Projecting mangrove forest resilience to sea-level rise on a Pacific Island: Species dynamics and ecological thresholds

Kevin J. Buffington; Joel A. Carr; Richard Mackenzie; Maybeleen Apwong; Ken Krauss; Karen M. Thorne

doi:10.1007/s12237-024-01422-y

Home/Archive/USGS

USGS · 70261031

Projecting mangrove forest resilience to sea-level rise on a Pacific Island: Species dynamics and ecological thresholds

Kevin J. Buffington·Joel A. Carr·Richard Mackenzie·Maybeleen Apwong·Ken Krauss·Karen M. Thorne

View original source ↗Export metadata ↓

Abstract

Mangroves can increase their elevation relative to tidal flooding through biogeomorphic feedbacks but can submerge if rates of sea-level rise are too great. There is an urgent need to understand the vulnerability of mangroves to sea-level rise so local communities and resource managers can implement and prioritize actions. The need is especially pressing for small islands, which have been identified as an area of concern by the IPCC. We developed a generalizable modeling framework for tidal wetlands (WARMER-3) that accounts for species interactions and the belowground processes that dictate soil elevation building relative to sea levels. The model was calibrated with extensive field datasets, including accretion rates derived from 29 soil cores, over 300 forest inventory plots, water level, and elevation. The model included five mangrove tree species and was applied across seven regions around the Pacific Island of Pohnpei, Federated States of Micronesia, where mangrove forest is a critical ecosystem that supports subsistence living for local communities. We explored mangrove resilience and carbon accumulation under six sea-level rise scenarios. We also conducted an analysis to determine the sea-level rise rate threshold above which mangroves would be lost. The results suggest that Pohnpei mangroves can build their elevations relative to low and moderate rates of sea-level rise to prevent submergence, with limited loss of mangrove area through 2150. Under higher sea-level rise rates, however, forest elevation decreased substantially relative to mean sea level and there was extensive loss of mangrove area by that year. Regarding mangrove community composition, for all sea-level rise scenarios, the model predicted a change to increasing relative abundance of flood tolerant species and decreasing relative abundance of high-elevation species, which started to being realized by 2100. Variation in sediment supply, water levels, and elevation capital led to differential vulnerability around the island. We identified a threshold for Pohnpei mangroves where if local sea-level rise rates exceed 7.8 ± 2.2 mm/year they are projected to eventually submerge and be lost. Our modeling framework is novel by addressing both species interactions and critical belowground processes to better understand potential tidal ecosystem responses to sea-level rise.

Explore related subjects

Estuaries and Coasts Article

Source-reported bounding extent: 6.766699284998481° to 7.009358068074874° latitude; 158.07984414004875° to 158.36071659346305° longitude. This indicates report coverage, not an exact sampling location. View area on OpenStreetMap.

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Kevin J. Buffington, Joel A. Carr, Richard Mackenzie, Maybeleen Apwong, Ken Krauss, Karen M. Thorne. 2024-10-04. Projecting mangrove forest resilience to sea-level rise on a Pacific Island: Species dynamics and ecological thresholds. https://doi.org/10.1007/s12237-024-01422-y

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

USGS2026

Drivers of ecosystem metabolism in an anthropogenically modified delta

Estuaries are under threat globally from a wide range of stressors, with rates of ecosystem metabolism – respiration (R), gross primary production (GPP), and net ecosystem metabolism (NEM) – representing a way for resource managers to quantify their impacts. In this study, we assessed how natural and anthropogenic stressors have altered metabolism within the extensively modified Sacramento-San Joaquin Delta (“Delta”) in California, USA. Metabolism was estimated using a diel open-water technique utilizing dissolved oxygen measurements from a high frequency (15 min) water quality monitoring array, and spanned 6 sites across 11 years (2014–2025). Our analyses were driven by a conceptual model of Delta carbon (C) cycling focused on 1) Invasive aquatic vegetation (IAV), 2) Phytoplankton, 3) Wastewater treatment plant (WWTP) upgrades, 4) Large-scale flow alterations, and 5) Wet vs. dry years. We found that allochthonous C fuels net heterotrophy in the outer riverine sections of the Delta, while the heavily vegetated central Delta produces autochthonous C remineralized elsewhere in the system. IAV were the dominant primary producers, with low phytoplankton production except during a historic bloom. WWTP upgrades imparted negligible changes to metabolism, with low chlorophyll a (< 5 µg L −1 ) both before and after. Introduction of a flow barrier – a management response to drought – hydrodynamically isolated deep-water channels from nearby shallow areas, resulting in decreased metabolism indicative of changes to C transport. Contrary to past research, low rainfall years were associated with increased metabolism. This study demonstrates how multi-sensor arrays can be used to quantify ecosystem responses to major environmental stressors.

California↗

USGS2026

Forecasting marsh resilience in the Plum Island Estuary: Integrating uncertainty to inform coastal management

Salt marshes provide critical ecosystem services yet face increasing threats from accelerated relative sea-level rise (SLR). The long-term persistence of these habitats depends on their ability to vertically accrete and laterally migrate into adjacent uplands. Predicting the location and viability of these future marsh corridors is therefore a primary challenge for coastal managers; one often hindered by uncertainties in future water levels and topographic data accuracy, particularly in vegetated wetlands where light detection and ranging (lidar) data systematically overestimate elevation. This study applies an uncertainty-informed modeling framework to forecast marsh resilience in the Plum Island Estuary, Massachusetts. We integrated the Sea Level Affecting Marshes Model (SLAMM) with a Monte Carlo error analysis, generating an ensemble of bias-corrected digital elevation models (DEMs) alongside three probabilistic SLR scenarios (Intermediate, Intermediate High, High) through 2100. We synthesized these simulations into a “marsh fate” classification map to distinguish between resilient core habitats, vulnerable zones requiring intervention, and likely migration corridors. While the marsh platform exhibits resilience under Intermediate scenarios, a critical tipping point emerges under the High SLR scenario. In this case, accelerating inundation overwhelms accretion capacities, driving a landscape-scale regime shift. By integrating the interacting uncertainties of SLR trajectories and initial marsh elevation, we project 2100 net marsh coverage outcomes ranging from a 10.5% gain to a 23.2% loss. These results demonstrate that incorporating elevation uncertainty is essential for detecting non-linear threshold behaviors that deterministic models often miss. The resulting framework empowers managers with a spatially explicit tool to prioritize land acquisition and restoration efforts in the face of climate uncertainty.

Massachusetts↗

USGS2026

Incorporating wetland drowning thresholds to assess the spatiotemporal variability in potential wetland migration and submergence from sea-level rise

In the coming decades sea-level rise rates could exceed the upper bounds of coastal wetland vertical movement leading to wetland submergence. Consequently, understanding the spatiotemporal variability of wetland migration and submergence at regional and national scales can inform decision making. In this study, we developed an approach that addresses elevation error in digital elevation models for wetlands, incorporates uncertainty in current and future water levels, and utilizes recent literature on coastal wetland drowning thresholds. We used our framework to predict wetland change with sea-level rise for the middle and upper Texas coast—a region that is experiencing some of the highest relative sea-level rise rates in the United States. Our elevation-based approach produced contemporary wetland maps that had a high level of agreement (> 87–91%) with existing national wetland maps. For the intermediate relative sea-level rise scenario, extensive change is predicted within the study area by 2100, including a net decrease in coverage of irregularly oceanic-flooded wetlands (-37%) and regularly oceanic-flooded wetlands (-23%). The wetland migration space of these two wetland zones is largely made up of existing wetlands (> 70%), but also includes other upland areas, such as pasture/hay and grassland. By breaking down our results for the full region and two distinct focal areas, we highlight the spatiotemporal variability of wetland change. Our scalable approach provides natural resource managers with information for developing location-specific strategies for balancing the trade-offs associated with wetland change from sea-level rise.

Texas↗