Weather surveillance radars routinely detect smoke of various origin. Of particular significance to the meteorological community are wildfires in forests and/or prairies. For example, one responsibility of the National Weather Service in the USA is to forecast fire outlooks as well as to monitor wildfire evolution. Polarimetric variables have enabled relatively easy recognitions of smoke plumes in data fields of weather radars. Presented here are the fields of these variables from smoke plumes caused by grass fire, brush fire, and forest fire. Histograms of polarimetric data from plumes contrast these cases. Most of the data are from the polarimetric Weather Surveillance Radar 1988 Doppler (WSR-88D aka NEXRAD, 10 cm wavelength); hence, the wavelength does not influence these comparisons. Nevertheless, in one case, simultaneous observations of a plume by the operational Terminal Doppler Weather Radar (TDWR, 5 cm wavelength) and a WSR-88D is used to infer backscattering characteristic and, hence, sizes of dominant contributors to the returns. To interpret these measurements, Computational Electromagnetics (CEM) tools are applied. For one wildfire from Oklahoma, radar and satellite (GOES-16, Geostationary Operational Environmental Satellite) images are analyzed. The case demonstrates a potential to forecast fire intensification caused by a very rapid cold front. Finally, we suggest a possible way to extract the smoke plume return from the class of nonmeteorological scatterers.