Abstract
This thesis investigated the application of spectroscopic methods and other methods involving the measurement of electromagnetic radiation for detecting and characterising stress in apple trees (Malus domestica Borkh.). Contemporary agriculture faces the conflicting goals of maintaining or increasing productivity while simultaneously reducing environmental impact and resource consumption. This is especially pertinent in regions like South Tyrol, Italy, where apple production, while achieving near-potential yields through sophisticated management practices, faces mounting pressure to reduce reliance on agrochemicals and to adapt to climate change. Through a series of experiments conducted between 2021-2024, this research demonstrated that analysing the optical properties of leaves using spectroradiometry combined with chemometric and machine learning data analyses can be used to effectively detect various biotic and abiotic stresses. Waterlogging, chemical inhibition of photosynthesis and apple scab could be detected before the appearance of symptoms. Preliminary results suggest that optical reflectance, thermal infrared and chlorophyll fluorescence measurements, in combination, provide good capacity to detect water deficit stress and estimate plant water status even when plants are also exposed to extreme heat. The work focused heavily on the detection of infections by ˈCa. Phytoplasma maliˈ which is a bacterial pathogen associated with the highly important disease, Throughout the work, key wavelength regions and spectral vegetation indices (SVIs) important for stress detection were identified. Regardless of the type of stress that was being investigated similar regions were found to be important. This included the UV region (365 nm) related to anthocyanin content, the visible region around 531 nm related to xanthophyll epoxidation, 550 nm and 680-700 nm related to chlorophyll content, and near-infrared regions (1240-2500 nm) for leaf water content. Selected SVIs related to the same physiological phenomena could often perform as well as full-spectrum analysis, suggesting potential for simpler and more cost-effective measurement systems. While the spectroradiometers used in this work are too expensive for commercial use, the identification of key wavelengths and SVIs provides a foundation for the development simpler, more affordable sensors that could be applied more readily within practical settings. This could enable more targeted and timely interventions in commercial orchards, potentially reducing the use of agrochemicals while managing crop health. Several important areas of future investigation are suggested, including the development of more affordable sensors and the investigation of remote sensing applications. In conclusion, this thesis provides comprehensive evidence that spectroscopic methods can detect and distinguish between various stresses in apple trees, often before visible symptoms appear. The work lays a solid foundation for the development of practical applications in commercial apple production, while also contributing to our understanding of how different stresses affect the optical properties of plants. The findings suggest that spectroscopic technology could play a crucial role in the future of precision agriculture, enabling more sustainable and efficient orchard management practices.