Introduction

Densely populated coastal regions face a serious threat from sea-level rise due to climate change (Dasgupta and Laplante 2009). The Intergovernmental Panel on Climate Change (IPCC) special report on Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation (SREX) has highlighted with high confidence that in the absence of adaptation, “locations currently experiencing adverse impacts such as coastal erosion and inundation will continue to do so in the future due to increasing sea levels” (IPCC 2012).

The median value of projected sea level rise for the year 2100 is estimated to be between 0.44 and 0.74 m for process-based projections and between 0.32 and 1.24 m for semi-empirical projections (IPCC 2013). For Vietnam, MONRE (2009, 2012 quoted in Thanh 2014) projected mean sea level rise ranging from 0.51 to 0.99 m by the year 2100. Vietnam is considered to be one of the countries that is most vulnerable to the impacts of climate change, in particular to floods, storms and sea level rise. The Mekong Delta is particularly vulnerable due to its topography, population density and economic importance (ADB 2013; ISPONRE 2009; Boateng 2012). Without adaptation, 1.0 m of sea level rise would inundate about 5.3 % (17,423 km2) of Vietnam’s total land area (IMHEN 2010a, quoted in UN 2012) and threaten about 39 % of the Mekong Delta (MONRE 2009, 2012 quoted in Thanh 2014).

Along the mangrove-mud coasts of the Mekong Delta, erosion (or the threat of erosion) and hence flooding is affecting the life of thousands, often poor farmers and fishers (Ghimire et al. 2015). The traditional response of building dykes and seawalls does not work on the soft soils of these mud coasts. To the contrary, such construction activities can be counterproductive because they disturb the sediment balance even more, and this option is generally too expensive to protect large rural areas (Temmerman et al. 2013). Hard structures often fail because the subsoil of mangrove-mud coasts and foreshores is too soft and prone to consolidation and base failure. The required soil improvement or bed protection is either too expensive or insufficient depending on the thickness of the soft mud layer (Fowler 1989; Hartlén 1996).

On coasts with low-lying floodplains covered with either marsh or a mangrove belt, the floodplain is an important stabilising element of the coastal protection system (Gedan et al. 2011). It decreases the incoming wave energy (Anderson et al. 2011; McIvor et al. 2012) and thus protects against erosion and flooding. The higher the floodplain, the greater its wave dissipation capacity, leading to significant decreases in the wave load on the dyke. In the presence of mangroves, the wave reduction effect is even larger. Decreased height and length of waves leads to a shortened wave run-up and thus the required dyke height and construction costs are correspondingly lower.

Sustainable coastal protection strategies need to be developed for the low-lying areas of the Mekong Delta. In this case, it is important to include a diverse and site-specific range of approaches to ensure that adaptation conflicts, maladaptations, or path dependencies can be avoided. Individual adaptation options, developed in isolation based on inconsistent strategies and lacking a vision for a preferred future, may create adaptation conflicts and ultimately lead to a reduction in adaptive capacity (Smith et al. 2013).

A sustainable and effective coastal protection strategy comprises a holistic approach which considers the floodplains or mangroves in front of the dyke as part of the coastal protection system. Therefore, floodplain and mangrove management are important components of effective area coastal protection. The design and height of the dyke should be optimised considering various boundary conditions such as the influence of floodplains and mangroves. However, at sites where the mangrove forests have been degraded or destroyed due to unsustainable use of natural resources and development in the coastal zone (Primavera 1997; Alongi 2002; Winterwerp et al. 2005; Thampanya et al. 2006) and the foreshore has been eroded, such a strategy can only be implemented following restoration of the eroded floodplains (Winterwerp et al. 2005; Gedan et al. 2011; Schmitt et al. 2013). This will re-establish site conditions conducive to mangrove forest growth.

This study focusses on optimising dyke design as part of area coastal protection strategy and on bamboo T-fences for the restoration of eroded floodplains as a precondition for natural regeneration of mangroves forests. It also discusses the importance of participatory involvement of local communities in the protection and management of mangroves. This focus on the combination of different coastal protection elements into a holistic strategy combines elements described in previous studies about ecological, hydrological, morphodynamic and socio-economic aspects of mangrove rehabilitation (Lewis 2005, 2009; Primavera and Esteban 2008; Winterwerp et al. 2005, 2013).

The aim of this paper is to present a holistic approach to area coastal protection which addresses the threats of erosion, flooding and storms along the mud-coast of the Mekong Delta, a threat which will be increased by the impacts of climate change such as rising sea levels and the increasing intensity and frequency of storm surges. It also addresses design of protection measures based on the specific local context and boundary conditions, limiting factors such as the load bearing capacity of the subsoil, and the involvement of local communities in mangrove co-management/governance as part of the holistic approach.

The study area

Human activities have influenced the natural processes of the coastal areas in the Mekong Delta since the beginning of the eighteenth century (Le Anh Tuan et al. 2007). Today, the coastal areas of the lower Mekong Delta are therefore cultural landscapes that have been shaped by technical activities such as dyke construction and drainage measures (Biggs 2010).

Soc Trang Province (Fig. 1) is one of 13 provinces in the Mekong Delta region with a population of 1,285,096. It covers an area of 331,176 ha, 62 % of which are used for agriculture, about 3.5 % for forestry and over 16 % for aquaculture (2008 figures from Soc Trang Statistics Office 2010). The coastline is 72 km long and subject to a dynamic process of accretion and erosion driven by the discharge regime of the Mekong River, the tides of the South China Sea (Vietnamese East Sea) and coastal long-shore currents. Accretion can reach up to 68 m per year (Cu Lao Dung) while about 11 km of coastline around Vinh Tan and My Thanh is currently subject to erosion of up to 30 m per year. In some places, the earth dyke that protects the hinterland from flooding is facing severe erosion endangering the people and farmland directly behind the dyke (Pham et al. 2009; Joffre 2010; Pham 2011; Schmitt and Albers 2014).

The hamlet of Mo O in Trung Binh Commune in Tran De District is located in Soc Trang Province north of the My Tanh River (Fig. 1). The dyke assessed in this study is an earth dyke about 1800 m in front of the national dyke at the mouth of My Tanh River. It protects about 45 ha of agricultural land from flooding.

Fig. 1
figure 1

Map of the study area and other sites mentioned in the text. (Color figure online)

Full size image

The frequency of high water overtopping the dyke at Mo O and flooding the fields in the hinterland has increased in recent years. The dyke height has been increased as an adaptation measure using material from just in front of the dyke. This extraction led to the creation of a channel, where the current velocities and even the wave heights at the dyke increased. As a result, erosion of the dyke increased dramatically. The outer slope of the dyke as well as parts of the dyke body eroded during recent high water events. Emergency measures (vertical poles, plastic foil, and sand bags) have failed to stop the erosion. At some locations the width of the dyke has decreased appreciable, and its stability is severely threatened. Complete dyke failure of a larger section is an imminent danger.

Methods

A holistic approach to dyke design was carried out for a dyke section along the coast of the hamlet Mo O in Soc Trang Province (Fig. 1). The existing dyke structure in Mo O, the mangrove forest belt and general hydrodynamic loads, including water levels and waves, were considered to create an adapted and optimized dyke design.

The hydrodynamic forces were assessed based on available data. The site-specific boundary conditions, including the condition and extent of mangrove forest, the dyke geometry and soil parameters, were determined through an on-site levelling survey, a geotechnical survey and a mangrove health assessment.

Data from the My Thanh stream gauge were used to calculate a design water level (data source: Southern Institute for Water Resources Research, Ho Chi Minh City, Vietnam). An extreme value prediction based on these data was carried out using the Pearson III, Log Pearson III and Gumbel distribution functions (Gumbel 1958; Todd and Walton 2000). Hydrodynamic parameters and boundary conditions were derived from previous studies along the coast of Soc Trang Province (Albers and von Lieberman 2011).

To calculate the design dyke height, a flood event with a recurrence interval of 5 years was chosen. To generate an optimised dyke design, the required freeboard (wave run-up height plus safety margin) was calculated using the critical values of overtopping water volume before the structural stability of the dyke is endangered. For most of the year, the dyke must be high enough to prevent flooding of the agricultural area. However, in the case of extreme events, a limited amount of overtopping can occur.

To assess the design wave height at Mo O, a numerical model simulating significant wave height setups for the coast of Soc Trang Province was used (Albers 2011). Based on a significant wave height (H s ) of 0.65 m, a peak period (T P ) of 5.5 s and a wave length (L) of 16.70 m for Mo O and using the Longuet-Higgins theory (EAK 2002), assuming an event with 1000 waves, the maximum wave height is calculated as:

$$H_{max} = 1. 8 6 \times H_{s} ,\quad {\text{hence}}\;H_{max} = 1. 8 6 \times 0. 6 5 = 1. 20 9\,{\text{m}}$$
(1)

The reduction rate (r) can be calculated as:

$$r = \frac{{H_{I} - H_{T} }}{{H_{I} }}\,\,\left( {{{\text{m}} \mathord{\left/ {\vphantom {{\text{m}} {\text{m}}}} \right. \kern-0pt} {\text{m}}}} \right)$$
(2)

where H I  = the initial wave height in front of the mangrove forest; here H max is applied, H T  = transmitted wave height after x meters of mangrove forest.

The transmitted wave heights H T have been calculated based on Eq. (2) using four wave reduction rates for four separate locations spaced 1, 40, 180, 290 and 350 m. Different reduction rates r have been applied. Wave reduction rates as a function of per meter mangrove forest vary widely depending on the location:

  • 0.001–0.004 per meter for Vinh Quang coast, Vietnam (Mazda et al. 2006)

  • 0.004–0.012 per meter for Red River Delta, Vietnam (Quartel et al. 2007)

  • 0.002–0.014 per meter for Katang/Pailan, Thailand (Horstman et al. 2012)

To determine the design dyke height, the wave overtopping rate was calculated based on the equation given in EAK (2002):

$$q = 0.038\gamma_{b} \sqrt {2gH_{T}^{3} } \frac{\tan \alpha }{{\sqrt {{{H_{T} } \mathord{\left/ {\vphantom {{H_{T} } L}} \right. \kern-0pt} L}} }}\exp \left( { - 3.7\frac{{R_{C} }}{{H_{T} }}\frac{{\sqrt {{{H_{T} } \mathord{\left/ {\vphantom {{H_{T} } L}} \right. \kern-0pt} L}} }}{\tan \alpha }\frac{1}{{\gamma_{b} \gamma_{f} \gamma_{\theta } }}} \right)\,\left( {{{{\text{m}}^{3} } \mathord{\left/ {\vphantom {{{\text{m}}^{3} } {{\text{m}}\,{\text{s}}}}} \right. \kern-0pt} {{\text{m}}\,{\text{s}}}}} \right)$$
(3)

where L = wave length (m), R C  = freeboard (m), γ f = empirical coefficient for the influence of the slope roughness [−], dyke surface covered with grass γ f = 0.9, \(\gamma_{\theta }\)  = empirical coefficient for the influence of the direction of the wave approach [−], perpendicular wave approach chosen \(\gamma_{\theta }\) = 1.0, γ b  = 1.0 empirical coefficient for the influence of a berm in the dyke profile [–], profile with no berm γ b  = 1.0, H T  = transmitted wave height (m), α = angle of the dyke slope on the seaward side, and g = gravity 9.81 (m/s2).

Equation (3) was used to calculate wave overtopping rates for different transmitted wave heights, dyke slopes and freeboard heights. Different wave attenuation rates (and relevant wave parameters) as well as various dyke slopes and heights were considered to assess an optimised dyke profile. For every combination of parameters, an assessment was carried out to check if the wave overtopping rate was lower than the critical overtopping rate.

The results of the calculations were used to derive the required freeboard (the difference between design water level and dyke crest) and the dyke slope. Critical values of average wave overtopping were assessed (CEM 2002).

To complete the design process, the design dyke profiles for the dyke section in Mo O were checked with regard to the geotechnical stability of the subsoil using the following parameters (EAK 2002):

  • Bearing capacity (Deutsches Institut für Normung 2006)

  • Hydraulic failure (Deutsches Institut für Normung 2010)

  • Sliding (Deutsches Institut für Normung 2010)

  • Embankment breakage according to the method of slices by Bishop (1955)

  • Settlement estimation (Deutsches Institut für Normung 2015)

The results of the holistic dyke design approach for Mo O were used, together with results of previous studies (Albers and von Lieberman 2011), for the development of a floodplain management approach for Vinh Tan Commune in Soc Trang Province. During the previous studies, data on bathymetry, topography, water levels, river discharges, sediment loads and characteristics were collected and assessed. Missing data about bathymetry, water levels, sediment loads and especially wave climate were recorded through field measurements. These data were also used to verify the results of numerical models and to understand the hydrodynamic and morphodynamic processes in the focus area. The effectiveness of conventional breakwater constructions as well as different designs using local materials was tested in a wave flume (Albers and von Lieberman 2011).

A comprehensive monitoring programme was initiated comprising wave measurements to quantify the wave damping effect of the bamboo fences during various storm and tidal conditions, measurements of mud density and mud elevation as well as shoreline changes (Albers et al. 2013). Mangrove monitoring was carried out by taking fixed-point photos at regular intervals.

Results

The results section is divided into three parts. The first part introduces the results of an improved and adapted dyke design at Mo O hamlet based on the design water level and wave height in relation to the width and health of the existing mangrove belt in front of the dyke. The second part looks at floodplain management in Soc Trang Province at sites where erosion has destroyed the mangrove forest in front of the dyke and has eroded the foreshore. It is based on previous work and provides results on how T-fences can be used to restore eroded floodplains. The third part shows the effects of floodplain restoration on mangrove regeneration. Mangrove management and protection, as an important element of coastal protection, is covered in the discussion section.

Improved and adapted dyke design at Mo O

The adapted dyke design requires a design water level with a certain recurrence interval. Figure 2 shows the recurrence intervals of different water levels based on the Pearson III, Log Pearson III and the Gumbel distribution functions. Due to short-term statistical series, the margin of error increases along with increasing recurrence intervals. For the chosen design event with a recurrence interval of 5 years, and using the Pearson III distribution function, the design water level is +2.29 m with regard to the Vietnamese national datum plane.

Fig. 2
figure 2

Extreme value prediction for water level

Full size image

Table 1 summarises transmitted wave heights H T in front of the dyke at Mo O attenuated by the mangrove forest and considering different reduction rates r for four separate locations spaced 1, 40, 180, 290 and 350 m. The initial wave height at the seaward end of the mangroves forest was H max  = 1.209 m according to Eq. (1).

Table 1 Transmitted wave heights (m) for different wave reduction rates (r) and widths of the mangrove forest (m)
Full size table

With increasing mangrove forest width and therefore decreasing wave heights directly in front of the dyke, the wave overtopping also decreases. The same effect is observed for a flatter dyke slope. The flatter the slope, the lower the wave overtopping. In the study area, the slope of the dyke is limited by the space between the existing dyke and the mangrove forest (~7 m) as well as the required dyke height.

The required dyke heights were determined based on the calculated freeboard. The design dyke height is the sum of the design water level of +2.29 m and the required freeboard. If a wave attenuation factor of r = 0.002 is applied and the width of the mangrove forest belt is 350 m, it results in a transmitted wave height of 0.363 m. If a dyke slope of 1:3 is assumed and the freeboard is set 0.90 m no damage occurs with reference to CEM (2002). Table 2 shows the design dyke heights, current dyke heights and required heightening ∆h for a scenario with a 350 m wide mangrove belt and a scenario without mangroves. In the latter case the required design dyke height would be 1.20 m higher.

Table 2 Design dyke heights and ∆h
Full size table

An adapted dyke design was developed based on the derived parameters (outer slope of 1:3, required dyke height of 3.19, and 4.39 m for the second scenario). Figure 3 shows a representative dyke cross-section at Mo O. Here the presence of mangroves reduces the required dyke height by 1.20 m.

Fig. 3
figure 3

Cross-sections of the dyke at Mo O for a design water level with a 5-year recurrence interval a with and b without mangroves

Full size image

The basic design of the improved dyke includes the existing dyke as part of the new dyke core. Sandy material can be used for the dyke core and is available nearby. A top layer made from clay covered with grass ensures impermeability of the dyke as well as a natural protection against erosion and suffusion of the dyke body. Subsoil material can be used for this layer. Compared with the standards for top layers given by EAK (2002), all characteristic parameters of the subsoil meet the required values.

All geotechnical verifications prove that the stability of the designed dyke and the subsoil is sufficient in Mo O. The subsoil can absorb the additional loads from the heightening of the dyke without losing its structural stability and the new dyke body itself is stable. However, the load bearing capacity in the investigation area is limited due to the soft and muddy soil. If the dyke height exceeds 4.60 m the soil cannot support the loads applied to the ground by the dyke leading to shear failure in the soil (Budhu 2011).

Floodplain management in Soc Trang

The design of T-groins/fences used for floodplain management in the Wadden Sea was adapted to local conditions and tested in a wave flume. T-shaped bamboo fences yielded the best results and have additional advantages due to the strength, availability and low costs of bamboo poles (Albers et al. 2013).

Floodplain restoration requires comprehensive knowledge about hydro- and morphodynamics. The results from field measurements and numerical modelling along the coast of Soc Trang Province from previous studies (Albers and von Lieberman 2011; Albers et al. 2013; Schmitt and Albers 2014) were used to define important boundary conditions for the design of coastal protection measures: soft soil with silty and clayey material; significant wave heights of 0.65 m; wave periods of 5–6 s; tidal range of 3.50 m; water depths at the dyke of up to 2 m at high tide.

Design parameters such as height of the bamboo fences (1.3 m), diameter of poles (8 cm for long-shore and 6 cm for cross-shore fences) and embedment depth (3.4 m; with 2.6 m in harder sub-soil) were based on field measurements, numerical modelling and static calculations. Field testing and monitoring results were used to determine the actual spacing of the poles (0.3 m), selection of the most suitable tying material (stainless steel wire) and composition of brushwood bundles (soft bamboo branches). Based on these, 5200 m of T-fences were built in Soc Trang and Bac Lieu Provinces in 2012.

The modelling was also used to ensure that downdrift erosion can be minimised as much as possible. The actual placement of the T-fences is done in such a way that they more or less recreate the original coastline by connecting existing headlands with mangrove vegetation (Fig. 4); i.e. they do not interfere with prevailing currents and thus will not cause downdrift erosion.

Fig. 4
figure 4

Restoration of eroded floodplains using bamboo T-fences in Bac Lieu Province (Mekong Delta, Vietnam). The long-shore elements close the eroded gap in the mangrove forest by connecting the remaining headlands (see white arrows). Photo Cong Ly and G. E. Wind 2013. (Color figure online)

Full size image

The permeable T-fences decrease the longshore currents and dampen the incoming wave energy by up to 80 % (average 66 %) independent of the height of the incoming waves at similar water levels (Albers et al. 2013). The reduction in wave height and thus in orbital velocities under waves leads to accelerated sedimentation rates. The results from benchmarks in Bac Lieu Province showed a deposition of approximately 17 cm of sediments within 7 months (Albers et al. 2013). The reduction of wave action on the landward side of the fences also accelerates the consolidation of the mud and thus increases the stability of the sediments against erosion. This was shown through mud density monitoring in Soc Trang Province (Albers et al. 2013) and can be seen in Fig. 5 (photo bottom left). The colour of the mud and the colonisation by Avicennia indicate consolidation of sediments from the land edge towards the gaps in the T-fences.

Mangrove regeneration on restored floodplain

The results of the fixed-point photo monitoring are shown in Fig. 5 which gives an example of sedimentation and natural regeneration of mangroves from the coast of Soc Trang Province at the sluice gate 4 area over a period of 28 months.

Fig. 5
figure 5

Natural regeneration of Avicennia on restored floodplain at sluice gate 4 (Soc Trang Province) from construction of the T-fences in October 2012 until January 2015. Photo top left (November 2012): coast parallel elements of the T-fences clearly visible (white arrow), gabions which protect the dyke front and row of tree poles to reinforce the dyke tow can been seen in the foreground; photo top right (February 2013) shows the beginning of the sedimentation; photo bottom left (November 2013): consolidation of sediments has started from the edge towards the gaps in the T-fences and natural regeneration of Avicennia starts to occur; photo bottom right (January 2015) shows the growth of mangroves which are not disturbed by wave action (due to the high/restored floodplain) or human impacts (Photos GIZ Soc Trang, R. Sorgenfrei 01/2015). (Color figure online)

Full size image

Discussion

The construction of coastal protection elements such as dykes is expensive and the possibility of increasing the dyke height is limited due to the load bearing capacity of the soil. Wave attenuation due to mangroves is therefore very important if dykes are to protect the hinterland when sea levels rise and storms increase in intensity and frequency. In locations where erosion has destroyed the mangrove forest in front of the dyke and eroded the foreshore, floodplain management is required to restore the eroded floodplain, thereby creating the pre-conditions for rehabilitation of the destroyed mangrove forest. Mangroves not only protect the coast from erosion, they also protect people living in the coastal zone from floods and storms, and provide ecosystem services, which generate benefits that extend far beyond the immediate coastal zone. This includes fisheries production, carbon sequestration and response to sea level rise (Barbier 2007; Nagelkerken et al. 2008; Alongi 2009, 2014; Lee et al. 2014). At certain sites, mangroves will be able to keep pace with sea level rise providing there is sufficient sediment supply and no change in hydrological conditions (Gilman 2006; McKee et al. 2007; Krauss et al. 2013; McIvor et al. 2013).

Floodplain management which uses T-groins/fences has become an important active coastal protection measure in the Wadden Sea over recent decades (Probst 1996; Kramer 1989). Sediment trapping fences have also been used in the USA (Boumans et al. 1997). More recently, Winterwerp et al. (2013) proposed the combined use of mangrove rehabilitation with engineering measures (building with nature) to stop and even reverse the erosion of mangrove-mud coasts through a strategy which combines ecological, hydrological and morphodynamic elements, as well as local socio-economic aspects. Floodplain management, including the stimulation of sedimentation using bamboo T-fences, is a cost-effective and sustainable approach which does not cause any major interference with natural coastal morphodynamics. However, sites where T-fences can be used for floodplain restoration must be carefully selected based on the existing soil type (the approach does not work under conditions dominated by grain fractions larger than medium sand) and bathymetry (extended periods of submergence significantly decrease the life span of bamboo T-fences). These criteria can only be assessed in the field and require data collection. Clough (2014) describes simple methods for soil and elevation analysis. To reduce the amount of field work required, sites should be pre-selected using remote sensing images.

Most of the design parameters derived for the investigation areas can be used for the southern Mekong Delta without modification. However, boundary conditions determined by local tides, currents and waves need to be taken into consideration for the T-fence design. This requires detailed measurements of wave, current and sediment parameters across both the rainy and dry seasons.

An area coastal protection strategy which combines floodplains (foreshore), mangrove forests and, where necessary, an appropriate earth dyke, provides a diverse and site-specific range of effective approaches to coastal protection. Such a strategy can ensure that the need for expensive engineering solutions is minimised and that path dependencies can be avoided. Thus the negative impact of dykes and shrimp ponds on sediment- and morphodynamic (Winterwerp et al. 2005; Thampanya et al. 2006) can be minimised or at least reduced. Furthermore, such a diverse strategy—which does not rely on concrete structures and which combines appropriate site-specific elements—can respond in a flexible way to future scenarios about flow regimes and sediment patterns of the Mekong River to the South China Sea which are predicted to diminish by 50 % in 2050–2060 mainly due to hydropower development in the catchment area (Nguyen et al. 2015). The need for a sound coastal defence strategy which is viable over time has also been identified as the solution for the dynamic mud-bank mangrove system along the coast of Guyana (Anthony and Gratiot 2012).

Maintaining or restoring a dynamic mud-coast mangrove system as part of an area coastal protection strategy requires a sound understanding of spatial and temporal sediment- and morphodynamic processes. Historic shoreline changes and simulation of future shoreline changes based on various boundary conditions needs to be considered. Such changes can be cyclic as has been shown for the coast of Vinh Tan (Soc Trang Province) where accretion of up to 23.6 m per year occurred between 1904 and 1965, and since then the coast has been subject to erosion of up to 14.1 m per year (Schmitt and Albers 2014). Understanding such processes will help with the decision when to intervene. Intervention should only become necessary if the natural dynamic (periodic disturbances such as destructive storms are natural processes and part of coastal dynamics) is disturbed by coastal and upstream developments (dykes, shrimp ponds, ports and dams) which lead to permanent changes.

Experience from Soc Trang and Bac Lieu Provinces has shown that bamboo T-fences are an effective way to restore eroded floodplains and the fine sediment balance required for natural regeneration of mangroves. However, site-specific differences in sediment supply and morphodynamic resulted in differences in sedimentation between sites ranging from 17 cm in 7 months up to about 100 cm in 12 months. Even if T-fences are unsuccessful in sites with limited sediment supply, path dependencies can be avoided because construction costs for bamboo T-fences are only about US$ 50–60 per meter, in contrast to US$ 2270 per meter for a 3.5 m high concrete dyke (Hillen 2008, price calculated based on an average exchange rate of 16,300 Vietnam Dong per US Dollar in 2008). In addition, the lifespan of bamboo fences—which is about 5–7 years (pers. comm. Worapol Douglomchan 2011, Khok Kha, Samut Sakhon Province, Thailand)—seems long enough for the restoration of floodplains along the east coast of the Mekong Delta and at the same time short enough to allow for flexible adjustments of elements of coastal protection if sediment—and morphodynamics change over time.

It is important that any area coastal protection strategy also contains provisions for long-term mangrove protection and management; otherwise the investments will be wasted if the mangrove forests on the restored floodplains are destroyed (again) by direct or indirect anthropogenic impacts, resulting in the floodplain eroding once again. Co-management, or shared governance, has shown to be an approach for sustainable and effective mangrove protection and management (Schmitt 2012). The protection of mangroves through co-management in the village Au Tho B, for example, has led to an increase of the mangrove area from nearly 70 ha in 2008 to almost 118 ha in 2014, mostly in the form of natural regeneration (based on comparison of satellite images and field observations).

The steps extending from eroded foreshore to floodplain restoration and mangrove regeneration/rehabilitation are summarised and illustrated in Fig. 6. The photo in Fig. 6 shows an example of an eroded foreshore from Soc Trang Province. The drawing to the right of the photo illustrates the effects of a low (i.e. eroded) floodplain on wave energy dissipation. Little wave energy is lost above the floodplain and most of the wave energy reaches the dyke. This increases the danger of severe erosion or even dyke failure and often leads to costly constructions such as revetments made from gabions.

Fig. 6
figure 6

The steps from eroded foreshore through floodplain restoration to mangrove regeneration/rehabilitation. Effective protection of the mangroves can prevent re-occurrence of erosion due to degradation or destruction of the mangroves. Photo K. Schmitt 2010; wave energy dissipation graphics modified from Albers et al. (2013) and mangrove zonation/root drawing from Duke and Schmitt (in print). (Color figure online)

Full size image

In order to protect the dyke and reduce the need to increase the height of the dyke, the eroded foreshores must be restored. The most effective way to do this along mud-coasts is to use permeable T-fences, which do not inhibit sediment input and create calm water conditions for sediment deposition. In addition, such T-fences, or groins, reduce erosion and therefore lead to the immediate solution of an acute threat in areas where the foreshore erosion has progressed all the way to the dyke. This is illustrated in the second drawing on the right side of Fig. 6. The high (i.e. restored) floodplain dissipates a lot of wave energy and much less energy (eroding force) reaches the dyke.

After successful restoration of sites suitable for mangrove growth, natural regeneration of mangroves will occur (Lewis et al. 2005). Similar outcomes were observed following the construction of detached breakwaters in Malaysia (Babak and Roslan 2011), and the same happened after the restoration of the eroded foreshores in the Mekong Delta through the use of bamboo T-fences (see Fig. 5). It is essential that the mangroves are protected from human impacts, otherwise the cycle of degradation/destruction and expensive restoration will continue uninterrupted. If rates of natural regeneration are insufficient, supplementary planting of mangroves may be necessary. In such cases, appropriate species need to be planted at the right sites at the correct time. This is most easily accomplished by learning from nature—mimicking how nature plants and the way nature creates a species zonation. Along muddy coasts, mangrove zonation (as shown in Fig. 6) leads to a complex structure of above and below-ground vegetation densities and thus to optimised coastal protection through wave attenuation by above-ground root structures along with soil stabilisation by below-ground root structures. The roots of the pioneer species at the seaward side are better adapted to protect the soil from wave action than the prop roots of the species in the middle of the mangrove belt. Here the larger above-ground root density is more effective at wave attenuation (Horstman et al. 2013; Duke and Schmitt in print).

Conclusion

Sophisticated, site-specific and appropriate approaches to coastal protection will become more and more important within the context of rising sea levels and the increasing frequency and intensity of extreme weather events such as storm surges.

A detailed calculation of wave overtopping rates shows the relevance of a wide (and healthy) mangrove forest belt in front of the dyke and its influence on the required dyke heights. The possibility for the increase of dyke heights in delta regions as response to rising sea levels is limited by the load bearing capacity of the muddy soils. Therefore, sustainable mangrove management plays an important role as an element of an area costal protection strategy.

Area coastal protection which includes floodplain and mangrove management and, where necessary, an appropriate earth dyke has shown to be effective for the mud-coasts and the low-lying areas of the Mekong Delta. A protection strategy which uses a diverse and site-specific range of approaches can ensure that coastal protection measures are effective and can respond timely to special and temporal changes in sediment- and morphodynamics.