Coastal salt marsh ecosystems provide crucial ecosystem services including fisheries production, storm surge abatement, and pollutant removal. They also are aesthetically significant ecosystems, positively impacting quality of life and coastal economy. Over 80% of coastal salt marshes in Southern New England have been lost to development, and the remaining marshes are being impacted by human activities to varying degrees. In particular, global warming and its effects of sea level rise are among the most important threats to coastal salt marsh ecosystems today. To keep pace with sea level rise, salt marshes must vertically accrete and/or migrate landward. Global warming poses a unique threat because it not only leads to accelerated sea level rise, but it may also indirectly impact accretion rates by slowing sediment accumulation as a consequence of changes in salt marsh productivity. If salt marshes are unable to keep pace with sea level rise through vertical accretion, they must then migrate landward or be lost, however this ability to migrate is constrained by the state of the surrounding land cover. Thus, a predictive understanding of how coastal salt marsh ecosystems will be impacted by and respond to global warming and sea level rise requires a comprehensive understanding of the controls of vertical accretion and limits to migration.
Given the many challenges that global warming and accelerated sea level rise poses for coastal marsh ecosystems and communities, the need for innovative solutions that couple scientific understanding with long-term planning is essential. Towards this end, our interdisciplinary research team proposes to initiate salt marsh research at multiple spatial and temporal scales to inform the development of ecosystem-based tools for management of coastal salt marsh ecosystems. This research will lay the groundwork for achieving our long-term goal of developing a “Coastal Resilience Plan” for Connecticut. With funding from YCEI, we plan to initiate the first phase of our research program, which includes a novel combination of site-level research, remote sensing analyses, and community outreach activities. Specifically, the site-based research will include the first experimental examination of the impacts of warming on Long Island Sound coastal salt marsh productivity and accretion, coupled with paleo-ecological studies assessing historic accretion rates and sea level rise. Remote sensing data analyses will be used to map and categorize existing coastal salt marsh in CT in the context of surrounding land cover. Using long-term Landsat and in situ data, marshes will be categorized based on adjacent land uses, historic salt marsh locations, topography, etc. and organized based on future adaptability. This categorization of extant salt marshes coupled with the site-based research will be the first step in establishing a spatial and science-based prioritization of existing conditions along the CT coastline. Our long-term goals are to develop these categories over time to serve as science-based land types for future research and planning purposes, and we will engage the scientific community and decision makers though all stages of our research to ensure these products meet research and planning goals. Collectively, our site-based research and large-scale assessment of coastal marsh conditions will build the necessary foundation for increasing our synthetic understanding of responses of coastal salt marshes to forecast warming and sea level rise and for guiding an integrated salt marsh ecosystem management plan for Connecticut.
Our research will integrate science-based research and solutions through the advancement of the underlying science of climate change, while providing state and town officials with tools for planning. Thus, we believe that the proposed project fits in well with YCEI’s vision to utilize its surrounds as a natural laboratory and test bed and to help build local and regional public relations. Funding from YCEI will allow our innovative team of social and natural research scientists at Yale that approach the problem of climate change from very different but complementary perspectives to forge new interdisciplinary research frontiers, and as a consequence Yale stands to gain a competitive advantage in climate change research aimed at integrating complex scientific understanding of the problem into local and regional planning and decision-making processes.
Past research in salt marshes has focused primarily on how herbivory, hydrology and sediment geochemistry control above and belowground production (e.g., Silliman et al. 2005), but little research has focused on the response of above and belowground production to global warming, despite evidence for loss of marsh species diversity with warming (Gedan et al. 2009), and for productivity responses (both positive and negative) to warming in other herbaceous-dominated ecosystems (e.g., Luo et al. 2001). Our preliminary research in a CT salt marsh suggests that a moderate increase in air temperatures over the growing season (average +0.4°C imposed with open-sided shelters) can increase salt marsh aboveground biomass by as much as sixty percent compared to un-warmed control plots (unpubl. data). This increase in production is comparable to experiments examining the impacts of eutrophication on salt marsh growth (Dai and Wiegert 1997), and could have important consequences for sediment accumulation and salt marsh accretion. Our plan is to expand upon our preliminary research by establishing a well-replicated warming experiment at several representative LIS salt marsh sites, in which we will manipulate temperatures (~1-2°C) using passive warming shelters (without excluding rainfall) and measure above and belowground production, sediment accumulation, and vertical accretion.
Along the lines of our pilot study, we expect to observe increased above and belowground production in response to a moderate warming as a result of a prolonged growing season. Accretion of the salt marsh surface results from increased mineral trapping by marsh plants and/or increased belowground organic matter (OC) sequestration. Warming may have opposite effects on these two processes: increased temperature could increase sediment trapping due to greater standing biomass or surface litter, but may also enhance belowground OC decomposition which could cause sediment compression. Thus, accretion of marsh elevation with warming could be minimal if the effects of these two processes offset each other. Our research will provide the first experimental test of the effects of warming on these two processes in the field and will shed insight into the potential impacts of warming on salt marsh accretion.
As the rates of sea-level rise, the equilibrium marsh elevation declines toward a lower limit, where the duration of flooding is prolonged and the vegetation is negatively impacted. This leads to marsh die-off and an increase in open water throughout the marsh. However, if the surrounding land cover has suitable elevation and pervious surfaces, the marsh may grow shoreward as lower elevation salt marsh is lost to open water and middle and upper marsh is replaced by low marsh vegetation (e.g., Donnelley and Bertness 2001). Therefore, the total coastal salt marsh area may be maintained in the face
of sea level rise. Currently, there is a lack of accurate assessments of CT coastal salt marsh elevation, the vulnerability of salt marsh ecosystems, and the potential for salt marsh migration. We plan to use high resolution salt marsh elevation maps and coastal land cover changes derived from LIDAR and remote sensing data to obtain the information on marsh surface elevation and the land cover change and to assess CT salt marsh vulnerability to sea level rise.
The potential of salt marshes deposits in determining rates of Holocene sea-level change stems from their characteristic vertical zonations of flora and fauna, which allows for the identification of sub-environments within the intertidal zone (Scott and Medioli 1978; Scott et al. 2001; Horton and Edwards 2006). Foraminifera are eukaryotic, unicellular, heterotrophic organisms that are ubiquitous throughout marine habitats). Agglutinant foraminifera construct their tests from quartz, mica, diatoms and other available grains that reflect the environmental setting in which they developed, and are dominant in marsh environments (Scott et al. 2001; de Rijk 1995). The distribution of foraminifera in intertidal settings is primarily controlled by physico-chemical conditions related to exposure time, i.e., the fraction of time that the salt marsh surface is not covered by sea water (Scott and Medioli 1980; Scott et al. 2001). The exposure time is controlled by the elevation of the local salt marsh surface above mean sea level and the local tidal range, and therefore marsh foraminiferal species are distributed in vertical zones, similar to the marsh vegetation (Scott and Medioli 1978; 1980; Scott et al. 2001; Horton and Edwards 2006). These zones can be expressed in vertical distance away from mean high water (MHW). Because each foraminiferal zone is indicative of a specific level of the marsh surface above mean sea level, foraminiferal assemblages in core samples can be used to indicate the position of each sample with regard to mean sea level at the time that the foraminifera were living. If we then know the age of each sample, we can reconstruct the position of each sample (representing the marsh surface at the time of deposition) with regards to modern mean sea level, and thus we can reconstruct the rate of sea level rise over the time of deposition of the studied samples (Scott and Medioli 1980; Thomas and Varekamp 1991; Nydick et al. 1995; Scott et al. 2001; Horton and Edwards 2006).
Because investigations of the rate of recent sea level rise due to global warming require a vertical resolution of less than 1m, marsh foraminifera have become the dominant method for evaluating sea-level variation for the late Holocene in the past five to ten years (Armstrong and Brasier 2005). The accuracy of the reconstruction depends on the width of the foraminiferal zones (thus on the tidal range), as well as on the exact location of the living foraminifera, i.e., whether they really were living on the sediment surface or in the sediments (Saffert and Thomas 1998). In addition, other factors such as salinity may also influence the zonation (Scott and Medioli 1980; de Rijk 1995). As described above, salt marsh plants also exhibit vertical zonations determined by elevation above sea level as described above, but they are a less accurate indicator than foraminifera because their roots extend through a rather long vertical distance (Scott and Medioli 1980). By contrast, foraminifera are more accurate indicators of mean sea level, are comparatively easy to identify and well preserved in marsh deposits. It has not yet been established whether foraminiferal zonations are influenced by marsh eutrophication, as has been suggested for the metaphyta (e.g., Bertness et al. 2009).
We plan to establish a warming experiment consisting of replicate (n = 5 per zone per site) passive warming shelters (and unsheltered control plots) located in the low, middle and high salt marsh zones at several LI sound sites along the CT coastline to determine how warming may alter salt marsh biomass production and marsh accretion processes in each zone. Our goal will be to select sites with different levels of sediment inputs (coastal vs. river) or with different surrounding land uses (site selection will depend on accessibility and permitting). One candidate site is the Branford marsh owned by Yale University, where our pilot study was located. Over the two-year duration of the experiment, we will measure the effects of our warming treatments on salt marsh productivity, sediment accumulation, and belowground respiration in each 5 marsh zone at each site. Included in our measurements will be continuous measurements of microclimatic conditions (air and soil temperature, humidity), as well as plant-level photosynthesis, chamber sediment CO2 efflux measurements and the changes in the marsh elevation.
A. Warming Shelter: The warming shelters are simple in design and execution. They consist of clear corrugated shielding (with holes to allow for rainfall inputs) attached to a 3 x 3 x 0.5 m wooden frame (Fig. 2). Hourly temperatures will be measured during growing season at the marsh surface and at depth of 10 cm inside the shelter and at the control plot using a temperature logger (HOBO H8, Onset Corporation, Woods Hole, MA). At the end of the growing season, the above- and belowground biomass inside the warming shelters and control plots will be directly harvested.
B. Plant Photosynthesis and Soil Respiration: Plant photosynthesis within the warmed and control plots will be measured bi-weekly during the growing season using the LICOR 6400 portable photosynthesis system. CO2 fluxes from the soil surface will be measured with the LICOR 6400 outfitted with a soil respiration chamber using the dynamic, closed-chamber method (Rolston 1986). Beveled-edge 25-cm diameter PVC rings will be installed at approximately 2-3 cm into the salt marsh soil to form a seal at the soil surface with minimal impact on the roots of marsh plants. Rings will be installed at three randomly selected points within each warming shelter and control plot. Ambient CO2 and soil respiration rate will be measured bi-weekly during growing season. The warming effect on plant photosynthesis and soil respiration will be evaluated based on the data comparison between the shelter and control plots.
C. Sediment Accretion and Changes in the Elevation of the Marsh Surface: Monthly changes in marsh elevation will be obtained by Sedimentation-Erosion Tables (SET; Perillo et al. 2003). In order to accurately monitor the change in marsh elevation, SETs will be set up at each study site to obtain elevation data within the warming shelter and the control plot. Treatment effects will be assessed by comparing sediment accretion rates between control and experimental plots. We will monitor foraminiferal populations at these plots of which elevation is known precisely to evaluate a possible response to floral changes.
The ability of salt marshes to migrate towards shore is constrained by the onshore land cover and land use. We will use Landsat satellite data and spectral mixture analysis to assess short and long-term changes in salt marsh ecosystem conditions, surrounding land cover and land uses, and landscape dynamics. This information will help identify conditions that will either control or accelerate coastal salt marsh migration shoreward.
The Landsat Earth observing program has been in existence since 1972. We plan to use a spectral mixture modeling algorithm that has been tested extensively with Landsat data to estimate the fractional surface cover of various land cover types. The method is ideal for use in coastal vegetation systems where sub-pixel cover variation is high.
Historical changes in salt marsh cover will be assessed using Landsat imagery from 1972 to the present. The Seto group has pioneered methods ranging from time series econometrics (Kaufmann and Seto 2001) to geostatistics (Boucher et al. 2006) and neural networks (Seto and Liu 2003) to develop historical reconstructions of land-cover dynamics. We will use a spectral mixture modeling algorithm (Asner and Lobell 2000) tested extensively with Landsat data (Asner et al. 2003) to estimate the fractional surface cover of the salt marsh vegetation canopies as well as bare soil. Spectral mixture analysis is a technique 6 for deriving sub-pixel cover fractions of surface materials using optical reflectance measurements collected from airborne or spaceborne sensors. The method is ideal for use in ecosystems where sub-pixel cover variation is high. Each endmember component contributes to the pixel-level spectral reflectance where ρ e(λ) is the reflectance of each land-cover endmember e at wavelength λ, C is the fraction of the pixel composed of e, and ε is the error of the fit. The second equation shows that the endmembers sum to unity.
Because there are a number of endmember combinations that can produce a particular spectral signal (ρ(λ)pixel), a wide range of numerically acceptable unmixing results for any image pixel are possible (Asner et al. 2000). SMA techniques that use endmember (ρ e(λ)) reflectance “bundles” account for this natural variability. We therefore will employ a general probabilistic method based on Monte Carlo analysis, which accounts for endmember reflectance variability. The model, known as Automatic Monte Carlo Unmixing (AutoMCU), uses three spectral endmember “bundles” to decompose each image pixel using equation (1). For this study, endmember bundles for photosynthetic vegetation (PV), nonphotosynthetic vegetation (NPV) and bare soil (ρ pv(λ), ρ npv(λ), ρ soil(λ)) will be constructed from field spectra collected in and around salt marshes in Connecticut.
We will use the AutoMCU method with historical and contemporary Landsat imagery to estimate the fractional cover of salt marsh vegetation canopies (PV), non-photosynthetic canopies, and bare soil at ~30m spatial resolution. We will assess net changes in salt marsh vegetation canopies as well as changes in the spatial distribution of the vegetation canopies. These analyses will indicate present-day areas of salt marsh and extent of invasives. The proposed methods differ from existing work by Civco and Gilmore in that we will identify fractional cover of vegetation in salt marshes over time. The methods can also be used to map marsh dieback. The information provided by the remote sensing analysis will be foundational to establish a coastal resilience plan. Dynamic maps that provide information about salt marsh expansion and invasives will help guide decision-making on an integrated coastal ecosystem management plan.
We will establish the modern zonation in the eukaryotic unicellular salt marsh foraminifera by sampling the SET sites for foraminiferal populations, so that we can reconstruct rates of relative sea level rise in the past, reconstruct how the marsh adapted to past changes in such rates, and predict whether the marsh will be able to adapt in the near future. The detailed information on marsh elevation from the SETs will make it possible to construct a more precise zonation of foraminifera than available today. We will use that information to reconstruct marsh environmental changes in the past by analyzing foraminiferal content of dated peat cores through a foraminiferal transfer function derived from the data on the modern marsh surface, and correlate the marsh paleoenvironmental curves with the climate record over the last few hundred years.
For this research, we plan to build upon the earlier work of Thomas in cooperation with Varekamp (Wesleyan University) (see Varekamp and Thomas, 1998 for methodology). We expect to be able to document foraminiferal marsh zonations in more detail and with greater precision, because during the earlier studies GPS was not widely available, and for most of the studied marshes tide gauges were not present in the marsh under study, so that data had to be extrapolated from the nearest tide gauge. We plan to study the marsh close to the Yale University field station in Branford, where a tide gauge is already installed, and we will be able to use the field station during the initial field observations. 7
We plan for a summer research period of 8 weeks in the summer of 2010. We will use a Dutch corer to obtain peat cores; about 30-40 cores will be described in the field (including identification of root remains), by 3-4 undergraduate students (Wesleyan and Yale) led by Varekamp and Thomas. The cores will be located on transects from high to low marsh adjacent to the warming experiment. The foraminiferal assemblages in the tops of all cores will be analyzed in order to establish the modern zonation in the marsh, with foraminiferal assemblage composition correlated to location within the tidal framework. The dominant vegetation present will be recorded for all sites. We will describe the lithology of the cores in the field in order to gain insight in the local marsh stratigraphy, we will select 3 sites (in low, middle and high marsh) to take cores for detailed analyses. We expect to analyze about 150 cm-long cores at each location, in order to obtain a record of the last 400-500 years, based on our studies of other Connecticut salt marshes. The upper section of 1 core will be dated by 210Pb analyses, then the other two cores will be correlated through analyses of the pollutant metal mercury (Varekamp et al., 2003, 2005). The cores will be sliced in 2 cm thick samples in the upper part of the cores, 5 cm lower in the cores. Foraminifera will be analyzed in all samples, and paleo-environments of the marsh determined using the faunal assemblages. We will then develop marsh paleo-environmental curves, and derive rates of relative sea level rise over the last 4-5 centennia.
With the goal of integrating this science-based salt marsh assessment into the local decision-making process, we propose organizing a series of public meetings including a working meeting with local science practitioners and strategic meetings with local government officials. Through these outreach meetings we will publicize the role of science-based coastal planning. Through a charette-style engagement, we will help stakeholders visualize and understand how they can make informed decisions about salt marsh conservation, land preservation, and coastal development to enable local and state decision makers to use the information in their planning, zoning, acquisition and permitting decisions. Ultimately, the multidisciplinary effort will generate data, maps, analysis, and the local buy-in to move forward with the development of a coastal resilience plan (with the future support from outside funding agencies). The future coastal resilience plan will prioritize key areas for conservation and wetland migration opportunities, address sea-level rise and storm hazards through science-based land-use planning, promote local regulations and tax incentives for landowners to pursue conservation options, identify vulnerable areas and generate proposals to avoiding disaster reconstruction in these areas.