The microwave scattering and emission properties of precipitation cells are investigated by comparing 118‐GHz radiometric observations with a planar‐stratified numerical radiative transfer model. Liquid and frozen hydrometeors are modeled as spherical Marshall‐Palmer and Sekhon‐Srivastava distributed Mie scattering polydispersions, respectively, with Henyey‐Greenstein phase functions. Comparisons are made between computed brightness temperatures based on weather radar observations of a convective precipitation cell couplet during the Cooperative Huntsville Meteorological Experiment (COHMEX), 1986, and brightnesses observed coincidentally by an airborne imaging spectrometer. Agreement between observed and computed brightness perturbations is within 10% over the radiometrically opaque, mature regions of the cell, although the model brightnesses over the partially transparent anvil region are highly sensitive to the assumed mean ice particle size. The sensitivity of the 118‐GHz channels to temperature at various levels in the troposphere is exhibited, in the presence of precipitation, through the perturbed temperature weighting functions. Calculations using the perturbed and nonscattering weighting functions suggest 118‐GHz transparent‐channel cell top reflectivities of up to 50% in the convective core region and 6% in the anvil region. A rain cell model parameterized by cell top altitude and total (liquid and ice) water density is used to analyze the information content of 118‐GHz precipitation cell spectra. The model suggests that at 118 GHz, cells with uniform water densities greater than 0.5 g m−3 are opaque, and cells with uniform densities less than 0.1 g m−3 are transparent. The dominant 118‐GHz spectral modes are found to contain useful information on the cell top altitude and can be used to detect transparent anvil regions. The physical retrieval of cell water density is shown to be facilitated by coincident observations using similar weighting‐function channels at 53 and 118 GHz, although the retrieval is sensitive to the assumed mean ice particle size.