- Email: [email protected]
Mercury in canopy leaves of French Guiana in remote areas
The Science of the Total Environment 311 (2003) 261–267
Mercury in canopy leaves of French Guiana in remote areas ´ ` a,*, M. Pourcheta, P. Charles-Dominiqueb, P. Gaucherc Marie-Antoinette Melieres a
´ ` Laboratoire de Glaciologie et Geophysique de l’Environnement, CNRS, Universite´ J. Fourier, BP 96, 38402 St Martin d’Heres, France b ´ ´ ˆ Laboratoire d’Ecologie Generale, MNHN, 4 rue du Petit Chateau, 91800 Brunoy, France c ´ Mission pour la Creation du Parc de la Guyane, BP 275, 97326 Cayenne, France Received 29 May 2002; accepted 8 February 2003
Abstract A study of total Hg concentration in the foliage of the canopy was carried out in two remote areas in French Guiana. The sampled canopy is representative of the French Guiana canopy. The concentration in the foliage, 64"14 ng gy1 (dry wt.), is used to estimate the annual input of total Hg to the soil through the litterfall, found to be 45"10 mg my2 yy1. As translocation is negligible, mercury in the canopy originates mainly from atmospheric uptake by the leaves and this litterfall deposit represents a direct atmospheric input from the background atmospheric load into the soil. 䊚 2003 Elsevier Science B.V. All rights reserved. Keywords: Mercury; Canopy; Rainforest; Hg atmosphere soil transfer
1. Introduction As in the Amazonian basin of Brazil, large areas of French Guiana (South America) have been subjected to gold mining since the 19th century. This activity uses large quantities of mercury (Hg) (1–4 kg Hg for 1 kg of gold produced), which are mainly lost in soil, river sediment and the atmosphere. In order to estimate, understand, and finally prevent the resultant Hg pollution, studying the Hg cycle has been a major goal of the last 20 years (see for example Schroeder W., 1998). To estimate the anthropogenic impact we must first *Corresponding author. Tel.: q33-4-76-82-42-11; fax: q 33-4-76-82-42-01. E-mail address: [email protected] ´ ` (M.-A. Melieres).
evaluate the contribution of the two natural origins of the Hg present in soil: geological soil content and atmospheric deposition. Atmospheric uptake is an important process by which Hg accumulates in plants (see for example Lindberg et al., 1992, 1994; Munthe et al., 1995; Benesh et al., 2001). We focus here on the contribution of the background atmospheric deposit, measuring specifically the contribution from the leaves of the canopy through litterfall deposit. Total Hg concentration in leaves has been measured by Roulet et al. (1999) in different parts of Brazil and in French Guiana; results have shown a higher content, close to double, at the French Guiana site (mean values in French Guiana ranging from 130 to 150 ng gy1, compared to 40–120 ng gy1 in Brazil). Is this high content representa-
0048-9697/03/$ - see front matter 䊚 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0048-9697(03)00142-6
262
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267
tive of the entire French Guiana canopy in remote areas? Our goal here is to establish a representative value for this total Hg deposit in litterfall for the tropical rainforest of French Guiana. This requires first sampling representative species in the canopy, and second, performing this sampling in remote areas free from gold mining Hg pollution and from human settlement, to detect the contribution from the background atmospheric level. Previous estimates of Hg content in the tropical forest canopy have been made but not systematically. 2. Sampling and measurements 2.1. Study area Two sampling sites were chosen. The first is the field station of ‘Les Nouragues’, the French CNRS observatory in a remote area (48059 N, 528419 W) situated 100 km away from the oceanic coast at an altitude of 50 m above sea level (asl). The site is under regular trade wind influence throughout the year with a prevailing northeast direction. This atmospheric circulation involves the first 1000 m of the atmosphere (the low troposphere), an air layer which carries most of the atmospheric load. This load in terms of gaseous compounds and aerosol is therefore of Atlantic origin and is representative of background levels. This has been tested at the site of Petit Saut in French Guiana ´ ` (Melieres et al., 2003) through a continuous monitoring of the aerosol load using the 210Pb radionuclide, a major tracer of aerosol transport and deposit. Monitoring began in July, 1999 and still continues. This site is under the influence of air masses of similar origin to the sites of Les Nouragues and Pic Matecho; results show that the 210 Pb air concentration at Petit Saut is close to 0.2 ´ ` Bq my3 during both wet and dry seasons (Melieres et al., 2003), a value which is characteristic of an oceanic air mass in tropical regions, and similar to what give data for others studies (see the data base on lead-210 by Preiss et al., 2000). The trajectory of this oceanic air mass arriving at Les Nouragues precludes any large influence from the vast surrounding Amazonian regions where gold mining could load the atmosphere with Hg vapor. The only gold mining activities which exist close
to the station are at ‘la Montagne Tortue’ (30 km north), and at Boulanger (60 km north-east). These places have restricted activity, and the atmospheric influence of Hg vapor released is assumed to be negligible over such a distance. The second site is situated in the center of French Guiana (03843944.50 N, 53802918.40 W), 3 km south of Pic Matecho (590 m asl), at an altitude close to 250 m. This appears to be an ideal pristine site: no gold mining existed in this area, and there has been no human settlement during the 20th century. It is 150 km south of the Atlantic coast, 30 km north-east of the small ¨ (f200 inhabitants). As at the village of Saul Nouragues station the air masses arriving from north-east do not cross any gold mining areas and come directly from the Atlantic; they can therefore be considered to be under the influence only of background atmospheric load. Thus at both sites atmospheric Hg can originate only from the natural background atmospheric load. 2.2. Sampling At Les Nouragues, a large area has been equipped with five observation platforms in the upper canopy, between 35 and 40 m above the ground. This area has been extensively studied for 15 years (ecology, vegetation dynamic, biodiversity,«). Canopy walkways, with a total length of 108 m, connect the different observation platforms. The upper canopies of trees belonging to six different species have been sampled. Foliage was collected by selecting fully developed undamaged leaves (between 1 and 3 years old), cutting the leaves at the stalk. The leaves were placed directly into plastic bags, where they were stored. Foliage samples were collected on the emergents of three trees on which the platforms had been installed: Sapotaceae on the first platform, Ampelozisyphus americanus (Rhamnaceae) on the second, and Quaribea sp. (Bombacaceae) on the third. The last three samples, Licania sp. (Chrisobalanaceae), Streculia sp. (Sterculiaceae) and Vouacapoua americana (Caealpiniaceae), were collected on trees whose canopies are between platforms 1 and 2. All these trees are situated on a hill (altitude 60–90 m) dominated by an ‘inselberg’ (a granite
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267
mountain of altitude 430 m) and are representative of high forest on drained soil, a type of forest well represented in French Guiana. The sampled trees form the high level of the forest, between 30 and 35 m above the ground (Poncy et al., 1998). At the Pic Matecho site, the tree canopies of five different species were sampled in July, 2001. These trees were cut at the edge of a recent natural treefall; we sampled the trees at the same height as at the other site (upper canopy). Trees were 30–35 m high, on drained soil. The forest of this region is well known for its high biodiversity. The collected species were Chrisobalanaceae, Eperua falcata, Liane, Pourouma sp. and Caraipa sp. (Clusiaceae). At Les Nouragues, collection was performed in 1999 and, for each tree, sampling was duplicated on two different leaves. In the Pic Matecho field campaign in 2001, 3–5 samplings of different leaves of the same tree were carried out in order to estimate more precisely the range of Hg concentration. In order to be able to consider the measurements at those two sites to be representative of French Guiana forest we have to address two questions: Is the sampled forest at each site representative of the tropical rainforest in French Guiana? The inventory at the Nouragues station indicates density per hectare of 64–125 trees with diameter above 30 cm, and 468–680 of diameter above 10 cm; 550 tree species corresponding to 63 families have been censused at Les Nouragues (Poncy et al., 1998, 2001). None of them is dominant. At the very most 20 of them (including Vouacapoua americana) are more frequently found. The basal area, which is well correlated to the total biomass, is between 30 and 45 m2 hay1. This is only slightly higher than the Amazonian forest (e.g. Valencia et al., 1994; Rankin-de Merona et al., 1992; Spichiger et al., 1996) but less than the value of 53 ¨ m2 hay1 found in the very beautiful forest of Saul, in central French Guiana, well known for its rich biodiversity (Mori et al., 1983). Pic Matecho, our second site, is situated in this region. All 11 specimens are common species and, taken together, constitute approximately 10% of the forested bio´ mass (Prevost and Sabatier, 1996). They can therefore be considered as representative of the
263
French Guiana forest and of the mean operating cycle of the tropical rainforest. 3. Analyses and results After each campaign, the collected leaves were dried following the same procedure as used for sediments (2 days in an oven at 60 8C). This procedure does not appear to affect the total Hg content in sediments. Dried leaves were then sealed in plastic bags and sent to the Centre d’Analyse du CNRS (Vernaison, France). The total Hg concentration was measured using the cold vapor technique with atomic absorption spectrometry detection (Ama 254, Spectra France). The dried leaves were directly placed in a cell and heated to 750 8C and the Hg vapor content determined. The limit of analysis is 0.5 ng of Hg (the detection limit is ten times less), and the relative standard deviation is a few percent for 10 ng of Hg. The estimated measurement error is 0.1 ng. Mercury measurements were performed on the blade of the leaves and, in two cases, in the central nervation of the leaf. Results are presented in Table 1. In this table, the different measurements for each species correspond to different leaves on the same tree. 4. Discussion 4.1. Mean leaf concentration At Les Nouragues, the total Hg concentration ranges from 32.4 to 114 ng gy1, whereas at Pic Matecho concentration ranges from 52.4 to 103.0 ng gy1. For a given species, dispersion appears to be small whereas leaves can be of different ages. Both measurements in the central nerve indicate that Hg concentration is systematically lower (5– 10 times lower) than in the blade. The mean values obtained at Les Nouragues (59"28 ngHg gy1) and at Pic Matecho (70"16 ngHg gy1) cannot be considered to be significantly different. This leads to assume an overall mean concentration in the canopy of 64"14 ng gy1 (with a confidence level of 95%), for the tropical rainforest of French Guiana.
264
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267
Table 1 Total Hg concentration in the blade of the leaves and in the central nerve of the leaves (given in brackets) Nouragues
Pic Matecho
Family
Dry matter (mg)
Total Hg (ng gy1)
Family
Dry matter (mg)
Total Hg (ng gy1)
Sapotaceae
106.9 105.7
36.7"1 35.7"1
Chrisobalanaceae
92.4 127.1 108.1 109.1 108.0 (43.2)
59.0"1 45.1"0.8 50.8"1.2 40.0"0.9 57"0.8 (13.3"2.2)
Licania sp.
30.5 63.4
40.7"3.6 46.6"1.7
Eperua falcata
47.8 56.7 50.2 54.0 87.8
77.7"2.1 52.4"2.1 82"1.9 58.6"1.7 81.0"1.1
Sterculiaceae
42.8 91.2
65.7"2.7 59.7"1.2
Caraipa sp.
80.7 122.2 87.0 (79.7)
80.7"1.2 75.6"0.8 81.4"1.1 (8.1"1.3)
Strychnos sp.
108.7 147.2
68.4"1.3 62.5"0.9
Liane
65.3 88.3 73.2
65.2"1.5 53.1"1.1 88.3"1.3
Vouacapoua americana
90.4 62.7
32.4"1.3 37.9"1.8
Pourouma sp.
94.4 63.0 80.5 93.0
84.8"1 82.9"1.6 91.6"1.2 103.0"1
Bombacaceae
56 94.8
114"1.9 105"1.1
4.2. Comparison with other measurements Total Hg concentration in leaves has been measured by Roulet et al. (1999) in different parts of Brazil (along the Tapajos Basin and in the Tocantins Basin) and at three sites in French Guiana (f80 km south of Cayenne), and results have shown a higher content (mean value 142"50 ng gy1) at the French Guiana site than in Brazil (mean value of 73"40 ng gy1). Our results indicate a lower value (55% smaller) for the remote areas of French Guiana. Moreover, the measurements performed by Roulet et al. (1999) suggest a Hg content in the foliage similar in the remote areas of Brazil (the Tapajos and the Tocantins Basins) and in the remote areas of French Guiana. Few measurements exist in Europe in remote areas. Iverfeld (1991) measured the pine needle concen-
tration in remote areas of Sweden and found systematically lower values (40 ng gy1). Other measurements in spruce needles performed in Scandinavia (see Goldbold, 1994, for a review article) show similar values in the range of 40–50 ng gy1. 4.3. Mean contribution of canopy to Hg content in soil The French Guiana forest is sempervirent. Most tropical trees lose their leaves continuously, after a mean time of 4 years, in an asynchronous manner (Hadlik and Blanc, 1987). The litter deposit on soil has been estimated in tropical forest in different parts of the world. Results are quite homogeneous and the annual deposit is of the order of 7 tons of dry matter per hectare (Proctor, 1983,
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267
1984; Leight, 1978). A more accurate estimate is in progress at Les Nouragues. The annual estimated litter deposit of 7 ton hay1, together with the mean concentration in the leaves of 64"14 ngHg gy1, leads to an annual mean deposit on soil of 45"10 mgHg my2 due to litterfall. 4.4. Origin of Hg in leaf canopy There are two possible origins of the Hg present in canopy leaves: atmospheric uptake by foliage, andyor translocation (i.e. transfer of Hg from soil to leaves through the roots of the tree). This second process is usually believed to be negligible (see e.g., Gilmour and Miller, 1973; Beaufort et al., ˆ 1977; Lindberg et al., 1979; Goldbold and Huttermann, 1988; Graydon et al., 2001; Rea et al., 2002), leaving Hg atmospheric load as the main contributor to Hg concentration in leaves. The Hg content of the atmosphere at the canopy level can be influenced by the soil emission. Literature suggests that this latter source is much less important than the global atmospheric load in unpolluted areas. Therefore, the Hg concentrations measured in the canopy foliage originate mainly from the atmosphere which, at these two sites, is representative of the background level. This leads to an estimate for the background atmospheric contribution to the soil via litterfall of 45"10 mgHg my2 in the French Guiana rainforest. 4.5. Atmosphere–soil transfer In tropical rainforests, the Hg present in the atmosphere can be deposited on the soil mainly via 3 processes: rainfall, throughfall and litterfall deposits. In a forested area the rain lands on the soil after passing through the canopy and becomes enriched in Hg (throughfall deposit); during this transfer, the wash-off of dry deposited Hg compounds from the leaf surface appears to be the most important source of Hg, rather than foliage leaching (Rea et al., 2000, 2001). The third type of atmosphere–soil transfer operates through the foliage depositionyuptake, followed by the litterfall deposit. Measurements performed over two growing seasons on foliar Hg concentration indicate
265
that atmospheric Hg accumulates during the growing season in living foliage (Rea et al., 2002). This may be illustrated by the 5–10 times lower concentration that we found in the central nervation of the leaf (Table 1), compared to the concentration found in the blade of the leaves (living cells). Absorbed Hg may be released by evaporation andyor leaching but these two mechanisms are shown to be of minor importance (Lodenius et al., 2001). Different studies have focused on these mechanisms. For example, Iverfeld (1991) estimates these three components separately at the site of ¨ Sweden. The measured components are Gardsjon, total Hg rainfall deposit (12 mg my2 yeary1 ), and throughfall deposit (17–19 mg my2 yeary1); litterfall deposit was estimated (25 mg my2 yeary1). This clearly shows the amplifying role of the foliage in transmitting the atmospheric mercury to the soil: the total Hg deposit through rainfall deposit would be 12 mg my2 yeary1 on unforested soil, but close to 40 mg my2 yeary1 (throughfallq litterfall) in forested areas. Here the throughfall represents 72% of the litterfall. A similar ratio was found in the review article by Schroeder and Munthe (1998) on mercury deposition in precipitation, throughfall and litterfall in forested ecosystems of boreal forest in Sweden (Iverfeld, 1991; Munthe et al., 1995) and of temperate forests in USA (Lindberg, 1996). When averaged separately, rainfall, throughfall and litterfall are, respectively, 10, 15 and 20 mg my2 yeary1. Here the throughfall represents 75% of the litterfall: the deposition in such forested areas (throughfallqlitterfall) is significantly greater than deposition in precipitation only (rainfall). A recent study estimating through fall and measuring litterfall in deciduous forests in USA (Rea et al., 2002) leads to a comparable order of magnitude. Here, in the tropical rainforest of French Guiana, the litterfall deposit mechanism, alone, represents ;45"10 mg my2 yeary1 of input to soil. This is therefore a low estimate of the atmospheric contribution to soil, as throughfall deposit (not measured here) should also be included. The fact that temperate (deciduous) and boreal (evergreen) forests display a similar ratio of litterfall to throughfall, is a good indication that this
266
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267
ratio can be extended to tropical forests. If the same ratio, (0.7), is assumed, the Hg flux due to throughfall may reach 30 mg my2 yeary1, thus yielding a total deposition flux of 75 mg my2 yeary1. 5. Conclusion The purpose of this study was to estimate the canopy contribution to the Hg present in the soils of the French Guiana tropical rainforest through the litterfall deposit only. This canopy contribution originates mainly from the atmospheric background load. A study of Hg concentrations in the foliage of the canopy was carried out in two remote areas where gold mining has not occurred and which are free from any other anthropogenic influences. The sampling (11 species) of the canopy is representative of the French Guiana canopy. The mean average concentration found in the foliage was 64"14 ng gy1 (dry wt.). Using an annual litterfall deposit of 7 ton hay1, a value typical of tropical rainforests, these results lead us to estimate the annual input of total Hg to the soil through the litterfall to be ;45"10 mg my2 yeary1. As translocation is believed to be negligible, mercury in the canopy originates mainly from atmospheric uptake by the leaves. This measured flux, therefore, represents a direct atmospheric input into the soil; in fact it represents the lower limit of this atmospheric deposit, as deposition due to throughfall must also be included in the total atmospheric deposit to soil in a forested area. This mechanism for transferring Hg from atmosphere to soil contributes partly to the Hg content in of remote areas in the French Guiana tropical rainforest. Acknowledgments We are very grateful to S. Mori for helpful comments. This project was supported by The Mission du Parc de la Guyane, the Programme ` de ‘Mercure en Guyane’ (INSU-CNRS, Ministere l’Environnement, Europe), and the University of Grenoble.
References Benesh JA, Gustin MS, Schorran DE, Coleman J, Johnson DA, Lindberg SE. Determining the role of plants in the biogeochemical cycling of mercury on an ecosystem level. Sixth International Conference on mercury as a global pollutant, Minamata, Japan; 2001. p. 40. Beaufort W, Barber J, Barringer AR. Uptake and distribution of mercury within higher plants. Physiol Plant 1977;39:261 – 265. Gilmour JT, Miller MS. Fate of a mercuric-mercurous chloride fungicide added turfgrass. J Environ Qual 1973;2:145. Goldbold DL. Mercury in forest ecosystems: risk and research needs. In: Watras CJ, et al, editor. Mercury pollution. London: Lewis publishers, 1994. p. 295 –303. ˆ Goldbold DL, Huttermann A. Inhibition of photosynthesis and transpiration in relation to mercury-induced root damage in spruce seedlings. Physiol Plantarum 1988;74:270 –275. Graydon JA, St Louis VL, Lindberg SE, Hintelman H. Mechanisms of mercury deposition under boreal forest canopies. Sixth International Conference on mercury as a global pollutant, Minamata, Japan; 2001 p. 41. Hadlik A, Blanc P. Croissance des plantes en sous-bois de ˆ dense humide (Makokou, Gabon). Rev Ecol (Terre et foret Vie) 1987;42:209 –234. Iverfeld A. Mercury in forest canopy throughfall water and its relation to atmospheric deposition. Water Air Soil Poll 1991;56:553 –564. Leight EG. Leaf production, leaf consumption, and the regulation of folivory on Barro Colorado Island. In: Montgomery GG, editor. The ecology of arboreal folivores. Washington: Smithsonian Institution Press, 1978. p. 33 –50. Lindberg SE, Jackson DR, Huckabee JW, Janzen SA, Levin MJ, Lund JR. Atmospheric emission and plant uptake of mercury from agricultural soils near the Almaden mercury mine. J Environ Qual 1979;8:572 –578. Lindber SE, Meyers TP, Taylor GE, Turner RR, Schroeder WH. Atmosphereysurface exchange of mercury in a forest: results of modeling and gradient approaches. J Geophys Res 1992;97:2519 –2528. Lindberg SE. Forests and the global biogeochemical cycle of mercury. In: Baeyens W, Ebinghaus R, Vasiliev O, editors. Regional and global mercury cycles: sources, fluxes and mass balances. Dordrecht, Netherland: Kluwer, 1996. p. 359 –380. Lindberg SE, Owens JG, Stratton WJ. Application of throughfall methods to estimate dry deposition of mercury. In: Watras CJ, et al, editor. Mercury Pollution. London: Lewis Publishers, 1994. p. 261 –271. Lodenius M, Tulisalo E, Soltanpour-Gargari A. Exchange of mercury between atmosphere and vegetation. Sixth International Conference on Mercury as a Global Pollutant, Minamata, Japan; 2001 p. 40. ´ ` Melieres MA, Pourchet M, Richard S. Surface air Concentration and Deposition of Lead-210 in French Guiana: two years of Continuous monitoring. J Environ Radioact 2003;66:261 –269.
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267 Munthe J, Hultberg H, Iverfeldt A. Mechanismes of deposition of mercury and methyl mercury to coniferous forests. Water Air Soil Poll 1995;80:363 –371. Mori SA, Boom BM, de Carvalho AM, dos Santos TS. Southern Bahian moist forests. Bot Rev 1983;49:155 –232. ´ B, Larpin D, Belbenoit P, Jullien M, Hoff M, Poncy O, Riera Charles-Dominique P. The permanent field research station les Nouragues in the tropical rainforest of French Guiana: current projects and preliminary results on tree diversity, structure and dynamics. In: Dallmeier F, Comiskey JA, editors. Forest Biodiversity in North, Central and South America and the Caribbean: research and monitoring. Man and the biosphere series, vol. 20, 22. Carnforth, Lancashire, UK: Unesco and The Parthenon Publishing Group, 1998. p. 289 –314. ´ Poncy O, Sabatier D, Prevost MF, Hardy I. The lowland high rainforest: structure and tree species diversity. In: Bongers ´ F, Charles-Dominique P, Forget PM, Thery M, editors. Nouragues: Dynamics and Plant–Animal Interactions in a Neotropical Rainforest. Biological Monographs series, Werger series ed., Kluwer, Dordrecht, Netherlands; 2001. pp. 31–46. ´ Prevost MF, Sabatier D. Variations spatiales de la richesse et ˆ Guyanaise. de la diversite´ du peuplement arbore´ en foret ´ In: Guillaumet JL, Belin M, Puig H, editors. Phytogeogra´ ´ et Perspectives, Actes du Coll. Intern. phie tropicale, Realites ´ Phytogeographie tropicale, Paris 6-8 juillet 1993, Ed. ´ ORSTOM, Colloques et Seminaires; 1996. pp. 263–277. ´ ` Preiss N, Melieres MA, Pourchet M. Data base on lead-210 concentration in surface air, lead-210 atmospheric deposition and water-sediment flux. Available on Webserver, http:yy www-lgge.ujf-grenoble.fryequipesyglaciersyDonneesDispy
267
datayfluatm.html, Laboratoire de Glaciologie et ´ ` Geophysique de l’Environnement, St Martin d’Heres. Proctor J. In: Sutton SL, Whitmore TC, Chadwick AC, editors. Tropical rainforest: ecology and management. Oxford: Blackwell, 1983. p. 267 –273. Proctor J. Tropical forest litter fall. II. The data set. In: Chadwick AC, Sutton SL, editors. Tropical rainforest: the leeds symposium. Leeds: Leeds Philosophical and Literary Soc. L.P.D, 1984. p. 83 –113. Rankin-de Merona JM, Prance GT, Silva MF, Rodrigues WA, Uehling M. Preliminary results of a large-scale tree inventory of upland rain forest in the central Amazon. Acta Amazonica 1992;22:493 –534. Rea AW, Lindberg SE, Keeler GJ. Developement of a washing technique for measuring dry deposition of mercury and selected trace elements to foliage and surrogate surfaces. Environ Sci Technol 2000;34:2418 –2425. Rea AW, Lindberg SE, Keeler GJ. Dry deposition and foliar leaching of mercury and selected trace elements in deciduous forest throughfall. Atm Environ 2001;35:3453 –3462. Rea AW, Lindberg SE, Scherbatskoy T, Keller GJ. Mercury accumulation in foliage over time in two northern mixedhardwood forests. Water Air Soil Poll 2002;133:49 –67. Roulet M, Lucotte M, Farella N, Serique G, Coelho H, Passos CJ, De Jesus da Silva E, de Andrade PS, Mergler D, Guimaraes JR, Amorin M. Effects of recent human colonization on the presence of mercury in Amazonia ecosystems. Water Air Soil Poll 1999;112(3–4):297 –313. Schroeder WH, Munthe J. Atmospheric mercury—an overview. Atm Environ 1998;32:809 –822. Spichiger R, Loizeau PA, Latour C, Barriera G. Candollea 1996;51:559 –577. Valencia R, Balslev H, Paz y Mino G. High tree-alpha-diversity in Amazonian Ecuador. Biodiv Conserv 1994;3:21 –28.
Mercury in canopy leaves of French Guiana in remote areas ´ ` a,*, M. Pourcheta, P. Charles-Dominiqueb, P. Gaucherc Marie-Antoinette Melieres a
´ ` Laboratoire de Glaciologie et Geophysique de l’Environnement, CNRS, Universite´ J. Fourier, BP 96, 38402 St Martin d’Heres, France b ´ ´ ˆ Laboratoire d’Ecologie Generale, MNHN, 4 rue du Petit Chateau, 91800 Brunoy, France c ´ Mission pour la Creation du Parc de la Guyane, BP 275, 97326 Cayenne, France Received 29 May 2002; accepted 8 February 2003
Abstract A study of total Hg concentration in the foliage of the canopy was carried out in two remote areas in French Guiana. The sampled canopy is representative of the French Guiana canopy. The concentration in the foliage, 64"14 ng gy1 (dry wt.), is used to estimate the annual input of total Hg to the soil through the litterfall, found to be 45"10 mg my2 yy1. As translocation is negligible, mercury in the canopy originates mainly from atmospheric uptake by the leaves and this litterfall deposit represents a direct atmospheric input from the background atmospheric load into the soil. 䊚 2003 Elsevier Science B.V. All rights reserved. Keywords: Mercury; Canopy; Rainforest; Hg atmosphere soil transfer
1. Introduction As in the Amazonian basin of Brazil, large areas of French Guiana (South America) have been subjected to gold mining since the 19th century. This activity uses large quantities of mercury (Hg) (1–4 kg Hg for 1 kg of gold produced), which are mainly lost in soil, river sediment and the atmosphere. In order to estimate, understand, and finally prevent the resultant Hg pollution, studying the Hg cycle has been a major goal of the last 20 years (see for example Schroeder W., 1998). To estimate the anthropogenic impact we must first *Corresponding author. Tel.: q33-4-76-82-42-11; fax: q 33-4-76-82-42-01. E-mail address: [email protected] ´ ` (M.-A. Melieres).
evaluate the contribution of the two natural origins of the Hg present in soil: geological soil content and atmospheric deposition. Atmospheric uptake is an important process by which Hg accumulates in plants (see for example Lindberg et al., 1992, 1994; Munthe et al., 1995; Benesh et al., 2001). We focus here on the contribution of the background atmospheric deposit, measuring specifically the contribution from the leaves of the canopy through litterfall deposit. Total Hg concentration in leaves has been measured by Roulet et al. (1999) in different parts of Brazil and in French Guiana; results have shown a higher content, close to double, at the French Guiana site (mean values in French Guiana ranging from 130 to 150 ng gy1, compared to 40–120 ng gy1 in Brazil). Is this high content representa-
0048-9697/03/$ - see front matter 䊚 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0048-9697(03)00142-6
262
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267
tive of the entire French Guiana canopy in remote areas? Our goal here is to establish a representative value for this total Hg deposit in litterfall for the tropical rainforest of French Guiana. This requires first sampling representative species in the canopy, and second, performing this sampling in remote areas free from gold mining Hg pollution and from human settlement, to detect the contribution from the background atmospheric level. Previous estimates of Hg content in the tropical forest canopy have been made but not systematically. 2. Sampling and measurements 2.1. Study area Two sampling sites were chosen. The first is the field station of ‘Les Nouragues’, the French CNRS observatory in a remote area (48059 N, 528419 W) situated 100 km away from the oceanic coast at an altitude of 50 m above sea level (asl). The site is under regular trade wind influence throughout the year with a prevailing northeast direction. This atmospheric circulation involves the first 1000 m of the atmosphere (the low troposphere), an air layer which carries most of the atmospheric load. This load in terms of gaseous compounds and aerosol is therefore of Atlantic origin and is representative of background levels. This has been tested at the site of Petit Saut in French Guiana ´ ` (Melieres et al., 2003) through a continuous monitoring of the aerosol load using the 210Pb radionuclide, a major tracer of aerosol transport and deposit. Monitoring began in July, 1999 and still continues. This site is under the influence of air masses of similar origin to the sites of Les Nouragues and Pic Matecho; results show that the 210 Pb air concentration at Petit Saut is close to 0.2 ´ ` Bq my3 during both wet and dry seasons (Melieres et al., 2003), a value which is characteristic of an oceanic air mass in tropical regions, and similar to what give data for others studies (see the data base on lead-210 by Preiss et al., 2000). The trajectory of this oceanic air mass arriving at Les Nouragues precludes any large influence from the vast surrounding Amazonian regions where gold mining could load the atmosphere with Hg vapor. The only gold mining activities which exist close
to the station are at ‘la Montagne Tortue’ (30 km north), and at Boulanger (60 km north-east). These places have restricted activity, and the atmospheric influence of Hg vapor released is assumed to be negligible over such a distance. The second site is situated in the center of French Guiana (03843944.50 N, 53802918.40 W), 3 km south of Pic Matecho (590 m asl), at an altitude close to 250 m. This appears to be an ideal pristine site: no gold mining existed in this area, and there has been no human settlement during the 20th century. It is 150 km south of the Atlantic coast, 30 km north-east of the small ¨ (f200 inhabitants). As at the village of Saul Nouragues station the air masses arriving from north-east do not cross any gold mining areas and come directly from the Atlantic; they can therefore be considered to be under the influence only of background atmospheric load. Thus at both sites atmospheric Hg can originate only from the natural background atmospheric load. 2.2. Sampling At Les Nouragues, a large area has been equipped with five observation platforms in the upper canopy, between 35 and 40 m above the ground. This area has been extensively studied for 15 years (ecology, vegetation dynamic, biodiversity,«). Canopy walkways, with a total length of 108 m, connect the different observation platforms. The upper canopies of trees belonging to six different species have been sampled. Foliage was collected by selecting fully developed undamaged leaves (between 1 and 3 years old), cutting the leaves at the stalk. The leaves were placed directly into plastic bags, where they were stored. Foliage samples were collected on the emergents of three trees on which the platforms had been installed: Sapotaceae on the first platform, Ampelozisyphus americanus (Rhamnaceae) on the second, and Quaribea sp. (Bombacaceae) on the third. The last three samples, Licania sp. (Chrisobalanaceae), Streculia sp. (Sterculiaceae) and Vouacapoua americana (Caealpiniaceae), were collected on trees whose canopies are between platforms 1 and 2. All these trees are situated on a hill (altitude 60–90 m) dominated by an ‘inselberg’ (a granite
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267
mountain of altitude 430 m) and are representative of high forest on drained soil, a type of forest well represented in French Guiana. The sampled trees form the high level of the forest, between 30 and 35 m above the ground (Poncy et al., 1998). At the Pic Matecho site, the tree canopies of five different species were sampled in July, 2001. These trees were cut at the edge of a recent natural treefall; we sampled the trees at the same height as at the other site (upper canopy). Trees were 30–35 m high, on drained soil. The forest of this region is well known for its high biodiversity. The collected species were Chrisobalanaceae, Eperua falcata, Liane, Pourouma sp. and Caraipa sp. (Clusiaceae). At Les Nouragues, collection was performed in 1999 and, for each tree, sampling was duplicated on two different leaves. In the Pic Matecho field campaign in 2001, 3–5 samplings of different leaves of the same tree were carried out in order to estimate more precisely the range of Hg concentration. In order to be able to consider the measurements at those two sites to be representative of French Guiana forest we have to address two questions: Is the sampled forest at each site representative of the tropical rainforest in French Guiana? The inventory at the Nouragues station indicates density per hectare of 64–125 trees with diameter above 30 cm, and 468–680 of diameter above 10 cm; 550 tree species corresponding to 63 families have been censused at Les Nouragues (Poncy et al., 1998, 2001). None of them is dominant. At the very most 20 of them (including Vouacapoua americana) are more frequently found. The basal area, which is well correlated to the total biomass, is between 30 and 45 m2 hay1. This is only slightly higher than the Amazonian forest (e.g. Valencia et al., 1994; Rankin-de Merona et al., 1992; Spichiger et al., 1996) but less than the value of 53 ¨ m2 hay1 found in the very beautiful forest of Saul, in central French Guiana, well known for its rich biodiversity (Mori et al., 1983). Pic Matecho, our second site, is situated in this region. All 11 specimens are common species and, taken together, constitute approximately 10% of the forested bio´ mass (Prevost and Sabatier, 1996). They can therefore be considered as representative of the
263
French Guiana forest and of the mean operating cycle of the tropical rainforest. 3. Analyses and results After each campaign, the collected leaves were dried following the same procedure as used for sediments (2 days in an oven at 60 8C). This procedure does not appear to affect the total Hg content in sediments. Dried leaves were then sealed in plastic bags and sent to the Centre d’Analyse du CNRS (Vernaison, France). The total Hg concentration was measured using the cold vapor technique with atomic absorption spectrometry detection (Ama 254, Spectra France). The dried leaves were directly placed in a cell and heated to 750 8C and the Hg vapor content determined. The limit of analysis is 0.5 ng of Hg (the detection limit is ten times less), and the relative standard deviation is a few percent for 10 ng of Hg. The estimated measurement error is 0.1 ng. Mercury measurements were performed on the blade of the leaves and, in two cases, in the central nervation of the leaf. Results are presented in Table 1. In this table, the different measurements for each species correspond to different leaves on the same tree. 4. Discussion 4.1. Mean leaf concentration At Les Nouragues, the total Hg concentration ranges from 32.4 to 114 ng gy1, whereas at Pic Matecho concentration ranges from 52.4 to 103.0 ng gy1. For a given species, dispersion appears to be small whereas leaves can be of different ages. Both measurements in the central nerve indicate that Hg concentration is systematically lower (5– 10 times lower) than in the blade. The mean values obtained at Les Nouragues (59"28 ngHg gy1) and at Pic Matecho (70"16 ngHg gy1) cannot be considered to be significantly different. This leads to assume an overall mean concentration in the canopy of 64"14 ng gy1 (with a confidence level of 95%), for the tropical rainforest of French Guiana.
264
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267
Table 1 Total Hg concentration in the blade of the leaves and in the central nerve of the leaves (given in brackets) Nouragues
Pic Matecho
Family
Dry matter (mg)
Total Hg (ng gy1)
Family
Dry matter (mg)
Total Hg (ng gy1)
Sapotaceae
106.9 105.7
36.7"1 35.7"1
Chrisobalanaceae
92.4 127.1 108.1 109.1 108.0 (43.2)
59.0"1 45.1"0.8 50.8"1.2 40.0"0.9 57"0.8 (13.3"2.2)
Licania sp.
30.5 63.4
40.7"3.6 46.6"1.7
Eperua falcata
47.8 56.7 50.2 54.0 87.8
77.7"2.1 52.4"2.1 82"1.9 58.6"1.7 81.0"1.1
Sterculiaceae
42.8 91.2
65.7"2.7 59.7"1.2
Caraipa sp.
80.7 122.2 87.0 (79.7)
80.7"1.2 75.6"0.8 81.4"1.1 (8.1"1.3)
Strychnos sp.
108.7 147.2
68.4"1.3 62.5"0.9
Liane
65.3 88.3 73.2
65.2"1.5 53.1"1.1 88.3"1.3
Vouacapoua americana
90.4 62.7
32.4"1.3 37.9"1.8
Pourouma sp.
94.4 63.0 80.5 93.0
84.8"1 82.9"1.6 91.6"1.2 103.0"1
Bombacaceae
56 94.8
114"1.9 105"1.1
4.2. Comparison with other measurements Total Hg concentration in leaves has been measured by Roulet et al. (1999) in different parts of Brazil (along the Tapajos Basin and in the Tocantins Basin) and at three sites in French Guiana (f80 km south of Cayenne), and results have shown a higher content (mean value 142"50 ng gy1) at the French Guiana site than in Brazil (mean value of 73"40 ng gy1). Our results indicate a lower value (55% smaller) for the remote areas of French Guiana. Moreover, the measurements performed by Roulet et al. (1999) suggest a Hg content in the foliage similar in the remote areas of Brazil (the Tapajos and the Tocantins Basins) and in the remote areas of French Guiana. Few measurements exist in Europe in remote areas. Iverfeld (1991) measured the pine needle concen-
tration in remote areas of Sweden and found systematically lower values (40 ng gy1). Other measurements in spruce needles performed in Scandinavia (see Goldbold, 1994, for a review article) show similar values in the range of 40–50 ng gy1. 4.3. Mean contribution of canopy to Hg content in soil The French Guiana forest is sempervirent. Most tropical trees lose their leaves continuously, after a mean time of 4 years, in an asynchronous manner (Hadlik and Blanc, 1987). The litter deposit on soil has been estimated in tropical forest in different parts of the world. Results are quite homogeneous and the annual deposit is of the order of 7 tons of dry matter per hectare (Proctor, 1983,
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267
1984; Leight, 1978). A more accurate estimate is in progress at Les Nouragues. The annual estimated litter deposit of 7 ton hay1, together with the mean concentration in the leaves of 64"14 ngHg gy1, leads to an annual mean deposit on soil of 45"10 mgHg my2 due to litterfall. 4.4. Origin of Hg in leaf canopy There are two possible origins of the Hg present in canopy leaves: atmospheric uptake by foliage, andyor translocation (i.e. transfer of Hg from soil to leaves through the roots of the tree). This second process is usually believed to be negligible (see e.g., Gilmour and Miller, 1973; Beaufort et al., ˆ 1977; Lindberg et al., 1979; Goldbold and Huttermann, 1988; Graydon et al., 2001; Rea et al., 2002), leaving Hg atmospheric load as the main contributor to Hg concentration in leaves. The Hg content of the atmosphere at the canopy level can be influenced by the soil emission. Literature suggests that this latter source is much less important than the global atmospheric load in unpolluted areas. Therefore, the Hg concentrations measured in the canopy foliage originate mainly from the atmosphere which, at these two sites, is representative of the background level. This leads to an estimate for the background atmospheric contribution to the soil via litterfall of 45"10 mgHg my2 in the French Guiana rainforest. 4.5. Atmosphere–soil transfer In tropical rainforests, the Hg present in the atmosphere can be deposited on the soil mainly via 3 processes: rainfall, throughfall and litterfall deposits. In a forested area the rain lands on the soil after passing through the canopy and becomes enriched in Hg (throughfall deposit); during this transfer, the wash-off of dry deposited Hg compounds from the leaf surface appears to be the most important source of Hg, rather than foliage leaching (Rea et al., 2000, 2001). The third type of atmosphere–soil transfer operates through the foliage depositionyuptake, followed by the litterfall deposit. Measurements performed over two growing seasons on foliar Hg concentration indicate
265
that atmospheric Hg accumulates during the growing season in living foliage (Rea et al., 2002). This may be illustrated by the 5–10 times lower concentration that we found in the central nervation of the leaf (Table 1), compared to the concentration found in the blade of the leaves (living cells). Absorbed Hg may be released by evaporation andyor leaching but these two mechanisms are shown to be of minor importance (Lodenius et al., 2001). Different studies have focused on these mechanisms. For example, Iverfeld (1991) estimates these three components separately at the site of ¨ Sweden. The measured components are Gardsjon, total Hg rainfall deposit (12 mg my2 yeary1 ), and throughfall deposit (17–19 mg my2 yeary1); litterfall deposit was estimated (25 mg my2 yeary1). This clearly shows the amplifying role of the foliage in transmitting the atmospheric mercury to the soil: the total Hg deposit through rainfall deposit would be 12 mg my2 yeary1 on unforested soil, but close to 40 mg my2 yeary1 (throughfallq litterfall) in forested areas. Here the throughfall represents 72% of the litterfall. A similar ratio was found in the review article by Schroeder and Munthe (1998) on mercury deposition in precipitation, throughfall and litterfall in forested ecosystems of boreal forest in Sweden (Iverfeld, 1991; Munthe et al., 1995) and of temperate forests in USA (Lindberg, 1996). When averaged separately, rainfall, throughfall and litterfall are, respectively, 10, 15 and 20 mg my2 yeary1. Here the throughfall represents 75% of the litterfall: the deposition in such forested areas (throughfallqlitterfall) is significantly greater than deposition in precipitation only (rainfall). A recent study estimating through fall and measuring litterfall in deciduous forests in USA (Rea et al., 2002) leads to a comparable order of magnitude. Here, in the tropical rainforest of French Guiana, the litterfall deposit mechanism, alone, represents ;45"10 mg my2 yeary1 of input to soil. This is therefore a low estimate of the atmospheric contribution to soil, as throughfall deposit (not measured here) should also be included. The fact that temperate (deciduous) and boreal (evergreen) forests display a similar ratio of litterfall to throughfall, is a good indication that this
266
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267
ratio can be extended to tropical forests. If the same ratio, (0.7), is assumed, the Hg flux due to throughfall may reach 30 mg my2 yeary1, thus yielding a total deposition flux of 75 mg my2 yeary1. 5. Conclusion The purpose of this study was to estimate the canopy contribution to the Hg present in the soils of the French Guiana tropical rainforest through the litterfall deposit only. This canopy contribution originates mainly from the atmospheric background load. A study of Hg concentrations in the foliage of the canopy was carried out in two remote areas where gold mining has not occurred and which are free from any other anthropogenic influences. The sampling (11 species) of the canopy is representative of the French Guiana canopy. The mean average concentration found in the foliage was 64"14 ng gy1 (dry wt.). Using an annual litterfall deposit of 7 ton hay1, a value typical of tropical rainforests, these results lead us to estimate the annual input of total Hg to the soil through the litterfall to be ;45"10 mg my2 yeary1. As translocation is believed to be negligible, mercury in the canopy originates mainly from atmospheric uptake by the leaves. This measured flux, therefore, represents a direct atmospheric input into the soil; in fact it represents the lower limit of this atmospheric deposit, as deposition due to throughfall must also be included in the total atmospheric deposit to soil in a forested area. This mechanism for transferring Hg from atmosphere to soil contributes partly to the Hg content in of remote areas in the French Guiana tropical rainforest. Acknowledgments We are very grateful to S. Mori for helpful comments. This project was supported by The Mission du Parc de la Guyane, the Programme ` de ‘Mercure en Guyane’ (INSU-CNRS, Ministere l’Environnement, Europe), and the University of Grenoble.
References Benesh JA, Gustin MS, Schorran DE, Coleman J, Johnson DA, Lindberg SE. Determining the role of plants in the biogeochemical cycling of mercury on an ecosystem level. Sixth International Conference on mercury as a global pollutant, Minamata, Japan; 2001. p. 40. Beaufort W, Barber J, Barringer AR. Uptake and distribution of mercury within higher plants. Physiol Plant 1977;39:261 – 265. Gilmour JT, Miller MS. Fate of a mercuric-mercurous chloride fungicide added turfgrass. J Environ Qual 1973;2:145. Goldbold DL. Mercury in forest ecosystems: risk and research needs. In: Watras CJ, et al, editor. Mercury pollution. London: Lewis publishers, 1994. p. 295 –303. ˆ Goldbold DL, Huttermann A. Inhibition of photosynthesis and transpiration in relation to mercury-induced root damage in spruce seedlings. Physiol Plantarum 1988;74:270 –275. Graydon JA, St Louis VL, Lindberg SE, Hintelman H. Mechanisms of mercury deposition under boreal forest canopies. Sixth International Conference on mercury as a global pollutant, Minamata, Japan; 2001 p. 41. Hadlik A, Blanc P. Croissance des plantes en sous-bois de ˆ dense humide (Makokou, Gabon). Rev Ecol (Terre et foret Vie) 1987;42:209 –234. Iverfeld A. Mercury in forest canopy throughfall water and its relation to atmospheric deposition. Water Air Soil Poll 1991;56:553 –564. Leight EG. Leaf production, leaf consumption, and the regulation of folivory on Barro Colorado Island. In: Montgomery GG, editor. The ecology of arboreal folivores. Washington: Smithsonian Institution Press, 1978. p. 33 –50. Lindberg SE, Jackson DR, Huckabee JW, Janzen SA, Levin MJ, Lund JR. Atmospheric emission and plant uptake of mercury from agricultural soils near the Almaden mercury mine. J Environ Qual 1979;8:572 –578. Lindber SE, Meyers TP, Taylor GE, Turner RR, Schroeder WH. Atmosphereysurface exchange of mercury in a forest: results of modeling and gradient approaches. J Geophys Res 1992;97:2519 –2528. Lindberg SE. Forests and the global biogeochemical cycle of mercury. In: Baeyens W, Ebinghaus R, Vasiliev O, editors. Regional and global mercury cycles: sources, fluxes and mass balances. Dordrecht, Netherland: Kluwer, 1996. p. 359 –380. Lindberg SE, Owens JG, Stratton WJ. Application of throughfall methods to estimate dry deposition of mercury. In: Watras CJ, et al, editor. Mercury Pollution. London: Lewis Publishers, 1994. p. 261 –271. Lodenius M, Tulisalo E, Soltanpour-Gargari A. Exchange of mercury between atmosphere and vegetation. Sixth International Conference on Mercury as a Global Pollutant, Minamata, Japan; 2001 p. 40. ´ ` Melieres MA, Pourchet M, Richard S. Surface air Concentration and Deposition of Lead-210 in French Guiana: two years of Continuous monitoring. J Environ Radioact 2003;66:261 –269.
´ ` M.-A. Melieres et al. / The Science of the Total Environment 311 (2003) 261–267 Munthe J, Hultberg H, Iverfeldt A. Mechanismes of deposition of mercury and methyl mercury to coniferous forests. Water Air Soil Poll 1995;80:363 –371. Mori SA, Boom BM, de Carvalho AM, dos Santos TS. Southern Bahian moist forests. Bot Rev 1983;49:155 –232. ´ B, Larpin D, Belbenoit P, Jullien M, Hoff M, Poncy O, Riera Charles-Dominique P. The permanent field research station les Nouragues in the tropical rainforest of French Guiana: current projects and preliminary results on tree diversity, structure and dynamics. In: Dallmeier F, Comiskey JA, editors. Forest Biodiversity in North, Central and South America and the Caribbean: research and monitoring. Man and the biosphere series, vol. 20, 22. Carnforth, Lancashire, UK: Unesco and The Parthenon Publishing Group, 1998. p. 289 –314. ´ Poncy O, Sabatier D, Prevost MF, Hardy I. The lowland high rainforest: structure and tree species diversity. In: Bongers ´ F, Charles-Dominique P, Forget PM, Thery M, editors. Nouragues: Dynamics and Plant–Animal Interactions in a Neotropical Rainforest. Biological Monographs series, Werger series ed., Kluwer, Dordrecht, Netherlands; 2001. pp. 31–46. ´ Prevost MF, Sabatier D. Variations spatiales de la richesse et ˆ Guyanaise. de la diversite´ du peuplement arbore´ en foret ´ In: Guillaumet JL, Belin M, Puig H, editors. Phytogeogra´ ´ et Perspectives, Actes du Coll. Intern. phie tropicale, Realites ´ Phytogeographie tropicale, Paris 6-8 juillet 1993, Ed. ´ ORSTOM, Colloques et Seminaires; 1996. pp. 263–277. ´ ` Preiss N, Melieres MA, Pourchet M. Data base on lead-210 concentration in surface air, lead-210 atmospheric deposition and water-sediment flux. Available on Webserver, http:yy www-lgge.ujf-grenoble.fryequipesyglaciersyDonneesDispy
267
datayfluatm.html, Laboratoire de Glaciologie et ´ ` Geophysique de l’Environnement, St Martin d’Heres. Proctor J. In: Sutton SL, Whitmore TC, Chadwick AC, editors. Tropical rainforest: ecology and management. Oxford: Blackwell, 1983. p. 267 –273. Proctor J. Tropical forest litter fall. II. The data set. In: Chadwick AC, Sutton SL, editors. Tropical rainforest: the leeds symposium. Leeds: Leeds Philosophical and Literary Soc. L.P.D, 1984. p. 83 –113. Rankin-de Merona JM, Prance GT, Silva MF, Rodrigues WA, Uehling M. Preliminary results of a large-scale tree inventory of upland rain forest in the central Amazon. Acta Amazonica 1992;22:493 –534. Rea AW, Lindberg SE, Keeler GJ. Developement of a washing technique for measuring dry deposition of mercury and selected trace elements to foliage and surrogate surfaces. Environ Sci Technol 2000;34:2418 –2425. Rea AW, Lindberg SE, Keeler GJ. Dry deposition and foliar leaching of mercury and selected trace elements in deciduous forest throughfall. Atm Environ 2001;35:3453 –3462. Rea AW, Lindberg SE, Scherbatskoy T, Keller GJ. Mercury accumulation in foliage over time in two northern mixedhardwood forests. Water Air Soil Poll 2002;133:49 –67. Roulet M, Lucotte M, Farella N, Serique G, Coelho H, Passos CJ, De Jesus da Silva E, de Andrade PS, Mergler D, Guimaraes JR, Amorin M. Effects of recent human colonization on the presence of mercury in Amazonia ecosystems. Water Air Soil Poll 1999;112(3–4):297 –313. Schroeder WH, Munthe J. Atmospheric mercury—an overview. Atm Environ 1998;32:809 –822. Spichiger R, Loizeau PA, Latour C, Barriera G. Candollea 1996;51:559 –577. Valencia R, Balslev H, Paz y Mino G. High tree-alpha-diversity in Amazonian Ecuador. Biodiv Conserv 1994;3:21 –28.