Climate and Dynamics
Assessing uncertainty in estimates of atmospheric temperature changes from MSU and AMSU using a Monte-Carlo estimation technique
Article first published online: 20 APR 2011
Copyright 2011 by the American Geophysical Union.
Journal of Geophysical Research: Atmospheres (1984–2012)
Volume 116, Issue D8, 27 April 2011
How to Cite
2011), Assessing uncertainty in estimates of atmospheric temperature changes from MSU and AMSU using a Monte-Carlo estimation technique, J. Geophys. Res., 116, D08112, doi:10.1029/2010JD014954., , , and (
- Issue published online: 20 APR 2011
- Article first published online: 20 APR 2011
- Manuscript Accepted: 9 FEB 2011
- Manuscript Revised: 7 FEB 2011
- Manuscript Received: 25 AUG 2010
- climate change;
- remote sensing
 Measurements made by the Microwave Sounding Unit (MSU) and the Advanced Microwave Sounding Unit (AMSU) provide a multidecadal record of global atmospheric temperature change, which have been used by several groups to produce long-term temperature records of thick layers of the atmosphere from the lower troposphere to the lower stratosphere. Here we present an internal uncertainty estimate for the Remote Sensing Systems data sets made using a Monte Carlo approach that includes contributions to the total uncertainty from sampling error, premerge adjustments to each individual satellite, and the merging procedure. The results can be used to estimate uncertainties in this product at all space and time scales of interest to any specific application. On small space and time scales sampling effects dominate. On the longer time scales intersatellite merging is important at all levels and the diurnal adjustment is a critical uncertainty for the two layers that have a significant surface component, particularly over land. A comparison of trends for the globe, tropics, and extratropics between the best estimate data set along with these error estimates and homogenized radiosonde estimates and available MSU/AMSU estimates from other groups is undertaken. This shows consistency between our product and those produced by others within the stated uncertainty for many regions and layers. In almost as many cases, however, the interdata set differences of the estimated trends are too large be accounted for by the internal uncertainty estimates derived herein.