Hydrodynamic characteristics and morphodynamic evolution

In the absence of mangroves, the mud profile continues to develop and accretes sediment with clear spring-neap tidal depositional signals near the boundary (Fig. 2), albeit at a very low rate. The introduction of mangroves has a clear impact on coastal morphology: The presence of mangroves promotes sediment deposition in the patch interior and erosion at the seaward fringe to a greater extent in scenarios B, C, and F and to a lesser extent in scenarios D and G. Scenario B even traps some sediment landward from the mangrove patch. Interestingly, the mangrove forests in the upper intertidal (above MHWN) promote a relatively high depositional rate in the interior. As shown in Fig. 3b, the presence of mangroves facilitates vertical accretion rate in the forest interior by up to two times compared to the no-vegetation scenario. The modelled mean accumulation rate in the mangrove interior is found to nearly correspond to the global average for undisturbed conserved mangrove56, about 3.6 ± 0.4 mm yr−1 (Fig. 3b), which illustrates that the model simulates a realistic sedimentation rate. Mangrove-induced drag can have an extended influence on reaching the seaward boundary close to the lowest tidal water level. The scenario with mangroves near MSL (B) shows profound erosion at the seaward fringe and deposition in the mangrove interior towards its landward surroundings. Higher, more landward mangrove planting decreases erosion near the fringe and tends to deposit sediment within the interior. Under those circumstances, mangroves drive the redistribution of sediment over the profile, causing the soil surface elevation to accrete until near the boundary seaward (Fig. 3b). For that reason, it indicates the extent of mangrove influences on morphology in this system. In scenarios with multiple vegetation patches, the morphodynamics patterns around the lower patch resemble the single-patch scenarios (Fig. 2). In most cases, the deposited volume in the mangrove patch located in the lower intertidal is even less than the eroded volume at the seaward portion of the fringe, suggesting the possibility of a negative impact on bare mudflat sediment stock by introducing mangroves.

Fig. 2: Simulated alongshore averaged bed-level development. Panels a–g show the spatiotemporal cumulative soil surface elevation development in each scenario. The mangrove extent is shown as black hatches overlaying the soil surface elevation. Panel h shows the ocean boundary with morphological spring-neap tide forcing. The tidal datum is presented in the lower right of the panel, with MHWN at 0.8 m and MHWS at 1.2 m. Full size image

Fig. 3: Initial profile, mangrove planting scenarios, and annual vertical erosion/ sedimentation rate. Panel a describes the initial profile, tidal levels, and scenario placement. The observation stations represent the seaward, fringe, and mangrove patch interior. The observation stations are arranged as follows: sta 1 at y-dir = 1000 m represents the offshore; sta 2 at y-dir = 1800 m represents the fringe for above MSL (scenario B); sta 3 at y-dir = 2000 m represents the forest interior of scenario B and fringe for below MHWN scenario C and F; sta 4 at y-dir = 2500 m represents forest interior for (scenario C and F) and fringe for above MHWN (scenario D and G); and sta 5 at y-dir = 3000 m represents the fringe for below MHWS (scenario E). We do not consider the forest interior of scenario E because it is close to MHWS and less frequently inundated. Panel b shows the annual alongshore average erosion/sedimentation rate (y-direction), mm/yr. The alongshore average yearly rate is calculated over the entire 20-year simulation period. Coloured lines above the graphs represent the extent of the mangrove forest after 20 years. Scenario line colours are consistent throughout the article. Full size image

The drag-inducing effect of the growing mangrove community is apparent and, therefore, influences the tidal asymmetry and the associated transport capacity. Together with fine sediment erosion and deposition lag effects, tidal asymmetry of water levels and velocities influences tide-residual sediment transport, where a slight asymmetry can result in a large net influx of sediment into the forest over a tidal cycle57,58. To explore these effects, Fig. 4 shows tidal stage versus velocity plots for the entire 20-year simulation period. The observation stations are selected to represent the tidal stages at the mudflat, the fringe, and the mangrove patch interior of each scenario.

Fig. 4: Tidal stage plots extracted from numerical modelling results. Panels a–h show the evolution of the cross-shore averaged tidal stage over simulation time, where water velocity (m/s) is plotted against water level (m) at 5 different observation stations of each scenario during spring tide. The plots show the evolution of the tidal stage for each scenario over the simulation period of 20 years with a colour palette ranging from red for year 0 to yellow for year 10 and blue for year 20. Observation stations on the initial profile are described in panel (h). Velocities are positive in the landward (flood) direction. Full size image

Near-equilibrium conditions prevail in the bare mudflat scenario A (Fig. 4a) where the velocities are similar along the entire profile during the tidal cycle, with a slightly larger flood velocity, indicating an accretive system. Figure 4b–g clearly illustrate the reinforcing effect of drag induced by growing mangroves on the tidal asymmetry. The systems become flood-dominated (larger flood velocity) in the mangrove forest interior and ebb-dominated (larger ebb velocity) at and seaward the fringes. Water levels and velocities are lower in and landward of the mangrove patches. These effects become more assertive when mangroves grow in size and colonize more areas. On average, mangroves start to substantially affect the hydrodynamics once they have passed 10 years.

Mangroves induced drag that limits flow and attenuates waves favouring sediment deposition in the forest interior7,59. Larger pressure gradients drive larger sediment fluxes during the flood. During ebb, a pressure gradient develops when water levels at the mudflat drop faster than in, and landward of, the mangrove patches. High friction delays the flow when ebb water drains, maintaining relatively high water levels in the patch interior. Turbulence from breaking waves enhances sediment resuspension, particularly during the low water near the fringe in the seaward direction. The increased bed slope and pressure gradient at the fringes lead to larger ebb velocities, favouring further offshore sediment transport. This pattern is consistent with previous observations and modelling studies, e.g., Van Maanen et al.60 and Bryan et al.61. Denser and larger-diameter mangrove populations enhance this effect. The effect of mangrove-induced drag on tidal asymmetry is valid in all scenarios regardless of the position of mangrove forests relative to MSL. In the case of multiple patches, the most seaward-located patch governs the mechanism. Under similar conditions, mangroves situated at higher elevations have less pronounced effects. Particularly for patches above MHWN, the transition from peak flood in the interior to lower ebb in the seaward or the other way around during flood is relatively short. Thus, the findings in our simulations support previous studies, which show that the sediment accumulation rate was greater near the fringe than in the interior.

The model results show a strong correlation between the position of mangroves relative to MSL and sediment deposition. Mangroves between the mean high-water spring (MHWS) and the mean high-water neap tidal elevation (MHWN) tend to favour deposition. In contrast, mangroves between MHWN and mean sea level (MSL) tend to induce scarp erosion at the seaward fringe edge. In contrast, scenarios featuring patches closer to MSL have a more pronounced scarping effect. This is attributed (a) to a larger, landward-located water volume that needs to be drained, (b) to faster-lowering water levels during ebb at elevations closer to MSL, and (c) to associated larger pressure gradients within the mangrove interior. Such pronounced fringe edge erosion has likewise been observed in field studies of mangrove environments62,63. The supplement presents additional numerical experiments on morphostatic profiles that explain this effect in more detail.

Mangrove biomass development

Scenarios with mangroves planted in two patches with a gap and above high water neap exhibit the highest biomass and widest extent. We observed a pronounced sensitivity of the canopy cover area (Fig. 5) and number of trees (Fig. 6) to mangrove placement. The final cover area can vary by a factor of two between mangroves located at the low water level and those placed at the high water spring level. This striking difference is attributed solely to the initial mangrove patch position relative to MSL. The maximum cover area is achieved in scenario G (patches below MHWS+above MHWN) followed by scenario F (patches below MHWN&above MHWN) (Fig. 5). The canopy area evolution can be divided into three distinct phases: During the first period until year 5, patched scenarios cover slightly larger areas than the single block scenarios. Next, between years 5 and 10, trajectories diverge depending on bed elevation within the two planting scenarios (single vs. multiple patches). As the community grows older, the higher soil surface elevation provides a shorter hydroperiod and less energetic wave conditions, enhancing the probability of seedling establishment. We simulate this process using the conceptual model of the Window of Opportunity (WoO)64, defined as the minimum disturbance-free days,the critical period to let the seedling’s root securely anchor in the soil. Once the parental trees have released the propagules, we estimate the dispersal to depend on the averaged current magnitude and direction, which is equal to two weeks of the morphological simulation period. This period considers the phase of obligate dispersal to achieve early anchorage in Avicennia sp. and Rhizophora sp.43 where the propagules will be distributed following the prevailing currents. We evaluated the seedling establishment in the model on the local inundation frequency, which requires 3 inundation-free days for Avicennia sp. and 5 days for Rhizophora sp.43. In addition, the model also distinguish the seedlings’ failure to settle due to uprooting and burial, described in more detail in the methods section. Since the competition remains low, mangroves grow optimally, resulting in a larger extent than those located lower on the profile, featuring less favourable conditions for establishment and growth. After 12.5 years of growth, leveraging the advantage of being in a higher topographical position, the canopy area in scenario C, located below MHWN, surpasses that in scenario B, where the patch is situated above MSL. The slightly better WoO due to lower inundation frequency in scenario C provides a relative advantage, which becomes more apparent when the species initiate propagule production. The improved window of opportunity increases the probability of stranded propagules surviving and growing into saplings. Therefore, scenario C contains a higher number of trees (Fig. 6) and a wider canopy cover than scenario B. Overall, mangrove placement at particular levels on the topographic profile exerts a strong influence on the number of trees in the community after 20 years. Mangrove communities located below MHWS are the largest in size, and community sizes tend to decline in the order: scenario above MHWN, below MHWN, and above MSL. Patchiness offers a clear advantage as tree numbers in those planting scenarios surpass those in single block scenarios after years 12.5–15.

Fig. 5: Canopy cover area development. Development of canopy cover area over 20 years in different planting scenarios. Full size image

Fig. 6: Mangrove populations’ development. a Mangrove populations’ (number of trees) development in space-available case. This plot captures each species’ aggregated mangrove tree population over the simulation period. b Distribution of mangrove diameter at breast height (\({D}_{130}\)) in cm every 5 years for Rhizophora apiculata (red bars) and Avicennia marina (blue bars). Propagules are produced when the tree has reached the flowering stage, with species-specific propagule density per crown surface area. c Overall biomass over the 20 years simulation period for the two-competing species Avicennia marina (solid line) and Rhizophora apiculata (dashed line). Biomass (kg) was calculated following the approximation by Comley and Mc Guinness 110 for Avicennia marina, Kauffman and Cole 111 and Ong et al. 112 for Rhizophora apiculata as in Murdiyarso et al. 68. Biomass content in the plot is the sum of all mangrove species individuals on each time stamp. Full size image

Canopy cover area development represents the top of the canopy and, therefore, corresponds to the behaviour of the tallest stands44. The development of species-specific number of trees, as shown in Fig. 6a provides additional insight into whether the mangrove community is declining, expanding, or heading towards (dynamic) equilibrium by including all life stages in the community.

Until generative reproduction commences within the first 5 years, the number of trees remains constant across all scenarios (Fig. 6a). Although Avicennia marina grows more slowly in biomass than Rhizophora apiculata65, it has a higher seedling production rate66 Additionally, Rhizophora apiculata propagules require a longer inundation-free period to settle than Avicennia marina. This causes the Avicennia marina population to dominate the forest regarding the number of trees.

As illustrated in Fig. 6a, b, after 15 years, the mangrove community in all single-patch scenarios stabilizes gradually. An exception is the scenario above MSL (scenario B), which shows a decline after year 12 in the Avicennia marina population and a transition to a steady number of Rhizophora apiculata trees. In contrast, the number of trees in the two-patches scenarios continues to rise. The increase in the number of trees in Fig. 6a reflects an expanding mangrove forest or a less competitive environment with more younger mangrove trees established. Conversely, a decrease in tree numbers indicates a more competitive mangrove forest, with high juvenile mortality and more growth of the mature trees or a more potent signature of the self-thinning mechanism. Despite the slower growth rate of the Rhizophora apiculata population, the increasing trend for all species is quite similar. Seedling mortality in the mangrove community is mainly controlled by the inundation period, where the hydro-morphodynamic feedback by the expanding and denser forest lowers the WoO. In the first 10 years, the biophysical properties (surface area, density, and \({D}_{130}\)) of the mangrove forest have only a minor effect on hydrodynamics, leading to a negligible influence on morphodynamic development (Figs. 2 and 4). Under these conditions, the WoO is lower at the seaward fringe and higher towards the interior and landward fringe. After 10 years, the increased mangrove density and \({D}_{130}\) cause higher friction, leading to tidal asymmetry and pressure gradient. This tidal asymmetry delays the interior water level, extending the inundation period and preventing propagule establishment. The increased velocity induced by the pressure gradient moves propagules from the seaward fringe offshore, reducing seedling establishment. As a result, the number of trees remains static, with no colonization occurring at either the seaward fringe or the interior. In contrast, scenarios with multiple patches continue to produce viable saplings in both colonization areas: the seaward fringe at the upper patch and the landward fringe at the lower patch. Plots of WoO and propagule production on observation plots are provided in the supplement.

Although Rhizophora apiculata has a smaller population size, the mean diameter of mature trees at the upper end of the diameter distribution is substantially larger for Rhizophora apiculata than that for Avicennia marina (Fig. 6b). We observe a dominance of the large Rhizophora apiculata population after year 10. However, given that Avicennia marina produces more propagules, we find an equivalent number of juveniles across all scenarios. In years 15 and 20, the dominant distribution of \({D}_{130}\) 30 cm category in year 20. This histogram illustrates that two-patch scenarios consistently have a wider distribution of tree diameters. All single-block scenarios experienced a decreasing number of juveniles, with scenario B having the lowest number compared to the patched scenarios. This indicates that the community will be dominated by mature trees with high competition.

Rhizophora apiculata (dashed lines) has developed a considerably larger biomass from year 2 onwards (Fig. 6c), similar to the observation by Berger and Hildenbrandt65. In contrast, Avicennia marina lags behind in biomass, as juveniles dominate the diameter distribution (Fig. 6b). Once again, scenarios with multiple patches and gaps produce more biomass due to their wider extent (Fig. 2) and higher number of trees (Fig. 6a). Beyond 20 years, the higher WoO in multiple patches scenarios leads to lateral expansion due to colonisation, higher survivability of juveniles progressing into mature trees, and ultimately higher biomass capacity.

Carbon mitigation capability and potential

We estimate potential mangrove carbon (C) stock in three carbon pools: a) aboveground biomass carbon (AGC) represents all living parts of vegetation above the soil, including stem, branches, foliage, part of stilt roots (R. apiculata), and pneumatophores (A. marina), b) belowground biomass carbon (BGC) represents biomass of belowground roots below the soil surface67, and c) soil organic carbon (SOC) which is the change in carbon accumulation due to the modelled cumulative erosion/sedimentation. We assume that sediment imported from the seaward boundary contains organic matter with 10% carbon content, in accordance with values reported for observation projects in Indonesia68,69. The carbon content was derived from 1m-depth sediment core samplings characterized by abundant fibrous roots and occasional shell fragments, in which the detritus contribution would be captured in the observation. The schematized domain corresponds to an accreting coastal system that feeds on the mud of riverine origin transported alongshore, where the allochthonous sediment dominates in-situ carbon burial. The methods section explains the estimation of AGC, BGC, and SOC.

Figure 7 shows that, in line with findings in the previous section, all scenarios below the level of MHWN show a reduced soil organic carbon accumulation compared to the no vegetation case (scenario A). Scenarios with all combinations situated above MHWN (scenarios D, E, G) accumulate more sediment than scenario A. Scenario D (above MHWN) and Scenario E (below MHWS) surpasses scenario A after year 10. At the end of the simulation, the two-patches scenario with a combination above MHWN and below MHWS (scenario G) has the highest SOC. Interestingly, although all two-patches scenarios have the largest canopy area, scenarios with a combination situated below MHWN (scenario F) feature less SOC than scenario A. Scenario B (above MSL) has the smallest amount of net SOC accumulation, almost vanishing near the end of the simulation. The SOC accumulation rate is quite similar in all scenarios. However, all scenarios below MHWN reverse their response pattern in year 10 and regain sediment afterward (Fig. 7), whereas only SOC in scenario B keeps decreasing.

Fig. 7: Carbon stock changes. Simulated carbon stock changes from three carbon pools (living vegetation above- belowground and soil organic carbon). The y-axis shows the value of carbon stocks changes (in 106 kg) with positive values in both directions, where a value above zero represents aboveground carbon and below zero represents belowground and soil organic carbon. Above and belowground biomass carbon are estimated based on the mangrove species properties. Soil organic carbon is estimated from cumulative sediment volume changes from the whole model domain. Figure panels provide the carbon stock changes for scenarios a no vegetation, b single patch planted above mean sea level, c single patch planted below mean high water neap, d single patch planted above mean high water neap, e single patch planted below mean high water spring, f multiple patches planted below and above mean high water neap, and g multiple patches planted below mean high water spring and above mean high water neap. Full size image

We calculate carbon stock changes for the carbon pools modelled above. All scenarios show an increase in carbon stock over time. However, scenario B has reached its limit within 20 years for the SOC due to soil loss by erosion. In comparison, scenarios below MHWN have apparently reached a saturated carbon content in living biomass after 20 years (AGC-BGC contribution). Interestingly, two patches scenarios are still in an expanding trend, where they have a chance to sequester more carbon until the forest reaches maximum mangrove community density, depending on the forest composition65. It shows the benefit of strategically providing a less competitive environment and colonization space to allow for larger successful landward seedling establishment.

Our simulations suggest that SOC is the highest contributor among the carbon pools for all scenarios. The SOC content exhibits a significant difference where scenarios above MHWN accumulated more, due to mangroves-induced drag and sediment trapping efficiency. Recent studies3,54,70,71 mention SOC is largest contributor to the total ecosystem carbon stock. However, it is essential to note that the SOC’s source in those studies is majorly of autochthonous origin, where we assume the availability is still limited within our domain. Additionally, our modelling exercise resembles mangrove afforestation within 20 years after planting, in which allochthonous carbon prevails54,72. Scenario B represents an exception due to its deep scarp erosion pattern and the reversal of the rising SOC trend after year 10. Planting above MHWN in general stores 10% more soil organic carbon, whereas the below MHWN scenarios always underperform in comparison to the scenario without a restoration attempt.

Consistent with Fig. 6c, there is no substantial difference in the biomass carbon pools among scenarios as shown in Fig. 7. The finding in biomass carbon seems inconsistent with the findings of the previous section, i.e., a twofold increase in canopy cover (Fig. 5) and a clear rise in tree number (Fig. 6a). The number of trees-diameter histograms can explain this phenomenon in Fig. 6b. The younger trees dominate the community of the forest with \({D}_{130}\) <5 cm. This is particularly evident when the community exceeds the age of 10 years. Although the younger trees are dominant in number, the total biomass is dominated by the contribution of surviving first-generation trees.

By the end of year 20, the potential total carbon benefit ranges from 64% (scenario C) to 103% (scenario G) increase compared to the scenario without restoration. The scenarios above MHWN consistently accumulate more carbon, gaining above 94% increase in carbon potential. The below MHWN scenarios can only achieve a maximum of 80% carbon (details are provided in the Supplementary 3). Considering that mangroves for scenarios above MHWN still show expansion, this subtle placement difference would potentially sequester more in the biomass pool when the mangrove community reaches maturity. The two-patches strategy helps to gain more carbon in comparison with a single and lower patch, i.e., Scenario G (patch below MHWS + above MHWN) vs Scenario E (above MHWN) and Scenario F (patch below MHWN + above MHWN) vs Scenario C (below MHWN). Although the lower patch in the two-patches strategy has similar hydrodynamic characteristics as the single patch at the same elevation, the gap allows sediment to settle, distributing the sediment along the gap and interior, thus increasing the soil surface elevation. The patches practically increase the probability of the seedlings establishing and expanding in the seaward direction of the upper patch and the landward direction of the lower patch. Consequently, the scenarios of higher WoO, higher propagules production, and less energetic environment allow two-patches scenarios to sequester more carbon than single patch strategy, thus providing a higher climate change mitigation potential. This study provides an evidence-based restoration practices, with a conservative estimate excluding the contribution of autochthonous carbon pools that is relatively small for an afforestation project like this setting70, but would be dominant in natural mangrove forests70,71

Implications for mangrove restoration management

Most studies link tree species to carbon sequestration potential73,74 or focus on the optimisation of restoration strategies in terms of ecology26,75, intervention approach76, species type(s)77, geomorphic settings78, or planting density79. These studies typically cover a large spatiotemporal scale, i.e., minimum at the forest scale, where gradual and slow changes occur. However, the restoration effectiveness and potential carbon sequestration are also related to the successful establishment43,80, tied to a short spatiotemporal scale81. This implies that smaller-scale processes need to be included to understand the system’s behaviour on a larger scale82. Our modelling results show that a fractional modification, such as strategic placement on a specific elevation with respect to tidal levels, provides a disparate outcome. Whilst many guidelines have provided detailed implementation of mangrove restoration, the analysis on predictive mangrove forest trajectories and carbon stock potential remains lacking. The knowledge gained in our modelling experiments can provide a quantitative indication of restoration pathways during the feasibility assessment or midterm adjustment. It is deemed necessary, especially given the changes in physical-environmental drivers, for instance due to climate change, requires mechanistic understanding of mangrove-mudflat biophysical feedback5,38,39.

The simulations reveal that assessing restoration success should not only be evaluated in terms of large mangrove extent or the potential biomass, particularly if the monitoring period after restoration is limited in time, which we assume is similar to a typical engineering time scale (20 years). As such, morphological development, referred to as SOC accumulation, is the sensitive yet potentially influential parameter demonstrated in this work. For instance, scenario B (above MSL) indeed has the lowest extent (Fig. 5) and the smallest number of trees (Fig. 6a). Nevertheless, the amount of simulated biomass carbon does not significantly differ from other scenarios (Fig. 6c). Our study suggests successful mangrove restoration should not be seen exclusively from biomass carbon accumulation, as it can be a false positive. For example, Fig. 7b shows that the total carbon stock can be even lower than in the no vegetation scenario due to soil loss and the associated soil organic carbon stock. The results indicate the importance of efficient mangrove placement in restoration works to avoid counterintuitive results, where positioning the planting above high water neap elevation promotes sediment deposition, increases seedling establishment probability, and increases the associated carbon stock capacity.

Notably, mangrove restoration is now becoming a global trend, following a global slowdown of forest loss24,35. Mangrove restoration can be a potential solution pathway for climate change mitigation when restoration failures are minimized. In this study, we demonstrate the possibility of modelling predictive behaviour of mangrove forests following their placement and the changes in carbon stocks. From the perspective of mangrove restoration projects, it can potentially reduce the cost by strategically planting the seedlings and avoiding excessive trial-and-error, enhancing outcome predictability and cost-effectiveness of mangrove restoration.

The findings suggest focusing on mangrove restoration in the high water neap elevation in the initial project implementation, supporting the past reports that most failures occurred in the lower intertidal zone25,36. As the intertidal flat has been experiencing rapid loss and degradation in a multitude of scales83, the conservation strategy should consider positive sediment accumulation and improvement of mangrove-mudflat connectivity84. We observed stable tree-number distributions and a consistent pattern of biomass production after year 15. The restored or created wetlands require, on average, 30 years to recover to pre-disturbance forest structure. However, the recovery state in ecosystem structure and function remains lower than reference state85. The recovery state period would also depend on spatial scale and specific indicators, such as the wetland’s hydrologic, biological, or biogeochemical characteristics. These conditions emphasize the need for monitoring to guide whether the objectives have been achieved or adaptive actions are required to be implemented.

Guidelines25,26,55 suggest monitoring is mandatory after the restoration project has commenced, with a minimum period of five years with a focus on assessing the success of seedling establishment and tree recruitment. At the same time restoration is a long-term effort and 10+ years of monitoring would be preferred. Our findings show distinct forest trajectory developments, in particular after 15 years of forest growth. We suggest monitoring an undisturbed restoration project that can be conducted at least within 20 years, as our findings corroborate. This way, we can assess tree recruitment, species composition, changes in forest structure, and morphological development. Additionally, selecting a 20-year operation and monitoring period is feasible to implement, following the current standard in engineering projects. Although a mangrove’s lifetime can take more than a century, based on our modelling experiment, a 20-year period is adequate to represent life stages, near optimum mangrove structure, and the quasi-equilibrium mangrove-mudflat biophysical dynamics. Our findings corroborate with observations by Salmo et al.86, who shows that the ecological characteristics of restored mangroves would stabilise after 11 years and resemble natural forests after 25 years.

Literature mentions a potentially substantial effect of large, infrequent disturbances, e.g., storms or tsunamis, on the development trajectory of mangrove ecosystems87,88. A combination of storm surge waves and wind defoliates the canopy and potentially topples the trees, altering the forest’s structure and composition. Intense storm surges and currents transport and redistribute sediment along and in the interior of the mangrove coast, changing the sediment properties and affecting local salinity. The changes in local topography and fallen tree debris will disrupt the mangrove wetland’s hydrological regime. Depending on the degree of the damage to the forest’s structural-compositional attributes, these events can lead to dieback and delayed recovery at various rates87,89,90. Thus, the forest trajectories will be temporarily interrupted or reset. Given the large variation of recovery and dependencies on local habitat conditions after a storm, it suggests the need to investigate tailored restoration techniques89 conducive to accelerating the recovery after a large disturbance. In this situation, readjustments of the conservation strategy can benefit from the lessons learned55,88 and quantitative ecosystem models, as applied in the current study. Additionally, low energetic wave conditions also contribute to successful mangrove establishment91. We parameterized this effect in a proxy of critical erosion and burial threshold. In future work, it will be worthwhile to explicitly investigate the effects of disturbance from wind and waves on seedlings’ survivability.