Evaluación agro-ambiental del olivar basada en técnicas espectroscópicas

Resumen

In the present PhD, the study of the agro-environmental system of the olive grove is approached, paying special attention to the problem of soil degradation. In addition, other related aspects have been considered such as the nutritional status of the olive tree and the use of organic amendments to improve soil fertility. For this type of studies, a large number of laboratory analyses are usually required. As an alternative to traditional wet chemical analysis, the possibility of using infrared spectroscopy techniques, in both middle (FTIR) and near infrared regions (FT-NIR), has been evaluated. In general, infrared spectra provide a large amount of information on the mineralogical composition of the samples and, to a lesser extent, on the organic matter. The strong overlapping between bands of inorganic compounds makes usually difficult the interpretation of the bands of organic compounds, especially in cases like this, where soils are poor in organic matter. Both techniques provide similar information, which may be in some way complementary. The use of FTIR and FT-NIR is presented as a good alternative for the characterization of olive groves, as well as the possible effects derived from soil management and lithological material. The use of chemometrics tools is mandatory to get useful information from a high number of spectra. First, a principal component analysis was used to study possible groupings among soil samples based on their infrared spectra. The PCA analysis together with discriminant analysis (PLS-DA) have been proved very useful to find patterns of grouping, as well as their source. In addition, these tools are able to differentiate between soil types and, to a lesser extent, between the soils management. The combined use of both regions of the infrared spectrum is the best option for the discrimination of soils according to their quality. On the one hand, it has been demonstrated, the ability of infrared spectroscopy to classify the soil samples according to their geological material or according to their management, and, on the other hand, that the effects derived from soil management are reflected in a remarkable way in both regions of the spectrum infrared. The importance of the systems of soil management and the lithological material in the soils quality has been also demonstrated. In this way, more respectful management of the agro-system such as organic management gives rise to soils with better properties, compared to the soils under conventional management. In addition, the effects of irrigation over soil properties have been studied, demonstrating that the continued use of low quality water causes salinization and sodification of the soil, with the consequent loss of fertility. However, the effect of irrigation and management depends on the geological material, since more favorable substratum such as colluvial limestones better support degradation processes compared to marls, while carbonated soils are better able to withstand sodification processes than siliceous soils. Furthermore, we used the spectra to predict some soil properties by using a partial least squares regression (PLS). We confirmed that infrared spectroscopy has a great ability to predict physical-chemical parameters of soils, with excellent results especially for those parameters that have a direct reflection in the spectra as OC, N and CaCO3, but also for parameters in certain point correlated with some of its bands, such as pH and CEC, closely related to the amount of carbonates and clays respectively, or P and K related in part to organic matter. However, the prediction of textural or biological parameters was more difficult because there is no direct influence in the spectra. We also found the great capacity of the infrared spectroscopy to predict a global biological soil quality index (GMea), that could be useful for a quickly characterization of the soil. Although both regions show predictive capacity, the FT-NIR spectra produce better results. For a prediction as accurate as possible, the ideal would be using both regions together, especially with the application of mid level data fusion strategies. Hence, these techniques are a great alternative to typical laboratory analysis, being much faster and cheaper. However, it is very important to take into account that not all parameters were predicted with the same accuracy. Furthermore, the calibrations that have been developed are only applicable to similar samples, and that the use of this type of models to other different soil samples have to be explored. Therefore, for its use as a substitute for the laboratory, it would be necessary to increase the number of samples as well as to use wider concentration ranges or even to develop individual prediction models depending on the origin of the sample. For a complete assessment of olive agro-system, we also need to characterize the plant nutritional status. This is currently made by leave nutrient analysis, which is usually expensive and time consuming. We predicted these nutrients by using two spectroscopic techniques, FT-NIR and energy dispersive X-ray fluorescence (EDXRF). To use both spectra working together, we also applied two different data fusion strategies, low and mid level. FT-NIR and EDXRF are capable of predicting the majority of essential nutrients for the olive tree in the leaf, and in this way, they could partially replace the traditional laboratory analysis, although these models are not totally accurate for some micronutrients. On the other hand, the data fusion strategy seems to be useful to combine spectroscopic techniques, especially at the mid level, that built the best possible models, although with slight improvement comparing to individual models. Finally, we studied the possibility of using organic amendments to restore soil organic matter level, as a possible solution for the degradation of the olive agro-system. First, the effect of the addition of spent coffee grounds (SCG) over soils was studied. SCG has a clear effect on the increase of soil organic matter, in all its fractions, which in the short term generate a clear improvement in soil fertility, especially in more clayey soils like the Vega soil, that is able to stabilize more amount of organic matter. However, the analysis of humic acids shows that the functionality of organic matter is reduced, due to a decrease of its complexity and stability, so a continued input of this amendment could negatively affect the ability of the organic matter to provide nutrients to the plant in the long term. To solve this problem, we tried the mixing of different amendments as an alternative, mixing SCG with olive mill pomace compost (COMP), in this way correcting the problem of organic matter quality, but also the acid pH of SCG and the high C/N of this particular COMP. By another side, we studied characteristics of different olive mill pomace with and without composting process. We conclude that the composting of olive mill pomace is a totally necessary process to improve this amendment, although even after this, it continues to present problems such as high water repellency. This problem was found in all the amendments studied and in order to eliminate this harmful effect, a thermal treatment was applied. We found that thermal treatment, especially at 275ºC, is effective to eliminate negative aspects such as toxicity and water repellency, while generating an amendment with a higher quality organic matter, with a greater proportion of more stable organic material. On the other hand, the mixture of amendments of SCG and COMP is demonstrated as a viable alternative, since its mixture manages to correct some non-optimal parameters such as pH or C/N ratio, resulting in amendments of quality improved with intermediate parameters between both. Furthermore, this amendment after thermal treatment becomes as an alternative of equal or higher quality in all aspects than the COMP and SCG amendments separately.