@article{fdi:010076667,
  title = {{E}stimating leaf mass per area and equivalent water thickness based on leaf optical properties : potential and limitations of physical modeling and machine learning},
  author = {{F}eret, {J}. {B}. and le {M}aire, {G}. and {J}ay, {S}. and {B}erveiller, {D}. and {B}endoula, {R}. and {H}mimina, {G}. and {C}heraiet, {A}. and {O}liveira, {J}. {C}. and {P}onzoni, {F}. {J}. and {S}olanki, {T}. and de {B}oissieu, {F}. and {C}have, {J}. and {N}ouvellon, {Y}. and {P}orcar-{C}astell, {A}. and {P}roisy, {C}hristophe and {S}oudani, {K}. and {G}astellu-{E}tchegorry, {J}. {P}. and {L}efevre-{F}onollosa, {M}. {J}.},
  editor = {},
  language = {{ENG}},
  abstract = {{L}eaf mass per area ({LMA}) and leaf equivalent water thickness ({EWT}) are key leaf functional traits providing information for many applications including ecosystem functioning modeling and fire risk management. {I}n this paper, we investigate two common conclusions generally made for {LMA} and {EWT} estimation based on leaf optical properties in the near-infrared ({NIR}) and shortwave infrared ({SWIR}) domains: (1) physically-based approaches estimate {EWT} accurately and {LMA} poorly, while (2) statistically-based and machine learning ({ML}) methods provide accurate estimates of both {LMA} and {EWT}. {U}sing six experimental datasets including broadleaf species samples of >150 species collected over tropical, temperate and boreal ecosystems, we compared the performances of a physically-based method ({PROSPECT} model inversion) and a {ML} algorithm (support vector machine regression, {SVM}) to infer {EWT} and {LMA} based on leaf reflectance and transmittance. {W}e assessed several merit functions to invert {PROSPECT} based on iterative optimization and investigated the spectral domain to be used for optimal estimation of {LMA} and {EWT}. {W}e also tested several strategies to select the training samples used by the {SVM}, in order to investigate the generalization ability of the derived regression models. {W}e evidenced that using spectral information from 1700 to 2400 nm leads to strong improvement in the estimation of {EWT} and {LMA} when performing a {PROSPECT} inversion, decreasing the {LMA} and {EWT} estimation errors by 55% and 33%, respectively. {T}he comparison of various sampling strategies for the training set used with {SVM} suggests that regression models show limited generalization ability, particularly when the regression model is applied on data fully independent from the training set. {F}inally, our results demonstrate that, when using an appropriate spectral domain, the {PROSPECT} inversion outperforms {SVM} trained with experimental data for the estimation of {EWT} and {LMA}. {T}hus we recommend that estimation of {LMA} and {EWT} based on leaf optical properties should be physically-based using inversion of reflectance and transmittance measurements on the 1700 to 2400 nm spectral range.},
  keywords = {{B}iophysical properties ; {L}eaf spectroscopy ; {EWT} ; {LMA} ; {R}adiative transfer model ; {S}upport vector machine ; {V}egetation},
  booktitle = {},
  journal = {{R}emote {S}ensing of {E}nvironment},
  volume = {231},
  numero = {},
  pages = {art 110959 [14 p.]},
  ISSN = {0034-4257},
  year = {2019},
  DOI = {10.1016/j.rse.2018.11.002},
  URL = {https://www.documentation.ird.fr/hor/fdi:010076667},
}