@article{fdi:010095905,
  title = {{D}evelopment of hourly resolution air temperature across {T}iticaca {L}ake on auxiliary {ERA}5 variables and machine learning-based gap-filling},
  author = {{S}irpa-{P}oma, {J}. {W}. and {C}alle, {J}. and {U}scamayta-{F}errano, {E}. and {M}olina-{C}arpio, {J}. and {S}atg{\'e}, {F}r{\'e}d{\'e}ric and {T}oledo, {O}. {C}. and {D}uran, {R}. and {M}ollinedo, {P}. {P}. and {H}ussain, {R}. and {P}illco-{Z}olá, {R}.},
  editor = {},
  language = {{ENG}},
  abstract = {{T}his article presents an innovative procedure that combines advanced quality control ({QC}) methods with machine learning ({ML}) techniques to produce reliable, continuous, high-resolution meteorological data. {T}he approach was applied to hourly air temperature records from six automatic weather stations located around {L}ake {T}iticaca in the {A}ltiplano region of {S}outh {A}merica. {T}he raw dataset contained time gaps, inconsistencies, and outliers. {T}o address these, the {QC} stage employed {I}nterquartile {R}ange, {B}iweight, and {L}ocal {O}utlier {F}actor ({LOF}) statistics, resulting in a clean dataset. {T}wo gap-filling methods were implemented: a spatial approach using time series from nearby stations and a temporal approach based on each station's time series and selected variables from the {ERA}5-{L}and reanalysis. {S}everal {ML} models were also employed in this process: {R}andom {F}orest ({RF}), {S}upport {V}ector {M}achine ({SVM}), {S}tacking ({STACK}), and {A}da{B}oost ({ADA}). {M}odel performance was evaluated on a validation subset (30% of station data). {T}he {RF} model achieved the best results, with {R}2 values up to 0.9 and {R}oot {M}ean {S}quare {E}rror ({RMSE}) below 1.5 degrees {C}. {T}he spatial approach performed best when stations were strongly correlated, while the temporal approach was more suitable for locations with low inter-station correlation and high local variability. {O}verall, the procedure substantially improved data reliability and completeness, and it can be extended to other meteorological variables.},
  keywords = {{T}iticaca {L}ake ; air temperature ; quality control ; machine learning ; gap-filling ; data consistency ; {PEROU} ; {BOLIVIE} ; {TITICACA} {LAC}},
  booktitle = {},
  journal = {{S}ensors},
  volume = {25},
  numero = {23},
  pages = {7165 [20 p.]},
  year = {2025},
  DOI = {10.3390/s25237165},
  URL = {https://www.documentation.ird.fr/hor/fdi:010095905},
}