It was used data from 19 locations of the Rio Grande do Sul State, Brazil, collected from 1953 to 2007, with the objective to determine the sample size to estimate the average monthly rainfall. The central tendency, variability and asymmetry was calculated for each of the 228 time series (12 months x 19 locations) and the aleatory and normality data were tested. Then it was verified the homogeneity of variance among months in each locality and among places in each month and it was determined the sample size to estimate the average monthly rainfall in each month and locality. The sample size (number of years) to estimate the average monthly rainfall is dependent on the month and locality. One concluded that 51 years of data are enough to predict the average monthly rainfall, with an estimation error equal to 25% of estimated average, with a degree confidence of 95%.
time series; error of estimation; sample size; sampling
