Home Archived October 29, 2018
(i)

South Florida Information Access (SOFIA)


Analysis and Simulation of Propagule Dispersal and Salinity Intrusion from Storm Surge on the Movement of a Marsh-Mangrove Ecotone in South Florida

Methods

> Methods

An ecotone model is first developed to describe possible mechanisms that can maintain a sharp ecotone and, by extension, the possibility of regime shifts causing a discrete jump in the location of the ecotone. The model is applied to a single USGS study site on a transect across the mangrove-freshwater marsh ecotone. While much relevant data are available at this site, only rough estimates of some model parameters are available at this time, so parameters that need further research are identified.

Study Site

The mangrove-marsh ecotone studied is adjacent to the Harney River estuary (25°25' N, 81°03' W), a distributary of the Shark River estuary from the freshwater Everglades into the Gulf of Mexico (Fig. 1). Two U.S. Geological Survey (USGS) paired surface/groundwater gauges along the Harney River bookend a north-south transect. Shark 4 (SH4) is located ~30 m south of the Harney River, within a mangrove fringe forest and USGS gauge Shark 5 (SH5) is located ~300 m south in an interior coastal marsh. The SH4-SH5 transect is located on the middle reach of the estuary, about 9.5 km from the Gulf of Mexico. Historical aerial images show there has been some mangrove migration into the freshwater marsh habitat in this area of the southwestern Everglades, occurring between the late 1920s and early 1940s (Smith III et al. 2013).

illustration showing study area of mangrove-marsh transition
Fig. 1 Study area of mangrove-marsh transition. SH4 is USGS gauge Shark 4, located in mangrove fringe forest; SH5 is USGS gauge Shark 5, located in sawgrass; HR is Everglades National Park gauge on Harney River. The inset graph shows study area and Hurricane Wilma pathway [larger image]

Model Spatial Scale

The model is based on a grid of 10x10 m spatial cells. Within each cell, water and salinity dynamics, as well as vegetation dynamics are described.

Model Water Budget and Salinity

The water budget was computed as a water balance on a daily basis. Vadose or rooting zone salinity, assumed homogeneous within a given spatial cell, is determined by infiltration rate, which is the difference between the precipitation, P, tidal input, T, and evapotranspiration, E. A groundwater lens, whose salinity can vary, sits on top of underlying seawater. Salinity of this groundwater lens in the same given spatial cell is determined by the difference between the infiltration rate and the daily change in the surface height of this groundwater. Infiltration rate (INF) is calculated as follows,

INF = E - P - T.

The dynamics of salinity in the rooting zone are the same as used by Sternberg et al. (2007),

equation 1 - dynamics of salinity in the rooting zone for capillary rising from the groundwater to the rooting zone

equation 2 - dynamics of salinity in the rooting zone for percolation downward into the groundwater

where ρ is the porosity, z is rooting zone height, and SV and SGW are the salinities of the pore water in the rooting zone and of the underlying groundwater, respectively, and INF has units of distance per unit time. Positive values of infiltration indicate capillary rising from the groundwater to the rooting zone. Conversely, negative values of infiltration indicate percolation downward into the groundwater. The underlying saline ocean water, on which the groundwater sits, is assumed to be 35 (SO). The dynamics of salinity in the groundwater are given by the equations,

equation 3 - dynamics of salinity in the groundwater

equation 4 - dynamics of salinity in the groundwater

equation 5 - dynamics of salinity in the groundwater

equation 6 - dynamics of salinity in the groundwater

where HGW is the groundwater gauge level. When the groundwater level changes to compensate upward infiltration to the rooting zone (i.e., equation - groundwater level changes to compensate upward infiltration rate   is positive), additional water from underlying saline ocean water will move into the groundwater. Groundwater salinities do not affect the much larger pool of underlying saline ocean water.

To parameterize the model, water level or stage data (NAVD 88) were obtained from USGS paired surface/groundwater gauges SH4 and SH5. Daily water level changes were calculated as the difference from 1 day to the next. Daily rainfall data were obtained from the USGS Everglades Depth Network (EDEN) website (http://sofia. usgs.gov/eden/). Monthly averages and standard deviations were used to generate precipitation input for the model. Tidal height and salinity data were obtained from a nearby Everglades National Park gauge, Harney River (HR). The evapotranspiration, E, used in the model was calculated by multiplying potential evapotranspiration (PET) by salinity effect (fS). Daily PET data, which represent the evapotranspiration rate without moisture or salinity limitation, were obtained from the EDEN website. Evapotranspiration reductions due to salinity effect (fS) differ significantly between halophytic and glycophytic species (Saha et al. 2012). Glycophytic species decrease transpiration when salinity begins to reach levels that the plants cannot tolerate, while halophytic species can continue to transpire at relatively high salinity. To represent the differences in glycophytic and halophytic vegetation, we modified two functions for the salinity effect (fS) based on empirical data (Lin and Sternberg 1992; Sternberg et al. 2007).

equation 7 - modified salinity effect

equation 8 - modified salinity effect

where fS1 is the salinity effect for freshwater marsh, and fS2 is the salinity effect for mangroves. Both functions decrease with increasing SV, but fS1 declines much more sharply.

Vegetation Dynamics

Competition between mangrove and freshwater marsh (sawgrass: C. jamaicense) was simulated by extending an existing individual-based model, the Spatially Explicit Hammocks and Mangroves (SEHM) model (Jiang et al. 2012a). The SEHM model was developed for two competing halophytic and glycophytic tree types, based on a neighborhood competition model between individual plants from Berger and Hildenbrandt (2000) and self-reinforcement between of soil porewater salinity and vegetation. The same concepts are assumed to play a role in mangrove-marsh ecotones. Direct competition (i.e., for light) of vegetation would likely favor sawgrass over mangrove seedlings, though the reverse would occur once mangroves reached large size. But indirect effects on vadose zone salinity favor mangroves. Mangroves are able to utilize both shallow soil water and deeper groundwater. During the dry season, when the upper soil layer can become highly saline, mangroves have been observed to switch to less saline groundwater (Ewe et al. 2007). Sawgrass appears to be restricted to utilizing only the shallow water. Although sawgrass has some salinity tolerance, it will ultimately be outcompeted when mangrove density is high enough to raise dry season vadose zone salinity to high levels (Ewe et al. 2007). These opposing mechanisms of competition are hypothesized here to be sufficient to maintain a sharp ecotone, though other processes, such as fire, which can destroy mangrove seedlings, may contribute to the boundary (Egler 1952). Fire processes are implicitly considered to have a positive effect on marsh ecosystems that maintains open canopies, as with some models in savanna- forest systems (Staver and Levin 2012). We use the Overview, Design Concepts, and Details protocol (Grimm et al. 2006; Grimm et al. 2010) to describe the model in detail (see Supplementary Material). A simple overview of vegetation dynamics is described as follows.

We simulated competition between mangroves and freshwater marshes, including the effects of the abiotic factors of hydroperiod and rooting zone porewater. Freshwater marsh (e.g., Cladium) is relatively spatially homogeneous at a scale of 10 by 10 m and was simulated as biomass per unit area. The biomass of marsh in a given cell was determined by monthly gains from photosynthesis and losses due to respiration or mortality. Photosynthesis was modeled as the maximum possible rate multiplied by limitation factors, including salinity, light, and number of days of flooding. Mangrove biomass has indirect effects of limiting marsh spread both through influencing soil salinity and retarding fire percolation (analogous to a forest-savanna ecosystem; e.g., Staver and Levin 2012). Similar to SEHM (Jiang et al. 2012a), mangrove dynamics were modeled as individual-based. For each monthly time step, every tree has a growth increment that is a function of light, number of days of flooding, salinity of the particular spatial cell, and fire damage induced by biomass of marsh, which promotes the spread of fire. When a mangrove tree reaches maturity, new propagules are produced at monthly intervals by the tree and disperse to new locations according to a probability function. The probability of each propagule to produce a successful new recruit depends on the salinity of the soil porewater at the location the seedling reaches. At the end of the monthly time step, the probability of tree mortality is related to size-dependent factors, such as too small a diameter at breast height of the tree, or from reduced growth rate caused by competition or salinity.

Analysis and Simulation

Model simulations were performed on a 30x30 grid of 10-m square cells applied to the study site. Elevations of all the cells were assumed to be 8 cm above sea level (NAVD 88). Tidal signal inputs were a maximum at the riverward edge of the simulation landscape and decreased exponentially inland, based on empirical relationship from Krauss et al. (2009). On a daily time scale, the hydrodynamics submodel updated salinities of the soil porewater and groundwater. Monthly average values of salinity in each cell, which affect tree growth and propagule establishment, were then used in the vegetation dynamics submodels.

The model was applied to an actual hurricane that produced a storm surge at the study site. This application was designed to demonstrate the feasibility of the model. Because the size of the impact of the storm surge was small at the study site and because salinities at the Harney River study sites are usually less than 5 during the wet season, an effect on the mangrove- marsh ecotone was not expected. Hydrology data from January 2000 to January 2010 were analyzed to investigate the possibility of effects from storm surges during that period. Hurricane Wilma, a category-3 (Saffir-Simpson scale) hurricane, made landfall upon the western facing Florida Everglades mangrove coast during low tide on the morning of 24 October 2005. The eyewall of the hurricane crossed over Everglades City, about 50 km north of the Harney River. Hurricane-induced floodwater levels along the mouth and lower reach of the Harney River exceeded 3 m (Smith III et al. 2009); however, along the middle reach (9.5 km from the Gulf of Mexico) of the Harney River measured water levels increases were dampened to only 11.2 and 4.9 cm (NAVD 88) at USGS gauge SH4 and SH5, respectively. Both gauge stations were fully operational during the storm event. Salinity and water level changes after Hurricane Wilma were used as inputs to the simulation model in order to test the potential of the storm surge to produce a regime shift across some area of marsh vegetation.

To investigate how salinity intrusion and mangrove dispersal from larger hurricane events might trigger ecotone movement, we supplemented the simulation of Hurricane Wilma with simulations of vegetation dynamics under given scenarios of both different levels of salinity disturbance duration and different amounts of mangrove propagules carried by the storm surge into freshwater marsh sites. Due to lack of studies on mangrove propagule dispersal after the storm surge, we assumed pulse dispersal input to freshwater marsh sites with Poisson distribution on the spatial grid. Three levels of propagule density transported by the storm surge, 0, 1,000, and 2,000 propagules/ha, were tested. Long-term vegetation dynamics were also simulated under given scenarios of salinity disturbance duration caused by storm surge salinity intrusion. For each level of mangrove propagule density, three levels of salinity intrusion duration were assumed, 0 (no salinity overwash), 1, and 2 years. It was assumed that a level of salinity in the groundwater (i.e., freshwater lens, SGW) remained at a level of 35 for these durations.

< Introduction | Results >


Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo USA.gov logo U.S. Department of the Interior | U.S. Geological Survey
This page is: http://sofia.usgs.gov/publications/papers/propdisp_salint/methods.html
Comments and suggestions? Contact: Heather Henkel - Webmaster or (727) 502-8028
Last updated: 28 May, 2014 @ 09:58 AM (KP)