Emissions of biogenic sulphur compounds from several wetland soils in Florida

A- Vd. 21. No. 7. pi. 1491-1495. hpcsdimoratRtilain. oo(-ml/x7 s3.00+0.OO ~1917l%qmcmJowmbLld. EMISSIONS OF BIOGENIC SULPHUR COMPOUNDS SEVERAL WETLAND SOILS IN FLORIDA w. J. COOPER+, D. J. FROM COOPERt, E. s. SALYZMAN, w. 2. DE MELLO, D. L. SAVOIE, R. G. ZIU and J. M. PROSPERO Romtstiel school of Marine and Atmoapbcrk sdena. University of Miami, Miami, Florida 33149. U.SA. and *Drinking Water Rcsarcb C+acr, Florida International University. Miami, Florida 33199, USA. (First reched 18 Seprtmlxr 1986 and tnjnal jii 5 lkcember 1986) AbUnct-Emission rates of tbc bio@c sulpbur pwr hydrogen sulpbidc, dimetbyl sulpbide, carbon disulpbidc and dimetbyl dirulpbide lmvc bcco maaurui from the cxposui soils of five wetland planet communitks in Florida Dimctbyl b Steudkrand Peterson, 1985; and rcfcrcnces therein) can be expected to release large quantities of dimcthyl sulphide (DMS) as a result of the plant’s high content of dimethylpropiothetin (Daczy et al., 1986), whose hydrolysis or enzymatic cka~ge leads to formation of DMS. The second, coastal mudfkts (J#rgcnscn and Okholm-Hansen. 1985, and references therein), have the tendency toward anoxia, with concomitant production of H2S via sulphate reduction. Tidal pumping mechanisms have been shown to enhance emissions of this gas in short, intense pulses (An+. 1984; J#rgcn.tcn and Okholm-Hansen, 1985; Cooper et al.. 1987. and re- ferences therein). Studies in more typical soil con- ditions generally show drastically lower emission rates of rcduccd S (Adams et al., 1981). The objective of the work reported here was to extend the studies of the emission of biogenic S gases to a variety of other wetland ecosystems. This paper reports measurements of the emission rates of the biogenic S compounds HIS, DMS. carbon disulphide t Author lo whom corrapondcnce should be addressed. (CS2) and dimcthyl disulphide (DMDS) from four different wetland soil environments in the State of Florida, complementing mcasurcmcnts reported prc- viously from a S. altermipora zone using the same analytical methods (Cooper et al.. 1987; de Mel10 et al.. 1987). The sites comprised wetland arcas with pre- dominant plant communities of Juncus roemerianus (black needle rush), Distichh spicata (spike grass). Avicennia germinans (black mangrove), Cladium jamaicense (sawgrass) and Boris nauririma (saltwort), which are found extensively in marsh environments of the SE U.S.A. SAMPLING SITES The study was conducted during the months of April, May and October 1985 and January 1986. Geographic locations are shown in Fig. 1. The J. roemerianw site was located at the St Marks National Wildlife Refuge, about 2 km inland from the coastline. The soil was an unconsolidated silica sand with an organic content of less than 2 %. and was only flooded by extremely high tides. The height of the needle rush was approximately 50-80 cm. The D. spicata site was located at the Maritt Island National Wildlife Refuge. The soil in this site was peat. to a depth of approximately 1.3 m. with the water table just below the soil surface during April and May 1985. It was above the soil on the later visits, with a salinity of less than 3 pa+ per thousand. Water kvcls were regulated as part of the Refuge water management program. The height of the emergent vegetation ~8s always < 20 cm.


INTRODUCrlON
Most models of the global sulphur cycle include biological sourca of volatile S gases from continental areas as a significant input to the atmosphere. However, extensive studies using direct measurement of the emission rates of S compounds have centered largely on two distinct coastal environments, both of which have the potential for extremely high emission rates of reduced S gases. The first, Spartina alterniporo salt marshes (Aneja. 1984;Steudkrand Peterson, 1985; and rcfcrcnces therein) can be expected to release large quantities of dimcthyl sulphide (DMS) as a result of the plant's high content of dimethylpropiothetin (Daczy et al., 1986), whose hydrolysis or enzymatic cka~ge leads to formation of DMS. The second, coastal mudfkts (J#rgcnscn and Okholm-Hansen. 1985, and references therein), have the tendency toward anoxia, with concomitant production of H2S via sulphate reduction. Tidal pumping mechanisms have been shown to enhance emissions of this gas in short, intense pulses (An+. 1984;J#rgcn.tcn and Okholm-Hansen, 1985;Cooper et al.. 1987. and references therein). Studies in more typical soil conditions generally show drastically lower emission rates of rcduccd S (Adams et al., 1981).
The objective of the work reported here was to extend the studies of the emission of biogenic S gases to a variety of other wetland ecosystems. This paper reports measurements of the emission rates of the biogenic S compounds HIS, DMS. carbon disulphide t Author lo whom corrapondcnce should be addressed.
(CS2) and dimcthyl disulphide (DMDS) from four different wetland soil environments in the State of Florida, complementing mcasurcmcnts reported prcviously from a S. altermipora zone using the same analytical methods (Cooper et al.. 1987;de Mel10 et al.. 1987). The sites comprised wetland arcas with predominant plant communities of Juncus roemerianus (black needle rush), Distichh spicata (spike grass).

SAMPLING SITES
The study was conducted during the months of April, May and October 1985 and January 1986. Geographic locations are shown in Fig. 1.
The J. roemerianw site was located at the St Marks National Wildlife Refuge, about 2 km inland from the coastline. The soil was an unconsolidated silica sand with an organic content of less than 2 %. and was only flooded by extremely high tides. The height of the needle rush was approximately 50-80 cm.
The D. spicata site was located at the Maritt Island National Wildlife Refuge. The soil in this site was peat. to a depth of approximately 1.3 m. with the water table just below the soil surface during April and May 1985. It was above the soil on the later visits, with a salinity of less than 3 pa+ per thousand. Water kvcls were regulated as part of the Refuge water management program. The height of the emergent vegetation ~8s always < 20 cm. The A. gnmiaans site was located at the Rookery Bay National Estuarine Sanctuary, approximately 50 m from the shore, and was only occasionally flooded with salt water by extreme high tides. The substrate was peat with numerous mangrove pneumatophores.
The C. jamaicense site was an inland freshwater marsh located in the Taylor Slough area of the Everglades National Park. This region was dry from December to June. The substrate was a silty calcarcous sediment (marl), with 15-20x organic matter. Standing water had a sulphate content of approximately 2.5 mg/-*. The height of the homogeneous vegetation surface was about 1 m.
The B. muritiwra site was also in the Everglades National Park, near Flamingo, approximately 20 m from the shore. The oxic calcareous substrate would not normally be fiushed by tides, and was the driest site studied. The height of the plant growth was < 20 cm.

SAMPLING AND ANALYTICAL METHODS
Sulphur gases were collected over the soil surface in a rigid polycarbonate chamber, lined with Teflon FEP to provide an inert internal surface. Scrubbed ambient air, used as sweep gas, was introduced into the chamber at a flow rate ranging from 2.2 to 3.2 / min-' (Cooper et cl., 1987). The chamber was allowed to equilibrate for at least 30 min prior to sampling.
Sites were selected within an ecosystem such that the damage to the living plants was minimized when the chamber was placed on the surface. Soil surface temperatures, inside and outside the chamber, were taken with mercury thermometers. Insolation was measured using a radiometer (The Eppley Laboratory, Inc., RI) equipped with a wide band solar spectrum filter (300800 nm). The radiometer was connected to a computer (IBM Corp., Armonk, NY) which averaged the solar flux every 10 min.
Samples were analysed for HIS, DMS, CSz and DMDS as described in Cooper et al. (1987) and de Mello et al. (1987). Calibration was performed daily. These methods allowed the determination of a minimum emission rate of 0.03, 0.01, 0.03 and 0.06 ~8 S me2 h-' for HIS, DMS, DMDS and CS2, respectively. Table I shows a summary of all emission measurements made from the five ecosystems, together with the range of soil temperatures. HIS and DMS were never below the detection limit of the methods, whereas CS2 and DMDS were often not detected. In all cases where sufficient measurements were made, a diurnal cycle in the emission rates of all compounds was found. Peak emissions were found in the early-to mid-afternoon, alightly later than the daily temperature and solar flux maxima.

Distichlis spicata site
Although this site was sampled four times over the entire study period, the soil was only exposed during the April and May visits. This site showed the highest fluxes of all compounds. up to 152 pg S(H#) m-' h-' and 23 pg S(DMS) rnbz h-'. A very strong odour of HZS was often noticed in the early morning before the breakdown of the nocturnal atmospheric inversion. Ambient concentrations of HIS at this time were measured at up to 86 pg mm3 (62 ppbv), the highest recorded during the entire study. The high emission rates of reduced biogenic S compounds probably result from a combination of sulphate availability, soil moisture and the high organic content of the soil. The diurnal cycle in emission rates is best illustrated with data from this site, the most comprehensive of which is shown in Fig. 2   The moist. sandy J. roemerionus site was the site most all&ted by seasonal temperature changes, with a range from a high of 38°C in May 1985 down to a low of 7°C in January 1986. In May, a limited data set was obtained during the middle of the day, when the DMS and HIS emission rates peaked at 6.4 and 2.8 PgS m-zh-'. respectively. In January 1986 the emissions were measured using two chambers placed within 20 cm of each other (Fig. 4, Chamber 1 and Chamber 2). The emission rates'from the two chamhers showed very little spatial variability and were considerably lower than the May visit, with maximum emissionsof0.9and 1.3pgSm-'h-' DMSand H$. respectively on the first day. The second day was significantly colder (Fig. 4). and the observed maxima in emission rates correspondingly smaller (0.2 and 0.5~gm-2h-1). May, Wig to much lower emission rates of all the S gases, a fWor of eight lower in the case of DMS, five for HIS.

DISC3JSdION
There is much uncertainty about the mechanisms amtrolling the observed die1 variation in the emission mtes of biologically gencmtcd S gases. This study was not designed for mechanistic work, but we can make some inferences W on our limited data. It is obvious from Table 1 that there is considcrabk variability in theemission rates of DMS. l-l$% DM DS and CS2 from the mtlaad soils studied. HZS and DMS emissions were generally amparable and were normally more than an order of~~it~e greater than these of C!& and DMDS, as found in previous sh~dies CoMtuctbd in  emission rates from the different sampling sites probably refkcts the availability of both sulphate and organic substrate. Most H2S is formed by dissimilatory sulphatc reducing bacteria under anoxic conditions, e.g. Demlphovibrio and Lksvlpho~~ulum (J$rgcnscn and Okholm-Hansen, 1985, and references therein), whenas DMS and DMDS originate through --_ S. ultcrmt~rta marshes. The relative magnitude of microbial degradation of methionim, and C& from cysteinc and cystine, all under waterlogged conditions (Bremner and Steele, 1978, and references therein). These conditions are certainly found at the peaty D.

spicuza 84
A. germinrms rites, where the highest emission rates were found. In addition, whilst the emission rates were not as great from the D. spicato site as those reported from a Florida Spartine alfemipora stand (Cooper et al, 1987;de Mel10 er al., 1987). it is conceivable that the Distichlis may use sulphonium compounds in its osmoregulatory system in a similar way to Sparcina species (Larher l r al.. 1977).
7'he relatively low emissions observed at the 1. roemeria~~us and B. maritiwta sites may be explained by the fact that they are low in organic matter, and that their elevation (upper marsh) is such that they are only inundated by extreme high tides, limiting the availability of sulphate.
Previous studies have shown that release of H2S can be controlled by a tidal pumping mechanism (Aneja, 1984, Jdrgensen andOkholm-Hansen, 1985;Cooper et ul., 19871 but this effect is not important in this study due to the upper marsh elevation or the impounded nature of all sites here. The only site where such a mechanism could be important is the J, roemerionus site, which may be periodically flushed by extreme high tides. Release of microbial mctabolites can be correlated with soil temperature (Adams er ol., 1981;de Mel10 et al, 1987). but this daily cycle is necessarily correlated with solar irradiation. It is therefore impossible to distinguish between the effects of increasing microbial activity and photosynthetic activity of the plant community without further investigation. In addition, where S compounds are involved in the osmoregulatory system of the plants, release can be a result of plant response to changes in the pore water salinity (de Mello et al., 1987).
Our study on the emissions of the S gases at the J. roemerionus site in January 1986 (Fig. 4) shows even though temperatures dropped to below 10°C. emissions were never completely shut off, in contrast to the findings of Hill et al. (1978). This may, however, simply reflect the lower detection limit of our method.
The major sources of error in this study are the unavoidable changes made to the environment by the use of an enclosure. The emissions of S compounds may be enhanced both by the use of a S-free sweep gas and by the daytime warming of enclosed soils relative to the surroundings.

SUMMARY AND CONCLUSIONS
Emission rates of the biogenic S gases HIS, DMS, CSx and DMDS have been measured from the exposed soils of five wetland plant communities in Florida.
The emission rates of DMS and HsS are generally IO-1OOtimeshigherthan thoseofDMDSandCS~.All the studied ecosystems showed die1 variation in the emission of rates of biogenic S gases, which is correlated loosely with the diurnal cycle in soil temperature and solar irradiation.
The measured emission rates of total reduced S from the individual ecosystems followed the trend Distichlis spicata > Acicennia gemGuns > Batis wwitima I Juwus roenkanus sx Cladium jandense.
Only the emission rates from the D. spicara site are comparable in magnitude to previous emission measurements in wetland ecosystems, most of which were conducted over SRartina alrmripora or areas of mud eats.