Mapping multiple plant species abundance patterns - A multiobjective optimization procedure for combining reflectance spectroscopy and species ordination

摘要: Abstract Nature conservation and ecological restoration crucially depends on the knowledge about spatial patterns of plant species that control habitat conversion disturbance regimes. Especially, abundances are capable indicating early development tendencies for setting management strategies. This study demonstrates transfer field spectroscopy to hyperspectral imagery map multiple in an open dryland area using two imaging spectrometers different phenological phases. We show can partially be described by gradients forming coordinates a contour map. For this purpose, were projected into ordination space non-metric multidimensional scaling subsequent interpolation. It was demonstrated modeled Partial Least Squares regression framework resulting distinct spectral features certain gradient directions. combine both objectives multiobjective NSGA-II procedure maximize quantitative determination abundance predictability related spectra, simultaneously. finally used select optimal models n = 35 single transferred mapping purpose. predictabilities evaluated basis individual model performances hold each designated phase. Finally, we present spatially explicit multi-species maps best n = 18 n = 8 could linked richness, coexistence, succession stages type conditions.