Potential and Limitations of Grasslands ?-Diversity Prediction Using Fine-Scale Hyperspectral Imagery

Plant biodiversity is an important feature of grassland ecosystems, as it related to the provision many ecosystem services crucial for human economy and well-being. Given importance grasslands, research has been carried out in recent years on potential monitor them with novel remote sensing techniques. In this study, optical diversity (also called spectral diversity) approach was adopted check using high-resolution hyperspectral images estimate ?-diversity ecosystems. 2018 2019, species composition surveyed canopy data were acquired at two sites: Monte Bondone (IT-MBo; species-rich semi-natural grasslands) experimental farm University Padova, Legnaro, Padua, Italy (IT-PD; artificially established plots a species-poor mixture). The relationship between (species richness, Shannon’s, evenness, Simpson’s indices) metrics (coefficient variation-CV standard deviation-SD) not consistent across investigated plant communities. Species richness could be estimated by R = 0.87 IT-PD site. more complex grasslands IT-MBo, estimation indices difficult failed accurately probably due higher number strong spatial heterogeneity. Therefore, results study confirmed ability proxies detect man-made ecosystems but highlighted limitations when natural are observed. Nevertheless, metric SD calculated from post-processed transformed spectra showed, red part spectrum, significant correlation (up 0.56, p 0.004) indices. Spatial resampling that sward optimal pixel size 1 cm, while IT-MBo mm. random extraction did improve performance both sites. Further needed fully understand links biochemical heterogeneity heterogeneous assess whether can scale ?-diversity. Such insights will provide robust information mechanisms linking