Restoration of Coastal Beach Forming Ecosystem Processes through Shoreline Armoring Removal of a Former Mine Site Increases Our Understanding of Coastal Resiliency and Large Scale Landslides Along the Northeast Pacific Coastline

Abstract

Coastal zones are important components of marine ecosystems that link upland and marine areas, and are often maintained by landslides, though these habitat forming processes are poorly understood. They are also often sites of commercial development, including mining, and restoration. In this ecosystem restoration project, we rapidly removed a large volume of shoreline armoring from the perimeter of an intertidal earthen filled mine structure. The remaining feature mimicked the toe of a large land slide characteristic of the region. Physical and ecological monitoring allowed us to understand how coastal ecosystems respond to large scale coastal sediment processes, including landslide function that forms them, and restoration actions to restore them. As a course of restoration, over 22,936 m³ of non-native armor totaling 0.47 hectares was rapidly removed from the perimeter of a 2.8 hectare earthen pier. Three years after removal, the feature is less than half its original configuration (area). Beach composition and ecological community of the restoration area responded positively relative to pre-project conditions and to the control site. Applying these results to historic deep seated landslides that define coastal zones of the region, we conclude that nearshore ecosystems are resilient and respond similarly to episodic natural and restored large-scale hydrodynamic processes.

Full text

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/349524545
Restoration of Coastal Beach Forming Ecosystem Processes through Shoreline
Armoring Removal of a Former Mine Site Increases Our Understanding of
Coastal Resiliency and Large Scale...
ArticleinInternational Journal of Mining, Reclamation and Environment · February 2021
DOI: 10.1080/17480930.2021.1872149
CITATIONS
0
READS
44
8 authors, including:
Some of the authors of this publication are also working on these related projects:
Elwha River Dam Removal Monitoring View project
Anne Shaffer
Coastal Watershed Institute
46 PUBLICATIONS340 CITATIONS
SEE PROFILE
Bob Oxborrow
University of Washington Seattle
1 PUBLICATION0 CITATIONS
SEE PROFILE
All content following this page was uploaded by Anne Shaffer on 30 May 2023.
The user has requested enhancement of the downloaded file.

Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=nsme20
International Journal of Mining, Reclamation and
Environment
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/nsme20
Restoration of Coastal Beach Forming Ecosystem
Processes through Shoreline Armoring Removal of
a Former Mine Site Increases Our Understanding
of Coastal Resiliency and Large Scale Landslides
Along the Northeast Pacific Coastline
Anne Shaffer , Dave Parks , Jamie Michel , Kirsten Simonsen , Katrina
Campbell , Bob Oxborrow , Jonathan Hall & Jennifer Weslowski
To cite this article: Anne Shaffer , Dave Parks , Jamie Michel , Kirsten Simonsen , Katrina
Campbell , Bob Oxborrow , Jonathan Hall & Jennifer Weslowski (2021): Restoration of Coastal
Beach Forming Ecosystem Processes through Shoreline Armoring Removal of a Former Mine
Site Increases Our Understanding of Coastal Resiliency and Large Scale Landslides Along the
Northeast Pacific Coastline, International Journal of Mining, Reclamation and Environment, DOI:
10.1080/17480930.2021.1872149
To link to this article: https://doi.org/10.1080/17480930.2021.1872149
© 2021 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Group.
Published online: 22 Feb 2021.
Submit your article to this journal
View related articles
View Crossmark data

Restoration of Coastal Beach Forming Ecosystem Processes
through Shoreline Armoring Removal of a Former Mine Site
Increases Our Understanding of Coastal Resiliency and Large
Scale Landslides Along the Northeast Pacic Coastline
Anne Shaer
a
, Dave Parks
a
, Jamie Michel
a
, Kirsten Simonsen
a
, Katrina Campbell
a
,
Bob Oxborrow
b
, Jonathan Hall
c
and Jennifer Weslowski
c
a
Coastal Watershed Institute, Port Angeles, Washington, USA;
b
University of Washington School of Aquatic and
Fishery Sciences, Seattle, Washington, USA;
c
Lafarge Holcim Seattle, Washington, USA
ABSTRACT
Coastal zones are important components of marine ecosystems that link
upland and marine areas, and are often maintained by landslides, though
these habitat forming processes are poorly understood. They are also
often sites of commercial development, including mining, and restoration.
In this ecosystem restoration project, we rapidly removed a large volume
of shoreline armoring from the perimeter of an intertidal earthen lled
mine structure. The remaining feature mimicked the toe of a large land
slide characteristic of the region. Physical and ecological monitoring
allowed us to understand how coastal ecosystems respond to large
scale coastal sediment processes, including landslide function that forms
them, and restoration actions to restore them. As a course of restoration,
over 22,936 m
3
of non-native armor totaling 0.47 hectares was rapidly
removed from the perimeter of a 2.8 hectare earthen pier. Three years
after removal, the feature is less than half its original conguration (area).
Beach composition and ecological community of the restoration area
responded positively relative to pre-project conditions and to the control
site. Applying these results to historic deep seated landslides that dene
coastal zones of the region, we conclude that nearshore ecosystems are
resilient and respond similarly to episodic natural and restored large-scale
hydrodynamic processes.
ARTICLE HISTORY
Received 9 May 2020
Accepted 3 January 2021
KEYWORDS
Nearshore; ecosystem;
coastal landslides; resilience;
hydrodynamics
1. Background and introduction
Shoreline fill and armoring are common features of coastal mine operations. Fill is placed along the
shoreline to create piers that allow easy access to barges. Shoreline armoring, in the form of large
boulders, sheet pile, and cement are often placed in very large quantities around these structures to
protect against marine erosion and provide structural integrity. As a result, shoreline fill and
armoring also constitute some of the most impactful nearshore ecosystem impediments of coastal
systems, primarily through the disruption of physical and ecological ecosystem processes [1].
Shorelines where these materials are placed are often characterised by large deep-seated landslides
(defined as landslides that have failure planes at least several metres below the ground surface and
involve sediment volumes on the order of thousands to millions of cubic metres) that result in
features that are of cultural importance, but their formation and ecological mechanics are poorly
CONTACT Anne Shaffer anne.shaffer@coastalwatershedinstitute.org Coastal Watershed Institute, Port Angeles,
Washington 98362, USA
INTERNATIONAL JOURNAL OF MINING, RECLAMATION AND ENVIRONMENT
https://doi.org/10.1080/17480930.2021.1872149
© 2021 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://
creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the
original work is properly cited, and is not altered, transformed, or built upon in any way.

understood. The Twins coastline, located on the north-central Strait of Juan de Fuca, Washington
state (Figure 1) is some of the most complex and diverse shorelines of the Salish Sea [2]. The area is
characterised by weak and easily erodible marine clay-silt-sandstone that results in large deep-
seated landslides that deliver significant amounts of sediment to the shoreline. These are evidenced
by the large hemispherical rock ‘ring’ features along the shallow subtidal and intertidal (Figure 2)
[3]. These are documented to be important cultural features, including historic fish weirs, for the
Tribes of the region [4–6]. The detail on the evolution of these coastal sites, including erosional rates
associated with their formation, is very poorly understood and has not been studied in the field.
The area is also the site of one of the largest shoreline restoration projects for the Salish Sea
(Figure 2). The goal of the Twins nearshore ecosystem restoration project is to restore a shoreline
significantly impacted by a large filled pier structure to a functioning intact ecosystem. The
restoration project consisted of removing non-native armoring, including riprap, sheet-pile, and
creosoted timbers from the perimeter of a 2.8 hectare earthen filled pier (hereafter called ‘the mole’)
from a former clay mine site and allowing clean native fill material to erode naturally to the central
Strait of Juan de Fuca shoreline. Details can be found in background materials [7–9].
The project resulted in approximately 22,937 m
3
of non-native material removed from the
nearshore. This included 17,585 m
3
of riprap, and 9175 m
3
of mixed soil and riprap, a combined
volume of approximately 5,352 m
3
of concrete, creosote-treated wood and sheet pile (Figure 2).
Immediately after the one-month long First phase of deconstruction, the mole footprint was
approximately 29%, smaller by area, and 31% smaller by volume relative to pre-project conditions
due to the significant portion of material moved to an upland disposal site, and a smaller proportion
transported by littoral processes (Figure 3a–c and Table 4). A smaller second phase event occurred
Figure 1. Twins nearshore ecosystem restoration site location (a) and comparative area Cline Spit (b).
2A. SHAFFER ET AL.

during 1 week in summer 2019 to remove additional remaining non-native material revealed
through erosion after the first phase of restoration.
The ecological response to drift cell scale restoration of physical sediment hydrodynamics was
defined by quantifying the change in standard coastal ecological metrics: beach profiles, erosion
rates of mole feature, and ecological metrics of Large Woody Debris (LWD) amount and size, beach
wrack sediment grain size, community composition and diversity indices, and surf smelt egg
distribution. Large woody debris is important for beach formation. Beach wrack (also known as
drift algae) is a critical food source and the basis for higher trophic systems including migrating
birds and fish. Similarly, surf smelt are one of a guild of fish known as ‘forage fish’ which are
a critical component to the Salish sea ecosystem and are prey species for Chinook and coho salmon
depended on by southern resident killer whales that are the focus of significant conservation
management and restoration efforts. Surf smelt have a very specific grain size requirement for
spawning, making them an extremely useful indicator of habitat suitability and change [10]. We
quantitatively documented the response of these metrics along both sides of the restored site before
and after site restoration and relative to pre-project and a comparative site. We then apply these
Figure 2. Looking west from Twins mole to large deep-seated landslide deposits, August 2018 (photo by Dave Parks and CWI).
Figure 3. Approximate transect locations for large woody debris (red) and beach wrack (yellow) surveys. Forage fish samples were
collected within approximate bounds of vertical lines (green).
INTERNATIONAL JOURNAL OF MINING, RECLAMATION AND ENVIRONMENT 3

results to an inventory of large-scale landslides characteristic of the region to provide important
insight into how large-scale landslides influence intertidal nearshore morphology.
2. Methods and materials
2.1. Restoration action
The majority of riprap, sheet pile, and large boulder armoring around the earthen fill pier structure
was removed from the perimeter of the structure over a month-long period using up to six
excavators and large off-road dump trucks that relayed the material to an upland stockpile site. It
was a large-scale restoration project.
2.2. Physical process
Beach profiles and mole size and position were assessed quarterly, beginning with pre-project
baseline. A backpack-mounted Arrow Gold Real Time Kinematic Global Positioning System (RTK-
GPS) was used in fixed mode, connected to the Washington State Reference Network (WSRN) to
collect horizontal and vertical positions of the surface of the Twin Rivers mole structure. These data
were then processed using computer software [11] to construct an interpolated digital elevation
model (DEM) of the surface topography of the site. Estimated horizontal and vertical accuracy of
these measurements is estimated to be 5 cm-10 cm (2–4 inches). Some variability in the data and
resulting DEM are due to variations in the location of survey points between surveys and the
differences in the quality of the GPS signal due to the number of available satellites and quality of
cellular phone connection.
By differencing model surfaces using GlobalMapper software, an estimate of the volume of
sediment lost from erosion over the time between sampling dates can be derived. Table 1 shows the
sampling dates and estimates of the Mole area and estimated volume of sediment comprising the
mole fill.
The area (hectares) and volume (cubic metres) of the mole were measured quarterly and
opportunistically following storm events using a fixed-course utilising an Arrow Gold real-time
kinematic global positioning system (RTK-GPS) connected to the Washington Reference Network.
Measurements of latitude, longitude and elevation were collected in the NAD 83 and M.S.L. datums
and projected using the Washington North, State Plane Coordinate system. Data were collected by
traversing the mole surface using a 10-metre grid. The resulting data were then input into Global
Mapper software to generate a digital elevation model of the mole topography. Global Mapper
software is then used to compare the area and volume of successive digital elevation models over
time to estimate area and volume loss from erosion. Horizontal and vertical root-mean-square
(RMS) accuracy is approximately 10 cm (3.93 inches).
Table 1. Parameters for sediment categories used to determine per cent coverage as
defined by Altizio, NOAA, 2010.
Category of sediment Size in mm
Cobble 64–256
Pebble 4–64
Granule 2–4
Sand 0.0625–2
Silt 0.002–0.625
Clay < 0.002
4A. SHAFFER ET AL.

2.3. Ecological monitoring
Three metrics were used to define the ecosystem function of the Twins nearshore: Surf smelt egg
distribution and abundance, beach wrack (per cent composition and invertebrate abundance), and
LWD composition were assessed pre and post-project using standardised techniques [10,12–15].
Sampling for these metrics occurs once/twice a year during July 2017- 2019 for LWD and surf smelt
spawning and 2017 and 2020 for beach wrack metrics at the restoration site and a comparative site
(Cline Spit, Dungeness Bay Washington). Sediment data included percent composition of cobble,
pebble, gravel, sand, and silt/clay from the beach surface (top layer), 5 cm depth, and a total
estimate. Data were analysed relative to pre-project (2017) data annually to quantify the ecosystem
response in a Before After, Control Impact (BACI) experimental design. We hypothesised that, as
a result of the project, beach wrack, LWD, and forage fish spawning values would change along the
two project site beaches, and trend towards becoming more similar as the mole erodes. We
predicted beach wrack and LWD metrics would respond at the restoration site immediately.
Specifically, we predicted that changes in the ecological metrics would occur within each of the
two reaches of the restoration site, with an increase in metrics relative to pre-project. Further, we
hypothesised the two reaches of the restoration site would become more similar as the two beaches
become more connected and uniform relative to pre-project. We predicted no significant con-
comitant changes at the comparative site over the same time period.
2.4. Large woody debris (LWD) composition
LWD surveys were conducted in three 18 metre long transects on either side of the mole, running
parallel to the beach length and extending from the bank to the waterline [15]. A total of three
transects each east and west of the mole were sampled for LWD in June and July 2017–2019.
Transects were measured at the approximate coordinates below, from west to east, in July each year
(Figure 2). All logs over 0.5-metre in length and 0.1-metre in diameter were measured for length
and diameter and recorded to the nearest 0.1-m. For each log measured, location on beach (high-
intertidal, mid-intertidal, or low-intertidal), stability (root wad, unstable, pinned, or buried), species
(deciduous or coniferous), relative orientation, composition (root wad or straight log) was also
recorded [15]. A per cent cover for pieces of wood under measurement criteria is also estimated.
2.5. Beach wrack characterisation
Beach wrack surveys were conducted in two 50-metre transects at the most recent location of wrack
deposition (last high tide) on either side of the mole [12,13]. Transects sampled in July in 2017 and
2020 (Figure 3). Ten random locations along each transect were chosen using a random number
generator, and a 0.9-m
2
quadrat was placed at each numbered (in metres) location. Each quadrat
was measured for per cent composition of algae, eelgrass, terrestrial plant material, and human
debris. A 15-cm benthic corer was then used to collect wrack and the top 2.5 cm of sediment at
a random 5 of the 10 quadrat locations. The collected samples were then frozen and later preserved
in formalin until analysis at School of Aquatic and Fishery Sciences at the University of
Washington. Beach sediment characterisation is also collected using these same transects, and
0.9-m
2
quadrat, at five random locations. Sediment is characterised by approximate grain size
(Table 1).
To test hypothesises, beach wrack invertebrate metrics, including total invertebrate count,
species count, and Shannon–Weiner diversity index were compared among Twins East, Twins
West, and the control site, Cline Spit. The Shannon–Weiner diversity index was created using the
vegan package (v 2.5–5) in R to examine species diversity. A generalised linear model (GLM) was
used to test for differences in each metric among sites and between years (2017 and 2020). The nine
most commonly collected taxa groups were also compared among sites and between years.
INTERNATIONAL JOURNAL OF MINING, RECLAMATION AND ENVIRONMENT 5

2.6. Surf smelt egg abundance
The shorelines east and west of the mole were sampled for surf smelt eggs a minimum of once
during 1 month of the spawning season (May through July) from 2017–2019 using methods
developed by WDFW [10]. A minimum of eight samples were collected from along approximately
1.6 km of shoreline for the east and west sites. The sample area covers a length spanning 527 metres
east and 311 metres west of the mole removal project (Figure 3). Bulk samples of beach sand/gravel
mix were collected in the upper intertidal (approximately +2.1 m to +2.7 m MLLW) zone of the
beach during June and July [10]. These bulk samples were collected using an approximately evenly
spaced collection of 8–10 samples on either side of the mole. Samples were then processed via
filtering the sediments through a series of 4, 2, and 0.5 mm screens, in order to obtain a grain size
that is most likely to obtain surf smelt eggs. Samples were then processed via the vortex method,
which sorts the lighter particles (eggs) and creates a sample that is a manageable size for further
analysis [14]. The samples were then preserved with Stockard’s Solution and sent to a private
accredited laboratory for inspection under a dissecting microscope.
3. Results
3.1. Physical metrics
In August 2017, right after project completion, the mole structure area was approximately 2.1 hec-
tares. As of mid-November 2019, after additional site de-compaction, wind and wave action
reduced the structure to 1.0 hectares, however, sediment from the north end of the mole was
redistributed to the southeastern side of the mole, expanding the mole area to 1.34 hectares by
February 2020 (Table 2 and Figure 4a–d, Figure 5). Mole volume had decreased by over 63%.
Regression analysis indicates that based on the past rate of erosion the site will require at least
another 4.66 years to fully erode (Figure 6).
3.2. Ecological metrics
Beach wrack, and Large Woody Debris (LWD) and surf smelt metrics increased along the
shorelines east and west of the Twins mole the first year after the 2017 rock removal and relative
to the control site (Cline Spit). Trends vary with metric, and in total indicate that the site is
functioning ecologically and continuing to evolve. Details by metric are as follows.
Table 2. Twins mole area and volume loss, pre-project- to February 2020.
Date Days after project Mole Area (ha) Mole Volume (m
3)
Volume (m
3
) Loss
5/26/17 0 2.582 206,975 0
8/26/17 0 2.112 142,580 64,395
9/6/17 31 2.032 136,516 6064
11/25/17 111 1.979 128,491 8025
1/17/18 163 1.866 127,171 1320
4/26/2018 262 1.845 122,541 4631
5/24/18 290 1.793 122,443 98
8/28/2018 362 1.683 116,279 6164
11/30/2018 461 1.483 101,589 14,690
1/26/2019 518 1.306 95,991 5598
3/2/2019 553 1.291 94,997 994
5/19/2019 631 1.085 91,131 3823
8/7/2019 711 1.340 87,952 3179
11/11/2019 807 1.012 85,849 2103
12/26/2019 868 1.356 81,305 4544
1/11/2020 884 1.358 76,960 4345
2/9/2020 913 1.335 76,157 803
6A. SHAFFER ET AL.

3.3. Beach wrack
3.3.1. Sediment
The majority of the beach wrack sediment was in the middle size classes, with pebble, gravel, and
sand making up the largest percentage of sediment, and very little cobble and silt/clay collected.
Figure 4. Twins mole pre-project a. August 2015 (top); b. August 2018 (middle); c. March 2019; d. March 2020 (bottom).
0.000
0.500
1.000
1.500
2.000
2.500
3.000
0 200 400 600 800 1000
Area (ha)
# Days Since Removal
Twins Mole Area (ha) vs. Time (days)
Figure 5. Mole area (ha) by time, days since armoring removal 2017-present.
INTERNATIONAL JOURNAL OF MINING, RECLAMATION AND ENVIRONMENT 7

Both cobble and silt/clay were only collected at Twins West, both with higher contributions in 2020
as compared to 2017. Results of the GLM analysis of both cobble and silt/clay indicated significant
differences, which are driven by the absence of these size classes at Twins East and Cline Spit. The
amount of pebble decreased between years at Twins West, but increased at both Twins East and
Cline Spit. Results of the GLM analysis indicate a significant site x year interaction term for the top
layer and total pebble composition. Post-hoc analyses confirm significant site differences between
Twins West and Twins East, and the significant interaction is being driven by the differences
between Twins West in 2017 and Twins East in 2020. No statistical differences were observed in the
amount of gravel observed among sites and between years (Table 3), and the total percentage of
gravel was relatively consistent among all samples (Figure 7). Significant differences were observed
in the percentage of sand observed among sites and between years. Twins East had the highest
percentage of sand of the three sites, which made up close to 100% of sediment found in the top
layer (Figure 7). Percentage of sand decreased between years at Twins East, but remained more
consistent at the other two sites. Results of the GLM and post-hoc analyses indicate the differences
in total sand were significant among all sites (p < 0.05, Table 3). Top layer sand differences were
driven the difference between Twins West in 2017 and Twins East in 2020, while Cline Spit was
relatively consistent between years.
In short, sediment composition became coarser on the west side and finer on the east side of the
mole after the restoration project. No such changes were observed along the Cline Spit (compara-
tive) site (Figure 7 and Table 3)
3.3.2. Percent composition
Algae made up the vast majority of beach wrack at all sites over both years sampled. Eelgrass made
up between 10–25% of beach wrack, while terrestrial plants were observed rarely, with the largest
percentage observed at the control sites in 2020 (Figure 7). Results of the GLM analysis indicate that
there were significant differences pre- and post-project for total wrack cover, eelgrass, and terrestrial
plants (p < 0.05, Table 4). Significant site differences were observed for algae, eelgrass, and terrestrial
plants, and a significant interaction between site and year was observed for all but eelgrass (Table 4).
Results of the post hoc analysis for total wrack indicate that the difference in the site x year
interaction was driven by the difference between Twins West and Cline Spit in 2020. For algae,
differences were driven by lower total coverage at Cline Spit, and the increase in coverage between
y = -88.63x + 150907
R² = 0.766
0
50000
100000
150000
200000
250000
0 200 400 600 800 1000
Fill Volume (m3)
Number of Days
Twins Mole Volume (m3) vs. Time (days)
Figure 6. Mole volume change by days since removal, and loss rate correlation, from project start August 2017 to February 2020.
8A. SHAFFER ET AL.

2017 and 2020 at Twins East. Algae coverage increased along the east side of the mole relative to the
west side and to pre-project. Overall differences for eelgrass were low, according to post hoc
analyses, and differences in terrestrial plants were driven by the large percentage observed at
Cline Spit in 2020.
3.3.3. Beach Wrack Invertebrates
Beach wrack invertebrate total abundance species richness, and species diversity also showed
upticks at the east restoration site post-project (Figure 8). Results of the GLM analysis indicate
that there were significant differences (p < 0.05) in the total invertebrate count among sites and
between years (Table 5). The interaction between site and year was also significant (p < 0.05, Table
5). This difference stems from large differences observed at Twins East, where counts were
significantly higher in 2020 (Figure 9, 10). There were no significant differences observed for either
Table 3. Beach wrack sediment characteristics by year and site. * = statistical significance.
Sediment Type Variable F-value p-value
Top Layer Cobble Year 0.33 0.75
Site 13.54 <0.0001 *
Year*Site 0.18 0.83
Top Layer Pebble Year 6.71 0.015 *
Site 9.91 <0.0001 *
Year*Site 3.73 0.024 *
Top Layer Gravel Year 1.49 0.1
Site 13.66 <0.0001 *
Year*Site 4.83 0.008 *
Top Layer Sand Year 18.9 <0.0001 *
Site 38.92 <0.0001 *
Year*Site 3.14 0.043 *
Top Layer Silt/Clay Year 0 N/A
Site 0 N/A
Year*Site 0 N/A
5 cm Cobble Year 3.65 0.09
Site 9.55 <0.0001 *
Year*Site 1.31 0.27
5 cm Pebble Year 15 0.00018 *
Site 3.48 0.031 *
Year*Site 2.17 0.11
5 cm Gravel Year 2.63 0.1
Site 0.11 0.9
Year*Site 0.8 0.45
5 cm Sand Year 9.56 0.004 *
Site 10.87 s *
Year*Site 2.42 0.089
5 cm Silt/Clay Year 0.4 0.56
Site 0.9 0.41
Year*Site 0.32 0.73
Total Cobble Year 2.07 0.25
Site 17.23 <0.0001 *
Year*Site 0.33 0.72
Total Pebble Year 14.38 0.00029 *
Site 9.01 0.00012 *
Year*Site 4.05 0.017 *
Total Gravel Year 0.18 0.83
Site 3.97 0.019
Year*Site 1.96 0.14
Total Sand Year 18.38 <0.0001 *
Site 26.71 <0.0001 *
Year*Site 1.56 0.21
Total Silt/Clay Year 0.4 0.56
Site 0.9 0.41
Year*Site 0.32 0.73
INTERNATIONAL JOURNAL OF MINING, RECLAMATION AND ENVIRONMENT 9

species richness or Shannon–Weiner diversity index among sites or between years (p > 0.05,
Table 5).
Acari, spp., Amphipoda, Chordata (fish eggs), Coleoptera (beetles), Diptera (flys), ostrocods,
nematode and oligochaete worms were the most abundant invertebrate taxa at all sites. Abundances
of many of these species increased significantly at the Twins East site post-project (Figure 10).
3.4. Large woody debris (LWD)
LWD count, size and volume varied by year some metrics were higher than the project start in 2017,
and relative to comparative site (Figure 11). LWD count was highest a year after the restoration project
and then returned to pre-project numbers (Table 6). LWD diameter was significantly larger post
project. Length of LWD was found to be significantly higher at the restoration site pre-project. LWD
abundance was higher at the restoration site relative to the control site the first year after restoration.
This abundance difference decreased 2 years post-project and was ultimately not significantly different
(Figure 11).
3.5. Surf smelt spawning
Surf smelt eggs were present along the east Twins shoreline and control area for all years (2017–-
2019) and both east and west of the mole beginning in 2019. Eggs numbers increased in abundance
along the east Twins shoreline during May through July. This is the first year eggs have been
detected west of the mole (Table 7 and Figure 12).
4. Discussion
The Twins ecosystem restoration project, as one of the largest ecosystem restoration projects in the
Salish Sea, continues to inform our understanding of how nearshore ecosystems respond to large-
scale events including ecosystem-scale restoration actions and large-scale landslides. The restora-
tion project itself was massive, requiring years of federal, state, and local permitting review for
approval. Implementation required up to six construction crew (excavator and off road dump truck
operators) over full time for 1 month. Despite the scale of the project, the restoration itself was of
little environmental impact. The removal of armoring from around the mole structure was rapid
(occurred in just under 1 month), and done almost completely from the upland. The work was done
during appropriate timing to avoid impacts to fish life. All the non-native material was removed
from the shoreline and transported to an appropriate upland site. As a result of this restoration
action, a large volume of native sediment and LWD has been contributed to the shoreline in
a configuration almost identical to large deep-seated landslides that are characteristic of the region.
Table 4. Beach wrack characteristics by year and site. * = statistical significance.
Beach Wrack Type Variable F-value p-value
Total Wrack Cover Year 0.73 0.81
Site 22.2 <0.0001 *
Year*Site 1.95 0.14
Algae Year 10.99 0.0088 *
Site 30.11 <0.0001 *
Year*Site 7.85 0.00039 *
Eelgrass Year 5.33 0.031 *
Site 22.04 <0.0001 *
Year*Site 0.16 0.85
Terrestrial Plants Year 7.98 0.0067 *
Site 4.89 0.0075 *
Year*Site 5.39 0.0046 *
10 A. SHAFFER ET AL.

Three years after the armoring removal, the site is just over half of the original configuration.
Complete erosion is modelled to be complete in less than 5 years. We also see that seasonal episodic
changes in erosion rates that redistribute material along the shoreline are important. In February
and March of 2020, we observed a brief rate of erosion of upwards of 12 metres in 40 days. The
erosion phase was brief, and the material redistributed along the shoreline. This redistribution of
material slows the net decrease in the area of material lost, but is important ecologically, and
Figure 7. Beach wrack metrics of the west and east Twins restoration site beaches, and comparative site before 2017) and after
(2020) project.
INTERNATIONAL JOURNAL OF MINING, RECLAMATION AND ENVIRONMENT 11

illustrates that erosion is not just a net loss of sediment, but also sustains beach conditions
important for ecological processes.
The redistribution of sediment liberated from the restoration site resulted in significant changes
in beach substrate, coarsening the beach up drift and fining the beach downdrift of the restoration
site. Despite this large-scale environmental event, and these marked changes in habitat, the
ecological response to a rapid unlocking and redistribution of a large amount of sediment to the
shoreline appears muted. Observed changes post-project included an uptick in beach wrack volume
and ecological metrics of species richness and diversity as well as some invertebrate species relative
to the control site. The highest increase was consistently observed at the east site, indicating
Figure 8. Beach wrack ecological metrics of the west and east Twins restoration site beaches and comparative site before (2017)
and after (2018–2019) project.
12 A. SHAFFER ET AL.

a reconnection of hydrodynamic processes that resulted in an increased distribution of beach wrack
downdrift of the structure. This increase in drift algae on the beach in turn fueled an increase in
invertebrate abundances at the restoration site. The species that increased are all important food for
migrating birds, forage fish, and salmon that use this shoreline. The similarity of metrics of
invertebrate species richness and diversity to control levels post-project confirms the variable
nature of nearshore invertebrate communities, and ecological resilience to project-related site
disturbance [16–18].
Another interesting response of ecological metrics was the change in distribution and abundance
of surf smelt eggs, which showed an increase at the restoration site after the project relative to pre-
Figure 9. Beach wrack invertebrate abundance averages of the west and east Twins restoration site beaches, and comparative
site before (2017) and after (2020) project.
INTERNATIONAL JOURNAL OF MINING, RECLAMATION AND ENVIRONMENT 13

project and the control site. Further, eggs appeared for the first time up drift of the site after
restoration action. This ‘expansion’ of surf smelt spawning area was likely a result of a coarsening of
the beach in this area which thereby made it suitable for surf smelt spawning substrate. We
hypothesise the spawning activity along this west/up drift reach of shoreline will persist and egg
abundance will continue to increase as time goes on.
Collectively the physical and ecological changes documented in this study are very informative to
understanding large-scale landslides which in large part define shorelines of the region (Figures 2,
12 and Table 8). Hydrodynamic processes associated with the introduction of large amounts of
native sediment to the shoreline include, not just erosion, but also sediment redistribution, change
in beach substrate composition and concomitant ecological response. When these large-scale
landslide events happen they liberate a large amount of sediment quite a distance from the shoreline
(Figure 12, Figure 13) and, except for the remnant toe features, erode fairly quickly-in less than
a decade. From our study, we now know that the shoreline ecological communities experiencing
these rapid changes in hydrodynamic processes are resilient, and respond positively within 2 years
to these large-scale ecosystem processes including increase in beach wrack distribution, invertebrate
abundances, and surf smelt spawning along newly suitable shorelines. Given the sheer size of the
project, the project location along the main corridor connecting coastal and inland waters of
Washington, and immediate increase in surf smelt spawn distribution and abundance we conclude
Table 5. Results of generalised linear models (GLM) for invertebrate metrics. Asterisk (*) indicates
significant difference.
Invertebrate Metric Variable F-value p-value
Total Invertebrates Year 3.40 0.07
Site 4.22 0.02 *
Year*Site 4.95 0.01 *
Species Count Year 1.34 0.25
Site 1.19 0.31
Year*Site 0.20 0.82
Shannon-Weiner Diversity Index Year 0.11 0.74
Site 2.55 0.08
Year*Site 0.02 0.98
Figure 10. Beach wrack sediment characteristics west and east of the Twins Mole and comparative site (Cline Spit) before (2017)
and after (2020) restoration project.
14 A. SHAFFER ET AL.

Figure 11. Large woody debris count, length, and diameter west and east of the Twins Mole and comparative site (Cline Spit).
before (2017) and after (2018–2019) restoration project.
Table 6. LWD statistical analysis by year and site. * = statistical significance.
LWD Type Variable F-value p-value
LWD Diameter Year 8.09 0.005 *
Site 0.82 0.44
Year*Site 1.23 0.29
LWD Length Year 4.33 0.058
Site 4.19 0.02 *
Year*Site 0.58 0.56
LWD Count Year 0.95 0.32
Site 0.58 0.56
Year*Site 0.17 0.84
INTERNATIONAL JOURNAL OF MINING, RECLAMATION AND ENVIRONMENT 15

Table 7. Surf Smelt egg summary east and west Twins and comparative area before (2017) and after
(2019) restoration.
Site Year Total # of samples Average of total # of eggs
Twins West 2017 13 0
2018 8 0
2019 16 0.31
Twins East 2017 17 4.53
2018 8 3.00
2019 16 107.50
Cline Spit 2017 6 189.33
2018 4 112.75
2019 4 193.25
Figure 12. (a) Samples with surf smelt eggs west and east of the Twins mole 2019; (b) Surf smelt spawn map prior to 2019
(reprinted from WDFW: Source: https://wdfw.wa.gov/fishing/management/marine-beach-spawning).
16 A. SHAFFER ET AL.

that the restoration will continue to evolve and provide significant positive feedback to the larger
Salish Sea. These results will inform future shoreline management considerations, and contribute to
our understanding of lessons learned from other ongoing large scale ecosystem restoring restora-
tion projects of the Salish sea region [19-21].
These results add to the knowledge base of nearshore ecosystem response to drift cell scale
restoration events, including smaller landslides [19], large-scale dam removals [20,21], and indicate
that properly designed ecosystem-scale restoration actions are an effective tool to restore both
physical processes and shoreline ecosystem function.
Acknowledgments
A number of people contributed to the restoration project and field monitoring. Bruch and Bruch Construction
performed extremely skilled and surgical restoration. Caroline Walls, Lindsey Howard, Breyanna Waldsmith, Tara
Table 8. Landslide inventory for the Pysht-Twins shoreline, landslide estimated volume, based on toe distance from shoreline.
Landslide number # Latitude Longitude Area (m
2
) Distance from MHWM (m)
1 48.160666 −123.921762 10387 60.6
2 48.160487 −123.902557 20391 85.0
3 48.164649 −123.968065 8113 50.8
4 48.164468 −123.973954 7871 50.5
5 48.16476 −123.977355 40437 139.3
6 48.165694 −123.978815 138906 292.6
7 48.167389 −123.986704 65580 272.6
8 48.166866 −123.983039 53588 181.6
9 48.168078 −123.992392 15801 74.0
10 48.16836 −123.996913 8792 60.5
11 48.168756 −123.997224 20,584 107.0
12 48.182024 −124.059879 4250 54.3
13 48.193618 −124.085111 5415 60.6
14 48.194923 −124.088013 9418 82.8
15 48.196629 −124.092172 2649 29.4
16 48.19683 −124.094964 3175 30.6
Figure 13. Landslide deposit area vs distance from MHWM.
INTERNATIONAL JOURNAL OF MINING, RECLAMATION AND ENVIRONMENT 17

McBride, CWI, led field components of ecological field monitoring. CWI/Peninsula College/WWU Student interns
Kendall Barton, Trevor Bermaster, Marissa Christopher, Bronwyn Davis, Josh Davis, Anthony Delorenzi, David
Harvey, Riley Philips, Sam Schotterback, Tony Thompson and Seren Weber completed field sampling. Dan Dafoe
and Chris Byrnes, WDFW (retired) contributed to site visits. Dan Penttila provided surf smelt egg sample analysis
and decades of technical assistance and goodwill. Funding for this project was provided by the Lafarge Corporation,
Patagonia (for internships) and Coastal Watershed Institute. Thank you to all.
Disclosure statement
No potential conflict of interest was reported by the author(s).
References
[1] M.N. Dethier, W.W. Raymond, A.N. McBride, J.D. Toft, J.R. Cordell, A.S. Ogston, S.M. Heerhartz, and H.
D. Berry, Multiscale impacts of armoring on Salish Sea shorelines: Evidence for cumulative and threshold eects,
Estuar. Coast Shelf Sci. 175 (2016), pp. 106–117. doi:10.1016/j.ecss.2016.03.033.
[2] B.S. Miller, C.A. Simenstad, J.N. Cross, K.L. Fresh, and S.N. Steinfort. Nearshore fish and macroinvertebrate
assemblages along the Strait of Juan de Fuca including food habits of the common nearshore fish. Final report
of three years sampling 1976-1979. EPA No. D6-E693-EN. Office of Environmental Engineering and
Technology Office of Research and Development, US EPA, Washington, DC, 1980.
[3] WSDOT, Landslide mapping of the twin rivers-deep creek area. Washington Department of Natural
Resources. D. Anderson & G. Taylor. Washington Department of Transportation, Olympia, WA, 2006.
[4] E. White, Heiltsuk stone fish traps on the central coast of British Columbia, in The Archaeology of North Pacific
Fisheries, M. ML and A. Cannon, eds., University of Alaska Press, Fairbanks, 2011 pp.75–90.
[5] D. Lepofsky and M. Caldwell, Indigenous marine resource management on the Northwest Coast of North
America, Ecol. Proc. 2 (1) (2013), pp. 12. doi:10.1186/2192-1709-2-12.
[6] J.T. Elder, D.M. Gilmour, V.L. Butler, S.K. Campbell, and A. Steingraber, On the role of coastal landscape
evolution in detecting fish weirs: A Pacific Northwest Coast example from Washington State, J. Island Coast.
Archaeol. 9 (1) (2014), pp. 45–71. doi:10.1080/15564894.2014.881933.
[7] J.A. Shaffer, J. Paul, P. Crain, M. McHenry, P. Jensen, T. Whitey, D. Parks., and A. Schouten, Nearshore
Restoration Strategy for Twin Rivers: A Revised Proposal by the Twins Nearshore Restoration Work Group 2004,
Revised 2009, Port Angeles, WA, 2004. www.coastalwatershedinstitute.org
[8] Puget Sound Nearshore Ecosystem Restoration Project. (PSNERP). Nearshore Restoration Strategy for Twin
Rivers. Strategic Restoration Conceptual Engineering-Design Report. USACE, Seattle, WA and WDFW,
Olympia, WA, 2011
[9] CWI blog. http://www.coastalwatershedinstitute.org/blog/.
[10] L.L. Moulton and D.E. Penttila, Field manual for sampling forage fish spawn in intertidal shore regions. San
Juan County Forage Fish Assessment Project. WDFW, LaConner, WA and MJM Research, Lopez Island, WA,
2006.
[11] Blue Marble Geographics, 2018. GlobalMapper Version 20.0., Hallowell, Maine.
[12] S.M. Heerhartz, M.N. Dethier, J.D. Toft, J.R. Cordell, and A. Ogston, Eects of shoreline armoring on beach
wrack subsidies to the nearshore ecotone in an estuarine fjord, Estuaries Coasts 37 (5) (2014), pp. 1256–1268.
doi:10.1007/s12237-013-9754-5.
[13] S.M. Heerhartz, J.D. Toft, J.R. Cordell, M.N. Dethier, and A.S. Ogston, Shoreline armoring in an estuary
constrains wrack-associated invertebrate communities, Estuaries Coasts 39 (1) (2016), pp. 171–188.
doi:10.1007/s12237-015-9983-x.
[14] P. Dionne, Vortex Method for Separation of Forage Fish Eggs from Beach Sediment Addendum to the 2006
Revision of Field Manual for Sampling Forage Fish Spawn in Intertidal Shore Regions, WDFW, Olympia,
Washington, 2016.
[15] S.L. Rich, J.A. Shaffer, M.J. Fix, and J.O. Dawson, Restoration considerations of large woody debris in the Elwha
River Nearshore, Olympic Peninsula, Washington, Project MUSE; Ecol. Restorat. 32 (3) (2014), pp. 306–313.
muse.jhu.edu/article/552446.
[16] W.P. Sous, The role of disturbance in natural communities, Annu. Rev. Ecol. Syst. 15 (1) (1984), pp. 353–391.
doi:10.1146/annurev.es.15.110184.002033.
[17] K.M. Dernie, M.K. Kaiser, and R.M. Warwick, Recovery rates of benthic communities following physical
disturbance, J. An. Ecol. 72 (6) (2003), pp. 1043–1056. doi:10.1046/j.1365-2656.2003.00775.x.
[18] S.R. Palumbi, K.L. McLeod, and D. Grünbaum, Ecosystems in action: Lessons from marine ecology about
recovery, resistance, and reversibility, BioSci. 58 (1) (2008), pp. 33–42. doi:10.1641/B580108.
18 A. SHAFFER ET AL.

[19] J.A. Shaffer and D.S. Parks, Seasonal variations in and observations of landslide impacts on the algal composi-
tion of a puget sound Nearshore Kelp Forest, Bot. Marina 37 (1994), pp. 315–323. doi:10.1515/
botm.1994.37.4.315.
[20] J.A. Shaffer, F. Juanes, T.P. Quinn, D. Parks, T. McBride, J. Michel, C. Naumann, M. Hocking, and C. Byrnes,
Nearshore fish community responses to large scale dam removal: Implications for watershed restoration and fish
management, Aquat. Sci. 79 (3) (2017), pp. 643–660.
[21] J.A. Shaffer, E. Higgs., C. Walls., and F. Juanes, Large-scale dam removals and nearshore ecological restoration:
Lessons learned from the Elwha dam removals, Ecol. Restorat. 35 (2) (2017), pp. 87–101. doi:10.3368/er.35.2.87.
INTERNATIONAL JOURNAL OF MINING, RECLAMATION AND ENVIRONMENT 19
View publication stats