Regional Scale Ecological Risk Assessment - Chapter 5 doc

115 This assessment examined risk to spring chinook salmon, rainbow trout, cutthroat trout, and summer steelhead in the mainstem upper Willamette and lower McKenzie Rivers in Oregon, an

Trang 1

Application of the Relative Risk Model

to the Upper Willamette River and Lower McKenzie River, Oregon Matthew Luxon and Wayne G Landis

CONTENTS

Introduction 92

Problem Formulation 93

Description of the Willamette–McKenzie Study Areas 93

Assessment Endpoints 94

Receptors of Concern 95

Sources of Stressors 95

Risk Regions 97

Risk Characterization 98

Habitat 101

Sources of Stressors 101

Landuse 102

Cumulative Landuse 102

Riparian Landuse 102

Water Rights 103

Industrial Effluent 104

Channel Modification 105

Hatchery Releases 105

Summary of Sources and Habitats of Stressor Ranks 106

Calculation of Relative Risks 106

RRM Results 107

Confirmation of Risk Ranks 110

Sensitivity Analysis 111

Discussion 113

L1655_book.fm Page 91 Wednesday, September 22, 2004 10:18 AM

Trang 2

92 REGIONAL SCALE ECOLOGICAL RISK ASSESSMENT

The Relative Risk Approach 113

Risk Confirmation 114

Uncertainty 114

Utility 115

References 115 This assessment examined risk to spring chinook salmon, rainbow trout, cutthroat trout, and summer steelhead in the mainstem upper Willamette and lower McKenzie Rivers in Oregon, an area of approximately 3,500 km2 Regional aspects of the project were addressed using the relative risk model (RRM) The study area was divided into nine risk regions with unique ecological and anthropogenic character-istics Stressor sources in each region were analyzed and compared to provide a regional perspective of risk

The RRM ranked risk in each risk region for chemical and physical stressors from multiple sources The rankings are testable hypotheses regarding the nature and location of risk Ongoing field studies of periphyton, macroinvertebrate, and fish communities will test these findings

The results of this assessment show the RRM to be robust for large-scale screen-ing-level ecological risk assessments (EcoRA) Uncertainty in risk predictions was high due to course scale analyses Site-specific data at the appropriate scale and spatially explicit process models could reduce uncertainty and make the model applicable to higher-tiered risk assessments

INTRODUCTION

This study is an EcoRA of the upper Willamette River and lower McKenzie River in Oregon It is designed to determine the relative contribution of natural and anthropogenic stressors to the risk of degradation of the aquatic community and specifically to fish of the family Salmonidae as represented by spring chinook salmon, rainbow trout, cutthroat trout, and summer steelhead The risk predictions from this assessment are testable hypotheses regarding the area and type of impacts likely to be affecting the ecological structure The risk analysis was conducted using the RRM developed for an EcoRA of Port Valdez, Alaska (Wiegers et al 1998) This project is a component of the National Council of the Pulp and Paper Industry for Air and Stream Improvement (NCASI) Long Term Receiving Water Studies (LTRWS) for pulp and paper mill effluents, an integrated effort to determine the effects of pulp mill effluents on receiving water ecological condition

The RRM was developed for an EcoRA of Port Valdez, AK (Landis and Wiegers 1997; Wiegers et al 1998) The RRM follows the risk assessment three-phase approach: problem formulation, analysis, and risk characterization Each phase is spatially explicit During the planning phase the location of the project area, habitats supporting potential endpoints, and potential sources of stressors are mapped In problem formulation the map facilitates discussion between stakeholders, risk man-agers, and risk assessors to determine assessment endpoints and to break the project area into smaller risk regions with unique combinations of habitats and sources of L1655_book.fm Page 92 Wednesday, September 22, 2004 10:18 AM

Trang 3

APPLICATION OF THE RRM TO THE WILLAMETTE AND MCKENZIE RIVERS 93

stress In the analysis phase a spatially explicit approach ensures that the pathwaysfrom release of a stressor to exposure are geographically feasible and accurate Inthe risk characterization phase mapping the risk regions provides a means of com-municating the nature, location, and extent of the risks of ecological impacts.Ranks are tied to specific locations within a landscape The landscape is broken

up into risk regions representing unique combinations of sources and habitats Ranksare assigned based on the distribution of a habitat or source of stressors in a riskregion relative to its magnitude in all other risk regions The numerical ranks forsources and habitats are combined into matrices The risk scores in the matricesshow the degree of overlap between sources of stressors and habitat within each riskregion The matrices thus provide an accounting of risks within a risk region and acomparison of risks among risk regions

One of the advantages of the relative risk procedure is that it produces testablehypotheses concerning which risk regions are relatively likely to be impacted andthe sources contributing to potential impacts The conceptual model provides mech-anistic connections between the measurement endpoints and instream measures ofeffect Thus risk predictions should correspond to instream measures of exposureand response relevant to linkages in the conceptual model Data from the study areathat was not used in risk modeling provide tests for these hypotheses

PROBLEM FORMULATION

Description of the Willamette–McKenzie Study Areas

This risk assessment covers the section of the Willamette River from its inceptionabove Eugene at the confluence with the Coast Fork Willamette River and MiddleFork Willamette River, 56 river miles (RM) down to Corvallis and the McKenzie

they are conducting monitoring of the aquatic ecological structure

The Willamette River is a ninth-order river, which is unconstrained by phy and flows through the middle of the Willamette Valley It is the 13th largestriver in the United States in terms of discharge (Kammerer, 1990) It flows northfor 187 RM from Eugene to Portland, draining the Cascade Mountains to the East,the Calapooya Mountains to the south, and the Coast Fork Mountains to the west.The Willamette Basin is commonly divided into three sections: the upper, middle,and lower basins (Willamette Basin Task Force 1969) This study focuses on theupper basin, which includes the river from Eugene to Corvallis, RM 187 to RM 128

topogra-At Corvallis, the Willamette River drains approximately 4000 square miles includingthe watersheds of the Coast Fork Willamette River, the Middle Fork WillametteRiver, the Long Tom River, the McKenzie River, and the Mary’s River At RM 161the average discharge for water years 1969 to 1998 is 11,490 cfs (USGS 1999) The McKenzie River is a seventh-order river It flows mainly west 93 miles throughthe Cascade Mountains Major tributaries include the Blue River, the South ForkMcKenzie River, and the Mohawk River The McKenzie Basin drains approximately

River from RM 34 to its confluence with the Willamette River near Eugene (Figure5.1) These boundaries represent the area to be considered in NCASI-LTRWS where

Trang 4

1300 square miles where it joins the Willamette River at RM 174, approximatelydoubling the flow of the Willamette River Average discharge at RM 6.5 is 5933 cfs.This study focuses on the lower 34 miles of the McKenzie River from DearhornPark, where the valley widens into a broad alluvial floodplain and the river becomesunconstrained, to the confluence with the Willamette River

Assessment Endpoints

To best represent the desired state of the ecological structure the assessmentendpoints must have social and biological relevance, be accessible to prediction andmeasurement, and be susceptible to the hazard being assessed (Suter 1990) For thisEcoRA, social relevance is assured by deriving endpoints from stakeholder concerns.Supporting the stated values with numerical criteria from state regulations ensuresthat endpoints for this EcoRA are accessible to prediction and measurement Criteria

to establish assessment endpoints are derived from values expressed through theWillamette Valley Livability Forum (WVLF) (WVLF 1999) and the Willamette

Figure 5.1 Risk regions of the project area and its location within the Willamette Valley.

Trang 5

Basin Reservoir Study (WBRS) (OWRD 1999a) State regulations provide Although there are other values expressed through the WVLF and WRBS studies,this assessment focuses on the aquatic environment and the directly related values.Maintenance of self-sustaining populations of native salmonids and providing arecreational and commercial fishery for salmonids are the primary values evaluated

quanti-in this risk assessment Water quality criteria for the support of aquatic life are usedbecause these criteria reflect physical conditions and concentrations of toxicantsthat, through testing and extrapolation, are believed to depress populations of valuedspecies

Stakeholder values for the Willamette Basin extend beyond the native salmonids

to include such uses as swimming; protection from flooding; providing sources ofwater for drinking, irrigation, and industry; maintenance of reservoirs for windsurf-ing and warm water fisheries; and providing a fishery for introduced salmonids such

as coho and summer steelhead These values may directly conflict with the statedvalue of maintaining a self-sustaining native fish population Inclusion of thesevalues in the relative risk assessment enables identification of areas where compro-mise in maintaining the resources may be necessary

Receptors of Concern

There are 28 species of fish reported in the project area (Altman et al. 1997), ofwhich eight are salmonids Five species of fish that use the project area at leastoccasionally, including river lamprey, pacific lamprey, spring chinook salmon, Ore-gon chub, and bull trout, are listed under the Endangered Species Act The salmonidsspring chinook salmon, cutthroat trout, rainbow trout, and summer steelhead areselected as receptors of concern because they spawn within the project area and arehighly valued as sport and food fish Habitat requirements for the salmonids selected

as receptors of concern (ROCs) overlap to a large degree with those of the otherlisted fish, thus assessing risk for these salmonids to some degree addresses risk tothe other listed fish

Sources of Stressors

The United States Geological Survey (USGS) recently included the WillametteBasin in their National Water Quality Assessment showing that it is moderatelydegraded compared with other rivers in the United States of similar size Theyshowed that fish community structure is degraded except in higher elevation streams.More pollution-tolerant species and more external anomalies were found in streamswith few riffles, poor riparian habitat, and elevated temperatures Water chemistrywas not strongly associated with fish community structure except that externalanomalies and pollution-tolerant fish were associated with agricultural and urbanstreams with the highest nutrient and pesticide concentrations Additionally, theyshow that nutrients in streams and groundwater are degrading water quality (Wentz

et al 1998)

tative criteria related to the expressed values (Table 5.1)

Trang 6

Table 5.1 Stakeholder Values and Assessment Endpoints

Stakeholder Values Assessment Endpoint Citation

Water Quality

River water is usable as a

source of drinking water

River water meets or exceeds Oregon drinking water quality criteria

1 River is swimmable River water meets or exceeds Oregon water

quality criteria for primary contact recreation

1 Avoiding or minimizing point

and nonpoint sources of

pollution (chemical input into

the river does not

compromise water quality)

Water meets or exceeds Oregon water quality criteria for toxics and nutrients

1

Conservation is the primary

means of ensuring an

adequate supply of water

Management by wise utilization of the water supply innate to the watershed

1

Fish caught from the river are

palatable and safe to eat

River meets or exceeds the Oregon water quality criteria for aquatic life

1

Fisheries

There are sufficient numbers

of desirable fish to support an

active recreational and

commercial fishery

No loss of fish production

No reduction of allowable catch of sport fish

Summer steelhead Population meets ODFW a basin fisheries plan:

maintain a potential sport catch of 250 in the mainstem above Willamette Falls

6

Maintain an annual catch of 1200 on the McKenzie River

3 Maintain a return of 2400 to the McKenzie sub-

basin

3

Native Fish Populations

River sustains thriving

populations of native fish

Populations of spring chinook salmon, rainbow trout, cutthroat trout, and winter steelhead meet ODFW basin fishery plans

2–5

Spring chinook salmon Increase production to 100,000 fish entering the

Columbia River

2 Increase the number of wild spring chinook salmon

to the McKenzie River to 10,000

2 Rainbow trout No hatchery rainbow trout found below Hayden

Bridge in the McKenzie River

3

No detectable loss of current production 3,5 Cutthroat trout No detectable loss of current production 3,5 Winter steelhead Maintain current annual sport catch in the upper

Willamette River of 190 fish

Floodplain protection and

enhancement for natural

functions and values

No net loss of riparian or floodplain vegetation

Trang 7

The Willamette River from its mouth to RM 190 was listed under Section 303d

of the Clean Water Act as water quality impaired for toxics due to elevated mercurylevels in fish (ODEQ 1998) In 1997 the Oregon Health division issued a mercuryadvisory for consumption of smallmouth bass, largemouth bass, and northern squaw-fish from the entire mainstem Willamette River Both the Willamette River and theMcKenzie River are listed under Section 303d as water quality impaired for tem-perature due to elevated summer temperatures The indigenous populations of springchinook salmon, Oregon chub, and bull trout have declined to the point that theyare listed as federally threatened or endangered under the Endangered Species Act.The sources of stressors being considered in this assessment are forestry, urban-ization, agriculture, water withdrawals, industrial effluent, introduced hatchery fish,

Risk Regions

based on the distribution of salmonid spawning, rearing, and migration habitat and

on the initial assessment of the distribution of sources of stressors The adjacentsubwatersheds draining to the specified sections of river delineate the boundaries of therisk regions Relative risk analyses compare the relative magnitude of sources of stressoccurring in and affecting habitats contained within these smaller watersheds

Table 5.1 Stakeholder Values and Assessment Endpoints (continued)

Habitat

Floodplain protection and enhancement for

natural functions and values

No net loss of riparian or floodplain vegetation

Potentially Conflicting Values

Floodplain management for human health

and windsurfing

No loss of recreation including

Boating

Values expressed through: Willamette Basin Reservoir Study (OWRD 1999a)

a Oregon Department of Fish and Wildlife.

and habitat alteration (Table 5.2)

The project area was divided into nine risk regions (Table 5.3 and Figure 5.1)

1) Oregon Administrative rules, chapter 340, Division 41, 1994

Trang 8

Risk Characterization

Several analyses were conducted to determine the extent of the above stressorsand habitats in each risk region The analyses are not exhaustive, but provide ameasure of the magnitude of a given source of stressors within each risk regionrelative to other risk regions within the project area Analyses measure the magnitude

of a given stressor within each risk region or within the cumulative watershedcontributing to the indicated risk region A cumulative watershed is the risk regionplus all land upstream of the risk region An example of a cumulative watershed isthat of Willamette (WB) The cumulative watershed contributing to WB includesthe entire McKenzie Basin, the upstream risk regions McKenzie A (MA), and theMiddle Fork and Coast Fork Willamette watersheds above MA Ranks show themagnitude of the source or habitat in a given risk region relative to the other riskregions in the project area

Each source and habitat was ranked for each risk region to indicate high, erate, low, or no magnitude Ranks are assigned using criteria specific to the projectarea Criteria are based on the size and frequency of the source and the amount andquality of available habitat Ranks are assigned to each source and habitat type on

mod-a two-point scmod-ale from 0 to 6 where 0 indicmod-ates lowest mmod-agnitude mod-and 6 the highest

Table 5.2 Sources of Stressors and Examples of Stressors Released

Source of

Stressors

Occurrence in and Upstream of Project Area Associated Stressors

Forestry About 90% of the lands draining to the

project area are forested and the majority of these lands are managed for timber harvest

Increased sediments Catastrophic debris flows Increased temperatures Urbanization Ten population centers including

Eugene–Springfield (combined population 190,000) and Corvallis (population 52,000)

Increased sediments and nutrients Metals and organic pollutants including pesticides, industrial and automotive wastes

Increased peak flows and lower low flows

Agriculture 41% of project area in agriculture

(primarily grass seed, mint, and filberts)

Pesticide runoff Increased sediments and nutrients Industrial

effluent

317 permitted waste dischargers in the project area and 22 upstream of the project area

Metals, organics, BOD, a TSS, b and nutrient input

Increased temperature Water

withdrawal

Two run-of-the-river dams, multiple industrial and agricultural water rights, numerous unpermitted withdrawals

Flow-related reduction in spawning and rearing habitat

Flow-related increases in temperatures Hatchery fish Four hatcheries that release rainbow

trout, summer steelhead salmon, and spring chinook salmon into or above the project area

Hybridization and competition with wild fish

Habitat

alteration

75% of historic shoreline lost Changes in current patterns

Loss of refugia Loss of instream cover Loss of bank cover

a BOD, biological oxygen demand.

b TSS, total suspended solids.

Trang 9

APPLICATION OF THE RRM TO THE

Table 5.3 Features of the Designated Risk Regions

Risk Region River Miles Physical Features Spawning Habitat Rearing Habitat Migration Habitat Stressors

McKenzie A (MA) 34.5 to 21.7 Floodplain:

moderately constrained Sinuousity: 1.16 Slope: 0.4%

Bed: cobble/gravel

Spring chinook Rainbow trout

Spring chinook Rainbow trout Cutthroat trout Summer steelhead

Walterville diversion canal

Forest landuse Upstream hatchery

McKenzie B (MB) 20.7 to 16 Floodplain: wide

unconstrained Sinuousity: 1.12 Slope: 0.4%

Bed: cobble/gravel

Spring chinook Rainbow trout

Urban, agricultural, and forest landuse Weyerhaeuser pulp and paper mill (NPDES major industrial discharger) McKenzie C (MC) 16 to 13 Floodplain:

Bed: cobble/gravel

Spring chinook Spring chinook

Rainbow trout Cutthroat trout Summer steelhead

Outskirts of Springfield Mohawk River (forest and agricultural landuse) Weyerhaeuser pulp and paper mill

18 permitted dischargers McKenzie D (MD) 13 to 0 Floodplain:

Bed: cobble/gravel

Rainbow trout Spring chinook

Rainbow trout Cutthroat trout Summer steelhead

Outskirts of Eugene–Springfield

on the left bank and agricultural and urban land on the right bank

13 NPDES dischargers

Trang 10

Table 5.3 Features of the Designated Risk Regions (continued)

Risk Region

River

Miles

Physical Features

Spawning Habitat

Rearing Habitat

Bed: cobble/gravel Depth: shallow

Rainbow trout

Eugene–Springfield Many small permitted dischargers Two sewage treatment plants

Willamette B

(WB)

175 to 160 Floodplain: unconstrained

Sinuousity: 1.08 Slope: 0.19%

Rainbow trout Cutthroat trout

Agricultural landuse City of Harrisburg runoff and STP

Willamette C

(WC)

Bed: cobble/gravel/sand Depth: deep

Rainbow trout Cutthroat trout

Agricultural landuse Small sewage treatment plant on tributary

Willamette D

(WD)

Bed: cobble/gravel/sand Depth: shallow Braided channel

Urban and agricultural landuse Industrial waste discharge (Amazon Creek) Long Tom River

Two pulp and paper mills

Willamette E

(WE)

Corvallis Mary’s River Muddy River Agriculture, forestry, and urban landuse

Note: Sinuosity: actual channel distance divided by straight-line distance

NPDES: National pollution discharge elimination system

Trang 11

Habitat

The quality of the instream habitat of the mainstem Willamette River and stem McKenzie River for the ROCs was determined for each of the risk regionssubjectively through discussion with an ODFW fish biologist (Wade 1999a) Ranksindicate the capacity of the mainstem river segment of each risk region to supportspawning, rearing, and migration lifestages for the various fish species used asassessment endpoints It is assumed that habitats supporting salmonids also supportother valued ecological attributes such as water quality Overall high ranks for a riskregion thus represent the ability of that risk region to support water quality valuesfor aquatic life beyond the salmonids indicated Table 5.4 shows the results of thehabitat ranking The general trend is increasing habitat quality and more life stagessupported proceeding from the downstream risk regions to upstream risk regions(i.e., from Willamette E [WE] to MA and WB)

main-Sources of Stressors

Ranks were assigned to all sources of stressors except hatchery fish based onthe statistical distribution of sources of stressors among the subareas This eliminatesbias associated with subjective assignation of ranks Jenk’s optimization statisticalanalysis (Groop 980) using ArcView™ 3.1 software was used to assign ranks Jenk’soptimization determines natural breakpoints in data by ordering the data from low

to high and interactively assigning groups by minimizing the variance within groupsand maximizing the variance between groups until an optimal grouping is determinedfor the number of groups the user specifies In general, data for each source wereassigned to four natural groups Ranks of 0, 2, 4, and 6 were assigned to these groupsfrom lowest to highest, respectively, to indicate no, low, medium, and high magnitude

of a given source in the indicated risk region

Table 5.4 Habitat Quality Ranks for Freshwater Lifestages of ROCs in the Upper

Rearing Migration Sum

McKenzie A McKenzie B McKenzie C McKenzie D

Trang 12

Landuse

Cumulative Landuse

Landuse provides an estimate of the types of activities occurring in the watershedthat may adversely affect river ecology The proportion of antropogenic landuses(urban, agriculture, or forestry) for the cumulative watershed was determined Agricultural landuse was calculated as a straightforward proportion of overalllanduse, whereas the magnitude of forestry and urban landuse in each risk regionwas determined slightly differently For these sources of stressors, the proportion ofeach watershed occupied by forest roads or urban roads* was calculated Thisanalysis assumes that watersheds with more roads in forest landuse, for example,have more logging activity and thus contribute more stress associated with loggingand the infrastructure supporting it Urban roads provide a surrogate for impermeablesurfaces, which through the resultant change in the hydrologic regime have beenshown to be a leading factor in degrading physical habitat in urban watersheds.Results reflect the relatively high proportion of forestry in the upper Willamette andMcKenzie watersheds, increasing agriculture in the lower portions of the watershed,and the influence of urbanization in the Eugene–Springfield area (Table 5.5)

Riparian Landuse

Anthropogenic landuse (agriculture and urban) in the riparian zone was ated Landuse in the riparian zone directly affects habitat quality in the adjacentriver by determining refugia during flood events, large woody debris recruitmentpotential, and shade The riparian zone is defined in this case as the land floodedduring the February 1996 flood, which was of a 30-year recurrence interval Analyses were carried out by overlaying geographic information system (GIS)coverage of the 1996 flood (ACOE 1998) with the aforementioned landuse GIS

evalu-Table 5.5 Percent Area in Agricultural Landuse, Urban Roads, and Forest Roads

for Risk Region Cumulative Watersheds Risk

Region

% Area in Agriculture Rank

% Area in Urban Roads Rank

% Area in Forest Roads Rank

* Urban landuse was calculated as a proportion of the risk region, not the cumulative watershed

Trang 13

coverages to determine percent of anthropogenic landuse within the floodplain foreach risk region

Results presented in Table 5.6 show that risk regions WB and WE had relativelylarge amounts of agriculture occurring close to the river The analysis also showsthat agriculture is largely confined to the riparian zone along the McKenzie River.Urbanization in the riparian zone is largely confined to risk region WA, but someurban riparian influence occurs in most risk regions

Water Rights

Impacts due to water withdrawal were calculated by determining the cumulativewater rights for agricultural and industrial water withdrawal Water rights representthe maximum amount of water that a permit holder may use This analysis assumesthat the proportion of a given water right used to that allotted is similar among riskregions The sum of water rights for the cumulative risk region was calculated based

on the Oregon Water Resources Department’s points-of-diversion GIS and waterrights information system database (OWRD 1999b) Total water rights for a givenrisk region (including upstream areas) were normalized to the modeled July-to-September low flow rate at the bottom of each risk region (Laenen and Risley 1997).This indicates that a large withdrawal is more important in an area with low flowthan an area with high flow

Water rights are awarded for storage of water as a volume (acre-foot [ac-ft]),while surface water and groundwater rights are awarded as a rate (cubic feet persecond [cfs]) Ranks for the relative magnitude water withdrawals among risk regionswere determined by assigning ranks to both stored and surface water withdrawalsseparately Stored and surface water ranks were then averaged and rounded up tothe nearest even number to determine the final rank

Results show that the permitted withdrawal rate from MA is much larger thandiversion canal Water is withdrawn for the length of MA to power turbines forelectricity production The water is returned to the river at the inception of riskregion MB, thus it is not reflected in rank scores for downstream risk regions Largeagricultural reservoirs in WD and WE represent the largest stored water rights forall risk regions

Table 5.6 Percent of 30-Year Floodplain in Agricultural or Urban Landuse

Risk Region % Agriculture % Urban Total % Anthropogenic Rank

that for other risk regions (Table 5.7) This withdrawal represents the Walterville

Định dạng
Số trang	27
Dung lượng	917,54 KB