The Effect of Epidermal Structures on Leaf Spectral Signatures of Ice Plants (Aizoaceae)

Remote Sensing

René Hans-Jürgen Heim²  Norbert Jürgens²  André Gro෾-Stoltenberg¹  Jens Oldeland²  Eyal Ben-Dor³  Clement Atzberger³ 

[1]Institute of Landscape Ecology, Heisenbergstraße 2, 48149 Münster, Germany

[2]Biodiversity Ecology and Evolution of Plants, Biocenter Klein Flottbek and Botanical Garden, University of Hamburg, Ohnhorststraße 18, 22609 Hamburg, Germany

[3]

[4]Biodiversity Ecology and Evolution of Plants, Biocenter Klein Flottbek and Botanical Garden, University of Hamburg, Ohnhorststraße 18, 22609 Hamburg, Germany

关键词: Aizoaceae;    biophysical properties;    leaf epidermis;    field spectrometry;    hyperspectral classification;    leaf traits;    optical types;    plant functional types;    succulent plants;   

DOI  :  10.3390/rs71215862

来源: mdpi

【 摘 要 】

Epidermal structures (ES) of leaves are known to affect the functional properties and spectral responses. Spectral studies focused mostly on the effect of hairs or wax layers only. We studied a wider range of different ES and their impact on spectral properties. Additionally, we identified spectral regions that allow distinguishing different ES. We used a field spectrometer to measure ex situ leaf spectral responses from 350 nm–2500 nm. A spectral library for 25 species of the succulent family Aizoaceae was assembled. Five functional types were defined based on ES: flat epidermal cell surface, convex to papillary epidermal cell surface, bladder cells, hairs and wax cover. We tested the separability of ES using partial least squares discriminant analysis (PLS-DA) based on the spectral data. Subsequently, variable importance (VIP) was calculated to identify spectral regions relevant for discriminating our functional types (classes). Classification performance was high, with a kappa value of 0.9 indicating well-separable spectral classes. VIP calculations identified six spectral regions of increased importance for the classification. We confirmed and extended previous findings regarding the visible-near-infrared spectral region. Our experiments also confirmed that epidermal leaf traits can be classified due to clearly distinguishable spectral signatures across species and genera within the Aizoaceae.

【 授权许可】

CC BY   
© 2015 by the authors; licensee MDPI, Basel, Switzerland.