Current biodiversity decline requires monitoring in an efficient and consistent fashion. Remote sensing has the potential of accessing biological variation in ways that overcome the spatial limitations of ground-based measurements. Additionally, it offers the possibility of repeatedly acquiring data to obtain information about temporal changes of vegetation. Temporal trajectories of physiological and morphological changes may express genetic constraints visible in highly-resolved spectral data; therefore, assessing temporal trajectories of plant spectral responses may improve biodiversity monitoring at the level of genetic diversity within species.
We investigated temporal trajectories of spectral data of individual trees to derive the genetic structure of a temperate forest at Laegern in the Swiss midlands (47°28N, 8°21E). We constructed distance matrices between 68 canopy trees of Fagus sylvatica based on the temporal spectral fingerprints and microsatellite information. The spectral fingerprints were derived from diurnal, seasonal and multi-annual acquisitions using the APEX (Airborne Prism Experiment) imaging spectrometer, covering a wavelength range between 350–2500 nm. Genetic data were derived from five microsatellite markers. We applied Mantel tests to assess the significance of correlations between genetic distances and distances calculated using different temporal trajectories across different spectral bands to identify spectral signals expressing the highest genetic constraints.
Preliminary Results/Conclusions
Multitemporal spectral responses reflect dynamic physiological, morphological and structural processes of tree canopies specific to diurnal, seasonal and multi-annual time scales. We observed positive diurnal trends in spectral reflectance and higher absorption in the spectral band of 1250–1450 nm and at 1750 nm during noon. On a seasonal basis, reflectance of 350–750 nm and 1450–1750 nm spectral regions showed a pronounced drop during the peak of growing season. The correlations between genetic distances and distances derived from spectral temporal trajectories indicated genetic constraints specific for the three temporal scales. The correlation between genetic and spectral distance matrices of the investigated trees reached 0.5 for diurnal and seasonal spectral data. We unraveled relatively higher correlations between genetic distances and diurnal reflectance distances in the spectral region of 2000–2500 nm as well as seasonal reflectance distances in the spectral region of 400–750 nm and at 1750 nm.
The observed spectral variation over time represent phenotypic expressions of individual trees in the studied beech population that are more similar for genetically more related individuals. Our results demonstrate that highly resolved imaging-spectrometer data have the potential to allow remote monitoring of genetic diversity at intraspecific level.