Mercury bioaccumulation and isotopic relation between Trichiurus lepturus (Teleostei) and its preferred prey in coastal waters of southeastern Brazil

An Acad Bras Cienc aabc Anais da Academia Brasileira de Ciências An. Acad. Bras. Ciênc. 0001-3765 1678-2690 Academia Brasileira de Ciências A transferência trófica de mercúrio total (THg) e sua bioacumulação das espécies de presas para o peixe predador Trichiurus lepturus foi analisada em águas costeiras do sudeste do Brasil para avaliar a dinâmica de elemento traço neste sistema predador-presa. A relação isotópica (δ15N and δ13C) entre o predador e suas presas permitiu inferências sobre a assimilação das presas e os hábitos alimentares do predador. O incremento THg variou de 4,5 a 19,5 vezes entre a presa e o predador, com uma força de biomagnificação de 0,59. As espécies de presas podem ser divididas em três grupos em relação aos valores de δ15N: i) 13,6 a 13,2‰ (juvenis co-específicos, Pellona harroweri e Peprilus paru); ii) 12,5 a 11,8‰ (Chirocentrodon bleekerianus, Lycengraulis grossidens e Dorytheuthis plei); e iii) 10,5‰ (Xiphopenaeus kroyeri). Com base nos valores δ13C, os grupos de presa foram: i) -15.3‰ (X. kroyeri); ii) -17,6 a -16,8‰ (C. bleekerianus, D. plei, P. harroweri, P. paru, e juvenis co-específicos); e iii) -18,7‰ (L. grossidens). Os valores de THg e δ15N destacaram os juvenis co-específicos como a principal via deste elemento traço e a presa mais assimilada. A relação isotópica entre o predador e suas espécies de presas demonstrou uma atividade alimentar preferencialmente costeira e pelágica. INTRODUCTION The species Trichiurus lepturus Linnaeus, 1758 is a teleost fish that forms shoals in brackish and marine waters along tropical and subtropical regions worldwide, with importance as fishery resource (Froese and Pauly 2015). During its ontogeny, there is a wide diet shift: juveniles are planktivores, while adults are top predators, feeding on the most abundant prey. In coastal waters of southeastern Brazil (~21-22°S), the feeding habits of adult specimens of T. lepturus was detailed by Bittar et al. (2008, 2012). Twelve-eight prey species were recorded for this predator, and seven were the most representative in its diet, in this order: juvenile conspecifics, which represented 33% of the diet; Pellona harroweri Fowler, 1917 (17%); Dorytheuthis plei Blainville, 1823 (13%); Chirocentrodon bleekerianus Poey, 1867 (11%); Xiphopenaeus kroyeri Heller, 1862 (11%); Lycengraulis grossidens Spix & Agassiz, 1829 (6%); and Peprilus paru Linnaeus, 1758 (3%). In general, prey identification and quantifi cation, and original size estimates are only possible by stomach content analysis, but it has limitations. Differences in prey digestion rates may lead to under- or overestimation of prey importance in predators' diet (Pierce and Boyle 1991). Many studies have applied multiple trophic markers together to minimize this bias, such as stomach content analysis, trace elements, and stable isotopes (Aubail et al. 2011, Di Beneditto et al. 2011, Kehrig et al. 2013, Connelly et al. 2014). Mercury is a trace element that undergoes biomagnification in animals' tissues through trophic transfer, i.e. its concentration increases over consecutive trophic levels (Lavoie et al. 2013). Then, this trace element is suitable as trophic marker. Isotopic ratios of animals' tissues provide records of the prey con tribution in the predator diet, indicating prey assi milation after digestion and excretion (Domi et al. 2005, Huckstadt et al. 2007). Furthermore, isotopic composition also indicates the contributions of different sources within a given trophic relation (e.g., aquatic versus terrestrial, coastal versus ocea nic, pelagic versus benthic) (Fry 2008). In this study, the trophic transfer of mercury and its bioaccumulation involving prey species and the fish T. lepturus was analysed to evaluate the trace element dynamic in this predator-prey system in southeastern Brazil. Furthermore, the isotopic relation between the predator and its prey allowed inferences on prey assimilation and feeding habits. MATERIALS AND METHODS Sampling The sampling of adult specimens of T. lepturus (body size >100 cm) was done between 21º35'S and 22º25'S. These specimens were targets of local commercial fisheries. The preferred prey were collected in the same area during local commercial fisheries (target or by-catch species), taking into account the prey size consumed by the predator (see Bittar et al. 2008, 2012 for details). A sample from the back dorso-lateral muscle (fish), mantle (squid), or abdomen (shrimp) removed from each specimen (predator and prey) was freeze-dried and homogenized with a mortar and pestle for total mercury and stable isotopes analyses. Total Mercury (THg) Determination The dried tissue samples (100 mg) were acid digested with 3 mL of H2SO4:HNO3 (1:1v/v) (Merck p.a.) and 1 mL of concentrated H2O2 (Merck p.a.) in a 50 mL centrifuge tube at 60 °C in water bath for 45 min. After addition of 5 mL of 5% KMnO4 (Merck p.a.) solution, the digested samples allowed to stand for overnight. The THg concentration in the acid digested solutions was determined by cold vapour atomic absorption spectrometry with a Flow Injection Mercury System (FIMS-400, Perkin Elmer) with auto sampler, using NaBH4 as reducing agent. The limit of detection for THg determination was 0.001 μg g-1. The results were expressed in ng g-1 (dry weight). Quality control involved replicates analysis, strict blank control, and certified reference material. The samples were analyzed in triplicate, and the coefficients of variation for analytical replicates were below 10%. In each ten triplicates, two blank controls were done to detect reagent impurities or external contamination signals. The accuracy was assessed through certified material DORM-2 (THg: 4.64±0.26 µg g-1) from the National Research Council Canada. The results for THg DORM-2 were 4.54±0.13 µg g-1, demonstrating high precision and accuracy of the analytical method. THg quantified in certified material was within 97% of the mean certified value. Isotopic Analysis of δ15N and δ13C Isotopic analyses of the dried tissue samples (0.5 mg) were completed in a ThermoQuest Finnigan Delta Plus (Finnigan MAT) mass spectrometer coupled to an elemental analyser. The reference values for δ15N and δ13C were atmospheric nitrogen and Pee Dee Belemnite (PDB), respectively, and the results were expressed in parts per thousand (‰). The analytical precision was ±0.3‰ for δ15N and ±0.2‰ for δ13C, determined by triplicates at each five samplings. The accuracy for elemental and isotopic composition were determined by a certified standard (Protein OAS/Isotope Cert 114859; Elemental Microanalysis), organic carbon and total nitrogen (99%), and δ13C and δ15N (100%). Lipids were not extracted from tissues prior to analysis; however, the C:N ratio in muscle samples from both predator and prey were lower or equal to 3.5, indicating low lipids levels. Therefore, the interpretation of δ13C was not compromised (Post et al. 2007). Data Analysis Differences among species regarding THg concen tration and isotopic ratios were analysed by Kruskal-Wallis test with Dunn's multiple comparison test a posteriori. A linear regression model was used to test the relation between δ15N on log-transformed THg concentrations, and the regression slope (b) represented the biomagnification power (Kidd et al. 1995). The analyses were performed in GraphPad Prism 5 for Windows, and p value <0.05 was chosen to indicate statistical significance. RESULTS The concentrations of THg were different among species (p< 0.0001). Higher values were registered in the predator T. lepturus (Table I). Taking into account the contribution of each prey species in the trophic transfer of THg to the predator, the juvenile conspecifics were the main via (Figure 1). Linear regression testing the relation between δ15N and THg was significant, showing the biomagnification of this trace element from prey to predator (b= 0.59). TABLE I - Body dimensions, total mercury concentration (THg), and isotopic ratios (δ15N and δ13C) of Trichiurus lepturus (adult specimens) and its preferred prey. Species N Body size (cm) Weight (g) THg ng g-1 (dry weight) δ15N(‰) δ13C(‰) Predator Trichiurus lepturus 20 142.0±11.0 2,326.0±517.0 1,440±931 14.8±0.4 -16.8±0.3 Prey T. lepturus (juvenile) 20 49.1±9.2 66.7±46.9 320±123 13.6±0.8 -17.0±0.4 Pellona harroweri 10 10.0±1.5 13.1±5.0 181±123 13.3±0.5 -17.3±0.8 Dorytheuthis plei 20 4.7±0.7 7.6±3.9 110±39 11.8±0.6 -17.1±0.5 Chirocentrodon bleekerianus 19 10.4±0.8 6.8±0.9 229±79 12.5±0.4 -17.2±0.3 Xiphopenaeus kroyeri 08 9.3±1.8 2.0±0.1 74±25 10.5±0.4 -15.3±0.4 Lycengraulis grossidens 15 11.4±1.1 16.3±4.0 178±45 12.3±0.4 -18.7±1.0 Peprilus paru 10 9.3±1.9 15.7±9.7 89±64 13.2±0.3 -17.6±0.3 Figure 1- Relationship between δ15N and THg concentration (log-transformed) for Trichiurus lepturus (adult specimens) and its preferred prey. Tl p: Trichiurus lepturus predator; Tl: T. lepturus prey; Cb: Chirocentrodon bleekerianus; Lg: Lycengraulis grossidens; Dp: Dorytheutis plei; Xk: Xiphopenaeus kroyeri; Pp: Peprilus paru; and Ph: Pellona harroweri. The δ15N and δ13C values differed among species (p< 0.05, and p< 0.0001, respectively). The predator presented the heavier δ15N values, and the lighter values were found in X. kroyeri. The carbon isotopic values were heavier in X. kroyeri and lighter in L. grossidens (Table I and Figure 2). According to average δ15N values, the prey species could be divided into three groups: i) 13.6 to 13.2‰ (juvenile conspecifics, P. harroweri, and P. paru); ii) 12.5 to 11.8‰ (C. bleekerianus, L. grossidens, and D. plei); and iii) 10.5‰ (X. kroyeri) (Figure 2). Based on average δ13C values, the prey species were also divided into three groups: i) -15.3‰ (X. kroyeri); ii) -17.6 to -16.8‰ (C. bleekerianus, D. plei, P. harroweri, P. paru, and juvenile conspecifics); and iii) -18.7‰ (L. grossidens) (Figure 2). Figure 2 - Relationship between δ13C and δ15N in Trichiurus lepturus (adult specimens) and its preferred prey. Tl p: Trichiurus lepturus predator; Tl: T. lepturus prey; Cb: Chirocentrodon bleekerianus; Lg: Lycengraulis grossidens; Dp: Dorytheutis plei; Xk: Xiphopenaeus kroyeri; Pp: Peprilus paru; and Ph: Pellona harroweri. Bars indicate the standard deviations. DISCUSSION The values of THg and δ15N, as well the relation between them highlighted juvenile conspecifics as the main via of this trace element to adult specimens of T. lepturus, and the most assimilated prey. Thus, cannibalism is confirming as an important feeding strategy to this predator in the study area, as previously recorded through stomach content analysis (Bittar et al. 2008, 2012). The THg increment varied from 4.5 to 19.5 times between prey and predator, revealing its biomagnification (Table I). Here, the biomagnifi cation power (b= 0.59) considered the increase of THg only from prey to predator, and not along the entire food chain. Jæger et al. (2009) argued that the increase of Hg from specific prey to predator is not an accurate measured of biomagnification. However, previous studies in the same area showed significant Hg transference in many trophic interactions, both predator-prey relations (Carvalho et al. 2008, Kehrig et al. 2009, Di Beneditto et al. 2011, 2013) as entire food chains (Di Beneditto et al. 2012, Kehrig et al. 2013), supporting our results. Di Beneditto et al. (2012) analysed data on biomagnification power of Hg in marine environments worldwide, and concluded no trend regarding latitude or water temperature. Therefore, the bioavailability of Hg in aquatic ecosystems should be the primary factor driving the magnitude of local biomagnification processes. The main source of Hg to coastal marine species in the study area is a river discharge. This area is permanently influenced by the Paraíba do Sul River (Souza et al. 2010), whose basin was widely impacted with practices of gold-mining and use of mercurial fungicides on plantations until 1980's (Lacerda et al. 1993). The nitrogen isotope values (δ15N) indicated the high trophic position of the predator relative to its prey, as expected (Fry 2008). The grouping of prey species (δ15N values) was not conclusive (Figure 2). These prey species are zooplanktivores (P. harroweri, L. grossidens, and P. paru) or carnivores whose prey have small size (juvenile conspecifics, D. plei, and C. bleekerianus) (Froese and Pauly 2015). Thus, their trophic positions would be similar. However, the prey species of T. lepturus varied in length and weight (and probably age) (Table I). Parameters as consumer class, size, and age may influence the isotopic ratios of a given specimen, either alone as combined (Jennings et al. 2002, Caut et al. 2009). For the shrimp X. kroyeri the feeding habit may explain the lower δ15N values, once this species is a benthic feeder whose main feeding resources are small invertebrates and sediment (Branco and Moritz-Júnior 2001). The isotopic values of δ13C were similar to those previously registered to marine coastal species in the study area (Di Beneditto et al. 2011, 2012, Kehrig et al. 2013), and showed that most prey species is pelagic. Heavier δ13C values in the shrimp X. kroyeri reflected its benthic habit (Branco and Moritz-Júnior 2001), and lighter values in fish L. grossidens indicated its anadromous habit, with seasonal movements between fluvial and marine areas for reproduction (Froese and Pauly 2015). Although δ13C is not usually applied to distinguish trophic levels, enrichment around ≤1‰ is generally expected from one trophic level to another (Peterson and Fry 1987). Therefore, the enrichment between prey and predator is within the expected interval. In the study area, the biomagnification process of Hg due to trophic transfer between the prey species and the predator T. lepturus is evident. The isotopic relation between them showed a feeding activity preferably coastal and pelagic. In marine tropical waters (our study area) the set of available prey is generally wide, and the model fitted may not represent the real feeding preference of a given predator when only one trophic marker is used. Therefore, the use of multiple trophic markers as complementary approaches tends to reduce bias, providing more reliable results and improving knowledge regarding predator-prey systems. ACKNOWLEDGMENTS We thank the fishermen from Atafona Harbour and technician Silvana Ribeiro Gomes who provided us with predator and prey specimens to this study. This study was supported by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, grant nº 301.405/13-1); Fundação Carlos Chagas de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ, grant nº E-26/201.161/2014); and CNPq INCT Material Transference from the Continent to the Ocean (grant nº 573.601/08-9). Aubail A, Teilman J, Dietz R, Rigét F, Harkonen T, Karlsson O, Rosing-Asvid A and Caurant Florence. 2011. Investigation of mercury concentrations in fur of phocid seals using stable isotopes as tracers of trophic levels and geographical regions. Polar Biol 34: 1411-1420. Aubail A Teilman J Dietz R Rigét F Harkonen T Karlsson O Rosing-Asvid A Caurant Florence 2011 Investigation of mercury concentrations in fur of phocid seals using stable isotopes as tracers of trophic levels and geographical regions Polar Biol 34 1411 1420 Bittar VT, Awabdi DR, Tonini Wct, Vidal Júnior MV and Di Beneditto Apm. 2012. Feeding preference of adult females of ribbonfish Trichiurus lepturus L.1758 through prey proximate-composition and caloric values. Neotrop Ichthyol 10: 193-203. Bittar VT Awabdi DR Tonini Wct Vidal Júnior MV Di Beneditto Apm 2012 Feeding preference of adult females of ribbonfish Trichiurus lepturus L.1758 through prey proximate-composition and caloric values Neotrop Ichthyol 10 193 203 Bittar VT, Castello Bfl and Di Beneditto Apm. 2008. Hábito alimentar do peixe-espada adulto, Trichiurus lepturus, na costa norte do Rio de Janeiro, sudeste do Brasil. Biotemas 21: 83-90. Bittar VT Castello Bfl Di Beneditto Apm 2008 Hábito alimentar do peixe-espada adulto, Trichiurus lepturus, na costa norte do Rio de Janeiro, sudeste do Brasil Biotemas 21 83 90 Branco JO and Moritz-Júnior HC. 2001. Alimentação natural do camarão sete-barbas Xiphopenaeus kroyeri (Heller) (Crustacea, Decapoda), na Armação do Itapocoroy, Penha, Santa Catarina. Rev Bras Zool 18: 53-61. Branco JO Moritz-Júnior HC 2001 Alimentação natural do camarão sete-barbas Xiphopenaeus kroyeri (Heller) (Crustacea, Decapoda), na Armação do Itapocoroy, Penha, Santa Catarina Rev Bras Zool 18 53 61 Carvalho Cev, Di Beneditto Apm, Souza Cmm, Ramos Rma and Rezende CE. 2008. Heavy metal distribution in two cetacean species from Rio de Janeiro State, south-eastern Brazil. J Mar Biol Assoc UK 88: 1117-1120. Carvalho Cev Di Beneditto Apm Souza Cmm Ramos Rma Rezende CE 2008 Heavy metal distribution in two cetacean species from Rio de Janeiro State, south-eastern Brazil J Mar Biol Assoc UK 88 1117 1120 Caut S, Angulo E and Courchamp F. 2009. Variation in discrimination factors (d15N and d13C): the effect of diet isotopic values and applications for diet reconstruction. J Appl Ecol 46: 443-453. Caut S Angulo E Courchamp F 2009 Variation in discrimination factors (d15N and d13C): the effect of diet isotopic values and applications for diet reconstruction J Appl Ecol 46 443 453 Connelly TL, Deibel D and Parrish CC. 2014. Trophic interactions in the benthic boundary layer of the Beaufort Sea shelf, Arctic Ocean: combining bulk stable isotope and fatty acid. Prog Oceanogr 120: 79-92. Connelly TL Deibel D Parrish CC 2014 Trophic interactions in the benthic boundary layer of the Beaufort Sea shelf, Arctic Ocean: combining bulk stable isotope and fatty acid Prog Oceanogr 120 79 92 Di Beneditto Apm, Bittar VT, Camargo PB, Rezende CE and Kehrig HA. 2012. Mercury and nitrogen isotope in a marine species from a tropical coastal food web. Arch Environ Contam Toxicol 62: 264-271. Di Beneditto Apm Bittar VT Camargo PB Rezende CE Kehrig HA 2012 Mercury and nitrogen isotope in a marine species from a tropical coastal food web Arch Environ Contam Toxicol 62 264 271 Di Beneditto Apm, Bittar VT, Rezende CE, Camargo PB and Kehrig HA. 2013. Mercury and stable isotopes (15N and 13C) as tracers during the ontogeny of Trichiurus lepturus. Neotrop Ichthyol 11: 211-216. Di Beneditto Apm Bittar VT Rezende CE Camargo PB Kehrig HA 2013 Mercury and stable isotopes(15N and13C) as tracers during the ontogeny of Trichiurus lepturus Neotrop Ichthyol 11 211 216 Di Beneditto Apm, Souza Cmm, Kehrig H and Rezende CE. 2011. Use of multiple tools to assess the fee ding preference of coastal dolphins. Mar Biol 158: 2209-2217. Di Beneditto Apm Souza Cmm Kehrig H Rezende CE 2011 Use of multiple tools to assess the fee ding preference of coastal dolphins Mar Biol 158 2209 2217 Domi N, Bouquegneau JM and Das K. 2005. Feeding ecology of five commercial shark species of the Celtic Sea through stable isotope and trace metal analysis. Mar Environ Res 60: 551-569. Domi N Bouquegneau JM Das K 2005 Feeding ecology of five commercial shark species of the Celtic Sea through stable isotope and trace metal analysis Mar Environ Res 60 551 569 Froese R and Pauly D. 2015. FishBase. www.fishbase.org (Accessed on February 15, 2015). Froese R Pauly D 2015 FishBase www.fishbase.org (Accessed on February 15 2015) Fry B. 2008. Stable Isotope Ecology, 1st ed., New York: Sprin ger Science Business Media, 316 p. Fry B 2008 Stable Isotope Ecology Sprin ger Science Business Media New York 316 p Huckstadt LA, Rojas CP and Antezana T. 2007. Stable isotopes analysis reveals pelagic foranging by the Southern sea lion in central Chile. J Experim Mar Biol Ecol 347: 123-133. Huckstadt LA Rojas CP Antezana T 2007 Stable isotopes analysis reveals pelagic foranging by the Southern sea lion in central Chile J Experim Mar Biol Ecol 347 123 133 Jæger I, Hop H and Gabrielsen GW. 2009. Biomagnifi cation of mercury in selected species from an Arctic marine food web in Svalbard. Sci Total Environ 407: 4744-4751. Jæger I Hop H Gabrielsen GW 2009 Biomagnifi cation of mercury in selected species from an Arctic marine food web in Svalbard Sci Total Environ 407 4744 4751 Jennings S, Pinnegar JK, Nicholas VC and Waarr KJ. 2002. Linking size-based and trophic analyses of benthic community structure. Mar Ecol Prog Ser 226: 77-85. Jennings S Pinnegar JK Nicholas VC Waarr KJ 2002 Linking size-based and trophic analyses of benthic community structure Mar Ecol Prog Ser 226 77 85 Kehrig HA, Fernandes Kwg, Malm O, Seixas TG, Di Beneditto Apm and Souza Cmm. 2009. Trophic transference of mercury and selenium in the Northern Coast of Rio de Janeiro. Quim Nova 32: 1822-1828. Kehrig HA Fernandes Kwg Malm O Seixas TG Di Beneditto Apm Souza Cmm 2009 Trophic transference of mercury and selenium in the Northern Coast of Rio de Janeiro Quim Nova 32 1822 1828 Kehrig HA, Seixas TG, Malm O, Di Beneditto Apm and Rezende CE. 2013. Mercury and selenium bio magnification in a Brazilian coastal food web using nitrogen stable isotope analysis: A case study in an area under the influence of the Paraíba do Sul River plume. Mar Poll Bull 75: 283-290. Kehrig HA Seixas TG Malm O Di Beneditto Apm Rezende CE 2013 Mercury and selenium bio magnification in a Brazilian coastal food web using nitrogen stable isotope analysis: A case study in an area under the influence of the Paraíba do Sul River plume Mar Poll Bull 75 283 290 Kidd KA, Hesslein RH, Fudge RJ and Hallard KA. 1995. The influence of trophic level as measured by d15N on mercury concentrations in freshwater organisms. Water Air Soil Pollut 80: 1011-1015. Kidd KA Hesslein RH Fudge RJ Hallard KA 1995 The influence of trophic level as measured by d15N on mercury concentrations in freshwater organisms Water Air Soil Pollut 80 1011 1015 Lacerda LD, Carvalho Cev, Rezende CE and Pfeiffer WC. 1993. Mercury in sediments from the Paraíba do Sul River continental shelf, SE Brazil. Mar Poll Bull 26: 220-222. Lacerda LD Carvalho Cev Rezende CE Pfeiffer WC 1993 Mercury in sediments from the Paraíba do Sul River continental shelf, SE Brazil Mar Poll Bull 26 220 222 Lavoie R, Jardine TD, Chumchal MM, Kidd KA and Campbell LM. 2013. Biomagnification of mercury in aquatic food webs: a worldwide meta-analysis. Environ Sci Technol 47: 13385-13394. Lavoie R Jardine TD Chumchal MM Kidd KA Campbell LM 2013 Biomagnification of mercury in aquatic food webs: a worldwide meta-analysis Environ Sci Technol 47 13385 13394 Peterson BJ and Fry B. 1987. Stable isotopes in ecosystem studies. Annu Rev Ecol Evol Syst 18: 293-320. Peterson BJ Fry B 1987 Stable isotopes in ecosystem studies Annu Rev Ecol Evol Syst 18 293 320 Pierce GJ and Boyle PR. 1991. A review of methods for diet analysis in piscivorous marine mammals. Oceanogr Mar Biol - An Annual Review 29: 409-486. Pierce GJ Boyle PR 1991 A review of methods for diet analysis in piscivorous marine mammals Oceanogr Mar Biol - An Annual Review 29 409 486 Post DM, Layman CA, Arrington DA, Takimoto G, Quattrochi J and Montaña CG. 2007. Getting to the fat of the matter: models, methods and assumptions for dealing with lipids in stable isotope analyses. Oecologia 152: 179-189. Post DM Layman CA Arrington DA Takimoto G Quattrochi J Montaña CG 2007 Getting to the fat of the matter: models, methods and assumptions for dealing with lipids in stable isotope analyses Oecologia 152 179 189 Souza TA, Godoy JM, Godoy Mldp, Moreira I, Carvalho ZL, Salomão Msmb and Rezende CE. 2010. Use of multitracers for the study of water mixing in the Paraíba do Sul River estuary. J Environ Radioact 101: 564-570. Souza TA Godoy JM Godoy Mldp Moreira I Carvalho ZL Salomão Msmb Rezende CE 2010 Use of multitracers for the study of water mixing in the Paraíba do Sul River estuary J Environ Radioact 101 564 570