Task 4: Future Scenarios

Dr. Wei Wu in the Ecospatial Lab at the Gulf Coast Research Laboratory is leading the future marsh structure and function simulations task.

Synopsis:

A mechanistic simulation model will be developed to predict the impact of future sea level rise (SLR) on salt marshes integrating the vegetation dynamics, as well as hydrological and sediment processes. Consequently, we aim to predict future changes of coastal marshes by 2100, including inundation and conversion of vegetation, under multiple environmental stressors that interact with each other. In the model, the elevation of the marsh surface changes due to sea-level rise, subsidence, sedimentation from organic and mineral matter, and erosion. Organic matter is added to the sediment column through growth of belowground biomass and the deposition of litter on the column’s surface, and is lost due to decomposition. Inorganic material is added to the column surface from particles settling from tidally induced floods and trapping of particles by salt marsh macrophytes. Erosion at the edge of marshes and extra deposition in the inner marsh due to tropical cyclones will also be simulated. We will conduct retrospective analysis to make sure the model can simulate the observed wetland changes. This will improve the predictability of the model through calibration and validation.

Methods

We developed a mechanistic model to predict the impact of SLR on the spatial distribution of coastal wetlands. The mechanistic model incorporated hydrodynamic, geomorphological, and ecological processes, important drivers for elevation change at wetland platform, and therefore wetland change. The mechanistic model was adapted from both the Marsh Equilibrium Model (MEM) for simulating accretion rates and a simplified hydrodynamic model for simulating erosion rates. We applied this model to a case study selecting a micro-tidal estuarine area with limited upland freshwater and sediment input in the Grand Bay National Estuarine Reserve. This study area represents conditions indicative of highly vulnerable coastal wetlands under SLR, namely low sediment inputs, a lack of freshwater supplied new sediments, and limited upland migration area for coastal vegetation.

We predicted whether wetland habitat would be kept or converted into open water at each time step in the model. We found the lower 2.5% quantile of elevation for salt marshes in our study area to be very close to mean low water, so we used mean low water as the lower elevation limit of salt marsh. We assumed the salt marshes were converted to open water if the elevation was lower than this lower limit. We compared the vegetation biomass, accretion rate, and velocity at the mud bed simulated from our model to the measured data and other models’ simulations. We also compared the simulated 2007 wetland distribution calculated by our newly developed model to the National Wetlands Inventory (NWI) map in 2007 (the reference map).

As the temporal scale is a critical factor affecting resilience of coastal wetlands to SLR, we applied the model to simulate wetland dynamics by 2050 and 2100 under the scenarios of a variety of SLR rates ranging from 4 mm/yr (current sea level rise rate) to 20 mm/yr (high end of SLR rate predictions from the IPCC 2013) using an increment of 0.5 mm/yr. This provided the predicted total wetland areas under 33 different scenarios of SLR rates, and from this we derived the thresholds of SLR rate beyond which coastal wetlands will fragment into a less desirable state with much smaller emergent wetland areas and loss of marsh to open water.

Results

Our model simulates wetland change over ~ 20 years from 1988 to 2007, both in location and amount of change. Land persistence is simulated correctly over 90.4% of the study area (Figure 2). During the validation period, the model could correctly simulate 48% of the reference (true) change that occurred between 1988 and 2007 at the Grand Bay NERR. For the time period 1988 to 2007, the simulated average biomass for the whole study area at the Grand Bay NERR was 808 g/m², consistent with the measurements at the same area in the literature. The simulated average accretion rate for the whole study area was 1.8 mm/yr, similar to the average measured data which was 1.4 mm/yr using the marker horizon method. The simulated average flow velocity, necessary to calculate erosion rate, at the mudflat bed in Grand Bay is ~ 7m/sec in the present day, and it increases to 11 m/sec by 2100 with a SLR rate of 4.1 mm/yr, very similar to the predicted velocity of 6.1 m/sec and 12.2 m/sec for the current time and 2100 derived from simulations using the more complex hydrodynamic Advanced Circulation model applied to Grand Bay.

There are two major processes that affect elevation of the wetland platform: deposition and erosion. We find through the dynamic modeling that deposition reduction in this retrograding delta has a larger impact on future wetland area compared to deposition increase or change in erosion rate. The change of erosion and increase of deposition have a similar effect for 2050 and 2100 but the reduction of deposition will have a larger impact on wetland loss further into the future (2100 vs. 2050). This finding is particularly important for this case study where storms may translocate sediments on to the marsh, in essence providing a mechanism to increase deposition within localized areas. Further, vegetation productivity was a more important factor than suspended sediments to determine the deposition rate in this freshwater limited estuary (Table 1), consistent with the main accretion mechanism in marine-dominated and sediment-deprived systems. We presume that coastal wetlands in less sediment–starved river delta systems, such as the Pascagoula River, would have better resilience against erosion losses as SLR increases, and highlight here that our study system should be considered on the highly vulnerable end of the coastal wetland continuum.

It is critical to account for temporal effects in assessing the resilience of coastal wetlands to SLR and deriving the thresholds of SLR rate and SLR acceleration rate. When our target year changes from 2050 to 2100, both threshold levels decrease substantially, indicating a higher likelihood of marsh habitat collapse in 2100 than 2050 (Figure 3). This has important implications for designing climate mitigation and adaption plans. While it seems in this case study that coastal wetlands are resilient to SLR by 2050 under both low and high emission scenarios, it is very likely that highly vulnerable coastal wetlands like Grand Bay could collapse by the end of the century, especially under the high warming scenario.

Summary

We presented the first study on the threshold of SLR acceleration rate, and the first comprehensive threshold analysis which accounts for the temporal scale, the interaction of SLR with other environmental factors, and landscape metrics used. We tested them in a highly vulnerable and sediment starved estuarine system. Based on the total wetland area, the threshold of SLR rate for our retrograding delta study area is 11.9 mm/yr for 2050, and it drops to 8.4 mm/yr for 2100. This study illustrates a transferrable and useful method for evaluating coastal wetlands’ nonlinear response to SLR, especially in marine-dominated systems, and facilitating design of mitigation and adaption policy under future climate change.

An education-outreach Wetland Loss Estimator web tool was developed based on these results to make this information more accessible to middle-school students.

Recommended Citations

Wu W, Biber P, Bethel M. 2017. "Thresholds of sea‐level rise rate and sea‐level rise acceleration rate in a vulnerable coastal wetland." Ecology and Evoution 7: 10890–10903.

https://doi.org/10.1002/ece3.3550

Wei W, P. Biber, D. R. Mishra, S. Ghosh. 2020. "Sea-level rise thresholds for stability of salt marshes in a riverine versus a marine dominated estuary." Science of The Total Environment 718: 137181

https://doi.org/10.1016/j.scitotenv.2020.137181