Biogenic Sulfur in the Environment - ACS Publications - American

Chapter 2

Downloaded via TUFTS UNIV on July 7, 2018 at 07:21:17 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

U.S. National Biogenic Sulfur Emissions Inventory A. Guenther, B. Lamb, and H. Westberg Laboratory for Atmospheric Research, College of Engineering, Washington State University, Pullman, WA 99164-2730 A U . S. national biogenic sulfur emissions inventory with county spatial and monthly temporal scales has been developed using temperature dependent emission algorithms and available biomass, land use and climatic data. Emissions of dimethyl sulfide (DMS), carbonyl sulfide (COS), hydrogen sulfide (H S), carbon disulfide (CS ), and dimethyl disulfide (DMDS) were estimated for natural sources which include water and soil surfaces, deciduous and coniferous leaf biomass, and agricultural crops. The best estimate of 16100 M T of sulfur per year was predicted with emission algorithms developed from emission rate data reported by Lamb et al. (1) and is a factor of 22 lower than an upper bound estimate based on data reported by Adams et al. (2). The predicted 16100 M T represents 0.13% of the 13 million M T annual emission of anthropogenic sulfur reported by Toothman et al. (3). 2

2

Recent concern for the role of sulfate in acidic deposition has intensified the need for a more accurate estimation of natural sulfur emissions in the United States. The magnitude and the spatial and temporal distribution of natural sulfur emissions must be quantified in order to be useful in efforts to predict the effectiveness of various control strategies for acid deposition. The inventory estimates described in this paper predict monthly sulfur emissions from each county in the contiguous United States. Improvements in the methodology used to calculate this natural emissions inventory may be a useful guide in the calculation of the national emissions inventories of other naturally emitted compounds which have the potential to make significant contributions to regional air quality. One of the major difficulties in estimating a national inventory is the high degree of variability in local natural sulfur emissions. This can be illustrated by a comparison of H S emission rates reported for tidal areas and salt marshes. Micrometeorological techniques have indicated an H7S flux of 80 ng • n r • min* over a tidal flat (4). Goldberg et al. (5) used a similar method to estimate emissions from a salt marsh that ranged from a minimum of 800 ng S • n r • min" in December to 3 x 10 ng S • n r • min in July. Other investigators have used dynamic enclosure methods and have reported a wide range of H2S 2

2

1

2

1

4

2

-1

0097-6156/89/0393-0014$06.00A) © 1989 American Chemical Society

Saltzman and Cooper; Biogenic Sulfur in the Environment ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

2. GUENTHER ETAL

U.S. National Biogenic Sulfur Emissions Inventory

15

emissions. Seasonal variations i n H S flux i n marine intertidal areas of Denmark have been estimated to range over several orders of magnitude (6) up to peak emissions of 6700 ng S • n r • min* (2) and 1700 ng S • n r • min"* (g) which were attributed to a tidal pumping action. Aneja et al. (9-11). and Adams et al. (2.12) have measured natural H S emissions at tidal and salt marsh sites along the U . S. Atlantic coast that varied from 10 to 1 x 10 ng S • n r • min* . Carroll (U) has reported lower rates of H S emission (0.6 to 20 ng S • n r • min* ) from a marsh. Annual variations i n the monthly H2S flux, measured by Steudler and Peterson (14) i n a salt water marsh, varied from removing 3 x 10 ng S • n r • min* to emitting 1 x 10 ng S • n r • min- into the atmosphere. Hourly fluctuations in H S ground to atmosphere flux rates for one day i n October varied from an uptake of 2 x 10 ng S • n r • min* to a release of 3 x 10 ng S • n r • m i n to he atmosphere. HoS emissions variability of over five orders of magnitude have been observea i n Florida salt water marshes where reported rates range from nearly 1 ng S • n r • min* to 10 ng S • n r • m i n (15-17). The large range of HbS emission rate estimates represented i n these various studies demonstrates trie variability of only one compound, H2S, i n one general source category, salt marshes and tidal zones. A t the present time, it is not known whether the variability in reported marshland H S emissions is due to environmental conditions or measurement problems, or a combination of both. The quantity of natural sulfur emitted to the atmosphere is dependent upon the availability of sulfur, the level of natural sulfur-reducing activity, and the environment into which the gases are released. A t present, there is a lack of information on specific mechanisms of biological sulfur release. A s a result, algorithms designed to predict natural sulfur emissions must be empirically based on analyses of correlations between observed natural sulfur emissions and environmental parameters. In order to extrapolate the available emission rate data, emission functions must be based upon parameters which are measurable and available on an appropriate scale of temporal and spatial resolution. The development of a natural sulfur emissions inventory for the contiguous U . S. is a necessary step toward understanding the natural component of acid deposition on a regional and national scale. A county length scale and a monthly time scale constitute appropriate levels of resolution for regional modeling. Emission estimates are based on the emission rate data described by Lamb et al. (1) which are limited to measurements of natural sulfur emissions in Ohio, Iowa, North Carolina, Washington and Idaho. The uncertainty associated with this inventory is carefully considered by analyses of model sensitivity. 2

2

1

2

2

5

2

1

2

2

1

3

2

1

4

2

1

2

4

4

2

1

2

2

1

-1

1

5

- 1

2

Methodology Over 300 field measurements of natural sulfur emission rates were collected for use i n this inventory. Most of these data were obtained from sites in Iowa, Ohio, and North Carolina during 1985 and have been described by Lamb et al. (1). Additional unpublished emissions data collected in Washington and Idaho have also been used. Bare M o l l i s o l , Histosol, and marshland soils were sampled in addition to surface areas with row crops (celery, carrots and onions) or natural vegetation. A b o v e ground emissions from agricultural crops (soybean, corn, and alfalfa) and forest canopies (ash, oak, maple, hickory, and pine) were also measured. These data were collected using a dynamic enclosure technique with capillary gas chromatography similar to that described by F a r w e l l et a l . (1979). Sulfur-free sweep air was blown through a polycarbonate or Teflon enclosure and a portion of the effluent air was cryotrapped at -183°C. Individual sulfur compounds were then separated using fused silica capillary gas chromatography and detected with a sulfur specific Saltzman and Cooper; Biogenic Sulfur in the Environment ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

16

BIOGENIC SULFUR IN THE ENVIRONMENT

flame photometric detector. The presence of water vapor in this system greatly reduces its ability to detect H7S and M e S H . In response to this, H S measurements were made using silver nitrate (AgN03) impregnated filters and fluorescence quenching as described by Natusch et al. (12) and Jaeschke and Herrmann (2Q). 2

Emission Rate Algorithms. In order to compile a natural emissions inventory, emission rate functions must be determined for the sources included i n the inventory. The emission rate for a specific source will vary depending upon certain environmental conditions. Analyses of sulfur emission measurements collected by Adams et al. (2) and later studies (2122) suggest that temperature plays an important role i n determining sulfur flux. While the mechanisms controlling the release of natural sulfur emissions are not well understood, field observations have demonstrated characteristic trends i n temperature-flux patterns. Sulfur emissions tend to increase logarithmically with increasing temperature for normal ambient temperatures (10°C to 35°C). A s temperatures fall below 0°C, sulfur emissions fall below the lower detection limit of the method which is approximately 1 ng S • n r • min* for surface fluxes and 4 x 10" ng S • kg* • h o u r i for vegetative emissions. The low emission rate below 0°C is not unexpected considering the minimal biological enzymatic activity which occurs at those temperatures (22). Although biological enzymatic activity normally reaches a plateau at high temperatures, this saturation point was not observed i n our studies which included numerous measurements taken at temperatures above 35°C. This is higher than any regional monthly average temperature reported for the contiguous U . S. during 1980(24). t h e emission characteristics observed in measurement studies suggest that the following form of the Michaelis-Menten equation could be used to mathematically represent the relationship between ambient temperatures and natural sulfur emissions: 2

2

log(F) =

+ k l + k

1

1

(1)

c

b

T 2

1

where F is the sulfur emission rate (ng S • n r • min* ). T is ambient temperature (°C), k and kj, are rate constants determined by a non-linear least squares fit to emission rate data, and k is a rate constant set at the lower detection limit. E q u a t i o n 1 is used to develop natural sulfur emission algorithms that predict very low emission near 0°C, increase logarithmically at intermediate temperatures, and approach a maximum emission rate at some saturation temperature which appears to be much greater than average ambient temperatures observed in temperate regions. a

c

Extrapolation to the Contiguous U . S. Rate constants for temperature dependent emissions functions for COS, C S ^ D M D S , D M S and H S have been developed for five surface categories which include wetland soils, organic soils, other soils, water, and agricultural crops (other than corn) and three vegetation categories which are deciduous and coniferous forest canopies and corn biomass. Additional natural sulfur compounds (e.g. mercaptans) are typically released at rates which are small (< 1%) relative to the total sulfur flux. The three soil categories correspond to the three sites visited by Lamb et al. (1), Goldan et al. (21), and MacTaggert and Farwell (22) in 1985 and Adams et al. 2


2.

GUENTHERETAL


(2) during 1977-78. Soil analyses indicate that these sites can be distinguished on the basis of the soil chemistry characteristics listed i n Table I. Emission rate data collected over wetlands were used to develop algorithms for predicting emissions from wetland soils which are defined as those soil map units which have a p H of less than 6, a cation exchange coefficient that is less than 25 meq • 100 g- and are designated as wet i n the U S D I National Atlas (2£). Orgamc soils are those soil map units which have more than 8% organic matter and a cation exchange coefficient that is greater than 35 meq • 100 g" . Emissions measured over Histosol soils provide the data for the organic soil emission rate algorithms and Mollisol emissions data were used to generate functions to predict emissions from all other soil map units. Using information available in the Geoecology Data Base (2£) and the U S D I National Atlas (2S), surface areas for each soil category are determined for all counties i n the contiguous United States. Water and agricultural cropland (other than corn) surface areas are also determined from the Geoecology D a t a Base. A lack of extensive measurements of sulfur emissions from water resulted in the emissions function for that category being developed from a combination of water and wetland measurements data. Although corn plants were found to significantly increase emission rates relative to bare soil or natural vegetation surface emissions, Lamb et al. Q) and Goldan et al. (21) found that other crops such as soybeans, alfalfa, oats and row crops produced only a slight increase over bare or naturally vegetated areas. The amount of additional sulfur resulting from these plants is estimated by subtracting out emissions predicted for nearby bare and naturally vegetated soils from emissions predicted for the croplands. Sulfur emissions from biomass are calculated only for the period between the last spring frost and the first fall frost. Data reported i n the Geoecology Data Base and in the U S D I National Atlas were used to make estimates of mean frost dates. The Geoecology Data Base also contains potential natural vegetation surface areas which are adjusted for current land-use practices occurring within counties. Deciduous and coniferous forest surface area estimates compiled in this data base were converted to leaf biomass estimates using the relationships listed i n Table II. Leaf biomass densities tend to be uniform throughout a vegetation association and are relatively insensitive to site-specific variables (22). C o r n is assumed to grow or increase biomass linearly through the growing season with periodic harvests. Estimates of corn biomass are based on the agricultural yield estimates reported on a county basis in the Geoecology Data Base which are then converted to total plant dry weight using relationships available i n the scientific literature (28-30). The amount of biomass present during each month increases monthly by a fraction of the total annual yield. For example, if corn requiring four months to mature is growing in a county with a seven month growing season from A p r i l to October, then monthly fractions of the total annual yield are assigned as follows: 1

1

Nov to Mar 0; Apr ^ ; May ^ ; Jun ^

; Jul | ; Aug ^ ; Sep ^ ; Oct ^

In the above example, the growing season is longer than the maturity period which results in simulated harvests occurring: in July and in October. The sum of the fractions assigned to each harvest (fj + 7) is equal to 1 which represents the total annual crop yield. Monthly emissions of COS, CS2, D M D S , D M S and H S are calculated for each of the 3071 contiguous U . S. counties. In addition to the source factors described above, mean monthly temperatures compiled in the Geoecology Data Base provide the inputs required for the estimation of natural sulfur emissions. 2


17

18


Table!. Soil Characteristics Soil Category

Site

1 (Wetlands) 2 (Organic) 3 (Other)

8

Cedar Island, N C Celleryville, O H Ames, I A

CEC* meq 100 g

PH

2.2-15.3 110 21-25

5.2-6.9 5.4 7.1-7.3

Organic Matter

%

1-65 >8 3.1-3.6

a

See Lamb et al. (1) for detailed site descriptions. •Cation Exchange Coefficient. Table II. Conversion Factors: Leaf Biomass Density Vegetation Type Coniferous Forest Deciduous Forest Sclerophyll Scrub Grasslands Tundra, Alpine fields Desert

Average Leaf Biomass Density (kg • ha-i) 6500 4500 3000 2500 1800 1000


2. GUENTHERETAL

US. National Biogenic Sulfur Emissions Inventory

19

Results and Discussion Inventory Estimates. The inventory estimates described in detail in this section are based on the emission rate data reported by Lamb et al. ( i ) . In order to obtain an uncertainty estimate, these inventory results are compared with estimates based upon data reported by Goldan et al. (21) and Adams et al. (2). The S U R E emissions data (2) referred to i n the following text have been corrected using the results of recovery efficiency studies conducted by S. Farwell (personal communication, 1985) using a G C system identical to that used during the S U R E program. For discussion purposes, the annual national flux of 16100 metric tons of sulfur (based on the data reported by Lamb et al. (!)) is broken down by source, season, compound and region in Tables III and I V . Seasons are defined as: spring - March, April, May; summer - June, July, August; autumn - September, October, November; and winter - December, January, February. Regional emissions are grouped according to the federal region designation scheme while sources are combined into eight categories: wetlands, organic soils, other soils, cropland (increase in emissions resulting from bare or naturally vegetated soils), and water surface areas, as well as corn, deciduous and coniferous biomass. Area and biomass estimates for the eight sources are listed in Table V . Above ground biomass is estimated to generate slightly more than one half of the national annual total flux. Coniferous canopies are the major biomass emitter with 30% of the total, while soil category 3 (other) is the largest surface source with 22% of all natural sulfur emissions. Although organic and wetland soils have higher per area emission rates, they each contribute only 7% of the total due to their smaller share of the total U . S. land area. Water and non-corn crops each contribute less than three percent of the total emissions on an annual basis. H^S is the dominant emission from wetlands, organic soils and water. Corn emits predominately D M S while forest canopies and non-com crops release more COS. A significant portion of the emitted C O S may be rapidly recycled into the biosphere by vegetative uptake (21). The total summer-time flux is predicted to account for 55% of the annual total while winter-time emissions are estimated to contribute only 4%. This is not unexpected considering the strong temperature dependence of the emission functions. In between these two extremes are autumn emissions (24%) and emissions during spring (18%). The higher autumn emissions are due both to increased agricultural biomass and higher temperatures. The autumn flux from agricultural crops is predicted to be more than twice the spring-time flux. The relatively large seasonal temperature variations in the northern regions (1,2,3, 5, 7 and 8) result in about 70% of their annual total being emitted i n the summer and less than 0.5% in the winter while southern regions (4, 6 and 9) are predicted to have more uniform seasonal emission rates. The relative contribution that each reduced sulfur compound makes to the total sulfur flux is often of interest because the various compounds behave differently once they enter the atmosphere. Terrestrial otogenic sulfur emissions are dominated by C O S (38%), D M S (35%) and H S (21%). Emissions of CS2 and D M D S together represent about 6% of the total. D M S emissions dominate during the summer season with 41% of the total. O n a regional basis, two-thirds of the total national flux originates in regions 4, 5, 6 and 7 which are in the southeastern and midwest portions of the United States. This encompasses about one-half of the total land area. The warmer southern regions (4, 6 and 9) generate about 90% of the inter-time emissions. By comparison, they produce about one-half of the sulfur flux on an annual basis. The regional average fluxes given in Table V I range from about 2 ng S • n r • m i n - i n region 8 to about 6 ng S • n r • m i n i n region 4, in 2

2

1

1

- 1


20


Table III. Source Summary: Surface and Vegetative Emissions Inventory Biogenic Flux in Metric Tons of Sulfur Season C O S CS2 D M D S D M S H S Total Sulfur

Source

2

Soil Category 1 Wetlands

Soil Category 2 Organic

Soil Category 3 Other

Water

Crop Biomass (non corn)

Corn Biomass

Deciduous Canopy

Coniferous Canopy

All Sources

Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual

60 260 80 10 410 8 81 13 0.1 100 150 530 190 31 890 17 79 24 3 120 13 53 39 1 100 1 4 3 .01 8 300 620 330 20 1300 750 1300 850 250 3200 1300 2900 1500 310 6100

9 23 11 3 46 2 6 2 0.1 10 52 110 60 18 240 3 8 3 1 15 2 11 5 0.1 18 2 9 5 .01 16 13 27 15 1 56 56 78 61 24 220 140 270 160 48 620

3 6 4 2 15 1 4 2 0.1 7 43 77 48 17 190 1 2 1 1 5 1 4 3 0.1 8 1 5 3 .01 10 10 19 11 0.6 41 34 45 37 16 130 96 160 110 36 400

42 170 55 8 270 4 29 7 0.1 40 260 930 340 48 1600 12 51 16 2 81 3 18 7 0.1 28 200 1800 490 0.6 2500 58 200 72 3 340 140 480 180 34 840 700 3600 1200 96 5600

63 280 85 10 440 190 560 260 0.5 1000 no 300 130 28 570 18 84 25 2 130 2 16 7 0.1 25 27 140 68 0.1 240 110 260 130 7 510 120 180 130 48 480 640 1800 840 96 3400


180 740 240 33 1200 200 680 290 1 1200 610 2000 770 140 3500 50 220 70 9 350 20 100 60 1 180 230 1900 570 1 2700 500 1100 560 31 2200 1100 2100 1300 370 4800 2900 8800 3800 590 16100

2. GUENTHERETAL


21

Table I V . Regional Summary: Surface and Vegetative Emissions Inventory Region

Season

Biogenic Flux in Metric Tons of Sulfur COS CS2 D M D S DMS H2S

1.2&3

Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual Spring Summer Autumn Winter Annual

89 260 120 4 480 310 540 330 99 1300 70 240 100 1 410 330 640 360 110 1400 61 190 79 4 340 130 420 160 1 710 200 390 240 78 900 110 260 140 22 520 1300 2900 1500 310 6100

4

5

6

7

8

9

10

National

10 22 13 1 45 28 44 29 13 110 10 28 15 0.5 53 33 56 36 16 140 9 23 12 2 46 17 41 21 0.2 79 20 33 23 11 87 12 22 15 4 52 140 270 160 48 620

7 14 9 1 31 17 24 18 9 68 7 17 10 0.5 35 22 32 23 12 88 7 14 9 1 31 13 27 15 0.2 55 14 21 15 8 59 9 14 10 3 36 96 160 110 36 400

46 300 93 2 440 180 520 210 33 940 99 850 240 0.5 1200 180 550 210 34 970 88 700 190 2 990 44 350 71 0.3 470 72 270 110 20 470 24 110 39 4 170 700 3600 1200 96 5600

130 430 90 2 760 130 300 150 33 620 69 270 110 1 450 120 300 130 30 580 84 220 110 2 420 35 130 45 0.3 210 51 110 61 20 240 24 58 31 6 120 640 1800 840 96 3400


Total Sulfur 280 1000 430 11 1800 660 1400 740 190 3000 260 1400 480 3 2100 680 1600 760 200 3200 250 1200 400 11 1800 230 970 320 2 1500 360 810 440 140 1800 180 460 230 39 900 2900 8800 3800 590 16100

22


Table V . Emission Sources Surface Area

2

Vfikm

Leaf Biomass

10 g

352

Deciduous Canopy Coniferous Canopy Corn

576

SoU Class 1 (Wetlands) Soil Class 2 (Organic) Soil Class 3 Water

7055 154

Total

7831

Cropland (Not Corn)

1281

270

12

1250 496

Based on data compiled in the Geoecology Data Base (26).

Table V I . Regional Summary: Area Flux From A l l Natural Sulfur Sources Federal Region

Emission Rate ng S • n r • min2

1,2 and 3 4 5 6 7 8 9 10

5.4 5.9 4.9 4.4 4.7 1.9 3.5 2.8

National

4.0

Based on emissions data reported by Lamb et al. (1).


:

2. GUENTHERETAL

US. National Biogenic Sulfur Emissions Inventory 1

23

1

comparison to the national average of 4 ng S • n r • min* . The spatial distribution of county averaged flux rates is shown i n Figure 1. In general, the highest flux rates are predicted to be in the agriculturally intensive midwestern corn belt region, in the New England deciduous forests, and over the wetlands and organic soils along the Atlantic and Gulf Coasts and the Mississippi river valley. The lowest flux rates are predicted for the Rocky Mountain region. Annual county average emission rates range over three orders of magnitude with a predicted minimum flux of 0.52 ng S • n r • m i n i n Denver county, Colorado and a maximum flux of 20 ng S • n r • m i n i n Monroe county, Florida. 2

2

- 1

- 1

A n t h r o p o g e n i c vs. B i o g e n i c E m i s s i o n Inventories. T h e N A P A P 1980 anthropogenic emission inventory has estimated an annual SO2 flux of 13 million M T or 3200 ng S • n r • min" (2). In comparison, terrestrial natural emissions are predicted to be 16100 M T which corresponds to 0.13% of the combined natural and anthropogenic total. Table V I I contains estimates of seasonal natural emissions and their contribution to total sulfur emissions, reported for the base year of 1980, in each state. Anthropogenic SO2 emissions are relatively constant throughout the year while biogenic sulfur emissions decrease substantially during the winter months. A s a result, the national average contribution of natural emissions ranees from 0.03% in winter to 0.28% in summer. Natural emissions are estimated to contribute less than 0.05% of the total annual emissions in Delaware, Indiana, Kentucky, New Jersey, Ohio and Pennsylvania, while over 1% of the total annual emissions are estimated to be from natural sources in Nebraska, Oregon, South Dakota and Vermont. Summer-time natural emissions contribute between 3% and 7% in Nebraska, Oregon, Idaho, South Dakota and Vermont. Natural sulfur emissions are from very diffuse sources spread over large areas while anthropogenic sources are point or concentrated area sources. The diffuse natural emissions are expected to have a negligible impact on urban sites but they should be included in r e g i o n a l budgets and may make significant c o n t r i b u t i o n s to sulfur concentrations in remote areas. 2

1

Assessment of Uncertainty Direct evaluation of the accuracy of the emission rate estimates compiled in this natural sulfur emissions inventory is difficult. Our limited understanding of n a t u r a l sulfur release mechanisms and the wide variety of possible environmental conditions to which the observed data must be extrapolated require a simplified approach to this complex process. A sensitivity analysis of the important components of the modeling procedure can be used indirectly to evaluate the uncertainty which should be associated with the model. The major components affecting the estimation of natural emissions in this inventory are source factors, temperature estimates, emission prediction algorithms and emission rate data. Source factor estimates are expected to make a relatively m i n o r contribution to the overall uncertainty (less than 10%). Surface area estimates of the various land use categories (e.g. water, marsh, urban, agriculture, and natural vegetation) compilea in the Geoecology Data Base were derived from several different data sources (26). The data are reported for different years so that it is possible to have some inconsistencies due to recent changes in land use (i.e. urbanization). In counties where the sum of the land use categories was greater than the reported county area, the amount of overlap was subtracted from the natural vegetation surface area.



2

1

2

1

2

1

Figure 1. County averaged annual natural sulfur emission rate estimates for all compounds from all sources. Empty = 0.5 to 3, ng S · m* · min* ; hatched = 3 to 6, ng S · m " · min- ; solid = 6 to 21, ng S · n r · m û r .

Η

§

S

Η

3

α

ο

Ι

5 δ ο

2. GUENTHERETAL


25

Table VII. Comparison of Anthropogenic and Biogenic Sulfur Emissions Biogenic Flux in Metric Tons of Sulfur Spring %*

State Alabama Arizona Arkansas California Colorado Connecticut Delaware Florida Georgia Idaho Illinois Indiana Iowa Kansas Kentucky Louisiana Maine Maryland Massachusetts Michigan Minnesota Mississippi Missouri Montana Nebraska Nevada New Hampshire New Jersey New Mexico New York North Carolina North Dakota Ohio Oklahoma Oregon Pennsylvania Rhode Island South Carolina South Dakota Tennessee Texas Utah Vermont Virginia Washington West Virginia Wisconsin Wyoming National

Summer % Autumni % Winter a

a

87 130 77 170 47 11 2 150 120 44 61 33 44 78 41 100 50 12 18 42 48 82 68 63 61 64 19 5 95 42 80 12 27 75 80 33 3 57 24 47 330 51 19 46 54 22 46 37

0.11 0.12 0.84 0.29 036 0.15 0.01 0.11 0.13 0.60 0.04 0.02 0.11 0.33 0.04 0.16 030 0.03 0.04 0.04 0.16 0.26 0.06 0.24 0.63 0.36 0.15 0.01 035 0.05 0.12 0.11 0.01 0.64 13 0.01 0.20 0.16 039 0.04 0.22 0.36 13 0.12 0.15 0.02 0.05 0.13

170 290 200 340 220 36 10 270 230 150 370 180 350 260 120 270 210 48 55 200 280 190 230 220 320 180 69 20 230 180 200 73 140 180 180 137 9 130 140 120 710 160 70 130 130 56 230 150

2900

0.10

8800

%

a

Annual

93 150 86 210 59 17 4 180 120 66 120 64 110 100 53 120 80 18 26 78 87 89 88 81 110 81 28 8 100 72 93 20 52 86 100 55 5 62 44 54 370 59 28 49 64 27 79 52

0.10 22 0.18 46 0.91 15 033 82 039 0.1 0.22 n 0.03 0.2 0.14 74 0.12 28 0.97 0.4 0.07 1 0.03 0.7 0.27 n 0.40 5 0.05 5 0.21 27 0.57 n 0.06 1 0.09 0.1 0.09 n 0.32 n 0.25 20 0.07 6 0.66 n 1.7 n 0.50 10 0.23 n 0.02 0.4 035 21 0.08 0.1 0.13 17 0.15 n 0.02 0.7 0.67 20 13 25 0.02 0.3 034 n 0.18 12 0.87 n 0.05 8 0.23 120 0.43 2 3.0 n 0.13 7 0.17 13 0.07 2 n 0.12 0.18 n

0.03 370 0.04 610 0.15 380 0.14 800 n 320 n 64 n 16 0.06 670 0.02 500 n 260 n 540 n 280 n 500 0.02 450 n 220 0.03 510 n 340 n 79 n 99 n 320 n 410 0.05 380 n 390 n 370 n 480 0.06 340 n 120 n 34 0.07 450 n 300 0.03 390 n 110 n 220 0.17 360 0.31 380 n 230 n 17 0.02 260 n 210 n 230 0.07 1500 n 270 n 120 0.02 240 0.06 260 0.01 110 n 360 n 2400

0.09 0.15 0.92 0.33 0.53 0.21 0.03 0.12 0.12 0.99 0.08 0.03 0.28 0.43 0.04 0.22 0.55 0.05 0.06 0.08 0.30 0.28 0.07 0.49 1.3 0.50 0.23 0.02 0.39 0.07 0.13 0.21 0.02 0.78 1.5 0.02 0.26 0.17 1.16 0.05 0.25 0.50 2.5 0.14 0.17 0.07 0.12 0.12

0.28 3800

0.13 590

0.03 16100

0.13

0.17 0.40 1.6 0.56 1.5 0.55 0.07 0.17 0.20 4.6 0.22 0.09 0.70 0.87 0.09 0.55 1.7 0.15 0.13 0.21 1.1 0.53 0.14 1.8 33 1.0 0.66 0.06 0.75 0.20 0.26 0.63 0.04 1.9 3.0 0.06 0.47 030 5.8 0.09 0.46 13 7.2 0.30 039 0.17 0.29 0.55

a

biogenic contribution to 1980 total sulfur emissions (Toothman 2). Represents negligible emission or contribution.


26


Vegetative emission source factors also depend on biomass factors to convert surface area estimates to biomass estimates. The leaf biomass factors used in this inventory (Table II) are those suggested by Zimmerman (21). A comparison of leaf biomass factors reported in the literature provides a simple illustration of the variability associated with the conversion factors used in this inventory. Monk et al. (22) reported a biomass factor of 8400 kg • ha" for trees with 14 cm diameter trunks in an oak hickory forest. Biomass factors of 5700 kg • ha* (2Q) and 5800 kg • ha- (22) have been estimated for mixed deciduous forests. Arnts et al. (24) reported a range of biomass factors between 4400 kg • ha* and 6340 kg • h a for a loblolly pine forest. The variability i n these biomass estimates is less than ±25%. The Geoecology Data Base reports arithmetic mean monthly temperatures for climatic divisions in the United States. This is not an ideal input to the inventory modeling procedure which assumes that sulfur emissions increase logarithmically with temperature. Extreme temperatures, which can generate a disproportionate amount of total emissions, are not well represented by the mean. The impact of this effect can be demonstrated by comparing a flux estimate calculated using a mean temperature to an estimate based on a range of temperatures. A mean temperature flux is based on the average monthly temperature. A n example of a weighted temperature flux is given in Equation 2 where a proportion of five fluxes predicted by five different temperatures are summed to determine the emission rate: 1

1

1

1

- 1

0.1 F + 0.2 F + 0.4 F + 0.2F + O.IF5 x

2

3

(2)

4

where is the flux resulting from the minimum temperature, F2 results from the average of minimum and mean temperatures, F3 is based on mean temperature, F is based on the average of mean and maximum temperatures, 4

Equation two has been usecf to estimate monthly and daily weighted emissions from monthly and daily weighted maximum, minimum and mean temperatures for a specific county. The daily temperatures were obtained from the U . S. Environmental Data Service (24). The total January and July monthly weighted flux estimates were within 4% of estimates predicted with only the mean monthly temperature. Daily weighted temperatures predict a July flux that is within 3% of July mean monthly estimates but the January flux is 25% greater than that predicted for January by monthly temperature statistics. The percent difference depends upon the flatness of tne flux vs. temperature curve over the temperature range 01 the month. The analysis indicates that the use of monthly average temperatures in the calculation of this natural emissions inventory is not likely to overestimate emissions but may underestimate individual emissions estimates by up to 25%. A comparison of emissions inventories which use the same emission rate data base but different modeling procedures can illustrate the sensitivity of the model to the flux prediction algorithms. Table VIII lists surface emission rates predicted by three different emission rate algorithms which have been fit to the corrected S U R E data (2) and extrapolated to cover the original S U R E area in the northeastern U . S. The Michaelis-Menten function (Equation 1) generates estimates that result in an annual average emission rate of 29 ng S • n r • min* . A segmented line emission algorithm ( S L E A ) has been used to predict an annual average emission rate of 70 ng S • n r • m i n (25) which is a factor of 2.4 greater than that predicted by the Michaelis-Menten function. The S L E A method fits a logjo least squares line to the emission rate and temperature data within the range of temperatures sampled in he study. A t lower temperatures, emissions are assumed to increase linearly from the lower 2

2

-1


1

2. GUENTHER ET AL


27

detection limit at 0 ° C to the least squares line. A t high temperatures the emission rate is constant at the level predicted for the highest measured temperature. The annual average estimate is based upon tnirty-five S L E A emission functions which were used to predict the monthly flux of five major sulfur compounds ( H S , D M S , COS, C S and D M D S ) from seven major sources (water, Histosols, Mollisols, Alfisols, Inceptisols, Ultisols and jnarshfands) on a county scale of resolution. 2

2

Table VIII. Model Sensitivity to Emission Algorithm: Annual Average Natural Sulfur Emission Rate (ng S • n r • min ) for Soil Surfaces in the Northeastern U . S. 2

Emission function

-1

Emission rate ng S • n r • min* 2

b

MM GMFIAb SLEA

1

29 15 70

C

a

Michaelis-Menten functions based on S U R E surface emissions data (2). G e o m e t r i c mean - forced intercept algorithms based on S U R E surface emissions data (2). Segmented line emission algorithms (25) based on S U R E surface emissions data (2).

b

The third emission function assumes a logarithmic increase i n sulfur emissions from the lower detection limit at 0 ° C to the geometric mean measured flux at the arithmetic mean sampling temperature. These geometric mean - forced intercept ( G M F I ) functions were developed from the S U R E data and used topredict monthly total sulfur emissions from nine soil groups in U . S. counties. The total annual average flux rate of 15 ng S • n r • m i n predicted for the S U R E region with this method is a factor of 2 lower than the flux predicted by the Michaelis-Menten functions. This comparison of these three emission functions suggests that the choice of temperature-flux function can cause a factor of two variation in emission rate estimates. In addition to the natural variability in sulfur emissions observed by individual investigators at single sites, there is also variability in emission rates observed at single sites by different investigators at different times. This variability could be due to differences in analytical techniques or changes in environmental conditions over a long period of time or combination of both. A comparison of the emission rate data recorded by Lamb et al. ( i ) and Goldan et al. (21) at one site demonstrates that the corrected S U R E H S emission rates (2) were approximately two orders of magnitude higher than comparable data recorded by Lamb et al. ( i ) and Goldan et al. (21). The emission rates of other natural sulfur compounds compiled in the corrected S U R E data base are about an order of magnitude greater than the emissions reported by Lamb et al. (1). Independent data collected simultaneously by Lamb et al. (1), Mactaggart and Farwell (22), and Goldan et al. (21) are within a factor of two. Segmented-line emission rate functions were developed from soil emission rates compiled in the Lamb et al. (1), Goldan et al. (21) and corrected Adams et al. (2) data bases and extrapolated to the national level using the inventory procedures described i n this paper. This analysis of model sensitivity to emission rate data is summarized in Table I X and demonstrates that the corrected 1978 S U R E data results in emission estimates which are a factor of 22 higher than estimates derived from the 1985 data bases. 2

- 1

2


28


Table I X . Model Sensitivity to Emission Data: National Annual Natural Sulfur Emission Rates ( M T S) Estimated for all Soil Categories Data Source

Annual Total (MT)

8

SURE NOAA WSU«

97200 5040 4210

b

Average Flux (ng S • n r • min ] 2

-1

25 1.3 1.1

fCorrected emissions data based on Adams et al. (2). "Emissions data reported by Golden et al. (21). Emissions data reported by Lamb et al. (1). b

Concisions Temperature dependent natural sulfur emission algorithms, along with biomass density data from the literature and land use and climatic data from the Geoecology Data Base, provide the basis of this national inventory of natural sulfur emission rate estimates. The annual national total of 16100 M T of naturally-emitted sulfur comprises 0.13% of the combined anthropogenic and natural sulfur flux. The biogenic contribution is as high as 7% in individual states during summer. C O S is estimated to be the major component of natural sulfur emissions followed by D M S , H S , CS2, and D M D S . Vegetative emissions are estimated to be slightly higher than those from soil surfaces. About 55% of the annual emission is predicted to occur during the high temperatures and biomass densities of summer. Three areas of uncertainty in this present inventory of natural sulfur emissions which need further work include natural variability i n complicated wetland regions, differences in emission rates in the corrected S U R E data and those reported by Lamb et al. (1) and Goldan et al. (21) for inland soil sites, and biomass emissions for which only a very limited data base exists. The current difficulty in determining the sources of variability emphasizes the need to better understand natural sulfur release mechanisms. A t present, it may be useful to consider the emission rates based on the corrected S U R E data as an upper bound to natural emissions and use the emission rates based on data described by Lamb et al. (1) as a more conservative estimate of natural sulfur emissions. However, this still leaves a factor of 22 difference between the suggested upper bound and our best current estimate. The temperature dependent algorithms used to predict natural sulfur emissions do not account for all of the variation in observed emissions. Other important environmental parameters may include, but are not limited to, tidal flushing, availability of sulfur, soil moisture, soil p H , mineral composition, ground cover, and solar radiation. A more accurate estimation of the national sulfur inventory will require a better understanding of the factors which influence natural emissions and the means to extrapolate any additional parameters which are determined to be important. 2

Acknowledgments T h i s work was supported by the N a t i o n a l O c e a n i c and A t m o s p h e r i c A d m i n i s t r a t i o n ( C o n t r a c t s N A 8 2 R A C 0 0 1 5 1 , N A 8 2 R A C 0 0 1 5 2 , and N A 8 5 R A C 0 5 1 0 5 ) as part of the National A c i d Precipitation Assessment Program.


2. GUENTHERETAL


29

Literature Cited 1. Lamb, B. K.; Westberg, H . H.; Bamesberger, W. L.; Allwine, G.; Guenther, A. B. J. Atmos. Chem. 1987,5,469-91. 2. Adams, D.; Farwell, S.; Robinson, E.; Pack, M.; Bamesberger, W. Environ. Sci. Technol. 1981,15,1493-8. 3. Toothman, D . A . ; Yates, J . C . ; Sabo, E . J . Status report on the development of the NAPAP emission inventory for the 1980 base year and summary of preliminary data. EPA-600/7-84-091, EPA Office of Research and Development, Industrial Environmental Research Laboratory, Research Triangle Park: NC, 1984. 4. Jaeschke, W. Atmos. Environ. 1978, 12, 715-22. 5. Goldberg, A . B.; Maroulis, P. J.; Wilner, L . A.; Bandy, A . R. Atmos. Environ. 1981,15,11-18. 6. Ingvorsen, K.; Jorgensen, B. B. Atmos. Environ. 1982,16,855-65. 7. Hansen, M. H.; Ingvorsen, K.; Jorgensen, B. B. Limnol. Oceanogr. 1978, 23,

68-76. 8. Jorgensen, B. B.; Okholm-Hansen, B. Atmos. Environ. 1985,19,1737-49. 9. Aneja V.; Overton, J.; Cupitt, L.; Durham, J.; Wilson W. Tellus 1979,

31,

174-8. 10. Aneja, V. P.; Overton, J. H.; Aneja, A. J. Air Pollut. Control Assoc. 1981, 31,

256-8. 11. Aneja, V. P. In Environmental Impact of Natural Emissions: Aneja, V. P., Ed.; Air Pollution Control Association: Pittsburg, PA., 1984. 12. Adams, D.; Farwell, S.; Robinson, E.; Pack, M. Biogenic Sulfur Emissions in the SURE Region, EPRI Report EA-1516, Palo Alto: CA, 1980. 13. Carroll, M. A. J. Atmos. Chem. 1986, 4, 375-82. 14. Steudler, P. A.; Peterson B. J. Atmos. Environ. 1985,19,1411-16. 15. Cooper, D. J.; de Mello, W. Z.; Cooper, W. J.; Zika, R. G.; Saltzman, E. S.; Prospero, J. M.; Savoie, D. L. Atmos. Environ. 1987a, 21, 7-12. 16. Cooper, W. J.; Cooper, D. J.; Saltzman, E . S.; de Mello, W. Z.; Savoie, D. L.; Zika, R. G.; Prospero, J. M. Atmos. Environ. 1987b, 21, 1491-5. 17. de Mello, W. Z.; Cooper, D. J.; Cooper, W. J.; Saltzman, E. S.; Zika, R. G.; Savoie, D. L.; Prospero, J. M. Atmos. Environ. 1987, 21, 987-90. 18. Farwell, S. O.; Gluck, S. J.; Bamesberger, W. L.; Schutte, T. M.; Adams, D. F. Anal. Chem. 1979, 51, 609-15. 19. Natusch, D.; Klonis, H . B.; Axelrod, H . D.; Teck, R. J.; Lodge J. P. Jr. Anal. Chem. 1972, 44, 2067-70. Saltzman and Cooper; Biogenic Sulfur in the Environment ACS Symposium Series; American Chemical Society: Washington, DC, 1989.

30


20. Jaeschke, W.; Herrmann, J. Int. J. Environ. Anal. Chem. 1981, 10, 107-20. 21. Goldan, P. D.; Kuster, W. C.; Albritton, D. L.; Fehsenfeld, F. C. J. Atmos. Chem. 1987, 5, 439-67. 22. MacTaggart, D.; Farwell, S. J. Atmos. Chem. 1987, 5, 417-38. 23. Tauber, H. The Chemistry and Technology of Enzymes: Wiley: New York, 1949. 24. Climatological data; North Carolina, 1975, National Oceanic and Atmospheric Agency, Environmental Data Service: Asheville, NC, 1980. 25. The National Atlas of the United States of America; U . S. Department of Interior Geological Survey: Washington, DC, 1970. 26. Olson, R. J. Geoecology: A County-level Environmental Data Base for the Coterminous United States. Environmental Sciences Division, Oak Ridge National Laboratory, Publication No. 1537; Oak Ridge, TN, 1980. 27. Satoo, T. In Primary Productivity and Mineral Cycling in Terrestrial Ecosystems: Symposium, 13th annual meeting of the Ecological Society of America; American Association for the Advancement of Science: New York, NY, 1967. 28. Hoffman, G. J. Trans. ASAE 1973, 16, 164-7. 29. Cooper J. P. Photosynthesis and Productivity in Different Environments; Cambridge University Press: Cambridge, 1975. 30. Whittaker, R. H.; Borman, F. H.; Likens, G . E.; Siccama, T. G . Ecol. Monogr. 1974, 44, 233-52. 31. Zimmerman, P. R. Testing for hydrocarbon emissionsfromvegetation leaf litter and aquatic surfaces, and development of a methodology for compiling biogenic emission inventories. PA Report 450/4-4-79-004, Environmental Protection Agency, 1979. 32. Monk, D. D.; Child, G. I.; Nicholson, S. A. Oikos 1970, 21, 138-41. 33. Lamb, B.; Westberg, H.; Allwine, G. J. Geophys. Res. 1985, 90, 2380-90. 34. Arnts, R. R.; Peterson, W. B.; Seila, R. L.; Gay, B. W. Jr. Atmos. Environ. 1981, 16, 2127-37. 35. Guenther, A . B. M . S. Thesis, Washington State University, Pullman, Washington, 1986. RECEIVED July 6, 1988


Biogenic Sulfur in the Environment - ACS Publications - American

Recommend Documents