Abstract:
In this paper, we quantify the effect of the temporal sampling frequency of commonly measured climatic variables on the estimation of the reference evapotranspiration. Using a set of data sampled on an intensive basis (i.e. one measurement each minute) during a period of 6 months, we first analyse the effect of the temporal sampling frequency on the estimation of the daily means of the shortwave solar radiation, the wind speed, the dry and wet temperatures, and on the estimation of the daily maximum and minimum dry temperature. Subsequently, a sensitivity analysis of a reference evapotranspiration model is carried out to determine the most sensible meteorological variables. The sensitivity coefficients were then combined with the errors due to the temporal sampling to quantify for each variable the impact of the sampling frequency on the estimation of daily ETo. The results showed that the solar radiation and the wind speed are the most sensitive to bias induced by inadequate temporal sampling frequency. Daily errors of 5.1MJm-2d-1 or 41.05% for the solar radiation, and 0.45ms-1 or 18% for the wind speed may be obtained if these variables are inappropriately sampled. Moreover, the impact of inappropriate temporal sampling on the estimation of ETo can be significant with respective maximum bias of 0.62mmd-1 due to inappropriate solar radiation sampling and 0.36mmd-1 due to inappropriate maximum temperature sampling. A non-intensive hourly temporal sampling schedule of all meteorological variables may induce errors on the daily ETo so high as -0.76mmd-1 or -27%. Fortunately, the errors generated on the estimation of the long-term integrated evapotranspiration are clearly lower (3.8%). Our study clearly demonstrates the importance of scheduling appropriately the sampling frequency of climatic variables to correctly estimate land surfaces fluxes as well in fundamental as in more practically oriented research studies.
