Quantifying the effectiveness of shoreline armoring removal on coastal biota of Puget Sound

Abstract

Shoreline armoring is prevalent around the world with unprecedented human population growth and urbanization along coastal habitats. Armoring structures, such as riprap and bulkheads, that are built to prevent beach erosion and protect coastal infrastructure from storms and flooding can cause deterioration of habitats for migratory fish species, disrupt aquatic–terrestrial connectivity, and reduce overall coastal ecosystem health. Relative to armored shorelines, natural shorelines retain valuable habitats for macroinvertebrates and other coastal biota. One question is whether the impacts of armoring are reversible, allowing restoration via armoring removal and related actions of sediment nourishment and replanting of native riparian vegetation. Armoring removal is targeted as a viable option for restoring some habitat functions, but few assessments of coastal biota response exist. Here, we use opportunistic sampling of pre- and post-restoration data for five biotic measures (wrack % cover, saltmarsh % cover, number of logs, and macroinvertebrate abundance and richness) from a set of six restored sites in Puget Sound, WA, USA. This broad suite of ecosystem metrics responded strongly and positively to armor removal, and these results were evident after less than one year. Restoration responses remained positive and statistically significant across different shoreline elevations and temporal trajectories. This analysis shows that removing shoreline armoring is effective for restoration projects aimed at improving the health and productivity of coastal ecosystems, and these results may be widely applicable.

Introduction

Worldwide, shorelines adjacent to bodies of fresh and salt waters face faster urbanization and population growth than other geographic regions (Neumann et al., 2015). Coastal regions have always experienced high immigration rates because of their ease of access to domestic and international shipping, military and defense uses, tourism, access to recreational activities, access to valuable ecosystem services, and employment opportunities (Bulleri, Chapman & Underwood, 2005; MEA, 2005; Gittman et al., 2015; Neumann et al., 2015). About 50% of the world’s population lives within 200 km from the coastlines and half of the world’s major city centers are located within 50 km from coasts (Stegeman & Solow, 2002; MEA, 2005). Many of these heavily populated coastal regions are in low-lying elevations. In 2000 these low-elevation coastal zones included nearly 11% of the world’s total coastal population, but by 2060 it is estimated that the population in these low-elevation coastal zones will be as great as 1.4 billion, or 12% of the world’s population (Neumann et al., 2015).

Coastal infrastructure and urban centers are exposed to various hazards including storms, large waves, flooding, sea level rise, and erosion (Jones & Hanna, 2004; McGranahan, Balk & Anderson, 2007). As a response, many coastal communities have established hardened structures such as bulkheads, jetties, riprap revetments and seawalls, a practice commonly called “shoreline armoring” that is part of ocean sprawl (Chapman, 2003; Bulleri & Chapman, 2010; Chapman & Underwood, 2011; Heerhartz et al., 2014; Gittman et al., 2015; Firth et al., 2016b). In some large urban centers such as San Diego Bay, Chesapeake Bay, Sydney Harbor, and Hong Kong’s Victoria Harbor, over 50% of shorelines have been armored, and the continuing growth of coastal immigration and urbanization is expected to increase the rate of shoreline armoring (Davis, Levin & Walther, 2002; Dugan et al., 2008; Lam, Huang & Chan, 2009; Patrick, Weller & Ryder, 2016). In the United States alone, about 14% of the lower 48 states’ shorelines are armored, and 64% of these armored shorelines are adjacent to estuaries and coastal rivers (Gittman et al., 2015).

Armored shorelines overall are associated with lower biodiversity, vegetation cover, and abundances of invertebrate and fish species (Moreira, Chapman & Underwood, 2006; Bilkovic & Roggero, 2008; Peterson et al., 2000; Firth et al., 2016a; Dugan et al., 2008; Morley, Toft & Hanson, 2012; Peters, Yeager & Layman, 2015; Gittman et al., 2016a). Armored shorelines can accelerate beach erosion as waves are reflected from armored structures (Heatherington & Bishop, 2012; Gittman et al., 2016b). They can reduce the overall ecological health of coastal ecosystems by degrading shallow intertidal habitats valuable for survival of juvenile fish and aquatic invertebrates (Bilkovic & Roggero, 2008; Seitz et al., 2006; Gittman et al., 2016a). Armored shorelines can also disrupt the transition between terrestrial and aquatic habitats as the gradual shoreline slope is abruptly steepened, which in turn can result in reduction of salt marsh habitats and submerged aquatic vegetation. Similarly, armoring also reduces deposition of woody debris and “wrack” or organic matter deposition on shorelines. This loss of organic debris can affect the aquatic–terrestrial food web, including fishes, macroinvertebrates associated with wrack and vegetated habitats, and birds (Bozek & Burdick, 2005; Dugan et al., 2008; Heerhartz et al., 2014; Harris, Strayer & Findlay, 2014; Heerhartz & Toft, 2015; Dethier et al., 2016; Wensink & Tiegs, 2016).

Recently, alternatives to shoreline armoring, including armoring removal, have emerged to simultaneously protect coastal urban infrastructure and restore ecological health (Davis et al., 2015; Gittman et al., 2016a; Bilkovic et al., 2017). Approaches include living shoreline techniques such as the creation of marsh sills in lieu of armoring shorelines in North Carolina (Bilkovic & Mitchell, 2013), restoring oyster reefs in Chesapeake Bay (Lawless & Seitz, 2014), restoring red mangrove colonization on riprap revetments in Biscayne Bay, Florida (Peters, Yeager & Layman, 2015), and managed realignment in Europe (Esteves & Williams, 2017). Few studies have assessed the effectiveness of armoring removal on restoring coastal ecosystems, but they generally demonstrate that shorelines without armoring can host higher abundances and diversity across different taxonomic groups. For example, marsh sills have higher abundance and diversity of fish and bivalves in shorelines of North Carolina (Gittman et al., 2016a) and introducing native riparian vegetation and logs after armoring removal can facilitate rapid response of macroinvertebrate assemblages in shorelines of Puget Sound, Washington (Toft, Cordell & Armbrust, 2014).

In Puget Sound, Washington, USA, there has been recent momentum to restore armored shorelines through removal of armoring structures, nourishment of sediments, re-planting native riparian vegetation, and addition of logs and woody debris (Toft et al., 2013b; Toft, Cordell & Armbrust, 2014). Such restoration efforts are driven by the need to protect Pacific salmon species such as endangered populations of Chinook salmon (Oncorhynchus tshawytscha) that are of cultural, ecological, and economic importance to the region (Rhodes et al., 2006; Munsch, Cordell & Toft, 2015). Juveniles of these and other salmon species use shallow intertidal areas as nursery habitats (Munsch, Cordell & Toft, 2016). Macroinvertebrate prey, both aquatic and terrestrial, are a vital part of Chinook diets and coastal food webs. Changes in their populations can negatively impact food availability for many fish species (Sobocinski, Cordell & Simenstad, 2010). Therefore, it is essential to ask whether shoreline restoration is having its intended effect, and to date, relatively little such analysis has been done.

Here we present an analysis of the effects of shoreline restoration, with the objective to determine how coastal biota respond when shoreline armor is removed, sediments nourished, and native vegetation planted. We assess responses across (a) study sites, (b) coastal biota type, (c) shoreline elevations, and (d) trajectories in time. Understanding these post-restoration dynamics can provide knowledge about factors that contribute to biological recovery that can be useful along other armored shorelines regionally and globally.

Materials and Methods Study sites and sources of data

Puget Sound is a fjordal estuarine ecosystem comprising the southern part of the Salish Sea, which encompasses over 30,000 km² in the Pacific Northwest, overlapping Washington, USA and British Columbia, Canada. This ecosystem comprises cold-temperate waters, river deltas, and shorelines mainly composed of clay, sand, mud, and gravel sediments originating from receded glaciers. Continued erosion of coastal bluffs contributes this sediment mix to Puget Sound beaches (Shipman, 2001). More than a quarter of the 4,000 km of shorelines in Puget Sound are armored (Puget Sound Partnership, 2016).

We assessed six sites in Puget Sound, along which armoring was removed, to determine coastal biota responses (Fig. 1). These restored sites from north to south were Cornet Bay of Deception State Park on Whidbey Island, Powel Property in Port Madison on the north side of Bainbridge Island, Salmon Bay Natural Area which is a park downstream from the Hiram M. Chittenden Locks of Seattle, Olympic Sculpture Park in downtown Seattle, and two locations in Seahurst Park in the city of Burien, WA (Fig. 1; Table 1). All sites were formerly armored with concrete or wooden bulkhead and riprap, and the Salmon Bay Natural Area also had an overwater structure (Fig. 2). These sites were restored from 2005 to 2014 (Table 1), and monitoring was accomplished opportunistically depending on the year that sites were restored and the capacity for monitoring that was available. All data analyzed herein were compiled from site-specific reports, as stated in Table 1. Our analyses are new, building from previous technical reports and publications that represent individual sites, to increase spatial and temporal scales including the most recent post-restoration data.

Figure 1: Map of the Puget Sound and the restored sites used for analysis. A, Cornet Bay; B, Powel Property; C, Salmon Bay Natural Area; D, Olympic Sculpture Park; E, Seahurst Park I (restored 2005); F, Seahurst Park II (restored 2014). Map Background: ArcGIS 10.2 Ocean Basemap (Credits: Esri, DeLorme, GEBCO, NOAA NGDC, and other contributors). Inset Map Background: ArcGIS Light Gray Canvas Map (Copyright: ©2013 Esri, DeLorme, NAVTEQ). DOI: 10.7717/peerj.4275/fig-1

Table 1: Sites used for analysis listed by pre-restoration year monitored (PR year), restoration year (Rest. year), post-restoration monitoring years and the types of coastal biota monitored.

Site	PR year	Rest. year	Post-restoration monitoring year							Coastal biota monitored					Elev.	Reference
<1	1	2	3	4	5	10	W %	L #	SM %	MIC	MIR
CB	2012	2013	–	X	–	–	–	–	–	X	X	–	X	X	–	Dethier et al. (2016)
PP	2012	2012	–	X	X	–	–	–	–	–	–	X	–	X	X	Adams, Padgham & Toft (2015)
SBNA	2004	2010	X	–	X	–	–	–	–	–	–	–	X	X	–	Toft et al. (2013a)
OSP	2005	2006	–	X	–	X	–	X	–	–	–	–	X	X	–	Toft et al. (2013b) and Cordell et al. (2017)
SHP I	2004	2005	–	X	–	X	–	X	X	–	–	–	X	X	X	Toft, Cordell & Armbrust (2014) and Toft (2016)
SHP II	2010	2014	–	X	–	–	–	–	–	–	–	–	X	X	X	Toft (2016)

DOI: 10.7717/peerj.4275/table-1

Notes:

CB, Cornet Bay; PP, Powel Property; SBNA, Salmon Bay Natural Area; OSP, Olympic Sculpture Park; SHP I, Seahurst Park restored in 2005; SHP II, Seahurst Park Restored in 2014; W %, wrack % cover; L #, number of logs; SM %, saltmarsh % cover; MIC, macroinvertebrate counts; MIR, macroinvertebrate richness; Elev., shoreline elevation sampling.

Figure 2: Three of the six restored sites used for this analysis. Frames (A, C, and E) show shorelines armored prior to their respective restorations and frames (B, D, and F) show shorelines in their restored state. Left to right: Cornet Bay (A, B), Salmon Bay Natural Area (C, D), Seahurst Park I (E, F). Photo Credit for frame (A): Sarah Schmidt. Photo Credit for frame (B): Lisa Kauman. Photo Credit for frames (C–F): Jason D. Toft. DOI: 10.7717/peerj.4275/fig-2

Five ecosystem metrics were monitored before and after restoration, with one site monitored up to 10 years after restoration (Table 1). Survey data varied by site due to goals and characteristics of each restoration site, and included counts and richness (number of individual taxa) of macroinvertebrates for both terrestrial and aquatic groups, wrack % cover, number of logs (which have both biotic and physical attributes), and saltmarsh % cover. Three sites were also monitored at two different shoreline elevations (Table 1).

While we sampled macroinvertebrates at all six sites, we also sampled non-macroinvertebrate ecosystem metrics at two sites, Cornet Bay and Powel Property (Table 1). At Cornet Bay, we sampled five replicates of terrestrial and aquatic macroinvertebrates before restoration (July 2012) and after restoration (July 2013) across a 50 m transect parallel to the shore (Dethier et al., 2016). During those same respective days, we also sampled 10 replicates of total wrack % cover and counted five replicates of total number of logs before and after restoration (Dethier et al., 2016). We collected wrack samples and the top 2.5 cm of sediment layer using a 15 cm diameter benthic corer; all invertebrates were separated from the wrack and counted and identified to the lowest possible taxonomic classification using a dissecting microscope (Dethier et al., 2016). At Powel Property, we sampled three replicates of macroinvertebrates for one year before restoration (2012) and two years after restoration (2013 and 2014). This sampling design was replicated across low (3.05 m mean lower low water or MLLW) and high (3.66 m MLLW) shoreline elevations (Adams, Padgham & Toft, 2015) corresponding to the placement of the armoring. In addition, we also sampled for saltmarsh % cover across three replicates for the same pre- and post-restoration years monitored for macroinvertebrates.

For the other four sites, we used macroinvertebrate counts and richness metrics for both aquatic and terrestrial types (Toft et al., 2013a, 2013b; Toft, Cordell & Armbrust, 2014; Toft, 2016; Cordell et al., 2017). At Salmon Bay Natural Area, we sampled seven replicates of aquatic invertebrates using sediment cores and terrestrial insects using plastic bin fall traps for one year before restoration (2004) and two post-restoration years (2010 and 2012) across April, May, June, and July (Toft et al., 2013a). We repeated this sampling approach in two shoreline elevations, 0.3 and 2.44 m MLLW. At Olympic Sculpture Park, we sampled seven replicates of aquatic invertebrates using an epibenthic pump that sampled the water–sediment interface, and terrestrial insects applying the same sampling gear used in the Salmon Bay Natural Area for one year before restoration (2005) and three post-restoration years (2007, 2009, 2011) across April, May, June, and July (Toft et al., 2013b). In Seahurst Park, we sampled two shorelines, one restored in 2005 and the other restored in 2014 (Table 1). For both the 2005 and 2014 restored shorelines, we sampled aquatic and terrestrial macroinvertebrates using the same sediment core and fallout trap sampling gear as the previous sites (Toft, Cordell & Armbrust, 2014; Toft, 2016). In the 2005 restored shoreline, we sampled seven replicates of one year before restoration (2004) and four separate years after restoration (2006, 2008, 2010, 2015) across the months of June, July, and September (Toft, 2016). Our sampling design for the 2014 restored shoreline was the same, except that we monitored one year before restoration in 2010 and one year after restoration in 2015 (Toft, 2016).

Quantitative analysis

To measure the effectiveness of shoreline restoration on coastal biota, we used Cohen’s D Effect Size (Cohen, 1992). This effect-size statistic has been widely used, for example to measure effects of stream engineering on increasing salmon abundances, pine forest restoration on native understory vegetation, and invasive vegetation removal on restoring native woody plants (Taylor, Smith & Haukos, 2006; McGlone, Springer & Laughlin, 2009; Stewart et al., 2009). Cohen’s D is calculated with the following equation: [Formula omitted. See PDF.] where μ_A is the mean value of measured variable (e.g., counts, richness, percent cover) after restoration, μ_B is the mean value of measured variable before restoration, and σ is the pooled standard deviation. Although values of D are likely to vary with context, as a general guideline, when D less than 0.2, the restoration is considered to have had no effect, while 0.2–0.8 indicate moderate effect, and 0.8 or greater indicates substantial effect (Rosnow, Rosenthal & Rubin, 2000). In our analysis, the greater the values of D, the greater positive response to restoration, in contrast to the armored state and presumably more toward a pre-disturbance natural state.

We calculated D for restoration effects in the following four major categories: (1) separately for each study site, (2) specific to each monitored type of coastal biota, (3) at the two elevations of the base and placement of armoring, and (4) trajectory in time of post-restoration years. For the study sites category, we calculated the effect sizes of all the respective coastal biota monitored for each site individually. For the monitored coastal biota, we calculated the effect sizes for the five types of coastal biota: wrack % cover, number of logs, saltmarsh cover, macroinvertebrate counts, and macroinvertebrate richness. For the shoreline elevation, we calculated effect sizes for the elevation at the base of the armoring and at the elevation where armoring formerly stood. Lastly, for the trajectory in time we calculated effect sizes for the six post-restoration monitoring years. To test for statistical significance of D, we performed one-sample two-tailed t-tests (α = 0.05), comparing the observed data against the null hypothesis (H₀: μ = 0). Where comparing the means of two different elevations, we used a two-sample two-tailed t-test.

Results

All six sites demonstrated positive responses with the mean effect size varying between 1.07 and 1.79 (Fig. 3). Four of the six sites had statistically significant responses (Fig. 3). These four sites were the Salmon Bay Natural Area, Olympic Sculpture Park, and the two Seahurst Park sites (Salmon Bay: t_0.05(2),54 = 9.9, p < 0.001, 95% CI = 0.95, 1.43; Olympic Sculpture Park: t_0.05(2),83 = 6.69, p < 0.001, 95% CI: 1.14, 2.12; Seahurst Park Restored 2005: t_0.05(2),55 = 8.49, p < 0.001, 95% CI = 1.37, 2.22; Seahurst Park Restored 2014: t_0.05(2),27 = 8.23, p < 0.001, 95% CI: 0.8, 1.34). Among sites and elevations, all five coastal biotic measures showed similarly positive effects, and two of the three that could be analyzed with t-tests were significant (Figs. 4 and 5). These two biotic measures were macroinvertebrate counts and richness (Macroinvertebrate Counts: t_0.05(2),140 = 10, p < 0.001, 95% CI: 1.23, 1.84; Macroinvertebrate Richness: t_0.05(2),144 = 12.01, p < 0.001, 95% CI: 1.58, 2.20). Mean effect size was stronger in higher shoreline elevations where armoring had previously been directly placed (Base of the Armoring: t_0.05(2),73 = 8.44, p < 0.001, 95% CI: 1.14, 1.85; On the Armoring: t_0.05(2),73 = 10.24, p < 0.001, 95% CI: 1.98, 2.93).

Figure 3: Cohen’s D Effect Sizes (±SE for error bars) across six restored sites of Puget Sound used in the analysis (site acronyms same as in Table 1). Data labels show the number of coastal biota types monitored and the sample sizes (the number of effect sizes for each site). Effect sizes with asterisks were significantly different from zero. DOI: 10.7717/peerj.4275/fig-3

Figure 4: Cohen’s D Effect Sizes (±SE for error bars) by five major types of coastal biota monitored. Data labels show number of restored sites that were monitored and the sample sizes (the number of effect sizes for each biota type). Coastal biota labeled in orange were not integrated for individual t-tests due to lack of replicates. Coastal biota effect sizes with asterisks were significantly different from zero. DOI: 10.7717/peerj.4275/fig-4

Figure 5: Cohen’s D Effect Sizes (±SE for error bars) by elevation monitored. Data labels show the number of restored sites and the sample sizes (the number of effect sizes for each elevation type). Both elevations’ effect sizes were significantly different from zero. DOI: 10.7717/peerj.4275/fig-5

All of the post-restoration years demonstrated positive and statistically significant responses (p < 0.001 for all post-restoration years), with year 10 showing the greatest positive response ( [Formula omitted. See PDF.] ) and year <1 showing the lowest ( [Formula omitted. See PDF.] ) (Year <1: t_0.05(2),27 = 6.07, p < 0.001, 95% CI: 0.71, 1.43; Year 1: t_0.05(2),91 = 8.02, p < 0.001, 95% CI: 1.16, 1.93; Year 2: t_0.05(2),31 = 6.5, p < 0.001, 95% CI: 0.95, 1.81; Year 3: t_0.05(2),55 = 6.71, p < 0.001, 95% CI: 1.30, 2.40; Year 5: t_0.05(2),55 = 9.2, p < 0.001, 95% CI: 1.22, 1.89; Year 10: t_0.05(2),27 = 6.25, p < 0.001, 95% CI: 2.24, 4.43) (Fig. 6).

Figure 6: Cohen’s D Effect Sizes (±SE) by post-restoration years monitored. Effect sizes reflect comparisons between pre-restoration and post-restoration for each respective post-restoration year monitored. Data labels show the number of restored sites and sample sizes (the number of effect sizes for each post-restoration year monitored). Effect sizes of all post-restoration years were significantly different from zero. DOI: 10.7717/peerj.4275/fig-6

Discussion

Armoring removal in Puget Sound has resulted in diverse, positive responses by coastal biota. All four major effect categories (restored sites, coastal biota, shoreline elevation, and trajectory in time) showed substantially positive responses (effect size >0.8; Rosnow, Rosenthal & Rubin, 2000). Furthermore, there was clear evidence that coastal biota responded quickly (within a year after restoration), with subsequent post-restoration years maintaining or even increasing biotic and abiotic gains. Of the five major types of coastal biota, recovery was strongest for macroinvertebrates; the responses of wrack cover, number of logs, and saltmarsh cover were also positive but they had smaller sample sizes and therefore had weaker statistical inferences.

The strong significant responses of macroinvertebrates were consistent with previous work on restoration of individual marine sites (Toft et al., 2013b; Toft, Cordell & Armbrust, 2014) and other aquatic habitats, most of which are in freshwater ecosystems. Some macroinvertebrate restoration response studies come from rivers and channelized streams, where restoring some habitat complexity results in greater abundances and diversity of macroinvertebrates (Korsu, 2004; Muotka & Syrjänen, 2007; Miller, Budy & Schmidt, 2010). Wetlands that have been restored or newly created can be quickly colonized by macroinvertebrates, especially those with greater dispersal capability such as aerial insects (Brown, Smith & Batzer, 1997; Stewart & Downing, 2008). Our analysis similarly showed that overall, macroinvertebrate responses to restoration in coastal ecosystems are positive and substantial, which in turn can enhance prey availability for migratory fishes and seabirds and improve ecosystem health as a whole (Dugan et al., 2003; Heerhartz & Toft, 2015).

With regards to pace of response, we found that coastal biota can quickly and positively respond to shoreline armor removal and that this trend can continue across multiple post-restoration years. Biotic recovery can be similarly rapid after restoration of channelized streams, saltmarsh habitats, and oyster reefs (Warren et al., 2002; Miller, Budy & Schmidt, 2010; La Peyre et al., 2014). However, this is not always the case, because abundances and diversity of macroinvertebrates can initially drop significantly after restoration of a channelized stream (Muotka & Syrjänen, 2007; Korsu, 2004). Muotka & Syrjänen (2007) found that macroinvertebrate recovery was not pronounced until four to six years, but the variability appeared to stabilize after eight years. Biotic recoveries in restored wetlands were also slow when compared with reference sites across the same temporal scales, and the richness of invertebrates and fishes did not approach their peak trajectories until three to five years after restoration (Simenstad & Thom, 1996; Morgan & Short, 2002). Similar perceptions have been recorded for seagrasses, as restoration shows a greater response when monitored for seven versus three years (Bell, Middlebrooks & Hall, 2014). While our sites experienced rapid recovery, the same might not hold for other similar coastal restoration projects outside the Puget Sound. As is the case with restored wetlands, restored coastlines may not replace the full natural functions of healthy coastlines, and may take longer than their respective monitoring periods to reach fully self-sustaining ecosystem functioning status (Zedler & Callaway, 1999).

We also found that coastal biota directly affected by armoring placement (higher shoreline elevations) responded more positively than biota that were indirectly affected (lower shoreline elevations). Toft, Cordell & Armbrust (2014) came to similar conclusions, with less response seen in biota below the footprint of armoring. This is an interesting comparison to eco-engineering of armoring, where diversity can be higher at lower elevations (Firth et al., 2016b), contrasting the expectations from hard versus soft restoration approaches. In rivers, biota within sections directly affected by dam impoundment can respond positively to dam removal within days to a few years; these biotic parameters include but are not limited to recolonization by native riparian plants and lotic organisms (Hart et al., 2002). Upstream areas that were indirectly affected by dam impoundment may not experience full recovery of aquatic–terrestrial ecosystem linkages until years to decades after the dam removal (Hart et al., 2002). In contrast, previous work on ecological responses to removal of dikes in estuarine tidal wetlands found that habitats indirectly affected by dikes are just as likely to respond positively during a similar temporal scale to habitats directly affected in terms of marsh vegetation, benthic macroinvertebrates, fish, and other megafauna (Hood, 2004).

One of the challenges in our analysis was the small sample size in certain parameters. Quantified effect responses of wrack % cover, saltmarsh % cover, and number of logs were weaker than for parameters with larger sample sizes. Small sample sizes can increase error rate and potentially distort response interpretations (Raudys & Jain, 1991), so we cannot necessarily generalize our results to other shorelines restored through similar means. However, it is important to note that wrack % cover can increase abundances and diversity of macroinvertebrates, facilitate saltmarsh growth, and support megafauna such as seabirds (Dugan et al., 2003; Chapman & Roberts, 2004; Smith, 2007; Harris, Strayer & Findlay, 2014; Heerhartz et al., 2014). Natural shorelines have higher woody debris counts and densities than armored shorelines, and these can enhance and retain wrack % cover but also reduce beach erosion (Angradi et al., 2004; Eamer & Walker, 2010; Harris, Strayer & Findlay, 2014; Heerhartz et al., 2014). Based on the suspected functions that wrack, saltmarsh, and logs provide in healthy coastlines, it is essential to increase the geographical scope and number of studies of these coastal biota types to assess the successful recovery of restored coastal ecosystems. Related to this is the added benefit that before–after control-impact (BACI) techniques would have in aiding interpretation (Underwood, 1994), emphasizing that increased sample size of restored and reference sites over time will be essential to fully understand restoration effectiveness.

While our sample size at the site scale was limited to six, this opportunistic sampling can be advantageous to efficiently produce information-rich results (Palinkas et al., 2015). Opportunistic sampling can also detect presence of certain ecosystem metrics in regions where systematic sampling would not detect them (De Barba et al., 2009). As our results have now demonstrated that armoring removal can elicit rapid ecosystem recovery, and site-specific studies that use BACI analysis have shown similar results (Toft et al., 2013b; Toft, Cordell & Armbrust, 2014), applying BACI in future monitoring studies will be helpful to increase the scale of inference. Furthermore, continuing to monitor the metrics through additional post-restoration years and expanding the pre- and post-restoration monitoring approach to other shorelines may reduce bias caused by opportunistic sampling, such as unequal sample sizes and preferential sampling (De Barba et al., 2009). This will be essential to expand the temporal scale, as most of our sites were restored for less than five years, with only one site restored for 10 years.

Coastal biota recovery after armoring removal may also be hindered or facilitated by abiotic variables and their responses to restoration, which we did not address in our analysis. Beach profiles and sediment grain size may change slowly in response to the placement of shoreline armoring (Dethier et al., 2016), which suggests that these two variables may likewise experience slow recovery across seasons to years after armoring removal. Furthermore, sediment changes following armor removal, such as increasing or decreasing deposition of fines, likely affect recovery patterns of biota. Lower shoreline elevations are also more susceptible to disturbances by hydrological and oceanographic processes, which in turn may prevent rapid recovery of associated coastal biota (Harris, Strayer & Findlay, 2014).

The responses of coastal biota observed in our analysis may be attributed to “passive” ecosystem management post-recovery. The passive approach does not require further action to be taken after a restoration project is completed (Simenstad, Reed & Ford, 2006). Instead, it allows disturbances to occur, potentially influencing the recovery trajectory of ecosystem components. Shorelines are exposed to frequent disturbances such as waves, flooding, and storm-related events (Gittman et al., 2016b). This can enhance recovery by depositing wrack along shorelines, providing habitats for grazing arthropods and saltmarsh vegetation (Dugan et al., 2003; Chapman & Roberts, 2004). Reestablishing the natural cycle of disturbance and recovery driven by various oceanographic processes has the potential to elicit positive responses by coastal biota for years after initial restoration.

Conclusion

Here we have shown that biotic metrics can respond strongly and positively to armor removal and restoration of beaches. Even with our pronounced results, reversing shoreline armoring is and will continue to be a management challenge. Coastal habitats around the world face unprecedented urban growth (Gittman et al., 2015). In the United States, shoreline armoring is primarily driven by development of residential properties, attempts to improve domestic and international shipping traffic, and protection against storm events (Gittman et al., 2015). With nearly half of the world’s population expected to live within 100 km from shorelines by 2030, it is safe to assume that armoring will continue to increase within and outside the United States in the next few decades (MEA, 2005; Gittman et al., 2015, 2016b). However, it is critical to recognize that shorelines without armoring can function as natural erosional barriers. For example, large woody debris protects from beach erosion but also enhances wrack accumulation, which in turn can enhance saltmarsh growth and improve aquatic–terrestrial connectivity (Chapman & Roberts, 2004; Eamer & Walker, 2010; Heerhartz et al., 2014). Removing shoreline armoring and improving aquatic–terrestrial connectivity is not only beneficial to the ecosystem but also can help coastal communities and livelihoods (Firth et al., 2016b), because ecosystem components that are harvested (such as fishes) rely on ample availability of macroinvertebrate prey for survival.

It is therefore critical for policymakers to consider numerous benefits of shoreline armoring removal before undertaking new shoreline development. While removal of armoring is not feasible in all cases due to financial or safety concerns, it is clear from this study that restoring shorelines through armoring removal can potentially benefit coastal ecosystem health and coastal populations by increasing ecosystem services. Furthermore, many shoreline homeowners are increasingly recognizing environmental impacts of shoreline armoring and express a preference for natural shoreline structures, as they can be aesthetically appealing and have many ecological benefits (Scyphers, Picou & Powers, 2014). Existing and new shoreline management policies should encourage homeowners and other stakeholders to protect natural shorelines and embrace shoreline restoration when it can simultaneously protect properties, coastal populations, biodiversity, and retain ecosystem services.

Additional Information and Declarations

Competing Interests

The authors declare that they have no competing interests.

Author Contributions

Timothy S. Lee conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, wrote the paper, prepared figures and/or tables, reviewed drafts of the paper.

Jason D. Toft conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, wrote the paper, prepared figures and/or tables, reviewed drafts of the paper.

Jeffery R. Cordell conceived and designed the experiments, performed the experiments, wrote the paper, reviewed drafts of the paper.

Megan N. Dethier conceived and designed the experiments, performed the experiments, wrote the paper, reviewed drafts of the paper.

Jeffrey W. Adams conceived and designed the experiments, performed the experiments, wrote the paper, reviewed drafts of the paper.

Ryan P. Kelly conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, wrote the paper, prepared figures and/or tables, reviewed drafts of the paper.

Data Availability

The following information was supplied regarding data availability:

The raw datasets have been provided as Supplemental Dataset Files.

Funding

This work was funded by a grant from Washington Sea Grant, University of Washington, pursuant to National Oceanic and Atmospheric Administration Award No. NA14OAR4170078. The views expressed herein are those of the authors and do not necessarily reflect the views of NOAA or any of its sub-agencies. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

Adams J, Padgham B, Toft JD. 2015. Habitat Restoration: Bainbridge Island Nearshore. Tacoma: Puget Sound Partnership.

Angradi TR, Schweiger WE, Bolgrien DW, Ismert P, Selle T. 2004. Bank stabilization, riparian land use and the distribution of large woody debris in a regulated reach of the upper Missouri River, North Dakota, USA. River Research and Applications 20(7):829-846

Bell SS, Middlebrooks ML, Hall MO. 2014. The value of long-term assessment of restoration: support from a seagrass investigation. Restoration Ecology 22(3):304-310

Bilkovic DM, Mitchell MM. 2013. Ecological tradeoffs of stabilized salt marshes as a shoreline protection strategy: effects of artificial structures on macrobenthic assemblages. Ecological Engineering 61:469-481

Bilkovic DM, Mitchell MM, La Peyre MK, Toft JD. 2017. Living Shorelines: The Science and Management of Nature-Based Coastal Protection (First Edition). Boca Raton: CRC Press.

Bilkovic DM, Roggero MM. 2008. Effects of coastal development on nearshore estuarine nekton communities. Marine Ecology Progress Series 358:27-39

Bozek CM, Burdick DM. 2005. Impacts of seawalls on saltmarsh plant communities in the Great Bay Estuary, New Hampshire USA. Wetlands Ecology and Management 13(5):553-568

Brown SC, Smith K, Batzer DP. 1997. Macroinvertebrate responses to wetland restoration in northern New York. Environmental Entomology 26(5):1016-1024

Bulleri F, Chapman MG. 2010. The introduction of coastal infrastructure as a driver of change in marine environments. Journal of Applied Ecology 47(1):26-35

Bulleri F, Chapman MG, Underwood AJ. 2005. Intertidal assemblages on seawalls and vertical rocky shores in Sydney Harbour, Australia. Austral Ecology 30(6):655-667

Chapman MG. 2003. Paucity of mobile species on constructed seawalls: effects of urbanization on biodiversity. Marine Ecology Progress Series 264:21-29

Chapman MG, Roberts DE. 2004. Use of seagrass wrack in restoring disturbed Australian saltmarshes. Ecological Management and Restoration 5(3):183-190

Chapman MG, Underwood AJ. 2011. Evaluation of ecological engineering of ‘armoured’ shorelines to improve their value as habitat. Journal of Experimental Marine Biology and Ecology 400(1–2):302-313

Cohen J. 1992. A power primer. Psychological Bulletin 112(1):155-159

Cordell JR, Toft JD, Munsch S, Goff M. 2017. Benches, beaches, and bumps: how habitat monitoring and experimental science can inform urban seawall design. In: Bilkovic DM, Mitchell MM, La Peyre MK, Toft JD, eds. Living Shorelines: The Science and Management of Nature-Based Coastal Protection. Boca Raton: CRC Press. 421-438

Davis JL, Currin CA, O’Brien C, Raffenburg C, Davis A. 2015. Living shorelines: coastal resilience with a blue carbon benefit. PLOS ONE 10(11):e0142595

Davis JLD, Levin LA, Walther SM. 2002. Artificial armored shorelines: sites for open-coast species in a southern California Bay. Marine Biology 140(6):1249-1262

De Barba M, Waits LP, Genovesi P, Randi E, Chirichella R, Cetto E. 2009. Comparing opportunistic and systematic sampling methods for non-invasive genetic monitoring of a small translocated brown bear population. Journal of Applied Ecology 47(1):172-181

Dethier MN, Raymond WW, McBride AN, Toft JD, Cordell JR, Ogston AS, Heerhartz SM, Berry HD. 2016. Multiscale impacts of armoring on Salish Sea shorelines: evidence for cumulative and threshold effects. Estuarine, Coastal and Shelf Science 175:106-117

Dugan JE, Hubbard DM, McCrary MD, Pierson MO. 2003. The response of macrofauna communities and shorebirds to macrophyte wrack subsidies on exposed sandy beaches of southern California. Estuarine, Coastal, and Shelf Science 58S:25-40

Dugan JE, Hubbard DM, Rodil IF, Revell DL, Schroeter S. 2008. Ecological effects of coastal armoring on sandy beaches. Marine Ecology 29(s1):160-170

Eamer JBR, Walker IJ. 2010. Quantifying sand storage capacity of large woody debris on beaches using LiDAR. Geomorphology 118(1–2):33-47

Esteves LS, Williams JJ. 2017. Managed realignment in Europe: a synthesis of methods, achievements, and challenges. In: Bilkovic DM, Mitchell MM, La Peyre MK, Toft JD, eds. Living Shorelines: The Science and Management of Nature-Based Coastal Protection. Boca Raton: CRC Press. 157-182

Firth LB, Browne KA, Knights AM, Hawkins SJ, Nash R. 2016a. Eco-engineered rock pools: a concrete solution to biodiversity loss and urban sprawl in the marine environment. Environmental Research Letters 11(9):094015

Firth LB, Knights AM, Bridger D, Evans AJ, Mieszkowska N, Moore PJ, O’Connor NE, Sheehan EV, Thompson RC, Hawkins SJ. 2016b. Ocean sprawl: challenges and opportunities for biodiversity management in a changing world. Oceanography and Marine Biology: An Annual Review 54:193-269

Gittman RK, Fodrie FJ, Popowich AM, Keller DA, Bruno JF, Currin CA, Peterson CH, Piehler MF. 2015. Engineering away our natural defenses: an analysis of shoreline hardening in the US. Frontiers in Ecology and the Environment 13(6):301-307

Gittman RK, Peterson CH, Currin CA, Fodrie FJ, Piehler MF, Bruno JF. 2016a. Living shorelines can enhance the nursery role of threatened estuarine habitats. Ecological Applications 26(1):249-263

Gittman RK, Scyphers SB, Smith CS, Neylan IP, Grabowski JH. 2016b. Ecological consequences of shoreline hardening: a meta-analysis. BioScience 66(9):763-773

Harris C, Strayer DL, Findlay S. 2014. The ecology of freshwater wrack along natural and engineered Hudson River shorelines. Hydrobiologia 722(1):233-245

Hart DD, Johnson TE, Bushaw-Newton KL, Horwitz RJ, Bednarek AT, Charles DF, Kreeger DA, Velinsky DJ. 2002. Dam removal: challenges and opportunities for ecological research and river restoration. BioScience 52(8):669-682

Heatherington C, Bishop MJ. 2012. Spatial variation in the structure of mangrove forests with respect to seawalls. Marine and Freshwater Research 63(10):926-933

Heerhartz SM, Dethier MN, Toft JD, Cordell JR, Ogston AS. 2014. Effects of shoreline armoring on beach wrack subsidies to the nearshore ecotone in an estuarine fjord. Estuaries and Coasts 37(5):1256-1268

Heerhartz SM, Toft JD. 2015. Movement patterns and feeding behavior of juvenile salmon (Oncorhynchus Spp.) along armored and unarmored estuarine shorelines. Environmental Biology of Fishes 98(6):1501-1511

Hood WG. 2004. Indirect environmental effects of dikes on estuarine tidal channels: thinking outside of the dike for habitat restoration and monitoring. Estuaries 27(2):273-282

Jones K, Hanna E. 2004. Design and implementation of an ecological engineering approach to coastal restoration at Loyola Beach, Kleberg County, Texas. Ecological Engineering 22(4–5):249-261

Korsu K. 2004. Response of benthic invertebrates to disturbance from stream restoration: the importance of bryophytes. Hydrobiologia 523(1–3):37-45

La Peyre MK, Humphries AT, Casas SM, La Peyre JF. 2014. Temporal variation in development of ecosystem services from oyster reef restoration. Ecological Engineering 63:34-44

Lam NWY, Huang R, Chan BKK. 2009. Variations in intertidal assemblages and zonation patterns between vertical artificial seawalls and natural rocky shores: a case study from Victoria Harbour, Hong Kong. Zoological Studies 48(2):184-195

Lawless AS, Seitz RD. 2014. Effects of shoreline stabilization and environmental variables on benthic infaunal communities in the Lynnhaven River System of Chesapeake Bay. Journal of Experimental Marine Biology and Ecology 457:41-50

McGlone CM, Springer JD, Laughlin DC. 2009. Can pine forest restoration promote a diverse and abundant understory and simultaneously resist nonnative invasion? Forest Ecology and Management 258(12):2638-2646

McGranahan G, Balk D, Anderson B. 2007. The rising tide: assessing the risks of climate change and human settlements in low elevation coastal zones. Environment and Urbanization 19(1):17-37

Millennium Ecosystem Assessment (MEA). 2005. Ecosystems and Human Well-Being: Synthesis. Washington, D.C.: Island Press.

Miller SW, Budy P, Schmidt JC. 2010. Quantifying macroinvertebrate responses to in-stream habitat restoration: applications of meta-analysis to river restoration. Restoration Ecology 18(1):8-19

Moreira J, Chapman MG, Underwood AJ. 2006. Seawalls do not sustain viable populations of limpets. Marine Ecology Progress Series 322:179-188

Morgan PA, Short FT. 2002. Using functional trajectories to track constructed salt marsh development in the Great Bay Estuary, Maine/New Hampshire, U.S.A. Restoration Ecology 10(3):461-473

Morley SA, Toft JD, Hanson KM. 2012. Ecological effects of shoreline armoring on intertidal habitats of a Puget Sound urban estuary. Estuaries and Coasts 35(3):774-784

Munsch SH, Cordell JR, Toft JD. 2015. Effects of seawall armoring on juvenile Pacific salmon diets in an urban estuarine embayment. Marine Ecology Progress Series 535:213-229

Munsch SH, Cordell JR, Toft JD. 2016. Fine-scale habitat use and behavior of a nearshore fish community: nursery functions, predation avoidance, and spatiotemporal habitat partitioning. Marine Ecology Progress Series 557:1-15

Muotka T, Syrjänen J. 2007. Changes in habitat structure, benthic invertebrate density, trout populations and ecosystem processes in restored forest streams: a boreal perspective. Freshwater Biology 52(4):724-737

Neumann B, Vafeidis AT, Zimmermann J, Nicholls RJ. 2015. Future coastal population growth and exposure to sea-level rise and coastal flooding—a global assessment. PLOS ONE 10(3):e0118571

Palinkas LA, Horwitz SM, Green CA, Wisdom JP, Duan N, Hoagwood K. 2015. Purposeful sampling for qualitative data collection and analysis in mixed method implementation research. Administration and Policy in Mental Health and Mental Health Services Research 42(5):533-544

Patrick CJ, Weller DE, Ryder M. 2016. The relationship between shoreline armoring and adjacent submerged aquatic vegetation in Chesapeake Bay and nearby Atlantic coastal bays. Estuaries and Coasts 39(1):158-170

Peters JR, Yeager LA, Layman CA. 2015. Comparison of fish assemblages in restored and natural mangrove habitats along an urban shoreline. Bulletin of Marine Science 91(2):125-139

Peterson MS, Comyns BH, Hendon JR, Bond PJ, Duff GA. 2000. Habitat use by early life-history stages of fishes and crustaceans along a changing estuarine landscape: differences between natural and altered shoreline sites. Wetlands Ecology and Management 8:209-219

Puget Sound Partnership. 2016. Puget Sound partnership: shoreline armoring.

Raudys SJ, Jain AK. 1991. Small sample size effects in statistical pattern recognition: recommendations for practitioners. IEEE Transactions on Pattern Analysis and Machine Intelligence 13(3):252-264

Rhodes LD, Durkin C, Nance SL, Rice CA. 2006. Prevalence and analysis of Renibacterium salmoninarum infection among juvenile Chinook salmon Oncorhynchus tshawytscha in North Puget Sound. Diseases of Aquatic Organisms 71:179-190

Rosnow RL, Rosenthal R, Rubin DB. 2000. Contrasts and correlations in effect-size estimation. Psychological Science 11(6):446-453

Scyphers SB, Picou JS, Powers SP. 2014. Participatory conservation of coastal habitats: the importance of understanding homeowner decision making to mitigate cascading shoreline degradation. Conservation Letters 8(1):41-49

Seitz RD, Lipcius RN, Olmstead NH, Seebo MS, Lambert DM. 2006. Influence of shallow-water habitats and shoreline development on abundance, biomass, and diversity of benthic prey and predators in Chesapeake Bay. Marine Ecology Progress Series 326:11-27

Shipman H. 2001. Beach nourishment on Puget Sound: a review of existing projects and potential applications.

Simenstad C, Reed D, Ford M. 2006. When is restoration not? Incorporation landscape-scale processes to restore self-sustaining ecosystems in coastal wetland restoration. Ecological Engineering 26:27-39

Simenstad CA, Thom RM. 1996. Functional equivalency trajectories of the restored Gog-Le-Hi-Te estuarine wetland. Ecological Applications 6(1):38-56

Smith SM. 2007. Removal of salt-killed vegetation during tidal restoration of a New England salt marsh: effects on wrack movement and the establishment of native halophytes. Ecological Restoration 25(4):268-273

Sobocinski KL, Cordell JR, Simenstad CA. 2010. Effects of shoreline modifications on supratidal macroinvertebrate fauna on Puget Sound, Washington beaches. Estuaries and Coasts 33(3):699-711

Stegeman JJ, Solow AR. 2002. Environmental health and the coastal zone. Environmental Health Perspectives 110(11):A660-A661

Stewart GB, Bayliss HR, Showler DA, Sutherland WJ, Pullin AS. 2009. Effectiveness of engineered in-stream structure mitigation measures to increase salmonid abundance: a systematic review. Ecological Applications 19(4):931-941

Stewart TW, Downing JA. 2008. Macroinvertebrate communities and environmental conditions in recently constructed wetlands. Wetlands 28(1):141-150

Taylor JP, Smith LM, Haukos DA. 2006. Evaluation of woody plant restoration in the Middle Rio Grande: ten years after. Wetlands 26(4):1151-1160

Toft JD. 2016. Benthic Macroinvertebrate Monitoring at Seahurst Park 2015: Year 10 Post-Restoration of South Seawall Removal and Year 1 of North Seawall Removal. Burien: King Conservation District.

Toft JD, Cordell J, Levy C, Morgan E, Armbrust E. 2013a. Salmon Bay Natural Area: Restoration Aquatic Monitoring. Seattle: Seattle Public Utilities.

Toft JD, Ogston AS, Heerhartz SM, Cordell JR, Flemer EE. 2013b. Ecological response and physical stability of habitat enhancements along an urban armored shoreline. Ecological Engineering 57:97-108

Toft JD, Cordell JR, Armbrust EA. 2014. Shoreline armoring impacts and beach restoration effectiveness vary with elevation. Northwest Science 88(4):367-375

Underwood AJ. 1994. On beyond BACI: sampling designs that might reliably detect environmental disturbances. Ecological Applications 4(1):3-15

Warren RS, Fell PE, Rozsa R, Brawley AH, Orsted AC, Olson ET, Swamy V, Niering WA. 2002. Salt marsh restoration in Connecticut: 20 years of science and management. Restoration Ecology 10(3):497-513

Wensink SM, Tiegs SD. 2016. Shoreline hardening alters freshwater shoreline ecosystems. Freshwater Science 35(3):764-777

Zedler JB, Callaway JC. 1999. Tracking wetland restoration: do mitigation sites follow desired trajectories? Restoration Ecology 7(1):69-73