Affordable Access

Publisher Website

Utility of spectral vegetation indices for estimation of light conversion efficiency in coniferous forests in Japan

Agricultural and Forest Meteorology
Publication Date
DOI: 10.1016/j.agrformet.2007.11.006
  • Deciduous Conifer
  • Evergreen Conifer
  • Light Use Efficiency (Lue) Model
  • Mixed Stand
  • Remote Sensing
  • Physics


Abstract To clarify the utility of spectral vegetation indices (VIs) for estimating light conversion efficiency ( ɛ) in Japanese coniferous forests, we investigated the relationships between six VIs (NDVI, EVI, SAVI, PRI, CI, and CCI) and ɛ in two mature monospecific forests of deciduous conifer (Japanese larch) and evergreen conifer (Japanese cypress) and one young mixed stand of deciduous conifer with evergreen undergrowth (hybrid larch and dwarf bamboo). In each forest canopy, we measured seasonal variations in CO 2 flux, radiation environment, and visible–near-infrared spectral reflectance during 1 or 2 growing seasons. We calculated ɛ as gross primary production (GPP) divided by the difference between incoming and reflected photosynthetically active radiation (PAR). VIs and ɛ under clear skies were averaged between 11:00 and 13:00 JST and their relationships were analyzed. In the larch forest, all calculated VIs were positively correlated with ɛ, and the highest correlation was that with CCI. Because of effects of extreme reduction in PRI in autumn with needle yellowing, the correlation of ɛ and PRI was relatively small in this forest. In the cypress forest, on the other hand, no significant correlation was found except with PRI and CCI. The highest correlation in this forest was that with PRI, suggesting that the leaf biomass-related VIs based on near-infrared reflectance are not sufficient for estimating ɛ of evergreen forest. In the mixed forest, with relatively sparse vegetation cover, all VIs were significantly correlated with ɛ, but the best correlation was that with SAVI, possibly owing to the reduction in the effect of the reflectance from background soil. Correlation analysis of the pooled data from all forests showed the highest correlation between ɛ and PRI. These results indicate that PRI is an effective VI in the remote estimation of ɛ in both deciduous and evergreen forests, although there are some sensitivity differences between vegetation types.

There are no comments yet on this publication. Be the first to share your thoughts.