IDL sees the difference between snow and clouds from space

J.M.Krijger, I. Aben, H. Schrijver
SRON, Netherlands Institute for Space Research, Utrecht, Netherlands

Satellite-based passive remote sensing is commonly used to derive global information about the composition of the Earth's atmosphere, such as total columns of greenhouse gasses. The presence of clouds can strongly affect the observation of constituents in the troposphere, because clouds effectively screen the lower part of the atmosphere. When clouds are not properly accounted for, and especially when a significant part of the airmass of interest is below the cloud, (large) errors are introduced. Therefore, cloud detection algorithms are of crucial importance in satellite remote sensing.

One of the few satellites currently monitoring the Earth's atmosphere is SCIAMACHY on ENVISAT, which measures different trace gases including those most abundant in the troposphere (e.g. CO2, NO2, CH4, BrO, SO2). Current employed cloud algorithms, however, suffer from the inadequacy of not being able to distinguish between clouds and snow-covered surfaces, as they were developed for older satellites which could only observe in visible wavelengths where both clouds and snow are completely white from space. As a result these areas are always flagged as clouded, and therefore often not used, while such areas, e.g., northern Europe during snowy winters, are regions people are very interested in. Using IDL's unique interactive structure and array management capabilities a method was developed to detect clouds simply and use infrared wavelength measurements to make a distinction between clouds and snow-covered surfaces at SRON.

Figure 1. Earth surface as seen from space with the SCIAMACHY satellite in the visible wavelengths. The image has been constructed from a full year of observations, yet still at some locations no cloud-free observations were obtained, showing as patchy white areas.

Figure 2. Earth surface, but now in infrared using fake colors. The image is only constructed from two months of data (February/march 2003), thus leaving many clouds, clearly visible over the oceans as green-grey. But note that large snow-covered areas, such as Antartica and the northern (winter) latitudes show as purple.

In summary, using the 'whiteness' or saturation (as determined from IDL color-routines) of an observation allows to detect clouds and snow-covered surfaces indiscriminately, as both are white and have strong contrast to the darker background earth surface. (see Figure 1). The behavior in the infrared allows distinguishing between clouds and snow-covered surfaces, as shown in Figure 2.

The above method is described in more detail in the scientific paper by Krijger et al, ACP, 2005. The method is currently employed by different international research groups.