“Flower power”: how flowering affects spectral diversity metrics and their relationship with plant diversity

Biodiversity monitoring is constrained by cost- and labour-intensive field sampling methods. Increasing evidence suggests that remotely sensed spectral diversity (SD) is linked to plant diversity, holding promise for monitoring applications. However, studies testing such a relationship reported conflicting findings, especially in challenging ecosystems such as grasslands, due to their high temporal dynamism and variety. It follows that a thorough investigation of the key factors, such as the metrics applied (i.e., continuous, categorical) and phenology (e.g., flowering), influencing such a relationship is necessary. Thus, this study aims to assess the applicability of SD for plant diversity monitoring at the local scale by testing six different SD metrics while considering the effect of the presence of flowering on the relationship and resampling the original data to assess how spatial resolution affects the results. Taxonomic diversity was calculated based on data collected in 159 plots with 1.5 m ×1.5 m experimental mesic grassland communities. Spectral information was collected using a UAV-borne sensor measuring reflectance across six bands in the visible and near-infrared range at ∼2 cm spatial resolution. Our results show that, in the presence of flowering, the relationship is significant and positive only when SD is calculated using categorical metrics. Despite the observed significance, the variance explained by the models had very low values, with no evident differences when resampling spectral data to coarser pixel sizes. Such findings suggest that new insights into the possible confounding effects on the SD∼plant diversity in grassland communities are needed to use SD for monitoring purposes. ### Competing Interest Statement The authors have declared no competing interest.