Atmospheric observations made with ground-based, multifrequency radiometers (f = 4.9, 10.4, 21, 35, and 94 GHz) for the original purpose of microwave signature research for remote sensing of snowpacks, sea ice, and vegetation were analyzed. The tipping-curve method was used both for radiometric calibration and for the determination of the zenith opacities. The data were grouped into three regions with different locations and different altitudes (10, 570, and 2550 m) above sea level. Key parameters are air temperature and atmospheric zenith opacities at each frequency. The large number of different (nonprecipitating) conditions provide statistical properties of atmospheric attenuation. Especially high correlation was found between the opacities at 35 and 94 GHz. For cloudless subsets of the data these opacities are also highly correlated with the ones at 21 GHz. Estimation of the total atmospheric contents of water vapor V and cloud liquid water L was made from the opacities at 21, 35, and 94 GHz. Whereas V is clearly correlated with the surface air temperature, large L values are clustered around temperatures near 0°C. Comparison with radiosonde measurements indicates a standard deviation in V of 0.22 cm and a bias of −0.11 cm. Consistency tests of liquid water estimates show a standard deviation of 0.03 mm.