Harrington_JMESthesis2014.pdf
Media
Part of Quantifying Land Use Disturbance Intensity (LDI) in the Skokomish River Watershed
- extracted text (extracttext:extracted_text)
-
QUANTIFYING LAND USE DISTURBANCE INTENSITY (LDI)
IN THE SKOKOMISH RIVER WATERSHED:
SALMONID HABITAT IMPLICATIONS
by
James W. Harrington
A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
September 2014
©2014 by James W. Harrington.
All rights reserved.
This Thesis for the Master of Environmental Studies Degree
by
James W. Harrington
has been approved for
The Evergreen State College
By
________________________
Carri LeRoy, Ph. D.
Member of the Faculty
________________________
Date
ABSTRACT
Quantifying Land Use Disturbance Intensity (LDI) in the
Skokomish River Watershed: Salmonid Habitat Implications
James W. Harrington
The degradation of in-stream habitat suitable for the spawning and rearing
activities of Pacific salmonids, resulting from anthropogenic land-use, has been
identified as a potential limiting factor for certain salmonid populations.
Quantifying the degree to which individual land-use practices disturb habitats is a
challenge to researchers. The research presented here was an attempt to quantify
land-use disturbances within a 100-meter riparian buffer on the Skokomish River
watershed in Washington State. A Land-Use Disturbance Intensity (LDI) index
was employed to classify disturbance by practice and by area of influence. A
watershed-scale LDI value was calculated to provide a snapshot of cumulative
land-use effects within the entire watershed. The Skokomish River is
characterized by a predominance of agricultural practices, forestry and timber
harvest-focused practices, low-impact residential areas, and recreational land-use
in its lower reaches. A substantial portion of the headwaters contributing to the
basin originate in protected areas or areas considered minimally impacted. It was
anticipated that a watershed-scale LDI value would be fairly low based on the
minimal impacts in the upper reaches of the watershed; however, the watershed
showed an LDI-value representing Medium Impact when examined in its entirety.
Resulting LDI values based on land-use were also applied to stream reaches
exhibiting coho activity in the lower segments of the Skokomish watershed.
Table of Contents
Chapter One: Introduction .................................................................................. 1
Human Land-use Patterns and Interactions with Anadromous Salmonids............. 4
Indigenous Northwest Populations and Salmonids......................................... 4
European/Euro-American Settlement and Timber Extraction ........................ 7
Historical Timber Harvest Around Hood Canal ............................................. 9
Coho Salmon (Onchorynchus kisutch) Characteristics: Understanding Freshwater
Spawning and Rearing Habitat Behavior.............................................................. 10
Spawning Habitat Preferences ...................................................................... 12
Juvenile Rearing............................................................................................ 12
Unique Characteristics of the Current Study Site: The Hood Canal Region ........ 13
Skokomish River Salmonid Habitat...................................................................... 15
Watershed Characteristics and Limiting Factors for Skokomish Coho ................ 16
Chapter Two: Methodological Literature Review........................................... 17
Land-Use Development Intensity Index ............................................................... 17
Proximity of Land-Use Effects on Stream Fish and Habitats In Great Lakes
Tributaries ............................................................................................................. 20
Land-use and Coho in Snohomish River (Pess et al., 2002). ................................ 22
Allan’s (2004) Four Challenges to Understanding Land-Use Change Effects on
Watersheds ............................................................................................................ 23
Covariation.................................................................................................... 24
Spatial Scales ................................................................................................ 24
Nonlinearities ................................................................................................ 25
Legacy Effects............................................................................................... 25
iv
LDI Applied Within Washington State Watersheds ............................................. 26
Chapter Three: Methods .................................................................................... 28
LDI Calculation Overview.................................................................................... 28
Primary Study Site: Skokomish Watershed-Land-use Classifications and LDI
Coefficients ........................................................................................................... 29
Calculating Catchment-Scale LDI Values for the Skokomish Watershed ........... 40
Special Considerations for LDI Value Assignments ............................................ 40
Performing The Area-Weighted LDI Calculation for the Skokomish
Watershed...................................................................................................... 41
Integrating Existing Salmonid-Related GIS Data ................................................. 43
Chapter Four: Analysis and Results ................................................................. 46
Skokomish River: Watershed-scale LDI .............................................................. 46
Individual Parcel LDI: Framing Disturbance Impact ........................................... 47
LDI-Values at Coho Distribution Sites on the Lower Skokomish River.............. 51
LDI-Value By Coho Activity Type-Lower Skokomish River.............................. 53
LDI at Coho Spawning and Rearing Reaches....................................................... 54
Spawning Reaches ........................................................................................ 54
Rearing Reaches............................................................................................ 56
Summary of Analysis and Results ........................................................................ 57
Chapter Five: Discussion.................................................................................... 58
Limitations and Challenges Within the Current Study ......................................... 58
Determining Scope and Framework ............................................................. 58
Focusing on the Skokomish River ................................................................ 59
v
Hood Canal Salmonid Recovery and Skokomish River Coho ..................... 60
Data Availability, Accuracy/Precision, Integration, Processing and Interpretation
............................................................................................................................... 62
Available Land-Use Datasets........................................................................ 62
Methodological Concerns Regarding “Area of Influence” ................................... 64
Utilizing Existing Datasets ........................................................................... 65
Potential for Comparative Studies ................................................................ 66
Implications for Informing Policy Decisions................................................ 67
Chapter Six: Interdisciplinary Aspects and Conclusions................................ 68
Interdisciplinary Nature of Salmonid Studies ....................................................... 68
The Intertwined Complexity of History, Habitat, Harvest, Hatcheries and
Hydrolectric Issues in the Hood Canal Region ..................................................... 69
Tribal Communities, Economies, and Land Use .......................................... 69
Pope and Talbot Establish Puget Mill Company at Port Gamble ................. 70
Effect of Expansion of Timber Harvest on Watersheds ............................... 72
Salmon Harvest in Western Washington and Hood Canal ........................... 74
Hatcheries...................................................................................................... 75
Hydroelectric Power ..................................................................................... 76
Summary of Interdisciplinary Aspects of the Current Study and Conclusions .... 77
APPENDIX .......................................................................................................... 78
REFERENCES.................................................................................................... 80
vi
List of Figures
Figure. 1. Hood Canal Region .................................................................................. 14
Figure 2. Equation to determine total LDI for a landscape unit ............................... 20
Figure 3. Combined South Fork Skokomish and Skokomish River-Frontal Hood
Canal Watersheds (From HUC-10 layer). ................................................................ 29
Figure 4. Flowline data showing all tributaries contributing
to Skokomish River watershed ................................................................................ 30
Figure 5. Complete Skokomish watershed and land use classifications. ................ 33
Figure 6. Lower Skokomish valley land-use classifications ................................... 34
Figure 7. Full Skokomish Watershed with 100 meter buffer applied and adjacent
land-use classifications ............................................................................................ 35
Figure 8. Linear county road data layer applied to the existing map ...................... 37
Figure 9. Road data integrated into the existing GIS ..................................................... 38
Figure 10. Data table from ArcMap GIS software converted
to Microsoft Excel.................................................................................................... 41
Figure 11. Coho distribution in the lower Skokomish River watershed ................. 43
Figure 12. Integration of coho activity data from SalmonScape with LDI
classification............................................................................................................. 44
Figure 13. Results of the basic statistic function from ArcMap using
“LDI_CALC_LOG” field ........................................................................................ 46
Figure 14. LDI coefficient distribution by area for Skokomish River .................... 48
Figure 15. Total number of parcels exhibiting
LDI value range by impact type............................................................................... 49
vii
Figure 16. LDI Values for Certain Parcels Intersecting
Coho Distribution Sites ............................................................................................ 50
Figure 17. Results of the basic statistic function from ArcMap using
“LDI_CALC_LOG” field: Coho Activity ............................................................... 51
Figure 18. Total number of parcels exhibiting LDI value range: Coho .................. 52
Figure 19. Mean LDI at three reaches of Lower Skokomish: Coho Spawning ...... 53
Figure 20. LDI at coho rearing reach ...................................................................... 55
List of Tables
Table 1. Low-Medium-High LDI Classifications..................................................... 25
Table 2. Land-use Classifications of the Skokomish Drainage ............................... 32
viii
Acknowledgements
I must acknowledge the following individuals for their various roles in the
completion of the work presented here. Without their contributions, the work
presented here would have been excessively more challenging or impossible to
complete.
My thesis reader and advisor, Dr. Carri LeRoy. Dave, Teresa, and Mendy at the
Pacific Northwest Salmon Center. Collyard and Von Prause with the Washington
Department of Ecology. Fellow members of my MES cohort and all faculty,
professors, and staff involved in the program, especially Krystle Keese for her
guidance in GIS software procedures. Barry Berejikian and others with NOAA
Fisheries. Elizabeth Greenleaf for her gravitas. Gail Wootan for her incredible
commitment to the program. My co-workers at the Gig Harbor Fly Shop. And my
parents for their unwavering support.
ix
Chapter One: Introduction
Beginning at the turn of the 20th century, anadromous Pacific salmonids
have faced continued declines across the majority of their range (Lichatowich,
1999; Montgomery, 2003). The reasons for these declines are multiple, the factors
complex, and synergisms among factors powerful.
Overarching themes for the
declines have emerged within the previous two decades and revolve around what
are frequently referred to as the four (or sometimes five) “H’s” (Montgomery,
2003). The five H’s are comprised of: a) harvest, b) habitat degradation, c)
hydroelectric power, d) hatchery issues, and e) history.
A tremendous amount of
research has been conducted examining these factors and the relationship to
salmonid declines at various scales and recent scholarship has indicated that there
is a need to study environmental issues at multiple spatial and temporal scales (Lele
& Norgaard, 2005; Cresswell, 2013). Furthermore, there is increasing evidence
that multidisciplinary, transdisciplinary, and interdisciplinary approaches may
prove effective when confronting environmental concerns because of the wide
array of factors involved and the complexity of ecological systems (Wells, 2013).
The issue of habitat degradation, particularly spawning and rearing habitat in river
systems, has been thoroughly researched with great emphasis placed on the effects
of timber harvest and the alteration of landscapes for human settlement purposes.
1
The current study was an examination of human- induced land-use
disturbance intensity (LDI) on the Skokomish River in western Washington State
and included an assessment of LDI in stream sections exhibiting coho salmon
activity.
The introductory chapter provides context and rationale for the current
research. Subject matter reviewed in Chapter One includes: (a) historical human
land-use patterns and the effects on salmonids in the Pacific Northwest in general
and the Hood Canal region in particular, from indigenous land-use through
Euro-American timber extraction, to current populations, (b) predominant habitat
attributes of anadromous salmonid spawning streams in general and biological
characteristics of coho salmon and their utilization of habitat, and (c) a review of
the habitat characteristics of the Skokomish River basin.
Chapter Two is an assessment of the literature which addresses
methodological approaches for quantifying the effects of land-use disturbance
intensity (LDI) within watersheds, and the manner in which quantified LDI can be
compared to additional ecological circumstances in effort to better model land-use
disturbance effects.
The basic principles of assessing land-use disturbance within
watersheds will be addressed and examples of how researchers utilize this
quantified data will be highlighted.
Chapter Three is a detailed account of the LDI quantification methodology
employed during the current study. The use of ArcMap software for conducting
these quantification procedures is documented and described. Specific procedures
2
developed by Collyard and Von Prause (2009, 2010) and amended during the
current research are explained.
The integration of existing GIS data regarding
salmonid behavior is also addressed.
Chapter Four is a presentation of the analysis of the data examined and the
results of the current study.
A quantification of human induced land-use
disturbance at the watershed-scale of the Skokomish River basin is provided.
Additional analysis examines quantified land-use disturbance within reaches of the
Skokomish watershed exhibiting coho salmon activity.
The intention of Chapter
Four is only to deliver results obtained through analysis of gathered data and within
the framework of the current study.
A discussion of the results and potential
implications is undertaken in Chapter Five.
Chapter Five is an expansive discussion of the current study itself, the
potential importance and implications of the results, and recommendations for
future study. The way in which land-use disturbance affects spawning and rearing
habitat, thus potentially affecting salmonid abundance in the Pacific Northwest is
addressed. The chapter also explores how future land-use studies may be
conducted and how additional factors may confound study efforts.
Chapter Six is a discussion of the interdisciplinary aspects which frame the
current study and provide context for developing future research questions.
Despite a tremendous body of research, the necessity of studies of Pacific
3
salmonids, and the way in which humans interact with them remains.
Focused
studies embedded deeply within a particular discipline continue to provide
researchers with baseline data and remain practical.
However, given the
complexity of the factors related to salmonid recovery efforts, interdisciplinary and
transdisciplinary approaches should prove germane when addressing
environmental concerns in general and may be particularly suited to salmonid
recovery.
Human Land-use Patterns and Interactions with Anadromous Salmonids
Indigenous Northwest Populations and Salmonids
Lichatowich (1999), in Salmon Without Rivers, utilized multiple anthropological
studies in describing the earliest inhabitants (arriving between 15,000 and 13,000
years ago) of the Pacific Northwest and their connections with anadromous fish.
One theory presents these earliest populations as unlikely to have utilized
salmonids as substantial forage because of the lack of suitable of salmon spawning
habitat among the oft-glaciated watersheds these groups inhabited (Matson &
Coupland, 1994). Rather, these early populations existed in small groups,
exhibiting a migratory lifecycle centered on opportunistic harvest of both terrestrial
and aquatic fauna, of which salmonids may have played only a small part. In
accordance with this theory, “salmon based economies” (Lichatowich, 1999, p. 27)
4
centered around salmon-bearing watersheds, likely grew as indigenous populations
increased and alternative resources became more scarce and less spatially
concentrated throughout newly opened post-glacial habitat in the migratory areas
the early populations traveled through and inhabited seasonally.
Alternate anthropological studies cited by Lichatowich (1999) suggest that
indigenous salmon economies increased and then flourished with shifting
postglacial habitats.
Improved habitat availability resulted in a robust increase in
stable salmon runs, thus decreasing the necessity for excessive migrations of
human populations (Faldmark, 1986, in Lichatowich, 1999).
Regardless of whether influenced by the deterioration of alternative forage
sources or the increase in harvestable salmon, Lichatowich (1999) and others have
extensively chronicled the increasing importance of salmon in the economies of the
growing indigenous populations from 9,000 years ago to the peaks of these
populations during the 18th and 19th centuries.
As climates stabilized and the
abundance of salmon increased, indigenous populations began to develop
communities at sites where harvestable salmon runs occurred. These sites would
include marine shorelines along salmon migratory routes and near the terminal
estuaries of rivers as well as streamside locations along larger rivers (Lichatowich,
1999; Mongtomery, 2003). The post-glacial Pacific Northwest during this peak of
early human populations has been described as “one of the most densely populated
nonagricultural regions of the world” (Boyd, 1990, in Lombard, 2006). Harvest
5
by indigenous populations at the peak may have actually approached numbers
similar to commercial harvests seen near the beginning of the twentieth century and
this extensive harvest of salmon could have stressed individual runs (Taylor, 1999).
Despite the shift from a migratory lifestyle to settlement based on salmon
resource availability, the indigenous inhabitants likely did not impose substantial
land-use disturbance in the areas of settlement which would have impacted
anadromous fish runs; however, land-use disturbances did occur. In an area of
settlement, trees would be felled to utilize for shelters and canoe building (Eells,
1889). Prescriptive fire was utilized as a means to promote new vegetative growth
of plant-derived food like camas and to entice megafauna like deer and elk into
grazing areas (Robbins, 1997).
Throughout the region, indigenous populations exercised similar lifestyles
based on harvesting available resources with minimal disturbance to the landscape.
During the 1800’s, three tribes inhabited the area of Hood Canal that was addressed
in the current study: the Suquamish, the S’Klallam, and the Skokomish.
Anecdotal accounts from members of these populations (Kitsap Sun, 1991) suggest
that the tribes adhered to the prevailing lifestyles of the region, with the
predominant resources utilized being timber, edible plants, salmon and other fish,
terrestrial fauna, and marine shellfish.
With the influx of Europeans,
Euro-Americans, and other immigrant ethnic groups, more substantial land-use
disturbances occurred throughout the region and near Hood Canal specifically.
6
European/Euro-American Settlement and Timber Extraction
The degree of shifting land-use patterns capable of affecting anadromous
salmonids intensified in the late nineteenth and twentieth centuries with the
large-scale influx of Europeans, Euro-Americans, and other nonindigenous people
into the region.
In the mid-1800’s, the California Gold Rush provided a catalyst
for a sharp increase in timber harvest in the Pacific Northwest (Chiang & Reese,
2003). The Puget Sound region was particularly suited for this purpose.
Extensive old growth forest in close proximity to protected marine harbors allowed
for the efficient harvest of timber, transportation of logs to mills, and the loading of
ships bound for San Francisco (Chiang & Reese, 2003.). In 1853, Pope and Talbot
established the Puget Sound Mill Company at Port Gamble, WA on the northern
end of the Kitsap Peninsula on Hood Canal (Chiang & Reese, 2003).
Mills were often located on streams to utilize hydrological energy
(Lichatowich, 1999). The effects of large-scale timber harvest on anadromous
salmon streams were substantial as mills created vast amounts of sawdust capable
of blanketing river bottoms where incubating salmon eggs were deposited and
excessive sawdust could disrupt returning adult fish and outmigrating juveniles by
clogging their gills with debris (Lichatowich, 1999). Timber harvesters also
utilized the streams for transporting logs and many harvest sites existed near
marine shorelines, the mouths of rivers, and within the riparian zone near the banks
farther upriver (Committee on the Protection and Management of Anadromous
7
Salmonids, 1996). Sedell and Luchessa (1981) indicated that trees had been
cleared 2 miles inland along Western Puget Sound and Hood Canal shorelines and
as many as 7 miles inland near streams and rivers.
Harvest efficiency decreased
during periods of low river flows therefore, splash dams were created as a means to
transport large amounts of timber at a single instance.
Splash dams were erected
above pool areas between riffles and as water collected, the pools would be filled
with logs.
At prescribed intervals the dams would be removed (frequently by
dynamite) and the presence of large amounts of logs in the streams and increased
unnatural flows disbursing sediment and other stream litter disrupted spawning and
rearing habitat (Lichatowich, 1999).
In addition to disruptions from splash dams, multiple effects from the close
proximity of harvest sites to rivers occurred when riparian timber and vegetation
were removed. Forest cover in riparian zones provide shade and keep rivers cool,
the root systems of large trees stabilize river banks and mitigate erosion, and
naturally fallen trees provide woody debris utilized as cover by juvenile salmonids,
as well as introducing nutrients which establish aquatic food chains. Additionally,
rivers and streams were often cleared of obstacles, primarily naturally-occurring
fallen trees and large woody debris, which would impede the transport of timber,
further altering spawning and rearing habitat (Committee on the Protection and
Management of Anadromous Salmonids, 1996; Lichatowich, 1999).
In the mid-20th century, technological advancements allowed for the
8
extraction of timber on a greater array of geographic landscapes. Heavy
equipment was capable of operating on steeper graded slopes and extensive road
systems were developed to provide access for the machines. Salmon streams were
now being impacted on greater spatial scales. The construction of roads and
clear-cutting of timber on steeper slopes caused soil and sediment destabilization
ultimately resulting in potential land-slides, depositing excessive sediment into
streams (Committee on the Protection and Management of Anadromous
Salmonids, 1996; Lichatowich, 1999). One study indicated that landslides were
approximately 25 times more likely to occur in clear-cut areas and areas near roads
than in forested areas (Lyons & Beschta, 1983).
Historical Timber Harvest Around Hood Canal
After Pope and Talbot’s 1853 establishment of the Port Gamble mill, the
succeeding years saw numerous logging camps established along the eastern and
southern shores of Hood Canal including present-day Seabeck, Bangor, Nelitta,
Holly, Dewatto, Union, and Belfair (Dunagan, 1991, p. 60). In the late 19th
century it was anticipated that several railroads would be routed through Union and
the cost of land parcels skyrocketed (Dunagan, 1991, p. 61). The town of Clifton
(now Belfair) also experienced a boom when roads connected the settlement with
Sydney (now Port Orchard). Also, in the late 1800’s, the community of Dewatto
experienced moderate growth due to its role as a connecting marine landing on the
east side of Hood Canal. In 1927, voters approved the formal establishment of a
9
port and subsequent pier and dock expansion but the plans never came to fruition
(Buchanan, 1930, as referenced in Dunagan, 1991; “Dewatto Citizens Petition…”
from
http://www.historylink.org/index.cfm?DisplayPage=output.cfm&file_id=9719).
The historical record clearly indicates that timber harvest was the primary
driver of land-use disturbance in the Hood Canal region post-settlement and the
effects of large-scale timber harvest on salmonid streams is addressed in the
previous section.
Although timber harvest likely held the predominant role in the
degradation of these streams, additional factors including the transformation of
cleared land to agricultural use and the harvest of salmon as a food supply for the
population of timber workers may have contributed as well (Platts, 1991).
Coho Salmon (Onchorynchus kisutch) Characteristics: Understanding
Freshwater Spawning and Rearing Habitat Behavior
Anadromous salmonids are those which are born in a freshwater setting,
rear in freshwater, begin a migration to marine waters as juveniles, exhibit
substantial growth during varying years at sea and return to spawn as
sexually- mature adults (Groot et al., 1995). For the purposes of framing the
current research, coho salmon are reviewed here; however, individual salmonid
species exhibit characteristics which are unique and varied. Currently, the
Skokomish River supports coho, chinook, chum, sea-run cutthroat, and steelhead,
10
with additional resident rainbow and coastal cutthroat trout, bull trout, stray pink
salmon, and there are current plans to establish a population of sockeye salmon
which would utilize the impoundment of Lake Cushman on the Skokomish.
Coho salmon are 1-3 “salt” fish meaning they spend one to three years
feeding and growing at sea. Research conducted from 1962 to 1970 indicated that
2-salt fish were the most common at 68.2% of the catch, while 1-salt and 3-salt fish
were measured at 22.0% and 7.5% respectively (Higgs et al., 1995). Additionally,
certain male individuals of a particular age-class may reach sexual maturity early
and return to spawning streams one year earlier than the remaining individuals of
the same age-class.
These individuals are frequently referred to as “jacks” or
“precocious males.”
Coho diet preferences in maturing, seagoing fish vary
dependent on region, however, larval and adult baitfish (including herring, sand
lance, and anchovies), euphausiids, amphipods, and squid constitute the majority of
adult coho diet (Higgs et al., 1995).
Populations of coho returning to spawn in Hood Canal typically enter their
natal streams in September, October, and November with spawning activities
occurring in November and December (Clarke & Hirano, 1995). Entrance to natal
streams is influenced by river flows and tidal fluctuations with large precipitation
events often prompting stream entry and subsequent upstream migration, as rising
water levels allow greater access and in some cases may move impediments such as
impassable beaver dams or excessive logjams.
Tabor et al. (1995) described this
11
phenomenon occurring on several Hood Canal streams during December 1992
when the city of Bremerton received over 14 cm of rain from Dec. 7-10.
Spawning Habitat Preferences
Coho typically exhibit a preference for smaller streams or the tributaries of
larger streams when selecting habitat for spawning.
Beds of deposited eggs
(redds) are often located in moderately paced sections of rivers between pools and
riffles. Moderate stream flow assists the female in removing sediment when
digging the redd and provides adequate oxygen to incubating eggs. Additionally,
moderate velocity sections are less likely to scour out substrates once eggs are
deposited and buried (Bjornn and Reiser, 1991). Gravel and pebble substrates are
often preferred for coho egg deposits with larger cobbles being infrequently
utilized (Mull, 2005).
Juvenile Rearing
Juvenile coho will typically spend the first year of life after emergence as
fry in the natal stream before outmigrating as smolts in late April and May;
however, Tschaplinski (1987) found that in some systems, coho fry may
outmigtrate with smolts and rear in the estuary before returning to the stream in fall.
Tschaplinski also noted that these fry exhibited more rapid growth than those
remaining in the streams.
Juvenile diets are typified by insects with chironomids
comprising a particularly important component of their diet. Yearling smolts
12
continue to feed on insects although other salmonid fry, particularly chum and pink
fry, are a substantial component of smolt diet. In estuarine settings, juvenile fry
also feed on insects; however, amphipods and copepods also contribute to diet.
Smolts in estuaries and marine environments feed primarily on insects and
amphipods, but at this life stage, small fish are also a substantial source of forage
(Higgs et al., 1995). Additional examinations of coho habitat suitability are
addressed in following sections.
Unique Characteristics of the Current Study Site: The Hood Canal Region
The Hood Canal region represents a unique area for watershed land-use
disturbance study. Hood Canal is bordered by on the west by the Olympic
Peninsula and the Kitsap Peninsula on the east. The canal is a glacially-carved
fjord with a general north-south orientation but “hooks” sharply to the east in its
southern reaches (Figure 1).
13
Figure. 1. Hood Canal Region showing major watershed drainages.
14
The region differs from other areas of Western Washington with salmonid streams
in several regards and in some ways may be represent a transitional area between
the highly populated urban areas which cover much of the eastern shore of Puget
Sound and the more sparsely populated rural or wilderness areas of the greater
Olympic Peninsula to the west. The eastern and northern edges of the Kitsap
Peninsula contain the population centers of Belfair, Port Orchard, Bremerton,
Silverdale, Poulsbo, Kingston, Seabeck and Port Gamble.
West and southwest of
Belfair, the southern shore of the “hook” of Hood Canal is far less densely
populated although cabins and homes are almost continuous along Highway 106.
Interestingly, although in general, the greater Olympic Peninsula has lower
population density, the western shore of Hood Canal is paralleled for much of it
range by Highway 101 and is far more developed than the eastern shore on the
Kitsap Peninsula.
The Skokomish River empties into Hood Canal in the southern
reaches near the town of Union.
Two highways cross the Skokomish, U.S. 101
and state Highway 106. The Skokomish Indian Reservation is situated between
the two highway crossings.
Skokomish River Salmonid Habitat
The Skokomish River, examined in the current study, is considered part of
Water Resource Area (WRIA) 16 and exhibits characteristics consistent with coho
behavior and habitats as explained in this section and the historical record indicates
an abundance of coho salmon under relatively undisturbed conditions.
Based on
15
the literature, the current study site is representative of prime habitat for coho,
provided the river had experienced minimal disturbance.
A 2003 report from the
Washington State Conservation Commission examined salmon and steelhead
habitat limiting factors in WRIA 16 (Correa, 2003). “Stocks were evaluated as to
status by the state and tribes in the 1992 Salmon and Steelhead Stock Inventory
(SASSI). Washington Department of Fish and Wildlife has updated the report in
the 2003 Salmon and Steelhead Inventory (SaSI)” (Correa, 2003). Both the 1992
and 2003 reports determined Skokomish River coho stocks to be rated as “healthy”
(Correa, 2003, p. 42).
Watershed Characteristics and Limiting Factors for Skokomish Coho
The Skokomish River drainage is the largest in the Hood Canal region and
encompasses 240 square miles with mainstem inputs of 80 miles and tributary
inputs of 260 miles (Correa, 2003). Two major forks (the North Fork and South
Fork) combine to create the “mainstem” which flows for 9 miles to the terminus in
Hood Canal. An extensive report prepared by Mason County characterizes land
cover, hazard areas, water quality and priority habitat and species areas by
watershed and reach (Mason County Shoreline Inventory and Characterization
Report (MCSIC), 2012) and could provide data to be used for comparison with the
results of the current research. Land- uses are classified primarily as “agricultural,
vacant, forestry, and/or residential.” For each stream reach identified in the report,
a percentage of land-use type within the reach is provided.
Additionally, critical
16
habitat for several salmonid species endemic to the Skokomish are identified at
each reach.
Chapter Two: Methodological Literature Review
Land-Use Development Intensity Index
Brown and Vivas (2005) determined that an anthropogenic land-use
development intensity index (LDI) could be calculated from land-use data and
corresponding energy usage in a measured geographical unit through the utilization
of a geographic information system (GIS). Units of geographical scale in which
the LDI could be determined include river, stream, or lake watersheds, and even
individual wetlands (Brown and Vivas, 2005). The LDI of a particular unit can
then be compared against other ecological data to determine land-use relationships
with species and habitats.
According to Brown and Vivas (2005), the greater the intensity of human
activities, the greater the ecological effects on a given landscape unit, and units
exhibiting the highest-energy land-uses may have extremely limited or no natural
ecological systems.
As the authors indicate, many human populated landscapes
exist between the extremes of highly developed human population centers and
“wildlands” (Brown & Vivas, 2005). The Hood Canal region seems particularly
representative of this assessment based on its geographic position between the
highly developed I-5 corridor and largely wild Olympic Peninsula.
17
Brown and Vivas (2005) define a landscape unit as “the ecological
community, drainage feature, or hydrologic system that is being studied. For
instance, the study unit could be an individual ecological community such as a
wetland, or a stream segment, or a sub-watershed drainage basin” (p. 302). It is
important to note that the authors determined a 100 m buffer was adequate to
quantify the disturbance surrounding a particular watershed and no significant
difference was found between LDIs calculated at 100 m and 200 m. This
determination was based on the area the authors examined, which were watersheds
in Florida exhibiting low-gradient drainages.
Brown and Vivas (2005), developed LDI coefficients for different land-use
classifications based on Emergy Accounting, a system developed by H.T. Odum
(1996). Thermodynamic bases of all forms of energy and materials are converted
into the equivalent of solar energy. “Emergy is the amount of energy required to
make something” (Brown and Vivas, 2005, p. 302). Units of emergy are referred
to emjoules, and quantify the energy consumed in transformations.
Thus, all
energy utilized in human development of environments (solar, fuel, electric, and
human work) can be defined by the amount of solar energy necessary for the
production of each. The unit used for expressing this energy utilization is the
solar emjoule (sej). The quantification of human-development intensity through
land-use can then be calculated from annual usages of non-renewable energies
(solar emergy joules [sej]) per unit (sej/ha yr-1 ). Brown and Vivas (2005) calculate
18
the “LDI coefficient as the normalized natural log of the empower densities
(p.292).” Calculating the natural log utilizes a base of “e” or 2.7182818. Brown
and Vivas (2005) continue, stating, "first the natural log of the empower densities
were calculated and then the resulting values were normalized on a scale from 1 to
10, with the LDI coefficient for natural lands equal to 1.0 and a LDI coefficient of
10.0 for the highest intensity land use, the Central Business District (p. 294)." The
use of a natural log transformation and the scaling of values from 1 to 10 indicates
that the coefficient values assigned to different land use types represent dramatic
increases in land use disturbance as values increase. The methodology used in the
current study (see Chapter 3) accounts for this disturbance intensity increase in the
models. Furthermore, when analyzing quantifications of LDI, it is imperative to
recognize the increasing influence that higher LDI coefficient values will have on
area-weighted results for units within the “area of influence.”
Because the LDI is a measure of human alterations, only usage of
non-renewable energies were included. Land-use or land-cover categories were
then defined and an LDI coefficient assigned to each category (See Appendix).
The equation represented in Figure 2 is then used to determine an area-weighted
land-use ranking for an individual unit.
19
Figure 2. Equation to determine total LDI for a landscape unit (Brown & Vivas,
2005).
Proximity of Land-Use Effects on Stream Fish and Habitats in Great Lakes
Tributaries
Stanfield and Kilgour (2012) conducted an extensive seven-year study
reviewing LDI relationships to stream fish and their habitats in tributaries to Lake
Ontario at 312 individual sites analysis sites.
Multiple spatial scales were used in
the analysis: catchment, eight varied riparian polygons differing in width and
length, upstream polygons of 1.6 and 3.2 km, and residual upland area unaccounted
for by the polygons.
A covariate analysis was conducted and fish assemblage
diversity, habitat suitability, and temperature variation were compared to LDI
measurements as well as with each other. The objectives of the data analyses were
to “(i) summarize the variations and covariations of the different measures of LDI
20
and template variables against measures of fish community composition and
in-stream physical conditions (response variables) and (ii) to determine the
magnitude of variation in the response variables that was uniquely explained by
LDI measured at different scales” (Stanfield and Kilgour, 2012). Data were
analyzed using methods similar to those proposed by Brown and Vivas (2005).
LDI values ranging from 0 to 50 were applied to parcels characterized from
undisturbed full forests to fully developed urban areas and examined in relation to
the response variables.
The results indicated that “sites with high LDI in the whole upstream
catchment also generally had high LDI in smaller polygons, regardless of the
proximity of the polygon to the stream site” (Stanfield and Kilgour, 2012). Stream
temperature was shown to be variable at low levels of LDI and consistently high at
LDI levels above 18. Pure fish biomass declined in linear fashion with increased
LDI. Additional findings indicated a greater influence of LDI in headwater
streams on downstream catchments and fish assemblages.
These findings were
contrary to a previous study (Frimpong et al., 2006) which determined that upland
influence decreases to almost zero beyond 150 m. Stanfield and Kilgour (2012)
suggest that:
“…land use in proximity of the headwater streams in these upland areas
directly influences the factors that influence both fish biomass and taxa richness of
21
downstream areas, and that future management activities should extend beyond the
main river and its valley.”
Land-use and Coho in Snohomish River (Pess et al., 2002).
22
Pess et al. (2002) studied a suite of habitat factors affecting adult coho
returns on the Snohomish River system.
Pess et al. (2002), utilized a hierarchical
linear model (HLM), a method frequently used in social sciences to model nested
units, to examine the relationships between habitat characteristics and fish
abundance; the HLM was applied to describe spawner counts within years.
Single-variable regression analyses were also used to explain the variation in fish
density as a function of an individual habitat characteristic, of which, land-cover
and land-use was one. Multiple-regression models were used to examine fish
densities as a result of a suite of habitat characteristics.
Models were applied to
both stream-reach and watershed-scale units (Pess et al., 2002).
The HLM indicated a negative correlation over time between spawner
abundance and percentage of land in urban or agricultural use, whereas a positive
correlation was shown between coho abundance and percent forest cover. Areas
designated with more than 50% forest cover showed 1.5-3.5 times greater
abundance in coho spawners than areas with less than 50% forested land. The 10
index reaches studied with highest coho spawner abundance exhibited more than
60% forest cover in riparian areas. Additionally, in stream reaches exhibiting less
than 10% agriculture, salmon abundance was two to three times greater (Pess et al.,
2004).
Allan’s (2004) Four Challenges to Understanding Land-Use Change Effects
on Watersheds
23
In a comprehensive analysis of land-use effects on watersheds titled,
Landscapes and Riverscapes: The Influence of Land-use on Stream Ecosystems,
Allan (2004) identified four difficulties for understanding these land-use effects:
(a) covariance between anthropogenic and natural gradients, (b) spatial scale
issues, (c) nonlinearities, and (d) legacy effects.
Covariation
Covariance often exists between the effects of land-use and natural landscape
features. Parent geological material, soil types, and topography will often dictate
suitability for anthropological land-uses. Therefore, caution should be exercised
to avoid overestimating the contribution of land-use influence on a particular
ecological variable (Allan, 2004).
Spatial Scales
Three common designations are frequently used in investigations of stream
conditions and land-use relationships.
A local reach is usually designated as the
buffer area 100 m to several hundred meters in width along both banks and some
hundreds of meters to one kilometer in length.
A riparian buffer area is usually
described as a similar width as the local reach (100 m or more) but contains a
greater length upstream or even an entire upstream distance on smaller waters. A
catchment is usually designated as the entire drainage upstream from a specific site.
24
There should be an expectation of variances of responses to large-scale or
local-scale factors.
Thus the effects of local inputs or influences may be
distributed along great distances of the stream. Therefore, the appropriate level of
scale for study should be chosen based on the best available literature (Allan,
2004).
Nonlinearities
Allan (2004) pointed to several studies which showed declines in species diversity
and the index of biotic integrity (IBI) with increasing urbanization or increasing
impervious area (IA). Generally, the evidence suggests that a range between
10%-20% of IA or urban land within the riparian buffer equates to a threshold for
stream health; however, some studies indicate that influences of hydrology based
on geography and resultant downstream flow are strong indicators and a single
threshold should not apply to characterizing an entire watershed.
Responses are
also likely to vary based on scale and thresholds will shift will based on the area of
the catchment examined and resultant thresholds (Allan, 2004).
Legacy Effects
Consequences from disturbances which are still influential long beyond the
occurrence of the disturbance are known as legacy effects and may be
underemphasized.
Wang et al. (2001 as cited in Allan, 2004) reported variations
in fish metrics along an urbanization gradient however, habitat quality varied little.
25
The effect was attributed to a legacy of prior influences of agricultural land-use.
Furthermore, previous anthropogenic geomorphological channel changes may
have had far reaching effects on physical structure and hydrology, and recovery
from these disturbances takes markedly longer than from changes in land-use
(Allan, 2004).
LDI Applied Within Washington State Watersheds
Collyard and Von Prause (2009) developed a methodology for the
utilization of LDI analyses for watersheds in Washington State using a 1-10 scale
with an LDI coefficient of 1 representing undisturbed landscapes and 10
representing the greatest amount of anthropogenic disturbance and (see Chapter
Three and Appendix). The researchers determined that a 100 meter riparian buffer
was appropriate for an LDI analysis of watersheds in Washington State. Based on
the 10-point scale, Collyard and Von Prause (2009) further delineated LDI ratings
into three categories represented in Table 1.
Table 1. Low-Medium-High LDI Classifications (from Collyard and Von Prause,
2009).
LDI Score
0-2.00
2.01-5.50
Description
Low Impact
Nearly pristine conditions
Medium Impact
26
Low Intensity Residential/ Agricultural
5.51-10.00
High Impact
Urban
27
Chapter Three: Methods
The current study was an attempt to quantify land-use disturbance in
watersheds draining to Hood Canal, based on an index proposed by Brown and
Vivas (2005) and adopted by Collyard and Von Prause (2009) of the Washington
State Department of Ecology (DOE). Land-use disturbance indices within
particular watersheds were calculated at the catchment level (entire system
drainage) utilizing a method which allows for more focused analysis of smaller
units of the watershed should a researcher desire to do so. Researchers Collyard
and Von Prause (2009) indicated the need for a statewide uniform classification of
land-use disturbance within all watersheds in Washington State. A complete
dataset of quantified land-use disturbance within watersheds would be beneficial
for the ease of comparing additional spatial watershed data with adjacent land-use
patterns.
LDI Calculation Overview
ArcGIS 10.2 software was the primary tool utilized for quantifying land-use
patterns and determining LDI values within a particular watershed.
An existing
dataset layer produced by the Washington State Department of Ecology and
Washington State Department of Revenue (2009), classifies land-use for the entire
state based on tax parcel units.
LDI coefficients can then be applied to each
land-use classification category using the methodology proposed by Brown and
Vivas (2005) and refined by Collyard and Von Prause (2009) for Washington State.
28
These data, when used in conjunction with GIS layers from the National Hydrology
Database (NHD) and scaled to Hydrological Units based on USGS Hydrologic
Unit Codes (HUCs), and with the application of a 100 meter riparian buffer in a
GIS, allow the user to calculate an area-weighted LDI for a chosen geographic unit
at a watershed reach based on the user’s needs.
Primary Study Site: Skokomish Watershed-Land-use Classifications and LDI
Coefficients
The primary study site was the Skokomish watershed which drains from the
southeast flanks of the Olympic Mountains on the Olympic Peninsula to southern
Hood Canal. For the purposes of the current the study (calculating a
watershed-scale LDI value), two hydrological units (defined under the USGS NHD
classification as South Fork Skokomish River and Skokomish River-Frontal Hood
Canal with HUC-10 digit codes of 1711001701 and 1711001702 respectively)
were combined and are represented in Figure 3. The NHD flowline data
delineates the drainage of tributaries into larger streams allowing for full
watershed-scale analyses and is represented in Figure 4.
29
Figure 3. Combined South Fork Skokomish and Skokomish River-Frontal Hood
Canal Watersheds (From HUC-10 layer).
30
Figure 4. Flowline data showing all tributaries contributing to Skokomish River
watershed
31
The DOE and DOR dataset uses a coding system (see Appendix) to classify land
use for particular parcels (based on tax data from DOR). Table 2 shows land-use
classifications and the codes for parcels within the Skokomish Hydrologic Unit.
The coded land-use layer for all parcels can then be added to the GIS (shown in
Figure 5 at the watershed-scale and in Figure 6 zoomed to the lower-Skokomish
reaches for interpretation). Using the “Buffer” and “Clip” tools in ArcMAP, a 100
meter buffer can be applied and the adjacent land- use classifications for each parcel
can be clipped to the buffer (Figure 7).
32
Table 2. Land-use Classifications of the Skokomish Drainage
Land-use
Code
11
15
18
Description
Household, single
family units
Mobile home parks
or courts
All residential not
elsewhere coded
Land-use
Code
Description
68
Educational services
74
Recreational activities
75
Resorts and group camps
76
Parks
19
Vacation Cabin
25
Furniture and
fixtures
79
45
Highway and street
right of way
81
46
Automobile parking
83
48
Utilities
84
51
Wholesale trade
87
53
54
58
Retail trade-general
merchandise
Retail trade-food
Retail trade-eating
and drinking
Other cultural,
recreational, church,
cemetery
Agriculture (not classified
under current use law)
Agriculture classified
under use Chapter 84.34
RCW
91
Fishing activities and
related services
Public
timberland/non-designated
forest
Designated forest land
under Chapter 84.33 RCW
Undeveloped land
92
Noncommercial forest
88
59
Other retail trade
94
63
Business services
95
67
Governmental
services
Open Space land classified
under Chapter 84.34 RCW
Timberland classified
under Chapter 84.34 RCW
33
Figure 5. Complete Skokomish watershed and land use classifications.
34
Figure 6. Lower Skokomish valley land-use classifications. View zoomed and full
parcels shown for interpretation purposes.
35
Figure 7. Full Skokomish Watershed with 100 meter buffer applied and adjacent
land-use classifications.
Methodological Additions To DOE Framework: Accounting For All Roads in
LDI Calculations
A primary addition to the methodology proposed by Collyard and Von
Prause (2009) used in the current study, and one which should prove useful in
future research, was the integration of accounting for the LDI values of all roads
(not only those present in the existing DOR dataset) in the calculation of the LDI
indices for watersheds examined in the current study. According to the values
from Brown and Vivas (2005), 2-lane highways are given an LDI coefficient of
36
7.81 and accounting for these roads may result in different LDI scores when
examining at various scales.
County road data were added to the GIS (Figure 8); however, because the
road data are linear, it was necessary to apply a 35-ft buffer to the linear data line to
achieve a polygonal dataset which could be analyzed at intersections with the
existing 100 m buffer applied to the stream data. Using ArcMAP, subsequent to
the application of the 35-ft buffer to the roads, executing a spatial merge with the
existing land-use layer, and clipping areas where the buffered roads intersect with
the 100 m riparian buffer, a new GIS layer was created which accounts for the
intersection of roads within the riparian buffer zone (Figure 9).
37
Figure 8. Linear county road data layer applied to the existing map. View is
zoomed to show where roads may exist in close proximity to the watershed.
38
Figure 9. Road data integrated into the existing GIS map accounting only for spatial area
where roads intersect the 100 m riparian buffer used for LDI analysis and allowing for the
integration of applied LDI coefficients for roads in the Skokomish watershed. View is
zoomed for interpretation purposes.
39
Calculating Catchment-Scale LDI Values for the Skokomish Watershed
With the land-use classifications clipped to a 100 meter riparian buffer,
calculation of an area-weighted LDI becomes possible. ArcMap layers generally
have associated data tables providing additional information about the geography
being displayed in the GIS. Because each tax parcel is given an identification
number (and thus is a distinct unit in the GIS) in the original statewide land-use
layer, it was possible to calculate the physical area of each buffered parcel utilizing
the X-Tools Pro extension for ArcMap. Working within the layer’s data table also
allowed for the assignment of the LDI coefficients listed in Table 2 to individual
parcels and subsequent calculations, based on the classification system and
methodology utilized by Collyard and Von Prause (2009) as adopted from Brown
and Vivas (2005). Figure 10 shows an Excel adaptation of the data table for the
GIS. The first column (labeled FID) identifies individual parcels, the fourth column
(labeled LANDUSE_CD) contains the numerical designations for land-use
classification as shown in Table 2, the sixth column lists the area in square meters
of each parcel clipped within the 100 meter riparian buffer and the 11 th column
(labeled LDI_Coef) shows the land-use disturbance coefficient applied to each
parcel.
Special Considerations for LDI Value Assignments
A review of the table in the Appendix
will show several redundancies
among the “Land- use-CD numbers” applied for various land-use types as proposed
40
by DOE (2010). Additionally, several different LDI value coefficients may be
applicable to a single “land-use CD number.”
During the current research, I used
satellite photograph layers, some level of ground truthing, and the existing DOE
GIS classifications in an attempt to appropriately classify certain parcels and assign
an LDI coefficient for each parcel. For example, a particular parcel with a
“land-use CD” classification of “82” in the DOR/DOE dataset (classified as
‘agriculture’) may actually represent one of several different types of agricultural
use. Accordingly, LDI values will change based on the intensity of the particular
type of agriculture practiced.
Because of the nature of the scope of work, time,
and financial resources for the current study, I relied best judgment when assigning
LDI coefficients to parcels with a “land-use CD” classification which could receive
several different LDI coefficients.
I am familiar with the study area and with the
associated land-use types, so it is my hope that the values are as accurate as
possible; however, extensive ground-truthing would likely improve the accuracy of
any models derived from the dataset. Furthermore, experts familiar with
determining land-use from satellite photography may be able to more accurately
determine and assign land-use classifications for future analyses.
Performing The Area-Weighted LDI Calculation for the Skokomish
Watershed
Figure 10 depicts the data table associated with the Skokomish River GIS and
containing the data required to perform the area-weighted LDI calculation for the
41
Skokomish watershed. Using the ArcMAP field calculator tools, an additional
column (LDI_Calc) was added to the attribute table for the GIS dataset and
area-weighted LDI scores were able to be calculated by entering a function which
multiplied values in the “Area” column by the associated “LDI_Coef” column and
designating the results be returned to the new “LDI_Calc” column. Following the
methodology of Collyard and Von Prause (2009), all values in the “LDI_Calc”
column were then natural log transformed and finally scaled from 1-10 using the
ArcMAP field calculator. The “LDI_CALC_L” column contains final, transformed
and scaled LDI values for land-use disturbance intensity within the 100 m riparian
buffer.
Figure 10. Data table from ArcMap GIS software converted to Microsoft Excel
table showing essential data required for an area-weighted LDI calculation of the
Skokomish watershed.
42
Integrating Existing Salmonid-Related GIS Data
The Washington Department of Fish and Wildlife (WDFW) operates an interactive
GIS accessible via the web known as SalmonScape. According to the homepage for
SalmonScape, “SalmonScape delivers the science that helps recovery planners
identify and prioritize the restoration and protection activities that offer the greatest
benefit to fish (WDFW, retrieved from http://apps.wdfw.wa.gov/salmonscape/).”
Data presented in the mapping system are culled from the research results of state,
federal, tribal, and local studies and integrated into a sinuous system which is
readily accessed by any interested entity.
A primary feature of the data available in the SalmonScape interface which
is pertinent to the current research, is the documentation of salmonid utilization of
habitat within specific stream segments. This utilization is classified by species and
by type of utilization (spawning, rearing, etc.). Figure 11 shows coho activity in
the lower reaches of the Skokomish River watershed as classified within the
SalmonScape interface GIS dataset.
43
Figure 11. Coho distribution in the lower Skokomish River watershed. Source:
http://apps.wdfw.wa.gov/salmonscape/map.html
44
Using the geoprocessing tools in the ArcMap software, it was possible to integrate
SalmonScape data into the LDI dataset for the Skokomish River created during the
current study (Figure 12).
Figure 12. Integration of coho activity data from SalmonScape with LDI
classification data rendered during the current study.
By employing the same procedure used to calculate a watershed-scale LDI value
for the entire Skokomish (as described throughout Chapter 3), it was also possible
to calculate mean LDI values for individual stream segments exhibiting specific
coho activity as classified in the SalmonScape GIS.
45
Chapter Four: Analysis and Results
A major component of this research involved the effort required to understand,
adapt, and apply the methodology for LDI quantification.
The methods are
GIS-intensive which necessitated that I develop skills using the software and
expend the time performing the functions. These efforts constituted a tremendous
amount of study and application in and of themselves. A fortunate outcome of the
methodological research was the rendering of some interesting and applicable
results.
This chapter is a review of the results and analysis.
Following this
chapter is a discussion of the study, limitations and difficulties of the methods and
results, contextual information, and recommendations for refinement of the
methodology for salmonid habitat assessment and future research opportunities.
Skokomish River: Watershed-scale LDI
After performing the GIS methodology explained in Chapter 3, new data
emerged which quantified LDI for the watershed examined and bear further
scrutiny. Running a simple statistics summary using ArcMap yielded some
compelling results. Figure 13 shows the initial results of the “Statistics” function
when isolating the “LDI_CALC_LOG” field from the data table.
46
Figure 13. Results of the basic statistic function from ArcMap using
“LDI_CALC_LOG” field for Skokomish watershed LDI values.
Because all LDI values were area-weighted and scaled, the mean value of
all values from the “LDI_CALC_LOG” field shows the LDI for the complete
watershed catchment within the riparian buffer. Within the Skokomish watershed
riparian buffer, the mean LDI was 4.16 (n=1142, m=4.162, SD=2.36). Based on
Collyard and Von Prause (2009), this value (4.16) indicates that at the watershed
scale, the Skokomish River would be classified as “Medium Impact” or exhibiting
an LDI-value between 2.00 and 5.50 (see Table 1).
Individual Parcel LDI: Framing Disturbance Impact
The frequency distribution graph in Figure 10 provides a simple visual
representation of the distribution of the number of parcels exhibiting various
LDI-values within the riparian buffer along the Skokomish. The distribution of the
47
number of parcels exhibiting particular log transformed LDI values as represented
in the graph is indicative on one hand and potentially misleading when viewed in
isolation. The distribution graph shows that the greatest number of parcels within
the riparian buffer are classified as a “1” (or “undisturbed” based on the LDI
coefficients).
The graph also shows a substantial number of parcels are classified
by LDI coefficients between “4.2” and “7.4”, which indicates “medium” or “high”
impact. However, the frequency distribution does not represent spatial area, only
the number of parcels classified by particular LDI values. Therefore, additional
analysis was required to more accurately model land-use disturbance within the
riparian buffer of the watershed.
When examining the distribution of LDI values by area, an excessive
amount of spatial area is classified with an LDI value of “1” and other “low impact”
values (Figure 14). It is only when analyzing the area-weighted distribution of
LDI values based on the log transformation, that LDI impacts are recognizable.
48
Figure 14. LDI coefficient distribution by area for Skokomish River.
It was hypothesized that the Skokomish watershed would likely show a disturbance
index value indicating “low” impact. The rationale for this thinking was based on
visual analysis and the perception that watersheds with headwaters within or
closely adjacent to protected or managed areas (thus relatively undisturbed) and
large segments of riparian area characterized as low impact, would show low
impact at the watershed scale. However, based on the methodology for
calculating area-weighted LDI values proposed by Brown and Vivas (2005) which
utilizes a natural log-transformation for values, the disturbance may be far greater
49
than what a simple visual analysis or count and area analysis would indicate when
evaluating based on LDI-coefficient distribution alone.
Figure 15 shows the count distribution of parcels exhibiting log-transformed LDI
values. Viewing this distribution shows how the Skokomish River is classified as
a system experiencing medium-impact land-use disturbance when analyzed at the
watershed-scale.
Distribution of Parcels Within
LDI Impact Ranges-Skokomish Watershed
500
400
407
376
Parcel Count
359
300
200
100
0
0-2.00
2.01-5.50
5.51-10.00
Low Impact
M edium Impact
High Impact
Log Transformed LDI-Value Ranges
Figure 15. Total number of parcels exhibiting LDI value range by impact type
when examining the Skokomish River at the watershed-scale.
50
LDI-Values at Coho Distribution Sites on the Lower Skokomish River
It was possible to modify the established method for calculating watershed-scale
mean LDI values to analyze LDI in individual stream segments. Using the spatial
coho distribution data available from SalmonScape, stream segments exhibiting
particular coho distribution characteristics could be analyzed for land-use
disturbance. Figure 16 shows coho behavior distribution, land-use classification
within the riparian buffer, and the area-weighted LDI value for the parcels in the
extreme lower Skokomish River.
Figure 16. LDI Values for Certain Parcels Intersecting Coho Distribution Sites in
the extreme Lower Skokomish River. (Note: This map does not represent all coho
distribution for the Skokomish nor are all parcels labeled with an LDI Value to aid
in data interpretation.)
51
Reviewing the summary stats provided by ArcMap provided the mean LDI for the
extreme Lower Skokomish watershed reach examined in relation to coho activity
documentation and the simple frequency distribution of LDI values in the
watershed reach. Figure 17 shows the summary data provided by ArcMap.
Figure 17. Results of the basic statistic function from ArcMap using
“LDI_CALC_LOG” field for lower Skokomish watershed LDI values in stream
reaches exhibiting coho activity.
Within the lower-Skokomish reaches exhibiting coho behavior, the mean LDI was
3.86 (n=279, m=3.861, SD=2.75). While this values is somewhat lower than the
watershed-scale LDI value for the Skokomish of 4.16, an LDI values of 3.86 is still
classified as “medium” impact.
Figure 18 shows the parcel distribution of all
parcels within the lower Skokomish reaches intersecting with stream reaches
exhibiting coho activity.
52
Distribution of Parcels Within LDI-Impact
Ranges-Lower Skokmish Reaches
With Coho Activity
140
128
118
120
Parcel Count
100
80
60
33
40
20
0
1-2.00
Low Impact
2.01-5.5
5.51-10.0
M edium Impact
High Impact
Log Transformed LDI-Value Ranges
Figure 18. Total number of parcels exhibiting LDI value range by impact
type-Lower Skokomish reaches with coho activity.
The parcel LDI-value data indicate that the lower Skokomish reaches with coho
activity are characterized by a substantial number of parcels exhibiting low or high
impact land- use disturbance. The number of parcels with high LDI- values is clearly
influencing the mean LDI within this reach.
LDI-Value By Coho Activity Type-Lower Skokomish River
Figure 11 (presented earlier in the text) showed coho distribution sites as
documented by the SalmonScape web interface.
Using this spatial distribution
data, it was possible to perform some final analyses on LDI impacts at smaller
scales of stream reaches with coho activity.
There were some limitations to how
53
the LDI quantification methods could be applied at these reaches but a preliminary
investigation is presented below. A section in the Discussion chapter will
expound on these and other limitations and challenges in the study.
LDI at Coho Spawning and Rearing Reaches
Spawning Reaches
Spawning reaches were isolated (stream segments in red) and mean LDI values
were calculated for each reach. Three reaches were identified and are
characterized with the associated mean LDI-values in the map in Figure 19.
Figure 19. Three reaches of the Lower Skokomish watershed exhibiting coho
Spawning Activity and associated mean LDI.
54
The site showing an LDI of 4.82 is a small drainage basin east of Highway 101 near
the Skokomish Indian Reservation. The site showing an LDI of 3.04 encompasses
the mainstem of the Skokomish River above the Highway 106 crossing and below
the Highway 101 crossing as well as spawning reaches of the tributary Purdy
Creek. The site showing an LDI of 4.30 comprised of the remaining drainage
above Highway 101. The mean LDIs for each site and associated statistics were:
(n=3, m=4.82, SD=2.70), (n=11, m=3.04, SD=2.73) and (n=30, m=4.30,
SD=2.97). Based on these analyses, all spawning reaches in the study area would
be classified as Medium impact.
It should be noted that these reaches do not
account for all coho spawning habitat within the Skokomish watershed but simply
the area of the lower watershed chosen for this project and as represented by
SalmonScape.
55
Rearing Reaches
A single rearing reach was isolated (Site A) due to the overlapping of LDI parcels
(stream segment in green) and the mean LDI values was calculated and is shown in
Figure 20.
Figure 20. Single reach of the Lower Skokomish watershed exhibiting coho rearing
activity and associated LDI.
The mean LDI for the rearing site was 4.09 (n=14, m=4.09, SD=3.15). Based on
this value, the rearing site identified, like the spawning sites, would be classified as
Medium Impact.
56
It should be noted that similarly to the spawning sites, this site does not account for
all coho rearing sites within the Skokomish watershed but simply within the area of
the lower watershed chosen for this project. Furthermore, spawning and rearing
habitat often overlap. In the current configuration of the SalmonScape web
interface, it was not possible to delineate areas of overlap for additional and more
robust analysis.
Further assessment of these limitations and potential options for
addressing these issues is included in the Discussion chapter.
Summary of Analysis and Results
It was anticipated that a quantification of LDI within watersheds would be possible
with the utilization of the required data and GIS software to conduct the
calculations and analysis and this proved to be true. Collyard and Von Prause
(2009) provided a basis for this quantification procedure and subsequent analysis
which built upon the work of Brown and Vivas (2005). The current research was
an attempt to refine the methods and apply them for the purpose of fisheries habitat
assessments. In this regard, the current research appears to be successful in
advancing spatial assessments of LDI for watersheds. As with most research, there
are certainly areas which can be further refined and improved to yield more
powerful and robust results. Chapter Five is an extensive discussion regarding the
implications of the current research, challenges and limitations within the current
methodology, and recommendations for further refinement of the techniques and
avenues for future study implementing LDI analyses within watersheds.
57
Chapter Five: Discussion
The contents of Chapter Five are intended to provide additional context about the
current research, address challenges and limitations encountered, and examine how
this research “fits” within the current body of scholarship and may inform future
study efforts. Challenges and limitations are discussed along with a framing of the
current research and recommendations for future.
Limitations and Challenges Within the Current Study
The current study was limited by multiple constraints including data availability,
accuracy, and precision.
Additional concerns surrounded the adaptation of the
methodology for examining watersheds in the Pacific Northwest.
I, as the
researcher, was also limited by time and budget constraints which are addressed
here. Based on these limitations, the Skokomish River watershed was chosen for
the focus of this research.
Determining Scope and Framework
A substantial challenge for researchers interested in examining issues surrounding
salmonid fisheries in the Pacific Northwest is the task of determining what aspect is
intended to be studied. The body of research on the issue is voluminous and
navigating the literature can be difficult. Based on personal experiences, potential
researchers must conduct ample “pre-research” to obtain what could be considered
even a tenuous grasp on the situation. Beginning with the five “H’s” proposed by
58
Montgomery (2003), and reviewed in Chapter One should be requisite. Within any
of the five “H’s,” there are ever-increasing magnitudes of focus.
In the current
research, “Habitat” become the focus. For myself, the process of arriving at this
focus resulted from my experience in the MES program at Evergreen, two research
internships with a local salmon enhancement organization, and finally a personal
connection with the issue I intended to focus on.
Focusing on the Skokomish River
Even after determining that ‘habitat’ was the “H” I wanted to examine, what
particular aspect of this factor to assess became an additional challenge.
Personal
experiences combined with graduate internship experiences sharpened my focus.
Hood Canal became the initial focal point and several salmonid streams in the area
garnered my attention.
Initially, I intended to focus on the Dewatto and Tahuya
Rivers based on personal interaction with these watersheds and their respective
fisheries.
However, as I was interested in habitat, and based on data availability
and additional constraints, the Skokomish River watershed became the area of
focus.
I was intrigued by the diversity of land-use adjacent to the riparian zone
within the Skokomish watershed (discussed in Chapter One) and personal
interactions with the fishery on the Skokomish justified my decision to focus my
study here.
59
Hood Canal Salmonid Recovery and Skokomish River Coho
Several Hood Canal salmonid species have garnered much attention based on their
ESA listing status and various recovery efforts.
I was involved in two of these
efforts as an intern with the Pacific Northwest Salmon Center. Winter steelhead
returning to Hood Canal streams, Hood Canal summer chum, and certain
populations of chinook salmon returning to Hood Canal streams are considered to
be distinct population segments (DPS) or evolutionarily significant units (ESU) and
are currently listed as “threatened” under the Endangered Species Act (ESA)
(NMFS, 2010). Based on these listings, study efforts and recovery plans under
cooperative direction from several agencies have been undertaken.
As such,
population dynamics of these salmonid species may exhibit greater influence from
the supplementation efforts and study protocol which dictate the manner in which
these populations are managed.
The winter steelhead project, in which I participated in as an intern, is
attempting to evaluate the potential impact of hatchery broodstock supplementation
on wild populations as well as the effectiveness of such supplementation as a
recovery measure (Hood Canal Winter Steelhead Supplementation HGMP, 2012).
Summer chum are considered to be extinct in the Skokomish River but are present
in the nearby Tahuya and Union Rivers and are the focus of additional study and
recovery efforts.
60
Chinook salmon spawning in the Skokomish River are considered part of
the ESU known as “Puget Sound Chinook” and are included in study and recovery
efforts for this population.
The Hood Canal Coordinating Council report (2005)
characterizes these recovery efforts as “unique and potentially challenging
scenarios” and posits “the status of chinook salmon stocks in Hood Canal is
confounded by a long history of artificial introduction and production of stocks into
Hood Canal systems, severely degraded habitat conditions, and an extremely
complex hydroelectric relicensing process” (p. 10).
Based on the status of the species covered above, I chose to focus on
Skokomish coho activity for the preliminary LDI analysis conducted in the current
research. The rationale for selecting coho habitat for review was rooted in the idea
that this particular stock may be under less influence from study and recovery
efforts and could potentially provide a clearer assessment of LDI impact on
salmonid habitat and fish activity.
This is not to suggest that the current findings
show any substantial LDI impact on fish behavior but rather present an opportunity
for a future study with fewer confounding factors. If LDI impacts on salmonid
activity and habitat can be established, the resulting methodology could potentially
be used to assess populations of greater concern.
61
Data Availability, Accuracy/Precision, Integration, Processing and
Interpretation
Quantifying land-use disturbance within watersheds using GIS-software is
reliant on data availability and/or the opportunity to gather, synthesize, and assess
new data. In the case of the current research, the utilization of existing data was
imperative based on time and resource constraints.
Multiple challenges and
limitations were encountered in the process of utilizing data for the purposes of the
current research and are examined here.
Available Land-Use Datasets
Determining land-use is a challenging task in itself and quantifying
human-induced ecosystem disturbance imposed by land-use is reliant on the
land-use classification schema employed by the reporting agencies.
The GIS
land-use dataset utilized for the current research was the “2010 Statewide Landuse”
layer created by the Washington State Department of Ecology (DOE)
(http://www.ecy.wa.gov/services/gis/data/planningCadastre/landuse.htm).
The
GIS layer dataset was derived from digital county tax parcel layers as specified by
the Washington State Department of Revenue (DOR). The synthesis of data
produced by two separate agencies to create this dataset is telling of the
interdisciplinary nature of environmental study and the interconnectedness of
ecosystems and human interaction.
It is important to acknowledge the fact that
land-use classifications were based on an economic framework (tax parcel
62
purposes) and may not be the most accurate representation of land- use disturbance.
DOE staff attempted to classify parcels that did not contain DOR coding and were
randomly checked for accuracy. Additional effort was made during the current
research to validate land-use classification data and was based on my best
judgment.
There were instances in which I felt it necessary to assign different
LDI-coefficients to particular parcels based on orthographic photography or
observational ground-truthing. Based on the potential to utilize data derived from
the current study and future efforts, it is recommended that consultation with
experts from multiple fields be conducted when assigning land-use classifications.
DOE technicians echo this sentiment and the following disclaimer and
recommendation is offered in the official description of the dataset:
…“The land use coding is only as good as each county was able to provide
or as good as we were able to ascertain editing from orthophoto imagery overlay.
We welcome and encourage edits, updates and corrections from the GIS user
community”…(http://www.ecy.wa.gov/services/gis/data/planningCadastre/landus
e.htm).
Additional temporal consideration may also increase the validity of the
dataset. For example, a forester may be able to more accurately assign the LDI
coefficient for a parcel classified as “Commercial Forest” or “Undeveloped Land”
based on assessment of historical timber harvest practices. Furthermore, the
inclusion of road data unaccounted for in the existing dataset may also bear
63
additional scrutiny.
Experts in erosion effects caused by road construction and
subsequent sediment deposition in streams may also choose to reclassify parcels or
recommend different LDI coefficients be applied to certain data points.
Methodological Concerns Regarding “Area of Influence”
Brown and Vivas (2005) created the methodology for assessing LDI within
watersheds and their research in the state of Florida posited that a 100 meter buffer
was sufficient to determine LDI effects.
Brown and Vivas (2005) indicated no
significant difference between applying a 100 meter or 200 meter buffer when
quantifying land-use disturbance influence on watersheds.
Given the differences
in physical geography between watersheds in Florida and the Pacific Northwest,
additional study to determine the appropriate area of influence in Pacific Northwest
watersheds may be required. Considerations of slope-gradient and subsequent
hydrological influence as contributing factors in land-use disturbance should not be
overlooked.
It should also be noted that land-use disturbance tends to decrease
with elevation gain within the Skokomish River watershed as land adjacent to
stream inputs is increasingly unsuitable for development and headwater areas exist
within protected lands of Olympic National Park. However, timber extraction has
occurred and continues to occur in upstream segments adjacent to streams in areas
with slope gradients unsuitable for development but suitable for timber harvest
practices. Based on these factors, a sliding scale of increasing riparian buffer for
64
upstream segments at increased elevations when determining the area of influence
in Pacific Northwest watersheds may be required.
Integrating LDI with Salmonid Studies
The current research was only capable of examining LDI relationships with
salmonid habitat at a finite spatial scale and regarding only a single salmonid
species. However, the methodology utilized in the current research may be
adapted to conduct more focused research on individual salmonid stocks and
salmonid stream habitat.
Utilizing Existing Datasets
During the current research, LDI quantification was conducted in relationship to
coho salmon activity in the lower reaches of the Skokomish River.
Coho activity
data utilized consisted only of “documented” activity at particular stream reach
sites and is not indicative of coho activity at the watershed-scale. Furthermore, the
current salmonid habitat data utilized for analysis exhibited some deficiencies.
The current SalmonScape web interface provides users with an efficient method for
reviewing baseline data regarding salmonid activity. It is my opinion that a recent
reconfiguration of SalmonScape has made the interface more “user-friendly” for
interested citizens and non-experts but has done so at the expense of readily
providing more extensive and potentially powerful data. For example, under the
current configuration, users are only given the option to select for fish activity
65
based on species.
The resulting data displayed indicates documented activity of a
particular species at specified spatial scales. There is no indication of overlap
among documented activities thus, a particular stream reach may be characterized
as exhibiting “documented spawning” and another stream reach characterized as
exhibiting “documented rearing.”
In some cases these stream reaches are
connected but only one type of activity per species is assigned to a particular reach
making it difficult to determine the variety of manners in which salmonid species
are utilizing particular stream habitat. The previous web-based incarnation of
SalmonScape, while complicated to use, provided a more robust array of datasets
and users could review individual stream reaches for species presence and activity,
all which could be displayed separately.
I suspect that these data could be
obtained by contacting administrators for the SalmonScape site or the individual
agencies which contributed the data comprising the dataset but doing so was
beyond the scope of the current research given time constraints.
Future
researchers interested in conducting studies regarding LDI relationships with
salmonid activity habitats would be encouraged to seek out more robust datasets to
conduct analyses.
Potential for Comparative Studies
Collyard and Von Prause of the Washington Department Ecology indicated that a
statewide database of LDI for all Water Resource Inventory Area s (WRIA’s) could
prove extremely useful for a multitude of studies (personal communication, Spring
66
2014). With the establishment of a uniform method for classification and
reporting of LDI within watersheds, the influence of or relationship of LDI in
regard to other ecological factors could be examined.
A scenario in which this
analysis might prove useful would be a comparative study of LDI, salmonid
spawner abundance, and smolt abundance in several watersheds exhibiting varying
levels of suitable spawning and rearing habitat availability and currently classified
quality.
Through such study, it may be possible to quantify the relationship or
influence of particular land-use types of parcels adjacent to salmonid streams with
fish activity and fish population dynamics or specific habitat characteristics such as
water quality, turbidity, stream sediment composition, or forage availability.
Implications for Informing Policy Decisions
The simplicity of the manner in which LDI study results are reported could prove
beneficial in informing policy decisions.
Most individuals are able to grasp the
concept of a “one to ten” scale as a measurement of influence or impact without the
requirement of a depth of knowledge regarding the factors which comprise the
impact.
A simple understanding of land-use disturbance effects may aid in
prioritizing fish recovery and management efforts in watersheds, selecting areas for
habitat restoration, or in the approval process of development projects.
67
Chapter Six: Interdisciplinary Aspects and Conclusions
Researchers interested in studying environmental issues are faced with a
challenging decision regarding the manner in which to undertake their
examinations.
Environmental concerns can frequently be approached from
multiple disciplinary frameworks and research efforts should attempt to
acknowledge the complexity of factors which may be influencing a particular issue.
This is not to suggest that tightly focused studies within a specific discipline have
diminished value nor to suggest that interdisciplinary efforts are superior.
However, the value of interdisciplinary study should not be trivialized and framing
a particular study in interdisciplinary terms may improve accessibility to study
results and allow for new knowledge to be utilized across disciplines.
The
interdisciplinary aspects of the current research are presented in this chapter and are
followed by a summation of the current research.
Interdisciplinary Nature of Salmonid Studies
Referring back to Montgomery’s (2003) five “H’s” offers an initiation point
for considering salmon studies from a variety of disciplines and are indicative of
the interdisciplinarity and transdisciplinarity of the of the issues. The five H’s are
comprised of: a) harvest, b) habitat degradation, c) hydroelectric power, d) hatchery
issues, and e) history. In the particular case of the Skokomish River watershed, all
five “H’s” have influenced the fishery and offer potential study avenues.
68
The Intertwined Complexity of History, Habitat, Harvest, Hatcheries and
Hydroelectric Issues in the Hood Canal Region
Tribal Communities, Economies, and Land Use
Eells (1887) described extensively the dominant cultural traits of indigenous
populations that existed in the Hood Canal region.
Additionally, in 1960,
Elmendorf produced an extensive volume on the Twana culture which was updated
in 1992. Both indicate that the original inhabitants of the Hood Canal region
engaged in largely subsistence- foraging lifestyles, similar to other populations in
the Northwest (Eels, 1887; Elmendorf, 1992). These populations were therefore
engaged in minimal economic exchange with outside communities.
Elemendorf
(1992) cites an instance of a Hood Canal population “selling” a particular type of
clam to a main-body Puget Sound population, whose members eventually would
utilize it as a tradable good with inland communities east of the Cascades.
Hood Canal indigenous populations usually utilized two primary
community establishment-types: the winter village consisting of plank houses at a
single site and summer foraging camps, usually associated with fishing grounds
(Elmendorf, 1992). Trees were harvested for plank house building and for the
construction of canoes but no additional extensive timber excavation occurred
(Eells, 1887; Elemendorf, 1992).
69
Surplus resources were occasionally gathered but were largely used for
community feasting or potlatches involving gathering of multiple communities
(Eells, 1887), rather than being harvested and distributed for economic gain.
Thus, it could be concluded that these indigenous populations utilized natural
resources beyond a deterministic existence and therefore began to shape the
landscape in a manner beneficial to the subsistence and growth of populations.
However, the rate of resource utilization was minimal compared to the rapid
changes which would occur upon European settlement of the region.
Richard White (1980) produced an extensive history of what is currently
called Island County, Washington State. As White indicated, the land-use patterns
exhibited by indigenous populations on present-day Whidbey Island likely
mirrored or at least paralleled those of other indigenous populations in the
immediate Puget Sound region.
White’s assertions support a view of indigenous
populations as the original shapers of the landscape in the Puget Sound region,
though caution should be taken in generalizing too liberally from the populations
on Island County to those of Hood Canal. What is clear, are the rapid alterations in
human land-use patterns which occurred after European and Euro-American
settlement.
Pope and Talbot Establish Puget Mill Company at Port Gamble
Andrew Jackson Pope and Frederic Talbot were sons of prominent logging families
based out of East Machias, Maine. The two made their way to San Franscisco in
70
1849, and with a third partner, Captain J.P. Keller initiated a transport company.
In March of 1850, William C. Talbot, brother of Frederic (who had returned East)
arrived from Maine and joined in the business.
The California Gold Rush was
underway and the city of San Francisco required lumber for the construction of
buildings. A.J. Pope and W.C. Talbot had heard from captains who had visited the
Puget Sound area, of the vast stands of timber in close proximity to protected inland
waters. Pope and Talbot resolved to build a steam-operated sawmill on Puget
Sound.
In 1853, the W.C. Talbot and Co. launched their new venture, the Puget
Mill Company at present-day Port Gamble. The demand for timber was high with
the leading markets in San Francisco and the Far East, and the Puget Mill Company
quickly flourished. In addition to the mill operations, the company was possession
of a fleet of early shipping freighters.
The company shipped not only timber
milled at Port Gamble but at others throughout the Puget Sound region and
eventually usurped the Usalaty mill on Camano Island and the mill located in Port
Ludlow.
Additional mills were located at Union and Seabeck and the utilization
of timber from the Hood Canal area began in earnest and a striking reshaping of the
land was underway. The Puget Mill Company had holdings of up to 170,000 acres
at the height of the timber boom.
71
Effect of Expansion of Timber Harvest on Watersheds
Early logging practices were highly destructive and even wasteful, with
clear-cutting the standard mode of practice. Excavation sites were located either
near marine areas or within river basins to assist in transport. Marine areas and
river basins were often the site of major timber excavation operations because
water provided easy transport. Trees felled near marine areas could be deposited
into the saltwater then gathered for transport to the mill.
assisted in timber harvest further inland.
Flowing water also
Trees could be fallen and then floated
downriver to the marine gathering site. Sedell and Luchessa (1981) indicated that
trees had been cleared 2 miles inland along Western Puget Sound and Hood Canal
shorelines and as many as 7 miles inland near streams and rivers.
Technological advancements allowed for increasing efficiency and
expansion of timber harvest practices. The advent of the automobile and
associated mechanized equipment meant the opening of the forests even further
inland and at higher elevations. No longer were timber operations limited to close
proximity to water. Road systems were developed to facilitate access of the
equipment to new logging sites. By the mid-20th century timber excavation had had
a tremendous impact on the ecosystems of the forests and the watersheds bore a
major brunt of the impact. Roads were often built without culverts or with
culverts unsuitable for migratory fish passage. The removal of streamside timber
in the riparian zone caused water temperatures to rise during the warm summer
72
months which was at times lethal for salmonids which require cooler water
temperatures for survival. Mass excavation of trees also had a major impact on the
soil systems as the root systems provide stability, especially on steep hillsides.
The newly loosened soil could easily wash off hillsides during major precipitation
events and cause a dramatic input of sediment into the river systems.
Analyzing Historical Timber Harvest-Economic Drivers and Lack of
Ecosystem Understanding Drive Land Alteration and Degradation
It is easy to take an environmental deterministic view when examining the
historical alterations of much of the Puget Sound region’s landscape. An
environmental deterministic view posits that the surrounding environment dictates
the behavior of its human inhabitants. Following this view, one would believe that
the harvest of timber was a product of necessity to adapt to the region. In doing so
however, one would be remiss in failing to recognize that it was frequently the
machinations of the economy that drove the alteration post-European settlement as
well as the activities of the indigenous populations prior to that.
As Richard White indicated, “In the northern conifer forests generally,
burning has for centuries shaped woodland ecology” and “changes such as these
were not readily apparent to the casual observer. Unless the environment bore
obvious marks of human handiwork, the first whites dismissed it as wilderness,
natural and untouched” (White, 1980, p. 25). It is therefore not unreasonable to
73
conclude that the first Europeans saw a vast, inexhaustible resource in timber. At
the time of the initial boom, the philosophies of the likes of John Muir and Gifford
Pinchot were still in their infancies and as Richard White states, “waste had little
economic meaning” (p. 89). Shipped lumber however, did have tremendous
economic meaning and wasteful practices continued as White again indicated
“waste did have environmental consequences” (p. 89). The reconciliation of the
economic gain versus the environmental consequences was decades away.
Furthermore, it is important to recognize that the early utilization of timber by
Euro-Americans was undertaken as a means to benefit individuals far from the site
of the excavation.
San Francisco was a burgeoning city at this time and much of
the lumber harvested from Washington State built that city 800 miles to the south.
Additional markets existed in Asia, even further removed from the source of the
raw materials.
Salmon Harvest in Western Washington and Hood Canal
American and indigenous tribal commercial fisheries have existed in the Puget
Sound region for more than a century and provided economic gain for those
engaged in the fisheries.
Additionally, recreational fishing opportunities serve as
an economic driver for the region. Estimates of the economic value of the
combined salmon fisheries on Hood Canal ranged from $500,000 to $2.2 million
between 1978 and 1988 (Washington Department of Fisheries archival data).
74
Estimated salmon catch values in 2006 were $3.77 million for the entire South
Puget Sound region which includes Hood Canal (TCW Economics, 2008).
Throughout Washington State in 2006, recreational angling fisheries for salmon
contributed a $128.4 million in net economic gain.
Currently, the Skokomish Tribe operates fisheries on the Skokomish River
for commercial, subsistence, and ceremonial purposes and are primarily centered
around chinook and chum salmon.
Recreational fisheries for both chinook and
chum on the Skokomish are also popular among anglers.
It is difficult to assess
the historical and current impacts these fisheries have had on abundances, but the
economic and intrinsic values of these fisheries for individuals utilizing them
should be taken into consideration when studying the region.
Hatcheries
The George Adams hatchery is operated on the Skokomish River and contributes to
supplementary efforts for restoring salmon and steelhead runs. The impact of
hatcheries and the fish produced at these facilities has been the focus of large
bodies of research and results show varying degrees of influence from these
programs. Currently, the George Adams hatchery is operated under stipulations of
the Hatchery Genetics Management Plans (HGMPS) for ESA-listed Puget Sound
Chinook and Hood Canal Summer Chum.
Multiple state, federal, tribal, and
non-governmental organizations are involved with projects that originate from this
75
hatchery facility and are representative of the various interest groups concerned
with Hood Canal and Skokomish River fisheries.
Hydroelectric Power
Two hydroelectric dams exist in the Skokomish River watershed: Cushman Dam
No. 1 and Cushman Dam No. 2. The dams were erected by Tacoma Power to
provide additional electricity to the City of Tacoma. The dams exist in the mid to
upper reaches of Skokomish watershed and were presumed to have blocked
anadromous fish passage as well as disrupted spawning and rearing habitat for
anadromous and non-anadromous fish.
According to the Tacoma Public Utilities
web page for the Cushman dams, “On Jan. 12, 2009, Tacoma Power, the
Skokomish Tribal Nation and state and federal agencies signed a settlement
agreement that resolved a $5.8 billion damages claim and long-standing disputes
over the terms of a long-term license for Cushman Hydroelectric Project”
(https://www.mytpu.org/tacomapower/about-tacoma-power/dams-power-sources/
hydro-power/cushman-hydro-project/).
Under the new licensure agreements,
multiple measures were put in place to place responsibility on the utility provider to
address and participate in fish restoration issues (Federal Energy Regulatory
Commission, 2010).
76
Summary of Interdisciplinary Aspects of the Current Study and Conclusions
The interconnectivity of issues surrounding salmonid abundance in the
Pacific Northwest is illustrated in brief by the review of the topics discussed in
Chapter Six. When examining the Skokomish River, these connections may
become apparent when delving into any particular aspect. For example, the
current study attempted to quantify land-use disturbance effects within the riparian
zone of the watershed.
A primary motivation for this research was my desire as a
researcher to understand how land-use may affect salmonid stream habitat.
This
motivation stemmed from my experience with a salmon enhancement group which
was involved in a collaborative effort to study and potentially restore diminished
stocks of particular salmonid species. Organizations involved in the collaborative
efforts of the salmonid studies included federal, state, and tribal agencies, as well as
independent non-governmental organizations.
The greater motivations for the
involvement of these organizations may include economic, biological, ecological,
cultural, or climatological concerns among many others.
The data I utilized in the current study included land-use classifications
based on a schema utilized for tax assessment purposes and collected by a
governmental organization concerned with state revenue. The individual land-use
classifications represent the various ways in which humans interact with and utilize
landscapes. Reviewing the history of the region indicates the ways in which
human populations have interacted with the land over time. Analyzing the
77
geography of the region shows how existing landscapes may have driven the
activities of individuals residing here and how previous and current activities
continue to alter the landscape and ecosystems.
In summary, the current research assessing the quantification of land-use
disturbances in the Skokomish watershed required an interdisciplinary approach in
the development, application, and analysis phases of the project. Based on my
experience during this research, and in conjunction with the education I received
during my involvement with the Graduate Program on the Environment at The
Evergreen State College, it is anticipated that research efforts guided by an
interdisciplinary approach will become more common and yield increasingly
robust results. It is hoped that results derived from such interdisciplinary research
projects may aid in improving awareness of environmental issues, help to frame
these issues in multiple contexts, and inform policy decisions regarding habitat and
species conservation, the management of ecosystems, and developmental planning.
78
APPENDIX
Land-use CD number
(Ecology, 2010)
Nonrenewable
empower density
(E14 sej/ha/yr)
91,92,93,94,95,96,
97,98,99
0
Natural open
water
93
0
Pine plantation
88,95
5
71,72,73,74,75,77,78,
79
7
81,82,83
8
81,82,83
17
81,82,83
33
Orchard
81,82,83
44
3.68
Improved pasture
– high-intensity
(with livestock)
81,82,83
47
3.74
Row crops
81,82,83
107
72,73,74,75,77,78,79
1077
11,18,19
1230
81,82,83
1349
7.00
11,18,19
2175
7.47
Land-use
Natural system
Recreational /
open space –
low-intensity
Woodland
pasture (with
livestock)
Improved pasture
(without
livestock)
Improved pasture
– low-intensity
(with livestock)
Recreational /
open space –
high-intensity
Single family
residential –
low-density
Agriculture –
high intensity
Single family
residential –
medium density)
LDI
coefficients
1.00
1.00
1.58
1.83
2.02
2.77
3.41
4.54
6.90
6.92
79
Single family
residential – high
density
Mobile home
(medium density)
7.55
11,18,19
2372
15
2748
7.70
42,45
3080
7.81
50,51,52,53,54,55,56,
57,58,5960,6162,63,6
4,65,66,67,68,69
3758
8.00
67,68,69
4042
42,45
5020
15
5087
20,21,22,23,24,25,26,
27,28,29,30,31,32,33,
34,35,36,37,38,39,82,
85
5211
8.32
Multi-family
residential (low
rise)
12,13,14,15,16,17
7392
8.66
High-intensity
commercial 12
50,51,52,53,54,55,56,
57,58,5960,6162,63,6
4,65,66,67,68,69
661
9.18
12,13,14,15,16,17
12825
9.19
16150
9.42
Highway (2 lane)
Low-intensity
commercial
Institutional
Highway (4 lane)
Mobile home
(high density)
Industrial
Multi-family
residential (high
rise)
Central business
district (average
2 stories)
Central business
district (average
4 stories)
50,51,52,53,54,55,56,
57,58,5960,6162,63,6
4,65,66,67,68,69
50,51,52,53,54,55,56,
57,58,5960,6162,63,6
4,65,66,67,68,69
29401
8.07
8.28
8.29
10.00
80
REFERENCES
Allan, J.D. (2004). Landscapes and riverscapes: The influence of land-use on
stream ecosystems. Annual Review of Ecology, Evolution, and Systematics,
35. (257-284).
Bjornn, T.C., and Reiser, D.W. 1991. Habitat requirements of salmonids in
streams. In Influences of forest and rangeland management on salmonid
fishes and their habitats. In W.R. Meehan., (Ed.) American Fisheries
Society Special Publication No. 19. Bethesda, Maryland. pp. 83-138.
Brown, M.T. & Vivas, B. (2005). Landscape development intensity index.
Environmental Monitoring and Assessment 101. 289-309.
Chiang, C. & Reese, M. (2003.) Evergreen State: Exploring the history of
Washington’s Forests. Seattle, WA: Center for the Study of the Pacific
Northwest, Department of History, University of Washington. Retrieved
from:
http://www.washington.edu/uwired/outreach/cspn/curforest/curformain.ht
ml
Clarke, W.G., & Hirano, T. (1995). Osmoregulation. In Groot, C., Margolis, L., &
Clarke, W.C. (Eds.). Physiological Ecology of Pacific Salmon (pp.
317-378). Vancouver, British Columbia: University of British Columbia
Press
Collyard, S. and M. Von Prause., (2010). Deschutes River multi-parameter total
maximum daily load effectiveness monitoring pilot project: Quality
Assurance Project Plan. Washington State Department of Ecology,
Olympia, WA. Publication No. 09-03-133. Retrieved from
www.ecy.wa.gov/biblio/0903133.ht
81
Collyard, S. and M. Von Prause., (2009). GIS procedure for creating an LDI index.
Unpublished document obtained through personal communication.
Committee On Protection and Management of Pacific Northwest Anadromous
Salmonids, Board on Environmental Studies and Toxicology, &
Commission on Life Sciences. (1996). Upstream: Salmon and society in the
Pacific Northwest (95-51662). Washington D.C.: National Academy Press
Correa, G. (2003). Salmon and steelhead habitat limiting factors: Water Resource
Inventory Area 16 Dosewallips-Skokomish basin. Final Report.
Washington State Conservation Commission. Retrieved from:
http://www.pugetsoundnearshore.org/supporting_documents/wria16_salm
onid_habitat_lfa_report.pdf
Creswell, J. W., & Creswell, J. W. (2013). Qualitative inquiry and research design:
Choosing among five approaches. Los Angeles: SAGE
Dunagan, C. (1991). Rising from toppled trees. In Hood Canal: Splendor at risk.
Bremerton, WA: The Sun Newspaper.
Eels, Rev. M. (1889). The Twana, Chemakum, and Klallam Indians, of
Washington Territory. In Annual Report of the Board of Regents of the
Smithsonian Institution. (pp. 605-679). Retrieved from
http://books.google.com/books?id=DVjvH99zHP8C&pg=PA605&lpg=P
A605&dq=rev+myron+eells+the+twana,+chemakum+and+klallam+indian
s+of+Washington+territory&source=bl&ots=Qc2mX_WXDc&sig=ys8Db
vAqWeVK08EKqv58tPyG1HY&hl=en&sa=X&ei=YjWlUseoCtjkoAT8
mIGwDw&ved=0CDIQ6AEwAg#v=onepage&q=rev%20myron%20eells
%20the%20twana%2C%20chemakum%20and%20klallam%20 indians%2
0of%20Washington%20territory&f=false
82
Elmendorf, W.W. (1992). The Structure of Twana Culture. Pullman, WA:
Washington State University Press.
Federal Energy Regulatory Commission. (2009). Settlement Agreement for the
Cushman Project. Retrieved from:
https://www.mytpu.org/file_viewer.aspx?id=712
Frimpong, E.A., Sutton, T.M., Lim, K.J., Hrodey, P.J., Engel, B.A., Simon, T.P.,
Lee, J.G., Le Master, D.C. (2005). A Determination of optimal riparian
forest buffer dimensions for stream biota-landscape association. Canadian
Journal of Fisheries and Aquatic Sciences, 62, 1–6.
Hicks, B. J. (2002). Gravel galore: Impacts of clear-cut logging on salmon and their
habitats. In B. Harvey and M. MacDuffee (Eds.), Ghost runs: The future of
wild salmon on north and central coasts of British Columbia. (pp. 97-118).
Victoria, British Columbia: Raincoast Conservation Society.
Higgs, D.A., Macdonald, J.S., Levings, C.D., & Dosanjh, B.S. (1995). Nutrition
and feeding habits in relation to life history stage. In Groot, C., Margolis,
L., & Clarke, W.C. (Eds.). Physiological Ecology of Pacific Salmon (pp.
159-316). Vancouver, British Columbia: University of British Columbia
Press
Hood Canal Coordinating Council. (2005). Salmon Habitat Recovery Strategy for
the Hood Canal & the Eastern Strait of Juan de Fuca. Retrieved from:
hccc.wa.gov/Downloads/Downloads_GetFile.aspx?id=116447&fd=1
Hood Canal Winter Steelhead Supplementation Hatchery Genetics Management
Plan (HGMP). (2012). Retrieved from:
http://wdfw.wa.gov/hatcheries/hgmp/pdf/puget_sound/hood_canal_winter
_steelhead_hgmp_apr2013.pdf
83
Kitsap Sun. (1991). Hood Canal: Splendor at risk. Bremerton, WA: The Sun
Newspaper.
Kuttel, Jr., M. (2003). Landscape assessment and conservation prioritization of
freshwater and nearshore salmonid habitat in Kitsap County. Retrieved
from Washington State Conservation Commission:
www-stage.scc.wa.gov/wp-content/.../wria-15-East-LFA-report_web.pdf
Lele, S., & Norgaard, R. B. (2005). Practicing interdisciplinarity. BioScience,
55(11). 967-975.
Lichatowich, J. (1999). Salmon with rivers: A history of the pacific salmon crisis.
Washington, D.C.: Island Press.
Lombard, J. (2006). Saving Puget Sound: A conservation strategy for the 21st
century. Bethesda, MD: American Fisheries Society
Lyons, J.K. & Beschta, R.L. (1983). Land-use, floods, and channel changes: Upper
Middle Fork Willamette River, Oregon (1936-1980). Water Resources
Research, 19. 463-471.
Mason County. (2013): Retrieved from:
http://mapmason.co.mason.wa.us/website/mason/viewer.htm.
Mason County Shoreline Master Program Update: Inventory and Characterization
Report (2012). SMA Grant Agreement No. G1100004. Retrieved from:
http://www.co.mason.wa.us/community_dev/shoreline_master_program/s
mp_update_revised_october_2012.pdf
Matson, R.G., & Coupland, G. (1994). The prehistory of the northwest coast. Lost
Coast Press.
84
Montgomery, D. (2003). King of fish: The thousand-year run of salmon.
Cambridge, MA: Westview Press.
Mull, K.E. (2005). Selection of spawning sites by coho salmon (Onchorynchus
Kisutch) in Freshwater Creek, California (Unpublished master’s thesis).
Humboldt State University
National Marine Fisheries Service. (2010). 5-Year Review: Summary & Evaluation
of Puget Sound Chinook Hood Canal Summer Chum Puget Sound
Steelhead. Retrieved from:
http://www.nmfs.noaa.gov/pr/pdfs/species/pugetsound_salmonids_5yearre
view.pdf
Pess, G.R., Montgomery, D.R., Steel, E.A., Bilby, R.E., et al. (2002). Landscape
characteristics, land-use, and coho salmon (Oncorhynchus kisutch)
abundance, Snohomish River, Wash., USA. Canadian Journal of Fisheries
and Aquatic Sciences; 59, (4). 613
Platts, W.S. (1991). Livestock grazing. In Meehan, W.R. (Ed.). Influences of forest
and rangeland management on salmonid fishes and their habitats.
(pp.389-423). American Fisheries Society Special Publication 19.
Bethesda, MD: American Fisheries Society.
Robbins, W.G. (1997). Landscapes of promise: The Oregon story, 1800-1940.
Seattle, WA: University of Washington Press
Sedell, J.R. & Luchessa, K.J. (1981). Using the historical record as an aid to
salmonid habitat enhancement. In Armantrout, A.B (ed.). Symposium on
Acquisition and Utilization of Aquatic Habitat Inventory Information.
Portland, OR: American Fisheries Society, Western Division
85
Stanfield, L.W. & Kilgour, B.W. (2012). How proximity of land-use affects stream
fish and habitat. River Research and Applications. doi: 10.1002/rra.2585
Tacoma Public Utilities. (n.d.) Retrieved from:
https://www.mytpu.org/tacomapower/about-tacoma-power/dams-power-so
urces/hydro-power/cushman- hydro-project/
Taylor, III, J.E. (1999). Making salmon: An environmental history of the northwest
fisheries crisis. Seattle, WA: University of Washington Press
TCW Economics. (2008). Economic analysis of the non-treaty commercial and
recreational fisheries in Washington State. December 2008. Sacramento,
CA. With technical assistance from The Research Group, Corvallis, OR.
State. December 2008. Sacramento, CA. With technical assistance from
The Research Group, Corvallis, OR.
Tschaplinski, P. (1987). The use of estuaries as rearing habitats by juvenile coho
salmon. In Chamberlin, T.W. (Ed.). Procedures of the workshop: Applying
15 years of Carnation Creek results. Pacific Biological Station, Nanaimo
BC.
Washington Department of Ecology/Washington Department of Revenue. (2010).
State Land Use Parcel Classifications (ArcGIS layer). Retrieved from:
http://www.ecy.wa.gov/services/gis/data/planningCadastre/landuse.htm#1
Washington Department of Fish and Wildlife’s SalmonScape Web Application.
Retrieved from: http://apps.wdfw.wa.gov/salmonscape/map.html#
Wells, J. (2013). Complexity and Sustainability. New York: Routledge.
White, R. (1980). Land Use, Environment, and Social Change: The Shaping of
Island County, Washington. Seattle, WA. University of Washington Press.
86
87