Measurements of the 3D distributions of cloud water are pressingly needed in many areas. The remote sensing techniques for determining cloud water distributions mainly make use of either microwave radiometers or cloud radars. These instruments are complementary in nature as the microwave radiometers measure path-integrated water content while cloud radars provide range-resolved reflectivities. Although radars have been used to study clouds since WWII, the use of microwave tomography to retrieve 3D cloud structures was made possible only by Warner et al.'s pioneer work in the 1980s. Since then the cloud tomography method has lain dormant until recently when it was revisited by Huang et al., who showed that the retrieval of cloud water distribution from microwave emission data is an under-constrained problem. We describe the use of radar data to constrain the tomographic retrieval: either to derive cloud boundaries (single-frequency radar); or to derive the initial estimates of the cloud liquid water content fields (dual-frequency radar). The simulation experiments show that adding radar data into cloud tomography greatly improves the retrieval accuracy. The use of radar data allows two radiometers to achieve the same retrieval accuracy as eight radiometers can achieve without radar data.