Filling missing daily data in climatological time series records

Abstract

The lack of daily data in meteorological stations is frequent and this does not allow the series to be used in agroclimatic studies. With the above, the temporal and spatial variation of the variables that make up the agroclimate of a region is not known. The objective of this work was to estimate and verify by means of the methods: normal ratio, Fourier series and square of the inverse of the distance, the method with the lowest error for filling in missing daily data of the variables precipitation, sunshine, evaporation, temperature maximum, minimum temperature and relative humidity of the weather stations surrounding the rice production area in the department of Valle del Cauca, Colombia. Nine stations were analyzed, which do not have distances greater than 50 km, nor altitudinal differences of more than 750 m. These were used with different study periods according to the variable in progress and the methods were evaluated with the statistical indices square root of the mean square of the error and the coefficient of determination, the first allowed knowing the maximum admissible value of error and the second , the level of fit between the observed and estimated values. Therefore, these allowed us to infer that the variable sunshine and evaporation obtained the best results with the normal ratio; the minimum temperature and relative humidity with the Fourier series and the inverse square of the distance for precipitation and maximum temperature.