Gross production exceeds gross consumption of ... - Wiley Online Library

GEOPHYSICAL RESEARCH LETTERS, VOL. 37, L18813, doi:10.1029/2010GL044341, 2010

Gross production exceeds gross consumption of methyl halides in northern California salt marshes Robert Rhew1 and Olivier Mazéas1 Received 15 June 2010; revised 23 July 2010; accepted 11 August 2010; published 28 September 2010.

[1] Coastal salt marshes are sources of CH3Br and CH3Cl to the atmosphere, but the wide range of reported emission rates illustrates the need to understand better the factors controlling net fluxes. Here we demonstrate the use of a stable isotope tracer method to separately evaluate gross production and consumption fluxes to determine their relative roles in the overall net flux. At two salt marshes in northern California, gross production exceeds gross consumption at all measured sites, leading to a large net source overall. Emission rates are within the range observed at other temperate salt marshes. By evaluating the consumption component separately, we explain how a typical salt marsh source might convert into a temporary net sink by exposing the ecosystem to uncharacteristically high concentrations of methyl halides. This circumstance may account for the reported net uptake of methyl chloride during the growing season at a coastal salt marsh in China. Citation: Rhew, R., and O. Mazéas (2010), Gross production exceeds gross consumption of methyl halides in northern California salt marshes, Geophys. Res. Lett., 37, L18813, doi:10.1029/2010GL044341.

1. Introduction [2] Methyl chloride (CH3Cl) and methyl bromide (CH3Br) have atmospheric lifetimes estimated at 1.0 and 0.7 years respectively, long enough to transport a significant fraction of these gases to the stratosphere, where the halogens can catalyze the destruction of ozone [Clerbaux and Cunnold, 2007]. Observations of tropospheric concentrations over the last decade show a notable decline in globally averaged CH3Br (coinciding with reduced agricultural emissions), but CH3Cl shows no clear trend [Clerbaux and Cunnold, 2007]. In the CH3Br budget, estimates of the sinks continue to outweigh their known sources, leading to large budget imbalances that are inconsistent with atmospheric concentration trends [Yvon‐Lewis et al., 2009]. In the CH3Cl budget, recent work has identified tropical forests (from live and decomposing plants) as potentially the largest global source [Keppler et al., 2005; Saito and Yokouchi, 2008], but major uncertainties remain owing to the limited number of field measurements. [3] Outside of the tropics, coastal salt marshes may be the largest natural terrestrial source of CH3Br and CH3Cl. Within salt marsh ecosystems, higher plants have been identified as the main sources of these compounds [Manley et al., 2006; Rhew et al., 2002] and anaerobic sediments have been identified as a sink [Oremland et al., 1994]. However, coastal salt 1 Department of Geography and Berkeley Atmospheric Sciences Center, University of California, Berkeley, California, USA.

Copyright 2010 by the American Geophysical Union. 0094‐8276/10/2010GL044341

marsh emissions show very large geographic variability, with high net emissions observed in southern California [Manley et al., 2006; Rhew et al., 2000] but relatively low net emissions observed in Scotland [Blei et al., 2010; Drewer et al., 2006] and Tasmania (Australia) [Cox et al., 2004]. The study of a salt marsh in eastern China is particularly surprising, as it concludes that some salt marshes can also act as a large sink for CH3Cl [Wang et al., 2006]. [4] This disparity of results highlights the importance of a) sampling a diverse range of salt marsh ecosystems; b) determining gross production and gross consumption fluxes in order to evaluate their contributions to net fluxes; and c) understanding how environmental factors can affect gross fluxes. In this field study of two different northern California salt marshes, we first verify the applicability of using dark chambers to measure methyl halide fluxes in the salt marsh, and then we apply a stable isotope tracer method to quantify gross production and consumption fluxes at vegetated salt marsh sites. This approach can help us interpret the relative importance of consumption and production as drivers of the overall net flux observed at these and other salt marsh sites.

2. Methods [5] Three separate field outings were conducted between 2004 and 2007 at two northern California salt marshes of the San Francisco Bay area (Figure 1): two at Bolinas Lagoon (37°55′N, 122°42′W) and one at China Camp State Park (38°0′N, 122°29′W). Bolinas Lagoon is a shallow estuary separated from the Pacific Ocean by a long spit, with salt marsh habitat on its margins. China Camp State Park is located within the San Francisco Bay estuary and encompasses over 100 acres of intertidal marsh on San Pablo Bay. [6] The two outings at Bolinas Lagoon were conducted on June 16–17, 2004 (during the growing season) and February 19–20, 2005 (during non‐growing season). In the first outing, two sites were chosen: site ‘A’ with a vegetation cover of Salicornia virginica only, and site ‘B’ with a vegetation mixture of Frankenia grandifolia, Distichlis spicata, Salicornia virginica, Limonium californicum, and Jaumea carnosa. Each site was sampled 4 times during the daytime using consecutive transparent and shrouded chambers (with one additional transparent chamber the next day). The purpose of this outing was to test the effect of sunlight on net fluxes through consecutive light/dark experiments at the same location (4 pairs). Weather conditions shifted from sunny (chambers 1–3) to foggy (chambers 4–10). [7] In the second outing, two different sites at the Bolinas Lagoon salt marsh were sampled: site ‘C’ (90% Frankenia grandifolia, 10% Distichlis spicata) was measured 4 times over 24 hours and site ‘D’ (95% Salicornia virginica 5%

L18813

1 of 5

L18813

RHEW AND MAZÉAS: SALT MARSH GROSS CH3BR AND CH3CL FLUXES

Figure 1. Map of the two study sites: Bolinas Lagoon and China Camp State Park. Distichlis spicata) was sampled twice on the second day only. The purpose of this outing was to verify that the stable isotope tracer method for quantifying gross fluxes would not significantly affect the net fluxes of CH3Br and CH3Cl. To do this, consecutive chamber experiments were run without/with the isotope addition to the chamber headspace (3 pairs of chambers). The weather during this outing shifted from rainy during the first two chambers (site C) to sunny the next morning (site D) to cloudy for the last pair of measurements (site C). [8] For the third outing, three sites were chosen from upper to lower marsh at China Camp State Park (hereafter referred to as CCSP), and nine daytime flux measurements were made on April 13, 2007, at the beginning of the growing season. Sites ‘E’ (Jaumea carnosa and Distichlis spicata), ‘F’ (100% Salicornia virginica) and ‘G’ (Spartina foliosa) were sampled alternately three times each over the course of a day. Sites E and F were located in the upland area of the salt marsh while site G was located in the lower marsh. The purpose of this last outing was to employ the stable isotope tracer method and assess the diurnal and intra‐marsh spatial variability in gross production and consumption rates. Plants were harvested for biomass determination. [9] Gas fluxes were measured using static flux chambers with the following features: two components (lid and open‐ ended base), internal mixing fans, a vent tube for pressure equilibration during sampling, and inlets for sample lines and temperature probes. Enclosure times ranged from 20–32 minutes long, and 3 or 4 air samples were drawn from the chamber into previously evacuated 1 L electropolished or 3 L fused silica lined stainless steel canisters. Chamber bases were placed in the soil 1 hour to 1 day prior to sampling. [10] Slightly different flux chamber designs and methods were employed at the two salt marshes. The chamber (152 L, 0.23 m2 basal area) used at Bolinas Lagoon consisted of an

L18813

aluminum base and a transparent Lexan lid, with the seal made via a Viton gasket on the external lip [Rhew and Abel, 2007]. Chilled water was pumped through an internal aluminum coil to maintain temperatures close to ambient. The chamber reduced surface photosynthetically active radiation by an estimated 15–50%. The chamber (188 L, 0.26 m2 basal area) used at the CCSP marsh was opaque and consisted of a base and lid made from mirror‐polished aluminum, with the seal made via a water channel surrounding the top of the base [Rhew et al., 2007]. For ‘dark’ chamber experiments, temperatures were modulated by covering the lid with a reflective shroud. These chambers did not release or consume significant amounts of methyl halides [Rhew and Abel, 2007; Rhew et al., 2007]. [11] After the first outing, a stable isotope tracer method was utilized for half of the chambers in the second outing and all of the chambers in the third outing. Briefly, a stable isotope 13CH3X (X: Cl or Br) tracer (“spiking gas”) was added to the chamber headspace, and subsequent monitoring of 12CH3X and 13CH3X concentrations allowed for the quantification of bi‐directional fluxes (see below). The added spiking gas (30–44 mL of a gas mixture consisting of 0.46 ppm 13CH3Br, 4.6 ppm 13CH3Cl, and 4.6 ppm CFC‐113 in nitrogen) yielded initial concentrations of ∼100 ppt 13 CH3Br, ∼1000 ppt 13CH3Cl, and ∼1000 ppt CFC‐113. [12] For unspiked chambers in the first and second outings, three air samples were drawn over 30 minute (t = 1, 15, 30) and 20 minute (t = 1, 10, 20) enclosure periods, respectively. For spiked chambers in the second and third outings, four air samples were drawn over 32 minute (t = 2, 12, 22, 32) and 26 minute (t = 2, 10, 18, 26) enclosure periods, respectively. Soil and flux chamber temperatures, ambient pressure, and depth of chamber placement in the soil and water were measured. [13] Two samples of air from each canister were analyzed by gas chromatography‐mass spectrometry (Agilent 6890N/ 5973). Details of the analytical system, gas standards and calibration methods are described elsewhere [Rhew et al., 2007]. For Bolinas Lagoon air samples, four parent ions were quantified in selective ion monitoring mode for both CH3Cl (mass to charge ratio m/z = 50–53) and CH3Br (m/z = 94–97). Separate injections of high concentration 13CH3X standards were performed to determine ion ratios and subsequently correct for peak overlap in flux chamber samples caused by fragmentation of the parent ions [Rhew and Abel, 2007]. 12CH3X concentrations derived from different isotopologues (e.g., m/z = 50 and 52 for 12CH3Cl) agreed to within 1% on average; 13CH3X concentrations agreed to within 5% on average. For CCSP samples, the procedure was simplified to monitor only 2 parent ions each for CH3Cl (m/z = 50 and 53) and CH3Br (m/z = 94 and 97). [14] Net fluxes were determined by applying a linear least squares fit to 12CH3X concentrations versus sampling time. Net emission rates, observed at all sites, were calculated by multiplying the linear regression slopes with the number of moles of air in the chamber. Flux errors were calculated as the standard error on the slope coefficient propagated with the uncertainty of the number of moles of air. [15] Gross fluxes were calculated using a stable isotope tracer technique [Rhew et al., 2003; Rhew and Abel, 2007] with a modified error minimization function, as described by von Fischer and Hedin [2002]. In this method, the combined

2 of 5

L18813


L18813

Figure 2. Net and gross fluxes of (a, b) CH3Cl and (c, d) CH3Br from northern California salt marsh sites (A to G). Gray circles represent net flux measurements, determined at all sites. Black circles (in Figures 2a and 2c) represent gross consumption rates; open circles (in Figures 2b and 2d) represent gross production rates. Repeat measurements are averaged; error bars represent 1 s. Note logarithmic scale for production rates (error bars omitted for clarity). Net fluxes from comparative studies are shown on the bottom of each panel. 12

CH3X and 13CH3X data were used to determine the best fit of concentration changes to a zero‐order production and first‐ order consumption [Goodwin et al., 2001] box model. Loss rates of 13CH3X represented both biological/chemical consumption as well as the physical (advective and diffusive) loss of the compounds from the chamber, the latter quantified by the loss of the inert tracer F‐113. However, when emission rates greatly exceeded consumption rates, the production of 13 CH3X (∼1% of the production of 12CH3X) could counteract or even exceed the loss rates of 13CH3X. Thus, we numerically solved for the best fit of production rate ‘P’ and uptake rate constant ‘k’(min−1) to the combined 12CH3X and 13 CH3X concentrations instead of assuming that 13CH3X changes only represent loss rates. [16] Gross consumption rates were calculated by multiplying the first order rate constants with background Northern Hemisphere concentrations (535.7 ppt for CH3Cl and 10.4 ppt for CH3Br from 1998–2001 [Simmonds et al., 2004]) and the number of moles of air in the chamber. Because these are first order uptake rates, they can be adjusted to different ambient concentrations simply by scaling the ratio of new to original concentrations. [17] For clarity, consumption (gross or net) rates were reported as negative values, while production (gross or net)

rates are reported as positive values. All fluxes are reported in units of mmol m−2 d−1 for CH3Cl and nmol m−2 d−1 for CH3Br, although enclosures were 20–32 minutes.

3. Results and Discussion 3.1. Paired Experiments [18] All sites showed net emissions of both CH3Cl and CH3Br (Figure 2). In the first outing at Bolinas Lagoon (sites A and B), the four pairs of light/dark experiments showed no significant difference for CH3Cl (light:dark = 1.02 ± 0.22, 1s), but a possible enhancement in CH3Br emissions in transparent chambers (light:dark = 1.38 ± 0.29). However, three of the four light/dark pairs in this study had net fluxes that were not significantly different from each other when considering the errors on individual fluxes, and the one pair that showed a significantly larger flux from the transparent chamber also coincided with higher internal chamber temperatures (4°C greater than the shrouded chamber) owing to a change in weather conditions. Thus, we found no consistent difference between transparent and opaque chambers in the net fluxes of methyl halides that could not be explained by higher chamber temperatures. In the second outing at Bolinas Lagoon, the 3 pairs of unspiked/spiked experiments showed

3 of 5

L18813


similar net fluxes without any consistent bias (Figure 2, sites C and D). This lack of an observed spiking effect on net fluxes has also been shown in other terrestrial ecosystems [e.g., Rhew and Abel, 2007]. Based on these results from the first two outings, the third outing utilized isotope spiking with opaque chambers at all sites. 3.2. Net Fluxes by Plant Species [19] Repeat net flux measurements at individual sites (A, B, E, F and G) demonstrated strong diurnal variability, but these fluctuations were relatively small compared to the difference in net fluxes between sites. CH3Cl net fluxes had standard deviations of 16% to 52% around the mean within a site over the course of a day (daylight hours), but varied by a factor of 200 between sites. CH3Br net fluxes had standard deviations of 17% to 43% around the mean within a site, but varied by a factor of 60 between sites. These findings are consistent with prior reports that indicate that the enclosed plant species is the most important factor in the emission rates of these compounds from salt marshes [Manley et al., 2006; Rhew et al., 2002]. [20] Frankenia grandifolia, one of the two largest emitting species in southern California salt marshes [Manley et al., 2006; Rhew et al., 2000], was represented in site C (and partly in site B) at Bolinas Lagoon. Despite being measured in the non‐growing season (February), site C showed the greatest net emissions of methyl halides in this study: 15 ± 1 mmol m−2 d−1 for CH3Cl and 1300 ± 200 nmol m−2 d−1 for CH3Br (n = 4). While large, these emission rates are on the lower end of those observed from this species during the non‐ growing season (November–March) in southern California: up to 80 mmol m−2 d−1 for CH3Cl and 11700 nmol m−2 d−1 for CH3Br [Manley et al., 2006; Rhew et al., 2000]. Although measured in the growing season (June), site B had only the second largest emissions of methyl halides: 6 ± 2 mmol m−2 d−1 for CH3Cl and 710 ± 380 nmol m−2 d−1 for CH3Br (n = 5). However, plot B was a mixture of five different species, not a monospecific stand of Frankenia. [21] Salicornia virginica, the predominant species at sites A, D, and F, was the second highest producer of methyl halides among the vegetation types included in this study. Net emissions were relatively large during the growing season: 0.9–2.1 mmol m−2 d−1 for CH3Cl and 90–160 nmol m−2 d−1 for CH3Br (sites A and F, n = 8), but reduced by 70% or more in the non‐growing season (site D, n = 2). These fluxes were also on the lower end of the range of emissions observed from S. virginica in southern California: 0.2–51 mmol m−2 d−1 for CH3Cl and 37–2800 nmol m−2 d−1 for CH3Br between April and June [Manley et al., 2006; Rhew et al., 2000]. However, the green biomass density of S. virginica at CCSP (64 gdwt m−2) was also much lower than observed in San Diego salt marshes (∼400 gdwt m−2 in November–December) [Rhew et al., 2002]). [22] Spartina foliosa (site G) yielded small net fluxes (0.22 ± 0.07 mmol m−2 d−1 for CH3Cl and 14 ± 4 nmol m−2 d−1 for CH3Br, n = 3) despite a relatively high biomass density of ∼300 gdwt m−2. These net emissions were also on the lower range of those observed from southern California salt marshes (0.1–1.4 mmol m−2 d−1 for CH3Cl and 36–120 nmol m−2 d−1 for CH3Br between March–May) [Manley et al., 2006; Rhew et al., 2000]. [23] The site E mixture of Jaumea carnosa (240 gdwt m−2) and Distichlis spicata (83 gdwt m−2) emitted methyl halides

L18813

at a slightly lower rate than Salicornia virginica, with average emissions of 0.9 ± 0.5 mmol m−2 d−1 for CH3Cl and 93 ± 17 nmol m−2 d−1 for CH3Br (n = 3). [24] Overall, the wide range of net emission rates observed in these northern California salt marshes is comparable to the range of net fluxes observed at two higher latitude salt marshes in Scotland [Blei et al., 2010; Drewer et al., 2006], but maximum rates are much lower than those observed in southern California salt marshes (Figure 2). 3.3. Gross Fluxes [25] Gross production rates dominated over gross consumption rates, as reflected in the large positive net fluxes; gross consumption represented 9 ± 10% of gross production for CH3Cl and 4 ± 3% for CH3Br (n = 12). The range of gross consumption rates was relatively narrow: 0 to −0.25 mmol m−2 d−1 for CH3Cl and −0.4 to −9.3 nmol m−2 d−1 for CH3Br. In contrast, gross production rates ranged widely: from 0.1 to 17 mmol m−2 d−1 for CH3Cl and 10 to 1600 nmol m−2 d−1 for CH3Br. These results are consistent with the idea that soil consumption of methyl halides is relatively constrained, while production rates can vary widely depending on the type of enclosed plant species. [26] A strong correlation is observed between CH3Cl and CH3Br gross production rates (r2 = 0.975, n = 12) as well as their net fluxes (r2 = 0.952, n = 25). The molar ratio of gross CH3Cl to CH3Br production rates is 12.5 ± 3.0 (excluding the flooded non‐growing season Salicornia plot D, which had a low ratio of 4.5). This ratio is comparable to the net emission ratios observed in the southern California salt marshes of San Diego (17 ± 14 [Rhew et al., 2000]) and Newport Beach (7.4 to 15, depending on enclosed plant species [Manley et al., 2006]). These ratios are all significantly larger than the CH3Cl:CH3Br net flux molar ratio of ∼4 reported from salt marshes in Tasmania [Cox et al., 2004] and Scotland [Blei et al., 2010]. Gross CH3Cl and CH3Br consumption rates are also correlated (r2 = 0.910, n = 12) and have a molar ratio of 27 ± 19. 3.4. How a Source Might Become a Sink [27] These gross flux measurements may provide insight into the surprising report of a salt marsh in the Jiangsu Province of eastern China (33°36′N, 120°36′E) that acted as an overall net sink for CH3Cl [Wang et al., 2006]. In that study, net uptake fluxes in the salt marsh ranged from −1 to −29 mmol m−2 d−1 during the growing season, far greater than the maximum gross consumption rates of −0.3 mmol m−2 d−1 observed in this study (Figure 2a). In addition, that salt marsh acted as a net source during the non‐growing season (ranging from 0.3 to 9 mmol m−2 d−1), when the marsh surface was largely frozen. Perhaps most surprising was that a non‐ vegetated, flooded mudflat site showed the largest net emissions (42 mmol m−2 d−1), larger than any source observed in this study (Figure 2b). [28] This summer sink and winter source is opposite of what one would expect if salt marsh plants act as large net sources and sediments act as sinks. Indeed, Wang et al. [2006] deduced the vegetation contribution by difference after measuring fluxes from the plots with and without vegetation, and plants were actually shown to be net producers of CH3Cl (up to +20 mmol m−2 d−1) and represented 17–68% of the net soil consumption observed. These plant emission

4 of 5

L18813


fluxes are in the same range as the gross production rates in the present study (0.1 to 17 mmol m−2 d−1 for CH3Cl). [29] One striking feature about the study in eastern China was the unusually high ambient concentrations of CH3Cl reported during the growing season: from 8679 to 58465 ppt. These concentrations are 16 to 110 times greater than typical background Northern Hemisphere concentrations of ∼540 ppt [Clerbaux and Cunnold, 2007; Simmonds et al., 2004]. Indeed, Wang et al. reported that the net CH3Cl fluxes have a strong linear correlation (r2 = 0.97) to initial CH3Cl concentrations. By scaling down their soil consumption rates to 550 ppt, they determine that the salt marsh would be a source under background atmospheric conditions. In our study, the tracer‐corrected loss rates of 13CH3Cl is assumed to be a combination of first order biological and chemical degradation losses, allowing uptake rates to be scaled to ambient concentrations. Thus the average gross consumption rate of 0.088 mmol m−2 d−1 at our sites under background concentrations would scale to 1.4 to 9.6 mmol m−2 d−1 under the range of conditions experienced in the salt marsh of eastern China, likely sufficient to turn this ecosystem into a net sink for CH3Cl.

4. Conclusion [30] Under natural conditions, coastal salt marshes act as large net sources of CH3Br and CH3Cl to the atmosphere, although emission rates vary widely from salt marshes around the world. This study demonstrates that gross consumption does occur, but that it represents