Format: ORAL

Rykkar Jackson1, Dominic Wood2, Jordan Wilson-Morrison2, Subbaiah Mechanda2, Shan Kothari3, Warren Cardinal-McTeague1

1 Department of Forest & Conservation Sciences, University of British Columbia, Vancouver, Canada 2 Ottawa Research & Development Centre, Agriculture & Agri-Food Canada, Ottawa, Canada 3 Département des sciences biologiques, Université de Québec à Montréal, Montréal, Canada

The estimation of leaf functional traits using leaf spectroscopy is a novel and underutilized technique in biological collections. With ~400 million plant specimens being held in ~3,500 herbaria worldwide today, there is massive potential for the use of spectroscopy on dried leaf samples. Once models are developed to quantify the limits of leaf spectroscopy on degraded herbarium specimens, such as those which have been stored for decades or centuries, we will better know the extent to which we can use herbaria to estimate leaf functional traits through time. This research expands the application of collections-based leaf spectroscopy by quantifying the limits of specimen degradation in leaf spectral analyses. To do so, this project utilizes spectra from both fresh and dried leaves of Ginkgo biloba. Representative specimens were collected in 2022 from a hedgerow and pressed, dried for 48 hours, and assembled into the four treatments: one in a standard herbarium cabinet (control conditions) and three in growth chambers (18C, 21C, 24C) to accelerate degradation. At this time, initial fresh-leaf spectral readings were conducted. In 2023, after 1 year within these treatments, the corresponding dried leaf spectra (350-2400 nm) were measured. This data will be used to analyze the effects of degradation on spectral analysis across treatment types by building an empirical model using multivariate techniques to estimate leaf functional traits via the use of least squares regression (PLSR). This research is among the earliest studies using PLSR modelling on intact dried leaves like those in herbaria, as previous models have been trained using the reflectance spectra of either fresh or ground, dried, but not intact, dried leaves.