Sulfur isotopic composition and concentration in …Sulfur Cycle and sulfur isotopic composition of...
Transcript of Sulfur isotopic composition and concentration in …Sulfur Cycle and sulfur isotopic composition of...
Sulfur isotopic composition and concentration in throughfall
and precipitation samples from a suburban forest
By
Heyfa Khenissi
Abstract
Gross precipitation and throughfall data were collected in October and November 2009 to measure the amount of throughfall and gross precipitation. The throughfall declined from 58% to 85% as the leaf cover declined approximately from 71% to 55% in November overall. The sulfate concentrations for the throughfall and the gross precipitation in November and December 2009, however, were similar and near the detection limit of the spectrophotometer. Thus, sulfate concentration estimates were made and the values ranged from 0.78+0.75 mg/L to 2.69+1.20 mg/L for storms with precipitation ranges between 3.3 and 41.5 mm from October to February for 8 storms in total. Analytical difficulties (low recovery) precluded s-isotope analysis of all but two samples. These two samples were from the same storm event. One sample of gross
precipitation yielded δ34S=6.07+0.40 ‰ and ∆33S=-0.029 ‰+0.020 ‰, and is similar to the mean value for precipitation reported by Cooney (2005), which is 6.2+0.3‰, suggesting a mix of coal and marine sulfur sources. A second sample of throughfall sulfate yielded a similar
δ34S=5.41+0.40 ‰ and ∆33S=-0.004+0.020 ‰, suggesting no difference in isotopic composition between throughfall and gross precipitation.
Table of Contents
Introduction
Statement of the Problem
Hypotheses
Previous Work
Sulfur Cycle and sulfur isotopic composition of sulfur sources
Throughfall measurement and prediction
Determination of Dry deposition from throughfall measurements
Previous studies of δδδδ34343434
S sulfur isotope ratios, sulfate concentrations and dry
deposition in Maryland
Methods
Field methods
Measurement of throughfall and rainfall
Sampling throughfall and rainfall for sulfate and isotope analysis
Determination of leaf cover
Laboratory Methods
Isotopic sample preparation and analysis
Analysis of NADP and Precipitation data
Relationship between rainfall amount and sulfate concentration
Determination of storm source from NOAA Hysplit analysis
Results and Discussion
Seasonal variations in Throughfall
Relationship between sulfate concentration and rainfall
Sulfur Isotope Ratios
Implications
Appendix
Introduction
Statement of the problem:
Sulfur in the atmosphere is a major contributor to acid rain and acid waters. According to Menz and Seip (2004), acid rain leads to soil acidification and therefore stream water acidification. Most freshwater aquatic species require pH values between 5 and 7; pH values that are below 5 do not support aquatic acidophobic animals such as fish to thrive. In Norway, acid waters contributed to the loss of various fish populations from the 1950s until today (Hesthagen et al., 1999). Sulfur and other elements in the atmosphere are deposited onto watersheds with precipitation and as dry deposition. Collection and measurement procedures for precipitation and dry deposition are quite different, so in many cases only one of these two forms of
precipitation is measured.
In the U.S., the NADP (National Acid Deposition Program) collects information on the pH and chemistry of precipitation at locations throughout the U.S. Eastern United States has higher concentrations of sulfate in precipitation than Western United States. The high sulfur concentrations are associated with the burning of fossil fuels, primarily in the Ohio River Valley (fig. 1). Sulfate concentrations in the atmosphere have shown significant improvement since the clean-up associated with the Clean Air Act of 1972 and the re-authorization of 1990. This improvement can be seen in sulfate concentration data (fig. 2). The station shown in fig.2 is at Wye Island, MD, East of Annapolis, MD. This station has the longest record in the Washington
D.C. area. A number of additional sites have recently been added to the network, but most of
Fig. 1: Sulfate concentration contour map of the United States. Sulfate concentration of precipitation is shown in mg/L. From NADP
Fig. 2: Decrease in sulfate concentrations observed at Wye Island, MD; Data from NADP
these sites have records less than 10 years in duration. One of these sites is at the USDA in Beltsville, just north of the University of Maryland. Precipitation removes sulfur from the
atmosphere where it falls on vegetation, the soil, or urban impervious surfaces. Sulfur is also deposited as dry particulates in between storm events. The NADP, however, does not make measurements of dry deposition. Dry deposition is the amount of sulfate from the atmosphere that is deposited on the ground or the leaves of trees between precipitation events. Measuring
dry deposition + wet deposition would provide a more accurate picture of the total sulfur flux.
Butler and Likens (2004) suggest that measurements of sulfate concentrations in throughfall, which is the precipitation that falls under a forest canopy, can be used to estimate dry deposition (dry deposition= throughfall concentration- gross precipitation concentration). The sulfate concentrations in gross precipitation are lower than in throughfall, especially when it has not rained for a significant period of time. The dry deposition may be even higher in urban areas than in the forested areas measured by Butler and Likens (2004). According to a study by Chiwa et al. (2002), net throughfall in urban regions can be three times the concentration of
sulfate in throughfall in adjacent non-urban areas.
The first purpose of this project is to determine whether measurements of sulfate in throughfall can be used to evaluate the amount and isotopic composition of dry deposition. The second purpose of this project is to determine how much sulfate is contributed by dry deposition. Finally, this project’s third purpose is to determine whether isotopic methods provide a way to
constrain the relative contributions of wet and dry deposition.
Previous work
Sulfur Cycle
The study of the sulfur cycle is the study of ways that sulfur is transferred between various reservoirs on Earth. These include the deep Earth, the atmosphere, the oceans and the parts of the Earth that are in-between such as sediments, crust, weathered rock, soils and fossil fuels (Schlesinger, 1997). The focus of this work is on the part of the sulfur cycle that impacts the Baltimore-Washington corridor, namely the transferred S between the atmosphere and the
surface waters and soils in the area.
The atmospheric part of the S-cycle starts with the introduction of S into the atmosphere. There is considerable research being conducted to quantify the flux of S from various sources into the atmosphere. Erickson et al. (1990) determined that dimethylsulfide, a biological flux flux is 15 X 1012 g S/yr. On average, the annual global flux of sulfur from volcanic eruptions is 10 X 1012 g S/yr (Stoiber et al. (1987), Berresheim and Jaeschke (1983), Bates et al. (1992)). 70% of the annual global flux of S from volcanoes comes from passive volcanic lava flows and the rest comes from periodic explosive events that inject sulfur into the atmosphere (Bluth et al., 1993, Allard et al. 1994). Sulfate aerosols are also introduced with seasalt and soils (Reheis and Kihl, 1995). The global flux of S from soil dust is 8 X 1012 g S/yr, and dust storms deposit large particles locally and smaller particles at longer distances (Ivanov, 1983). Terrestrial biogenic sources of sulfur include H2S, which is emitted from freshwater wetlands, anoxic soils and COS, which is emitted from a variety of sources. It is thought to play a lesser role in the sulfur cycle, and emissions from plants are poorly understood. The total flux of biogenic sulfur from land is
less than 1 X 1012 g S/yr (Bates et al., 1992), and forest fires emit an additional 3 X 1012 g S/yr
(Andreae, 1991).
Important anthropogenic sources of sulfur have been major targets of the original Clean Air Act of 1972. Sulfur is a common constituent of oil and coal. When ores such as copper are smelted and when fossil fuels such as oil and coal are burned, SO2 is emitted, which in turn leads to air pollution and acid rain. At the present, these anthropogenic emissions are the largest source of sulfur gases in the atmosphere and they range from 50 to 100 X 1012 g S/yr globally (Möller 1984, Hameed and Dignon 1988, Spiro et al. 1992, Müller 1992), but the estimate is slightly lower today due to pollution reduction measures in the U.S. (e.g. fig. 2) and Western Europe.
The history of anthropogenic sulfur is recorded in ice cores in glaciers and ice sheets. Ice
cores from Greenland indicate that there has been a large increase in sulfate deposition from the atmosphere since the beginning of the Industrial Revolution (Herron et al. 1997, Mayewski et al. 1986, 1990). Most of the SO2 emissions are deposited locally in dry deposition and precipitation. Weathering processes and the transport of sulfur from the atmosphere both contribute sulfur from watersheds to streams, lakes, and coastal regions. The weathering of pyrite and gypsum add sulfate to rivers through groundwater flow. Human activities have raised the river transport of sulfur to more than 200 X 1012 g S/yr (Brimblecombe et al., 1989).
Sulfur isotope ratios and prediction of regional storm sources and local sources of sulfur
In Xiao and Liu (2002), light rainfalls contained δ34S sulfur isotope ratios of −4.9±2.8‰
whereas heavy rainfalls contained δ34S sulfur isotope ratios of +4.6±5.0‰. Negative δ34S values
were associated with coal and biogenic sulfur whereas positive δ34S values were associated with marine sources. Sulfur isotopes are categorized as mass-independent sulfur when they have an
amount of ∆33S that was significantly different from 0. Otherwise, the sulfur is categorized as
mass-dependent δ34S. ∆33S isotope ratios are present in atmospheric aerosols as well (Romero and Thiemens, 2003).
Measurement and prediction of throughfall
Throughfall is rainfall that falls through the forest canopy. There are two types of throughfall: direct throughfall is defined as rain that goes through the gaps in forest canopies. The second type of throughfall drips off the leaf surfaces after the storage capacity of the leaves is saturated. According to Link et al. (2003), direct throughfall and interception loss during saturated canopy conditions are not the same. When the canopy is saturated, the leaves cannot intercept anymore, and the throughfall drips off the leaves. The amount of throughfall dripping from saturated canopies varies according to the leaf area index during the seasons (Link et al.,
2003).
The canopy storage for a forest can be determined from measurements of multiple storm events that have varying amounts of precipitation for the same values of leaf cover. Linear regressions are fit to a scatterplot of throughfall vs. gross precipitation in order to determine the canopy saturation point (fig. 3). This value of canopy storage can be used to determine the
minimum amount of rain required to saturate the leaves. Rain in excess of this amount will then
result in removal of water and solutes from the leaves (Link et al. 2003).
Fig. 3 Estimate of the canopy saturation point in order to determine the minimum amount of throughfall with linear regressions fit a scatterplot of throughfall vs. gross precipitation for a given storm event (Link et al., 2003) . An estimate of canopy storage can also be obtained plotting gross precipitation versus throughfall for multiple storms (Guevara-Escobar et al., 2007) and determining the x axis intercept (which on this plot would give a smaller canopy storage
term of 2.7 mm).
Determination of dry deposition from throughfall measurements
The NADP program measures wet deposition, but not dry deposition. Global estimates of dry deposition of sulfate indicate that it may be as high as 120 X 1012 g S/yr (Andreae and Jaeschke 1992). Therefore, dry deposition cannot be ignored in sulfur budget calculations. ). Measurements from several studies indicate that sulfate concentrations in gross precipitation are lower than in throughfall, especially when it has not rained for a significant period of time. According to Butler and Likens (1994), throughfall sulfate measurements can be used to estimate dry deposition (dry deposition= throughfall concentration- gross precipitation concentration). According to a study by Chiwa et al. (2002), net throughfall sulfate concentrations in urban regions in Japan can be three times the concentration of sulfate in throughfall in nearby non-
urban areas. Therefore, dry deposition in urban areas may be higher than in the forested areas
measured by Butler and Likens (2004).
The previous work indicates that measurements of sulfate in throughfall and gross precipitation might provide a method to determine the amounts and isotopic compositions of sulfate in dry deposition. These data could then be used to determine sulfur sources of dry
deposition and determine whether it is similar to sources of precipitation.
Previous studies of δδδδ34343434S sulfur isotope ratios, sulfate concentrations and dry deposition
in Maryland
Cooney (2005) found that sulfate concentration ranges in Frederick, Maryland ranged
from 7+5 µM to 75 +5 µM. Moreover, with the exception of Hurricane Isabel (with values of
8.4+0.5‰, 8.5+0.5‰, and 12.3+0.5‰), the δ34S isotope ratios of precipitation sulfate were from
4.4+0.3‰ to 7.6+0.4‰. The δ34S isotope ratios were also relatively constant, and they did not
have any seasonal variations.
Mann (1998) worked on throughfall and dry deposition as well. However, she used a different method to obtain throughfall. She used a garbage can with a funnel to collect throughfall in a non-urban forested area near Annapolis, Maryland. Her concentration measurements of sulfate ranged 0.052 meq/L to 0.083 meq/L for gross precipitation. As for
throughfall, the concentration measurements are from 0.052 to 0.351 meq/L. She found δ34S isotope ratios between 3.5 and 4.8‰. These values indicate a mix of sulfur of marine and coal
origin, and the δ34S sulfur isotope ratios did not have any seasonal variations, but it is likely that there was more coal in the 1990s than in 2005.
Methods
Field measurements and precipitation sampling
Measurement of gross rainfall, throughfall and canopy cover
A suburban site near the University of Maryland was chosen to measure throughfall and gross precipitation (fig.,4) Both forested and open sites are available at the site. Gross precipitation and throughfall were measured in gauges (Styrofoam cups) that were pegged to the ground so that they remained upright. Throughfall was measured under the canopies of two sweetgum (Liquidambar) trees and two willow oak (Quercus phellos) trees (fig. 5). Three precipitation gauges were placed in an open area to measure gross precipitation. After each measured storm, the depth of water in each gauge was measured to determine the amount of throughfall and gross precipitation. The gauge readings were converted to point precipitation
values by dividing the volume of water in each gauge by the surface area of the gauge.
Depth of precipitation (cm) = Volume (cm3)/Surface area (cm2)
100 ft (30 m)
A
B
Fig. 4: Site map showing A) open precipitation measurement site; B) forest canopy site.
Image date: 3/1/2007; USGS air photo accessed through Google Earth
These precipitation measurements are used to determine mean and standard deviation of open and throughfall. The duration of each storm event was recorded. These data are also used to
determine the average intensity of each storm:
Intensity (cm/hr) = rainfall (cm)/storm duration (hr)
Estimation of canopy cover
Canopy cover for each of the throughfall measurement sites was determined by visual estimation. Leaf cover was photographed for each site on the same date that throughfall measurements were made. (fig. 6). The area of the sky covered by leaves and
Sweetgum Willow Oak
~ 4m
Rain gauge
Sheets
4
5
6
7
8
Figure 5: Schematic diagram showing locations of the two different types of trees, rain gauge locations, numbering system, and locations of 2.5m2 Nalgene sheets that were used to
collect rainwater.
Fig.6 Photos show canopy above each plastic sheet photographed on Oct. 14, 2009 and
on November 12, 2009. There was a decline in canopy as autumn advanced.
branches was estimated with a canopy visual estimation standard available from Cornell University (http://www.birds.cornell.edu/bfl/percentcover.pdf). Leaf cover decreases in the fall
as leaves drop from the trees. Examples of leaf cover photographs are shown in figure 6.
Throughfall decreases as deciduous trees lose their leaves. Therefore, throughfall and gross precipitation was measured at least several times per month over the study period. Measurements of throughfall that were made without leaf cover indicates the interception of
rainfall by branches, which usually turns into stemflow, not throughfall.
Collection of throughfall and rainfall samples for sulfate and isotope analysis Samples of throughfall and gross precipitation were required for sulfur and stable isotope
analysis. The concentration of sulfur is likely to be dilute, therefore, large volume (1 L) samples were required. To collect a sufficient volume of rain and throughfall samples, precipitation was sampled on Nalgene sheets. Acid-washed Nalgene sheets (2.25 m2 each) were placed next to each precipitation gauge. Three sheets were placed in the open and 5 sheets were placed under tree canopies. After the storm, the water collected on the plastic sheets was placed in acid-
washed 1L Nalgene bottles. The samples in the Nalgene bottles were then stored in the freezer.
Sulfate concentration and Isotopic composition Measurements
1. Analyses of NADP data to determine relationship between sulfate concentration and rainfall.
I acquired the Beltsville NADP data (http://nadp.sws.uiuc.edu/) and analyzed these data
to determine the range of sulfate concentration in gross precipitation samples. NADP data are available for the Beltsville station for the period 2004 until 2009. In that period, the mean sulfate concentration is 2.40 mg/L, which is 25.0 micromoles per liter. The concentration range is between 0.47 mg/L and 4.34 mg/L (standard deviation: 1.94 mg/L). In micromoles per liter, the concentration range is between 4.87 micromoles per liter and 45.2 micromoles per liter
(standard deviation: 20.1 micromoles per liter).
The NADP Beltsville data were plotted to determine whether there were consistent relationships between sulfate concentration and gross precipitation. These data show significantly higher concentrations of sulfur for small storm events (< 13 mm) than for large storm events, suggesting that rain-out of sulfur in the urban atmosphere occurs relatively
quickly during the storm event.
2. Analyses of precipitation data to determine sulfate concentrations
Examination of precipitation during a calendar year with the archived NADP and Weather Underground data from March 2009 to March 2010 will provide data that can be analyzed in order to probability distributions of sulfate concentrations. Analysis of annual precipitation data will also be used to evaluate the probability of isolated versus sequential storm events. If the storms are isolated and smaller than 13 mm, then it will be possible to
get dry deposition and wet deposition.
Throughfall will have higher sulfate concentrations than gross precipitation, and throughfall will have dry deposition in addition to wet deposition. Thus, sulfur isotope ratios will be different for both throughfall and gross precipitation. If the storms are sequential and larger than 13 mm, then dry deposition will be rained out, and throughfall
and gross precipitation concentrations and sulfur isotope ratios will be the same.
3. Preparation of sulfate standards
Based on the concentration data from the NADP data, I prepared sulfate samples that would provide accurate analyses for the expected sulfate concentration range. Harry Oduro helped me make concentration calibration standards of sodium sulfate. The range of concentrations was
from 1-80 micromoles per liter, and standards were made with the following concentrations: 1, 5, 10, 20, 40, 60 and 80 micromoles per liter, which is consistent with the estimated range of sulfate
concentrations from the NADP data.
4. Measurement of sulfate concentrations
I measured the sulfate concentrations of throughfall and gross precipitation with a spectrophotometer, which has a detection limit of 0.9 mg/L. The SulfaVer 4 Reagent Powder would react if sulfate was present. Also, the sample has to be in room temperature in order for the reagent powder to dissolve. The detection limit of the spectrophotometer is close to the
NADP’s detection limit of 0.6 mg/L.
5. Estimation of dry deposition as throughfall sulfate minus rainfall sulfate
The concentrations of sulfate in throughfall in sulfate and rainfall sulfate will be
compared to see the amount of dry deposition that occurs on leaves between storms:
Dry deposition = Gross concentration – TF concentration
Dry deposition should increase with the amount of time between storm events. The dry deposition rate could be estimated as the concentration/time interval between storm events. Different storms will be analyzed to determine the amounts of dry and wet deposition with storm from different directions and with different intervals between storms. There should
be little dry deposition when there are sequential storms.
Sample preparation for Sulfur Isotope analysis
Samples are prepared for isotopic analyses by precipitation of barium sulfate. 1 M of acidic barium chloride was introduced into the Nalgene bottles with precipitation samples in order to precipitate barium sulfate. The samples were then filtered with a filtration apparatus (fig. 5). After that, I reduced barium chloride to barium sulfide using a reduction apparatus (fig. 6). I added 25 ml of a general reduction solution to a N2 purged distillation line in order to chemically reduce BaSO4 to H2S. The H2S is captured as ZnS and then converted to Ag2S with 2 ml 1 M AgNO3, which was then rinsed with 250 ml Milli-Q and 15 ml NH4(OH). The reduction solution contains 125 ml HI, 205 ml HCl and 61 ml H2PO4. The reduction solution is boiled under a N2 atmosphere for 3 hours and then filtered to remove more impurities and
precipitates (Johnston et al, 2007).
Fig. 7 Filtration apparatus used before the reduction phase of the isotopic sample
preparation to remove impurities.
Fig. 8 Reduction apparatus used to chemically reduce BaSO4 to H2S, which in turn would
be captured as ZnS and converted to Ag2S.
The samples were then filtered and weighed. While weighing the samples, I compared the masses of the samples to the concentration measurements and estimates. For 1 mg/L of sulfate, there would be 7.5 mg of silver sulfide. For the 6 sufficiently large samples, Andrew Masterson helped fluorinate the silver sulfide with a manifold and analyze SF6 for isotope ratios with the mass spectrometer. A 10X excess of F2 was used to fluorinate the silver sulfide into SF6. The SF6 was then purified cryogenically (distilled at -110 degrees C) and chromatographically. Isotope ratios are measured in comparison to either V-CDT or the composition of paired sulfide in a specific experiment (Johnston et al, 2007). The samples were placed in tubes and burned. Then the sample was collected, transferred and purified. Then it was collected and transferred to small tubes. Then, the samples were ready to be analyzed in the mass spectrometer. A set of small samples were mixed with vanadium powder and wrapped in tin cups. Craig Hebert used the SO2 method to combust and analyze the sulfur isotope ratios. The CF-IRMS (Continuous Flow Isotope Ratio) mass spectrometer introduces a sample as a discrete pulse carried in a flow of helium to the ionization chamber. The standard gas is also introduced as a discrete pulse to the ionization chamber. Thermal decomposition in the element analyzer generates SO2 that is entrained in the helium stream. The SO2 is purified in a chemical separation process in the GC column, and it would be swept by inert carrier gases via open split through the ion source of the mass spectrometer. In the ion source of the mass spectrometer, the gas is ionized and a beam of ions is created. The ion beam is focused into a flight tube where it is separated into ion beams with discrete masses by a turning magnet and the charging ion currents on each of these ion beams are recorded with multiple collectors. (Pier de Groot, 2004). In this project, a Eurovector elemental analyzer was used to combust the sulfate samples and to separate the SO2 for 34S/32S analyses. An injector introduces timed pulses of the SO2 reference gas, and in order to determine isotope ratios, ion beam intensities are compared to background values. Small samples of 100-200 micrograms are weighed and folded into small tin cups with vanadium. Then there is a pulse of O2 of 12 ml into the catalytic combustion furnace. The frosted quartz reaction tube allows oxidation and O2 resorption. The magnesium perchlorate column removes water from the combustion products, and the GC column separates the SO2 from the
other gases. The cycle time is 210 seconds for each analysis (Grassineau et al., 2001).
Analysis of precipitation data and determination of precipitation source regions
1. Evaluation of precipitation data
Annual precipitation data were evaluated to determine the percentage of storms that occur in a year that would be useful for the measurement of dry deposition through throughfall samples. The considerations for this analysis include seasonal variations in leaf cover over the course of the seasons and the percentage of isolated, moderate-sized storms. Sources of precipitation data
include NADP archives and other nearby sites (e.g.Weather Underground stations)
The determination of storm source directions
The sulfur in the atmosphere in an urban area is likely derived from both local and distant sources. Storm trajectories can be used to evaluate source regions for storms. Therefore, I used a meterological program used to evaluate storm trajectories. NOAA’s Hysplit model (Draxler
et al., 2009) was used to evaluate storm trajectories. With Hysplit 4, I was able to identify the source of the regional storm. The trajectories provide an idea of the source region and travel path
for storms, which provide information about possible sulfur isotopic compositions.
In the Maryland Coastal Plain, storms can originate from a variety of sources. Storms that travel over the Ohio Valley can bring with them the isotopic signature of sulfate derived from the coal power plants of the Ohio Valley, which have negative values of δ34S (Xiao and Liu, 2002). A storm from the Atlantic Ocean might bring with it marine-derived sulfate, which might would be expected to have positive δ34S sulfur isotope ratios of +10‰ (Xiao and Liu, 2002). Sulfur derived from local urban sources (oil and gas) will likely also be negative (Xiao
and Liu, 2002).
Results and Discussion
Evaluation of precipitation data
Precipitation data for 2009-2010 were used to evaluate the percentage of storms in a year that could be provide estimates of dry deposition via the throughfall analysis technique. . Considerations that were used in this evaluation this included: a) presence of leaves on the trees; b) storm events greater than the minimum required to saturate leaf storage; c) isolated storm events—there is little accumulated sulfate on leaves for sequential storms (storms within 2 days of a pervious storm).
In 2009, leaf drop occurred by December 1st. Spring leaf-out occurred around April 7. Therefore, I evaluated the percentage of total storms that occurred between April and December. Storms during this period with leaves on the trees were 69 /127 or 54% of the
total storms.
Precipitation events that wash solutes from leaves must be larger than the leaf storage term, which varies over time as leaf out and leaf drop occur. Storm events that are less than the maximum leaf storage capacity will not be useful for calculating dry deposition from throughfall because a minimal amount of rainfall is required to get the dry deposition washed off from the leaves. The relationship between throughfall and precipitation data is shown in figure 9. These data were collected in October –December and the interception losses include stem and branch interception, which appears to be significant for this site. The regression equation for this trend indicates a canopy storage term of 0.078 cm (0.7 mm). This is a very low value compared with other estimates, which reflects the decline of leaf cover during the measurement period. This estimate for the late fall is unlikely to reflect leaf storage during the summer months, therefore estimates from similar studies (e.g. Link, 2007; Guevara-Escobar et al., 2007) suggest that leaf storage is approximately 1.5- 4 mm. If we use 3 mm as an estimate of canopy storage, this would eliminate 11 of the 69 storms that
occur during April to December.
TF = 0.7645 (GP) - 0.0783
R2 = 0.9955
0
1
2
3
4
5
0 2 4 6
Gross Precipitation, cm
Th
rou
gh
fall, c
m
Fig. 9: Estimate of canopy storage from the relationship between throughfall and gross precipitation. Canopy storage for the period between October and December averages 0.7
mm. The value of canopy storage for October would be higher than this average value.
Sequential storms also would not be useful for analysis of dry deposition. Between April and November, 23 of the remaining 58 storms were sequential storms, with less than 2 days between events. Therefore, only 35 of the 127 storms (27%) would be suitable for
analysis of dry deposition by the throughfall analysis.
Measurements of Throughfall and gross precipitation
Throughfall and gross precipitation were measured with the 8 rain gauges for 5 storm events in October -December. An example of these data are shown on the following diagram. For each storm event, these measurements were used to determine the average and standard deviation for gross precipitation and throughfall. The error bars shown on the diagrams represent the standard deviation for the gross precipitation and throughfall samples. These data
are also summarized in table I.
Fig. 10: Measurements of gross precipitation (open sites) and throughfall (tree sites) on October 15, 2009. Data from this sampling date will be discussed later in the isotope analysis
section.
Table I: Summary of gross precipitation and throughfall data
Date Gross
precipitation
Standard
deviation,
gross
precipitation
(1σ)
Throughfall Standard
deviation,
throughfall
(1σ)
Throughfall, %
gross precipitation
10/14/09 1.76 cm 0.06 cm 1.01 cm 0.11 cm 58%
10/15/09 2.14 cm 0.06 cm 1.42 cm 0.40 cm 66%
11/11/09-
11/12/09 10.1 cm 0.52 cm 7.17 cm 0.71 cm 71%
11/12/09 1.04 cm 0 cm 0.88 cm 0.15 cm 85%
11/13/09 1.18 cm 0.05 cm 0.85 cm 0.07 cm 72%
12/09/09 5.03 cm 0 cm 3.87 cm 0.16 cm 80%
For most of the storms there was a significant difference between gross precipitation and the throughfall. The measurement of gross precipitation in the 3 rain
gauges showed little or no variation, but the variation in throughfall was significant. The percentage of gross precipitation appearing as throughfall varied from storm to storm, but in general increased from 58% in October to 80% in December. The high value of 85% in
November appears to be associated with high winds that removed water from leaves.
Canopy Cover Estimation
I photographed the tree canopy cover of 5sites for each storm event and determined canopy cover using a visual estimation procedure using canopy visual estimation standards from Cornell University (http://www.birds.cornell.edu/bfl/percentcover.pdf). The visual estimation is approximate, even with the canopy visual estimation standards. On average, the leaf cover decreased from 70% on October 14th to 55% on November 12th. The decrease in leaf cover documents the expected leaf cover decrease during the autumn season. By December, the trees
no longer had any leaves, but branches still provided 20% interception of gross precipitation.
Table II: Leaf canopy estimates
Site Oct 14 Nov 12
#4 sweetgum 75% 50%
#7 sweetgum 70% 60%
#5 willow oak 60% 55%
#6 willow oak 80% 50%
#8 willow oak 70% 60%
Average 71% 55%
Sulfate Concentration Measurements and Estimates
Sulfur concentration of precipitation samples is available for the NADP station in Beltsville, MD. These data are presented in fig. 11 and they were used to estimate the sulfate
concentration of gross precipitation samples.
Fig. 11 The NADP sulfate concentration data in Beltsville, MD. This graph indicates that
sulfate concentration is inversely proportional to precipitation in Beltsville.
Table III: precipitation and sulfate concentrations
Storm date Precipitation mm
Throughfall Estimated sulfate (mg/l)
Measured sulfate (mg/l)
(1σ) mm
14-Oct 17.6 10.1 2.16+1.9
0
15-Oct 21.4 14.2 1+0.69
11-Nov 101 71.7 1.3+0.6
12-Nov 5.4 8.8 1+1.1 2+1.32
13-Nov 11.7 8.5 1+1
9-Dec 50 40 0.78+0.7
5 1.3+1.32
1-Feb *3.3 2.69+1.2
0
10-Feb *18 0.9+0.75
0.1
1
10
0 20 40 60 80 100
Precipitation, mm
Sulfate
, m
g/l
* Snow sample measurements from Weather Underground converted into
precipitation values.
The sulfate concentration values obtained for the November and December storms were near the 0.9 mg/L detection limit of the spectrophotometer, and the values were similar for both throughfall and gross precipitation. Thus, sulfate concentration estimates were calculated with the help of the NADP data in order to have a more accurate picture of the amounts of sulfate in the samples. The sulfate concentration estimates and the uncertainties of the estimates are based on the average and lower bounds of the sulfate concentration graph based on the NADP data (fig. 12). When the storm was the first storm of a series, the average was selected. However, when the storm was a sequential storm in a series, the lower bound was selected because I assumed that dry deposition was washed off. The sulfate concentration estimates ranged from 0.78+0.75 mg/L to 2.69 mg/L+1.20 for storms of precipitation ranges between 3.3
and 101 mm from October to February for 8 storms in total.
Table IV: Jackie Mann (1998)’s Throughfall and Gross Precipitation Samples in a non-
urban site near Annapolis, MD (SERC)
Date ppt.,
cm
(gross
ppt.)
Sulfate
meq/L
(gross
ppt.)
Sulfate
mg/L
(gross
ppt.)
Sulfate
meq/L
(TF)
Sulfate
mg/L
(TF)
dry
deposition
meq/L
dry
deposition
mg/L
7/15/1995 0.51 0.052 2.49756 0.052 2.24016 0 0
7/23/1995 0.25 0.086 4.13058 0.067 2.88636 -0.019 -0.9126
8/13/1995 1.96 0.083 3.98649 0.109 4.69572 0.026 1.24878
9/21/1995 1.75 0.046 2.20938 0.134 5.77272 0.088 4.22664
10/12/1995 - 0.041 1.96923 - - - -
3/30/1996 3.68 0.072 3.45816 0.105 4.5234 0.033 1.58499
5/19/1996 - - - 0.351 15.12108 - -
9/21/1996 - - - 0.228 9.82224 - -
When Jackie Mann estimated sulfate concentration in throughfall and gross precipitation in the 1990s, she found higher sulfate concentration values, especially for throughfall (from 1.96 mg/L to 4.13 mg/L for gross precipitation and from 2.24 mg/L to 15.12 mg/L for throughfall). Therefore, a clean-up of the atmosphere may have taken place between
the 1995 and 2009.
Fig.12 Graph based on the NADP precipitation and sulfate concentration data with averages, upper bounds and lower bounds. The upper bounds are useful for estimating sulfate concentrations in small, isolated storms. The averages are useful for estimating the first day of a storm series, and the lower bounds are useful for the sequential storms in a storm series of
several days.
Silver Sulfide mass measurements
Before weighing the samples, I needed to collect the silver sulfide by scraping the filters. The low recovery rate and the high loss rates (64%-99%) are likely due to the washing, the filtering and the scraping of the samples. This explains why there was a need to try two different
methods to obtain sulfur isotope values, the SF6 method and the SO2 method.
Table V: silver sulfide mass measurements
Sample Mass (mg) (2 Anticipated predicted (mg)
% loss
10/15/09 #5 willow oak 2.7+0.2 7.5 64%
11/12/09 #3 open 0.2+0.2 7.5 97%
12/9/09 #1 open 0.1+0.2 5.85 98%
10/15/09 #6 willow oak 0.9+0.2 7.5 88%
10/15/09 #7 sweetgum 0.4+0.2 7.5 95%
11/12/09 #2 open 0.1+0.2 7.5 99%
11/12/09 #8 willow oak 0.1+0.2 7.5 99%
12/9/09 #8 willow oak 0.1+0.2 5.85 98%
12/9/09 #3 open 0.4+0.2 5.85 93%
11/12/09 #6 willow oak 0.1+0.2 7.5 99%
12/9/09 #7 sweetgum 0.1+0.2 5.85 98%
10/15/09 #2 open 19.8+0.2* 7.5
10/15/09 #4 sweetgum 0.6+0.2 7.5 92%
10/15/09 #3 open 18.6+0.2 * 7.5
11/12/09 #5 willow oak 3+0.2 7.5 60%
11/12/09 #4 sweetgum 0.3+0.2 7.5 96%
11/12/09 #5 willow oak 0.5+0.2 7.5 93%
10/15/09 #1 open 1.2+0.2 7.5 84%
10/15/09 #8 willow oak 1.5+0.2 7.5 80%
*: There was white precipitate in the two samples. So they were reduced again, and they were centrifuged 6 times in order to get a smaller and more accurate result. Thus, the mass is
actually smaller than what I initially weighed.
Sulfur Isotope Measurements
The sulfur isotope samples were difficult to analyze due to small sample size, and loss of samples. Three of 6 of the samples that were analyzed with the SF6 method did not yield enough sulfur to analyze. Of the remaining three samples that did yield enough to analyze, one was lost accidently. Two samples were successfully analyzed; one throughfall and one gross
precipitation sample from October 15th, 2009. Both samples have similar δ34S values with estimated uncertainties. (Table IV). The sulfur isotope values of the two samples indicate that the throughfall and the gross precipitation have the same regional sources of sulfur during this storm. Since this was a sequential storm where the precipitation was 21.4 mm and well over the possible maximum (13 mm based on the NADP data) to get dry deposition, the dry deposition was probably washed off and only the wet deposition of sulfur remained.
The δ34S values indicate a mixing of sulfur sources (urban+ coal and Atlantic marine sulfate mixed would provide values less than +10 (Atlantic); but significantly larger than the negative values expected for coal or urban sources alone. The isotopic values of sulfate can be compared with two previous studies. Mann, 1998, measured sulfate in precipitation and throughfall samples at SERC, approximately 35 miles to the east of the study site in this report (Table IV). Mann also had problems with isotopic analysis of these dilute samples and conducted the analysis using TIMS techniques that required smaller sample volumes (Mann, 1998). The samples that Mann collected were from 1995-1996, when atmospheric sulfur concentrations were higher than at the present. Mann, 1998, found higher values of sulfur in throughfall than in precipitation and both values were higher than observed in 2010. Mann, 1998 did not see seasonal variations in sulfur isotopic composition, but the slightly lower isotopic values might reflect higher contributions of sulfur from coal. In another study of isotopic values of gross precipitation, Cooney, 2005 found that the
δ34S values do not vary seasonally (Cooney, 2005) and the isotopic values were similar to those measured in this study. This suggestions that sulfur compositions are well-mixed and stable. The NOAA Hysplit trajectories also indicate a mix of regional storm sources for the October 14th, 2009 and October 15th, 2009 storms. While the October 14th, 2009 storm came from the Ohio Valley, the October 15th storm came from the Atlantic (fig. 13 and 14). Thus, samples obtained on October 15th should reflect a combination of coal-derived sulfur isotopes from the Ohio Valley and marine-derived sulfur isotopes from the Atlantic Ocean.
∆33S isotope ratios were not significantly different from 0. Therefore, the amount of mass-independent sulfur is not significant in the atmospheric aerosols of the region.
Table VI: Sulfur Isotope Measurements
Sample δ34S ∆33S
10-15-09 #5 willow oak TF 5.41+0.40 ‰ (2σ) -0.004+0.020 ‰ (2σ) 10-15-09 #1 open 6.07+0.40 ‰ (2σ) -0.029+0.020 ‰ (2σ) *07/23/95 open 3.5
*03/30/96 open 4.8
**07/23/95 TF 3.9
*03/30/96 4
* Mann, 1998
Fig. 13 NOAA Hysplit trajectory model of the October 14th, 2009 storm. This indicates that the
storm came from the Ohio Valley. Therefore, the sulfate came from coal.
Fig. 14 NOAA Hysplit trajectory model of the October 15th, 2009 storm. This indicates that the storm came from the Atlantic Ocean. Therefore, the sulfate came from a marine source. There was a mix of regional storm sources from the October 14th and the October 15th, 2009 storms. So this explains why the sulfur isotope values indicate a mix of coal and marine sulfur sources.
Implications
The values of δ34S isotopes were between 5.4 and 6, which indicate possible mixing of coal and urban-derived sources with marine sulfur. The sulfur isotope values were similar to those previously measured by Cooney, which suggests considerable stability in the isotopic value, which perhaps reflects well-mixed sources during precipitation events. Comparison with the earlier analyses by Mann, 1998 suggests that the isotopic composition of rainfall has shifted upward a 1-2 ‰, which may reflect clean-up of coal-derived sources. This project was designed to evaluate the amount and source of sulfur in dry deposition. Analysis of precipitation events in 2009-2010 suggest that only about 17% of the precipitation events can be used to evaluate dry deposition. Although this technique has been successful in other regions, this may not be an effective way to quantify dry deposition in this region.
Nonetheless, δ34S isotopes can still be used to quantify the source of sulfur in a storm, and dry deposition still needs to be taken into account in quantifying the total deposition of sulfur. Collecting the rainfall was not very time-consuming. However, the sulfur isotope preparation required to conduct a sulfur isotope analysis takes a considerable amount of time. Thus, it is best to focus on one storm instead of trying to estimate the sulfate concentrations and sulfur isotope ratios of many storms. Moreover, there is a need to ensure that the collected storm is small and isolated. If the predicted storm happens to be sequential or massive, it can be quickly discarded before even measuring the concentrations in order to be able to focus fully on a future storm that may include dry deposition. Therefore, further tests should be done to see if dry deposition and wet deposition are similar or different in other precipitation events. I also suggest doing further research on the leaf storage and the canopy saturation point in order to measure the minimal amount of throughfall required to remove solutes from the leaves of the trees in this region. It is better to study leaf storage during the summer because the leaf storage capacity would be constant whereas during the fall, the leaf storage capacity declines.
Appendix
Fig. 1 Precipitation values from March 2009 to March 2010, including samples from the study period. NADP and Weather Underground archived data were used to estimate the probability
of finding an isolated small storm from April to November.
October 14, 2009
00.10.20.30.40.50.60.70.80.9
1
open open open tree
edge
willow
oak
willow
oak
willow
oak
sweet
gum
pre
cip
itati
on
, cm
Fig. 2 Measurements of gross precipitation (open sites) and throughfall (tree sites) on October 14,
2009.
Fig. 3 Measurements of gross precipitation (open sites) and throughfall (tree sites) on November
11-12, 2009.
Fig. 4 Measurements of gross precipitation (open sites) and throughfall (tree sites) on November 12, 2009.
Fig. 5 Measurements of gross precipitation (open sites) and throughfall (tree sites) on November
13, 2009.
References
Allard, P., J. Carbonnelle, N. Métrich, H. Loyer and P. Zettwoog. 1994 Sulphur output and
magma degassing budget of Stromboli volcano. Nature 368: 326-330.
Andreae, T.W. G.A. Cutter, N. Hussain, J. Radford Knoery and M.P. Andreae, 1991. Hydrogen sulfide and radon in and over the western North Atlantic ocean. Journal of
Geophysical Research 96: 18753-18760.
Andreae, M.O. and W.A. Jaeschke. 1992. Exchange of sulphur between biosphere and atmosphere over temperate and tropical regions, pp. 27-61. In R.W. Howarth, J.W.B. Stewart, and M.V Ivanov (eds.), Sulphur Cycling on the Continents. Wiley, New
York.
Bates, T.S., R.J. Charlson and R.H. Gammon. 1987. Evidence for the climatic role of marine
biogenic sulphur. Nature 329: 319-321.
Bates, T.S., B.K. Lamb, A. Guenther, J. Dignon and R.E. Stoiber. 1992. Sulfur emissions to the
atmosphere from natural sources. Journal of Atmospheric Chemistry 14: 315-337.
Berresheim, H. and W. Jaeschke. 1983. The contribution of volcanoes to the global atmospheric sulfur budget. Journal of Geophysical Resarch 88: 3732-3740.
Bluth, G.J.S., C.C. Schnetzler, A.J. Krueger and L.S. Walter. 1993. The contribution of explosive volcanism to global atmospheric sulphur dioxide contributions. Nature 366: 327-329.
Brimblecombe, P., C. Hammer, H. Rodhe, A. Ryaboshapko, and C.F. Boutron. 1989. Human influence on the sulphur cycle. Pp. 77-121. Evolution of the Global Biogeochemical
Sulphur Cycle. Wiley, New York.
Butler, T.J. and Likens, G.E. A Direct Comparison of Throughfall plus Stemflow to estimates of dry and total deposition for sulfur and nitrogen. Atmospheric Environment Volume 29,
Issue 11, June 1995, pages 1253-1265.
Cooney, K.S., 2005. A Stable Isotopic Investigation of Precipitation Nitrate. Master's of Science
thesis in Geology. University of Maryland, 173 pp.
De Groot, Pier. Handbook of Stable Isotope Analytical Techniques, 2004.
Draxler, Roland R, Wang, Y.Q. and Zhang, X.Y (2009). TrajStat: GIS-based software that uses various trajectory statistical analysis methods to identify potential sources from long-term air pollution measurement data. Environmental Modelling & Software, Volume
24, Issue 8, August 2009, Pages 938-939.
Grassineau, N.V., Mattey, D.P. and Lowry, D. (2001) Sulfur isotope analysis of sulfide and sulfate minerals by continuous flow-isotope ratio mass spectrometry. Anal. Chem. 73, 220-225.
Guevara-Escobar, A, González-Sosa, E, Véliz-Chávez, C, Ventura-Ramos, E, Ramos-Salinas,
M. (2007) Rainfall interception and distribution patterns of gross precipitation around an
isolated Ficus benjamina tree in an urban area. Journal of Hydrology, volume 333, issues 2-4, 15 February 2007, pages 532-541.
Hackley, K.C. and T.F. Anderson. 1986. Sulfur isotopic variations in low-sulfur coals from the Rocky Mountain region. Geochemica et Cosmochimica Acta 50: 1703-1713.
Hameed. S. J. Dignon. 1988. Changes in the geographical distributions of global emissions of NOx and SOx from fossil fuel combustion between 1966 and 1980. Atmospheric
Environment 22: 441-449.
Herron, M.M., C.C. Langway, H.W. Weiss and J.H. Cragin. 1977. Atmospheric trace metals and
sulfate in the Greenland ice sheet. Geochemica and Cosmochimica Acta 41: 915-920.
Hesthagen, T., Sevaldrud, I.H. and Berger, H.M., 1999. Assessment of damage to fish population in Norwegian lakes due to acidification. Ambio 28, pp. 112–117.
Ivanov, M.V. 1983. Major fluxes of the global biogeochemical cycle of sulphur. Pp. 449-463. The
Global Biogeochemical Sulphur Cycle. Wiley, New York.
J.F. Kasting. "The Rise of Atmospheric Oxygen" (Perspective) Science 293: 819-820 (2001).
Link, Timothy E., Unsworth, Mike, Marks, Danny. The dynamics of rainfall interception by a seasonal temperature rainforest. 2003-2004. Agricultural and Forest Meteorology 124
(2004), 171-191.
Mann, J.L., 1998, Effects of Hydrologic flow paths on the relative amounts of atmospheric and aquifer-derived sulfur in surface waters of a Maryland Coastal Plain catchment.: Univ.
of Maryland M.S. Thesis 212 pp.
Masaaki Chiwa, Do Hoon Kim and Hiroshi Sakugawa. Rainfall, Stemflow and Throughfall Chemistry at Urban and Mountain-facing at Mt. Gokurakuji, Hiroshima, Western Japan.
Atmospheric Environment Volume 37, Issue 3, January 2003, Pages 327-337.
Mayewski, P.A., W.B. Lyons, M.J. Spencer, M. Twickler, W. Dansgaar, B. Koci, C.I. Davidson and R.E. Honrath. 1986. Sulfate and nitrate concentrations from a south Greenland ice
core. Science 232: 975-977.
Mayewski, P.A., W.B. Lyons, M.J. Spencer, M. Twickler, W. Dansgaar, B. Koci, C.I. Davidson and R.E. honrath. 1990. An ice-core record of atmospheric response to anthropogenic
sulphate and nitrate Nature 346: 554-556.
Menz , Fredric C, Seip, Hans M. Acid rain in Europe and the United States: an update Environmental Science & Policy, Volume 7, Issue 4, August 2004, Pages 253-265
Möller D. 1984. On the global natural sulphur emission. Atmospheric Environment 18: 29-39.
Müller, J.F 1992. Geographic distribution and seasonal variation of surface emissions and deposition velocities of volcanic trace gases. Journal of Geophysical Research 97:
3787-3804.
Reheis M.C. and R. Kihl, 1995. Dust deposition in southern Nevada and California, 1984-1989: Relations to climate, source area and source lithology. Journal of Geophysical Research
100: 8893-8918.
Romero, Alvaro B. and Thiemens, Mark H (2003). Mass-independent sulfur isotopic compositions in present-day sulfate aerosols. Journal of Geophysical Research, vol. 108,
7 pp.
Schlesinger, William H. Biogeochemistry: An Analysis of Global Change, 1997 (2nd edition).
Spiro, P.A., D.J. Jacob and J.A. Logan, 1992. Global inventory of sulfur emissions with 10 X 10
resolution. Journal of Geophysical Research 97: 6023-6036.
Stoiber, R.E., S.N. Williams and B. Huebert. 1987. Annual contribution of sulfur dioxide to the
atmosphere by volcanoes. Journal of Volcanology and Geothermal Research 33: 1-8
Xiao, H.-Y and Liu, C-Q. Sources of nitrogen and sulfur in wet deposition in Guiyang, Southwest China. Atmospheric Environment, Volume 36, Issue 33, November 2002,
pages 5121-5130.
Visual percentages of percent cover. http://www.birds.cornell.edu/bfl/percentcover.pdf
Weather Underground Weather Archive for Washington DC. http://www.wunderground.com/history/airport/KDCA/2010/4/16/DailyHistory.html?req_city=Washington&req_state=DC&req_statename=District+of+Columbia
NADP website: http://nadp.sws.uiuc.edu/
NOAA Hysplit trajectory modeling: http://ready.arl.noaa.gov/HYSPLIT_traj.php