Decades of Atmospheric Measurement Errors Corrected

What if you suddenly found that everything you thought you knew was wrong? Okay: What if you suddenly found that everything you thought you knew was slightly wrong?

For years, climate researchers have mined radio occultation data to produce a picture of the atmosphere. Just as refraction makes the sun appear to drop tendrils down to the horizon as it sets, the atmosphere bends radio waves arcing through the upper air from medium-Earth-orbit GPS satellites (circling at altitudes of about 20 200 km) to low-Earth-orbit satellites (non-geostationary communications and remote-sensing craft at altitudes of about 2000 km). Picture a GPS satellite beaming a ray over the surface of the Earth to a low-flying satellite playing peek-a-boo around the curvature of the earth, with the atmosphere acting like a prism to bend the light path downward.

Environmental scientists can track the exquisitely timed GPS signals to measure how a wave’s path and phase shift, analyzing it to give a picture of atmospheric pressure and, especially, temperature at almost every altitude.

The basic formula for calculating the refractive bending angle dates back to 1953, and includes terms for deflection by dry atmosphere, moist atmosphere (humidity), water droplets, and ionospheric effects, all as functions of altitude. The atmospheric, humidity, and water droplet terms all tend to increase the beam’s downward bend. The ionospheric term—a function of electron density (number of electrons per cubic meter) and the beam’s frequency—tends to bend the beam back up towards the horizontal. In effect, the ionospheric term masks some of the effects of temperature and humidity.

Despite later refinements in the bending angle calculation, subsequent studies may not have made enough allowance for the wide fluctuations in high-atmosphere ionization caused by the ever-changing solar wind. 

Researchers at the Wegener Center for Climate and Global Change at Austria’s University of Graz analyzed 11 years worth of radio occultation and solar particle flux data to conclude that we have made longstanding and systematic errors in our pictures of the atmosphere. After reviewing CHAMP (Challenging Minisatellite Payload) and COSMIC  (Constellation Observing System for Meteorology, Ionosphere, and Climate) radio occultation data from 2001 through 2011, Graz physicists J. Danzer, B. Scherlin-Pirscher, and  project leader Ulrich Foelsche calculate, for example, that temperature measurements for the air 35 km above the equator have been off by about -3.9 Kelvin during the peak of the solar cycle (as in 2002) and -1.4 K during the solar minimum (2008). And, though it has long been known that bending angles change at night, the variation is, in fact, larger than had been previously suspected, with daytime path “undeflections” as much as 0.4 microradians ( 0.000023 degrees) greater than previously calculated.

The purpose of the research, the authors note, is to allow more accurate interpretation of very large-scale, very long-term climate monitoring and projection. Foelsche readily concedes that the effect is small, but it does affect our understanding of the current situation. For example, atmospheric warming near the Earth's surface seems to be accompanied by a cooling in the stratosphere. The CHAMP satellite was launched in 2000 (just in time for the maximum) and re-entered the atmosphere in 2010 (just after the minimum). So the gradual decline of ionospheric bias would have masked anthropogenic cooling of the upper atmosphere, leading us to underestimate the magnitude of the effect, Foelsche says.

Image: Wegener Center

(Post revised to correct 0.4 microradian bias figure.)