Effects of a prescribed fire on dissolved inorganic carbon dynamics in a nutrient-enriched Everglades wetland

eschweizerbartxxx

Prescribed fi re on DIC dynamics in an Everglades wetland 263

Introduction

Fire has signifi cant impacts on the biological commu-nity and biogeochemical processes in various ecosys-tems (Minshall et al. 1989, Gresswell 1999, Slocum et al. 2003) and has been used as an effective manage-ment and restoration tool (Wan et al. 2001, Miao & Carstenn 2006). Compared to terrestrial systems, there are few studies of the effects of fi re on aquatic sys-tems and most of these studies focus on changes in the biological community (Minshall et al. 1989, Gresswell

DOI: 10.1127/1863-9135/2008/0171-0263 1863-9135/08/0171-0263 $ 2.50 © 2008 E. Schweizerbart’sche Verlagsbuchhandlung, D-70176 Stuttgart

Effects of a prescribed fi re on dissolved inorganic carbon dynamics in a nutrient-enriched Everglades wetland

Binhe Gu1*, ShiLi Miao1, Chris Edelstein2 and Thomas Dreschel1

South Florida Water Management District

With 4 fi gures and 3 tables

Abstract: The purpose of this study was to assess the short-term response of dissolved inorganic carbon (DIC) chemistry to a prescribed fi re in a nutrient-enriched area of the Everglades. Surface water samples were analyzed for water temperature, pH, total carbon (TC) and DIC concentrations in the burn area, in addition to one upstream and two downstream stations, pre- and post-fi re for a period of 14 weeks. Dissolved free carbon dioxide (CO2), the partial pressure of CO2 (pCO2) and CO2 fl ux between surface water and the atmosphere were calculated. Although a large amount of unburned litter remained in the burned area and an immediate increase in the DIC concentration was observed post-fi re, the average DIC concentration at the burn area was only slightly higher than the upstream station and lower than the two downstream stations. Similarly, CO2 concentrations and pCO2 at the burn area were lower than those of the three control stations. The low CO2 and pCO2 immediately post-fi re were likely attributed to the elevated pH due to the addition of basic ash. However, the continuously low CO2 and pCO2 were the combined results of high pH and increased CO2 sequestration by the growth of periphyton triggered by increased availability of light and growth space. These factors also explain why the DIC concentration at the burn area did not show a dramatic increase after the fi re. Overall, our results suggest that despite decreases in pCO2 post-burn, this wetland continues to act as a net source of CO2 to the atmosphere.

Key words: Dissolved inorganic carbon (DIC), CO2, pCO2, CO2 fl ux, Everglades, prescribed fi re, wetland.

Fundamental and Applied Limnology Archiv für Hydrobiologie Vol. 171/4: 263–272, May 2008 © E. Schweizerbart’sche Verlagsbuchhandlung 2008

1999) and nutrients (Minshall et al.1989, Spencer et al. 2003, Meixner et al. 2004). Research on the response of carbon cycling to fi re is limited and focused largely on dissolved organic carbon in the aquatic ecosystems (Carignan et al. 2000, Enache & Prairie 2000, Allen et al. 2003). Little is known regarding the response of dissolved inorganic carbon (DIC) in freshwater wet-lands due to fi re, although DIC can be a limiting nutri-ent to submerged macrophytes and periphyton and is potentially a source of CO2 to the atmosphere (Kling et al. 1991, Cole et al. 1994, Duarte & Prairie 2005).

1 Authors’ addresses: Everglades Division, South Florida Water Management District, 3301 Gun Club Road, West Palm Beach, FL 33406, USA.

2 TBE Group, 2257 Vista Parkway, Suite 19, West Palm Beach, FL 33411, USA.

* Corresponding author; E-mail: [email protected]

eschweizerbartxxx

264 Binhe Gu et al.

Dissolved inorganic carbon in water consists of dissolved free carbon dioxide (CO2), bicarbonate (HCO3

–), and carbonate (CO32–). Transformation be-

tween these forms is dynamic and their relative abun-dance is strongly affected by biological activities (pri-mary production and respiration), water temperature, and pH (Wetzel 2001). In freshwater wetlands, sub-merged macrophytes and periphyton remove dissolved free CO2 from the water column during photosynthe-sis, whereas microbial respiration releases CO2 into the water column. If primary production is greater than respiration, the wetland acts as a sink for CO2 from the atmosphere. Alternately, if primary production is low-er than respiration, the wetland is a source of CO2 to the atmosphere. Many aquatic systems receive exter-nal organic carbon for microbial respiration, leading to surface water being supersaturated with CO2 (Kling et al. 1991, Cole et al. 1994, 2000).

Fire can affect DIC cycling in wetland ecosystems in several ways. First, fi re can cause the temperature and pH of the surface water to increase due to reduced shading by emergent plants, and due to basic ion ad-dition. These may lead to increases or decreases in the DIC concentration and changes in DIC speciation. Second, fi re removes large amounts of standing bio-mass and provides additional nutrients (phosphorus and nitrogen), growth space and solar radiation to the surface water, which may result in increased CO2 se-questration by stimulating the growth of submerged macrophytes and periphyton. This process tends to lower dissolved CO2 concentrations and the pCO2 in surface water. Third, incomplete burning, which is often the case in wild and prescribed fi res, will leave varying amounts of dead organic matter behind for mi-crobial decomposition. Therefore, fi re may add more CO2 to the surface water by providing organic matter for microbial respiration. The changes in DIC dynam-ics depends on the magnitude of biomass removal by fi re and the plant response to open water and nutrient enrichment.

The Everglades ecosystem in Florida, USA, has experienced increasing anthropogenic infl uences dur-ing the past 100 years. Surface water runoff from agri-cultural and urban areas provides nutrients to the Ever-glades ecosystem and has resulted in the replacement of native sawgrass (Cladium jamaicense Crantz) by cattail (Typha spp.) in highly-impacted areas (Reddy et al. 1998). In recent years, Stormwater Treatment Areas (STAs) have been constructed to reduce total phosphorus (TP) from the Everglades agricultural area runoff to reduce the impacts to the Everglades Protec-tion Area. Other management alternatives are also be-

ing investigated to accelerate the recovery process of the impacted wetlands. One of these potential alter-natives is the use of prescribed fi res to eliminate cat-tail biomass since fi re is a natural phenomenon that has contributed to the shaping of the historical Ever-glades landscape (Wu et al. 1996, Slocum et al. 2003, Miao & Carstenn 2006). A long-term study has been initiated to examine the effects of prescribed fi re on an ecosystem in a nutrient-enriched region of the Ev-erglades (Miao & Carstenn 2006). In this paper, we investigated the short-term (days to months) effects of a prescribed fi re on DIC dynamics. We specifi cally ex-amined (1) the effect of fi re on concentrations of DIC and dissolved free CO2; (2) the effect of fi re on pCO2 and CO2 fl ux between the surface water and the atmo-sphere and (3) the underlying mechanisms for changes in DIC cycling.

Methods

Site description and design

The Everglades is located in a subtropical region and its season-ality is typically divided into a wet period from May through November, and a dry period from December through April. This study was conducted in the northern section of Water Con-servation Area 2A (WCA-2A) of the Florida Everglades (Fig. 1) which has been impacted by nutrient-enriched surface water runoff from agricultural and urban areas, and is a region charac-terized by monospecifi c cattail stands (Davis 1991, King et al. 2004). To examine the effects of a prescribed fi re at an ecosys-tem level, six 300 × 300 m (9 ha) research plots were selected in 2005. Two plots were placed in a highly P-enriched area with dense cattail (also referred to as highly-impacted sites, H-sites), two in a moderately P-enriched area with both sawgrass and cattail (also referred to as moderately impacted sites, M-sites), and two in an unimpacted area (also referred to as the refer-ence sites, R-sites) with sawgrass dominance (Miao & Carstenn 2006). Each plot consists of six sampling stations, one upstream (25 m upstream) of the plot, three within the plot (also termed the burn area), and two downstream of the plot, at 25 m and 100 m, respectively (Fig. 1). This study was conducted at H2 which is one of the two highly P-enriched research plots.

Sampling and analysis

The prescribed fi re was initiated in H2 (Fig. 1) on 25 July 2006. Our sampling scheme was designed to capture the immedi-ate response in the burn area to the prescribed fi re with daily sampling during the fi rst 2 weeks and approximately weekly to monthly sampling for the remaining period from 18 July 2006 to 31 October 2006. Water samples were taken from the middle of the water column using Sigma portable pumps (Hach Com-pany, Loveland, CO) or Sigma Streamline 800SL autosamplers (Hach Company, Loveland, CO). The automatic samplers were deployed in one sampling station at the burn area (C1), as well as upstream and 25 m and 100 m downstream of the burn area and programmed to collect samples on an hourly basis. The day of the fi re, samples were composited into two samples, one at

eschweizerbartxxx

Prescribed fi re on DIC dynamics in an Everglades wetland 265

nine hours and the other at 21 hours after fi re initiation. For samples collected the week pre-fi re and on days two through nine post-fi re, hourly samples were composited into one daily sample. To examine the immediate response of surface water to the prescribed fi re, grab samples were collected three hours after ignition at the burned site, upstream and downstream sam-pling locations. Samples were analyzed for total carbon (TC) and DIC following standard methods (APHA 1998). Grab sam-ples were also collected from two other stations (C2 and C3) within the burn area for the measurement of water temperature, pH, DIC and TC concentrations on selected dates to assess eco-system heterogeneity.

In situ water quality parameters

YSItm 600XLM data sondes (YSI Corp. Yellow Springs, CO) were used to collect in situ pH, conductivity, dissolved oxygen (DO) and temperature readings at 30 minute intervals. Sondes were deployed during surface water sample collection events at each of the automatic sampler locations. The sondes were

deployed the day before the burn with the dissolved oxygen sensor at the midpoint of the water column. The sonde in the burn site was installed so that the entire sonde was below the water surface to prevent any fi re damage to the sonde and/or sensors. The sondes were collected, checked for drift, cleaned, calibrated and redeployed on a weekly basis with a typical turn around time of less than one day.

Calculations of CO2, pCO2 and CO2 fl ux

The concentrations of dissolved free CO2 (CO2(aq), mol m–3) were calculated using the following equation (Butler & Cogley 1998):

[CO2(aq)] = CT[H+]2

(1) ([H

+]2 + Ka1[H+] + Ka1Ka2)

where total DIC concentrations (CT) were measured directly. [H+] is the hydrogen ion concentration (mole m–3). First and second dissolution constants Ka1 and Ka2 at the ambient water

Fig. 1. Location of the study site in Water Conservation Area 2A (WCA-2A) in south Florida. The site layout is for H2, a highly phosphorus impacted prescribed burn research plot. The study site is 300 m × 300 m and consisted of three sampling stations (C1, C2 and C3) within the burn area, in addition to one upstream (25 m north of the plot), two downstream sampling stations, one at 25 m and the other at 100 m. M1 and M2 stand for moderately phosphorus impacted plots; Rs and Rc stand for reference sites oc-cupied by sawgrass and cattail, respectively.

eschweizerbartxxx

266 Binhe Gu et al.

temperature were determined using the discrete values at tem-peratures from 10– 40 °C (Butler & Cogley 1998).

The pCO2 in the surface water was calculated using Henry’s Law:

pCO2 = [CO2(aq)] (2)

KH

where KH is a dissolution constant of CO2 (mole m–3 atm–1) which is a function of water temperature:

KH = 0.0279T2 – 2.3895T + 75.819 (R2 = 0.99) (3)

The fl ux of CO2 (FCO2, mol m–2 s–1) between the atmosphere and the surface water was estimated using the following equa-tion:

FCO2 = ([CO2(atm)] – [CO2(aq)]) DCO2 (4)

z

where [CO2(atm)] is the CO2 concentration at the top of the stag-nant layer z (m) which is in equilibrium with atmospheric CO2; [CO2(aq)] is the CO2 concentration at the bottom of the stagnant layer. DCO2 is a diffusion coeffi cient of CO2 (m

–2 s–1) which var-ies with water temperature. We used the linear regression model (Akgerman & Gainer 1972) to estimate temperature-dependent DCO2:

DCO2 = (–12.2048 + 0.04752T) (R2 = 0.91) (5) 109

The stagnant layer z (m) was estimated using a modifi ed regression equation provided by Kling et al. (1992):

z = 10(2.56 – 0.133 W) (R2 = 0.71) (6)

106

where W is the daily average wind speed (m s–1) recorded in situ at 5-minute intervals using HOBOtm microstations.

Periphyton biomass

Periphyton samples were collected once pre-fi re, one month post-fi re, and three months post-fi re from the burn plot (H2) and the control plot (H1) dominated by cattail, and the refer-ence plot (Rs) dominated by sawgrass (Fig. 1). Samples were also collected from the upstream, the burn area and downstream sites for the burn plot (H2) one month post-fi re, following a modifi ed procedure of McCormick & O’Dell (1996). Sam-ples were analyzed for total organic carbon following standard methods.

Results

Fire effects on water temperature, DO and pH

Daily average water temperature during the study pe-riod varied between 22 and 30 ºC (Fig. 2a). Water tem-perature measured in the burn area increased from 25 to 28 ºC during the day of fi re and to approximately 29 ºC on day 1 and 2 post-fi re. The daily maximum increase of water temperature was 5 ºC by day 2. The

average water temperature in the burn area was about 1.5 ºC higher than the average for the upstream and 0.5 to 1.3 ºC higher than those of the two downstream stations (Paired t test, all P < 0.01) during this study period (Table 1).

Measurements of DO concentration for the control stations were incomplete due to instrumental errors; therefore, all DO data from the three upstream and downstream sites were pooled and presented in Fig. 2b along with the DO data for the burn area. Pre-fi re, daily average of the DO concentrations at the burn site was low and decreased to near depletion post-fi re, while DO concentrations at the control sites increased slightly during the same period (Fig. 2b). Dissolved oxygen concentration in the burned area increased dra-matically on Day 13 and remained elevated compared to the pre-fi re and the control sites throughout the re-mainder of the study period (Fig. 2b).

A post-fi re pH peak (8.4) was observed as early as 15 minutes after the fi re reached the burn area sam-pling station (Miao et al., unpubl.). An increase in pH was not observed at the other sites (Fig. 2c). The rapid increase in pH in the burn area subsided somewhat by day two post-fi re although it remained elevated. The average pH at the burn area was 0.17 unit higher than the upstream station and 0.33 and 0.30 unit higher than the 25 and 100 m downstream stations (paired t test, all P < 0.01), respectively (Table 1).

Fire effects on total and dissolved inorganic carbon chemistry

Total carbon concentrations at the burn site increased by nearly 1 mol m–3 for four days post-fi re and then fl uctuated during the remaining study period (Fig. 3a). Increases in TC concentrations in the two down-stream stations were also apparent after the fi re, while a slight increase in TC concentration was observed in the upstream station. The average TC concentration decreased from upstream to the two downstream sta-tions (Table 1).

Table 1. Mean and standard deviation (in parenthesis) of water temperature, pH, total carbon (TC) and dissolved inorganic car-bon (DIC) concentrations at different locations of the study area from 18 July to 31 October 2006. N = 18 days.

Location Water temp(oC)

pH(SU)

TC(mol m–3)

DIC(mol m–3)

Upstream 26.5(1.3) 7.11(0.09) 9.75(0.97) 4.20(0.59)Burn site (C1) 28.0(1.8) 7.28(0.19) 9.97(0.88) 4.32(0.52)25 m Downstream 27.5(1.5) 6.95(0.25) 10.27(1.19) 4.48(0.49)100 m Downstream 26.7(1.5) 6.98(0.30) 10.18(1.01) 4.37(0.54)

eschweizerbartxxx

Prescribed fi re on DIC dynamics in an Everglades wetland 267

Fig. 2. Time-series plots of (a) water temperature, (b) dissolved oxygen concentration (DO) and (c) pH for different sampling loca-tions in the study area. The DO concentrations from upstream and downstream stations (termed as control) were pooled to provide a complete set of time series data. The legends for subplot c are the same as subplot a.

The concentration of DIC at the burn area during the day of fi re was not much higher than the average of the two pre-fi re samples, but increased by 0.50 mol m–3 one day post-fi re and remained elevated for the re-mainder of the study period (Fig. 3b). The average DIC concentration (4.32 mol m–3) at the burned area was signifi cantly higher than that of the upstream station and lower than the two downstream stations (paired t test, all P < 0.05). Dissolved free CO2 concentrations calculated from water temperature, pH and DIC were

Table 2. Concentration, partial pressure (pCO2) and evasion rate (CO2 fl ux) of dissolved carbon dioxide (CO2) at different locations of the study area from 18 July to 31 October 2006. N = 18 days.

Location CO2

(mol m–3)pCO2

(µatm)CO2 fl ux(mol m–2 d–1)

Upstream 1.16(0.12) 36000(3600) 0.57(0.06)Burn site 0.94(0.24) 30500(7700) 0.48(0.12)25 m Downstream 1.46(0.26) 46990(8600) 0.74(0.14)100 m Downstream 1.40(0.40) 43900(9400) 0.70(0.16)

eschweizerbartxxx

268 Binhe Gu et al.

Fig. 3. Time-series plots of (a) dissolved inorganic carbon (DIC), (b) dissolved free CO2, (c) partial pressure of CO2 (pCO2) in the sur-face water and (d) CO2 effl ux at different sites of the study area.

eschweizerbartxxx

Prescribed fi re on DIC dynamics in an Everglades wetland 269

about 30 % of the total DIC concentration and were lowest at the burned site (Table 2). Concentration of CO2 at the burn site dropped quickly from 1.0 to 0.29 mol m–3 one-half day post-fi re and then increased, while no such immediate drop in CO2 concentration was observed at the other stations (Fig. 3c).

The pCO2was extremely high in the study area with a range of 30100 to 43400 µatm (Table 2), easily exceeding the average atmospheric pCO2 (380 µatm) by two orders of magnitude. Similar to CO2, pCO2 at the burn site was lowest and showed an immediate de-crease post-fi re (Fig. 3d). Decreases in pCO2 were also found at all stations about two months after the fi re. The average pCO2 was lowest in the burned area when compared to other sites (P < 0.01). As a result of high CO2 and pCO2, the effl ux rate of CO2 from the wet-land to the atmosphere was high (Fig. 3e), with a range from 0.47 to 0.69 mole m–2 d–1 (Table 2). The average CO2 fl ux at the burn site was signifi cantly lower than at the other three stations (P < 0.01) (Fig. 3e).

Variations within the burn site

Data for water temperature, pH, DIC and TC concen-trations from the three stations within the burn site, i.e., C1, C2 and C3 (Fig. 1) on selected dates, were nearly identical (Table 3). Variations for these variables with-in the burn site were considerably smaller than in the upstream and downstream stations (Table 1).

Table 3. A comparison of the average and standard deviation (in parenthesis) of water temperature, pH, dissolved inorganic carbon (DIC) and total carbon (TC) concentrations in the three stations of the burn area for the same sample collection dates (N = 6 days).

Station Water temp.(oC)

pH(SU)

DIC(mol m–3)

TC (mol m–3)

C1 26.5(1.7) 7.41(0.13) 4.76(0.74) 10.03(1.46)C2 26.2(1.2) 7.40(0.16) 4.74(0.91) 9.63(1.59)C3 26.4(1.5) 7.37(0.18) 4.83(0.84) 9.85(1.60)

Fig. 4. Periphyton organic carbon was measured at a nutrient unim-pacted site dominated by sawgrass (Rs), the control site (H1) and the burn site (H2) from the highly nu-trient-enriched area pre- and post-prescribed burn (a). Missing bars indicated no detectable periphyton biomass. Periphyton organic car-bon was also measured at differ-ent locations in the burn plot three months post-fi re (b).

eschweizerbartxxx

270 Binhe Gu et al.

Changes in periphyton biomass

Periphyton organic carbon was high at the reference site dominated by sawgrass (Rs), but was not detect-ed at either the control or the burn site pre-fi re, and was found at the burn site 29 days post-fi re (Fig. 4a). Concentration of periphyton organic carbon decreased between day 29 and day 66 at the reference and burn sites. The lower periphyton carbon in the reference site post-fi re than pre-fi re is likely a result of hetero-genic distribution of periphyton biomass. Periphyton samples collected one month post fi re displayed the highest organic carbon content within the burn area compared to the upstream and downstream stations (Fig. 4b).

Discussion

Our results demonstrate that the prescribed fi re altered the carbon chemistry in the surface water, evidenced by the changes in TC, DIC, and CO2 concentrations, and the pCO2 and CO2 fl ux. In addition to microbial respi-ration, the incompletely burned cattail biomass present as ash and plant fragments may contribute to the in-crease in the water-column TC concentration, which included all forms of particulate and dissolved carbon in the analysis. The fi ne ash from this prescribed fi re, which contained on average 22 % carbon (Qian et al., unpubl.), could represent a portion of the new carbon observed in the surface water. Our measurements indi-cate that ash contributed 2.4 mol m–3 of TC to the sur-face water during the fi re. However, the TC concentra-tion at the burn site did not have an average increase of 2.4 mol m–3 relative to the upstream station. Together with higher TC concentrations in both downstream stations, this suggests that a portion of the particulate carbon might have fallen out of the water column and a small amount transported downstream.

The dissolved free CO2 concentrations decreased signifi cantly at the burn site one day post-fi re and grad-ually increased to the pre-fi re level by day eight before its fi nal decline. No such pattern was observed at the other stations. The immediate reduction in dissolved free CO2 post-fi re was not due to biological uptake because total DIC concentration increased, instead of decreasing, during this time period. We attribute the rapid decrease in CO2 to the increase in pH as a result of the addition of basic ions in the ash. The propor-tion of CO2 in DIC decreases as pH increases (Wetzel 2001). The daily average pH increased from 7.16 to 7.72 at one-half day post-fi re, to 7.40 one day post-

fi re. The effect of pH on CO2 concentration is dem-onstrated by calculating the CO2 concentration for the burn site and comparing it to the upstream station as a control. This calculation resulted in a 20 % increase in CO2 when the lower pH values from the upstream station were used.

Water temperature affects the solubility of CO2 (Stumm & Morgan 1981). However, our calculation indicates that the magnitude of temperature increase (1–2 °C) during the study period did not signifi cantly infl uence the CO2 concentration (P > 0.05). Another process for CO2 removal is biological carbon seques-tration. Although the initial CO2 decrease was likely due to an elevated pH by ash addition, the subsequent decrease in CO2 was likely the combined effect of high pH and periphyton photosynthesis. This mechanism is supported by increases in periphyton biomass at the burn site (Fig. 4) which was apparently triggered by the increased light, nutrients and growth space post-fi re (Miao et al., unpubl.).

The second phase of CO2 decrease at the burn site during the study period was due to the changes in pH, and biological uptake, which was also supported by the corresponding decreases in surface water pCO2 and CO2 fl ux to the atmosphere. This is consistent with fi ndings from the lacustrine systems where phy-toplankton blooms use CO2 as their carbon source, leading to lower pCO2 in the surface water (Schindler & Fee 1973, Maberly 1996). However, the appearance of periphyton at the burn site failed to change the di-rection of CO2 fl ux due to the excessive carbon storage in the impacted Everglades wetland. This wetland re-mained supersaturated with CO2 and acted as a source of CO2 to the atmosphere even though Bridgham et al. (2006) suggested that North American wetlands as a whole were small to moderate carbon sinks.

The DIC and CO2 concentrations, pCO2 and CO2 fl ux in the surface water refl ect both processes of ad-dition and removal of carbon by respiration and pho-tosynthesis. These processes were likely affected by ash addition, elevated pH, water temperature, light and nutrient availability. The effects of elevated pH, water temperature, and ash addition directly due to fi re are limited to a short-time scale (days to weeks). However, we expect that water temperature and light penetration would remain elevated due to reduced canopy prior to the full recovery of the cattail standing crop. Concur-rent studies of plant growth from the same experiment site indicated that 30 % of the biomass was recovered within a 6-month period (Sindhøj & Miao, unpubl. data). Increased light availability may be the greatest infl uencing factor that determines periphyton growth

eschweizerbartxxx

Prescribed fi re on DIC dynamics in an Everglades wetland 271

(McCormick et al. 1997). Further data collection on a longer-time scale is currently ongoing to provide an understanding of the magnitude of temporal carbon uptake and the interaction between cattail regrowth and other environmental forcing functions.

The most intriguing question regarding the effects of fi re on DIC cycling and ecosystem metabolism may be whether the ecosystem in question becomes more autotrophic or more heterotrophic. A more autotrophic system dominated by periphyton is the desired direc-tion for Everglades restoration. Findings from the present study indicate that this nutrient-enriched Ever-glades wetland became less heterotrophic post-fi re due to enhanced periphyton growth. However, periphyton growth will likely be inhibited by rapid cattail reestab-lishment. Repeat burns may be needed to provide sus-tained and suffi cient solar irradiation and growth space for periphyton. The study of DIC dynamics provides another dimension to our knowledge of ecosystem re-sponse to prescribed burns.

Acknowledgements

We thank the South Florida Water Management District for supporting the Fire Project, and the Fire Project team, Robert Johnson, Christina Stylianos, Erik Sindhøj, Dan Salembier, Cassondra Thomas, and Manuel Tapia for logistical support, Sara Neugaard for editorial suggestions and Susan Hohner for the preparation of Figure 1. ShiLi Miao was the principal in-vestigator of the Fire Project; Binhe Gu was responsible for the experimental design of this study and writing the manuscript with the assistance from ShiLi Miao, Chris Edelstein and Tho-mas Dreschel.

References

Akgerman, A. & Gainer, J. L., 1972: Predicting gas-liquid dif-fusivities. – J. Chem. Eng. Data. 17: 372–277.

Allen, E. W., Prepas, E. E., Gabos, S., Strachan, W. & Chen, W., 2003: Surface water chemistry of burned and undisturbed watersheds on the Boreal Plain: an ecoregion approach. – J. Environ. Eng. Sci. 2 (S1): S73–S86.

APHA (ed.), 1998: Standard Methods for the Examination of Water and Wastewater, American Public Health Association, American Water Works Association and Water Pollution Control Federation. – Washington, D.C., pp.1162.

Bridgham, S. D., Megonigal, J. P., Keller, J. K., Bliss, N. B. & Trettin, C. 2006: The carbon balance of North American wetlands. – Wetlands 26: 889–916.

Butler, J. N., & Cogley, D. R., 1998: Ionic equilibrium solubil-ity and pH calculation. – John Wiley & Sons., pp.559.

Carignan, R., D’Arcy, P. & Lamontagne, S., 2000: Compara-tive impacts of fi re and forest harvesting on water quality in Boreal Shield lakes. – Can. J. Fish. Aquat. Sci. 57 (S2): 105–117.

Cole, J. J., Caraco, N. F., Kling, G. W. & Kratz, T. K., 1994: Carbon dioxide supersaturation in the surface waters of lakes. – Science 265: 1568–1570.

Cole, J. J., Pace, M. L., Carpenter, S. R. & Kitchell, J. F., 2000: Persistence of net heterotrophy in lakes during nutrient addi-tion and food web manipulations. – Limnol. Oceanogr. 45: 1718–1730.

Davis, S. M., 1991: Growth, decomposition, and nutrient reten-tion of Cladium jamaicense Crantz and Typha domingensis Pers. in the Florida Everglades. – Aquat. Bot. 40: 203–224.

Duarte, C. & Prairie, Y. T., 2005: Prevalence of heterotrophy and atmospheric CO2 emissions from aquatic ecosystems. – Ecosystems 8: 862–870.

Enache, M. & Prairie, Y. T., 2000: Paleolimnological recon-struction of forest fi re induced changes in lake biogeochem-istry. – Can. J. Fish. Aquat. Sci. 57 (S2): 146–154.

Gresswell, R., 1999: Fire and aquatic ecosystems in forest-ed biomes of North America. – T. Amer. Fish. Soc. 128: 193–221.

King, R. S., Richardson, C. J., Urban D. L. & Romanowicz, E. A., 2004: Spatial dependency of vegetation-environment linkages in an anthropogenically infl uenced wetland ecosys-tem. – Ecosystems 7: 75–97.

Kling, G. W., Kipphut, G. W. & Miller, M. C., 1991: Arctic lakes and streams as gas conduits to the atmosphere: Implica-tions for tundra carbon budgets. – Science 251: 298–301.

– – – 1992: The fl ux of CO2 and CH4 from lakes and rivers in arctic Alaska. – In: Obrien, W. J. (ed.): Toolik Lake – Ecol-ogy of an Aquatic Ecosystem in Arctic Alaska. – Hydrobio-logia 240: 23–36.

Maberly, S. C., 1996: Diel, episodic and seasonal changes in pH and concentrations of inorganic carbon in a productive lake. – Freshwat. Biol. 35: 579–598.

McCormick, P. V. & O’Dell, M. B., 1996: Quantifying Periphy-ton Responses to Phosphorus in the Florida Everglades: A Synoptic-Experimental Approach.. – J. N. Amer. Benthol. Soc. 15: 450–468.

McCormick, P. V., Shuford III, R. B. E., Backus, J. G. & Kenne-dy, W. C., 1997: Spatial and seasonal patterns of periphyton biomass and productivity in the northern Everglades, Florida, U.S.A. – Hydrobiologia 362: 185–208.

Meixner, T., Poth, M., Stephens, S. L., Payne, D. & McGurk, B., 2004: Prescribed fi re, soils, and stream water chemistry in a watershed in the Lake Tahoe Basin, California. – Internat. J. Wildland Fire 13: 27–35.

Miao, S. & Carstenn, S., 2006: A new direction for large-scale experimental design and analysis. – Front. Ecol. Environ. 4: 227–227.

Minshall, G. W., Brock, J. T. & Varley, J. D., 1989: Wildfi res and Yellowstone’s stream ecosystems. – BioScience 39: 707–715.

Reddy, K. R., Wang, Y., DeBusk, W. F., Fisher, M. M. & New-man, S. 1998: Forms of soil phosphorus in selected hydro-logic units of the Florida Everglades. – Soil Sci. Soc. Amer. J. 62: 1134–47.

Schindler, D. W. & Fee, E. J., 1973: Diurnal variation of dis-solved inorganic carbon and its use in estimating primary production and CO2 invasion in Lake 227. – J. Fish. Res. Board Can. 30: 1501–1510.

Slocum, M. G., Platt, W. J., & Cooley, H. C., 2003: Effects of differences in prescribed fi re regimes on patchiness and intensity of fi res in subtropical savannas of Everglades Na-tional Park, Florida. – Restor. Ecol. 11: 91–102.

Spencer, C. N., Gabel, K. O. & Hauer, F. R., 2003: Wildfi re effects on stream food webs and nutrient dynamics in Gla-cier National Park, USA. – Forest Ecol. and Manag. 178: 141–153.

eschweizerbartxxx

272 Binhe Gu et al.

Stumm, W. & Morgan. J. (ed.), 1981: Aquatic chemistry. An introduction emphasizing chemical equilibria in natural wa-ters. – John Wiley & Sons, pp. 780.

Wan, S., Hui, D. & Luo, Y., 2001: Fire effects on nitrogen pools and dynamics in terrestrial ecosystems: A meta-analysis. – Ecol. Appl. 11: 1349–1365.

Wetzel, R. G., (ed.), 2001: Limnology: Lake and River Ecosys-tems. – Academic Press, San Diego, CA, pp. 1006.

Wu, Y., Sklar, F. H., Gopu, K. & Rutchey, K., 1996: Fire simu-lations in the Everglades landscape using parallel program-ming. – Ecol. Model. 93: 113–124.

Submitted: 21 September 2007; accepted: 14 February 2008.