*Note: This dataset version has been superseded by a newer version. It is highly recommended that users access the current version. Users should only use this version for special cases, such as reproducing studies that used this version.* This Climate Data Record (CDR) includes lower tropospheric, mid-tropospheric, and lower stratospheric temperatures over land and ocean derived from microwave radiometers on NOAA and NASA polar orbiting satellites. The temperatures are from measurements produced by Microwave Sounding Units (MSU) since 1978 and Advanced Microwave Sounding Unit-A (AMSU-A) since 1998 flying on NOAA polar orbiting satellites, on NASA Aqua satellite (operating since mid-1998) and on the European MetOp satellite (operating since late 2006). The instruments are cross-track through-nadir scanning externally-calibrated passive microwave radiometers. Brightness temperature measurements are derived at microwave frequencies within the 50-60 GHz oxygen absorption complex, and (in the case of AMSU-A) at a few microwave frequencies above and below that absorption complex. There are three atmospheric layers for which intermediate products are processed: (1) lower-tropospheric (TLT) deep-layer average temperature, computed as a weighted difference between view angles of AMSU-A channel 5, whose heritage comes from MSU channel 2, (2) mid-tropospheric (TMT) deep-layer temperature, computed as an average of the central portion of the scan of AMSU-A channel 5, whose heritage also comes from MSU channel 2, and (3) lower-stratospheric (TLS) deep layer temperatures, computed from the central portion of the scan of AMSU channel 9, whose heritage comes from MSU channel 4. This CDR includes several products. The global monthly anomaly data data are averaged onto a 2.5 x 2.5 degree latitude-longitude grid for each of the three atmospheric layers. Monthly anomalies are averaged for each of the three atmospheric layers over multiple regions, including Global, hemispheric, tropic, extratropic, polar and contiguous U.S. A mean annual cycle of monthly mean layer temperatures is also included. Anomalies are deviations from 1981-2010 mean. The datasets have been converted from the native ASCII format to CF-compliant netCDF-4 format.