Análisis de factores que influyen en el secuestro de carbono en los suelos y su relación con la composición molecular de la materia orgánica

Tesis doctoral inédita leída en la Universidad Autónoma de Madrid. Facultad de Ciencias. Departamento de Geología y Geoquímica. Fecha de lectura: 22-11-2019The progress of desertification, which is very pronounced in the Mediterranean area, and the emission of greenhouse gases into the atmosphere, especially CO2, are taking the attention of the researchers. The establishment of the factors involved in these processes as well as the development of emergent technologies to solve these issues is a main objective in different scientific fields. In this perspective, soil conservation plays an important role, due to the content of soil organic matter (SOM) and its stability are important factors. Biogeochemical processes involved in the stabilization of soil organic carbon (SOC) are being the subject of study in this field. Some studies focus on the organo-mineral interactions, while others are interested in the relationship between the molecular structure or the SOM and its resilience. This thesis deals with the molecular characterization of SOM accumulated in different types of soil to establish its relationship with the potential for carbon storage in the corresponding soils, as well as the factors with a bearing on SOM quality. For this purpose, 35 soils with high variability in their organic carbon content (17–157 g·kg-1) have been selected. A detailed characterization of the organic matter has been carried out using destructive and non-destructive techniques such as analytical pyrolysis (Py-GC/MS), 13C solid-state nuclear magnetic resonance (NMR), infrared (IR) spectroscopy, visible spectroscopy and Fourier transform ion cyclotron resonance mass spectrometry (FTICR-MS). In particular, the attention has been focused on certain types of biomarker compounds that could act as environmental indicators of soil biogeochemical processes. The families of alkanes and methoxyphenols were analyzed in detail; its molecular composition was used to distinguish between microbial synthesis and transformation of plant biomass. This study examines the utility of the Shannon diversity index (H’), calculated from the abundance of alkanes analyzed by analytical pyrolysis (C9–C31) from whole soil samples, for the evaluation of the C storage potential and quality of SOM. A series of multivariate data treatments showed a significant relationship between the H' diversity of alkanes and the concentration of SOC. In particular, a significant relationship was found between SOC levels and the percentage of long-chain alkanes, while the percentage of short-chain alkanes was correlated with specific descriptors of SOM quality. Finally, partial least squares regression (PLS) successfully predicted SOC content using exclusively the information provided by the alkane patterns. In a parallel study, the molecular assemblages of methoxyphenols released by analytical pyrolysis from whole soil samples were also examined using the Shannon diversity index to describe the complexity of their pyrolytic patterns. A series of exploratory statistical methods (linear regression, PLS, multidimensional scaling (MDS), etc.) were applied to analyse the relationships between pyrolysis products and the chemical and spectroscopic characteristics of SOM and with the total SOC content. These results showed significant correlations between the progressive molecular diversity of the pyrolytic methoxyphenols and the SOC levels in the corresponding soils. The fact that the diversity of the phenolic signature provides information about the potential for carbon storage in soils can be interpreted as the progressive structural complexity of plant macromolecules modified by soil microorganisms, which makes them more difficult to be recognized by enzymes.From a quantitative point of view, PLS regression models based exclusively on the total abundance of the 12 major methoxyphenols were especially effective in predicting carbon storage in the soil. After studying the information provided by analytical pyrolysis of SOM from different scenarios of carbon storage activity, the study was focused to the fraction traditionally considered most representative of the SOM, humic acids (HA). This fraction corresponds to a colloidal product of advanced transformation of plant and microbial biomass. The HA characterization was carried out using visible, IR and NMR spectroscopies. A PLS study using the intensities of digital IR spectra points (4000–400 cm-1) as descriptors showed that there is a relationship between IR spectral patterns and SOC content. This was also the case with E4 index (i.e., indicative of progressive humification, and based on the optical density of HAs at 465 nm). The use of principal component analysis (PCA) and MDS suggested that the bands assigned to carbonyl and amide groups were characteristic in HA of soils with low C content, while the spectra of HAs from soils with high levels of C showed a typical pattern of lignin bands, which indicates accumulation of less transformed plant residues. The IR spectral patterns were analyzed in detail by digital treatments including weighted subtraction of spectra obtained by averaging those of HAs from soils classified in the upper and lower quartiles according the SOC distribution, respectively, and calculating the statistical significance level of the differences. The results showed significant differences between the molecular composition of the HAs, according the SOC and E4 values. Peaks corresponding to aromatic, carboxyl and amide groups showed comparatively high intensity in HAs from soils with low SOC content, while peaks corresponding to lignin-derived structures were more marked in the spectra of the HAs from soils with high SOC content. In the second level of study of the HAs, its characterization was carried out using FTICR-MS. The application of PLS to FTICR-MS data showed a great potential for identifying the molecular components of HA that varies in terms of the SOC levels in different environments. A significant model to predict the SOC was obtained through PLS using as descriptors the 131 compounds detected by FTICR-MS in all HAs. In order to identify the compounds with the highest value as indicators of the SOC levels, the importance of the variables for prediction (VIP) was represented in the space defined by their atomic ratios using van Krevelen diagrams. The results showed a significant relationship between the molecular composition of HA and the C content stored in the soil: HA in soils with high levels of organic C displayed significantly higher proportions (P < 0.1) of molecular formulas corresponding to unsaturated lipids and lignin-derived compounds. On the other hand, low SOC levels were associated with comparatively higher proportions of saturated aliphatic structures.La realización del presente trabajo ha sido posible gracias al apoyo económico recibido a través del proyecto CGL2013-43845-P y a la concesión de una beca de Formación de Personal Investigador (FPI, BES-2014-069238) concedidos por el Ministerio de Economía y Competitividad