Differences in trends and anomalies of upper-air observations from GPS RO, AMSU, and radiosondes

Observations of thermodynamic variables are sparse in the upper-air
region, especially when considering measurements capable of detecting
changes in the climate state. In the upper troposphere/lower
stratosphere, temperature measurements with global coverage, high
vertical resolution, and high quality are provided by the GPS Radio
Occultation (RO) satellite dataset. GPS RO now provides 15 years of
observations, and while this is still a short record from a climate
perspective, its properties enable a long-term stable and consistent
global data record. Due to its unique properties, also other upper-air
measurement systems such as the Advanced Microwave Sounding Unit
(AMSU) or the global radiosonde network can be validated.

In this study we use the whole GPS RO record to calculate differences
in anomalies between GPS RO, AMSU, and high-quality radiosondes
(Vaisala) to determine potential biases. We base these comparisons on
layer-integrated brightness temperatures for the lower stratospheric
channel of AMSU, where we expect best quality from GPS RO. We
investigate irregularities in the anomaly difference time series, and
compare to ECMWF model results.

We also take advantage of the good vertical resolution of the
radiosonde dataset to investigate differences in the observed
vertically resolved atmospheric trend signals between GPS RO and
radiosondes.