Direct Atmospheric Inputs versus Runoff Fluxes of ... - ACS Publications


Direct Atmospheric Inputs versus Runoff Fluxes of...

0 downloads 78 Views 111KB Size

Environ. Sci. Technol. 2000, 34, 4058-4063

Direct Atmospheric Inputs versus Runoff Fluxes of Mercury to the Lower Everglades and Florida Bay W O O - J U N K A N G , †,‡ J O H N H . T R E F R Y , * ,† TERRY A. NELSEN,§ AND HAROLD R. WANLESS# Division of Marine and Environmental Systems, Florida Institute of Technology, 150 West University Boulevard, Melbourne, Florida 32901, NOAA/AOML/OCD, 4301 Rickenbacker Causeway, Miami, Florida 33149, and Department of Geological Sciences, University of Miami, Miami, Florida 33124

Age-dated sediments from the lower Everglades and Florida Bay provide a record of inputs of excess Hg from direct atmospheric input versus runoff. Direct atmospheric fluxes of excess Hg to sediments in the lower Everglades and Florida Bay, calculated using a mass balance model for excess 210Pb, currently average 24 ( 9 µg m-2 yr-1 and are comparable with recent results from bulk atmospheric deposition. In contrast, present-day runoff fluxes of excess Hg to area sediments are variable, ranging from about 4-160 µg m-2 yr-1. The runoff flux now carries 60-80% of the total flux of excess Hg to the sediments in areas near river sloughs but less than 20% of the total flux of excess Hg in more remote areas of Florida Bay. These results show the greater importance of runoff relative to direct atmospheric deposition for Hg inputs to many areas of the lower Everglades and immediately adjacent Florida Bay. Thus, the choice of future water management strategies can play an important role in controlling Hg inputs to the lower Everglades and portions of Florida Bay.

Introduction Mercury is transported to remote and urbanized aquatic environments via the atmosphere and runoff. The importance of atmospheric transport of Hg is well documented in sediment cores from remote lakes in North America and Europe (1-6). Such records have helped identify Hg as a global pollutant (5-7). Other studies also emphasize the significance of runoff to Hg inputs (8-12). The amount of mercury transported via runoff varies as a function of the characteristics of the catchment basin including soil type, hydrology, land use, and human activities such as discharge of wastewater (8). Knowledge of the proportions of atmospheric and runoff inputs of Hg to a depositional environment provides a valuable scientific and management tool. Mercury contamination in south Florida has received increased public attention since 1989 when Hg levels in game * Corresponding author phone: (321)674-7305; fax: (321)674-7212; e-mail: [email protected]. † Florida Institute of Technology. ‡ Present address: Chemistry Laboratory, Southwest Florida Water Management District, Brooksville, FL 34609. § National Oceanic and Atmospheric Administration. # University of Miami. 4058

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 19, 2000

fish from the Florida Everglades were found to exceed the advisory value for human consumption of 1.5 µg g-1 wet wt (13). To date, efforts to explain elevated levels of Hg in south Florida have included measurements of Hg in the atmosphere, water, sediment, and biota. Several studies have examined atmospheric transport and deposition of Hg in the Everglades and suggested contributions from both local and global sources (14-16). A study in the upper Everglades showed higher concentrations of total Hg, total P, and organic C in water from more northern canals and higher levels of Hg, P, and organic C in sediments from more southern canals, suggesting Hg remobilization and transport to the south (17). Furthermore, drying and flooding cycles, linked with recent drainage strategies, may have caused Hg losses from organicrich soils in selected portions of the Everglades (18, 19). Quantitative estimates of Hg loading from runoff versus direct atmospheric input for depositional areas in south Florida are presently unavailable. In this study, we used sediments and seagrasses collected from river sloughs and various sites in the lower Everglades and Florida Bay to evaluate the relative importance of direct atmospheric input versus runoff to Hg loadings. Such information on transport pathways for Hg supports current efforts to reduce Hg levels and to identify key waterways that need pollutant load reduction.

Methods Sampling. Sediment cores, about 100-150 cm long, were carefully collected in the lower Everglades and Florida Bay at water depths of 0.5-4 m between October 1994 and May 1998 (Figure 1). To obtain interpretable cores with good stratification, test cores were split and examined in the field before final decisions on sample locations were made. Surface sediments (top 2 cm) and the seagrass, Thalassia testudinum, also were collected from Florida Bay during June 23 and 24, 1998, using a grab sampler deployed from a small boat and by direct collection using plastic gloves (Figure 1). Upon return from the field, the sediment cores were sectioned into 1-cm intervals. Surface sediment and subsectioned core samples were stored at 4 °C until analysis. Seagrass leaves were stored frozen after epiphytes were removed by scraping all leaves with a cover-slip. Radiometric Determination. The activities of 210Pb, 214Pb, 214Bi, and 137Cs were determined using a well-type, intrinsic germanium detector (WiGe, Princeton Gamma Tech). Vials containing about 10 g of freeze-dried sediment were counted for about 2 days until all peak areas were sufficient to provide 100-120 yr) for application of the excess 210Pb technique. northwestern Oyster Bay (Figure 2). These records of nonsteady-state deposition and sediment disturbance by hurricanes and other sedimentary process for these cores are discussed in detail by Kang (28) and Nelsen et al. (29). Fortunately, a 35-year, post-1960 record of undisturbed sediment was found in all cores with varying degrees of preservation dating back as far as ca. 1900 in the core from Coot Bay. Concentrations of Total and Excess Mercury. Concentrations of total Hg in recent (post-1990) sediments from this study range from 236 ng g-1 at Avocado Creek to 14 ng g-1 at First National Bank (Table 1). These levels of total Hg are 2-4 times above background levels found in deeper sections from each core, yet similar to total Hg levels of 10-330 ng g-1 reported by Rood et al. (18) for sediment from corresponding areas of the Everglades. The onset of marked Hg enhancement in our sediment cores began between the 1930s and 1950s (Figure 3), depending on sample location and the record preserved at that site. In the deepest portion of these cores (prior to ∼1900), concentrations of total Hg ranged from 64 ng g-1 in the organic-rich (18-26% TOC) sediments at Avocado Creek to 3 ng g-1 in sediment with ∼1% TOC from First National Bank. Background levels of Hg (pre-1900) correlated with concentrations of TOC (r2 ) 0.88, n ) 24) and TOC plus Al (r2 ) 0.92, n ) 24; Figure 4). Previous studies reported that 70-90% of the total Hg in fjord and lake sediments was associated with an organic fraction in preference to an aluminosilicate fraction (30, 31). However, we found that calculated concentrations of excess Hg, as described below, yielded negative numbers in a few instances if we normalized only to TOC, especially in sediment where levels of TOC and Al were both ∼1%. Therefore, even though aluminosilicates generally play a minor role in explaining natural Hg levels in the lower Everglades and Florida Bay, we obtained more consistent and logical values for excess Hg for all samples by 4060

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 19, 2000

normalizing to (Al+TOC). Concentrations of excess Hg above natural levels were determined using the formula

[Excess Hg]z ) [Total Hg]z - {[Hg/(Al+TOC)]bgk × [Al+TOC]z} (1) where z is the sediment depth of interest (cm) and background (bgk) values of [Hg/(Al+TOC)]bgk are calculated using data from sediments for each core deposited prior to 1900. The average [Hg/(Al+TOC)] ratio in pre-1900 sediments from six different locations is 2.9 ( 0.8 × 10-7 (n ) 24). At a given site, this ratio for pre-1900s sediments varied by generally less than 0.4 × 10-7. Despite the potential for loss of TOC during early chemical diagenesis, use of [Hg/(Al+TOC)] ratios is supported by the spatial and temporal uniformity in values for pre-1900 sediments. Concentrations of excess Hg in recent (post-1990) sediment range from 13 to 133 ng g-1 in the area of Shark River Slough and 2-8 ng g-1 in Florida Bay (Table 1). At these levels in recent sediment, excess Hg accounts for 20-60% of the total Hg. Values for excess Hg approach zero during the late 1800s in the well-preserved sediment of Coot Bay. However, in less-well preserved deeper sediment from the other locations, excess Hg levels approach zero during the 1930s-1940s. Fluxes of Mercury to the Sediment. Values for the total flux of excess Hg (ΣFHg) to post-1990 sediment, calculated as the product of sediment accumulation rate and excess Hg concentration for each layer in a core, range from 199 µg m-2 yr-1 in Avocado Creek to 18 µg m-2 yr-1 at First National Bank (Table 1). This 10-fold decrease in the flux of excess Hg at increasing distances from river mouths into Florida Bay is most likely coincident with comparable shifts in the fraction of excess Hg that comes from direct atmospheric fluxes versus runoff inputs as described below.

FIGURE 3. Vertical profiles for concentrations of total Hg and fluxes of excess Hg in age-dated sediments from Coot Bay, NW Oyster Bay, Jimmie Key, and First National Bank. The dotted lines with open circles are for concentrations of total Hg, and the solid lines with closed circles are for fluxes of excess Hg. particle (Hgparticulate and 210Pbparticulate) phase transformations in the atmosphere. Fitzgerald et al. (7), in their critical review, developed a case for steady-state emission of Hg0 and 222Rn from soils and comparable chemical stability for Hg and 210Pb in sedimentary environments. If we assume reasonably similar behavior for Hg and 210Pb from the atmosphere to a sedimentary repository, a simple mass balance model for excess 210Pb can be used to differentiate direct atmospheric inputs of Hg from runoff fluxes of Hg. Benninger (33) previously showed that direct atmospheric deposition and riverine inputs of excess 210Pb were the two dominant sources for his model of Long Island Sound. Applying these conditions to our study area, the direct atmospheric input of excess Hg (FHg atm) at present can be calculated using the formula

FHg atm ) S × [Excess Hg]top × FIGURE 4. Concentrations of total Hg versus (Al+TOC) in cores collected at six locations. Open symbols and closed circles represent data points for post-1990 sediments and pre-1900 sediments, respectively. The open symbols include Avocado Creek (O), NW Oyster Bay (0), NE Oyster Bay (]), Coot Bay (4), Jimmie Key (3), and First National Bank (+). The solid line indicates the best-fit line obtained from the pre-1900 data (n ) 24). The figure insert shows an expanded view of the data with lower concentrations. Recent studies support the hypothesis that Hg and 210Pb follow relatively similar pathways from soils to the atmosphere to rainwater to deposition and burial in sediments (7, 32). For example, Lamborg et al. (32) showed a good correlation between Hg and 210Pb in rainwater (r2 ) 0.72) that they believe results from similar gas (Hg0 and Rn) to

(

FPb-210 atm FPb-210 total

)

(2)

where S is the sediment accumulation rate (g cm-2 yr-1) based on excess 210Pb, [Excess Hg]top is the concentration of excess Hg in the top layer of sediment from a given core (µg g-1), (FPb-210 atm) is the local long-term atmospheric flux of excess 210Pb (1.0 ( 0.1 dpm cm-2 yr-1) obtained from seven different lake sediments in Florida (24), which corresponds to values for other areas along the east coast of the U.S.A. (33-35), and (FPb-210 total) is the total flux (i.e., atmospheric + runoff) of excess 210Pb (dpm cm-2 yr-1) to the sediment at each site, obtained by multiplying the sediment accumulation rate for each core by the computed initial activity of excess 210Pb. Thus, the fraction of total excess Hg in present-day sediments that came from direct atmospheric deposition is proportional to the fraction of total, excess 210Pb that was directly deposited VOL. 34, NO. 19, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4061

TABLE 2. Data for Hg in Surface Sediment (Post-1990)

local name

station

relative accumulation ratea (g cm-2 yr-1)

Barnes Sound East FL Bay Near Taylor Slough Central FL Bay SW FL Bay

S1-4 S6, 7 S10-13, 23, 24 S30 S17, 18

0.21-0.30 0.19-0.29 0.45-0.75 0.63 0.23-0.66

total Hg (ng g-1)

excess Hg (ng g-1)

relative excess Hg fluxb (µg m-2 yr-1)

14-20 9-10 10-26 12 15-16

7-15 5-6 4-8 3 1-2

22 ( 7 14 ( 5 30 ( 6 17 5(3

a The relative sediment accumulation rate (RS) at each site was obtained using the best-fit equation for sediment accumulation rate versus excess 210Pb/Al ratio at the top layer for each core: RS ) -0.0019 (210Pb/Al)0 + 1.13; (r2 ) 0.88, n ) 5). b Mean ( standard deviation.

FIGURE 5. Excess Hg flux via direct atmospheric input (FHg atm) versus runoff (FHg runoff) to recent sediments (>1990) from the lower Everglades and Florida Bay. Sample stations are listed based on their proximity to Everglades National Park (ENP) or Florida Bay as follows: Avocado Creek (Avo Crk), northeastern Oyster Bay (NE OB), northwestern Oyster Bay (NW OB), Coot Bay (Ct Bay), Jimmie Key (Jim Key), and First National Bank (1st NB). from the atmosphere. The calculated fluxes are reported in µg m-2 yr-1 to allow comparison with data from other studies. The error associated with the fluxes was obtained through a propagation of error analysis for the four terms in eq 2 using the approach described by Meyer (36). The direct atmospheric flux of excess Hg (FHg atm) calculated for post-1990 sediments at Avocado Creek is 40 ( 7 µg m-2 yr-1 and comprises only 20% of the ΣFHg. In contrast, the FHg atm for First National Bank is lower at 14 ( 5 µg m-2 yr-1, yet the direct atmospheric flux of excess Hg at this site accounts for >80% of the ΣFHg. A southward trend of decreasing values for FHg atm (Figure 5) suggests that local influence from the greater Miami area is superimposed on a more regional (i.e., North America) flux of excess Hg. Overall, an average FHg atm of 24 ( 9 µg m-2 yr-1 throughout the study area is similar to values of 21 and 30 µg m-2 yr-1 for bulk atmospheric deposition in south Florida at eastern ENP and Tamiami Trail (northern ENP), respectively (16), and the atmospheric flux of 12.5-20 µg m-2 yr-1 in North America (37-39). Runoff fluxes of excess Hg (FHg runoff ) ΣFHg - FHg atm) range from 4 to 159 µg m-2 yr-1 with a trend of decreasing values from the river sloughs to more isolated and remote sites (Figure 5). The largest runoff flux of 159 ( 24 µg m-2 yr-1 at Avocado Creek is about 4 times greater input than at NW Oyster Bay and Jimmie Key and about 8-40 times greater than for isolated Coot Bay and more remote First National Bank. Overall, values for FHg runoff near river mouths (AC, NW OB, JK) are 2-7 times greater than the average direct atmospheric flux of excess Hg. Mercury in Surface Sediment and Seagrass. Data for surface sediment support the quantitative importance of Hg loading via runoff obtained from sediment cores. Fluxes of excess Hg were calculated from sites where only surface sediments were collected as the product of [Excess Hg]top 4062

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 19, 2000

and the sediment accumulation rate. Excess Hg levels for surface sediments were determined by subtracting natural Hg concentrations calculated using the equation shown in Figure 4 from total Hg levels. Sediment accumulation rates were obtained using the inverse relationship between sediment accumulation rate and the excess 210Pb/Al ratio at the top layer of individual cores (see Table 2 for details). The relationship between the excess 210Pb/Al ratio in surface sediment and sediment accumulation rate was validated using the complete cores analyzed for this study; however, we use the term relative flux of excess Hg because we do not have an exact value. The largest relative flux of excess Hg (37 µg m-2 yr-1) was calculated for station S11, near Taylor Slough and the lowest flux (3 µg m-2 yr-1) was determined for SW Florida Bay (S17). Overall, the relative fluxes of excess Hg for surface sediments show a distinct spatial pattern with ∼2-6 times greater accumulation of excess Hg near Taylor slough relative to more remote sites in southwestern and eastern Florida Bay (Table 2). Total Hg concentrations in leaves of the seagrass, Thalassia testudinum, complement data for Hg in sediment. The highest level of total Hg (53 ng g-1, dry wt) was found for leaves collected near the mouth of Taylor Slough (G12), and the lowest Hg level (17 ng g-1) was obtained for leaves from southwestern Florida Bay (G17 and G18) away from river sloughs where seagrass beds were very dense. This spatial pattern for total Hg in seagrass, although limited, indicates Hg enrichment near Taylor Slough, most likely the result of Hg in runoff. The importance of runoff to the ΣFHg in our study area is consistent with past water management strategies (i.e., water flow reduced by 40% since the 1900s) and increased agricultural use of organic-rich soils in the upper Everglades (40). Such local activities increase the chance for oxidation, vaporization, and leaching of Hg bound in preexisting, contaminated organic soils. Losses of Hg from dry sites in the northern portion of ENP described by Rood et al. (18) are certainly consistent with increased transport via runoff to our study area. In summary, direct atmospheric inputs of excess Hg obtained using sediment cores collected in the lower Everglades and Florida Bay are relatively uniform (24 ( 9 µg m-2 yr-1) and comparable with bulk atmospheric deposition in south Florida (21-30 µg m-2 yr-1). However, fluxes of excess Hg via runoff to the sites near Shark River Slough and Taylor Slough in the lower Everglades and Florida Bay are 2-7 times greater than direct atmospheric inputs. These results show that runoff is the most important means for transporting Hg to the lower Everglades and immediately adjacent Florida Bay, whereas atmospheric transport is the primary pathway for carrying Hg to more isolated or remote areas. In contrast with most studies for remote lakes (7), historical trends for ΣFHg are not synchronous throughout the lower Everglades and Florida Bay due to the varying influences of runoff. To prevent expansion of the Hg problem now being experienced in the upper Everglades, the choice of future water manage-

ment strategies must provide an adequate volume of water with less terrigenous sediments.

Acknowledgments We thank Simone Metz and Robert Trocine for help with sample analysis and Carlos Alvarez-Zarikian and Geoffrey Ellis for assistance in field sampling. Mark Rials, John Windsor, and Iver Duedall provided helpful comments for the manuscript. We thank Carl Lamborg for sending his forthcoming Hg paper. This manuscript was improved by the insightful suggestions of the three anonymous reviewers. The South Florida Ecosystem Restoration Prediction and Modeling Program of the National Oceanic and Atmospheric Administration (NOAA) supported this research.

Literature Cited (1) Evans, R. D. Arch. Environ. Contam. Toxicol. 1986, 15, 505-512. (2) Meger, S. A. Water, Air, Soil Pollut. 1986, 30, 411-419. (3) Rada, R. G.; Wiener, J. G.; Winfrey, M. R.; Powell, D. E. Arch. Environ. Contam. Toxicol. 1989, 18, 175-181. (4) Verta, M.; Tolonen, K.; Simola, H. Sci. Total Environ. 1989, 87/ 89, 784-787. (5) Lindqvist, O.; Johansson, K.; Aastrup, M.; Anderson, A. Water, Air, Soil Pollut. 1991, 55, 7-261. (6) Landers, D. H.; Gubala, C.; Verta, M.; Lucotte, M.; Johansson, K.; Vlasova, T.; Lockhart, W. L. Atmos. Environ. 1998, 32, 919928. (7) Fitzgerald, W. F.; Engstrom, D. R.; Mason, R. P.; Nater, E. A. Environ. Sci. Technol. 1998, 32, 1-7. (8) Balogh, S. J.; Engstrom, D. R.; Almendinger, J. E.; Meyer, M. L.; Johnson, D. K. Environ. Sci. Technol. 1999, 33, 3297-3302. (9) Bishop, K. H.; Lee, Y. H.; Pettersson, C.; Allard, B. Water, Air, Soil Pollut. 1995, 80, 445-454. (10) Hurley, J. P.; Benoit, J. M.; Babiarz, C. L.; Shafer, M. M.; Andren, A. W.; Sullivan, J. R.; Hammond, R.; Webb, D. A. Environ. Sci. Technol. 1995, 29, 1867-1875. (11) Sanudo-Wilhelmy, S. A.; Gill, G. A. Environ. Sci. Technol. 1999, 33, 3477-3481. (12) Balogh S. J.; Meyer M. L.; Johnson, D. K. Environ. Sci. Technol. 1997, 31, 198-202. (13) Ware, F.; Royals, H.; Lange, T. Proc. Am. Conf. SEAFWA 1990, 44, 5-12. (14) Landing, W. M.; Perry Jr, J. J.; Guentzel, J. L.; Gill, G. A.; Pollman C. D. Water, Air, Soil Pollut. 1995, 80, 343-352. (15) Dvonch, J. T.; Vette, A. F.; Keeler, G. J.; Evans, G.; Stevens, R. Water, Air, Soil Pollut. 1995, 80, 169-178. (16) Guentzel, J. L.; Landing, W. M.; Gill, G. A.; Pollman, C. D. Water, Air, Soil Pollut. 1995, 80, 393-402. (17) Stober, Q. J.; Jones, R. D.; Scheidt, D. J. Water, Air, Soil Pollut. 1995, 80, 991-1001. (18) Rood, B. E.; Gottgens, J. F.; Delfino, J. J.; Earle, C. D.; Crisman, T. L. Water, Air, Soil Pollut. 1995, 80, 981-990.

(19) Lodge, T. E. In The Everglades Handbook: Understanding the Ecosystem; Lodge, T. E., Ed.; St. Luice Press: Delray Beach, FL, 1994. (20) Kim, K. H.; Burnett W. C. Anal. Chem. 1983, 55, 1796-1800. (21) Koide, M.; Bruland, K. W.; Goldberg, E. D. Geochim. Cosmochim. Acta 1973, 37, 1171-1187. (22) Robbins, J. A.; Edgington, D. N. Geochim. Cosmochim. Acta 1975, 39, 285-304. (23) Appleby, P. G.; Oldfield, F. Catena 1978, 5, 1-8. (24) Binford, M. W.; Brenner, M. Limnol. Oceanogr. 1986, 31, 584595. (25) Peirson, D. H. Nature 1971, 234, 79-80. (26) Hermanson, M. H.; Christensen, E. R. J. Great Lakes Res. 1991, 17, 33-50. (27) Trefry, J. H.; Metz, S. Anal. Chem. 1984, 56, 745-749. (28) Kang, W.-J. Ph.D. Dissertation, Florida Institute of Technology, Melbourne, FL, 1999. (29) Nelsen, T.; Wanless, H. R.; Trefry, J. H.; Alvarez-Zarikian, C.; Hood, T.; Blackwelder, P.; Swart, P.; Tedesco, L.; Kang, W.-J.; Metz, S.; Garte, G.; Featherstron, C.; Souch, C.; Pachut, J. F.; O’Neal, M.; Ellis, G. In Linkages Between Ecosystems in the South Florida Hydroscape: The River of Grass Continues; Porter, J. W., Porter, K. G., Eds.; in press. (30) Smith, J. N.; Loring, D. H. Environ. Sci. Technol. 1981, 15, 944951. (31) Nater, E. A.; Grigal, D. F. Nature 1992, 358, 139-141. (32) Lamborg, C. H.; Fitzgerald, W. F.; Graustein, W. C.; Turekian, K. K. J. Atmos. Chem. 2000, 36/3, 325-338. (33) Benninger, L. K. Geochim. Cosmochim. Acta 1978, 42, 11651174. (34) Turekian, K. K.; Nozaki, Y.; Benninger, L. K. Annu. Rev. Earth Planet. Sci. 1977, 5, 227-255. (35) Urban, N. R.; Eisenreich, S. J.; Grigal, D. F.; Schurr, K. T. Geochim. Cosmochim. Acta 1990, 54, 3329-3346. (36) Meyer, S. L. In Data Analysis for Scientists and Engineers: Propagation of Error and Least Squares; Meyer, S. L., Ed.; John Wiley & Sons: 1975; pp 39-48. (37) Swain, E. B.; Engstrom, D. R.; Brigham, M. E.; Henning, T. A.; Brezonik, P. L. Science 1992, 257, 784-787. (38) Engstrom, D. R.; Swain, E. B. Environ. Sci. Technol. 1997, 31, 960-967. (39) Benoit, J. M.; Fitzgerald, W. F.; Damman, A. W. In Mercury Pollution: Integration and Synthesis; Watras, C. J., Huckabee, J. W., Eds.; Lewis Publishers: Boca Raton, FL, 1994; pp 187202. (40) Smith, T. J., III.; Hudson, J. H.; Robblee, M. B.; Powell, G. V. N.; Isdale, P. J. Bull. Mar. Sci. 1989, 44, 274-282.

Received for review March 1, 2000. Revised manuscript received July 4, 2000. Accepted July 7, 2000. ES0010608

VOL. 34, NO. 19, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4063