Estudos da APA – Biblioteca Digital Itupararanga

Multi-proxy approach involving ultrahigh resolution mass spectrometry and self-organising maps to investigate the origin and quality of sedimentary organic matter across a subtropical reservoir

Este trabalho é somente para uso privado de atividades de pesquisa e ensino. Não é autorizada sua reprodução para quaisquer fins lucrativos. Esta reserva de direitos abrange a totalidade dos dados do documento bem como seu conteúdo. Na utilização ou citação de partes do documento é obrigatório mencionar o nome da pessoa autora do trabalho.

Abordagem multiproxy envolvendo espectrometria de massa de ultra-alta resolução e mapas auto-organizados para investigar a origem e a qualidade da matéria orgânica sedimentar em um reservatório subtropical

ABSTRACT

Humic substances (HS) in sediments play an important role in carbon and nutrient biogeochemical cycles and fate of contaminants in the environment. However, information regarding HS quality and transformations that may affect their behaviour in reservoirs is still limited. The aim of this investigation was to track sources and changes in sedimentary HS across a subtropical reservoir, connecting them to in lake processes and land-use influences. Surface sediments were collected at seven sampling sites in Itupararanga Reservoir (Brazil).

Humic (HA) and fulvic (FA) acids (components of HS) were extracted from the sediment samples followed by in-depth characterisation via UV/VIS, fluorescence spectroscopy, elemental (C, N) and isotopic analysis (d13C, d15N), nuclear magnetic resonance (13C NMR) and Fouriertransform ion cyclotron resonance mass spectrometry (FT-ICR-MS). All data were analysed by selforganising maps. The results showed that samples from the upstream part of the reservoir were older and more decomposed. They likely originated from C3 land-plants (d13C varied from  26.3‰ to  22.4‰), having more aromatic, oxygen-poor (O/C < 0.5) and unsaturated compounds (H/C < 1.1).

In contrast, near-dam samples were younger and had larger contributions of autochthonous material. This was confirmed by oxygen-rich (O/C > 0.5) and partly more unsaturated compounds for FA as well as oxygenpoor and saturated compounds with H/C > 1.1 for HA. Self-organising maps pointed out these differences between upstream and dam areas and indicated that agriculture lands were related to microbiallyderived HS. Changes in HS composition revealed that internal reservoir processes may have influenced HS quality across the reservoir.

Keywords: Kohonen neural network, 13C NMR, FT-ICR-MS, Stable isotopes, Humic substances, Source discrimination

RESUMO

Substâncias húmicas (HS) em sedimentos desempenham um papel importante nos ciclos biogeoquímicos de carbono e nutrientes e no destino de contaminantes no ambiente. No entanto, informações sobre a qualidade e as transformações de HS que podem afetar seu comportamento em reservatórios ainda são limitadas. O objetivo desta investigação foi rastrear fontes e mudanças em HS sedimentar em um reservatório subtropical, conectando-as a processos no lago e influências do uso da terra.

Sedimentos superficiais foram coletados em sete locais de amostragem no reservatório de Itupararanga (Brasil). Ácidos húmicos (HA) e fúlvicos (FA) (componentes de HS) foram extraídos das amostras de sedimento, seguidos por caracterização em profundidade via UV/VIS, espectroscopia de fluorescência, análise elementar (C, N) e isotópica (δ 13 C, δ 15 N), ressonância magnética nuclear ( RMN de 13 C ) e espectrometria de massas por ressonância cíclotron de íons com transformada de Fourier (FT-ICR-MS). Todos os dados foram analisados ​​por mapas auto-organizáveis. Os resultados mostraram que as amostras da parte a montante do reservatório eram mais antigas e mais decompostas.

Elas provavelmente se originaram de plantas terrestres C3 (δ 13 C variou de −26,3‰ a −22,4‰), tendo mais compostos aromáticos, pobres em oxigênio (O/C < 0,5) e insaturados (H/C < 1,1). Em contraste, as amostras próximas à barragem eram mais jovens e tinham maiores contribuições de material autóctone. Isso foi confirmado por compostos ricos em oxigênio (O/C > 0,5) e parcialmente mais insaturados para FA, bem como compostos pobres em oxigênio e saturados com H/C > 1,1 para HA. ​

Mapas auto-organizáveis ​​apontaram essas diferenças entre as áreas a montante e da barragem e indicaram que as terras agrícolas estavam relacionadas a HS derivados de micróbios. Mudanças na composição de HS revelaram que os processos internos do reservatório podem ter influenciado a qualidade de HS em todo o reservatório.

Introduction

Sediments in lakes and reservoirs store a substantial amount of organic matter (OM) and play an important role in the carbon cycle (Chen and Hur, 2015, Guillemette et al., 2017). This OM records processes and conditions from their surrounding catchments and its characterisation provides valuable information about sources and sinks of organic carbon (Torres et al., 2012).

A significant amount of OM in sediments is composed of humic substances (HS) (Chen and Hur, 2015). They are a mixture of organic acids of different molecular sizes with complex structures and mostly derive from decomposition of plants and animal biomass (Raposo et al., 2016). HS can be divided according to their solubility at different pH ranges in fulvic acids (FA, soluble at all pH), humic acids (HA, insoluble at pH below 2) and humin (insoluble at all pH). These substances are also responsible to store refractory organic carbon in sediments and complex to organic pollutants, nutrients, and trace metals, changing their mobility and bioavailability (Nguyen, 1999, Elkins and Nelson, 2002, Li et al., 2018).

The composition and quality of HS depend on their source, which can be allochthonous or autochthonous (Hur et al., 2014, Derrien et al., 2018). While allochthonous material is composed predominantly of lignin and aromatic structures derived from surrounding soils, trees and grasses, autochthonous material is produced from in-lake processes including leachates of dead organisms, macrophytes and phytoplankton.

The latter has a more aliphatic character with higher amounts of H and N in its molecules (McDonald et al., 2004, Derrien et al., 2017). The relative contribution of these two sources at each location depends on local and seasonal patterns and can affect OM quality and reactivity (Guillemette et al., 2017). Therefore, investigations of composition and structures of HS can help to understand sources and processes taking place in sediments (e.g. loss of aromatic compounds derived from Fe reduction) (Lovley et al., 1989, Roden et al., 2010, Kulkarni et al., 2018) and their influence on water quality, besides their role on carbon and nutrient cycling (de la Rosa et al., 2011).

Due to the complex characteristics of HS, a combination of several techniques is required to study their composition and structure. Each technique provides complementary pieces of information about the source, structure, and reactivity of HS (de la Rosa et al., 2011).

A broad range of methods can be used for HS characterisation, which include elemental composition, fluorescence spectroscopy, UV/VIS absorption, nuclear magnetic resonance (NMR), stable isotopes and recently Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS) (Hertkorn et al., 2013, Minor et al., 2014, Melendez-Perez et al., 2016, Herzsprung et al., 2017b). FT-ICR-MS combined with electrospray ionisation (ESI) provides detailed information on the molecular composition of HS typically resolving several thousand different molecular masses in HS. It enables molecular formula assignment to constituents of complex mixtures including non-fluorescent compounds of OM (Koch and Dittmar, 2006, Zherebker et al., 2016, Derrien et al., 2018).

The acquisition of large volumes of data by several techniques requires tools that simplify the interpretation of the results with synoptic visualisation capabilities. Artificial neural networks such as self-organising maps meet these needs because they can cluster and recognise patterns in complex data sets. This technique has been successfully used for discriminating organic matter fluorescence components (Ejarque-Gonzalez and Butturini, 2014), the evolution of fluorescent dissolved organic matter in treatment plants (Yang et al., 2019), water quality and identification of specific biogeochemical processes in waters (desorption of Fe, Mn and nutrients from sediments and algal growth) (Çinar and Merdun, 2009, Melo et al., 2019).

Past investigations have combined different techniques (e.g. 13C NMR, fluorescence, pyrolysis gas chromatography–mass spectrometry and stable isotopes) to study OM from sediments (Wolfe et al., 2002, de la Rosa et al., 2011, Guillemette et al., 2017, Li et al., 2018). However, only a few of them included FT-ICR-MS to characterise FA and HA and to track their sources (Dadi et al., 2017, Valle et al., 2018). Therefore, further studies about the potential of FT-ICR-MS to discriminate the sources of sedimentary OM are promising for getting a better picture of local and global carbon cycles (Derrien et al., 2018).

The main purpose of this study is to track sources, molecular composition and structural information of sedimentary FA and HA along a tropical reservoir with absorbance, fluorescence spectroscopy, stable isotopes and to combine them with 13C NMR, FT-ICR-MS and self-organising maps.

Possible changes in the HS were then investigated with the aim to identify in-lake processes and land-use influences around the reservoir. A novel focus on FT-ICR-MS intended to discriminate molecular changes in sedimentary HS and the combination of this information with self-organising maps is promising to provide integral information of all data collected. This multi-proxy study was promising to yield new information about refractory organic matter in sediments to identify possible biogeochemical processes including oxygen consumption and redox reactions. These processes may also influence the water quality of reservoirs that are important for drinking and irrigation (Meyers and Ishiwatari, 1993).

Section snippets

Study site and sampling

Itupararanga Reservoir is located in Sorocaba River catchment in the state of São Paulo, southeast of Brazil, with a storage volume of 286 million m3, a catchment area of about 934 km2 and a reservoir area of 19 km2. Its length is 26 km and maximum and average depths are 21 and 7.8 m, respectively (Ribeiro et al., 2014, Frascareli et al., 2018). The average water residence time in the reservoir is 250 days (Frascareli et al., 2015, Rodrigues et al., 2019). It was constructed in 1912 originally…

Particle size and organic matter content

The particle size distribution of the samples showed higher proportion of sand (particle size between 63 and 2000 µm) in the surface sediments at P1 (59.5%). This sand proportion decreased within the samples towards the dam (13.6% at P7). In contrast, silt (particle size between 2 and 63 µm) and clay (particle size < 2 µm) proportion increased within samples towards the dam (Fig. A.2, Appendix A). The OM content in the surface sediments analysed varied between 4.1 (at P1) to 17.2% weight (at…

Sedimentary fulvic and humic acids

The differences between FA and HA were dominant when compared to the internal differences among FA samples and HA samples (Fig. A.6). This trend was expected due to sample preparation (extraction procedure) and the typical distinctions between FA and HA. The FA samples for instance presented smaller average molecular size, higher content of oxygen compounds (CHO) and higher C:N ratio than HA. The 13C NMR data showed that the larger oxygen content of FA may be related especially to the greater…

Conclusions

Our multi-proxy approach in combination with self-organising maps pointed out significant differences between sedimentary FA and HA from the dam zone when compared to upper reaches of the Itupararanga Reservoir. In summary, samples from areas in the upstream reaches of the reservoir were older, more decomposed and had a larger contribution of land plants as precursor material. Furthermore, these samples had larger presence of oxygen-poor compounds in FA and HA and more unsaturated compounds…

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This investigation was supported by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) and Deutscher Akademischer Austauschdienst (DAAD) through the project “Organic carbon cycling in water reservoirs – ORCWAR” (DAAD-ID 57414997; CAPES 99999.008107/2015-07, 88887.141964/2017-00 and 88887.303733/2018-00). We also acknowledge the financial assistance of Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP, grant number 18/20326‐1) and Conselho Nacional de Desenvolvimento

References (88)

Há mais referências disponíveis na versão completa deste artigo.