total fish density and biomass estimates, and may be closely confounded with shading

within a particular stream reach. This situation lowered the integrity ranking for the

Upper Buffalo River and raised that of the Kings River relative to ideal conditions.

Water quality was important in determining if parameters were within standards

established by the Arkansas Department of Environmental Quality (ADPCE 1988).

Nitrate concentration was the most important water quality metric differentiating rivers

in this study. The Boston Mountain ecoregion is nutrient poor, and there is a known

source of enrichment from animal waste in some of these watersheds. The relationship

between pasture land and increased nitrate concentration was eluded to by this study.

Some of the instream habitat variables appear to have limited value in

determining ecological integrity of Boston Mountain ecoregion rivers. Mean depth and

velocity were very low at base flow, and it is difficult to tie differences among rivers to

disturbance. There is a known relationship between pool/riffle ratios and bankful width

in gravel bed streams, but gravel was not the predominant substrate in any of these

streams. Thus, those relationships are less applicable. The percentage of eroded bank

was high in many of these rivers, but there was no evidence that this was due to recent

human disturbance. It is unclear whether this is a natural phenomenon associated with

geomorphology (e.g. elevation gradient or bedrock/boulder substrate) in this region or

an artifact of intensive logging in the past. Because there was no clear indication that

eroded bank was indicative of disturbance, a median value was used in the model of an

ideal river.

99

Watershed attributes, including land use, ownership, and road density, were the

most important component in this study, playing a major role in discriminating among

rivers with ecological integrity closest to ideal conditions versus those more distant

from ideal. Under the Wild and Scenic Rivers Act, river segments are assigned wild,

scenic, or recreational status based on accessibility by road, and management plans are

based on assigned status. Hence, information on road density is useful in the ecological

and political assessment process, as well as for future management should protection be

conferred.

The North and Middle Forks of Illinois Bayou had the highest ecological

integrity relative to ideal, and therefore represent the least-disturbed reference

conditions for headwater reaches of Boston Mountain ecoregion streams. They were

not recommended for protective status in a suitability study completed by the U.S.

Forest Service (U.S. Forest Service 1991). At that time, impoundment of the North

Fork was being considered, and the Middle Fork did not rank as high as other rivers in

the U.S. Forest Service study in scenic value or recreational potential. Ecological value

of these rivers was not addressed in that study. The results of my research suggest that

consideration of some form of protection should be considered for these rivers on the

basis of their ecological value.

All of the river reaches selected for this study belong to a group representing a

small percentage of high quality, free-flowing rivers in the United States. Hence, the

rankings of the White River and War Eagle Creek should not be construed to indicate a

lack of ecological integrity, but rather a lower degree than those to which they were

compared. Changes in ecological integrity that were identified in this research are

100

minimal in comparison to the effects that disturbance has had on many of the Nation’s

rivers. Efforts to minimize and prevent further change in these rivers would ensure that

they remain worthy of continued recognition for their high quality.

Multidimensional scaling proved to be an effective analytical method for

achieving the stated goals of this research. Least-disturbed reference conditions were

established for Boston Mountain headwater streams. Rankings provided a relative

comparison of ecological integrity, and the attributes responsible for these rankings

were identified. While this study addressed ecological integrity of rivers in a specific

ecoregion, these methods, with appropriate refinement, could be applied to other areas.

This may facilitate the process of identifying rivers that have retained high ecological

integrity for protection under the Wild and Scenic Rivers Act or other mechanism, use

as least-disturbed reference streams in biomonitoring, and providing benchmarks in

restoration efforts.

101

LITERATURE CITED

ADPCE, Arkansas Department of Pollution Control and Ecology. 1998. Regulation establishing water quality standards for surface waters of the state of Arkansas. Regulation 2, as amended. Little Rock.

Allan, J.D., and A.S. Flecker. 1993. Biodiversity conservation in running waters. BioScience 43:32-43. Allan, J.D., and L.B. Johnson. 1997. Catchment-scale analysis of aquatic ecosystems. Freshwater Biology 37:107-111. Bailey, R.G. 1983. Delineation of ecosystem regions. Environmental Management 7:365-373. Barbour, M.T., and J. Gerritsen. 1996. Subsampling of benthic samples: a defense of

the fixed-count method. Journal of the North American Benthological Society 15:386-391.

Barbour, M.T., J. Gerritsen, B.D. Snyder, and J.B. Stribling. 1999. Rapid

bioassessment protocols for use in streams and wadeable rivers: periphyton, benthic macroinvertebrates, and fish. 2nd edition. EPA 841-B-99-002. U.S. Environmental Protection Agency, Office of Water, Washington, D.C.

Benke, A.C. 1990. A perspective on America’s vanishing streams. Journal of the North American Benthological Society 9:77-88.

Bohlin, T., S. Hamrin, T.G. Heggberget, G. Rasmussen, and S.J. Saltveit. 1989. Electrofishing – theory and practice with special emphasis on salmonids. Hydrobiologia 173:9-43. Bovee, K.D., and R. Milhous. 1978. Hydraulic simulation in instream flow studies:

theory and technique. U.S. Fish and Wildlife Service FWS/OBD-77/63. Fort Collins, Colorado.

Cairns, J. Jr., and J.R. Pratt. 1993. A history of biological monitoring using benthic macroinvertebrates. Pages 10-27 in D.M. Rosenberg and V.H. Resh, editors. Freshwater biomonitoring and benthic macroinvertebrates. Chapman and Hall, New York. Cassidy, T.J. Jr. 1994. 25 years of the Wild and Scenic Rivers Act. Fisheries 19(3):24-25.

102

Courtemanch, D.L. 1996. Commentary on the subsampling procedures used for rapid bioassessments. Journal of the North American Benthological Society

15:381-385. Davis, W.S. 1995. Biological assessment and criteria: building on the past. Pages

15-30 in W.S. Davis and T.P. Simon, editors. Biological assessment and criteria: tools for water resource planning and decision making. CRC Press, Boca Raton, Florida.

Davis, W.S., and T.P. Simon, editors. 1995. Biological assessment and criteria: tools for water resource planning and decision making. CRC Press, Boca Raton, Florida.

Echevarria, J.S., and J. Fosburgh. 1988. The American Rivers outstanding river list.

American Rivers, Inc., Washington, D.C. Forbes, S.A., and R.E. Richardson. 1913. Studies on the biology of the upper Illinois

River. Illinois State Laboratory of Natural History Bulletin 9:1-48. Guttman, L. 1946. An approach for quantifying paired comparisons and rank order.

Annals of Mathematical Statistics 17:144-163. Hughes, R.M. 1995. Defining acceptable biological status by comparing with

reference conditions. Pages 31-48 in W.S. Davis and T.P. Simon, editors. Biological assessment and criteria: tools for water resource planning and decision making. CRC Press, Boca Raton, Florida.

Hughes, R.M., D.P. Larsen, and J.M. Omernik. 1986. Regional reference sites: a

method for assessing stream potentials. Environmental Management 10:629-635.

Hughes, R.M., T.R. Whittier, and C.M. Rohm. 1990. A regional framework for

establishing recovery criteria. Environmental Management. 14:673-683. Johnson, L.B., and S.H. Gage. 1997. Landscape approaches to the analysis of aquatic

ecosystems. Freshwater Biology 37:113-132.

Karr J.R. 1981. Assessment of biotic integrity using fish communities. Fisheries 6(6):21-27. Karr, J.R. 1991. Biological integrity: a long neglected aspect of water resource management. Ecological Applications 1:66-84.

103

Karr, J.R. 1995. Protecting aquatic ecosystems: clean water is not enough. Pages 7-14 in W.S. Davis and T.P. Simon, editors. Biological assessment and criteria: tools for water resource planning and decision making. CRC Press, Boca Raton, Florida.

Karr, J.R. 1998. Rivers as sentinels: using the biology of rivers to guide landscape management. Pages 502-528 in R.J. Naiman and R.E. Bilby, editors. River ecology and management: lessons from the Pacific coastal ecoregion. Springer-Verlag, New York.

Karr, J.R., and D.R. Dudley. 1981. Ecological perspective on water quality goals. Environmental Management 5:55-68. Karr, J.R., K.D. Fausch, P.L. Angermeier, P.R. Yant, and I.J. Schlosser. 1986. Assessing biological integrity in running water: a method and its rationale.

Illinois Natural History Survey, Special Publications 5. Kwak, T.J. 1992. Modular microcomputer software to estimate fish population parameters, production rates and associated variance. Ecology of Freshwater

Fish 1:73-75. Lammert, M., and J.D. Allan. 1999. Assessing biotic integrity of streams: effects of

scale in measuring the influence of land use/cover and habitat structure on fish and macroinvertebrates. Environmental Management 23:257-270.

Lenat, D.R. 1993. A biotic index for the southeastern United States: derivation and

list of tolerance values, with criteria for assigning water-quality ratings. Journal of the North American Benthological Society 12:279-290.

McNab, W.H., and P.E. Avers. 1994. Ecological subregions of the U.S. section descriptions. U.S. Forest Service Publication WO-WSA-5. Washington,

D.C. Merritt, R.W., and K.W. Cummins. 1996. An introduction to the aquatic insects of

North America. 3rd edition. Kendall/Hunt Publishing, Dubuque, Iowa.

Naiman, R.J., J.J. Magnuson, D.M. Knight, J.A. Stanford, and J.R. Karr. 1995. Freshwater ecosystems and their management: a national initiative. Science 270:584-585. National Research Council. 1992. Restoration of aquatic ecosystems: science, technology, and public policy. National Academy Press, Washington, D.C.

104

Norris, R.H., and A. Georges. 1993. Analysis and interpretation of benthic macroinvertebrate surveys. Pages 234-286 in D.M. Rosenberg and V.H. Resh, editors. Freshwater biomonitoring and benthic macroinvertebrates. Chapman and Hall, New York.

Novotny, V., and H. Olem. 1994. Water quality: prevention, identification, and

management of diffuse pollution. Van Nostrand Reinhold, New York. Ohio EPA. 1987a. Biological criteria for the protection of aquatic life: Volume I.

Ohio Environmental Protection Agency, Columbus, Ohio.

Ohio EPA. 1987b. Biological criteria for the protection of aquatic life: Volume II and addendum dated September 30, 1989. User’s manual for the biological assessment of Ohio surface water. Ohio Environmental Protection Agency Columbus, Ohio.

Ohio EPA. 1989. Biological criteria for the protection of aquatic life: Volume III.

Standardized biological field sampling and laboratory methods for assessing fish and macroinvertebrate communities. Ohio Environmental Protection Agency, Columbus, Ohio.

Omernik, J.M. 1987. Ecoregions of the conterminous United States. Annals of the Association of American Geographers 77:118-125. Omernik. J.M. 1995. Ecoregions: a spatial framework for environmental

management. Pages 49-62 in W.S. Davis and T.P. Simon, editors. Biological assessment and criteria: tools for water resource planning and decision making. CRC Press, Boca Raton, Florida.

Palmer, T. 1993. The Wild and Scenic Rivers of America. Island Press, Washington, D.C. Petersen, J.C. 1998. Water-quality assessment of the Ozark plateaus study area,

Arkansas, Kansas, Missouri, and Oklahoma – Fish communities in streams in the Ozark plateaus and their relations to selected environmental factors. Water-Resources Investigations Report 98-4155. National Water-Quality Assessment Program. U.S. Department of the Interior, U.S. Geological Survey.

Plafkin, J.L., M.T. Barbour, K.D. Porter, S.K. Gross, and R.M. Hughes. 1989. Rapid bioassessment protocols for use in streams and rivers. U.S. Environmental Protection Agency, EPA/444/4-89-001, Washington, D.C.

105

Poulton, B.C., and K.W. Stewart. 1991. The stoneflies of the Ozark and Ouachita mountains (Plecoptera). Memoirs of the American Entomological Society 38:1-116.

Rambo, R.D. 1998. Ozark stream fish assemblages and black bass population dynamics associated with watersheds of varying land use. M.S. thesis. University of Arkansas, Fayetteville.

Rankin, E.T. 1989. The qualitative habitat evaluation index (QHEI). Rationale, methods, and applications. Ohio Environmental Protection Agency, Division of Water Quality Planning and Assessment, Ecological Analysis Section, Columbus, Ohio.

Rankin, E.T. 1995. Habitat indices in water resource quality assessments. Pages 181 -208 in W.S. Davis and T.P. Simon, editors. Biological assessment and criteria: tools for water resource planning and decision making. CRC Press, Boca Raton, Florida.

Resh, V.H., and J.K. Jackson. 1993. Rapid assessment approaches to biomonitoring

using benthic macroinvertebrates. Pages 195-233 in D.M. Rosenberg and V.H. Resh, editors. Freshwater biomonitoring and benthic macroinvertebrates. Chapman and Hall, New York.

Resh, V.H., and E.P. McElravy. 1993. Contemporary quantitative approaches to biomonitoring using benthic macroinvertebrates. Pages 159-194 in D.M. Rosenberg and V.H. Resh, editors. Freshwater biomonitoring and benthic macroinvertebrates. Chapman and Hall, New York. Robison, H.W., and T.M. Buchanan. 1984. Fishes of Arkansas. University of

Arkansas Press, Fayetteville. Rohm, C., J.W. Giese, and C.C. Bennett. 1987. Evaluation of an aquatic ecoregion classification of streams in Arkansas. Journal of Freshwater Ecology 4:127-

140. Rosenberg, D.M., and V.H. Resh, editors. 1993. Freshwater biomonitoring and

benthic macroinvertebrates. Chapman and Hall, New York. Roth, N.E., J.D. Allan, and D.L. Erickson. 1996. Landscape influences on stream

biotic integrity assessed at multiple spacial scales. Landscape Ecology 11(3):141-156.

SAS Institute Inc. 1990. SAS/STAT User’s Guide, version 6, 4th edition, Volume 1

and 2, Cary, North Carolina.

106

Schlosser, I.J. 1991. Stream fish ecology: a landscape perspective. BioScience 41:704-712.

Seber, G.A.F. 1982. The estimation of animal abundance and related parameters, 2nd edition. Charles Griffin, London.

Shackleford, B. 1988. Rapid bioassessments of lotic macroinvertebrate communities:

biocriteria development. Arkansas Department of Pollution Control and Ecology, Little Rock.

Stanford, J.A., and J.V. Ward. 1979. Stream regulation in North America. Pages 215-236 in J.V. Ward and J.A. Stanford, editors. The ecology of regulated streams. Plenum, New York.

Steedman, R.J. 1988. Modification and assessment of an index of biotic integrity to

quantify stream quality in southern Ontario. Canadian Journal of Fisheries and Aquatic Sciences 45:492-501.

Suter, G.W. II. 1993. A critique of ecosystem health concepts and indices. Environmental Toxicology and Chemistry 12:1533-1539. USACERL, U.S. Army Corp of Engineers Construction Engineering Research

Laboratories. 1993. GRASS (Geographical Resources Analysis Support System) 4.1 User’s Reference Manual. Champaign, Illinois.

U.S. Forest Service. undated. Ozark-St.Francis National Forests final environmental impact statement for the land and resources management plan. U.S. Forest Service Southern Region, Arkansas.

U.S. Forest Service. 1987. Land and resources management handbook, Chapter 8.

Wild and Scenic River evaluation. Forest Service Handbook 1909.12.

U.S. Forest Service. 1988. Wild and Scenic River study report and draft environmental impact statement on thirteen rivers in the Ozark National Forest. U.S. Forest Service Southern Region, Ozark National Forest. Management Bulletin R8-MB26.

U.S. Forest Service. 1991. Wild and Scenic River study report and final environmental impact statement on thirteen rivers in the Ozark National Forest. U.S. Forest Service Southern Region, Ozark National Forest. Management Bulletin R8-MB55.

U.S. National Park Service. 1982. Nationwide Rivers Inventory. U.S. Government

Printing Office, Washington, D.C.

107

Vitousek, P.M. 1994. Beyond global warming: ecology and global change. Ecology 75:1861-1876.

Wang, L., J. Lyons, P. Kanehl, and R. Gatti. 1997. Influences of watershed land use

on habitat quality and biotic integrity of Wisconsin streams. Fisheries 22(6):6-12.

Yoder, C.O., and M.A. Smith. 1999. Using fish assemblages in a state biological assessment and criteria program: essential concepts and considerations. Pages 17-56 in T.P. Simon, editor. Assessing the sustainability and biological integrity of water resources using fish communities. CRC Press, Boca Raton, Florida.

Documents

ECOLOGICAL INTEGRITY ASSESSMENT OF OZARK ...biscweb.uark.edu/Coop/Publications/andreasthesis.pdfreaches of 10 Ozark rivers, five with Wild and Scenic status. Thirty -four variables