As a first cut at removing the seasonality of the nottem data, we compute average temperatures across years within a month. This is done below and the results are plotted.
> nottem.matrix_matrix(nottem,ncol=12,byrow=T) > # This arranges the data in a matrix. The columns are the data in a > # given month (first column is all the Januaries) and the rows are > # the data in a given year. > nottem.monthly.means_apply(nottem.matrix,2,mean) > plot(1:12,nottem.monthly.means) > #1:12 is the sequence of integers 1 to 12 > nottem.dev.mm_nottem-rep(monthly.means,20) > # Note that there are 20 years worth of data, so we have to repeat the > # sequence of monthly means 20 times. > ts.plot(nottem.dev.mm)
The last two plots appear as Figures 5 and 7.