During the past few decades, the Sahara Desert has become even hotter. Satellite observations suggest that this warming has led to a rise in the frequency of extreme storms in the Sahel region of West Africa. See Letter p.475
One of the most frequently asked questions regarding climate change is how a warming climate will affect weather in the future. Many disastrous weather events in the past few decades, including Hurricane Katrina (2005) and Hurricane Sandy (2012), have driven scientists to seek a better understanding of the occurrence, frequency and intensity of such events. For example, there has been debate over whether warming will lead to an increase in the number of intense tropical cyclones1,2. A major obstacle in reaching a conclusion from these discussions is that extremely destructive weather events are rare, making it difficult to obtain robust statistics. On page 475, Taylor et al.3 make progress in this direction. They use 35 years of satellite observations to show that there has been a persistent increase in the frequency of extreme storms called mesoscale convective systems in the Sahel — the semi-arid region to the south of the Sahara Desert (Fig. 1).
It is always tricky to identify a long-term trend from satellite observations, because of complications associated with calibration. In particular, the instrumental sensitivity of a satellite can vary if its sensor wears out or if its orbit drifts. When multiple satellites are involved, the concern is with the compatibility of calibrations between sensors that have different resolutions, geometries and sensitivities: newer satellites usually have higher resolution and sensitivity than older ones.
Taylor and colleagues analysed satellite images of the Sahel produced by thermal infrared sensors on board the Meteosat geostationary satellites over 35 years (1982–2016). The authors address the calibration issues by downgrading all the satellite images to a coarser resolution. They then focus on populations of clouds that are large (with areas exceeding 25,000 square kilometres) and cold (with temperatures below −40 °C), because these are typical features of mesoscale convective systems4. This approach removes, to a reasonable extent, the possible biases caused by different sensors, making the statistical analysis more convincing.
The authors found a three- to fourfold increase in the number of intense mesoscale convective systems in the Sahel in recent years, compared with 1982. If this trend continues, it will have implications for both agriculture and infrastructure in the region, because these storms are responsible for about 90% of the Sahel's rainfall5.
Over the past two decades, the effect of a warming climate on the organization of convection has been a topic of debate6,7. In particular, it has been speculated that convection will become more organized in the future — under the assumption that a warmer climate would give rise to an environment that supports less entrainment (mixing of dry environmental air into clouds), a stronger updraft (small-scale currents of rising air) and more-intense convection. If true, this could lead to an increase in the frequency of extreme weather events, because these are often tied to large, intense convective systems. To go beyond speculation, however, it will be necessary to understand the distribution and characteristics of extreme precipitation events across many different regions. This is because there will be substantial regional variations in how the atmosphere responds to the changing climate7.
Taylor and colleagues show that the number of intense mesoscale convective systems in the Sahel is highly correlated with global land temperatures. Because Sahelian temperatures have not risen during the past few decades, the authors propose that their results are instead caused by an increase in the temperature gradient across the region, which is driven by a warmer Sahara. They argue that this warming has led to increased convection through enhanced wind shear (the difference in wind speed over a relatively short distance, either vertically or horizontally, in the atmosphere) and changes to the Saharan air layer. These are reasonable speculations that could probably be validated using simulations of the regional climate. The authors' conclusions confirm the complicated nature of how regional weather patterns respond to climate change.
There are two key messages from Taylor and colleagues' work. First, the authors have demonstrated that severe weather events have substantially increased in certain regions over the past few decades. This observation is clear evidence of a variation in weather patterns, and is likely to be related to climate change. Second, the authors have shown that, with a careful analysis, satellite observations can be used to monitor such long-term variations.
In addition to the infrared images used by Taylor et al., observations of microwave radiation8, dating back to the mid-1980s, can reveal details about the distribution of precipitation on Earth's surface. Furthermore, space-borne radar9,10 has been in orbit since the late 1990s. Observations from these satellites contain a great amount of information about the process of precipitation and the structure of precipitation systems. It is therefore anticipated that more evidence of variations in weather patterns in different regions will be revealed by these valuable observations.