Mangrove forests can be an effective coastal defence in the Pearl River Delta, China

The study addresses how mangrove forests function as nature-based coastal defences within the complex hydrodynamics of an urban delta. While mangroves and salt marshes are known to attenuate waves and can reduce storm surges, the controlling factors of surge attenuation in estuary/delta systems remain insufficiently understood, limiting large-scale implementation in mega-cities. The Pearl River Delta (PRD), home to cities such as Shenzhen and Guangzhou, experiences significant flood risk from tides, river discharge, and typhoon-driven storm surges. The region has seen substantial historical loss of mangroves followed by partial restoration. The research aims to quantify both local (within-wetland) and non-local (upstream along estuary channels) water-level attenuation by mangroves during an extreme event (Typhoon Hato, 2017) and to assess sensitivity to vegetation properties, patch geometry, location within the delta, and future sea level rise scenarios. The central hypothesis is that mangrove effectiveness varies in space and time, depends strongly on patch width and drag properties, and that patches in estuary channels can also provide upstream attenuation while potentially causing downstream amplification (blockage effects).

Prior research has established wave height reduction by coastal wetlands, including mangroves, primarily at local scales. However, factors governing storm surge reduction by vegetation are less well constrained, with most empirical and modelling work focused on salt marshes. Reported within-wetland surge attenuation rates range from roughly 1.7–25 cm km−1 in salt marshes and up to 18–50 cm km−1 in mangroves. Effectiveness depends on vegetation density, height, width, fragmentation, storm characteristics (intensity, duration, speed, track), and larger-scale geomorphic settings. Only a few studies have investigated estuary-scale, non-local attenuation (upstream effects) and these largely concern salt marshes in European estuaries. There is a lack of similar estuarine-wide analyses for mangrove systems, especially in complex deltas. Furthermore, mangroves are often combined with hard infrastructure (seawalls/levees) to reduce overtopping and loads, potentially lowering engineering costs. Designing effective nature-based defences requires understanding how vegetation interacts with estuary morphodynamics, tides, surges, and river discharge at system scales.

The study uses the Finite Volume Community Ocean Model (FVCOM) for the South China Sea and the Pearl River Delta with an unstructured grid (coarse >10 km offshore, refined to ~100 m in delta channels) and 25 sigma layers. The model has been previously validated against coastal and riverine tide gauges and during Typhoon Hato. Typhoon Hato (2017) atmospheric forcing (winds and pressure) was generated using a Holland parametric model driven by IBTrACS best-track data to better capture peak winds. Vegetation parameterization: Mangrove drag is represented as a momentum sink within the water column (quadratic drag law), distributing drag across vertical layers occupied by vegetation. The force F = -½ ρ C_D A |u| u N, where ρ is water density, C_D the drag coefficient, A the projected frontal area per tree, u the velocity vector, and N plant density per model element. A homogeneous dense mature tall forest density of 0.5 plants m−2 is assumed. C_D values are based on flume measurements for PRD mangrove species: C_D = 4 (with leaves, high-drag) and C_D = 2 (without leaves, low-drag). Projected frontal area per tree is A = 3 m² (with leaves) and A = 1 m² (without leaves), accounting for branch realignment and partial submergence; mangrove height is 3 m. No vegetation-induced turbulence generation/dissipation is included. Study sites and experiments: - Shenzhen Bay: Existing mangrove reserve represented with patch widths of 300 m (theoretical), 600 m (approx. present), and 900 m (theoretical), each with high- and low-drag properties. The effect on total water level, tide, and surge is evaluated at coastal nodes during the Hato event. - Upper estuary near Guangzhou: Two hypothetical 300 m-wide mangrove corridor scenarios placed within confined channels: a short 6 km patch (Hualong) and a long 16 km patch (Hualong–Lianhuashan), each tested with high- and low-drag. Width sensitivity was not tested here due to channel constraints. Sea-level rise scenarios: High-drag experiments were repeated under uniform boundary mean sea level increases of 30, 50, and 90 cm. The model assumes a fully protected coastline (mangroves in front of seawalls), no additional land inundation, and no sediment accretion. The SLR scenarios align with regional projections for RCP8.5 and RCP4.5 time horizons. Diagnostics: Separate tide-only and fully forced simulations were used to compute surge (difference between total water level and tide-only). Spatial differences in surge and currents between vegetation and no-vegetation cases were mapped to assess local and non-local effects, including potential blockage and current speed changes relevant to erosion risk.

- Local attenuation in Shenzhen Bay during Typhoon Hato: - A 600 m high-drag mangrove patch reduced the coastal maximum total water level from ~3.3 m to ~2.1 m. - The maximum surge at the coast was halved from ~2.8 m to ~1.4 m. - A 600 m low-drag patch and a 300 m high-drag patch provided reductions before peak but did not reduce the surge peak at the coast (peak reduction ~0–0.05 m). - A 900 m high-drag patch strongly reduced the peak, bringing the total water level to ~0.5 m and suppressing the surge to ~−0.3 m at the coast; a 900 m low-drag patch reduced surge by ~1 m. - Tides were phase-delayed by vegetation (up to ~0.5 h for 600 m and ~1 h for 900 m high-drag), contributing to apparent pre-peak changes in tidal elevation and surge. - Prior to peak (before 23 Aug 2017 04:00 UTC), vegetation caused transient increases in total water level (~0.3 m) and surge (~0.6 m) due to delayed inflow and tidal modulation effects. - Spatially, vegetation reduced surge and current speed within the patch during peak conditions, while transient surge increases and current speed changes elsewhere indicated temporary piling up in the bay before the peak. Current speed reductions within the patch imply local erosion mitigation, with some increases outside the patch. - Sensitivity to sea-level rise (Shenzhen Bay): - Under 30 cm SLR, mangroves still reduced surge, but peak surge rose to ~2.2 m (greater than no-SLR case). - With 50 cm and 90 cm SLR, mangrove friction became less effective due to deeper water; the peak surge was ~2.5 m for both cases with diminishing vegetation impact. - At 90 cm SLR, surge reduction at the peak was nearly zero, though reductions persisted before the peak. Simultaneously, SLR reduced wind effectiveness in generating surge (deeper water attenuates wind stress effect). - Non-local upstream attenuation in upper estuary (near Guangzhou): - A 16 km high-drag corridor (Hualong–Lianhuashan) produced upstream surge reductions exceeding 0.2 m near Guangzhou and total water level reductions of about 0.2–0.3 m along the affected shores. - Along the west shore (within/adjacent to the patch), maximum reductions reached ~0.3 m (total) and ~0.2 m (surge); along the east shore (channel toward Guangzhou), reductions were ~0.2–0.3 m (total) and ~0.1–0.2 m (surge). - Shorter 6 km patches yielded smaller benefits (<0.1 m reductions in both total water level and surge). High-drag outperformed low-drag. - Blockage effects: When surge approached (07:00–08:00 UTC), the patch acted as a partial barrier, increasing water levels downstream of the patch and diverting flow into eastern channels. This effect diminished after the peak. - Current speeds decreased within the vegetation and in parts of adjacent branches, potentially reducing erosion locally, but increased in some upstream branches near Guangzhou, implying possible localized erosion risk there. - Attenuation rates: For a 600 m high-drag patch, apparent surge attenuation reached up to ~1.4 m (≈233 cm per km) under typhoon conditions, exceeding typical salt marsh values reported in literature. - Under SLR in the upper estuary, vegetation still reduced upstream total water levels with spatial patterns similar to the present-day case, though interactions with tides and river discharge make local changes complex and non-uniform across channels.

The findings confirm that mangrove forests can substantially mitigate extreme water levels in a mega-delta but that their effectiveness is highly location- and configuration-dependent. Wide, high-drag patches along open coasts/bays produce strong local within-wetland attenuation of storm surges and currents, lowering peak levels and potentially reducing coastal erosion. In contrast, narrower patches in confined estuary channels can provide meaningful non-local (upstream) attenuation that benefits vulnerable urban areas, such as Guangzhou, although they may temporarily amplify levels downstream due to blockage effects. These insights address the key research question by demonstrating that a single linear attenuation factor is inadequate for complex deltas: attenuation varies in space and time, depends on patch width and vegetation drag, and interacts with tides, river discharge, and storm timing. The results have practical relevance for integrated coastal management in the PRD, where combined nature-based and engineered defences are likely needed given the limited protection standards of existing hard infrastructure and rapid urban expansion. The modelling provides a basis for designing mangrove corridors and coastal patches that maximize protection while minimizing adverse redistribution of water levels. Under SLR, mangrove friction becomes less effective in deeper waters under a fully protected shoreline scenario; nevertheless, pre-peak benefits and upstream attenuation in channels persist, highlighting the need to co-design vegetation with adaptive coastal infrastructure and consider room for inland wetland migration where feasible.

This work demonstrates, via a validated 3D hydrodynamic model, that mangrove forests can be effective nature-based coastal defences in the Pearl River Delta. Key contributions include: quantification of local attenuation in Shenzhen Bay during an extreme typhoon, demonstration of significant upstream attenuation from narrow mangrove corridors in confined channels near Guangzhou, and assessment of sensitivity to vegetation properties, patch geometry, and sea-level rise. Wide, high-drag patches yield large local reductions in surge and total water level; strategically placed narrower patches in estuary channels can reduce upstream extremes by up to several decimeters. Implications are that delta-scale planning should incorporate both local and non-local effects of mangroves, potentially combining them with hard defences to lower overtopping and structural requirements. Future research should include: in situ validation of surge attenuation during extreme events; exploration of a broader range of storm characteristics and tracks; dynamic vegetation responses (mortality, defoliation) and turbulence effects; sediment dynamics and vertical accretion/inland migration under SLR; optimization of patch layouts to minimize blockage-induced amplification; and integration of flood risk reduction with other mangrove ecosystem services (carbon storage, biodiversity, fisheries, water quality).

- Coastline assumption: A fully protected shoreline with seawalls is assumed; no additional land inundation or accommodation space was considered. - Sediment dynamics: No sediment transport, accretion, or wetland vertical growth under SLR was modelled. - Vegetation parameterization: Uniform plant density (0.5 m−2), simplified projected area (A = 1–3 m²) and heights (3 m) were used; species-specific variability and site-specific compositions were not explicitly represented. - Drag coefficients: C_D values (2 and 4) were chosen from the lower range of flume-derived estimates for PRD species; real-world variability, defoliation dynamics, and storm damage were simplified. - Turbulence: Vegetation-induced turbulence generation/dissipation was not included. - Geometry constraints: Upper estuary analyses did not vary patch width (channel too narrow) and Shenzhen Bay did not vary patch length (matched existing reserve). - Event scope: Focused on one typhoon (Hato) and a typical track; broader storm variability (duration, intensity, alternate tracks) was not exhaustively explored. - SLR scenarios: Limited to 30, 50, and 90 cm; assumed uniform boundary increase and drowned, ineffective mangroves beyond ~1 m SLR under the protected-shoreline assumption. - Validation: Prior validation exists for water levels and tides, but direct in situ surge attenuation measurements within mangroves during extreme events were unavailable due to safety/logistical constraints.