This dataset depends on the following independent variable (i.e, grid):
grid: /T (days since 1960-01-01) ordered [ (6-15 Jan 1960) (16-25 Jan 1960) (26 Jan 1960 - 5 Feb 1960) (6-15 Feb 1960) (16-25 Feb 1960) (26 Feb 1960 - 5 Mar 1960) (6-15 Mar 1960) (16-25 Mar 1960) (26 Mar 1960 - 5 Apr 1960) (6-15 Apr 1960) (16-25 Apr 1960) (26 Apr 1960 - 5 May 1960) (6-15 May 1960) (16-25 May 1960) (26 May 1960 - 5 Jun 1960) (6-15 Jun 1960) (16-25 Jun 1960) (26 Jun 1960 - 5 Jul 1960) (6-15 Jul 1960) (16-25 Jul 1960) (26 Jul 1960 - 5 Aug 1960) (6-15 Aug 1960) (16-25 Aug 1960) (26 Aug 1960 - 5 Sep 1960) (6-15 Sep 1960) (16-25 Sep 1960) (26 Sep 1960 - 5 Oct 1960) (6-15 Oct 1960) (16-25 Oct 1960) (26 Oct 1960 - 5 Nov 1960) (6-15 Nov 1960) (16-25 Nov 1960) (26 Nov 1960 - 5 Dec 1960) (6-15 Dec 1960) (16-25 Dec 1960) (26 Dec 1960 - 5 Jan 1961)] :grid
so there will be 36 links below for the different files.
You may want to get a utility like wget which can download sets of files. For example (all one line)
wget -r -l2 -nd --no-parent -e robots=off -v --relative 'http://digilib.icpac.net/SOURCES/.ENACTS/.version0/.dekadal/.rainfall/.rfe_merged_vs_meanrfe_merged/%255BT+%255Dminover/%255BX+%255Dmaxover/%5BY+T+%5Ddetrend-bfl/yearly-climatology/T+differences/T/%28days%20since%201960-01-01%29streamgridunitconvert/ngridtable/downloadsarcinfo.html'
single quotes are necessary in unix to protect the url from expansion; probably necessary in other cases as well