publications > paper > groundwater control of mangrove surface elevation: shrink and swell varies with soil depth > materials & methods
Materials and Methods
The Surface Elevation Table (SET), based on the
design of Boumans and Day (1993), allows for
precise measurements of soil surface elevation
(±1.4 mm total error; Cahoon et al. 2002a). The
SET consists of a mechanical arm that is attached to
a benchmark and leveled, establishing a fixed
measuring point. Typically each SET has four fixed
measurement locations (directions), where nine
measuring pins are lowered to the soil surface to
obtain a relative soil elevation. The elevation is the
mean of 36 measuring pin readings per benchmark.
SETs have been successfully used to monitor
changes in elevation in a number of wetland
environments (Cahoon et al. 1999). They have been
used to monitor mangrove vertical accretion and
subsidence (Cahoon and Lynch 1997) and to follow
the response of soil elevation to season (Childers et
al. 1993), water management (Boumans and Day
1994; Hensel et al. 1999), vertebrate herbivores
(Ford and Grace 1998), and hurricane disturbance
(Cahoon et al. 2003).
New SET designs have recently been described that measure the change in soil elevation of specific parts of the soil profile (e.g., root zone, below the root zone; Cahoon et al. 2002b). At the Shark River, the shallow-rod surface elevation table (shallow-RSET) benchmarks were installed to a depth that measures elevation change in the majority of the active root zone (top 0.35 m of the soil profile). The deep-rod surface elevation table (deep-RSET) benchmarks were driven into bedrock and measure the full soil profile. The original design SET (original-SET) benchmarks used by Smith and Cahoon (2003) were driven to approximately 4 m (Fig. 1). Further information on the design and accuracy of the original-SET and RSETs can be found in Cahoon et al. (2002a,b). By using a combination of SET designs at a single study site, it is possible to partition changes in soil elevation among specific parts of the soil profile, such as the shallow root zone and deeper soil zones (Fig. 1). By determining the absolute change for each depth zone we can calculate expansion and contraction for each zone (surface [accretion and erosion, above 0 cm], shallow [active root, 0-0.35 m], middle [0.35-4 m], and bottom [4-6 m]) of the profile.
The study site, SH3 of Smith and Cahoon (2003), is located near the mouth of the Shark River (25°21'50.3" N 81°4'42.2" W, 1984) in a mature mixed mangrove riverine forest comprised of Rhizophora mangle (L.) (red mangrove), Laguncularia racemosa (L.) Gaertn. (white mangrove), and Avicennia germinans (L.) Stearn (black mangrove). The site has a sparse understory. The canopy ranges in height from 13 to 17 m. The site has mixed tides. During the study period Shark River had a daily average conductivity of 40 mS cm-1 and varied between a low of 25 mS cm-1 and a high of 51 mS cm-1. Shark River discharge was greatest at the end of the wet season, from September to November for 2002.
The soil profile of this site was determined from the well drilling log (Anderson unpublished data; Fig. 1). The mangrove peat was 5.5 m in depth. The peat matrix lay directly on top of limestone, into which the well was drilled 1.8 m. The transition between the peat matrix and limestone was rapid. The limestone-peat interface was difficult to drill but had softer material below it. Otherwise, the entire peat layer was of similar constituency. No clay deposits were encountered during the drilling.
Cohen (1968) described the stratiography of the mangrove soil column at the mouth of Little Shark River, a location approximately 2.5 km away from SH3. He found that the mangrove peat was 3.81 m in depth and the total depth to bedrock at the site was 3.86 m. The peat types did not have recognizable petrographic constituents. All of the peat types were marine or brackish and dominated by R. mangle. There was a general increase of fine granular debris at the top and bottom of the profile. Fine granular debris comprised approximately 35% of the sample at the top and bottom of the core. At the top of the core it was suggested that an increase in fine-grained marine carbonates were responsible for this high number. The increase in fine granular debris at the bottom of the core may be due to greater amounts of degradation of the organic constituents of the peat. Pyrite content was relatively high (2-18%) throughout the core suggesting reducing conditions. Fusinite only occurred at the bottom of the core and comprised a small percentage of the constituents. There were no clays reported from this core.
Preliminary sampling of the mangrove peat hydraulic conductivity (at a site 4 km away) yielded relatively low values (hydraulic conductivity field saturation method [Guelph permeameter] = kfs = 1.87 m d-1, see Hughes et al. 1998), which suggest slow water transmittance through the surface layer of the peat (Anderson et al. 2001).
We installed three groups of SETs within 18 m of each other and 45 m of Shark River. All groups were within 15 m of a U.S. Geological Survey (USGS) hydrological monitoring station (USGS station #252149081044301, described below). Each group included one shallow-RSET, one original-SET, and one deep-RSET along with four feldspar marker horizons (Cahoon and Turner 1989). The three original-SETs, used in the Smith and Cahoon study (2003), were installed on July 16, 1998. Three shallow-RSETs and three deep-RSETs were installed on February 28, 2002 (Table 1). On March 18, 2002, four separate layers of feldspar (0.5-3 mm deep) were laid as marker horizons with each group for a total of twelve new marker horizons. Shallow-RSET benchmarks were installed to a depth of 0.35 m. The original-SET benchmarks (76-mm [3"] diameter aluminum pipe, 1-mm thick wall) were driven approximately 4 m deep. The deep-RSET benchmarks (1.43-cm [9/16"] diameter stainless steel rods) were driven to approximately 6 m deep (Table 1). All SETs and feldspar markers were measured monthly from March 18, 2002, to March 21, 2003. Measurements were taken during low tide exposure on the same day. Two sampling events occurred with minimal water (a few puddles) present on the soil surface. On November 9, 2002, and February 10, 2005, a period of 2 yr and 4 mo, we surveyed the elevation of only the group number 3 shallow-RSET, original-SET, and deep-RSET with standard survey methods (±3 mm). There was no movement of the SET devices in relation to an established benchmark, suggesting that the assumption of a stable datum (Childers et al. 1993; Cahoon et al. 1995, 2002b; Cahoon and Lynch 1997) was valid during the study (Table 1).
The hydrological conditions investigated were daily rate of change in groundwater piezometric pressure and river stage. Groundwater head pressure was collected from a USGS station installed at the site in 1996 (Anderson and Smith 2005; Fig. 1). A piezometer recorded groundwater head pressure of the shallow coastal aquifer in a layer of limestone (hereafter referred to as groundwater). The 7.33-m piezometer consisted of threaded 7.62-cm diameter PVC pipe that was screened (0.20 slot PVC) from 5.7 to 7.2 m depth. The slotted part of the well was entirely within the limestone. The well was sealed with formation packer at 5.5 m depth, the interface of the limestone and the peat layer, to prevent vertical flow. Piezometric head pressure measurements were collected at hourly intervals. The pressure transducer was located at the depth of the well screen (for further details see Anderson and Smith 2005).
Shark River stage data were obtained from the Shark River hydrological monitoring station of Everglades National Park located 2.37 km downstream from SH3. This station records tidal influences as well as seasonal changes in river discharge for the area. Tidal flooding occurred at the site when the Shark River stage was above 0.07 m (Fig. 2). Shark River stage data were collected hourly. The groundwater piezometric head pressure and the Shark River stage were reported in North American Vertical Datum (NAVD) 88 datum (Geiod 99) (Fig. 2). Hourly Shark River stage and groundwater head pressure for the interval from December 13, 2002, to January 9, 2003, are included in Fig. 2. We used daily averages of the above parameters in order to remove the diurnal tidal signal. The daily averaged signal of these parameters shows the monthly lunar influences on the tide (Provost 1973), annual change in sea level (Provost 1973), and the seasonal changes in water level due to the regional wet season (Fig. 2). The hourly tidal signal was assumed to have minimal effect on our SET measurements because elevation data were always collected at low tide. Sensor malfunction resulted in the loss of daily groundwater piezometric head pressure data from October 7, 2002, to November 8, 2002, an interval that included the October 10 SET sample measurement.
Soil elevation at each SET benchmark was averaged across all measuring pins in four directions (n = 36) for each sampling event. To determine the average daily rate of change (DRC) in the soil elevation between sampling events we used the following formula:
Where Xt and Xt+1 is the average elevation at time t + 1. The DRC for all hydrological metrics were determined in a similar fashion. River stage averaged for day Xt+1 was subtracted from river stage averaged for day Xt and divided by the number of days in the interval. The daily average hydrological metrics were used in the analysis to remove hourly tidal effects (Fig. 2).
Within the three SET types, we used forward stepwise multiple regression to investigate the relationship between daily rate of change in soil elevation for each of the three benchmarks and the rates of change in the hydrological parameters and accretion. Stepwise multiple linear regression was used in order to discern the most important hydrologic variable associated with incremental elevation change. Stepwise regression not only allows for the identification of the most parsimonious model, but accounts for correlation among two or more variables (Zar 1999). All parameters included in the models were tested for collinearity and normality of the residuals (Quinn and Keough 2002). All models were analyzed using STATISTICA 5.0 (Statsoft Inc. 1996) and SPSS 11.0 (SPSS Inc. 2001). The final models included the two hydrological parameters: DRC in groundwater piezometric pressure and DRC in river stage. Within each SET type, we used a data set reduced from 36 data intervals (12 monthly intervals X 3 benchmarks) to 30 data intervals as a result of the hydrological data gap for groundwater piezometric pressure. Because there was only one well at the site, the hydrologic data was used three times, once for each SET type analysis. This may call into question the independence of the hydrology well data. We felt justified in presenting the hydrologic data with individual SET data to emphasis small scale spatial variation in soil surface elevation, and we had no reason to expect hydrological variation over this small distance mainly due to consistency in the soil matrix.
We felt that a regression using interval rate of change (as opposed to a regression of cumulative change) was justified because the focus of the study was to discover the relationship between elevation change and hydrologic variable from one sampling interval to the next. Interval data should reduce the influence of any serial correlation; due to the length of time between samples (monthly intervals), we felt that there was little influence of prior values on the relationships within a given interval. Regressions between the interval rate of change of soil elevation and the interval rate of change of hydrologic variables have been used previously (Childers et al. 1993).
By using the absolute change for each benchmark depth sampled by the three types of SET, we could calculate expansion and contraction for each component of the soil profile using the following formula:
Thickness of the entire soil profile is equal to the
sum of surface accretion (above 0 m) and changes
in thickness of the active root zone (0-0.35 cm), the
middle zone (0.35-4 m), and the bottom zone (4-6 m).
U.S. Department of the Interior, U.S. Geological Survey
This page is: http://sofia.usgs.gov/publications/papers/mang_surf_elev/matnmethods.html
Comments and suggestions? Contact: Heather Henkel - Webmaster
Last updated: 15 January, 2013 @ 12:43 PM(KP)