CCL: charge densities

Hi CCL'ers,
 I would like to compute highly accurate charge densities for atoms.
 Gaussian can calculate densities (for population and other analysis
 procedures), but if I understand the manual correctly, this can only be
 done for MP2, MP3, MP4(SDQ), QCISD, CCD, CCSD, CID and CISD and SAC-CI
 methods. I would like to have densities for methods that include triple
 excitations, like CCSD(T). As far as I am aware, Molpro and NWChem also
 do not calculate densities for highly correlated methods. Is there a
 theoretical reason why these programs do not compute densities for,
 say, CCSD(T) or CCSDT? Would it be at all possible?  And if so, does
 anybody know a program that can compute charge densities for highly
 correlated methods?
 Thanks a lot,
    Tanja van Mourik
    Royal Society University Research Fellow
    Chemistry Department
    University College London  phone:  +44 (0)20-7679-4663
    20 Gordon Street          e-mail: work:
    London WC1H 0AJ, UK               home: