Lobit P. auteur aut IRD Perez L. L. auteur aut IRD Lhomme Jean-Paul auteur aut IRD Tagle A. G. auteur aut IRD text journalArticle eng text/pdf born digital access This study evaluates the dew point method (Allen et al. 1998) to estimate atmospheric vapor pressure from minimum temperature, and proposes an improved model to estimate it from maximum and minimum temperature. Both methods were evaluated on 786 weather stations in Mexico. The dew point method induced positive bias in dry areas but also negative bias in coastal areas, and its average root mean square error for all evaluated stations was 0.38 kPa. The improved model assumed a bi-linear relation between estimated vapor pressure deficit (difference between saturated vapor pressure at minimum and average temperature) and measured vapor pressure deficit. The parameters of these relations were estimated from historical annual median values of relative humidity. This model removed bias and allowed for a root mean square error of 0.31 kPa. When no historical measurements of relative humidity were available, empirical relations were proposed to estimate it from latitude and altitude, with only a slight degradation on the model accuracy (RMSE = 0.33 kPa, bias = -0.07 kPa). The applicability of the method to other environments is discussed. specialized MEXIQUE 072 020 133 3-4 751-762 2018 0177-798X https://www.documentation.ird.fr/hor/fdi:010073628 10.1007/s00704-017-2211-9 0177-798X [F B010073628] https://www.documentation.ird.fr/hor/fdi:010073628 https://www.documentation.ird.fr/intranet/publi/2018/07/010073628.pdf Accès réservé (Intranet de l'IRD) IRD - Base Horizon / Pleins textes 2018-09-03 2018-09-03 fdi:010073628 fre