**John M. Montgomery \*, Karin R. Bryan, Erik M. Horstman and Julia C. Mullarney**

Faculty of Science and Engineering, School of Science, University of Waikato, Private Bag 3105, Hamilton 3240, New Zealand; karin.bryan@waikato.ac.nz (K.R.B.); e.m.horstman@utwente.nl (E.M.H.); julia.mullarney@waikato.ac.nz (J.C.M.)

**\*** Correspondence: jmontgom31@gmail.com; Tel.: +64-021-036-7446

Received: 29 June 2018; Accepted: 17 August 2018; Published: 23 August 2018

**Abstract:** Mangroves have been suggested as an eco-defense strategy to dissipate tsunamis, storm surges, and king tides. As such, efforts have increased to replant forests along coasts that are vulnerable to flooding. The leafy canopies, stems, and aboveground root structures of mangroves limit water exchange across a forest, reducing flood amplitudes. The attenuation of long waves in mangroves was measured using cross-shore transects of pressure sensors in two contrasting environments in New Zealand, both characterized by mono-specific cultures of grey mangroves (*Avicennia marina*) and approximate cross-shore widths of 1 km. The first site, in the Firth of Thames, was characterized by mangrove trees with heights between 0.5 and 3 m, and pneumatophore roots with an average height of 0.2 m, and no substantial tidal drainage channels. Attenuation was measured during storm surge conditions. In this environment, the tidal and surge currents had no alternative pathway than to be forced into the high-drag mangrove vegetation. Observations showed that much of the dissipation occurred at the seaward fringe of the forest, with an average attenuation rate of 0.24 m/km across the forest width. The second site, in Tauranga harbor, was characterized by shorter mangroves between 0.3 and 1.2 m in height and deeply incised drainage channels. No attenuation of the flood tidal wave across the mangrove forest was measurable. Instead, flow preferentially propagated along the unvegetated low-drag channels, reaching the back of the forest much more efficiently than in the Firth of Thames. Our observations from sites with the same vegetation type suggest that mangrove properties are important to long wave dissipation only if water transport through the vegetation is a dominant mechanism of fluid transport. Therefore, realistic predictions of potential coastal protection should be made prior to extensive replanting efforts.

**Keywords:** eco-defense; coastal defense; coastal morphodynamics; mangroves; flood attenuation; natural defense

### **1. Introduction**

Mangroves are the dominant species of vegetation in many tropical and sub-tropical intertidal environments. These salt-tolerant trees provide a valuable habitat for a range of animal species, reduce hydrodynamic forces, promote sedimentation, and provide protection from floods [1]. Additionally, mangroves are significantly more efficient than many terrestrial ecosystems at sequestering carbon [2]. Mangroves thrive in the zone between mean sea-level and high water and thus are sensitive to changes in inundation regime. Their zonation and ability to prevent erosion or increase sedimentation may provide a mechanism for mangroves to adapt to sea level rise and alleviate the threat of coastal retreat [3]. Despite the diverse array of valuable services, worldwide mangrove populations are in steep decline, with the loss of over one-quarter of global mangrove cover since 1980 [1,4].

Extreme flooding events are projected to increase with sea level rise [5,6]. Additionally, coastal populations and infrastructure are increasing [7], driving demand for effective coastal protection. Conventional engineering solutions are often costly and may have a limited lifespan, destroy or fragment sensitive habitat, and have been associated with enhanced erosion [8,9]. Coastal vegetation has been proposed as an alternative to hard engineering solutions. Mangroves can provide coastal protection by reducing storm waves, dissipating currents, and stabilizing sediments [10,11]. Additionally, sedimentation in mangrove forests may provide a mechanism to maintain present coastlines with respect to sea level rise [12].

The reduction in the wave height of short period wind-generated waves due to interaction with mangroves is well established [10,13,14]. Less well established are the protective benefits of mangroves with respect to storm surge [15–17]. Mangroves reduce peak flood levels by limiting fluid exchange across the forest [18]. Dissipation of storm surges through coastal vegetation has previously been quantified as a reduction in peak water level (cm) per distance of flood propagation (km) with values categorized by vegetation type [15–18]. Although providing an easily accessible solution, using fixed dissipation rates over wide-ranging sites may oversimplify flood protection provided by coastal vegetation.

Alongi [12] noted that flood protection provided by coastal vegetation is dependent on vegetation properties, local bathymetry, and storm parameters. At forest-wide scales applicable to coastal inundation issues, obtaining mangrove properties is problematic. Vegetation can be heterogeneously distributed [19], and quantifying the drag-inducing elements (leaves, stems, trunks, and pneumatophores) can be unwieldy. Several different summary statistics are used for large-scale hydrodynamics, including frontal area density, the proportion of volume occupied by the solid canopy, and the blockage factor [20,21]. However, Nepf [21] comments that at reach scales in vegetated rivers, the patch distribution plays a larger role in determining flow resistance than individual plant geometry. Typically, vegetation drag is large relative to bed drag and therefore in heterogeneously vegetated environments, flow is channelized and deflected away from vegetation/high-drag patches [21,22].

The influence of channelization on mangrove flood attenuation is explored through the comparison of high water events in two contrasting New Zealand mangrove forests. The study sites are similar in length, with the forest extending ~1 km in the direction of flood propagation, and both sites are comprised of the same mangrove species, *Avicennia marina* var. *australasica* [23]. The key distinction is that the Tauranga mangrove forest is highly channelized in comparison with the Firth of Thames site.

#### **2. Materials and Methods**

#### *2.1. Study Sites*

#### 2.1.1. Firth of Thames

The Firth of Thames (FoT) is a ~800 km<sup>2</sup> estuary on New Zealand's North Island (37◦12 S, 175◦27 E) (Figure 1b). The mesotidal estuary has a spring tidal range of 2.8 m and due to a shallow bed slope and plentiful fine-sediment supply, a large intertidal mudflat has developed [24]. The basin is bounded to the east and west by mountain ranges and the Hauraki Plains to the South. A stopbank (visible as a diagonal track on Figure 2a) prevents the inundation of the Hauraki Plains to the south of the Firth. The basin is exposed to moderate waves from the North and subject to a high terrigenous sediment supply from the Waihou and Piako rivers. The southern boundary of the Firth is colonized by a 1 km wide forest of grey mangroves (*Avicennia marina* var. *australasica*).

**Figure 1.** (**a**) The North Island of New Zealand with a panel (**b**) outlined; (**b**) Section of North Island of New Zealand showing the proximity of the Firth of Thames and Tauranga mangrove sites.

**Figure 2.** (**a**) The Firth of Thames study site with 9 (#1 seaward–#9 landward) instrument locations noted and bathymetry survey transect (brown). Station #1 is on the unvegetated mudflat, station #2 is in the vegetation fringe, stations #3–#5 are in the gently sloping intertidal, and stations #6–#9 are on the intertidal flat; (**b**) The Tauranga study site with 8 instrument locations noted with channel thalweg bathymetry survey (brown) and central mangrove survey (green). Station #A is on the vegetation fringe on the seaward boundary, stations #C and #D are in the western channel, stations #B, #E, and #H are in the central mangrove forest, stations #D and #F are in the eastern channel.

The cross-shore profile of the vegetated region (Figure 3a) consists of a level mangrove forest ~1.7–1.9 m above mean sea level (MSL) extending ~800 m seaward of the stopbank [24]. The sloping vegetation extends an additional ~100 m seaward to the mudflat. The topography and forest characteristics are relatively homogenous in the longshore direction. The elevation of the seaward fringe of the forest is close to a mean high water neap tide level (0.98 m MSL), so the tidal prism within the forest is relatively small and no substantial creeks have developed (Figure 3a) [25].

Mangrove characteristics vary throughout the forest. Along the forest fringe, trees are characterized by open spreading forms (Figure 4a,b). Within the forest, trees tend to have straight vertical trunks (Figure 4c). Tree height ranges from 0.5 to 3.5 m. Dense pneumatophores, as many as ~500 m<sup>−</sup>2, emerge from the bed up to 25 cm in height and ~1 cm in diameter (Table 1).

In November 2016, a supermoon and low-pressure event occurred to produce an unusually large flood event in the Firth of Thames (Figures 4d and 5a,b). The water levels reached 2.36 m above MSL, corresponding to an event with a ~10-year return period for the Firth of Thames. The study area was flooded for several tidal cycles prior to the peak high water and remained flooded for several tidal cycles after the peak water level.

**Figure 3.** (**a**) The elevation profile along the instrument transect in the Firth of Thames. Mean High Water Spring (MHWS) and Mean High Water Neap (MHWN) are noted. The main forest is higher than normal tidal levels and therefore no drainage channels have been scoured by tidal water flow; (**b**) the Tauranga RTK survey of transect through central mangrove forest (green) and thalweg (brown). High Water Spring (HWS) and High Water Neap (HWN) are marked (blue). The semi-diurnal tidally-driven flow through the forest is responsible for channelization at the site.

**Figure 4.** The images of Firth of the Thames study site. (**a**) Forest fringe at low tide; (**b**) Fringe at mid tide, trees are characterized with open spreading branches; (**c**) Interior mangrove forest with two researchers for scale. Trees are tall with vertical trunks; (**d**) Two researchers in the mangrove forest during the flood event.

**Figure 5.** (**a**) The Firth of Thames water level at each instrument station for five consecutive spring tidal cycles. The upper intertidal flat (stations #6–#9) did not fully drain for several tidal cycles; (**b**) Firth of Thames water level during maximum inundation event; (**c**) Tauranga water level for five consecutive spring tidal cycles. Note that only station #G was submerged at low tide; (**d**) Tauranga water level during the largest tidal cycle.

#### 2.1.2. Tauranga Site

Tauranga harbor is a 200 km2 barrier-enclosed lagoon on the North Island of New Zealand (37◦39 S, 176◦ E) (Figure 1). The mesotidal estuary has an average spring tidal range of 1.62 m and neap range of 1.24 m [26]. Due to the complexity of the estuary, exact tidal ranges are location-dependent [27]. The shallow lagoon, with an average depth of 3 m at low tide, has extensive intertidal areas that make up nearly 2/3 of the estuary area [28]. The estuary has two entrances and is comprised of many sub-estuarine basins. The mangroves in Tauranga have expanded rapidly, from 13 hectares in the 1940s to 168 hectares in 1999 [29]. The mangroves in Tauranga are at the southern boundary of their latitudinal range, which causes the forests to be less productive and the trees to be shorter [23]. The focus of the presented work is a basin north of Pahoia (Figure 1) that nearly drains at low tide.

The Pahoia field site is comprised of a ~1 km long intertidal mangrove forest that occupies ~2/3 of the basin surface area (Figure 2b). Two unvegetated steep-sided channels, along the eastern and western sides of the forest, maintain a near-uniform depth throughout the study site and dominate water flow into the area (Figures 3b and 6b). The western channel bifurcates around a ~300–400 m wide central mangrove platform. The vegetated regions are at the same elevation as high-water neap tidal levels and are approximately flat. A small creek drains into the western channel and further divides the central mangrove forest. The significant tidal prism in the forest is likely responsible for creating the channel network [3].

**Figure 6.** (**a**) The Firth of Thames LiDAR devoid of a channel network. Patchy higher elevations likely indicate a vegetation canopy; (**b**) Tauranga LiDAR data. Deep, incised channels and a level vegetated intertidal characterize the site. High elevation along channels displays a dense mangrove canopy. The color bar shows an elevation scale for both subplots.

The forest is comprised of small shrub-like grey mangroves less than 1.2 m in height (average 0.41 m). Individual trees have complex geometry (Figure 7a) and present a low dense canopy (Figure 7b). The pneumatophore density averages 75 per square meter, with individual pencil-roots of similar dimension to the pneumatophores in the Firth of Thames (Table 2).

**Figure 7.** The images of the Tauranga study site. (**a**) Example mangrove tree. Vegetation is characterized by a complex trunk structure and a low canopy height; (**b**) Mangrove forest at mid tide; (**c**) Mangrove lined channel with steep densely vegetated banks; (**d**) Weather station recording barometric pressure and wind speed during high spring tide with canopy almost submerged.

Typical spring tides nearly fully submerge the Pahoia mangrove forest (Figure 7d). Figure 5c,d displays data from a series of spring tides in June 2017. The peak water level reached ~1.25 m above MSL. Note that the study site nearly drained at low tide, leaving most of the instruments exposed.
