Elevated CO2 stimulates marsh elevation gain ...

Elevated CO2 stimulates marsh elevation gain, counterbalancing sea-level rise J. Adam Langleya,b, Karen L. McKeec, Donald R. Cahoond, Julia A. Cherrye, and J. Patrick Megonigala,1 aSmithsonian

Environmental Research Center, 647 Contees Wharf Road, Edgewater, MD 21037; bDepartment of Biology, Villanova University, 800 Lancaster Avenue, Villanova, PA 19085; cU.S. Geological Survey, National Wetland Research Center, 700 Cajundome Boulevard, Lafayette, LA 70506; dU.S. Geological Survey, Patuxent Wildlife Research Center, 10300 Baltimore Avenue, Beltsville, MD 20705; and eDepartment of Biological Sciences and New College, University of Alabama, Box 870206, Tuscaloosa, AL 35487

Tidal wetlands experiencing increased rates of sea-level rise (SLR) must increase rates of soil elevation gain to avoid permanent conversion to open water. The maximal rate of SLR that these ecosystems can tolerate depends partly on mineral sediment deposition, but the accumulation of organic matter is equally important for many wetlands. Plant productivity drives organic matter dynamics and is sensitive to global change factors, such as rising atmospheric CO2 concentration. It remains unknown how global change will influence organic mechanisms that determine future tidal wetland viability. Here, we present experimental evidence that plant response to elevated atmospheric [CO2] stimulates biogenic mechanisms of elevation gain in a brackish marsh. Elevated CO2 (ambient ⴙ 340 ppm) accelerated soil elevation gain by 3.9 mm yrⴚ1 in this 2-year field study, an effect mediated by stimulation of below-ground plant productivity. Further, a companion greenhouse experiment revealed that the CO2 effect was enhanced under salinity and flooding conditions likely to accompany future SLR. Our results indicate that by stimulating biogenic contributions to marsh elevation, increases in the greenhouse gas, CO2, may paradoxically aid some coastal wetlands in counterbalancing rising seas. coastal wetlands 兩 nitrogen pollution 兩 tidal marsh loss 兩 root productivity 兩 salinity

T

he world currently loses thousands of hectares of low-lying coastal wetlands to shallow open water each year (1–3), attributable, in part, to a recent acceleration of sea-level rise (SLR) (4–6). Loss of coastal wetlands threatens critical services these ecosystems provide, such as supporting commercially important fisheries, providing a wildlife habitat, improving water quality, and buffering human populations from oceanic forces (3). Recent catastrophes, such as Hurricane Katrina and the Asian Tsunami, have underscored the importance of understanding factors that govern sustainability of coastal wetlands in the face of climate change and accelerating SLR. Marshes must build vertically through accumulation of mineral and organic matter to maintain a constant elevation relative to sea level (7). To explain the dynamics of coastal wetland elevation, researchers have traditionally focused on abiotic factors, such as reductions of mineral sediment loads from hydrologic modifications (8). However, organic matter dynamics have a clear importance in peaty soils, which are composed mostly of live and dead plant tissues (9, 10) and may also play an important role in stabilizing mineral soils (11). Organic mechanisms may be especially sensitive to other global change factors and may determine the fate of tidal wetlands. Rising atmospheric CO2 is largely responsible for recent global warming and will continue to contribute to accelerating SLR through thermal expansion and ice melt (12). Elevated CO2, in addition to accelerating SLR, may have important biologically mediated effects on coastal wetland ecosystems, such as stimulating plant productivity (13). The effects of elevated CO2 must be considered along with other regionally or locally important www.pnas.org兾cgi兾doi兾10.1073兾pnas.0807695106

factors that are changing simultaneously, such as nitrogen enrichment (14), flooding (15), and salinity (13). We manipulated atmospheric CO2 and soil nitrogen availability in a tidal marsh with low mineral sediment inputs, allowing us to isolate organic controls on marsh elevation in Chesapeake Bay. This experiment was designed to track sensitively how changes in plant productivity influenced soil elevation. Determining soil surface elevation change is important for 2 reasons. First, the loss in soil elevation relative to local sea level may provide an early indication of the syndrome of tidal wetland collapse (16). Second, tracking elevation changes in discrete strata through time, along with measurements of plant productivity and other environmental variables, allows identification of specific mechanisms critical to the persistence of tidal wetlands under accelerating SLR. To examine how CO2 may interact with other factors that will accompany SLR, we manipulated CO2, salinity, and flooding in a companion greenhouse study (17). Results and Discussion Elevated CO2 stimulated above-ground productivity in the marsh by 30% in 2006 (Table 1; two-way ANOVA, CO2 effect: P ⬍ 0.05) but not in 2007 (P ⬎ 0.70) during a severe drought. These responses were consistent with a 20-year record of elevated CO2 treatment in a previous CO2 study on the same marsh (13). Strikingly, elevated CO2 accelerated the rate of soil elevation gain over 2 years of treatment. Least-squares linear trends revealed a slight loss of elevation in ambient CO2 (⫺0.9 mm yr⫺1) compared with an elevation gain (3.0 mm yr⫺1) in the elevated CO2 treatment (Fig. 1A; two-way ANOVA comparing linear trends, n ⫽ 5, CO2 effect: P ⬍ 0.05). Nitrogen addition tended to affect elevation negatively, although the overall nitrogen effect was not significant. Deposition of mineral sediment on the marsh surface is negligible at this site. Root zone expansion by accumulation of plant material is essential to maintaining a constant surface elevation relative to rising sea level. Indeed, changes in marsh surface elevation mirrored changes in root zone thickness (Fig. 1B). Elevated CO2 caused an increase in thickness of 4.9 mm yr⫺1 compared with 0.7 mm yr⫺1 in the ambient CO2 treatment, an effect that could increase marsh tolerance for SLR. Increases in soil elevation were driven by a stimulation of subsurface plant productivity. Changes in soil elevation over each growing season closely followed the pattern of treatment Author contributions: J.P.M., D.R.C., and J.A.L. designed the field study; J.A.L., J.P.M., and D.R.C. performed and analyzed the field study; K.L.M. and J.A.C. designed and analyzed the mesocosm study; J.A.C. performed the mesocosm study; and J.A.L., K.L.M., D.R.C., J.A.C., and J.P.M. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. Freely available online through the PNAS open access option. 1To

whom correspondence should be addressed. E-mail: [email protected].

This article contains supporting information online at www.pnas.org/cgi/content/full/ 0807695106/DCSupplemental.

PNAS Early Edition 兩 1 of 5

ECOLOGY

Edited by Christopher B. Field, Carnegie Institution of Washington, Stanford, CA, and approved February 20, 2009 (received for review August 13, 2008)

Table 1. Above-ground plant biomass and fine root production over 2 growing seasons of the field study

Aboveground Treatments

Fine roots

Aboveground Oct-07

Fine roots

Jul-06

Oct-06

2006

July-07

2007

Ambient CO2

387 (27)a

677 (68)a

56 (11) ab

524 (53)a

449 (25)a

108 (17)ab

Ambient CO2 +Nitrogen Elevated CO2

603 (127)b

952 (144)a

40 (11) a

842 (149)b

697 (167)b

63 (9)a

387 (26)a

883 (111) a

98 (39)ab

530 (39)a

399 (34)a

146 (14)b

Elevated CO2 +Nitrogen

560 (44)ab

1393 (82)b

110 (13) b

902 (65)b

702 (55)b

104 (16)ab

Values represent means (SE) in g m⫺2. Treatment means without common superscript symbols were different (P ⱕ 0.05) according to a Student’s t test multiple comparison.

effects on fine root productivity (Fig. 2; r2 ⫽ 0.86, P ⬍ 0.05). By the end of the second year, above-ground litter also may have entered the soil and contributed to a CO2 effect on elevation. Rhizome and subsurface stem productivity were likely also important but could not be quantified nondestructively in the field. Although the presence or absence of an elevated CO2 effect on above-ground production each year depended on climatic conditions, elevated CO2 strongly stimulated fine root productivity both years, by 48% in 2006 (two-way ANOVA, n ⫽ 5, P ⬍ 0.05) and by 26% in 2007 (P ⬍ 0.05). This finding concurs with previous work at this site and many other CO2 studies that report an inordinately large stimulation of root growth (18).

A

The timing of the below-ground response further supports the important role of plant growth in driving soil elevation gain. The CO2 effects on root zone elevation arose during the first growing season (Fig. 1), concurrent with the production of live plant biomass. CO2 effects on elevation during the growing season were not reversed by plant senescence and decomposition. Because anaerobic conditions in saturated soils substantially slow soil organic matter decay (11), the effects of productivity on elevation during the growing season persist through the nongrowing season. Consequently, elevated CO2 forced the soil surface upward by stimulating the accumulation of both live and dead organic matter. Other perturbations, particularly those that may accompany SLR, could modify the influence of elevated CO2 on marsh elevation gain. To understand better how the CO2 response under current sea-level conditions might differ with accelerating SLR, we examined the effect of CO2 in combination with flooding and salinity in a mesocosm study. This experimental approach allowed isolation of root-zone controls on soil elevation dynamics and manipulation of factors known to influence root growth. The objective of our treatments was to vary the depth and duration of flooding in combination with a range of salinities anticipated under different SLR scenarios (17).

B

Fig. 1. Relative surface elevation (A) and root zone thickness (B) over 2 growing seasons in a brackish marsh (Chesapeake Bay) exposed to ambient (Amb) or elevated (Elev) atmospheric [CO2] in open-top chambers and to factorial soil nitrogen (⫹N) enrichment. The slopes of the linear trends in elevation are inset in each panel (mean ⫾ 1 SE). Elevated CO2 increased elevation (two-way ANOVA, CO2 effect: P ⬍ 0.05), an effect that mirrored CO2 stimulation of root zone thickness (P ⬍ 0.05). Nitrogen addition had a nonsignificant tendency to reduce total elevation gain (P ⫽ 0.18) and root zone expansion (P ⫽ 0.13), despite stimulating above-ground plant productivity. 2 of 5 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0807695106

Fig. 2. Relation of subsurface biovolume production to vertical change in the field. Values represent treatment means (⫾1 SE) of relative elevation change and fine root productivity over each field season, 2006 (circles) and 2007 (squares). Amb, ambient; Elev, elevated; N, nitrogen.

Langley et al.

As found in the field study, CO2 had a significant effect on marsh elevations in mesocosms (17). However, the patterns of soil expansion, driven by below-ground production, varied in a complex manner because of interactions of CO2, flooding, and salinity (Fig. 3). Under intermittent flooding, CO2 enhanced elevation at lower salinities but had no effect or a negative effect at higher salinities (Fig. 3). Under constant flooding, however, the pattern of CO2 response was different, with expansion greatest both above and below the current salinity range for this marsh type (Fig. 3). In particular, CO2 treatment increased elevation in flooded mesocosms at high salinities (i.e., conditions that simulated future SLR). Although the precise CO2 response in nature will depend on many other factors in addition to salinity and flooding, our findings indicate that future CO2 effects on elevation in Kirkpatrick Marsh will likely be modified by conditions that accompany SLR. Agricultural and municipal runoff increases nitrogen (N) loading in estuaries, particularly those that drain densely populated watersheds. In the field study, nitrogen addition, simulating coastal eutrophication, tended to negate elevation gains caused by elevated CO2 alone (Fig. 1 A, Inset). Although nitrogen addition stimulated above-ground biomass, it decreased root productivity (Fig. 2 and Table 1). N addition commonly decreases carbon allocation to root productivity (19). Furthermore, N addition has been shown to stimulate decomposition in other nitrogen-limited peat-based wetlands (20). Nitrogen pollution may contribute to marsh loss (5), although the mechanisms have not been described. In our study, N addition tended to reduce elevation gain, partially by reducing root productivity, but also may have stimulated organic matter decomposition. Plant community composition will modify ecosystem responses to elevated CO2. Plants with the C3 photosynthetic pathway typically exhibit a stronger CO2 response than C4 plants (13). The CO2 effect may be particularly important for C3dominated marshes experiencing SLR, which includes many brackish marshes and mangrove swamps. In this marsh community, the responsive C3 species is more flood-tolerant and less salt-tolerant than the C4 species (21, 22). The mesocosm experiment also provided insight into C3 vs. C4 contributions to soil Langley et al.

expansion (17). Separation of roots by species was not possible, but species contributions to soil volume in the mesocosms could be assessed by measuring the final volume of shoot bases (i.e., the below-ground portion of the shoot that serves as a storage organ in graminoids). Vertical change in soil elevation was positively correlated with subsurface shoot volume of the C3 but not the C4 species (17). Consequently, the potential for CO2 stimulation under an accelerating SLR scenario may depend on the inherent tolerance of component species to salinity, flooding, and other factors as well as their potential to respond to CO2 (17). Our findings bear particular importance given the threat of accelerating SLR to coastal wetlands worldwide. A recent report suggests that a 2-mm increase in the rate of SLR will threaten or eliminate a large portion of mid-Atlantic marshes (6). The threshold rate of SLR (i.e., the SLR rate at which the marsh converts to open water) depends on the rate of marsh elevation gain by accumulating sediment or endogenous organic material (7). For instance, models predict that increasing mineral sediment load should increase this threshold rate (7, 23). Nevertheless, many marshes, like Kirkpatrick Marsh, naturally receive very little exogenous sediment. Therefore, the critical SLR threshold depends almost entirely on biological mechanisms that have been more difficult to describe and predict. We suggest that elevated CO2, by stimulating below-ground plant productivity and biogenic elevation gain, may also increase the capacity of coastal wetlands to tolerate SLR (Fig. 4). Atmospheric CO2 has risen 40% from preindustrial levels (⬇280 ppm) to the current concentration (⬇388 ppm) over the past 200 years and has enhanced plant productivity globally. A CO2 effect on coastal wetland elevation may have already PNAS Early Edition 兩 3 of 5

ECOLOGY

Fig. 3. Elevated CO2 effects on marsh elevation in greenhouse mesocosms subjected to different salinity and flooding treatments. Values are the difference in elevation (mm) between mesocosms under elevated CO2 (720 ppm) and ambient CO2 (380 ppm). The y axis values represent mean differences in CO2 response (⫾ 1 SE) under intermittent (circles) and constant (squares) flooding in relation to mean interstitial salinity (⫾ 1 SE) measured biweekly in mesocosms (x axis). Second-order polynomial curves were fitted to the data; t tests of quadratic coefficients indicated a significant change over salinity for both flooding treatments (P ⬍ 0.05). Patterns of response for the 2 flooding treatments were significantly different (F ⫽ 3.03, P ⫽ 0.10; analysis of covariance).

Fig. 4. Conceptual comparison of SLR and atmospheric [CO2] scenarios from the Inter-governmental Panel on Climate Change (IPCC) to hypothetical thresholds of marsh tolerance for increasing SLR. The black line approximates a simplified SLR ⫻ CO2 relation for the past 140 years (4, 12, 29, 30). Red and orange lines represent most probable increases in atmospheric [CO2] and SLR rates for the years 2090 –2100 according to model projections for 2 moderate IPCC socioeconomic scenarios, A2 and B2 (31). The difference between the 2 lines depends on socioeconomic uncertainties. Each model output has its own error based on physical uncertainties that are not shown here. The relation between SLR and CO2 appears approximately linear in the 100-year projections, but other relations are possible. The dotted green line represents a threshold SLR rate fixed at 4 mm yr⫺1, a rate that would threaten many marshes (6). The dashed green line represents a hypothetical threshold rate that depends on atmospheric [CO2] and is based on our finding of increased marsh elevation gain with elevated CO2. The CO2 effect (difference between the 2 green lines) portrayed here (⬇2 mm yr⫺1 at 720 ppm) is conservative compared with that measured in the field experiment (3.9 mm yr⫺1) because of uncertainties in extrapolating from short-term CO2 effects on elevation to actual marsh thresholds. Marshes collapse when the rate of SLR (red or orange lines) exceeds the threshold of marsh tolerance for SLR (green lines). Thresholds may have been lower in the past when [CO2] was at preindustrial levels (⬇280 ppm), because many plants likely exhibited reduced below-ground growth compared with today, and thus made smaller contributions to marsh elevation gain.

mitigated a significant amount of marsh loss acting through the biological mechanisms demonstrated in our study (Fig. 4). However, the stimulatory effects of CO2 on leaf-level photosynthesis are known to diminish with each marginal increase in CO2 concentration (24), and a similar attenuation of the stimulatory effect of elevated CO2 on soil elevation gain should follow. Consequently, further increases in atmospheric CO2 beyond 720 ppm may yield diminishing benefits for plant productivity and marsh elevation, while further accelerating SLR through greenhouse warming (12). A simultaneous increase in plant stressors, such as salinity and flooding, however, may sustain or enhance the CO2 effect, as found in the mesocosm study. Our findings show that elevated CO2 stimulates plant productivity, particularly below ground, thereby boosting marsh surface elevation. In addition to providing evidence for organic control of wetland topography on the landscape scale, we show how elevated CO2 may strengthen the organic mechanisms that sustain marsh elevation relative to SLR in the future. These effects likely extend beyond biogenic marshes, because organic processes can exert control over elevation dynamics in mineralsoil wetlands as well (11). For instance, marshes in the Mississippi River delta have high rates of subsidence that were naturally countered by mineral sediment deposition (25). However, hydrological modifications by humans have diminished sediment deposition in this system. Such sediment-deficient deltas, which once would have been buffered against increasing SLR, have been rendered more vulnerable, and hence more dependent on organic contributions to sustain a constant elevation relative to SLR. To manage or restore such ecosystems to a self-sustaining state, we must explore further how global change interactions will influence key biological mechanisms that maintain soil elevations in coastal wetland ecosystems. Materials and Methods The field study was conducted in Kirkpatrick Marsh (38°53⬘ North, 76°33⬘ West), located on a microtidal subestuary of Chesapeake Bay. The site is dominated by a perennial C3 sedge, Schoenoplectus americanus, and the C4 perennial grasses Spartina patens and Distichlis spicata and is representative of brackish high marshes of mid-Atlantic North America (26). The soils at the site are organic (⬎80% organic matter) to a depth of ⬇5 m. Mean tidal range is 44 cm. The high marsh zone is 40 – 60 cm (26) above daily mean low water level and is inundated by 28% of high tides. Salinity averages 10 parts per thousand (ppt) and ranges seasonally from 4 to 15 ppt. The average daily air temperature reaches a low of ⫺4 °C in January and a high of 31 °C in July. The 80-year trend of SLR is ⬇3.4 mm yr⫺1 (tidesandcurrents.noaa.gov, Annapolis), which outpaces the long-term local eustatic trend attributable to glacial isostatic adjustment (27). Beginning in May 2006, we exposed 20 3.3-m2 marsh plots to 2 atmospheric CO2 concentrations (ambient and ambient ⫹ 340 ppm) to simulate a likely atmospheric CO2 concentration projected for the year 2100. The 2 CO2 levels were crossed with 2 levels of N addition (0 and 25 g of nitrogen m⫺2 yr⫺1). Ammonium chloride was dissolved in 5 L of water from the Rhode River, the subestuary adjacent to the site. On 5 dates (approximately monthly, avoiding high tides) throughout the growing season, we used backpack sprayers to deliver a fertilizer solution to 10 plots. The fertilizer solution was then rinsed

1. Dahl TE (2005) Status and Trends of Wetlands in the Conterminous United States, 1998 to 2004 (U.S. Department of the Interior, Fish and Wildlife Service, Washington, DC). 2. Valiela I, Bowen JL, York JK (2001) Mangrove forests: One of the world’s threatened major tropical environments. Bioscience 51:807– 815. 3. Zedler JB, Kercher S (2005) Wetland resources: Status, trends, ecosystem services, and restorability. Annual Review of Environment and Resources 30:39 –74. 4. Church JA, White NJ (2006) A 20th century acceleration in global sea-level rise. Geophys Res Lett 33, 10.1029/2007GL029703. 5. Hartig EK, Gornitz V, Kolker A, Mushacke F, Fallon D (2002) Anthropogenic and climate-change impacts on salt marshes of Jamaica Bay, New York City. Wetlands 22:71– 89. 6. Reed DJ, et al. (2008) Site-Specific Scenarios for Wetlands Accretion as Sea Level Rises in the Mid-Atlantic Region. Section 2.1. Background Documents Supporting Climate Change Science Program Synthesis and Assessment Product, eds Titus JG, Strange EM (EPA 430R07004, U.S. EPA, Washington, DC).

4 of 5 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0807695106

from standing vegetation with another 5 L of unamended river water applied with backpack sprayers. Each fertilizer application simulated 5 g of nitrogen m⫺2 in the equivalent of 0.5 cm of river water. The 10 unfertilized chambers received 10 L of unamended river water applied in the same manner (see SI Text). We outfitted each plot with a surface elevation table (SET) to measure soil elevation change sensitively (Figs. S1 and S2). Deep-rod SETs, used in conjunction with shallow benchmarks, enabled elevation changes to be partitioned between the root zone (0 – 0.3 m in depth) and the deep zone (0.3– 4.5 m in depth). We used elevation measurements in adjacent unchambered reference plots to account for chamber effects and to minimize the confounding effects of spatial variability.

Greenhouse Methods To examine how CO2 might affect elevation response under simulated SLR scenarios, data from a greenhouse mesocosm experiment (17) were used. This experiment used the same marsh type as the field study at the mid-Atlantic location but was conducted using soil and plant material from the Mississippi River delta and carried out at the Wetland Elevated CO2 Experimental Facility (National Wetlands Research Center, Lafayette, LA). A detailed description of the facility and the experimental design is given elsewhere (17, 28) but is briefly repeated here to provide context for this study. Blocks of peat with intact vegetation (sods) were collected from a brackish marsh community (ca. 50:50 mixture of S. americanus and S. patens) in coastal Louisiana and established in mesocosms (0.1-m2 area ⫻ 0.4 m in depth, 25-L volume). In March 2005, 60 replicate mesocosms were randomly assigned to 1 of 4 greenhouses with atmospheric concentrations of 380 (ambient) or 720 (elevated) ppm of CO2, 3 flooding levels (constantly flooded, intermittently flooded, drained), and 5 salinity levels (0, 5, 10, 15, 20 ppt of sea salts). All above-ground material was removed on senescence so that the only inputs were from below-ground production. Expansion or contraction of the marsh sods was measured for 1 year with mini-SETs [modeled after SETs used in the field study (17)]. CO2 response was calculated as the difference between elevated and ambient CO2 treatments and plotted vs. interstitial salinity (averaged over experimental period). The drained treatment was not included in the analysis, because our objective in this study was to understand better how CO2 response to current SLR (observed in the field) might change under a future SLR scenario of increased flooding. Second-order polynomial curves were fitted to the data, and overall differences were tested with analysis of covariance (using salinity as a continuous effect and flooding as a categorical effect in the model); t tests of quadratic and linear coefficients (for each fitted curve) also were used to assess differences. ACKNOWLEDGMENTS. We thank A. Anteau, H. Baldwin, T. McGinnis, E. Travis, and W. Vervaeke for assistance with data collection and maintenance of mesocosm experimental treatments in the Wetland Elevated CO2 Experimental Facility. We thank J. Duls, J. Keller, M. Sigrist, G. Peresta, B. Drake, E. Sage, A. Martin, D. McKinley, J. Lynch, and N. Mudd for construction and maintenance of the field experiment at the Smithsonian Climate Change Facility. We appreciate comments from M. Erwin, M. Kirwan, H. Neckles, S. Chapman, B. Hungate, T. To¨rnqvist, and an anonymous reviewer. The field study was supported by the U.S. Geological Survey Global Change Research Program, U.S. Department of Energy (Grant DE-FG02–97ER62458), the U.S. Department of Energy’s Office of Science (BER) through the Coastal Center of the National Institute of Climate Change Research at Tulane University, and the Smithsonian Institution. The U.S. Geological Survey Global Change Research Program provided financial support for the mesocosm study. Any use of trade, product or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

7. Morris JT, Sundareshwar PV, Nietch CT, Kjerfve B, Cahoon DR (2002) Responses of coastal wetlands to rising sea level. Ecology 83:2869 –2877. 8. Turner RE (2004) Coastal wetland subsidence arising from local hydrologic manipulations. Estuaries 27:265–272. 9. Nyman JA, Chabreck RH, Kinler NW (1993) Some effects of herbivory and 30 years of weir management on emergent vegetation in brackish marsh. Wetlands 13:165–175. 10. McKee KL, Cahoon DR, Feller IC (2007) Caribbean mangroves adjust to rising sea level through biotic controls on change in soil elevation. Global Ecol Biogeogr 16:545–556. 11. Nyman JA, Walters RJ, Delaune RD, Patrick WH (2006) Marsh vertical accretion via vegetative growth. Estuar Coast Shelf Sci 69:370 –380. 12. Bindoff NL, et al. (2007) Climate Change 2007: The Physical Science. Basis Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, eds Solomon S, et al. (Cambridge Univ Press, New York), pp 387– 429. 13. Erickson JE, Megonigal JP, Peresta G, Drake BG (2007) Salinity and sea level mediate elevated CO2 effects on C-3-C-4 plant interactions and tissue nitrogen in a Chesapeake Bay tidal wetland. Global Change Biol 13:202–215.

Langley et al.

24. Jacob J, Greitner C, Drake BG (1995) Acclimation of photosynthesis in relation to Rubisco and non-structural carbohydrate contents and in situ carboxylase activity in Scirpus olneyi grown at elevated CO2 in the field. Plant Cell Environ 18:875– 884. 25. Coleman JM, Roberts HH, Stone GW (1998) Mississippi River delta: An overview. J Coast Res 14:698 –716. 26. Jordan TE, Correll DL (1991) Continuous automated sampling of tidal exchanges of nutrients by brackish marshes. Estuar Coast Shelf Sci 32:527–545. 27. Douglas BC (2005) Gulf of Mexico and Atlantic coast sea level change. Geophysical Monograph 161:111–121. 28. McKee K, Rooth JE (2008) Where temperate meets tropical: Multi-factorial effects of elevated CO2, nitrogen enrichment, and competition on a mangrove-salt marsh community. Global Change Biol 14:971–984. 29. Gehrels WR, et al. (2005) Onset of recent rapid sea-level rise in the western Atlantic Ocean. Quaternary Science Reviews 24:2083–2100. 30. Donnelly JP, Cleary P, Newby P, Ettinger R (2004) Coupling instrumental and geological records of sea-level change: Evidence from southern New England of an increase in the rate of sea-level rise in the late 19th century. Geophys Res Lett 10.1029/2003GL 018933. 31. Meehl GA, et al. (2007) Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, eds. Solomon S, et al. (Cambridge Univ Press, New York), pp 749 – 844.

ECOLOGY

14. Morris JT, Bradley PM (1999) Effects of nutrient loading on the carbon balance of coastal wetland sediments. Limnol Oceanogr 44:699 –702. 15. Donnelly JP, Bertness MD (2001) Rapid shoreward encroachment of salt marsh cordgrass in response to accelerated sea-level rise. Proc Natl Acad Sci USA 98:14218 –14223. 16. Cahoon DR, et al. (2003) Mass tree mortality leads to mangrove peat collapse at Bay Islands, Honduras after Hurricane Mitch. J Ecol 91:1093–1105. 17. Cherry JA, McKee K, Grace JB (2009) Elevated CO2 enhances biological contributions to elevation change in coastal wetlands by offsetting stressors associated with sea-level rise. J Ecol 97:67–77. 18. Rogers HH, Runion GB, Krupa SV (1994) Plant-responses to atmospheric CO2 enrichment with emphasis on roots and the rhizosphere. Environ Pollut 83:155–189. 19. Agren GI, Franklin O (2003) Root:shoot ratios, optimization and nitrogen productivity. Ann Bot (London) 92:795– 800. 20. Bragazza L, et al. (2006) Atmospheric nitrogen deposition promotes carbon loss from peat bogs. Proc Natl Acad Sci USA 103:19386 –19389. 21. Broome SW, Mendelssohn IA, McKee KL (1995) Relative growth of Spartina-patens (Ait) Muhl and Scirpus-olneyi Gray occurring in a mixed stand as affected by salinity and flooding depth. Wetlands 15:20 –30. 22. Rasse DP, Peresta G, Drake BG (2005) Seventeen years of elevated CO2 exposure in a Chesapeake Bay wetland: Sustained but contrasting responses of plant growth and CO2 uptake. Global Change Biol 11:369 –377. 23. Kirwan ML, Murray AB (2007) A coupled geomorphic and ecological model of tidal marsh evolution. Proc Natl Acad Sci USA 104:6118 – 6122.

Langley et al.

PNAS Early Edition 兩 5 of 5

Supporting Information Langley et al. 10.1073/pnas.0807695106 SI Methods Field Experiment Methods Open-Top Chamber Design. The chamber design was modified from previous open-top chamber designs (1). The base was an octagonal frame made of aluminum angle stock inserted 10 cm into the ground. Mounted on the base was a hollow octagonal manifold (cross section: 30 cm high ⫻ 10 cm wide) consisting of an aluminum frame and transparent acrylic walls. The purpose of the manifold was to distribute air equally to the 8 sides of the chamber. Two hundred 1-cm diameter holes were drilled on the inside of the manifold so that air would flow equally from each side of the octagon. Ambient air was forced into the manifold with a blower mounted on a stand about 1.5 m above the marsh surface to avoid high tides. The blowers were placed at least 6 m away from each chamber and oriented to avoid shading the study area (Fig. S2). A 20-cm diameter duct carried air from the blower to the chamber manifold. On top of the manifold sat an octagonal prism (120 cm high) built from 2.5-cm PVC pipe. These PVC frames were placed into welded supports on the aluminum manifold to create an octagonal-prism chamber 1.5 m tall, 2 m in diameter, and with 1-m sides. Removable rectangular panels were made from aluminum, covered with IR-transparent film (Aclar 22A; Honeywell), and attached to the PVC frame with custom-fitted PVC snaps.

the quadrats, each S. americanus stem was counted and nondestructively measured for total height, green height, and width at half-height. In the corner of each quadrat, we clipped and removed all vegetation and litter in a 5 ⫻ 5 cm area. Vegetation was sorted according to species. We measured the clipped S. americanus stems for total height and width. Clippings were dried for 72 h at 60 °C and weighed. To estimate total S. americanus mass, we developed an allometric relation using height and width to estimate dry mass of individual stems [mass ⫽ (height ⫻ width ⫻ 0.0028) ⫺ 0.14, r2 ⬎ 0.9]. To estimate S. patens and D. spicata mass, we scaled up from mass in the clipped areas to total chamber area. To capture late-season growth, we repeated the above measurements using 4 quadrats in each chamber in October. Three soil cores (30 cm in depth ⫻ 5 cm in diameter) were taken from a prescribed location in each plot. Cylindrical ingrowth bags were constructed (30 cm in height ⫻ 5 cm in diameter) from 1-cm pore-size mesh and filled with milled moistened peat to achieve the bulk density of in situ peat, 0.12 g cm⫺3. Bags were implanted in winter and removed in November the following year. Contents were washed over a 1-mm sieve. Larger organic fragments were picked out by hand. Root mass was separated into fine (⬍2 mm in diameter) and coarse (⬎2 mm in diameter) categories, dried for 72 h at 60 °C, and weighed. Results of plant biomass and productivity measurements are given in Table 1.

CO2 Delivery and Sampling. CO2 was delivered to each of the 10

elevated CO2 chambers at a rate of ⬇6 L min⫺1 to achieve a target concentration of 720 ppm. Each CO2 delivery line was controlled with metered valves and fed into the intake chimney on the blower for each respective elevated chamber. Infusing the CO2 upstream of the blower ensured sufficient mixing before air entered the chamber manifold. Two sample lines continuously delivered air from each chamber: one line sampled manifold air, and the other sampled the chamber atmosphere. Gas flow from each of the 40 lines entered a separate solenoid valve (model 3V1; Sizto Tech Corporation) that could either direct the flow into a gas analyzer or exhaust it into the atmosphere. The system was programmed so that only 1 chamber at a time was being measured, with samples from the remaining chambers going to exhaust. The solenoids for 1 manifold line and 1 chamber line from a single chamber were activated simultaneously, feeding samples to a pair of IR gas analyzers. Each chamber was sampled at least once every 40 min. We used the [CO2] of the manifold line to monitor the amount of CO2 delivered to each manifold. The chamber air sampling line was used to monitor the chamber atmosphere and to fine-tune the delivery rate to achieve our target concentration (ambient ⫹ 340 ppm). Average daily mean CO2 during daylight hours in 2006 was 395 and 669 ppm in ambient and elevated chambers, respectively. Values improved to 394 and 707 ppm in 2007, which is a treatment difference of 313 ppm. SE among chambers was 1.2 and 6.0 ppm for ambient and elevated treatments, respectively. The SD among daily means for individual chambers averaged 21.9 and 59.0 ppm for individual ambient and elevated chambers, respectively. Above-ground Biomass and Root Productivity. We estimated above-

ground biomass by a combination of allometry and harvest subplots (2). At the end of July, 8 quadrats that were each 30 ⫻ 30 cm were placed in prescribed locations in each plot, 6 inside the chamber and 2 in an adjacent unchambered control plot. In Langley et al. www.pnas.org/cgi/content/short/0807695106

Elevation Measurements. Measurements of soil elevation at each

plot were made with a SET (3) (Fig. S1 and Fig. S2) modified to accommodate plot dimensions. Outside each experimental chamber, a posthole was dug roughly 15 cm in diameter and 20 cm deep. A 30-cm segment of PVC pipe (15 cm in diameter) was placed into the hole. In the center of the PVC pipe, a series of attachable stainless steel rods was driven with an electric hammer through the entire profile of organic matter (4–5 m) until the point of refusal (6–7 m); at that point, the PVC pipe was anchored in a mineral deposit underlying the marsh. Concrete was poured into the PVC pipe, securing the top of the SET rod. To isolate the influence of root zone processes on elevation, we implanted ‘‘shallow benchmarks’’ to a depth of 30 cm. The vertical movement of these benchmarks results from processes that occur below a depth of 30 cm. The benchmarks were made of aluminum pipe (5 cm in diameter ⫻ 40 cm long). Several 1-cm diameter holes were drilled into the sides of the lower 10 cm of the pipes so that roots would grow through and anchor the benchmarks in place. Six benchmarks were implanted to a depth of 30 cm under the path of the SET arm in each chamber, 3 inside the chamber and 3 outside. After placement, solid caps were placed on the top of the pipes. All these perturbations as well as boardwalks to service each plot were completed in the summer of 2005, at least 9 months before the beginning of the experiment. At intervals ranging from 1 to 3 months, a modified horizontal aluminum SET arm (4 m long compared with ⬍0.5 m long for the original rod SET design) was attached to the top of the SET rod, leveled precisely, and affixed to an aluminum support post at the other end. The arm provided a horizontal reference of known elevation across the soil surface; changes in the distance from this reference surface to the soil surface are a sensitive measure of changes in soil elevation. Fiberglass pins (0.5 cm in diameter), all exactly 91.0 cm in length, were placed through precision-drilled holes in the SET arm at 4-cm increments. In 1 of 4

each chamber, ⬇40 individual measurements were made, and in each adjacent unchambered reference plot, 40 individual measurements were made. Each pin was carefully lowered to the soil surface and gently placed so that no litter or live plant material lay between the pin and the soil surface. The height from the SET arm to the top of each pin was measured to the nearest millimeter, providing a measurement of change in total elevation. Changes in total elevation were partitioned into either the root zone (top 30 cm of soil) or the deep zone (below 30 cm of soil). To measure changes in the thickness of the deep zone, we lowered 2–4 pins to the surface of each of the 6 shallow benchmarks (3 inside and 3 outside each chamber) and measured height in the same manner. Elevation measurements were performed by the same person to minimize sampling error. We calculated the change in elevation resulting from processes occurring in the root zone, ⌬ER, from the 2 measured variables following the equation ⌬ER ⫽ ⌬ET ⫺ ⌬ED, where ⌬ET repre-

sents the change in total elevation and ⌬ED represents elevation change attributable to change in thickness of the deep zone. Surface accretion was measured using feldspar marker horizons in each plot (4). Because roots that grew above soil accumulated above the marker horizon, this surface horizon was not functionally distinct from the root zone. Therefore, surface accretion was not mathematically separated from root zone dynamics in the present study. Changes in total soil elevation were strongly related to innate spatial and temporal variability in deep zone dynamics. Specifically, the thickness of the deep zone changed directly in response to mean monthly sea level through time, and distance from the bank predicted the amplitude of that oscillation. To isolate treatment effects on soil elevation, we accounted for variation by referencing within-chamber SET measurements to SET measurements in the adjacent unchambered plots. Individual linear trends were estimated for individual plots. All treatment comparisons, including those of plant response, were made using two-way ANOVA (CO2 ⫻ N, n ⫽ 5).

1. Drake BG, Leadley PW, Arp WJ, Nassiry D, Curtis PS (1989) An open top chamber for field studies of elevated atmospheric CO2 concentration on saltmarsh vegetation. Funct Ecol 3:363–371. 2. Erickson JE, Megonigal JP, Peresta G, Drake BG (2007) Salinity and sea level mediate elevated CO2 effects on C-3-C-4 plant interactions and tissue nitrogen in a Chesapeake Bay tidal wetland. Global Change Biol 13:202–215.

3. Cahoon DR, et al. (2002) High-precision measurements of wetland sediment elevation: II. The rod surface elevation table. Journal of Sedimentary Research 72:734 –739. 4. Cahoon DR, Reed DJ, Day JW (1995) Estimating shallow subsidence in microtidal salt marshes of the southeastern United States—Kaye and Barghoorn revisited. Mar Geol 128:1–9.

Langley et al. www.pnas.org/cgi/content/short/0807695106

2 of 4

Fig. S1.

A schematic of the SET in the field experiment (not drawn to scale).


3 of 4

Fig. S2. SET measurements on an experimental chamber and the adjacent reference plot. Note that the blower shown in the background is located ⬇6 m from the plot. (Photograph by M. Sigrist.)


4 of 4