Morphological Performance of Vegetated and Non-Vegetated Coastal Dunes with Rocky and Geotextile Tube Cores under Storm Conditions

Abstract

Coastal dune restoration projects are increasingly seen as a means of protecting the coast against storms, especially with the use of natural (plants), artificial (geotextile tube cores) and hybrid (plants and geotextile tube) elements. As geotextile tube cores have been found to negatively affect dune and beach natural morphodynamics, rocky cores are thus seen as a potential alternative. We carried out laboratory experiments to compare the performance of dunes with rocky cores, geotextile cores, and both types when coupled with planted vegetation. We investigated these elements in the context of scaled mild, moderate, and intense storm conditions. The results showed that dunes with either type of core lost more sand than dunes without cores. The addition of plants generally reduced the erosion across the various options, but most strongly for rocky cores under mild and moderate storm conditions. We also found that dunes with a high density of plants were best for intense conditions. Overall, the use of rocky dune cores, when coupled with plants, is the most suitable and sustainable alternative option in hybrid engineering projects.

1. Introduction

Coastal dunes act as natural buffers and protect the coast from hydrometeorological phenomena [1,2,3]. During high-energy conditions, dunes dissipate wave energy, prevent erosion, and retard flooding [4]. Thus, there is growing interest in using coastal dunes as barriers to protect human settlements.

This strategy is especially appropriate on sandy beaches where human assets are exposed to high-energy waves and where natural dune formation is possible [1]. On beaches where dunes have been flattened by human activities (e.g., tourism, ports, resorts, etc.) or where there is a natural sediment deficit, dunes could be restored or rehabilitated to act as a barrier, or to support other protection measures, as part of Coastal Green Infrastructure projects [5]. In these areas, dunes could be enhanced with natural (plants), artificial (geotextile tube cores), or hybrid (plants with rocky or geotextile tubes cores) elements to improve their performance (e.g., Feagin [6] and Nordstrom [7]).

To better design and use dunes as barriers, a comprehensive understanding of wave-driven erosional processes is required. Most of the information related to these processes comes from wave flume experiments and numerical models. For example, van Rijn [8] described three main processes that control dune erosion (the generation of low-frequency effects, the production of turbulence due to wave breaking, and the failure of the dune face due to wave impact). Kobayashi et al. [9] and Figlus et al. [10] documented the effect of dune and berm geometry in mediating wave overwash and overtopping. Authors such as Van Gent et al. [11], van Thiel de Vries et al. [12] and D` Alessandro [13] showed that dune retreat and erosion increases as the wave period increases. More recently, researchers have documented the effect of plants on these processes (e.g., [14]), but information about the effect of dune cores, and dune cores in combination with plants, is limited.

Wave flume experiments have demonstrated that plants mitigate wave erosion, empirically [14,15,16,17,18,19]. During high-energy conditions, aboveground plant parts reduce the water level and velocity in the swash zone [20], the wave up-rush, and the undertow current [14]. On the other hand, roots increase sand cohesion and decrease erosion and scarp retreat rates [18,21]. Furthermore, Mendoza et al. [22] observed that plants can shift the wave-breaking point seaward. Feagin et al. [23] showed that shifting this location seaward can create a scarp, and then accelerate dune erosion during extreme storm events, although they also maintained that the vegetation was still critical in initially building the dune. Thus, the dune could protect landward properties and assets in a sacrificial manner. Consequently, the revegetation of dunes is becoming more attractive as means of increasing the resistance of dunes to erosion and protecting landward features. Nonetheless, the effectiveness of plants is time-dependent [24] and factors such as plant architecture and maturity, species, number of leaves, or density influence the amount of sand removed from the dune [18,21,24,25].

Recently, dunes have been constructed with cores that contain geotextile tubes to counteract the shortage of high-quality sand, and to reduce construction times and costs. Geotextile tubes (hereafter “geotextile cores”) consist of water permeable, sand-sealed geotextiles filled with sand or other granular material [26]. On this basis, the use of geotextile cores has been considered to provide more environmentally friendly dunes than other types of hard solutions [6,7,27]. Certainly, they are effective in preventing flooding and erosion, but there is evidence that they cause adverse effects on beach and dune evolution. For example, beaches with geotextile-core dunes are narrower, with steeper slopes and less sand volume [27]. Under these conditions, when the dunes erode, the wave reflection is increased. As a result, the shoaling distance is very short and wave breaking occurs earlier and more aggressively [28], reducing the system’s ability to recover from high-energy conditions. As a result, continuous sand nourishment is required to re-bury the geotextile cores and prevent their damage by wind, waves, and solar radiation [6,29,30,31]. Due to the fact that the cross-shore profile is interrupted by the presence of the artificial geotextile core, embryonic dune formation is also arrested [29], plant habitat zonation is affected (especially for pioneer species) [6], and animal nesting sites are lost. Finally, social concerns arise, as geotextile cores restrict access and affect the aesthetic and recreational value of the beach [6]. Consequently, some authors, such as Gibeaut et al. [29] and Harris and Ellis [27], have suggested that the performance of geotextile-cored dunes is closer to that of a hard solution than that provided by a natural beach dune system.

Rocky-cored dunes could be an alternative to address the problems associated with geotextile cores. In high-energy conditions, the rocky structure reduces the overwash and erosion, and retards dune destruction, especially in the final stages of the storm [32]. A potential advantage of the rocky cores is that they do not alter the morphodynamic response or the erosion regime, as noted by Sallenger, Jr. [4]. Also, the sediment exchange between the dune and the beach in this case is more similar to that found in natural conditions. Thus, the beach slope and the recovery of the system are unaffected, and embryonic dunes can continue to grow. In addition, rocky cores can improve the success of dune revegetation efforts because the porous core allows root penetration and water supply for plants to grow more vigorously. Examples of dunes and beaches enhanced with hybrid solutions (plants and cores) are found on beaches in the Mexican Caribbean, and Texas, Los Angeles, Florida, and New Jersey in the USA (see Feagin, [6]; Nordstrom, [7]; Harris and Ellis, [27]) as well as Figueira da Foz, Portugal [30].

As the effectiveness of hybrid dunes is still unknown, it is necessary to examine the morphological performance of the dunes enhanced with different elements. To the best of the authors’ knowledge, there has been no previous comparison between the performance of dunes with rocky cores versus and geotextile cores. There have also been no studies on how each of these two types of cores interact with vegetation. Moreover, there is little information about the performance of dunes with geotextile cores more broadly. Thus, we carried out new experiments to compare the hydrodynamic and geomorphological responses of beach-dune profiles enhanced with natural (plants), artificial (rocky and geotextile cores), and hybrid (plants with rocky or geotextile cores) elements under storm conditions.

2. Materials and Methods

Three physical experiments were carried out in a wave flume at the Engineering Institute of the National Autonomous University of Mexico (IIUNAM) to study the response of a beach-dune profile supported with plants and/or dune cores, under storm conditions. The results of two of the three experiments have been published previously. The first, Experiment 1, by Silva et al. [20], studied the response of vegetated dunes with different plant densities. The second, Experiment 2, by Odériz et al. [32], studied the response of vegetated dunes with a rocky core, and with both elements combined. Experiment 3 is presented in this paper, and is a study of the response of vegetated dunes with a geotextile core, and with both elements combined. Because the three experiments were conducted in the same facility, using the same beach-dune geometry and the same sediment, their results can be compared. In all cases, a Froude’s Similarity Law was followed, with a linear scale factor of 20 used as a reference for the waves, rock, and geotextile tests. The sand and plants could not be scaled.

2.1. Experimental Set-Up

The wave flume at the IIUNAM is 0.8 m wide, 1.2 m high, and 37 m long. It is equipped with a piston-type wavemaker, featuring a dynamic wave absorption system. The wave flume was divided longitudinally into two sections of 0.395 m for the three experiments. The division was made with a 1 cm thick acrylic plate from x = 21 to 29.1 m (Figure 1a), allowing two profiles to be tested simultaneously. A gravel passive absorption system was placed at the other end of the flume.

The beach-dune profiles were built with sediment from a beach in the Gulf of Mexico (Tuxpan, Veracruz). This sediment is classified as fine sand mainly composed of silica, with 6% fine material, a mean diameter (D₅₀) of 0.142 mm, and a uniformity coefficient (C_u) of 0.142 [20,32]. Sand chemical composition was out of the scope of this paper and then it was not analysed, however, it does not alter our results.

2.2. Beach-Dune Profile

Experiment 1 was carried out with two beach-dune profiles, the main differences being the presence/absence of a berm, and dune slope and height. This experiment had two profiles, A (with berm) and B (without berm). In the present research, we have only used the results of profile A, as this is the profile that was used in Experiments 2 and 3. Profile A consisted of a submerged zone, a horizontal berm, and a narrow dune [20]. The submerged part of the profile had a gentle slope (1:32) from the bottom of the flume to a depth of 8.5 cm below the still water level (SWL = 0.5 m), starting at x = 15, followed by a second slope of 1:7 beginning at x = 26.06 m, 2 cm above the SWL [20]. From there, a horizontal berm, 35 cm long, extended to the toe of the berm, where the dune face rose to 22 cm above the SWL, with a 1:2.25 slope [20]. The back of the dune had a 1:1.36 slope, which descended again to reach the SWL, giving a total dune base width of 75 cm. Behind the dune, a horizontal section of 1.2 m was left (Figure 1b). It is worth noting that the profiles of the three experiments could not be replicated identically due to the technical conditions of the facility. In particular, in Experiment 2 the dune was higher and had a greater volume of sand.

After shaping the profiles, the dunes were exposed to a three-minute mild wave train (T_p = 1.0 s and H_s = 0.05 m) to achieve sand saturation. Following this, significant alterations were rectified to match the original slope, and the experiments were conducted.

2.3. Plant Propagation

Plants of Ipomoea pes-caprae (L.), commonly known as Beach Morning Glory, were used in the three experiments. This pantropical trailing vine populates the backshore due to its high resistance to salinity, solar radiation, and burial [33]. I. pes-caprae plays a significant role in coastal geomorphology as it aids in sand accretion and the formation of embryo dunes beyond the high tide line [33].

The plant samples for the experiments were reproduced from vegetative fragments that were cut from individuals of I. pes-caprae thriving at El Farallon and Playa Paraíso, two beaches situated on the central coast of the Gulf of Mexico, in Veracruz. The fragments were cut with growing meristem, roots, and leaves to guarantee successful plant establishment. The plant fragments were positioned in plastic bags containing sand from the dunes, and then transferred to a greenhouse situated in La Mancha, Veracruz, at the CICOLMA field station of the Institute of Ecology, A. C. The plants were then watered every three days and after three months, they were transported to Mexico City and transferred to a greenhouse to recover and be ready for the experiments. Finally, prior to every plant test, the plants and sand were carefully extracted from the bag and planted in previously prepared holes on the dune in the wave flume.

2.4. Dunes with Core

The gravel used for the dune in Experiment 2 had a D₅₀ grain size of 18.28 mm. It was arranged in a pyramidal form to enable the layer of sand above it to replicate the dune from Experiment 1. The bottom length of the gravel core was a 0.58 m, and situated 0.45 m above the SWL. The core slopes were built parallel to the dune faces, and the sand cover above them was 0.06 m thick (Figure 1c). In turn, the geotextile core in Experiment 3 consisted of a geotextile filled with the same sand found on the beach and dune. Geotextile was covered by a layer of sand of ±0.2 m.

2.5. Instrumentation

In the three experiments, the free water surface elevation was recorded with eleven wave gauges (WG). WG 1 was placed 14 m away from the wave paddle, while the other ten (WG 2:11) were installed on the centre line of each profile as shown in Figure 1a. Three WG were installed according to the methodology of Baquerizo [34], to separate incident and reflected waves, and two were installed in the surf zone (Figure 1a).

2.6. Experimental Program

Table 1. Experimental program for the three experiments.

Experiment	Storm Condition	Plant Density	Number of Plants	Core	ID	Hs (m)	Tp (s)	Duration (s)
Experiment 1 (plant densities)	S1	N	0	-	S1N	0.1	1.118	900
		L	6	-	S1L	0.1	1.118	900
		M	10	-	S1M	0.1	1.118	900
		H	15	-	S1H	0.1	1.118	900
	S2	N	0	-	S2N	0.1	1.5652	900
		L	6	-	S2L	0.1	1.5652	900
		M	10	-	S2M	0.1	1.5652	900
		H	15	-	S2H	0.1	1.5652	900
	S3	N	0	-	S3N	0.15	2.012	240
		L	6	-	S3L	0.15	2.012	240
		M	10	-	S3M	0.15	2.012	240
		H	15	-	S3H	0.15	2.012	240
		Plant location		Rocky
Experiment 2 (rocky core and plant position)	S1	VN	0	R	S1VNR	0.082	1.565	900
		VT	10	R	S1VTR	0.088	1.565	900
		VL	10	R	S1VLR	0.094	1.565	900
	S2	VN	0	R	S2VNR	0.082	1.789	900
		VT	10	R	S2VTR	0.085	1.789	900
		VL	10	R	S2VLR	0.082	1.789	900
	S3	VN	0	R	S3VNR	0.98	2.012	522–718
		VT	10	R	S3VTR	0.083	2.012	522–718
		VL	10	R	S3VLR	0.082	2.012	522–718
	S1	VN	0	-	S1VN	0.082	1.565	900
		VT	12	-	S1VT	0.088	1.565	900
		VL	12	-	S1VL	0.094	1.565	900
	S2	VN	0	-	S2VN	0.082	1.789	900
		VT	12	-	S2VT	0.085	1.789	900
		VL	12	-	S2VL	0.082	1.789	900
	S3	VN	0	-	S3VN	0.98	2.012	522–718
		VT	12	-	S3VT	0.083	2.012	522–718
		VL	12		S3VL	0.082	2.012	522–718
		Plant density		Geotextile
Experiment 3 (geotextile core)	S1	N	0	G	S1G	0.1	1.565	900
		H	12	G	S1GP	0.1	1.565	900
		H	12	-	S1P	0.1	1.565	900
	S2	N	0	G	S2G	0.1	1.788	900
		H	12	G	S2GP	0.1	1.788	900
		H	12	-	S2P	0.1	1.788	900
	S3	N	0	G	S3G	0.1	2.012	240
		H	12	G	S3GP	0.1	2.012	240
		H	12	-	S3P	0.1	2.012	240

Note: In storm condition S1 = mild, S2 = moderate and S3 = intense; in plant density N = non vegetated, L = low, M = medium and H = high; in plant location VN = void of vegetation, VT = vegetation on the exposed and protected face of the dune, and VL = vegetation on the protected face of the dune; in core R = Rocky core and G = geotextile core.

shows the 39 tests performed across the three experiments. In each experiment, three storm intensities were tested: mild (S1), moderate (S2) and intense (S3). Experiment 1 involved 12 tests with four plant densities, namely non-vegetated (N), low (L), medium and high (H) vegetated. The specific plant numbers are shown in Table 1. Experiment 2 included 18 tests with dunes devoid of vegetation (VN), dunes with plants on both sides, exposed and protected faces (VT), and dunes with plants on the leeside (VL), combined with a rocky core (R) and without a core. Experiment 3 comprises dunes with high-density plants (P), dunes with a geotextile core (G), and dunes with high-density plants and a geotextile core (PG). Note that the high plant density values in Experiment 3 are the same as those of Experiment 1. The dune devoid of vegetation without a core in Experiment 2 is equivalent to the non-vegetated dune from Experiment 1. In addition, in Experiments 2 and 3, S1 and S2 storm conditions (H_s, T_p, and duration) were slightly different than those of Experiment 1.

The waves generated in all the experiments corresponded to the JONSWAP spectrum (γ = 3.3). The SWL was set at 0.45 m and the storm surge at 0.05 m (Figure 1b).

2.7. Beach-Dune Profile Response and Wave Reflection

The morphological impact of vegetation and dune cores on beach and dune profiles in storm conditions was evaluated using three key parameters: the proportion of sand volume eroded from the dune, the displacement of the coastline, and wave reflection. To assess the morphological response, we recorded the pre- and post-storm profiles using a Leica^® total station. This system measured the sand surface within the wave flume and had a precision of 0.5 × 10⁻³ m. The outputs were processed and analysed in R Studio (R v. 2.23 Wooden Christmas-Tree; [35]). Then, in accordance with Sallenger, Jr. [4], the erosion regimes were categorized as Swash (run-up restricted to the foreshore), Collision (run-up colliding with the base of the foredune), Overwash (run-up surpassing the height of the dune) or Inundation (the dune being continuously and fully submerged).

We also computed the area, volume, and proportion of sand lost from two sections of the cross-shore profile: section A, encompassing only the dune (exposed and protected face), and section B, which includes the exposed and protected face of the dune and the portion behind the dune. As previously mentioned, it was not possible to replicate the profiles identically, and therefore, this research does not cover the evolution of the submerged part of the beach profile, and no results are presented regarding this section.

The change in the location where the water line intersected the slope was calculated by subtracting the initial intersection of the SWL with the profile in the X direction from the final intersection. Positive values therefore indicate beach regression, whereas negative values indicate beach progradation. The separation of incident and reflected waves was performed using the method of Baquerizo [34]. Thus, the reflection coefficient ($C_r$) was computed using Equation (1):

$$C_r=\frac{H_r}{H_i}$$

(1)

where H_r is the reflected wave height and H_i is the incident wave height.

3. Results

3.1. Erosion Regime

Overall, collision dominated the erosion regime for the S1 storm in the three experiments (Figure 2a–c). For S2 and S3 (Figure 2d–i), in Experiments 2 and 3, collision mostly developed into overwash, causing dune flattening due to wave overtopping. As can be seen in Figure 2, in S1, S2, and S3, most of the sediment removed from the dune was deposited at the toe, the berm disappeared, and a new slope was produced. This slope helped to dissipate wave energy in the final stages of the storm and consequently, erosion at the highest part of the profile ceased.

The details of the erosion regime documented for the 39 dune conditions evaluated in the three experiments are described in the following section. First, we describe the trends in each experiment and then compare the beach responses of each.

In Experiment 1, a collision regime was observed in 11 of the 12 tests conducted on dunes with different plant densities. Only in S3N (the dune devoid of vegetation in storm intensity 3) was overwash documented, and the waves transported the sediment in a landward direction. It is worth noting that among all of the 39 tests, the dunes in Experiment 1 were the most resistant to overwash. These dunes were scarped on the exposed side, but were not flattened (Figure 2 a,d,g). This could be attributed to the lower T_p used in S1 and S2 of Experiment 1 (T_p = 1.118 and 1.5652, respectively) as compared that used in Experiments 2 and 3 (T_p = 1.118 and 1.5652, respectively).

In contrast, for Experiment 2 (Figure 2b,e,h), we observed that in S1, all of the dune conditions exhibited a collision regime, but the highest waves reached the crest of the dune and slightly eroded it, transporting the sand to the back portion of the dune (Figure 2b). In S2, with dunes with rocky cores (S2VNR, S2VLR, and S2VTR), we observed a collision regime, whereas the dunes lacking a core (S2VN, S2VL, and S2VT) were overwashed and suffered severe damage, especially when the vegetation was restricted to the protected face of the dune (Figure 2e). Similarly, in S3, all the dunes were overwashed, and the effects of wave attack increased in the absence of a rocky core and plants on the seaside of the dune (Figure 2h). In Experiment 3, for S1, all of the dunes showed a collision regime (Figure 2c), while for S2 they were overwashed (Figure 2f). Both conditions showed that the dunes with plants were the most resistant, followed by the dunes with plants and a geotextile core, and finally by the dunes with a geotextile core but no plants. In S3, all the dunes were flattened, but no continuous inundation was recorded (Figure 2i). S3G had the most erosion-resistant profile, while S2P had the most flattened profile.

The comparison of the impact of both cores (Experiment 2 and Experiment 3) on the erosion regime gave the following outcomes. In S1, S1VNR had a collision regime (Figure 2b), while S1G was overwashed (Figure 2c). S1VLR, S1VTR, and S1GP also had the same morphological response. In S2, S2VNR (Figure 2e) and S2G (Figure 2f) underwent overwashing, but in the former, only the highest waves reached the crest of the dune and scarped it, while in the latter a continuous slope appeared due to sand deposition at the toe of the dune. No differences were seen in the morphological response of S2VLR, S2VTR, and S2GP, although erosion increased greatly in the vegetated dunes (S2VL and S2VT, Figure 2e), where the rocky core was absent, especially when vegetation was restricted to the protected face of the dune (S2VL). In S3, no significative differences in the erosion regime were observed with the dunes with the core only, nor when they had plants (Figure 2h,i).

3.2. Dune Erosion

The percentages of sand erosion from zones A (dune) and B (dune and area behind the dune) for the three storm intensities (S1, S2, and S3) were next compared (

Table 2. Dune erosion percentages in zone A (the seaward and landward face of the dune) at the end of the three storm intensities.

Experiment	Dune Condition (ID)		Dune Initial Volume (m3)	ID	Post-Storm Volume (m3)			Sand Erosion (%)
	Plant Density				S1	S2	S3	S1	S2	S3
Experiment 1 (plant densities)	None (N)		0.0565	N	0.0543	0.0516	0.0504	3.8392	8.6098	10.7669
	Low (L)			L	0.0544	0.0541	0.0527	3.8045	4.3054	6.7065
	Medium (M)			M	0.0542	0.0530	0.0526	4.1163	6.2201	6.8641
	High (H)			H	0.0551	0.0540	0.0534	2.5320	4.3440	5.4364
Experiment 2 (rocky core and plant position)	Plant location	Structure
	None (VN)	Rocky	0.0662	VNR	0.0605	0.0581	0.0569	8.6339	12.2139	13.9741
	Top (VT)	Rocky		VTR	0.0578	0.0548	0.0559	12.6614	17.1962	15.5218
	Landward (VL)	Rocky		VLR	0.0577	0.0558	0.0533	12.7957	15.7157	19.5531
	None (VN)	No structure		VN	0.0601	0.0577	0.0541	9.2306	12.8404	18.2851
	Top (VT)	No structure		VT	0.0557	0.0534	0.0503	15.8537	19.3181	23.9834
	Landward (VL)	No structure		VL	0.0561	0.0483	0.0492	15.2048	27.0898	25.6653
Experiment 3 (geotextile core)	Plant density	Structure
	None	Geotextile	0.0588	G	0.0514	0.0514	0.0560	12.5861	12.6029	4.8104
	High	Geotextile		GP	0.0557	0.0545	0.0499	5.2806	7.3967	15.1220
	High	No structure		P	0.0563	0.0558	0.0518	4.2730	5.1631	11.9143

). It is important to note that (i) the dunes used in Experiment 2 were slightly higher than those in Experiments 1 and 2, and (ii) the wave periods for S1 and S2 were longer in Experiments 2 and 3, so a higher percentage of erosion is expected in Experiment 2.

In zone A of Experiment 1, the highest density of plants (S1H) provided strong protection against dune erosion (2.53% of the total volume was eroded). The protective effect of the plants lessened as the plant density fell to medium (S1M eroded 4.11%) and low levels (S1L eroded 3.80%). It should be noted that the dunes without any vegetation (S1N eroded 3.83%) eroded less than those with medium density (Table 2). Moreover, in all three experiments, the dunes with plants were the most resistant to erosion in all the storms. In Experiment 1, the erosion percentage ranged from 2.53 to 4.11%, while in Experiments 2 and 3, it ranged from 4.2 to 15.8% (Figure 2 and Table 2). In Experiment 2, for S1, the dunes without plants were the most resistant (only 8.63% of sand eroded), regardless of the presence or absence of the rocky core (Table 2). However, among the vegetated dunes, those with rocky cores had the lowest erosion, particularly when the plants were on the seaward side of the dune (Table 2). In Experiment 3, S1P (4.27% of erosion) and S1GP (5.28% of erosion) had the most resistance, while S1G eroded the most (12.58%). Under this storm condition, S1G showed greater erosion than S1VNR (8.63% of erosion). It was also found that the dunes with geotextile core and plants (S1GP) were more effective in preventing erosion than those with a rocky core and plants (S1VTR and S1VTL) (Table 2). There were no significant differences in the results of zone B (

Table 3. Dune erosion percentages in zone B (the seaward and landward face of the dune, and the area behind the dune) at the end of the three storm intensities.

Experiment	Dune Condition (ID)		Dune Initial Volume (m3)	ID	Post-Storm Volume (m3)			Sand Erosion (%)
	Plant Density				S1	S2	S3	S1	S2	S3
Experiment 1 (plant densities)	None (N)		0.0700	N	0.0679	0.0652	0.0687	3.0071	6.8985	1.9265
	Low (L)			L	0.0678	0.0676	0.0663	3.1389	3.4111	5.3135
	Medium (M)			M	0.0676	0.0665	0.0658	3.3903	4.9558	5.9992
	High (H)			H	0.0686	0.0676	0.0669	2.0567	3.4446	4.4402
Experiment 2 (rocky core and plant position)	Plant location	Structure
	None (VN)	Rocky	0.0838	VNR	0.0779	0.0756	0.0745	7.0831	9.7463	11.1397
	Top (VT)	Rocky		VTR	0.0749	0.0721	0.0739	10.6102	13.9974	11.7996
	Landward (VL)	Rocky		VLR	0.0747	0.0731	0.0731	10.8400	12.8226	12.7379
	None (VN)	No structure		VN	0.0764	0.0742	0.0740	8.8094	11.4072	11.7407
	Top (VT)	No structure		VT	0.0719	0.0729	0.0707	14.2095	12.9926	15.6140
	Landward (VL)	No structure		VR	0.0749	0.0685	0.0690	10.6698	18.2425	17.6968
Experiment 3 (geotextile core)	Plant density	Structure
	None	Geotextile	0.0722	G	0.0658	0.0647	0.0698	8.8682	10.3591	3.3289
	High	Geotextile		GP	0.0689	0.0677	0.0645	4.5873	6.2832	10.6873
	High	No structure		P	0.0702	0.0692	0.0653	2.7271	4.1086	9.5248

).

In zone A for S2 of Experiment 1, the S2H and S2L conditions proved the most successful in protecting the dune from erosion (4.34 and 4.30% of erosion, respectively), whereas most erosion occurred in S2N (8.60% of erosion). In the experiments concerning the impact of a rocky core (experiment 2), S2VN and S2VNR (12.84 and 12.21% of erosion, respectively) had the least sand loss. In this experiment, for dunes with plants, the presence of a rocky core mitigated erosion, especially when the plants were on the seaside of the dune. In Experiment 3, S2G was the most eroded, while S2P was the most resistant, with erosion values similar to those found in Experiment 1. On comparing both cores, S2G was slightly more eroded than S2VNR; however, the geotextile core was more effective in the presence of plants than the rocky core and plants. Once again, under these storm conditions, no significant differences were observed in the sand loss proportions between zones A and B.

In Experiment 1 for storm S3, there were differences between zones A and B (Table 2 and Table 3). In zone A, S3H was the most efficient (5.43% of erosion), while S3N was the most eroded (10.76% of erosion). However, in zone B, we found that most of the sediment removed from S3N was deposited landward, while for the vegetated dune the sand travelled seawards (Figure 2i and Table 3). On the other hand, in Experiment 2, the dunes lacking vegetation were highly resistant to erosion, and resistance was greater when the rocky structure was present. Furthermore, on the vegetated dunes, erosion considerably increased when the rocky core was absent. The most erosive conditions on the beach were in the condition without a rocky core with plants on the leeside of the dune. These trends in Experiment 2 were almost the same for section B. In Experiment 3 during the most intense storm, the dune with geotextile core and without plants was the most resistant (only 4% of erosion). For dunes with plants and dunes with both plants and geotextile cores, the erosion increased (11.91 and 15.12% of erosion, respectively). Finally, in S3 dunes with both geotubes and plants, they performed similarly to dunes with rocky cores and plants on the seaside. However, the dunes with rocky cores and plants on the protected face eroded even more (15.12%, 15.52 and 19.55% of erosion, respectively).

3.3. Coastline Displacement

Figure 3a–c show the shoreline displacement found for each condition. In 38 of the 39 conditions, the beach and dune profile moved seaward. The profile displacement was directly correlated to the storm intensity, except for S1G where the shoreline advanced landward.

No linear trends or relations between the dune erosion and profile displacement could be established (Figure 3a–c). In S1, the most eroded dunes (S1VNR, S1VTR, S1VLR, S1VN, S1VT, S1VL, and S1G) showed less displacement than those that were less eroded (S1N, S1L, S1M, S2H). In S2 and S3, among all the plant densities, the dunes had less sand loss than those with the most seaward displacement of the profile. In these cases, the sediment was transported to the submerged part of the profile, or landward.

3.4. Wave Reflection

The sand loss was directly correlated to wave reflection across all S1, S2 and S3 storm intensities (Figure 3d–f). However, no clear trends were found for each individual experiment, and the general trend could be due to the differences in T_p between the experiments. In Experiments 2 and 3, the wave period was greater than in Experiment 1 and the wave reflection coefficients increased and removed more sand volume. However, these observations could not be attributed to the plants or cores.

In S1, the dunes with different plant densities had the lowest values of wave reflection (Figure 3d–f) and lower values of erosion (Table 1). The dune with a rocky core and no plants (VNR) and the dune with geotextile and no plants (G) had almost the same wave reflection coefficients, but erosion was greater in the first. Dunes with both geotextile cores and plants (GP) had a lower wave reflection coefficient and less erosion in comparison with dunes with both rocky cores and vegetation (VTR).

For S2, there was no clear correlation between sand loss and the reflection coefficient. In contrast to S1, wave reflection increased considerably among the dunes with different plant densities; however, for the dunes with less erosion (H and L), they were no greater in terms of their reflection coefficients. The dunes with either rocky cores or geotextile cores, but without plants, performed similarly. Yet, the dunes with geotextile cores and plants (GP) were more eroded than the dunes with rocky cores and plants (VLR and VTR). Finally, when considering the three experiments under the S3 storm condition, there was a general trend of increasing erosion and wave reflection, but it was not as clear as for S1. Dunes with geotextile cores and no plants (G) had a greater value of wave reflection but lower erosion than their counterparts with rocky cores and no plants (VNR). Finally, dunes with geotextile cores and plants (GP) had lower wave reflection coefficients than dunes with rocky cores and plants on the exposed face of the dune (VTR), though they had the same erosion percentage. Only the dunes with both rocky cores and plants on their leeward side eroded more.

4. Discussion

Coastal dunes are important reservoirs of sediment and regulate the morphological response of beaches in high energy conditions caused by storms. We investigated the resistance of dunes to erosion, when they were enhanced with natural solutions (plants), artificial solutions (rocky and geotextile cores), and hybrid solutions (plants with rocky or geotextile cores) that are commonly used in engineering for dune rehabilitation (particularly geotextile cores), for example on beaches in Yucatán (Mexico), Florida and Texas (USA).

4.1. The Effect of Plants in Abating Erosion

The results demonstrate that plants were the most effective element for reducing erosion, as compared to the use of rocky or geotextile cores alone. For all of the plant densities that we studied, the erosion on the seaward side of the dune was reduced and dune overwashing was either prevented or retarded. This finding is important because when overwashing occurs, erosion rates increase and dunes become more severely damaged [32]. These findings coincide with those of several previous studies, showing that plants generally mitigate erosion (e.g., Kobayashi, [14]; Sigren et al. [15]; Charbonneau et al. [25,36]; Bryant et al. [19]; De Battisti and Griffin [18]). Other studies have demonstrated that the effectiveness of plants in mitigating beach erosion depends on factors such as the plant species used, for example Charbonneau et al. [25] showed that plants that grow near the sediment bed are more effective. Similarly, Innocenti et al. [24] and Feagin et al. [37] showed that plant flexibility and architecture can determine how effective vegetation could be in attenuating wave erosion.

This study demonstrates that the effect of plants on dune resistance to erosion can also be more complex. We observed that the density of the plants does not correlate particularly well with dune erosion. Similarly, in a recent 1:1 scale experiment, Feagin et al. [23] demonstrated that plants can actually increase erosion during extreme storms, mainly by causing the creation of irregular scarps and micro-slopes. This sequence of events is related to an increase in wave reflection caused by biomass, meaning that more energy is available to remove sediment from the dunes or beach [20]. Still, Feagin et al. [23] also found that adding vegetation helps the dune grow in height before the storm, thus providing a protective barrier for landward locations. Coastal managers must therefore consider many sources of information when designing effective strategies. However, it is clear that adding plants to a dune provides greater protection for landward locations, whether by reducing the erosion on the dune itself, or by enhancing its protective function.

4.2. The Performance of a Rocky versus a Geotextile Core

In general, we observed greater erosion for the dunes with cores, whether those cores were rocky or geotextile. This result may be related to the longer wave periods and duration of Experiments 2 and 3. For Experiments 2 and 3, the reflection coefficients were greater, the dunes were overwashed, and the percentages of sand loss were considerably higher. However, the experiments with comparing these two core types were still similar in terms of the hydrodynamic conditions; these direct comparisons showed that for S1 and S2, the rocky core performed better than the geotextile core (and wave reflection was similar in both experiments). In S3, the geotextile core reduced erosion slightly more. Both cores were effective in preventing the complete destruction of the dune, and their effectiveness was better when plants were present, especially on the seaward side.

The results were somewhat expected. In the literature, Feagin et al. [23,38] and Gibeaut [29] both found that the geotextile cores altered the erosion regimes established by Sallenger, Jr [4]. However, in the experiments discussed in this study, we observed that despite the differences in the percentage of sand eroded, both cores showed the same erosion regime. This is important because it means that the geotextile cores can be substituted by rocky cores, thus helping to avoid some of the serious problems associated with the use of geotextile cores. Rocky cores can (i) allow the penetration of roots and facilitate plant establishment and fitness, (ii) promote a more natural dune system in which embryo dunes can develop and habitat zonation from plants is not affected, (iii) avoid the added economic costs of sand nourishment required to cover the geotextile material, or repair damage to it from solar radiation or wave attack, (iii) facilitate dune recovery in calm conditions and protect the coast in subsequent storms, and (iv) avoid pollution by polyethylene terephthalate (PET) geotextile fibres [39]. However, we recommend new research to validate these hypotheses be conducted.

Finally, the results show that both rocky and geotextile cores increased erosion in comparison with dunes that did not have these cores. Still, in many cases these types of solutions are necessary elements for projects where human assets are exposed. Here, we show that in such locations, that it is also beneficial to cover them with plants to minimize these negative effects.

4.3. Caveats of the Study

It is important to note that the sediment grain size and plants could not be scaled. However, the experiments allowed us to explore the qualitative and general response of the dunes by controlling variables that are otherwise difficult to manipulate under field conditions.

5. Conclusions

To the best of the authors’ knowledge, this is the first time that the performance of rocky versus geotextile cores has been tested. Our results demonstrate that rocky cores are a suitable option to substitute geotextile cores for increasing dune resistance to wave attack. Our results show that the use of plants could further increase the resistance of both core types. However, the effectiveness depends on the location of the plants and the storm intensity. It is important to note that plants cannot protect the dunes under all types of conditions, and this fact must be understood by stakeholders, coastal engineers, and the general public. Consequently, we recommend the use of ecologically enhanced, hybrid engineering solutions on sites where the use of purely natural solutions are not feasible. In summary, rocky cores appear to be a comparatively sustainable option when coupled with vegetation.

Due to the limitation in the scaling of some biophysical processes, although this work provides new information, in order to have design parameters it is important to perform 1:1 tests in the laboratory and/or in the field.