Nature-based flood risk mitigation in river deltas is most effective with upstream and channel-fringing mangroves

doi:10.21203/rs.3.rs-4350184/v1

Download PDF

Article

Nature-based flood risk mitigation in river deltas is most effective with upstream and channel-fringing mangroves

https://doi.org/10.21203/rs.3.rs-4350184/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Global changes induce increased flood risks in river deltas, which concentrate a large portion of the global population and economic activity. Nature-based solutions, such as the conservation and restoration of mangroves in tropical river deltas, are considered a sustainable addition to expensive engineered flood defence systems. However, we currently lack science-based guidelines on where, in a river delta, mangrove conservation and restoration should be prioritised to attenuate extreme water levels most effectively and, conversely, where mangrove conversion to human land use would have limited impact on flood risks. Here, we apply state-of-the-art models to (1) an idealised generic geometry of a river delta and (2) a complex geometry of a real river delta case, which explicitly represents the complex pattern of mangroves, channels and human land use, to show that upstream-concentrated mangroves and channel-fringing mangroves are more effective in attenuating extreme water levels. Conversely, converting mangroves to human land use in zones located farther away from river channels and more downstream, has much less impact on extreme high water levels. Our study provides first step advice towards effective spatial planning that conciliates conservation of natural wetlands, economic development and flood risk mitigation.

Coastal cities concentrate a large portion of the global population and economic activity but climate change confronts them with rising flood risks due to long-term mean sea level rise and increasing frequency and intensity of extreme sea level events¹. Especially low-lying river delta plains are home to dense population and economic centres that are highly vulnerable to sea-born flood risks ². Consequently, future projections suggest increasing risks for important human and economic loss, with impacts disproportionately concentrated in developing tropical regions ^2–4. With continued global warming, this calls for cost-effective and sustainable coastal defence strategies. Nature-based approaches, consisting of the conservation and restoration of natural coastal wetlands, such as saltmarshes in temperate regions and mangroves in the tropics, are increasingly proposed as an effective addition to expensive engineered flood defence systems for coastal flood risk mitigation ^5–7.

Coastal wetlands naturally occur in low-lying coastal zones, in particular in river delta plains, where they can contribute to attenuate extreme sea levels through a number of processes ^5,8. Extreme sea levels often originate from storm surges ⁹ or climatic fluctuations such as the El Niño Southern Oscillation (ENSO) ¹⁰. When extreme sea levels enter a river delta, they propagate inland through the network of channels. When rising water levels overtop the channel banks, water flows from channels into the fringing wetlands ^11,12, which act as temporary lateral water storage, thereby reducing upstream propagation of extreme sea levels through the channels ^13,14. Furthermore, the dense vegetation and shallow water depths in the wetlands cause friction on the water motion, which hinders the propagation of extreme sea levels inside the wetlands ^11,12. Six of the top 10 cities most economically affected by coastal flood hazards ¹⁵ are positioned more than 50 km inland along a delta or estuary and could benefit from protection through the conservation or restoration of intertidal wetlands ⁷. Notably, cities such as Ho Chi Minh (Vietnam), Khulna (Bangladesh), and Palembang (Indonesia), ranked 4th, 8th, and 9th, respectively, among the world’s most flood-vulnerable coastal cities ¹⁵ but are separated from the ocean by mangroves.

In the past decades, deltaic wetlands, dominated by mangroves in the tropics, have been replaced on large scales by human land use ^16,1718. The main driver of mangrove loss is the expansion of aquaculture, which mainly consists of deforestation and construction of ponds. As the aquaculture ponds are surrounded by levees, they do not flood and the flood water storage capacity provided by the former mangroves is lost ^16,19. Other land use changes driving mangrove loss include conversion to agricultural fields, urban and industrial development. Since the early 21st century, however, mangrove loss is globally declining, which can be attributed to both a lack of remaining mangroves suitable for aquaculture or other land use conversion, and to a raising awareness of the multiple benefits of mangrove conservation, such as for carbon sequestration, fishery production or coastal defence ²⁰. This offers an opportunity to develop effective spatial planning that balances the coastal protection benefits of mangroves in certain parts of a delta plain, with the economic benefits of aquaculture or other human development in other parts. However, at present, there is limited understanding of how extreme sea level propagation through river deltas is affected by the spatial configuration of zones of mangrove conservation vs. conversion to human land use.

The flood buffer effect of mangroves at the full scale of large river deltas or estuaries (10s to 100s of km²) is typically assessed by means of hydrodynamic modelling. Model studies typically consider scenarios where the entire spatial extent of mangroves in a river delta or estuary are included or excluded ^21–26. However, comparing systems with the presence or complete absence of mangroves does not reflect the wide variety of spatial configurations of mangrove patches vs. human land use types that typically occur in low-lying coastal areas ²⁷. Other model studies only focus on small-scale (< 100 km²) mangrove areas ²⁸ or investigate the effects of specific wetland properties such as topography ²⁵ or the amount of vegetation-induced friction ^29,30. However, spatial patterns of mangroves vs. human land use on the scale of large deltas and their effect on extreme sea level propagation remain understudied. As a consequence, we lack science-based guidelines on where in a river delta, the conservation or restoration of mangroves should be prioritised to maximise the nature-based mitigation of extreme water levels, versus where mangrove conversion to human land use would have limited impact on extreme water levels mitigation.

In this study, we show for the first time how spatial patterns of mangroves and aquaculture affect the propagation of extreme sea levels. First, we present an idealised analytical 1D model describing the propagation of an (extreme) tidal wave through a funnel-shaped estuary. We use scenarios with different spatial patterns of channel-fringing mangroves, to demonstrate how spatial mangrove patterns strongly affect estuarine high water levels, depending on the length and shape of the estuary. Whilst such an idealised model can illustrate fundamental principles of extreme water level propagation, it does not include the complex geometry of mangroves and intertwining channels that is typical for deltas. Hence, as a next step, we present a hydrodynamic 2D numerical model for a representative delta with large mangrove and aquaculture areas. We simulate high water level propagation under normal and extreme tides for scenarios with varying spatial patterns of mangroves and aquaculture. The results of all simulations show that upstream-located and channel-fringing are much more effective in attenuating extreme sea levels than downstream mangroves.

2.1 Impact of varying mangrove width along estuaries

To identify the impact of spatial patterns of mangroves on high water levels in a delta, we applied as a first approximation the 1D analytical model of Van Rijn (2011) and adapted by ¹¹. The equations assume that a tidal wave can be amplified (i.e. resulting in higher high water levels) or attenuated (lower high water levels) when propagating landward through a funnel-shaped estuary, as governed by the balance between convergence of the estuary width and friction exerted by the estuarine bed and the mangrove vegetation. The model considers a single channel with continuous fringing mangroves. We run model scenarios with nine different idealised estuaries (with varying estuary length and convergence) combined with three different configurations of channel-fringing mangroves: uniform mangroves where the width of mangroves remains constant along the estuary length; upstream-concentrated mangroves, where the width of mangroves increases upstream; and downstream-concentrated mangroves, where the width of mangroves increases downstream. The total area of mangroves is equal among all scenarios.

The scenarios with mangroves concentrated upstream result in the lowest upstream water levels (Fig. 1). The model predicts slightly higher water levels in case of the uniform mangroves while water levels are substantially higher in case of downstream-concentrated mangroves, for all considered estuary shapes and dimensions. The analytical model shows larger effects of spatial mangrove configuration in shorter and strongly converging deltas.

2.2 Impact of spatial patterns of mangroves vs. human land use in a large tropical delta

While an analytical approach demonstrates the primary impact of varying mangrove width along estuaries, it does not include the impact of complex spatial patterns of mangroves intertwined by channels, which is intrinsic to real deltas. Therefore, we present a hydrodynamic model of the tropical Guayas delta in Ecuador (Fig. 2). The Guayas delta shows a typical spatial pattern of complex networks of channels and intertidal mangroves and mudflats (Fig. 2). There are two major estuarine branches, the western and eastern, where only the latter is fed by substantial freshwater discharge. Extreme water levels regularly occur in the delta and are related to the positive phase of the El Niño Southern Oscillation (ENSO), widely known as El Niño ^10,31. During an El Niño event, both seaborn water levels and river discharges can increase substantially. For instance, during the strong 1997–1998 El Niño, the mean sea level rose up to 50 cm near the seaward front of the delta and water levels inside the delta reached 90 cm above their normal tidal values ¹⁰. In the north of the delta lies the city of Guayaquil, which hosts about 3 million people. Due to its low-lying position, partly built over former deltaic mangrove soils, as well as the regular occurrence of extreme sea levels, Guayaquil is ranked as the third most vulnerable city to coastal flood hazards globally ¹⁵. Aquaculture expansion, which began in the 1960s, has led to the conversion of mangroves in the delta, with 54% of the natural mangroves remaining in 2015 ³².

Here we present a numerical modelling study, elaborating on a 2D hydrodynamic model presented, calibrated and validated against observed water level ³³. We simulate a normal spring tide, an extreme sea level event 60 cm higher than normal spring tide (like the 1997–1998 El Niño event) and a very extreme sea level event 120 cm higher than normal spring tide. We ran 18 scenarios with different spatial mangrove configurations: one where we replace all mangroves by aquaculture (Fig. 2b), one with the entire delta plain turned into mangroves (Fig. 2c) and four sets with varying mangrove configurations referred to as (1) ‘channel fringing’ (i.e. concentrating all mangroves fringing the channels; Fig. 2d-g), (2) ‘upstream’ (i.e. all mangroves concentrated upstream; Fig. 2h-k), (3) ‘downstream’ (Fig. 2l-o), and (4) ‘reference’ (i.e. laterally expanding or shrinking the current mangrove areas; Fig. 2p-s). For each of these four mangrove configurations, we run 4 scenarios, for which the mangrove coverage accounts for 25, 50, 54 (current coverage) and 75% of the delta plain, respectively. Note that the reference scenario with 54% mangrove coverage represents the current land use in the delta.

In the reference scenario with current mangrove coverage, the model predicts landward high water level amplification (i.e increase in high water levels) in line with observations (Section 5.2.5.;³³). With the current 54% total mangrove coverage, high water level profiles in the western (Fig. 3) and eastern branches (Figure S1) of the delta vary strongly among the mangrove configuration scenarios. The scenario with channel-fringing mangroves results in the lowest high water levels. The scenario with upstream mangroves and the reference scenario result in higher high water levels that are very similar for both scenarios, while the downstream mangrove scenario shows substantially higher high water levels. Differences in high water levels in the first 20 km (i.e. most downstream section of the delta) are minimal for all mangrove scenarios, despite the contrasting spatial mangrove patterns. Between km 30 and 60 (i.e., the upstream section of the delta), high water levels of the scenario with downstream mangroves are clearly higher than for the reference and upstream scenarios while for the scenario with channel-fringing mangroves, high water levels are lower than all other scenarios.

The spatial patterns of mangroves vs. aquaculture strongly affect the amount of landward high water level amplification for both the neutral and the El Niño cases (Fig. 4). The channel-fringing, reference and upstream mangroves scenarios consistently result in lower amplification rates than the downstream mangroves scenario. In case of a neutral spring tide and + 60 cm extreme sea level, the channel-fringing scenario, reference, and upstream scenarios show similar amplification rates. In case of the + 120 cm extreme sea level, the upstream scenario results in substantially lower amplification rates, already from a 25% mangrove coverage. All scenarios except the upstream mangroves result in amplification rates which decrease exponentially with increasing mangrove coverage. Furthermore, for the channel fringing scenarios, there is a threshold mangrove coverage above which amplification rates do not further decrease (50% coverage for the neutral spring tide, and 75% coverage for the + 60 cm extreme sea level scenario). For the + 120 cm extreme sea level scenario, there is no such threshold and attenuation continues with increasing mangrove coverage.

In order to elucidate a general mechanism explaining the different results among the different scenarios, we divided the 60 km transect along the western branch into six subtransects of 10 km, with the model domain also divided into six corresponding zones (Fig. 5e). For each subtransect, we calculated amplification rates, and for each corresponding zone, we calculated the water volume stored in the surrounding mangroves, ${V}_{m}$, relative to the total water volume in mangroves and channels, ${V}_{t}$, as: $V{{\prime }}_{m} = {V}_{m}/{V}_{t}$. From hereon, we refer to this as the relative mangrove water volume, $V{{\prime }}_{m}$. Amplification rates decrease exponentially with increasing relative mangrove water volume for the all mangroves scenario, reference, channel fringing and upstream mangrove scenarios (Fig. 6a, b & c). For the downstream mangrove scenarios, there is no clear decrease of amplification rates with increasing relative mangrove water volume (Fig. 6d). Furthermore, amplification rates in the middle subtransects are higher for the downstream mangrove scenarios compared to the channel fringing and upstream mangrove scenarios, regardless of having similar relative mangrove water volume.

Mangroves are increasingly valued for nature-based protection of low-lying coastal zones against flood risks from extreme sea levels, in particular in river deltas ^5,34. However, mangroves have been converted to human land use, such as for aquaculture development, on large scales in many delta areas ¹⁸. Nevertheless, the effects of resulting spatial patterns of mangrove patches and human land use on extreme sea level attenuation are understudied. Consequently, we lack guidelines on where in a river delta to prioritise mangrove conservation or restoration for most optimal reduction of extreme sea levels, versus where mangrove conversion would have limited impact on extreme water levels. By combining both analytical and numerical modelling, we show that the propagation of extreme sea levels in deltas strongly varies depending on spatial pattern scenarios of mangroves vs. human land use (Figs. 1, 3). Especially upstream-concentrated mangroves and channel-fringing mangroves are more effective in lowering upstream water levels, as compared to downstream-concentrated mangroves (Figs. 1, 3 and 4). Moreover, our results show that high water level reduction decreases exponentially with mangrove coverage, implying that conversion to aquaculture has a stronger effect on flood protection in deltas where a large fraction of the mangroves has been lost already compared to deltas with large mangrove forest still intact. In what follows we discuss the mechanisms that can explain our findings and discuss implications for how mangrove conservation and restoration can be optimised for nature-based mitigation of deltaic flood risks.

3.1 Effect of tidal prism in mangroves vs channels

In this study, we show that upstream-concentrated mangroves are most valuable in mitigating sea-born flood risks in a tropical river delta (Figs. 3 and 4). This can be explained as in funnel shaped estuaries, channels typically become narrower upstream. Therefore, the volume of water which flows out of the channels into the mangroves is proportionally larger relative to the water volume that propagates upstream through the channels, thereby leading to a stronger reduction of high water levels (conceptually illustrated in Fig. 6). In addition, when water flows from the channel into the mangroves in downstream areas, it is replaced by water from the open ocean. Upstream, this replacement is less efficient due to the larger distance to the ocean. While our experiments do not allow us to distinguish between both mechanisms, we argue that both the larger relative water volume flowing to the mangroves and the larger distance to the ocean contribute to the larger effect on reduction of extreme water levels by upstream mangroves compared to downstream mangroves. Therefore, upstream positioned mangroves are more effective in reducing extreme water levels compared to downstream mangroves. Similarly, using a 2D hydrodynamic numerical model of a single-channel funnel-shaped estuary, Smolders et al. ¹³ showed a higher reduction of peak water level for a storm surge in scenarios where a single small tidal marsh (< 30 km²) was located more upstream. Here, we show that in large deltas with a typically complex spatial mosaic of wetlands, human land use and channels, upstream-concentrated mangroves are most effective in lowering extreme water levels.

3.2 Effect of mangrove width on extreme water level reduction

Mangroves can lower upstream water levels by temporarily storing water when rising water levels exceed the channel banks ^13,28. Therefore, the flow rate of water from the channels into the mangroves is important for determining the reduction of upstream high water levels ³³. In case of the channel fringing scenarios, all channels are bordered by mangroves and as such, water can flow out of the channels over a continuous mangrove-channel border. For all the other scenarios, the extent of mangroves directly bordering channels is more limited, allowing less water to flow into the mangroves and be temporarily stored outside the channels. As such, the mangrove-fringing scenarios maximise the flow from the channels into the mangroves, leading to substantial reduction in upstream water levels. In addition, with increasing mangrove coverage, the width of the channel-fringing mangrove belt also increases (Fig. 2d - g). Wider mangroves fringes can store a larger volume of water and therefore can for a longer time sustain flow from the channels into the mangroves. Consequently, high water amplification rates decrease with increasing channel-fringing mangrove coverage. However, that decrease is exponential with less added reduction in amplification rates for the widest mangroves (Fig. 4). This implies that there seems to be a threshold width of mangroves above which more mangrove coverage does not result in a stronger reduction of upstream water levels (Fig. 4 and Supplementary Fig. 2). That threshold increases with the height of the extreme sea level event (Fig. 4). We explain this exponential decrease by the dense and complex network of mangrove roots which exerts much friction on the water flow, typically much more than the bottom friction ^35,36. As such, we argue that for wider mangroves, the potential water storage capacity of the mangroves is not fully reached as this friction by the mangrove roots limits the flow from the channel into the mangroves.

3.3 Implications

Over 67% of the global mangroves are located in estuaries or deltas ¹⁸. From the top 10 cities most economically impacted by coastal flood hazards ¹⁵, six cities are located more than 50 km inland along a delta or estuary and could be protected by conservation or restoration of intertidal wetlands ⁷. Cities such as Ho Chi Minh (Vietnam), Khulna (Bangladesh) and Palembang (Indonesia), which rank 4th, 8th and 9th among the world’s most flood-vulnerable coastal cities, respectively¹⁵, are all located along funnel-shaped estuaries (Supplementary Fig. 3). However, legal protection of mangroves in these estuaries is typically concentrated in downstream parts of the estuaries, while past conversions of wetlands to human land use is typically concentrated upstream ^37–40. Globally, mangrove conversion to human land use has indeed been concentrated near cities due to short transportation time among other practicalities which, for estuaries with large upstream cities, has inherently resulted in the loss of upstream located mangroves ⁴¹. These observations contrast with our finding that especially upstream located mangroves are most effective in reducing high water levels in deltas. Hence, for estuaries and deltas with upstream located, flood-vulnerable human population centres, our findings call for targeted mangrove conservation and restoration efforts between the sea and cities at risk, and in the more upstream parts of the estuaries or deltas.

While precise mangrove spatial distribution and extreme sea level propagation are specific to the geometry of each estuarine and deltaic system, our results emphasise the disproportionate potential of upstream located mangroves to protect against extreme sea levels. Nevertheless, the conservation and restoration of mangroves are not a one-size-fits-all solution for flood risk mitigation. For instance, studies have indicated that wetlands such as mangroves are less effective in reducing high water levels in short wide estuaries or open bays ^5,42,43. Furthermore, our results indicate a smaller role for downstream-concentrated mangroves for extreme water level reduction in estuaries but yet it is important to note that downstream mangroves can protect against other coastal hazards, safeguarding economic activities taking place in these locations such as aquaculture ponds. Mangroves can protect against wind gusts during storms ⁴⁴ and attenuate wind-driven waves ⁴⁵, both of which are typically strongest in the downstream parts of estuaries and deltas where wider channels have larger wind fetch lengths. Also, downstream mangroves can locally attenuate tsunamis and storm surges to mitigate flood risks for communities right behind the mangroves ^22,46. As such, the conservation of downstream mangroves can still be a valuable addition to engineered flood defences such as dikes. Moreover, mangroves do offer other valuable ecosystem services such as carbon sequestration, natural fish nurseries and water quality regulation, which motivate the conservation of downstream located mangroves ^47,48.

In addition to upstream mangroves, our results indicate that channel fringing mangroves are also most effective in reducing upstream propagation of extreme sea levels through a delta or estuary (Fig. 4). The channel fringing scenarios imply that, if mangroves are converted to human land use such as aquaculture ponds, this should be located far from the channels, to minimise its impact on flood risks. Aquaculture ponds often rely on water exchange with the natural tidal channels to sustain good water quality ^49,50 and boat transport ³² and therefore, the channel fringing scenarios might not be a practical solution in many deltas. Hence, in deltas and estuaries where large mangrove areas remain, conservation efforts in upstream mangroves appear more favourable as nature-based flood risk mitigation. With the global recognition of mangroves as a coastal protection strategy ^5,51, mangrove restoration is becoming a popular practice ⁵². Our findings imply that such conservation efforts should focus on upstream areas, especially in funnel-shaped estuaries that host low-lying upstream cities being vulnerable to extreme sea level events. Ultimately, our study provides first step advice toward proposing effective spatial planning that conciliates the conservation of natural wetlands while still ensuring economic development. We therefore call for the pressing need to implement these strategies particularly in delta areas that are highly exposed to coastal hazards.

5.1 Analytical model

The analytical model used in this paper is of ⁵³ and adapted by ¹¹ and describes the propagation of a single progressive sinusoidal wave through an idealised converging funnel-shaped channel. The along-channel variation of tidal range is estimated based on the wave energy flux and is governed by the following equation:

$$\frac{dH}{dx}=0.5\beta H\left(x\right)-\frac{f {c}^{2} cos\phi }{12\pi g {h}^{3}}H{\left(x\right)}^{2}$$

where $H\left(x\right)$ is the tidal range at distance x from the estuary mouth, $\beta$ is the width convergence coefficient, $g$ is the gravitational acceleration, $c=\sqrt{gh}$ is the local wave speed, $\phi$ is the phase lead of velocity maximum with respect to water level maximum and equals 1.05 radians, h is the mean water depth in the channels and is assumed to be constant over the delta and $f$ summarises the frictional effects following

$$f=\frac{8g}{{C}^{2}}$$

where $C$ is the the width-averaged Chézy coefficient:

$$C=\alpha {\left(x\right)C}_{c}+\left(1-\alpha \left(x\right)\right){C}_{m}$$

With $\alpha$(x) is the ratio between the channel width and the total width of channels and fringing mangroves, which was used to describe the three land use pattern scenarios tested in the analytical model with a being constant for the uniform mangrove scenario, linearly increasing for the upstream concentrated scenario and linearly decreasing for the downstream scenario, ${C}_{c}$ and ${C}_{m}$are the Chézy coefficient for the channel and mangroves respectively. The first can be expressed in relation to the commonly used Manning’s coefficient (Baptist et al. 2007):

$${C}_{c}={{h}^{\frac{1}{6}} n}^{-1}$$

Where $n$ is the Manning coefficient (s/m^1/3), representing the bed roughness in the channel and is set equal to 0.0175. The Chézy coefficient inside the mangroves ${C}_{m}$ can be expressed following the formulation for emerged vegetation:

$${C}_{m}=\sqrt{\frac{2g}{{h}_{mangr}{C}_{D}DM}}$$

Where ${h}_{mangr}$ is the inundation depth inside the mangroves and is set at 60 cm, ${C}_{D}$ is the dimensionless bulk drag coefficient and is assumed to equal 1, $D$ is the average diameter of prop roots and is assumed to equal 0.035 m and $M$ is the density of prop roots and is set equal to 85 m^− 2 (i.e. the number of prop roots per unit surface area). Values for the latter two are obtained from literature on observations of Rhizophora trees in Australia and Japan ^35,54,55.

The analytical solution of Eq. (1) is:

$$H\left(x\right)={\left(\left(\frac{{C}_{a}+{C}_{b}{H}_{0}}{{H}_{0}{C}_{a}}\right){e}^{-{C}_{a}x}-\frac{{C}_{b}}{{C}_{a}}\right)}^{-1}$$

Where ${H}_{0}$ is the tidal range at the seaward boundary and ${C}_{a}$ and ${C}_{b}$are along-channel constants:

$${C}_{a}=0.5\beta$$

$${C}_{b}=-\frac{f {c}^{2} cos\phi }{12\pi g {{h}_{eff}}^{3}}$$

Where ${h}_{eff}$ is the effective water depth, which is incorporated to include the effect of channel-fringing wetlands (e.g. mangroves) to the width-averaged friction and is described by:

$${h}_{eff }= \alpha h$$

5.2 Numerical model description

The numerical model is an extended version of the hydrodynamic model of the Guayas delta presented by Pelckmans et al.⁵⁶. The methods and data sources to define the computational mesh, model boundary conditions and model bathymetry remain the same. As the model domain is extended to include scenarios with more extensive mangrove areas than in reality (Fig. 1), the model is recalibrated and revalidated for this more extensive computational mesh, following the same procedure as described by Pelckmans et al. ⁵⁶.

5.2.1 Model equations

The hydrodynamics of the Guayas delta are modelled using TELEMAC 2D v8p2r0 ⁵⁷, which solves the depth-averaged shallow water equations:

$\frac{\partial h}{\partial t}+\nabla \bullet (h\upsilon$ ) = 0 (11)

$$\frac{\partial \upsilon }{\partial t} + \upsilon \bullet \nabla \upsilon = -g\nabla \eta +\frac{1}{h}\nabla \bullet (h\nu \nabla \upsilon ) -\frac{{\tau }_{b}+{\tau }_{v}}{\rho h}$$

where $h$ is the water depth (m), $\nabla$ is the differential operator (/m), t is the time (s), $\upsilon$ is the depth-averaged flow velocity (m/s), $g$ is the gravitational acceleration and equals to 9.81 m/s², $\eta$ is the water surface elevation above a reference level (m), $\nu$ is the diffusion coefficient and equals to 0.01 m²/s, ${\tau }_{b}$ is the bed shear stress (kg/ms²), ${\tau }_{v}$ is the vegetation-induced drag force per unit surface area (kg/ms²) and $\rho$ is the water density and equals to 1000 m²/s. The bed shear stress is parameterized using the Manning formulation:

$${\tau }_{b} = \frac{\rho g {n}^{2}}{{h}^{1/3}} \upsilon \left|\upsilon \right|$$

where $n$ is the Manning coefficient (s/m^1/3), representing the bed roughness and is calibrated (see 2.2.5). The vegetation-induced drag force is estimated as drag induced by rigid vertical cylinders with uniform properties ^36,58:

$${\tau }_{v} = \frac{1}{2}\rho {C}_{D}DM h \upsilon \left|\upsilon \right|$$

where ${C}_{D}$, D and M are considered equal to the values used in the analytical model.

5.2.2 The Guayas delta

The Guayas delta features two large branches exposed to semidiurnal tides, with a mean tidal range increasing from about 2 m at the mouth up to about 5 m upstream ³³. While the western branch does not receive any significant fresh water, the eastern branch is fed by the Guayas river, where discharge varies from 200 m³/s in the dry season to 1600 m³/s in the wet season⁵⁹. The delta is regularly exposed to extreme water levels related to the positive phase of the El Niño Southern Oscillation (ENSO), widely known as El Niño ^10,31. In addition, intense precipitation during strong El Niño events causes the Guayas river discharge to increase by a factor of 2.5¹⁰. While 46% of the mangroves have been replaced by aquaculture since the 1960’s, stricter regulation has significantly slowed down the replacement of mangroves by aquaculture since the early 21st century^32,60,61 and, especially since 2010, the extent of mangroves has increased⁶².

5.2.3 Model domain & mesh

The model domain covers the Gulf of Guayaquil and Guayas delta (Fig. 1). The downstream boundary coincides roughly with the edge of the continental shelf and the upstream boundaries are located at the approximate tidal limit along the Daule and Babahoyo rivers. In addition to the currently present mangroves and intertidal mudflats ⁵⁶, we included aquaculture zones that were considered to be suitable areas for mangrove growth in our model scenarios (Section 2.3). We included all aquaculture zones within 7 km of a subtidal channel, which was calculated as the maximum distance between any current mangrove area and a subtidal channel, and therefore assumed to be suitable areas for mangroves before conversion to aquaculture basins. The delineation of current mangroves and aquaculture are based on ESA’s Sentinel 2 satellite imagery. We refer to the mangrove and aquaculture areas together as the delta plain. For the different scenarios, mangroves and aquaculture areas are differently distributed over the delta plain (Section 2.3).

For the calibration and validation procedure, flooding of all current aquaculture zones was prohibited by increasing their elevation to 10 m above the reference level. The mesh resolution varies as a function of delta morphology with finer resolution (1) along the bathymetric gradient from the open sea towards the delta mouth area, (2) with decreasing channel width in the channels dissecting the delta, and (3) with decreasing distance to channels in the mangroves. The resulting mesh resolution ranges from 250 m at the open sea to 3 m in the narrowest channels.

5.2.4 Model bathymetry & boundary conditions

Open ocean bathymetry is based on the General Bathymetric Chart of the Ocean (GEBCO, 2020). Channel bathymetry in the delta is interpolated based on data obtained from the Oceanographic Institute of the Navy in Ecuador (INOCAR) inside the delta. We estimated intertidal mudflat topographies through remote sensing ³³. On all tidal flats, we extracted the waterline based on the Modified Normalised Difference Water Index (MNDWI) computed from Sentinel 2 satellite images captured at different times. These water lines were then considered as topographic contour lines for which the height was estimated based on nearby tide gauge measurements (³³ for more details). No data was available on the mangrove platform elevation but measured inundation levels in the mangroves of the Guayas showed very similar maximum flood depths over three spatially distributed locations ¹⁰. Therefore, we calibrated the mangrove platform elevation in order for the inundation depth to be constant over the entire delta and to match observed inundation depths during a spring tide in September 2019¹⁰. All bathymetries and topographies are referenced to the global geoid XGM2019 model.

At the open sea boundary, water levels and velocities are extracted from the global tidal models TPXO9 ⁶³ and used as is for model validation. For the scenarios, we imposed neutral conditions, El Niño conditions and extreme El Niño conditions. For the neutral scenarios, we imposed the water levels from TPXO9 as is. For the El Niño scenarios we added 60 cm to all oceanic water levels, to represent extreme sea levels similar to observed water levels during the strong 1997–1998 El Niño event ¹⁰. For the extreme El Niño scenario, we added 120 cm to TPXO9 water levels. Regarding the river boundaries, discharge data was available for 73% of the Guayas delta for the period 1990–2015 through INAMHI (Ecuador’s national meteorological and hydrological institute). To quantify discharge for October 2019, we applied a linear precipitation-weighted interpolation with monthly precipitation data from OpenLandMap ⁶⁴. El Niño driven discharge anomalies were not included as they do not influence water levels in the western branch ⁵⁶.

5.2.5 Model calibration and validation

The Manning coefficient $n$ was calibrated to fit observed water levels at 10 tide gauge stations uniformly spread over the delta. To exclude uncertainties on vegetation-induced friction and mangrove platform elevation, we calibrated the bed friction on 5 consecutive high and low waters around neap tide (22–24 september 2019), during which mangroves in the Guayas do not flood. This resulted in a value of 0.0175 s/m^1/3 in the outer delta and the western branch, and a value of 0.0100 s/m^1/3 in the eastern branch.

The model is validated against water levels at the same tide gauge stations as during calibration, during a spring tide on October 29, 2019. Model performance is considered excellent with Nash and Sutcliffe model efficiency ^65,66 of 0.85 ± 0.10, and root mean square error of $0.194 \pm 0.09$ m (tidal range ranges from 3 to 5 m).

5.2.6. Scenario design

We simulate scenarios with 18 different spatial patterns of mangroves vs. aquaculture, resulting in a total of 18 scenarios, each forced with extreme sea levels associated with an El Niño event (Section 2.2.3). The spatial patterns include a reference scenario with the current mangrove extent where 54% of the delta plain is covered by mangroves, the remaining 46% are aquaculture and in the model are raised with 10 m to prevent them from being flooded. We include a scenario without any mangroves, where the entire delta plain is raised 10 m, and a scenario where the entire delta plain is considered mangroves (Fig. 1a-c). Furthermore, we apply four scenarios where mangroves are only considered in areas within certain distances from the subtidal channels. These scenarios were constructed based on a calculated map of the shortest distance of any delta plain location to a channel (10 x 10 m raster, Figure S4). In the four scenarios, either the 25th, 50th, 54th or 75th percentile of delta plain closest to the channels are considered as mangroves, and all other areas farther away from the channels are considered aquaculture zones (Fig. 1d-g). The remaining scenarios are simulating the presence of mangroves more downstream or upstream along the delta. To construct these scenarios, we calculated a map of the so-called path length (10 x 10 m raster, Fig. 7b) which, for each delta plain cell, corresponds to the sum of its distance to the nearest channel and its shortest distance along a channel to the mouth of the eastern or western branch of the delta. From then, we included four scenarios with the delta plain cells contained within 25th, 50th, 54th and 75th percentile of the highest (Fig. 2h-k) and lowest (Fig. 2l-o) path lengths, resulting in scenarios where mangroves are concentrated upstream and downstream, respectively. For each scenario, the mangrove topography was recalibrated based on inundation records during the spring tide on 29 October 2019 to ensure inundation depths at all mangroves near the channels match the observed inundations¹⁰. We do this, as tidal wetlands such as mangroves typically build up their soil surface elevation by sediment accretion processes, up to an elevation that is in balance with the tidal inundation ⁶⁷, resulting in mangrove tidal inundation depths that are in line with those measured in the field at three mangrove locations distributed in the delta ¹⁰. Consequently, inundation depths at all mangroves near the channels are uniform across the delta.

Here, we further point out that the model scenarios where current aquaculture is replaced by mangrove areas do not include a channel network, which is typically present in mangroves. As the presence of channels has been shown to enhance high water level reduction ^33,36, we therefore argue that these simulated attenuation rates by these aquaculture-to-mangrove converted areas are underestimated in our study.

6. Acknowledgements

We thank Jan De Nul and INOCAR for sharing their tide gauge data and INHAMI for sharing river discharge data. The computational resources and services used in this work were provided by the HPC core facility CalcUA of the Universiteit Antwerpen, and VSC (Flemish Supercomputer Center), funded by the Research Foundation - Flanders 425 (FWO) and the Flemish Government. Furthermore, we would like to thank the Research Foundation Flanders (FWO, Belgium) for the PhD fellowship for fundamental research for Ignace Pelckmans (11E0723N). J.-P. Belliard is supported by FED-tWIN ABioGrad. O. Gourge was supported by the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No 798222. The study was locally supported in the context of the VLIR-UOS Ecuador Biodiversity Network 430 project.

7. Author contributions

Ignace Pelckmans, Jean-Philippe Belliard, Olivier Gourgue, Luis E. Dominguez-Granda and Stijn Temmerman contributed to the design of the study and collecting the necessary data. Ignace Pelckmans and Olivier Gourgue performed the model setup and scenario design with contributions and feedback from Jean-Philippe Belliard and Stijn Temmerman. Ignace Pelckmans wrote the first draft of the manuscript. All authors contributed to writing and revising the manuscript and approved the submitted version.

8. Competing interests

The authors declare that they have no conflict of interest.

9. Data availability

The python code to reproduce the analytical model and the simulated high water levels are available at 10.5281/zenodo.11032362.

Glavovic, B. C. et al. Cross-Chapter Paper 2: Cities and Settlements by the Sea. 2163–2194 (2022) doi:10.1017/9781009325844.019.2163.
Edmonds, D. A., Caldwell, R. L., Brondizio, E. S. & Siani, S. M. O. Coastal flooding will disproportionately impact people on river deltas. Nat Commun 11, 4741 (2020).
Neumann, B., Vafeidis, A. T., Zimmermann, J. & Nicholls, R. J. Future Coastal Population Growth and Exposure to Sea-Level Rise and Coastal Flooding - A Global Assessment. Plos One 10, e0118571 (2015).
Tessler, Z. D. et al. Profiling risk and sustainability in coastal deltas of the world. Science 349, 638–643 (2015).
Temmerman, S. et al. Marshes and Mangroves as Nature-Based Coastal Storm Buffers. Annu Rev Mar Sci 15, 95–118 (2023).
Narayan, S. et al. The Effectiveness, Costs and Coastal Protection Benefits of Natural and Nature-Based Defences. Plos One 11, e0154735 (2016).
Temmerman, S. et al. Ecosystem-based coastal defence in the face of global change. Nature 504, 79–83 (2013).
Hespen, R. van. Establishment and survival of coastal mangrove trees under mechanical disturbances. (2023) doi:10.33540/1315.
Fox-Kemper, B. et al. Ocean, Cryosphere and Sea Level Change. 1211–1362 (2021) doi:10.1017/9781009157896.011.
Belliard, J.-P. et al. El Niño driven extreme sea levels in an Eastern Pacific tropical river delta: Landward amplification and shift from oceanic to fluvial forcing. Global Planet Change 203, 103529 (2021).
Stark, J., Oyen, T., Meire, P. & Temmerman, S. Observations of tidal and storm surge attenuation in a large tidal marsh. Limnol Oceanogr 60, 1371–1381 (2015).
Horstman, E. M., Dohmen-Janssen, C. M., Bouma, T. J. & Hulscher, S. J. M. H. Tidal-scale flow routing and sedimentation in mangrove forests: Combining field data and numerical modelling. Geomorphology 228, 244–262 (2015).
Smolders, S., Plancke, Y., Ides, S., Meire, P. & Temmerman, S. Role of intertidal wetlands for tidal and storm tide attenuation along a confined estuary: a model study. Nat Hazards Earth Syst Sci 15, 1659–1675 (2015).
Fairchild, T. P. et al. Coastal wetlands mitigate storm flooding and associated costs in estuaries. Environ. Res. Lett. 16, 074034 (2021).
Hallegatte, S., Green, C., Nicholls, R. J. & Corfee-Morlot, J. Future flood losses in major coastal cities. Nature Climate Change 3, 802–806 (2013).
Thomas, N. et al. Distribution and drivers of global mangrove forest change, 1996–2010. Plos One 12, e0179302 (2017).
Bunting, P. et al. The Global Mangrove Watch—A New 2010 Global Baseline of Mangrove Extent. Remote Sens. 10, 1669 (2018).
Worthington, T. A. et al. A global biophysical typology of mangroves and its relevance for ecosystem structure and deforestation. Sci. Reports 10, 14652 (2020).
Richards, D. R. & Friess, D. A. Rates and drivers of mangrove deforestation in Southeast Asia, 2000–2012. Proc. Natl. Acad. Sci. 113, 344–349 (2016).
Goldberg, L., Lagomasino, D., Thomas, N. & Fatoyinbo, T. Global declines in human‐driven mangrove loss. Global Change Biol 26, 5844–5855 (2020).
Zhang, K. et al. The role of mangroves in attenuating storm surges. Estuar Coast Shelf Sci 102–103, 11–23 (2012).
Deb, M. & Ferreira, C. M. Potential impacts of the Sunderban mangrove degradation on future coastal flooding in Bangladesh. J Hydro-environ Res 17, 30–46 (2017).
Sheng, Y. P., Rivera-Nieves, A. A., Zou, R. & Paramygin, V. A. Role of wetlands in reducing structural loss is highly dependent on characteristics of storms and local wetland and structure conditions. Sci. Rep. 11, 5237 (2021).
Li, L., Wang, X. H., Williams, D., Sidhu, H. & Song, D. Numerical study of the effects of mangrove areas and tidal flats on tides: A case study of Darwin Harbour, Australia. Journal of Geophysical Research: Oceans 117, 1–12 (2012).
Pelckmans, I. et al. Mangrove ecosystem properties regulate high water levels in a river delta. Nat. Hazards Earth Syst. Sci. 23, 3169–3183 (2023).
Fanous, M., Daneshkhah, A., Eden, J. M., Remesan, R. & Palade, V. Hydro-morphodynamic modelling of mangroves imposed by tidal waves using finite element discontinuous Galerkin method. Coast. Eng. 182, 104303 (2023).
Bryan-Brown, D. N. et al. Global trends in mangrove forest fragmentation. Sci. Rep. 10, 7117 (2020).
Dominicis, M. D., Wolf, J., Hespen, R. van, Zheng, P. & Hu, Z. Mangrove forests can be an effective coastal defence in the Pearl River Delta, China. Commun Earth Environ 4, 13 (2023).
Chen, Q. et al. Improved modeling of the role of mangroves in storm surge attenuation. Estuar Coast Shelf Sci 260, 107515 (2021).
Hu, K., Chen, Q. & Wang, H. A numerical study of vegetation impact on reducing storm surge by wetlands in a semi-enclosed estuary. Coastal Engineering 95, 66–76 (2015).
Colas, F., Capet, X., McWilliams, J. C. & Shchepetkin, A. 1997–1998 El Niño off Peru: A numerical study. Prog Oceanogr 79, 138–155 (2008).
Hamilton, S. E. Mangroves and Aquaculture: A Five Decade Remote Sensing Analysis of Ecuador’s Estuarine Environments. (Springer, 2019). doi:10.1007/978-3-030-22240-6.
Gijsman, R. et al. Nature-Based Engineering: A Review on Reducing Coastal Flood Risk With Mangroves. Frontiers in Marine Science 8, (2021).
Mazda, Y. et al. Drag force due to vegetation in mangrove swamps. Mangroves and Salt Marshes 1, 193–199 (1997).
Horstman, E. M., Bryan, K. R. & Mullarney, J. C. Drag variations, tidal asymmetry and tidal range changes in a mangrove creek system. Earth Surface Processes and Landforms 46, 1828–1846 (2021).
Kuenzer, C. & Tuan, V. Q. Assessing the ecosystem services value of Can Gio Mangrove Biosphere Reserve: Combining earth-observation- and household-survey-based analyses. Appl. Geogr. 45, 167–184 (2013).
Giri, C., Pengra, B., Zhu, Z., Singh, A. & Tieszen, L. L. Monitoring mangrove forest dynamics of the Sundarbans in Bangladesh and India using multi-temporal satellite data from 1973 to 2000. Estuar., Coast. Shelf Sci. 73, 91–100 (2007).
Eddy, S., Milantara, N., Sasmito, S. D., Kajita, T. & Basyuni, M. Anthropogenic Drivers of Mangrove Loss and Associated Carbon Emissions in South Sumatra, Indonesia. Forests 12, 187 (2021).
Nguyen, T. T. N. et al. Nutrient budgets in the Saigon–Dongnai River basin: Past to future inputs from the developing Ho Chi Minh megacity (Vietnam). River Res. Appl. 36, 974–990 (2020).
Hagger, V. et al. Drivers of global mangrove loss and gain in social-ecological systems. Nat. Commun. 13, 6373 (2022).
Haddad, J., Lawler, S. & Ferreira, C. M. Assessing the relevance of wetlands for storm surge protection: a coupled hydrodynamic and geospatial framework. Nat. Hazards 80, 839–861 (2016).
Cassalho, F. et al. Coastal Wetlands Exposure to Storm Surge and Waves in the Albemarle-Pamlico Estuarine System during Extreme Events. Wetlands 41, 49 (2021).
Das, S. & Crépin, A.-S. Mangroves can provide protection against wind damage during storms. Estuar., Coast. Shelf Sci. 134, 98–107 (2013).
Horstman, E. M. et al. Wave attenuation in mangroves: A quantitative approach to field observations. Coast. Eng. 94, 47–62 (2014).
Montgomery, J. M., Bryan, K. R., Mullarney, J. C. & Horstman, E. M. Attenuation of Storm Surges by Coastal Mangroves. Geophys Res Lett 46, 2680–2689 (2019).
Alongi, D. M. Carbon Cycling and Storage in Mangrove Forests. Annu. Rev. Mar. Sci. 6, 195–219 (2014).
Rogers, A. & Mumby, P. J. Mangroves reduce the vulnerability of coral reef fisheries to habitat degradation. PLoS Biol. 17, e3000510 (2019).
Davis, R. P., Boyd, C. E. & Davis, D. A. Resource sharing and resource sparing, understanding the role of production intensity and farm practices in resource use in shrimp aquaculture. Ocean Coast. Manag. 207, 105595 (2021).
Hukom, V., Nielsen, R., Asmild, M. & Nielsen, M. Do Aquaculture Farmers Have an Incentive to Maintain Good Water Quality? The Case of Small-Scale Shrimp Farming in Indonesia. Ecol. Econ. 176, 106717 (2020).
Hespen, R. van et al. Mangrove forests as a nature-based solution for coastal flood protection: Biophysical and ecological considerations. Water Sci Eng 16, 1–13 (2023).
Su, J., Friess, D. A. & Gasparatos, A. A meta-analysis of the ecological and economic outcomes of mangrove restoration. Nature Communications 12, 5050 (2021).
Rijn, L. C. van. Analytical and numerical analysis of tides and salinities in estuaries; part I: tidal wave propagation in convergent estuaries. Ocean Dyn. 61, 1719–1741 (2011).
Mazda, Y., Kobashi, D. & Okada, S. Tidal-Scale Hydrodynamics within Mangrove Swamps. Wetlands Ecology and Management 13, 647–655 (2005).
Vandenbruwaene, W., Bouma, T. J., Meire, P. & Temmerman, S. Bio‐geomorphic effects on tidal channel evolution: impact of vegetation establishment and tidal prism change. Earth Surf. Process. Landforms 38, 122–132 (2013).
Hervouet, J. Hydrodynamics of Free Surface Flows. (Hoboken, NJ, United States: Wiley-Blackwell, Hoboken, NJ, United States: Wiley-Blackwell, 2007). doi:10.1002/9780470319628.
Baptist, M. J. et al. On inducing equations for vegetation resistance. Journal of Hydraulic Research 45, 435–450 (2007).
INHAMI. Official INAMHI website. (2019).
Beitl, C. M. The Changing Legal and Institutional Context for Recognizing Nature’s Rights in Ecuador: Mangroves, Fisheries, Farmed Shrimp, and Coastal Management since 1980. J. Int. Wildl. Law Polic. 19, 317–332 (2016).
López-Angarita, J., Roberts, C. M., Tilley, A., Hawkins, J. P. & Cooke, R. G. Mangroves and people: Lessons from a history of use and abuse in four Latin American countries. For. Ecol. Manag. 368, 151–162 (2016).
Morocho, R., González, I., Ferreira, T. O. & Otero, X. L. Mangrove Forests in Ecuador: A Two-Decade Analysis. Forests 13, 656 (2022).
Egbert, G. D. & Erofeeva, S. Y. Efficient Inverse Modeling of Barotropic Ocean Tides. J Atmos Ocean Tech 19, 183–204 (2002).
Hengl, T. & Parente, L. Monthly precipitation in mm at 1 km resolution (multisource average) based on SM2RAIN-ASCAT 2007-2021, CHELSA Climate and WorldClim. (2022) doi:10.5281/zenodo.6458580.
Allen, J. I., Somerfield, P. J. & Gilbert, F. J. Quantifying uncertainty in high-resolution coupled hydrodynamic-ecosystem models. J Marine Syst 64, 3–14 (2007).
Nash, J. E. & Sutcliffe, J. V. River flow forecasting through conceptual models part I — A discussion of principles. J Hydrol 10, 282–290 (1970).
Temmerman, S., Govers, G., Wartel, S. & Meire, P. Modelling estuarine variations in tidal marsh sedimentation: response to changing sea level and suspended sediment concentrations. Mar. Geol. 212, 1–19 (2004).
Pelckmans, I. et al. Mangrove ecosystem properties regulate high water levels in a river delta. Nat. Hazards Earth Syst. Sci. 23, 3169–3183 (2023).

There is NO Competing Interest.

SupplementaryMaterials.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Nature-based flood risk mitigation in river deltas is most effective with upstream and channel-fringing mangroves

Status:

Version 1

Abstract

Figures

1. Introduction

2. Results

2.1 Impact of varying mangrove width along estuaries

2.2 Impact of spatial patterns of mangroves vs. human land use in a large tropical delta

3. Discussion

3.1 Effect of tidal prism in mangroves vs channels

3.2 Effect of mangrove width on extreme water level reduction

3.3 Implications

Methods

5.1 Analytical model

5.2 Numerical model description

5.2.1 Model equations

5.2.2 The Guayas delta

5.2.3 Model domain & mesh

5.2.4 Model bathymetry & boundary conditions

5.2.5 Model calibration and validation

5.2.6. Scenario design

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1