Coastal environments and coastal communities are vulnerable to climate change impacts such as sea level rise, higher rates of erosion, and more intensified storms. These effects will increase flooding and expose more inhabited land to the ocean, damaging livelihoods, infrastructure, and ecological systems. Historically, issues of erosion and flooding have been mitigated in populated areas using “hard” infrastructure such as seawalls, groins, jetties, and other concrete- or steel-based structures (Smallegan et al., 2016; Morris et al., 2018). Eroding beaches along more natural coasts have been managed with nourishments as well as hard structures to encourage sand accumulation (van Duin et al., 2004; Stive et al., 2013; de Schipper et al., 2020). These traditional measures are being challenged by the uncertain impacts of climate change both in terms of sea level rise and weather patterns, as hard infrastructure is difficult to design or adapt to uncertain climate futures (Sutton-Grier et al., 2015). Nature-based solutions (NBS) offer alternative coastal resilience measures which have the potential to adapt to a changing climate.

NBS leverage aspects of the natural environment to buffer against negative effects of climate change. Coastal examples include efforts to restore marshes and mangrove forests (Quartel et al., 2007; Howard et al., 2017), and designs to deploy oyster reefs near urban areas (Morris et al., 2018). These examples attempt to build natural barriers of vegetation and benthic organisms between ocean forces (e.g., storm waves) and coastal communities. The physical premise for the effectiveness of these NBS is reasonably well-understood. Research into seagrass and mangrove forests shows that their effectiveness at damping wave energy comes from forcing complex vortical structures in the landward flow (Quartel et al., 2007; Crooks et al., 2011; Howard et al., 2017; Bergstrom et al., 2019; Menendez et al., 2020). Similarly, submerged oyster reefs and even submerged artificial reefs operate under the principle of enhanced bed roughness, increasing the friction between the bed and incoming waves or surge and reducing the energy of the flow before it reaches the shoreline (Morris et al., 2018). NBS are organic, living entities and as such have the potential to adjust to, and mitigate, climate change effects like sea level rise and storm intensity.

These natural types of NBS require a significant amount of space to be effective: mangrove forests and marshes demand shoreline real estate and the research into the footprint of an oyster reef needed to fully damp a storm wavefield is ongoing, but likely in the range of hundreds of meters to kilometers. In urban environments, the necessary real estate may not be readily available. This presents both a challenge and an opportunity to develop new types of NBS which are engineered to incorporate natural features in infrastructure designed for urban environments and flexible enough to evolve as the climate changes. The research presented here focuses on one such engineered NBS, the Emerald Tutu, as a model for how these types of solutions may operate and be evaluated for effective deployment in urban environments.

The Emerald Tutu consists of floating vegetated mats that are interlinked and anchored to the seafloor to provide a modular network of protective canopies encircling urban coastlines (Figure 1). The Tutu was developed through collaboration between landscape architects, coastal engineers, climate scientists, and water resource specialists to provide flood mitigation as well as ecological and social co-benefits to coastal communities. Floating wetlands similar to the Tutu have been used thus far as individual elements in quiescent water bodies such as wastewater treatment plants, rivers, and lakes to improve water quality (Faulwetter et al., 2011; Hwang and LePage, 2011; Bi et al., 2019). Isolated vegetated canopies seeded with emergent and submerged vegetation have also been tested in laboratory conditions, showing that vegetated canopies have the potential to damp incident wave energy (Chen et al., 2016; Zhu et al., 2021). The Tutu extends the floating canopy concept to increase robustness for exposed coastal conditions, and adds the element of network effects to damp storm wave energy.

The modular, floating nature of the Tutu design makes it ideal for deployment in areas where land real estate is difficult to use for coastal resilience. Its structural reliance on vegetation provides benefits to native plants and animals. However, the efficacy of the Tutu, or any similarly engineered NBS, needs to be carefully evaluated to determine its optimal design and deployment. At present, there are few site-specific examples of NBS performance and few studies which dive into the physical underpinnings of NBS operations to the extent that coastal scientists could use them to improve existing numerical models or create new models which accurately predict and quantify NBS effects (Chen et al., 2016; Sarabi et al., 2019; Chausson et al., 2020; Zhu et al., 2021). The Emerald Tutu is an opportunity to begin to bridge this knowledge gap between NBS ideation and performance.

In this paper, we report on initial tests of the physical effectiveness of the Tutu system. The objective of the study was to evaluate the efficacy of an interlinked network of floating mats in attenuating nearshore waves. We designed and built scaled physical analogues of the Tutu and observed their behavior under a variety of wave conditions in the Natural Hazards Engineering Research Infrastructure (NHERI) O.H. Hinsdale Wave Basin at Oregon State University. In particular, we tested the hypothesis that the attenuation of waves owing to a network of Tutu mats is a function of mat network geometry, wavelength, and mat density. Our results show the existence of a relationship between network geometry and incident wavelength, as opposed to network density and incident wave conditions. Further, our results support the idea that a truly effective NBS requires a significant footprint along either the shoreline or the nearshore environment, suggesting that approaches to NBS which incorporate multiple, tiered solutions may be optimal for urban coastal communities in the face of climate change.

These lab experiments show the potential for a modular network of floating mats to reduce incident wave energy. We see a strong relationship between the wave steepness and the attenuation, where steeper waves are more attenuated than shallower ones (Figure 7A). We divide this into wavelength and wave height (Figure 7B) to better understand this relationship, and find that wavelength is a primary indicator of wave attenuation. The test conditions with both regular waves and random waves suggest that the geometry of the network relative to the incident wavelength drives wave attenuation. The majority of the wave attenuation occurs at wavelengths shorter than or equal to the length of individual mats, though wave energy continues to be reduced until the wavelengths grow beyond the cross-shore footprint of the entire mat network. Further, the results show that wave attenuation is not appreciably different between sparse and dense network layouts (Figure 7, squares versus circles). Instead, the size of individual mats and the network footprint relative to the wavelength (Figure 7, red dashed lines) have a stronger impact on wave attenuation than the separation between mats.

The mechanics of wave orbital velocity explains some of this dependence. The dissipation of wave velocity in the vertical water column scales with wavelength, with shorter (longer) wavelengths losing energy at shallower (deeper) water depths. A generalization from linear wave theory describes the depth of the wave velocity dissipation as half of the wavelength. The Tutu network is suspended at the surface with a draft of about 0.3 m maximum, thus we would expect the Tutu canopies to impact the shortest waves more than longer waves. To explain the dependence relative to the network footprint (Figure 7B, red dashed lines), we note that if a wave is shorter than the diameter of the Tutu mats then the crest and trough of the wave encounter a mat simultaneously which leads to dissipation. If the wave is longer than an individual mat, but shorter than the network, its crest and trough encounters two or more mats linked together, and energy dissipates owing to tension between them as the wave moves the mats in opposition. If the wave is longer than the entire network, however, the network rides along the wave and dissipation is negligible.

Our initial hypothesis was that the network spacing–sparse and dense–would be a significant factor in wave attenuation, with the denser cluster of mats dampening more wave energy than the sparse cluster. Our results show that there is not a significant difference between the mat spacing in the network for these two configurations, suggesting that the mechanism of mats moving in opposition to dissipate wave energy can work even if that tension is transferred on a rope between them. We cannot generalize to larger spacings between mats than tested here, though we anticipate that at the extreme a Tutu design with too much space between mats will not be in enough tension to dissipate energy. The comparisons between sparse and dense networks, however, do suggest that a field deployment of the Tutu could attenuate waves while allowing space for light to penetrate the water (a concern for ecological systems in the water column and in the benthic region of shallow areas) and space for small craft to move between the network (for maintenance and leisure).

Whether or not the wave energy dissipates at the initial row of mats or after several rows appears to depend on wave height rather than wavelength (Figure 8A). The spatial variance of wave energy within the mat network, calculated using the pressure sensors, shows how much the waves change in the direction of wave propagation. A higher (lower) variance indicates that the wave energy changes within (at the first row of) the network. In this experiment, we found that smaller incident waves had little energy variance within the mat network, suggesting that the first row of mats dissipate the bulk of the energy. Higher energy waves often had more energy variance within the mat network, suggesting that the waves change as they moved through the network. The wavelength does not appear to be as relevant as the wave height in determining where within the network the waves dissipate. Wavelength indicates whether dissipation will happen at all, and wave height directly correlates to the wave velocity (Figure 8B). the amount of energy in a wave, and thus likely the number of mat rows the wave needs to encounter for that energy to attenuate.

The results presented here are relevant for dense floating mats with minimal, short, stiff “emergent vegetation” represented by the rain screen, but not for mats with submerged vegetation. We found that the strips of tulle sewn to the bottom were not influencing the wave attenuation. We tested this by flipping the mats and re-running the Dense network regular waves, which produced similar results. At most 12 bunches of tulle were sewn onto each mat, and the bunches extended down to less than 20% of the water depth, so it is perhaps unsurprising that the effect was small. Observations from cameras placed underwater around the mats showed that the tulle was moving freely with the waves, and thus had little rigidity to contribute to energy attenuation. However, a concurrent full-scale prototype of a Tutu mat deployed in East Boston in summer 2021 showed dense vegetation undergrowth after 2 months in the water, which would have translated to a column of vegetation at least a meter deep in our wave basin experiments. Previous studies with kelp suggest that this type of submerged vegetation will enhance the dissipation of waves. These studies measured and modeled the dissipation of waves by suspended canopies of submerged kelp (fixed to a rigid platform above the water surface) (Zhu et al., 2021; Zhu et al., 2022) and suggest that the wave attenuation would improve with more realistic submerged vegetation structures, possibly extending the wavelengths at which the scaled Tutu is effective.

Our results build on previous work studying wave attenuation by floating individual canopies, notably experiments simulating floating shellfish farms in the field (Zhu et al., 2020). These experiments focused on dense buoys tethered individually to the sea floor and subjected to random wave attack. Attenuation measured from these tests using random waves are similar to Figure 9, with suspended canopies attenuating waves at higher frequencies better than at lower frequencies and plateauing at high frequencies. The plateau evident in Figure 9A occurs at wave frequencies greater than 1 Hz, which for a basin water depth of 1.36 m translates to a wavelength of 1.5m, which is also the diameter of our mats. The plateau appears to occur when the wavelength is at or smaller than the mat diameter. Zhu et al. (2020) further shows that the number of floating elements within the canopy influences attenuation, with more floating elements providing greater attenuation. This suggests that adding more mats to a Tutu network will improve performance at longer wavelengths.

In particular, Zhu et al. (2020) test the potential of combining floating shellfish farms (buoyant elements on individual tethers) with submerged seagrass or other vegetation patches anchored to the seafloor. Their results suggest that the floating farms can attenuate short waves and the submerged vegetation can attenuate longer waves (whose wave orbital velocities are more likely to penetrate the water column to the bed). This supports the idea that a tiered NBS approach to coastal protection should be used to protect against storm waves, with floating elements tackling higher wave frequencies and submerged elements tackling lower wave frequencies. The results presented above suggest that we can extend the frequency range attenuated by floating elements through interconnections between dense mats and network geometry. Future work will investigate how dense, suspended undergrowth from the mats impacts wave attenuation, and will likely expand the range of wave frequencies that a floating structure such as the Emerald Tutu can dissipate.

We tested the efficacy of a new NBS called the Emerald Tutu to attenuate waves in nearshore environments. Physical analogues of the mats were tested in the OH Hinsdale Wave Basin at Oregon State University for both regular and random wave conditions. Results from these tests show that wave attenuation depends on the geometry of the mat system, notably mat size and network size relative to the wavelength. These results are in line with previous work studying wave attenuation in the field owing to individually anchored floating shellfish farms. Both results suggest that floating nearshore networks of mats have the potential to protect the coastline from wave attack. Future research into the Tutu will include testing the impact of submerged vegetation, suspended from the mats, on additional wave attenuation, and developing a non-hydrostatic numerical model to replicate these lab results and scale them to the field.