Radio occultation data show that in the first two decades of the 21st century, the Earth’s lowest atmospheric layer – the troposphere – warmed significantly, by up to 0.5 K per decade, while the layer above that – the stratosphere - cooled by about the same amount.

Over the last century, the amount of greenhouse gases has accumulated in the Earth’s atmosphere at a gradually accelerating rate. It is expected that the warming caused by greenhouse gases impacts the structure of Earth's atmosphere (Meng et al., 2021). This is precisely what we observe. During the past decades the atmospheric layer closest to the Earth surface, the troposphere, has warmed, while the layer above it, the stratosphere, has cooled (see Figure 1), and related to these changes, the tropopause height has risen globally (IPCC, 2021).

Definition

The atmosphere of planet Earth: The tropopause is between the troposphere and the stratosphere.
Rising from the planetary surface of the Earth, the tropopause is the atmospheric level where the air ceases to become cool with increased altitude and becomes dry, devoid of water vapor. The tropopause is the boundary that demarcates the troposphere below from the stratosphere above, and is part of the atmosphere where there occurs an abrupt change in the environmental lapse rate (ELR) of temperature, from a positive rate (of decrease) in the troposphere to a negative rate in the stratosphere. The tropopause is defined as the lowest level at which the lapse rate decreases to 2°C/km or less, provided that the average lapse-rate, between that level and all other higher levels within 2.0 km does not exceed 2°C/km.[1] The tropopause is a first-order discontinuity surface, in which temperature as a function of height varies continuously through the atmosphere, while the temperature gradient has a discontinuity.

Quantifying the rate of warming and cooling in the different layers of the free atmosphere is a difficult task, as evidenced by the low confidence the Intergovernmental Panel on Climate Change (IPCC) assigned, in its fifth assessment report, to the temperature changes in the free atmo­sphere (IPCC, 2013). However, in its sixth assessment report IPCC saw an improved confidence assigned to these temperature changes, in particular for the period after 2001 (IPCC, 2021). This is largely due to the addition of a new satellite-based data record, i.e. the Global Navigation Satellite System (GNSS) Radio Occultation (RO) data.

Since around the 1850s, the temperatures in the lowest layers of the troposphere have been measured with ground-based thermo­meters. However, these time series are quite uncertain in the earliest decades, and only provide measurements close to the surface. Apart from sparse balloon-borne measurements gathered since the 1940s, getting a detailed view of upper-atmosphere temperatures had to wait for the start of the satellite era.

Since the end of the 1970s, satellite-borne instruments, flown in Earth's orbit, have been providing regular observations of upper-atmosphere temperatures. Currently, several satellite-based techniques exist to measure these temperatures. Most techniques are based on radiance measurements of the Earth’s atmosphere at infrared or microwave wavelengths, such as those derived from measurements from the IASI and AMSU-A instruments onboard Metop satellites.

Atmospheric radio occultation
Atmospheric radio occultation relies on the detection of a change in a radio signal as it passes through a planet's atmosphere, i.e. as it is occulted by the atmosphere. When electromagnetic radiation passes through the atmosphere, it is refracted (or bent). The magnitude of the refraction depends on the gradient of refractivity normal to the path, which in turn depends on the density gradient. The effect is most pronounced when the radiation traverses a long atmospheric limb path. At radio frequencies the amount of bending cannot be measured directly; instead the bending can be calculated using the Doppler shift of the signal given the geometry of the emitter and receiver. The amount of bending can be related to the refractive index by using an Abel transform on the formula relating bending angle to refractivity. In the case of the neutral atmosphere (below the ionosphere) information on the atmosphere's temperature, pressure and water vapour content can be derived giving radio occultation data applications in meteorology.[1]

At beginning of this millennium, a relatively new technique, called GNSS Radio Occultation, was introduced. This technique derives atmospheric temperature profiles from bending of GNSS signals caused by atmospheric refraction. The unique long-term stability of these profiles makes them excellently suited for studying changes in tropospheric and stratospheric temperatures over long time period.

Figure 4 shows vertical distributions of monthly-mean temperature anomalies between 8km and 40km for the low-latitude region between 20°S and 20°N over the period 2002 to 2022. The title of the plot is dry temperature as an indication that the temperatures are retrieved under the assumption that the humidity is negligible. This is also why the lowest 8km is masked out. Sampling-error corrected data were selected from the data files, and we used the 10-year period 2007 to 2016 as reference for the anomaly calculations.

The figure clearly reveals a lot of structure in the temperature data. The height of the tropopause is seen as a relatively sharp boundary around 18km. Below about 18km, in the troposphere, the temperature anomalies exhibit variations associated with the ENSO and other modes of variability. ENSO is the most influential natural climate pattern on Earth, it refers to three to seven years repeating phases of warming (El Niño) and cooling (La Niña) sea surface temperatures across the tropical Pacific Ocean (NOAA, assessed 20 Jan 2023). Above about 18km, in the stratosphere, the temperature anomalies are dominated by the so called Quasi-Biennial Oscillation (QBO) — these are the regular oscillations in the equatorial stratosphere with an average period of ~28 months. Any long-term trend in the data must be detected against a background of variability at a range of different time scales, which is why the time series need to be sufficiently long.

One set of findings of the RO contribution is summarised in Figure 6, which was included in Chapter 2 of the IPCC WG1 AR6 report (IPCC, 2021). It tells us that the observed warming rates are faster in the tropical upper troposphere than at, or near, the surface. Furthermore, the trends found in RO data agree reasonably well with the trends found in AIRS data and in radiosonde data, apart from a discrepancy between RO and radiosonde data around 15km. Steiner et al. (2020) suggests that this discrepancy is reduced to near zero if the radiosonde data is restricted to only high-quality instruments.

Another set of findings is presented in the Technical Summary of the IPCC WG1 AR6 report. In this summary the RO observed trends are related to two different climate modelled projections (see Figure 7). The similarities between the observations and the climate model projections are pointed out in in Chapter 2 of the IPCC WG1 AR6 report (IPCC, 2021) as follows:

In the tropics, since at least 2001 (when new techniques permit more robust quantification), the upper troposphere has warmed faster than the near-surface (medium confidence). There is medium confidence that most CMIP5 and CMIP6 models overestimate the observed warming in the upper tropical troposphere .

The conclusions from these assessments clearly show the potential of the RO climate data records to contribute to an improved understanding of the atmospheric changes associated with global warming.