Cloud analysis from satellite data is an important contribution to climate research

Clouds affect both the weather and climate, and they play an important role in calculations of a changing climate. Despite obstructing the view, without clouds we would have it even hotter on earth, as they reduce the overall amount of incoming solar radiation. At SMHI, researchers are studying clouds based on satellite data. 

“For many who use satellite data, the clouds are just in the way. If you are interested in obtaining satellite information on, for example, sea surface temperature or the condition of the earth's surface, then clouds can obstruct the view and make the process difficult,” says Karl-Göran Karlsson. He is a researcher at SMHI and works with the study of clouds using satellite data.

Clouds are important for the climate

Clouds are part of the system that controls the temperature on earth. The clouds in the atmosphere deflect some of the incoming solar radiation. At the same time they absorb some of the radiation leaving the earth and prevent it from traveling further up into the atmosphere and into space. So even if the clouds absorb and retain some heat and thus help to heat up the atmosphere, their cooling effect is currently greater.

“Clouds are important for the climate. If the clouds would completely disappear, it could result in greater climate change, and here we are talking about changes that may correspond to almost a tripling of the carbon dioxide concentration in the atmosphere. Now, this scenario is unlikely, but even small changes in the cloud cover can still have great significance. It is partly this that is behind the wide distribution of the climate model results in terms of the estimate of how much the temperature will rise over the next century. It is therefore interesting to analyse the clouds,” says Karl-Göran Karlsson.

Studying clouds with a satellite

How does the process of studying clouds with a satellite work? It can sometimes be difficult to say what is a cloud in satellite images. Smoke from fire and aerosols, tiny particles in the atmosphere, can look like clouds in satellite images. The thinnest clouds can maybe only be detected indirectly through them casting shadows on the earth's surface, such as on Greenland's snow surfaces.

Most clouds can be identified by studying how sunlight is reflected and the thermal radiation emitted in comparison with how the surfaces of land and sea appear.

“An interesting alternative technique is to study the reflections of light cast by the clouds when illuminated with a laser radar, known as lidar. Clouds are fairly easy to identify with this technology, and these observations are therefore invaluable in the development and validation of satellite-based technology.” 

Since 2006, lidar measurements have been made from the Calipso satellite, a French-American satellite for environmental monitoring.

“At present we are working to create a probability-based cloud analysis that describes the cloud cover in a more objective manner. It also provides an uncertainty assessment in the cloud information. It then makes it easier to use the information for further research, for example, to study how the cloud cover is affected when the climate changes,” concludes Karl-Göran Karlsson.

Satellite image of clouds over Greenland and a cloud analysis
Looking at a satellite image and determining what is or is not a cloud can be difficult. However, through analysing additional satellite information, you can create an image that shows the probability that there are clouds over a specified area. The left image shows clouds over Greenland, which can be difficult to distinguish from snow-covered ground and ice. The image shows information from three of the five spectral bands of the AVHRR sensor. The image on the right shows the probability of clouds in the same area, where the white areas show a high probability and the black areas show a low probability of clouds. This is obtained by using information from two additional spectral bands.