In the course of 1997, DFD evaluated and tested a variety of methods for determining ground temperatures from satellite images, and has now developed an appropriate procedure which is intended to lead to routine generation of maps with mean land temperatures twice a day for all of Europe. The image on this page is one of the first results of this activity.

For a higher resolution image click here (~280k)

The parts of the Alps covered with snow or glaciers are shown in white, and a red line overlay shows national boundaries. The colors indicate temperatures from 10 degrees C (blue) to 22 degrees C (rust). The enormous temperature difference between northern Italy and land to the north of the Alps is immediately evident, and particularly noteworthy are the finely resolved temperature distributions, whether in the various river valleys, in the distinction between cool forested regions and slightly warmer cropland, or surrounding the heat islands, such as in the upper Rhine valley near the intersection of Germany, France and Switzerland. All in all, a thermal image of amazing variety.

For each day between February 18 and 28, 1998, three afternoon images from the AVHRR (Advanced Very High Resolution Radiometer) on the US weather satellite NOAA-14 were processed and combined, by selecting for each pixel the one recorded under the best atmospheric conditions. The 11 resulting daily composites were averaged to obtain the mean afternoon temperature for the entire observation period.

Knowing the temperature of the land surface is essential information for a wide range of scientific studies as well as for practical decisions related to climate or geography. Examples are determining the agricultural suitability of particular regions, or monitoring the moderating influence of vegetation on temperature. Such information is typically available from meteorological stations which regularly measure the temperature at predetermined locations at an altitude of 2 meters. Although this solution is adequate for heavily populated areas with a dense station network, for most parts of the world equivalent information is lacking.

Remotely sensing surface temperatures from instruments located in space offers the significant advantages of systematic, consistent, global collection of data which can then be made available worldwide, rapidly and continuously. However, calculating ground temperatures from these data involves not only correcting for atmospheric disturbances but is also dependent on the kind of land cover and on changes to the vegetation over time. The complexity of this task has meant that until now global daily land surface temperatures obtained from satellite data have not been available.

For information about the image processing, contact Ms. P. Tungalagsaikhan.