Area 4: Integration and Forecasting

Description Objectives Outcomes Show All  

Overview

In Area 4 our objectives are to synthesize information collected in Areas 1-3 to produce an integrated picture of habitat provisioning and carbon flow across the landscape, and evaluate how changes in salinity and inundation may change these services in the future. We will accomplish this with a combination of integrative modeling, empirical observations, and remote sensing.

We use a combination of integrative modeling, empirical observations, and remote sensing to produce an integrated picture of habitat provisioning and carbon flow across the landscape, and evaluate how changes in salinity and inundation may change these services in the future. Major activities include A) develop an integrative model that uses a hydrodynamic model (FVCOM), a soil model, and 3 different semi-empirical plant models to predict salinity and inundation patterns, porewater salinities, and plant responses over different time scales, and B) use combined model output to evaluate habitat provisioning and C flow under different scenarios.

Components

Area 4A. Develop an integrative model

Hydrodynamics play a critical role in the distribution and transport of water and materials across the GCE domain. Consequently, an accurate hydrodynamic model is a necessary first step for our integrative modeling efforts. FVCOM will provide information on salinity and inundation patterns of the water that floods the marsh, which will be used in a soil model to predict porewater salinity and soil water content (described below), and in 3 different semi-empirical plant models. The Spartina productivity model (Area 3A) will provide information on individual plant responses to salinity and inundation; the plant community model (Area 3B) will provide spatial information on vegetation dynamics during transitions; and a modified version of the SLAMM model will predict habitat shifts at the landscape level.

Although these models will exchange information offline, they could inform each other, and our goal in GCE-3 is to lay the groundwork for fully integrated mechanistic models with multiple feedbacks. The effects of parameter sensitivity (e.g. bottom friction on the marsh in the hydrodynamic model, below-ground production parameters in the Spartina productivity model) will be investigated for each model individually and for how sensitivities and uncertainties propagate between models.

Area 4B. Evaluate habitat provisioning and carbon flow under different scenarios

We will use these models to run a series of scenarios to evaluate, through hindcasting and forecasting, how pulses and presses in our major drivers (sea level, river flow, precipitation, temperature, groundwater input, and overland runoff) will affect the domain. The models will be used to predict salinity and inundation patterns, porewater salinities, and plant responses over different time scales. Predictions will be evaluated in terms of habitat provisioning and C flow (see below). Climate change will be examined using bias-corrected, downscaled projections of the IPCC model results (Maurer et al. 2007). Human alterations will be evaluated by simulating potential modifications to shoreline armoring and overland runoff based on build-out scenarios from McIntosh County, as well as modifications in the greater Altamaha watershed (e.g. new reservoirs upstream). We will also consider scenarios with feedbacks to human behavior (e.g. building sea walls as sea levels rise, green developments). Our archeological studies will allow us to consider patterns in the pre-development landscape. We will also be able to modify our scenarios in response to experimental results showing changes in creek geomorphology or the top-down effects of blue crabs on plant production.

To evaluate dynamic habitat, we will use the continuous salinity data from our sonde network to map the locations of fresh, oligohaline (<5 PSU), mesohaline (5-15) and polyhaline (15-30) conditions across the domain, and determine how these habitat locations vary over lunar, seasonal, and annual cycles. These will be compared with results from FVCOM which will also be used to evaluate how these salinity ranges might shift given different scenarios.

Two major questions for global and regional C cycling are 1) whether the coastal zone is a net source or sink of CO2 to the atmosphere and the ocean, and 2) how such C fluxes might change over time. We will, in combination with the habitat analyses, evaluate how the changes that might occur in response to changes in salinity and inundation will affect carbon cycling, with the goal of developing new hypotheses about the implications of climate and human activities for the coastal C budget.

 

Research Objectives

  • 4A.1  Run FVCOM to predict salinity and inundation (yr 3-4)
    • Description:  Hydrodynamics play a critical role in the distribution and transport of water and materials across the GCE domain. Consequently, an accurate hydrodynamic model is a necessary first step for our integrative modeling efforts. FVCOM will be run for the GCE domain to provide information on salinity and inundation patterns of the water that floods the marsh, which will be used in a soil model to predict porewater salinity and soil water content, and in 3 different semi-empirical plant models.
    • Participants:  Renato Castelao, Daniela Di Iorio
  • 4A.2  Run the soil model to predict porewater salinity (yr 4-5)
    • Description:  Marsh porewater salinity varies over relatively long time scales (seasonal rather than tidal) and is generally higher than water column salinity due to evaporation and transpiration. We will build a soil model that tracks the inundation history (provided by FVCOM) of intertidal and subtidal cells in the hydrodynamic model on an aereal basis. The model will initially be run in each cell individually, and will use basic ET information, with plant transpiration parameterized from the literature (e.g. Giurgevich & Dunn 1982), to predict porewater salinity and soil water content in the top 10 cm of soil for each cell individually. Model calibration and validation will be performed using 2 existing ground surveys of porewater salinity and soil moisture collected in different vegetation classes (n=370 plots each). Porewater salinities predicted for areas of Spartina will be used as input to the Spartina production model to predict individual plant growth, which in turn will guide improvements to the parameterization of plant response to salinity in the community model. At the community level, we will use the porewater data to predict vegetation composition and class, which can be calibrated with our observations that have shown good separation of plant species based on porewater salinity and soil moisture (Lynes 2008).
    • Participants:  Christof Meile
  • 4A.3  Run the plant models to predict vegetation response yr (2-6)
    • Description:  The Spartina productivity model (Area 3a) will be run to provide information on individual plant responses to salinity and inundation; the plant community model (Area 3b) will provide spatial information on vegetation dynamics during transitions.
    • Participants:  Adrian Burd, Marc Garbey, Steve Pennings
  • 4B.1  Develop scenarios (yr 3)
    • Description:  We will use these models to run a series of scenarios to evaluate, through hindcasting and forecasting, how pulses and presses in our major drivers (sea level, river flow, precipitation, temperature, groundwater input, and overland runoff) will affect the domain. The models will be used to predict salinity and inundation patterns, porewater salinities, and plant responses over different time scales. Predictions will be evaluated in terms of habitat provisioning and C flow (see below). Climate change will be examined using bias-corrected, downscaled projections of the IPCC model results (Maurer et al. 2007). Human alterations will be evaluated by simulating potential modifications to shoreline armoring and overland runoff based on build-out scenarios from McIntosh County, as well as modifications in the greater Altamaha watershed (e.g. new reservoirs upstream). We will also consider scenarios with feedbacks to human behavior (e.g. building sea walls as sea levels rise, green developments). Our archeological studies will allow us to consider patterns in the pre-development landscape. We will also be able to modify our scenarios in response to experimental results showing changes in creek geomorphology or the top-down effects of blue crabs on plant production.
    • Participants:  Merryl Alber, Clark Alexander, Adrian Burd, Victor Thompson
  • 4B.2  Evaluate C flow (yr 3-6)
    • Description:  Two major questions for global and regional C cycling are 1) whether the coastal zone is a net source or sink of CO2 to the atmosphere and the ocean, and 2) how such C fluxes might change over time. A C budget for the South Atlantic Bight based on preliminary studies in the GCE domain (Cai 2011) suggested that the marsh is a sink for atmospheric CO2 and that the marsh-estuarine complex exports large quantities of carbon to the coastal ocean (Fig. 3). Two of the uncertainties in this budget, estimates of atmospheric exchange and lateral exchange from the marsh, are being addressed directly in Area 3a. We will also have a greatly improved understanding of water transport from the hydrodynamic model, and can use plant distributions and our monitoring data to estimate NPP throughout the domain. Finally, we are adding observations of DIC, DOM composition and source and soil C storage (Area 3a) to better quantify other aspects of the C budget. Taken together, these measurements will allow us to greatly improve our estimates of C flow, and in particular to test the hypotheses that 1) marsh vegetation is the dominant source of OC that drives net heterotrophy and CO2 degassing in estuarine waters whereas most riverine OC is exported to the coastal ocean; and 2) the importance of lateral transport of C from the marshes varies spatially (across the 3 GCE Sounds) and with time. We can also, in combination with the habitat analyses, evaluate how the changes that might occur in response to changes in salinity and inundation will affect these conclusions, with the goal of developing new hypotheses about the implications of climate and human activities for the coastal C budget.
    • Participants:  Merryl Alber, Wei-Jun Cai, Chris Craft, Chuck Hopkinson, Monique Leclerc, Janet Reimer
  • 4B.3  Evaluate habitat provisioning (yr 3-6)
    • Description:  We are interested in how habitat availability varies across the domain over different time scales (tidal to decadal), and the strength of forcing that triggers community reordering and habitat shifts. To evaluate dynamic habitat, we will use the continuous salinity data from our sonde network to map the locations of fresh, oligohaline (
    • Participants:  Merryl Alber, Clark Alexander, Renato Castelao, Daniela Di Iorio

Research Outcomes by Objective

  • 4A.1  Run FVCOM to predict salinity and inundation (yr 3-4)
    • 2014 report:

      Activities:  Begins yr 3

    • 2015 report:

      Activities:  Maps of residence time and connectivity have been produced, and we are investigating how those vary with forcing. Changes in salinity in the estuary in response to sea level rise and storms are also being investigated.

    • 2016 report:

      Activities:  FVCOM has been implemented for both the Duplin River and the larger GCE domain.

      Results:  Wang et al. (submitted) used the FVCOM model to evaluate salinity variability, residence times, and connectivity in the GCE domain (Fig. 12).

  • 4A.2  Run the soil model to predict porewater salinity (yr 4-5)
    • 2014 report:

      Activities:  Begins yr 3

    • 2015 report:

      Activities:  Model simulations have been performed to assess seasonal patterns and inter-annual variations. We also performed a sensitivity analysis to quantify the role of different external forcings (e.g. precipitation), or model parameters (e.g. soil hydraulic conductivity).

    • 2016 report:

      Activities:  We have run the soil model over several years to assess seasonal and interannual variability, and have begun a comparison of the model with patterns seen in Landsat data.

      Results:  A sensitivity analysis conducted on the soil model shows that porewater salinity in the Spartina zone is controlled by tidal salinity whereas high marsh plants are sensitive to changes in ET and precipitation.

  • 4A.3  Run the plant models to predict vegetation response yr (2-6)
    • 2014 report:

      Activities:  We have developed scenarios and the model infrastructure for the Spartina model. However, biomass predictions in initial runs diverged from observations after 18-24 mo, partly because resource allocation was not incorporated into the model. (See Obj. 3A7).

    • 2015 report:

      Activities:  The Spartina model is currently being validated against biomass data that was collected in the field as part of the GCE LTER and also against literature data from other locations.

    • 2016 report:

      Activities:  The plant model is being revised to include mechanistic transport from above to below ground tissues based on field observations. We have also been working on developing links to the GCE soil model.

  • 4B.1  Develop scenarios (yr 3)
    • 2014 report:

      Activities:  Begins yr 3

    • 2015 report:

      Activities:  Calendar year daily statistics of Altamaha River discharge over the period 1932-2014 were calculated and binned into 3 levels of discharge in order to define periods of wet, dry, normal and variable years for use in modeling efforts (Activities Fig. 4).

      Plans:  We plan to develop further scenarios that can be used in model runs.

    • 2016 report:

      Activities:  The hydrodynamic model for the Duplin is currently running for the time period Aug 2012 to Dec 2015 which will simulate dry, wet, and normal years of river discharge effects. The model for the entire domain has been run for these years as well and is being used to run simulations representing different levels of sea level rise.

  • 4B.2  Evaluate C flow (yr 3-6)
    • 2014 report:

      Activities:  Begins yr 3

    • 2015 report:

      Activities:  To better understand the factors controlling seasonal CO2 fluxes and the extent of autotrophy/respiration in the coastal South Atlantic Bight (SAB), we measured pCO2 from the GCE domain in each season (April, July, Sept., Dec.) as part of the series of oceanographic cruises conducted this past year. Underway pCO2 was measured in cross-shelf transects, and discrete samples were also collected for DIC,Total Alkalinity (TA), and pH measurements. These samples are currently being processed. We also collected and dated 4 cores at GCE monitoring sites to examine carbon sequestration rates.

      Results:  High resolution maps of sea surface pCO2 over the region suggest that while the estuarine zones are a strong source of CO2 to the atmosphere, the SAB shelf is a net sink of atmospheric CO2 during all seasons (Results Fig. 8). These data will also be used to evaluate the extent of in-situ DIC generation and export from estuarine zones to the coastal ocean.

    • 2016 report:

      Activities:  Samples for DIC, Total Alkalinity, and pH collected during cruises conducted in the South Atlantic Bight during year 3 are being processed.

      Results:  Observations of DIC, TA, and pH collected during GCE cruises are being used to validate the NOAA Gray's Reef National Marine Sanctuary CO2 time series.

  • 4B.3  Evaluate habitat provisioning (yr 3-6)
    • 2014 report:

      Activities:  Begins yr 3

    • 2015 report:

      Activities:  We are starting to use FVCOM to evaluate how salinity ranges (and hence dynamic habitat) will vary with sea level rise (Activities Fig. 5).

      Results:  McFarlin et al. (2015) evaluated the effects of the loss of Spartina alterniflora on habitat provisioning for benthic epifauna, macroinfauna and meiofauna. In the GCE domain, abundances of all invertebrate groups and the diversity of macroinfauna were lower in bare plots, with clear separation between infaunal assemblages in bare and reference (=vegetated) plots (Results Fig. 9). In contrast, there was overlap between the assemblages and the abundance of some groups (i.e. meiofauna) increased in bare plots in Louisiana, suggesting that the role of S. alterniflora is context-dependent.

    • 2016 report:

      Activities:  C. Hladik led an effort to correct tidal marsh digital elevation models for salt, brackish, and tidal fresh marshes based on field observations obtained with an RTK GPS.

      Results:  Hladik et al. produced improved habitat classifications for the tidal fresh marshes in the GCE domain (Fig 13).

 
LTER
NSF

This material is based upon work supported by the National Science Foundation under grants OCE-9982133, OCE-0620959 and OCE-1237140. Any opinions, findings, conclusions, or recommendations expressed in the material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.