@article{fdi:010085883,
  title = {{M}apping soil organic carbon stocks in {T}unisian topsoils},
  author = {{B}ahri, {H}. and {R}aclot, {D}amien and {B}arbouchi, {M}. and {L}agacherie, {P}. and {A}nnabi, {M}.},
  editor = {},
  language = {{ENG}},
  abstract = {{B}etter knowledge of the amount and spatial distribution of soil organic carbon ({SOC}) stock at national level is a key element for monitoring, planning and decision-making regarding soil quality management, agriculture or carbon storage options. {T}he present study proposes for the first time a {D}igital {S}oil {M}apping ({DSM}) initiative to map {SOC} stocks in {T}unisian topsoils (0-30 cm) at 100 m resolution, using a {Q}uantile {R}egression {F}orest ({QRF}) algorithm, a range of environmental covariates, and a national database of 1540 {SOC} stock profiles. {O}ur results provided a revised assessment of the {SOC} stock on the {T}unisian territory at 391{T}g {C} in the first 30 cm of soil profile, i.e. an average of 2.53 kg m-2. {T}he map of {SOC} stocks outperformed global {DSM} products such as {S}oil{G}rids 2.0 in both {R}2 (0.44 vs. 0.15) and {RMSE} (1.94 vs. 2.52 kg m-2) and can be used as a benchmark against changes of land use and climate. {T}he importance of the environmental covariates tested indicates the major role of bioclimatic data and, to a lesser extent, remote sensing images and topography-related variables. {O}ur study did not show a significant added value of using additional covariates in relation to nationally available variables or the {SOC} map predicted by {S}oil{G}rids2.0. {F}inally, our results showed that increasing the quality and quantity of soil profile observations is most likely the best way to improve the future {SOC} map, starting with the northern region of {T}unisia, which has the highest {SOC} stock predictions and uncertainties in the country. {A}n alternative way would be the exploration of new covariates through sub-national approaches.},
  keywords = {{D}igital soil mapping ; {S}oil organic carbon stocks ; {T}unisian soils ; {Q}uantile regression forest ; {M}achine learning spatial prediction ; {N}ational approach ; {M}ultiple soil classes ; {TUNISIE}},
  booktitle = {},
  journal = {{G}eoderma {R}egional},
  volume = {30},
  numero = {},
  pages = {e00561 [10 ]},
  ISSN = {2352-0094},
  year = {2022},
  DOI = {10.1016/j.geodrs.2022.e00561},
  URL = {https://www.documentation.ird.fr/hor/fdi:010085883},
}