@article{fdi:010049080,
  title = {{U}se of high-resolution satellite imagery in an integrated model to predict the distribution of shade coffee tree hybrid zones},
  author = {{G}omez, {C}{\'e}line and {M}angeas, {M}organ and {P}etit, {M}ichel and {C}orbane, {C}hristina and {H}amon, {P}erla and {H}amon, {S}erge and {K}ochko, {A}lexandre de and {L}e {P}ierres, {D}aniel and {P}oncet, {V}al{\'e}rie and {D}espinoy, {M}arc},
  editor = {},
  language = {{ENG}},
  abstract = {{I}n {N}ew {C}aledonia (21 {S}, 165 {E}), shade-grown coffee plantations were abandoned for economic reasons in the middle of the 20th century. {C}offee species ({C}offee arabica, {C} canephora and {C} liberica) were introduced from {A}frica in the late 19th century, they survived in the wild and spontaneously cross-hybridized. {C}offee species were originally planted in native forest in association with leguminous trees (mostly introduced species) to improve their growth. {T}hus the canopy cover over rustic shade coffee plantations is heterogeneous with a majority of large crowns, attributed to leguminous trees. {T}he aim of this study was to identify suitable areas for coffee inter-specific hybridization in {N}ew {C}aledonia using field based environmental parameters and remotely sensed predictors. {D}ue to the complex structure of tropical vegetation, remote sensing imagery needs to be spatially accurate and to have the appropriate bands for monitoring vegetation cover. {Q}uickbird panchromatic (black and white) imagery at 0.6 to 0.7 m spatial resolutions and multispectral imagery at 2.4 m spatial resolution were pansharpened and used for this study. {T}he two most suitable remotely sensed indicators, canopy heterogeneity and tree crown size, were acquired by the sequential use of tree crown detection (neural network), image processing (such as textural analysis) and classification. {A}ll models were supervised and trained on learning data determined by human expertise. {T}he final model has two remotely sensed indicators and three physical parameters based on the {D}igital {E}levation {M}odel: elevation, slope and water flow accumulation. {U}sing these five predictive variables as inputs, two modelling methods, a decision tree and a neural network, were implemented. {T}he decision tree, which showed 96.9% accuracy on the test set, revealed the involvement of ecological parameters in the hybridization of {C}offea species. {W}e showed that hybrid zones could be characterized by combinations of modalities, underlining the complexity of the environment concerned. {F}or instance, forest heterogeneity and large crown size, steep slopes (>53.5%) and elevation between 194 and 429 m asl, are favourable factors for {C}offea inter-specific hybridization. {T}he application of the neural network on the whole area gave a predictive map that distinguished the most suitable areas by means of a nonlinear continuous indicator. {T}he map provides a confidence level for each area. {T}he most favourable areas were geographically localized, providing a clue for the detection and conservation of favourable areas for {C}offea species neo-diversity.},
  keywords = {{D}ecision tree ; {M}ulti-scale textural analysis ; {N}eural networks ; {V}egetation index ; {Q}uickbird imagery ; {F}orest classification ; {C}anopy cover ; {T}ree crown size ; {N}ew {C}aledonia ; {H}ybrid zone ; {C}offea},
  booktitle = {},
  journal = {{R}emote {S}ensing of {E}nvironment},
  volume = {114},
  numero = {11},
  pages = {2731--2744},
  ISSN = {0034-4257},
  year = {2010},
  DOI = {10.1016/j.rse.2010.06.007},
  URL = {https://www.documentation.ird.fr/hor/fdi:010049080},
}