Models Warm the Lower Troposphere Too Much:
A Fingerprint Test with Updated Data

Testing of climate model results is an important but difficult problem. One of the key model results is the presence of a tropical troposphere "hotspot" in which the troposphere warms faster than the surface under conditions of enhanced greenhouse gas forcing. Previous studies have produced disagreement over whether data were consistent with models on this question. In this study, the authors made several advances by doing the following: 1) enhancing the data for surface trends, 2) extending the data to a 31-year length, 3) evaluating the wind-based temperature estimates, and 4) clarifying the meaning of "best estimate" multi-data warming trends from data and models.

Two prior studies had derived tropospheric temperature trends from the Thermal Wind Equation (TWE) -- which uses radiosonde measurements of wind speed to calculate temperature -- on the theoretical basis that warmer air should move faster than cooler air. They found that there were biases in the data for this type of calculation. For example, particularly for older radiosonde observations, on days when the upper wind was stronger, the balloons would tend to blow out of receiver range. This created a bias by causing missing data for high winds for older observations, leading to a spurious warm trend over time. Overall, the TWE-based trends were three times greater than trends derived from all other types of data. In addition, they did not agree with other wind data, and were also based on much sparser data. This type of data was therefore not used in the authors' analysis, which also identified a small warm bias in the RSS satellite data that was explained by Christy and his colleagues.

The next innovation was to use the Scaling Ratio (SR), which is the ratio of atmospheric temperature trend to surface temperature trend. The SR attempts to factor out the effect of the lack of actual (historic) El Niño's or other oscillations in climate model runs, and different such simulated events in different computer runs.

The nine researchers found that the SR for real-world data was 0.8 ± 0.3, whereas the model simulations had a SR of 1.38 ± 0.08 (a significant difference). That is, the data show a lower rate of warming for the lower troposphere than for the surface (though not statisically different), whereas the models show amplification, as theorized. In fact, the SR value for the middle troposphere data was 0.4, which is even more different from the model predictions. Only the SR for RSS data, which has the documented warming bias noted above, overlaps with any model SR results.

This study thus suggests that current state-of-the-art climate models have something fundamentally wrong with how they represent earth's atmosphere.