Climate Change Research Spotlight

Identifying Biases in Satellite Temperature and Humidity Records

Researchers estimated the sampling biases that affect temperature and humidity climatologies derived from detectors on NASA’s Aqua satellite.

By

NASA’s Aqua satellite has been in orbit for more than a decade, and current projections expect it to hold out until the early 2020s. The satellite’s long lifetime of continuous operation has made its observations particularly valuable for testing climate models.

Getting the full value out of Aqua’s sensors, however, requires having a clear understanding of how measurement techniques can bias climatologies derived from the satellite’s observations. In their research, Hearty et al. estimated the sampling biases that affect temperature and humidity climatologies derived from two of Aqua’s detectors: the atmospheric infrared sounder (AIRS) and the advanced microwave sounding ­unit-A (­AMSU-A).

Clouds, Sun glint, high surface emissivities, or measurement gaps due to the satellite’s orbit can result in some atmospheric states not being sampled, which can skew climatological averages. To calculate the extent of the sampling biases, the authors compared temperature and humidity climatologies derived from a model reanalysis data set that was sampled as if it were being measured by AIRS and ­AMSU-A against the full model reanalysis.

The authors found that the instrumental sampling biases, which stem from limitations in the instruments’ ability to make observations during atmospheric states, tend to outweigh the temporal sampling biases that stem from the observational limitations imposed by the diurnal cycle. In locations that are particularly difficult to measure, such as right above the Earth’s surface, temperature observations can be off by as much as 2 K. Instrumental sampling biases can also skew humidity measurements by more than 20%. In some regions, the authors note, the biases are not static and thus cannot be corrected with a blanket fix.

The authors point out that uncertainties in climate models generally outweigh biases in the AIRS/­AMSU-A climatologies. As models improve, however, the sampling biases should be taken into consideration when comparing climate models to AIRS/­AMSU-A climatologies. (Journal of Geophysical Research: Atmospheres, doi:10.1002/​2013JD021205, 2014)

—Colin Schultz, Writer

© 2014. American Geophysical Union. All rights reserved.