|publications > paper > controls on mangrove forest-atmosphere carbon dioxide exchanges in western Everglades National Park > discussion
Controls on mangrove forest-atmosphere carbon dioxide exchanges in western Everglades National Park
 Mangrove forest NEP values (1170 ± 145 g C m-2) estimated for 2004-2005 are substantially greater than those reported for terrestrial ecosystems [e.g., Baldocchi et al., 2001; Luyssaert et al., 2007; Hirata et al., 2008]. In general, the annual NEP of tropical ecosystems tends to be greater than that of temperate ecosystems due, in part, to the year-round productivity [Luyssaert et al., 2007]. The high NEP values reported here are reflected in the high leaf litter and wood production, which is reported at 1170 g C m-2 yr-1 [Twilley et al., 1992]. Relatively low respiration rates (Rd) in the mangrove ecosystem are largely responsible for the high NEP estimates. Nighttime Rd values varied from 1.71 ± 1.44 to 2.84 ± 2.38 µmol (CO2) m-2s-1 at soil temperatures of 15 ± 2°C and 20 ± 2°C, respectively. These Rd values are lower by a factor of 2 compared to terrestrial AmeriFlux and EuroFlux sites whose respiration rates range from 3.72 ± 2.20 to 5.92 ± 4.40 µmol (CO2) m-2s-1 at soil temperatures of 15°C and 20°C, respectively [Falge et al., 2001]. Slow biomass decomposition rates associated with saturated soils and anaerobic conditions result in reduced carbon respiratory losses in mangrove forests [McKee, 1993; Ferreira et al., 2007; Romero et al., 2005; Poret et al., 2007].
 When integrated over annual periods, the low respiratory fluxes determined at the study site resulted in annual ecosystem respiration (RE) rates during 2004 that were similar to midlatitude terrestrial biomes and significantly lower than other tropical or subtropical evergreen forests (Figure 11). The annual GPP (the sum of NEP and RE) reported for these tropical systems frequently exceeds 3000 g C m-2 [Kato and Tang, 2008], with a global average value of 3551 ± 160 g C m-2 [Luyssaert et al., 2007]. These GPP values are significantly higher than our estimates in the Florida Everglades mangrove forest. Other tropical and subtropical systems also typically exhibit higher RE (3061 ± 162 g C m-2 [Luyssaert et al., 2007]) compared to our site. As a result, the annual NEP values reported for most other tropical systems are also lower than the present 2004-2005 estimates in the mangrove forest.
 Tidal activity in mangrove forests, such as those found along Shark River, often results in substantial lateral fluxes of particulate and dissolved carbon. This export of carbon will tend to lower estimates of ecosystem respiration derived from EC measurements. For example, benthic microbial decomposition of particulate and dissolved organic carbon (DOC) exported from mangroves [Souza et al., 2009] and from tropical terrestrial forests [Mayorga et al., 2005] results in respiratory fluxes outside of the EC footprint. Similarly, dissolved inorganic carbon (DIC), often found in high concentrations in estuarine waters [Bouillon et al., 2007a; Miyajima et al., 2009] and derived from below-ground respiration, is removed by tidal flushing and does not contribute to atmospheric CO2 fluxes in the forested intertidal zone.
 Tidal export of dissolved and particulate organic carbon (POC) from the EC footprint was not measured in 2004-2005. However, the potential magnitude of these fluxes and their influence on our estimates of NEP can be constrained. For example, an extreme upper bound on these fluxes can be estimated as the difference between RE observed and the RE values that would be expected in this forest if NEP = 0 (i.e., RE = -GPP). A forest with NEP = 0 would lie along the 1:1 line in Figure 11. Therefore, the distance along the x axis from our observations to this line in Figure 11 represents the difference between RE and GPP and the potential carbon export assuming NEP = 0 in this system. This provides an upper limit on annual tidal export of ~1000 g C m-2 yr-1. However, we consider this an overestimate since this forest is known to accumulate biomass and soil carbon (i.e., NEP ≠ 0). We further constrain the magnitude of tidal carbon export using a combination of direct measurements obtained near our site and a literature review. For example, in a flume study near our site, Romigh et al.  estimated a net DOC export rate of 56 g C m-2 yr-1. In other mangrove forests along the Everglades Gulf Coast, Twilley  and Heald  estimated POC exports from 64 to 186 g C m-2 yr-1. No direct measurements of DIC fluxes in this region are available. However, in their review of data from other systems, Bouillon et al. [2007b, 2008] suggest DIC export can be as much as 3 to 10 times the amount of DOC exported from tidal mangrove forests. Using this relationship between DIC and DOC and the Romigh et al.  estimate of DOC, we estimate the DIC export can be 170 to 560 g C m-2 yr-1 (Table 4). Therefore, a better estimate of the total dissolved and particulate carbon export from our site is 550 ± 260 g C m-2 yr-1.
 Adding this estimate of total DOC, DIC, and POC export to the estimates of RE derived from our NEE measurements yields a GPP/RE ratio for this forest similar to values reported for other tropical forests [Kato and Tang, 2008] (Figure 11). Adding all of the POC and DOC fluxes to RE may, however, slightly overestimate the influence of these fluxes on NEE since in a system without tidal influences, some fraction of the POC and DOC may not be respired into CO2 and would instead accumulate in the system. We are unable to quantify the potential bias this introduced into our estimates of tidal carbon export and RE, but we do not consider this to be a significant term, primarily because the magnitude of POC and DOC fluxes relative to DIC fluxes measured in other mangrove systems is typically small. Estimates of high DIC flux from the mangrove forests at our site are supported by measurements of high partial pressures of carbon dioxide (pCO2) at the mouth of Shark River [Clark et al., 2004]. We conclude that between 25% and 70% of NEP is exported into the estuary with the remainder accumulating in tree biomass and soil carbon.
 There are several important challenges to measuring total carbon export at this site. Commonly applied methods used for determining DOC fluxes on an aerial basis have focused on water-soil surface exchanges across relatively well-defined tidal creeks or man-made flumes. However, high tides often inundate the entire island at our site, and overwash occurs around the island perimeter. The carbon fluxes via this overwash may be significant and will vary over time depending on the amplitude and duration of the tidal cycle. This aspect of the carbon budget at our site requires further examination.
 An independent estimate of NEP derived from EC measurements can be calculated as the difference between net primary productivity (NPP), based on biometric data and soil respiration (Rs [Luyssaert et al., 2009]). In tidal systems, the estimates of NEP derived from biometry and Rs do not account for dissolved and particulate carbon export and can therefore be compared directly to our estimates derived from EC. Bouillon et al.  suggest an average annual NEP of 1100 ± 644 g C m-2 for mangrove ecosystems based on the difference between globally averaged NPP (1363 ± 450 g C m-2) and Rs (263 ± 194 g C m-2). Komiyama et al.  provide a similar NEP estimate of 852 g C m-2 yr-1 for a mangrove forest in eastern Thailand. At our site, Ewe et al.  measured aboveground NPP, including increases in basal area and leaf litter, to be 1100 ± 45 g C m-2 yr-1. We estimate belowground NPP of 520 ± 360 g C m-2 based on a review by Bouillon et al.  of results from four studies in southwest Florida close to our site. The locations of these studies share many characteristics with our site and include a fringing forest of R. mangle; two mixed species basin forests of R. mangle, L. racemosa, and A. germinans; and an aggregate of sites located in mangrove forests along the east and west coasts of Everglades National Park. Two sets of direct Rs measurements were made at our site in six 20 cm2 plots using a soil CO2 flux system (model 8100, LI-COR, Inc., Lincoln, Nebraska). The soil CO2 efflux rates from these observations ranged between 0.5 to 2.0 µmol m-2s-1 (T. Troxler, Florida International University, unpublished data, 2009), and from these data we estimate an annual Rs of 360 ± 180 g C m-2 at our site. Subtracting this Rs value from the combined aboveground and belowground NPP values yields a biometric NEP estimate of 1000 ± 400 g C m-2. This value is within the confidence limits of the EC-derived NEP quantities for this ecosystem.
 Aboveground respiratory fluxes contributed by foliage, boles, and prop roots are expected to outweigh the belowground components of Rd. During high-tide periods when the soil surface is submerged, the average reductions in Rd from low-tide periods (Figure 8) are roughly equivalent to Rs derived from chamber measurements, which suggest the tides suppress belowground respiratory CO2 efflux to the atmosphere. There is substantial variability and overlap in Rd across tidal cycles, suggesting temperature effects on aboveground respiratory fluxes throughout the year have a greater effect than tidal influences on Rd. Dark respiration rates in red mangrove foliage are estimated at 1.62 ± 1.32 µmol (CO2) m-2s-1 at 30°C [Barr et al., 2009]. The leaf area index (LAI) at this site in 2008 was measured at 2.29 ± 0.18 (V. Rivera-Monroy, Louisiana State University, personal communication, 2009). Multiplying the foliage dark respiration rate by this estimate of LAI suggests that foliage respiration alone can contribute to 73% of total Rd during low-tide periods. A recent study by Lovelock , using data from 10 mangrove forests distributed throughout the Caribbean, Australia, and New Zealand, supports the hypothesis that soil respiration is a relatively minor term in Rd in mangrove forests. For example, applying the Lovelock  parabolic relationship between Rs and temperature at the Everglades site yields Rs values of 1.23 and 1.30 µmol (CO2) m-2s-1 at 20°C and 30°C, respectively. These values agree with the direct measurements of soil CO2 efflux at our site, and represent at most 25 to 41% of nighttime Rd.
 Synoptic-scale salinity effects are apparent when relating NEE to PAR and TA (Figure 6). There is a linear decrease in LUE with increasing salinity across all seasons (Figure 7). Other studies [Kozlowski, 1997; Ball and Farquhar, 1984; Sobrado, 1999; Parida and Das, 2005; Lopez-Hoffman et al., 2006] also report negative effects of salinity on mangrove physiological functioning and growth.
 Consistent with findings in terrestrial forests [Gu et al., 2002], increases in diffuse solar irradiance (i.e., decreasing Kt) were associated with increasing canopy LUE. However, the positive effects of diffuse solar irradiance were notable only at lower TA (≤21°C). At higher TA, high Kt was usually associated with PAR values above the saturation value of 1000 µmol (photons) m-2s-1 reported for R. mangle [Barr et al., 2009], Rhizophora mucronata and Ceriops tagal [Theuri et al., 1999], and Avicennia marina [Naidoo et al., 1997]. Therefore, a substantial proportion of the mangrove foliage during the summer months can function at or near light saturation conditions, and this process can reduce any positive effects of decreasing Kt on NEE. Leaf orientation is another factor contributing to the lack of Kt effects on NEE during the summer months. Sunlit mangrove foliage orients itself in a more vertical position compared to shaded foliage [Farnsworth and Ellison, 1996], and foliage in the canopy crown can be nearly vertical [Clough et al., 1982]. This adaptation mechanism allows efficient penetration of solar irradiance into deeper regions of the mangrove forest canopy resulting in comparable rates of photosynthesis above and below the forest crown. In the summer months, this strategy can be most effective at dispersing direct beam irradiance throughout the canopy at peak solar elevation angles. During this time, differences in absorption profiles of direct and diffuse solar beam can be small. When solar elevation angles are lower during the dry season months, the penetration of the direct solar beam into deeper regions of the canopy is reduced, and the differences in NEE due to differences in Kt are large (3 to 6 µmol (CO2) m-2s-1) compared to those (~2 µmol (CO2) m-2s-1) observed in the summer.
U.S. Department of the Interior, U.S. Geological Survey
This page is: http://sofia.usgs.gov/publications/papers/mang_exchange/discussion.html
Comments and suggestions? Contact: Heather Henkel - Webmaster
Last updated: 04 September, 2013 @ 02:04 PM(TJE)