ABSTRACT
Accurate estimation of fuel moisture content (FMC) is essential for determining the physiological status of vegetation and assessing fire risk. As FMC is not directly or physically linked to light absorption processes, its accurate estimation from hyperspectral data remains a challenge. Inspired by the recent boosting of applying fractional order derivative (FOD) spectra–based indices into leaf mass per area estimation, we investigated the potential of using FOD indices to retrieve FMC across a range of plant species of different vegetation types. We developed fractional indices to retrieve FMC based on a composite dataset consisting of 2891 leaf samples from different species. The evaluation results showed that the newly identified ND(1900, 2095) index based on the 0.8-order fractional derivative spectra had a high prospect, with an overall R2 value of 0.90, to capture a wide range (42–1321%) of FMCs in different plant species. Although the R2 values dropped when applied to each specific vegetation type, the regressions remained highly significant. Our results demonstrated the great potential of FOD-derived indices to estimate a wide range of live FMCs from hyperspectral reflected information.
Acknowledgements
The senior author wants to thank the members of the Laboratory of Macroecology and the Institute of Silviculture, Shizuoka University, for their support on developing the field Datasets. We are also grateful to the public online datasets.
Disclosure statement
No potential conflict of interest was reported by the authors.
Data availability statement
The public datasets used in this study are available at https://doi.org/doi:10.21232/9nr6-sq54 and https://ecosis.org/package/e88b832d-d7da-48b5-af59-25a8079a0ab6. The observed datasets collected from two deciduous broadleaf forest sites (Naeba and Nakagawane) used in this study are available from the corresponding author upon reasonable request.