The intermittency of rainfall in space and time introduces uncertainty to rainfall estimates based on limited observations in space and time. This is particularly true for rainfall estimation from space with infrequent observations made at intervals of several hours only. Using a multi-year data set of merged radar observations (WSI radar mosaic) made over the United States east of the Rocky Mountains, at a resolution of 2 km and 15 min, we are studying the sampling error of infrequent observations made at regular intervals. Partial visits of a given area are not considered as part of this analysis. In particular, we are trying to quantify the sampling error as a function of space and time domains, the rainfall intensity, and the sampling resolutions in space and time. Moreover, we are investigating how the fractional coverage may affect the sampling error from satellite sensors, where only part of the beam is filled with rainfall. Our analyses will also show how the spatial variability of estimated rainfall distributions may be affected by the sampling resolutions in space and time. This study aims at putting uncertainty measures (sampling error) on remotely sensed rainfall estimates. The anticipated results will provide guidance for interpretation of rainfall estimates from satellites, such as the Tropical Rainfall Measuring Mission (TRMM), and planning of future satellite missions, like the Global Precipitation Mission (GPM).