how can i use arrayfun to split a timeseries with multivariates into separate timeseries?

hi all - looking for some help to do this quickly with arrayfun or something similar a is multivariate a = [ value zoneId] [ time] with multiple zoneId for each "time", but only 1 value... I'd like to split out into a circa like a table with ZoneId across the colum and value down the row linked with the "time value"... but this can be achieved with a "set" of time series for each zoneId. any help gratefully appreciated!

if true
  a = [ 1.5 101 ; 2 102 ; 1 101 ; 2 102 ; 3 101; 2 102; 3 103];
  time = datenum( ['1/1/2005' ; '1/1/2005' ;'1/2/2005' ;'1/2/2005'...
   ; '1/3/2005' ; '1/3/2005'; '1/3/2005']);
  ts = timeseries(a, time);
  zoneList = unique(a(:,2));
  idx101 = a(:,2) == 101;
  idx102 = a(:,2) == 102;
  idx103 = a(:,2) == 103;
  % for simplicty i have labled idx101 , but in actualitye the size of zoneList
  % may vary making it required to handle this differnetly...
  tsgroup = repmat(timeseries, size(zoneList));
  tsgroup(1) = timeseries( a(idx101,1) , time(idx101))
  tsgroup(2) = timeseries( a(idx102,1),  time(idx102))
  tsgroup(3) = timeseries( a(idx103,1) , time(idx103))

end