@article{fdi:010094381,
  title = {{H}igh spatial resolution soil moisture mapping over agricultural field integrating {SMAP}, {IMERG}, and {S}entinel-1 data in machine learning models},
  author = {{T}ola, {D}. and {B}ustillos, {L}. and {A}rragan, {F}. and {C}hipana, {R}. and {H}ostache, {R}enaud and {R}esongles, {E}l{\'e}onore and {E}spinoza-{V}illar, {R}. and {Z}olá, {R}. {P}. and {U}scamayta, {E}. and {P}erez-{F}lores, {M}. and {S}atg{\'e}, {F}r{\'e}d{\'e}ric},
  editor = {},
  language = {{ENG}},
  abstract = {{S}oil moisture content ({SMC}) is a critical parameter for agricultural productivity, particularly in semi-arid regions, where irrigation practices are extensively used to offset water deficits and ensure decent yields. {Y}et, the socio-economic and remote context of these regions prevents sufficiently dense {SMC} monitoring in space and time to support farmers in their work to avoid unsustainable irrigation practices and preserve water resource availability. {I}n this context, our study addresses the challenge of high spatial resolution (i.e., 20 m) {SMC} estimation by integrating remote sensing datasets in machine learning models. {F}or this purpose, a dataset made of 166 soil samples' {SMC} along with corresponding {SMC}, precipitation, and radar signal derived from {S}oil {M}oisture {A}ctive {P}assive ({SMAP}), {I}ntegrated {M}ulti-satellit{E} {R}etrievals for {GPM} ({IMERG}), and {S}entinel-1 ({S}1), respectively, was used to assess four machine learning models' ({D}ecision {T}ree-{DT}, {R}andom {F}orest-{RF}, {G}radient {B}oosting-{GB}, {E}xtreme {G}radient {B}oosting-{XGB}) reliability for {SMC} mapping. {F}irst, each model was trained/validated using only the coarse spatial resolution (i.e., 10 km) {SMAP} {SMC} and {IMERG} precipitation estimates as independent features, and, second, {S}1 information (i.e., 20 m) derived from single scenes and/or composite images was added as independent features to highlight the benefit of information (i.e., {S}1 information) for {SMC} mapping at high spatial resolution (i.e., 20 m). {R}esults show that integrating {S}1 information from both single scenes and composite images to {SMAP} {SMC} and {IMERG} precipitation data significantly improves model reliability, as {R}2 increased by 12% to 16%, while {RMSE} decreased by 10% to 18%, depending on the considered model (i.e., {RF}, {XGB}, {DT}, {GB}). {O}verall, all models provided reliable {SMC} estimates at 20 m spatial resolution, with the {GB} model performing the best ({R}2 = 0.86, {RMSE} = 2.55%).},
  keywords = {sentinel-1 ; {SMAP} ; high spatial resolution ; soil moisture content ; machine learning},
  booktitle = {},
  journal = {{R}emote {S}ensing},
  volume = {17},
  numero = {13},
  pages = {2129 [19 p.]},
  year = {2025},
  DOI = {10.3390/rs17132129},
  URL = {https://www.documentation.ird.fr/hor/fdi:010094381},
}