The study of spatial climate variability or agricultural potentialities requires continuous fields of climate variables. Temperatures are collected from point sources by weather stations. It is necessary to aggregate these data in space in order to estimate the value of the variable in all points of the study area (Burgundy, France). For this purpose, a deterministic method was used: partial thin plate splines. In order to study the time step sensibility of these estimations, this method is applied at yearly, monthly and daily time steps, for the years 1991 and 2003. Elevation was used as an environmental covariate, to improve the interpolation process. The performance of the interpolation method was assessed by means of jack-knife cross-va...