Temporal variability of carbon and nutrient burial, sediment accretion, and mass accumulation over the past century in a carbonate platform mangrove forest of the Florida Everglades
4.1. Temporal and Spatial Variability
The spatial variability between cores decreases substantially from short to long timescales as reflected in the decreasing mean rates and standard deviations with increasing age (Table 3; Figure 4). All of the rates measured here are highest during the most recent 10 year period, and decrease over the 50 and 100 year periods (Figure 5), thus supporting our first hypothesis. This has important implications for sampling strategies. This evidence of high spatial variability over short time spans supports the practice of using multiple observation stations for these types of measurements, similar to those using surface marker horizons [e.g., Cahoon and Lynch, 1997; Lovelock et al., 2013; Saintilan et al., 2013]. If longer-term measurements are made then the spatial variability is reduced and the likelihood of accounting for this is much higher with just 1-2 sampling locations. Additionally, this evidence indicates the importance of assigning a timescale of observation to any flux rates when making comparisons between sites. For example, even if each of the rates are reported in units of g m-2 yr-1, it is not appropriate to compare carbon sequestration rates between sites when the measurements have been made using different dating techniques such as surface marker horizons (≤10 years), 137Cs (~50 years), and 210Pb (~100 years or greater [e.g., Chmura et al., 2003; Mcleod et al., 2011; Breithaupt et al., 2012]. Rather these rates should be reported as g m-2 yr-1 per specified timescale when making these comparisons.
Because soil accretion and carbon sequestration in coastal wetlands are primarily of interest in the context of global sea level rise and climate change mitigation [Grimsditch et al., 2012], we find that the 50- to 100-year timeframes provide the most conservative forecast of the regional long-term rates. Short-term measurements would likely overestimate the capacity for OC burial and accretion as surface sediments are most susceptible to processes such as erosion, remineralization, or pulse deposition events to name a few. However, there may be circumstances when the short-term rates are indicative of future trends. If a recent, long-lasting change has occurred in a sampling location (such as natural or anthropogenic alterations to wetland structure or hydrology), then it is conceivable that a short-term record that captures acceleration or deceleration in sedimentation rates as a response to a disturbance might provide the best indication of future trends. Examples include road construction [Harmon et al., 2014], shrimp farm construction [Suá rez-Abelenda et al., 2014], abrupt tectonic change [Patel and Agoramoorthy, 2012] or direct impact of a large storm [Cahoon et al., 2003]. But, unless such short-term changes can be categorically differentiated from the longterm steady-state cycles and processes of delivery and degradation, then observations from a longer period of record are more likely to provide a reliable prediction of future sediment accumulation rates.
|Figure 5. Ten year, 50 year, and 100 year integrated mean rates of inorganic matter accumulation, organic matter accumulation, organic carbon burial, and accretion for each of the six cores. [larger image]|
One mechanism that has been well identified is the influence of Hurricane Wilma. That event is one line of evidence for increased delivery as a driver of high rates in the near-surface intervals. While the magnitude of the increase relative to previous intervals is most pronounced in core SH3-1 (Figure 3), each of the cores has its generally highest rates in intervals that have occurred following the year 2000. The Wilma signature of inorganic (largely carbonate) sediment is not uniform throughout these cores spatially nor is it uniform temporally. The inorganic accumulation rates in two cores have decreased following 2005 (Figure 3: SH3-1, SH3-5) while the others have remained at the same level or even increased since 2005. The OM accumulation rate in each of the cores is highest in these uppermost intervals; however, in cores 3, 7, 8, and 9 the increase is only slight. It is conceivable that these uppermost layers simply have higher OM mass in them because of the temporal proximity to a fresh, labile litter supply that over time will degrade and eventually become more like the rates present in the middle to lower depths of these cores. However, in cores SH3-1 and especially SH3-5, the near-surface rates of organic matter accumulation are substantially elevated above the underlying intervals. There are several possibilities for the elevated OM in these intervals including poststorm production [Whelan et al., 2009] either caused by storm sediment provision of P [Castañeda-Moya et al., 2010], normal poststorm recovery production, storm surge deposition of previously buried OM [Smoak et al., 2013] or some combination of all of these.
The nutrient and stable isotope characterization of the soil similarly reflects a difference between the short- and long-term records. The average δ13C values for all dated soil intervals are 26.4 ± 0.9‰ (Table 1). The highest variability is found in the top three intervals (Figure 2 and Table 3), indicating the presence of both autochthonous and allochthonous OM and the influence of hurricane Wilma's storm surge deposition. The surface intervals for all cores except SH3-5 and 9 show a strong positive excursion in δ13C values of 25 to 22 (Figure 2). Enriched values like these suggest a nonmangrove source as evidenced by assessment of global mangrove litter averaging 28 to 30‰ [Kristensen et al., 2008] and Shark River mangrove leaves and wood ranging from 27 to 31‰ [Fry and Smith, 2002; Mancera-Pineda et al., 2009].
Additionally, core SH3-1 has the highest means and standard deviations of TN, δ15N, C:N, and N:P, indicative of its close proximity to the river's edge and the supply of allochthonous marine and algal sources of organic [Smoak et al., 2013] and inorganic material, including Ca-bound P [Castañeda-Moya et al., 2010]. The importance of Wilma on the soil characteristics of this site are reflected in the high flux rates of TN and TP in the last 10-50 years (Table 3). While the TP burial rates are somewhat lower than the global median, TN burial rates over the last five decades (especially for core SH3-1) exceed global median rates of 8.5 g m-2 yr-1 for mangrove wetlands (Table 4). The input of nutrients to the soil contributes to high forest productivity [Castañeda-Moya et al., 2010] but potentially also leads to increased soil microbial respiration [Deegan et al., 2012; Kirwan and Megonigal, 2013; Krauss et al., 2014] as well as a more diverse source of OM in the surface soil including labile microalgae and detritus with low carbon burial efficiencies [Sanders et al., 2014].
|Sampling Site||TN (g m-2 yr-1)||TP(g m-2 yr-1)||Sourceb|
|Shark River, FL||8.8||0.23||1|
|Shark River, FL||6.6||0.2||1|
|Shark River, FL||8.6||0.3||1|
|Shark River, FL||5.7||0.12||1|
|Shark River, FL||7.0||1|
|Shark River, FL||5.4||0.1||1|
|Sawi Bay, Thailand||8.3||2|
|Sawi Bay, Thailand||6.8||2|
|Sawi Bay, Thailand||11.0||2|
|Sawi Bay, Thailand||7.0||2|
|Jiulongjiang Estuary, China||49.0||31.4||4|
|Jiulongjiang Estuary, China||10.1||4.3||4|
|Jiulongjiang Estuary, China||12.8||5.5||4|
|Jiulongjiang Estuary, China||15.0||11.2||4|
|Jiulongjiang Estuary, China||16.3||12.2||4|
|Jiulongjiang Estuary, China||75.0||48.1||4|
|Irian Jaya, Indonesia||28.6||5|
|Irian Jaya, Indonesia||14.3||5|
|Irian Jaya, Indonesia||17.1||5|
|Irian Jaya, Indonesia||8.5||5|
|Yucatan Peninsula, Mexicoc||3.9||6|
|Yucatan Peninsula, Mexicoc||4.0||6|
|Yucatan Peninsula, Mexicoc||4.0||6|
|Yucatan Peninsula, Mexicoc||3.7||6|
|Yucatan Peninsula, Mexicoc||4.9||6|
|Terminos Lagoon, Mexico||5.8||0.8||7|
|Terminos Lagoon, Mexico||1.6||0.7||7|
|Terminos Lagoon, Mexico||3.9||0.7||7|
|Terminos Lagoon, Mexico||4.8||0.5||7|
|Terminos Lagoon, Mexico||2.7||0.1||7|
|Rookery Bay, Florida||5.3||0.2||7|
|Rookery Bay, Florida||4.2||0.2||7|
|Rookery Bay, Florida||4.7||0.2||7|
|Rookery Bay, Florida||6.0||0.2||7|
|Ilha Grande, Brazil||3.1||8|
|aThe mean and median TN rates are 12.5 and 8.9 g m-2 yr-1, respectively.
The mean and median TP rates are 6.5 and 0.7 g m-2 yr-1, respectively.
If values from anthropogenically disturbed locations (Jiulongjiang,
China and Cubatáo, Brazil) are excluded, then the respective mean rates
for TN and TP are 8.9 and 0.5 g m-2 yr-1.
b1 = This study; 2 = Alongi et al. ; 3 = Alongi et al. ; 4 = Alongi et al. ; 5 = Brunskill et al. ; 6 = Gonneea et al. ; 7 = Lynch ; 8 = Sanders et al. [2010a]; 9 = Sanders et al. [2010b]; 10 = Sanders et al. ; and 11 = Sanders et al. .
cValues from figures were estimated using Get Data Graph Digitizer (http://getdata-graph-digitizer.com/).
4.2. Carbon Burial
The 100 year mean OC burial rate for this site is 123 ± 19 g m-2 yr-1 (Table 3), which equates to 10.5% of Net Ecosystem Production [Barr et al., 2010]. This mean rate is relatively low compared to the global geometric mean of 163 (+40; 31) g OC m-2 yr-1 [Breithaupt et al., 2012] and is not greatly different from rates of OC burial at other mangrove sites in Florida, thus rejecting our second hypothesis that carbon burial is a function of high primary production and peat soil with a high percentage of OC. At a site approximately 7 km north of this location in the Everglades, Smoak et al.  measured a centennial rate of 168 g OC m-2 yr-1 in a single core, a rate slightly higher than any single core at this site over the same time period. Smoak et al.  proposed that a portion of the OC from that site could be attributed to the substantially higher deposition of organic material during Hurricane Wilma at the Harney River site compared to the Shark River site. In Rookery Bay, FL, 96 km northwest of this location, Lynch  used 210Pb-based geochronologies from four cores to measure a mean centennial OC burial rate of 86 ± 13 g m-2 yr-1. Our 50 year mean rate of 176 ± 31 is similar to the rate of 147 g OC m-2 yr-1 that was measured in the Florida Keys using the 137Cs dating method based on a peak fallout signature in 1963 [Callaway et al., 1997].
It is somewhat surprising that this location within the Everglades has a low OC burial rate. Even though the aboveground characteristics indicate a highly productive forest [Barr et al., 2010; Castañeda-Moya et al., 2013; Simard et al., 2006] and the peat soil has a high percentage of OC (Table 1) [Castañeda-Moya et al., 2011], these traits alone are insufficient to make a qualitative prediction of a site's OC burial rate (i.e., higher or lower than the global mean). Recently assessed OC burial rates in mangroves in Port Aransas, Texas provide a strong contrast based on these site characteristics. In Port Aransas rates of 270 ± 12 g OC m-2 yr-1 were measured using 137Cs and 253 ± 11 g OC m-2 yr-1 based on the 210Pb-derived 1963 date [Bianchi et al., 2013]. Our 50 year mean OC burial rate is 176 ± 31 (Table 3), substantially lower than the Texas rates. The mangroves in Texas are between 1 and 2m tall whereas those at our site in the lower Everglades are between 13-19m tall. Additionally, in Port Aransas the soil OC% for four cores range from 0.1 to 11.37% but the majority of the values are less than 6% [Bianchi et al., 2013]. At our Everglades site the mean OC% is 24 ± 6 (Table 1). The OC burial rates from these two sites disprove a notion raised in the literature that soil OC% and the rate of OC burial are positively correlated [Kristensen et al., 2008; Breithaupt et al., 2012]. More research is needed to specifically examine the relationship of soil OC density (as opposed to OC%) and productivity to OC burial rates, but the limited data mentioned here demonstrate that other factors exert a significant control. In addition to the potential increase in soil respiration from elevated nutrient concentrations mentioned earlier, these other factors likely include belowground productivity and its ratio to aboveground productivity [Castañeda-Moya et al., 2013; Lovelock, 2008], root density/turnover times [Adame et al., 2014; Castañeda-Moya et al., 2011], the rate of particulate and dissolved organic and inorganic carbon export [Bergamaschi et al., 2012; Maher et al., 2013; Sanders et al., 2010b], the extent of burrowing and bioturbation [Smith et al., 1991; Andreetta et al., 2013], and the supply of inorganic sediments which may aid in sealing organic matter off from remineralization.
The accumulation rates of OM, IM, and OC each influence accretion rates somewhat differently (Figure 6). There is a substantial amount of spatial and temporal variability in IM accumulation, and thus, the data have been subjected to three separate age-related regressions with accretion. Note that these regressions are for the measured rates of each individual soil interval in the specified age ranges; they are not the depth-integrated averages (i.e., Figure 5). While there is nearly a 1:1 relationship between OM accumulation and accretion, the slopes and R2 values for IM over the 0-10 and 10-50 year intervals are much shallower. Inorganic matter has a much lower and less predictable influence on site accretion over these timescales at this site in the Everglades. The density of OM is lower than that of IM, and as a result OM can contribute to a higher accretion rate. The y intercept for OM is approximately 0; however, for IM the y intercept for all three age classes is 0.6 to 6.4 mm yr-1. This range indicates the average magnitude of the accretion contribution made by OM relative to IM. While the evidence here indicates that IM accumulation has only a minor direct influence on sediment accretion, this does not preclude a substantial indirect influence that may occur because of P-fertilization in the storm surge deposition [Castañeda-Moya et al., 2010] and a subsequent increase in organic matter production. The IM accumulation is relatively well correlated with accretion for the intervals between the ages of 50-100 years, although the slope is about 25% shallower than that for OM. The difference in slopes may be attributed to greater compaction and consolidation for these depths, or it might indicate an increase in the supply of marine carbonate sediments from storm surge in recent years.
Mangrove soil OC is most often discussed in the context of greenhouse gas sequestration, but here it also serves as a soil-building component of vital importance to the ecosystem. If OC burial rates are regressed with accretion rates the R2 is 0.80 and the slope is 2.05 (Figure 6), reflecting the importance that OC provides to the soil structure here. Without the 25% of soil mass that is OC (Table 1), the forest floor of this site would be considerably lower than its present position and would consequently endure greater physical stress in addition to altered redox conditions from substantially increased periods and extents of inundation [Gilman et al., 2007; Lu et al., 2013]. As previous research has noted, the autochthonous production of organic material is of considerable importance to mangroves growing in locations with little to no terrigenous sediment supply [McKee, 2011; McKee et al., 2007; Parkinson et al., 1994; Woodroffe, 1990]. As it is, this OM and OC does not contribute uniformly to surface accretion but is able to shrink and swell in response to various hydrologic conditions at different depths of the soil column [Whelan et al., 2005].
Although these findings indicate that this location in the Everglades is keeping pace with sea level rise, research of this nature is needed on a much wider spatial scale to assess the net soil accumulation rates across the greater coastal Everglades. As has been shown in estuaries in China and Malaysia, single site assessments of accretion and accumulation may not accurately account for cumulative losses occurring elsewhere in the system [Alongi, 2011]. As was similarly noted by Smoak et al. , one of the significant implications may be that the addition of material at this site occurs at the expense of other locations on the seaward edges of the Everglades.
|Figure 6. Relationship between (a) organic matter accumulation, (b) organic carbon burial, and (c) inorganic matter accumulation and accretion rates for all dated soil intervals (≤ 100 years) (n = 161 for OM and OC; n = 158 for IM because three outlier samples from Hurricane Wilma in core SH3-1 have been removed for this analysis). In Figure 6c, trend line equations and R2 values are for the three interval age classes 0-10 years (equation A), 10-50 years (equation B), and 50-100 years (equation C). If the three age classes are analyzed together the linear regression trend line equation is: y = 0.1584x + 0.2787 and the R2 is 0.33. [larger image]|