Jacobaea through the eyes of spectroscopy : identifying plant interactions with the (a)biotic environment by chemical variation effects on spectral reflectance patterns

Plants interact with a wide array of aboveground and belowground herbivores, pathogens, mutualists, and their natural enemies. These interactions are important drivers of spatio-temporal changes in vegetation, however, they may be difficult to be revealed without extensive sampling.In this thesis I investigated the potential of visible and near-infrared spectral measurements to detect plant chemical changes that may reflect interactions between plants and biotic or abiotic soil factors. First, I examined the relative contribution of pyrrolizidine alkaloids (PAs; these are defence compounds of Senecio-type plants against generalist herbivores) to the spectral reflectance features in the visible and short-wave infrared region. My hypothesis was that PAs can be predicted from specific spectral features of aboveground plant tissues. Since PA profiles and their relation to spectral features could be species specific I compared three different species, <em>Jacobaea vulgaris</em>, <em>J. erucifolia</em> and <em>S. inaequidens</em> subjected to nutrient and water treatments to stimulate plant chemical variation. Pyrrolizidine alkaloids were predicted best by spectral reflectance features in the case of <em>Jacobaea vulgaris. </em>I related the better results obtained with <em>J. vulgaris </em>to the existence of the correlation between PAs and nitrogen and the presence of the epoxide chemical structure in <em>J. vulgaris</em>. I also examined if different soil microbial communities influenced plant shoot spectral reflectance. I grew the same three plant species as before in sterilized soil and living soil collected from fields with <em>J. vulgaris</em>. I expected that soil biota would change shoot defence content and hyperspectral reflectance in plant species-specific ways. Indeed, the exposure to different soils caused plant chemical profiles to change and both chemical and spectral patterns discriminated plants according to the soil biotic conditions. I studied how primary and secondary plant metabolites varied during the growing season and vegetation successional stages. I used a well-studied chronosequence of abandoned arable fields and analysed the chemistry of both leaves and flowers of <em>Jacobaea vulgaris </em>throughout the seasons in fields of different successional status. My general hypothesis was that seasonal allocation of nutrients and defence metabolites to reproductive organs fitted the optimal defence theory, but that pattern was dependent on the successional stage of the vegetation. I found an interaction between season and succession stage, as plants from longer abandoned fields generally had flowers and leaves with higher N-oxides, especially in late Summer. Independent of the succession stage there was a seasonal allocation of nutrients and defence metabolites to flowers. Analyses of spectral reflectance of the field plants showed thatdefence compounds could be estimated more reliably in flowers, while in leaves primary compounds could be predicted best. Succession classes were successfully discriminated by the spectral patterns of flowers. Both chemical and spectral findings suggested that flowers are more sensitive to field ageing processes than leaves. <strong>Conclusions</strong> The estimation of pyrrolizidine alkaloids by spectral reflectance features was better in <em>Jacobaea vulgaris</em> than in <em>Senecio inaequidens </em>or <em>Jacobaea erucifolia </em>(chapter 2). Differences in soil communities affect plant leaves’ chemistry and spectral reflectance patterns (chapter 3). <em>Jacobaea vulgaris</em>plants from recent and longer-abandoned fields showed the largest differences in chemical concentration, composition of defence compounds, and spectral reflectance patterns. Flowers were more discriminatory than leaves (chapters 4 and 5). There is a potential to detect plant-biotic interactions by analyzing spectral reflectance patterns (this thesis).