Improved Algorithms for Combining Satellite Imagery and Geographic Basemaps

Traditional algorithms for combining satellite imagery with topographic or vegetation basemaps have been limited to using infrared data. Here, an infrared brightness temperature (T_ir) is chosen to determine a cloud/no cloud threshold. The product created displays a basemap where T_ir is greater than the threshold and IR satellite data where T_ir is less than the threshold. Additionally, a temperature-based transparency is assigned to the satellite data, where colder clouds are more opaque. This method exaggerates cloud coverage over polar-regions and underestimates stratiform cloud coverage over lower latitudes. More recently, we have begun to see visible imagery combined with basemaps using an albedo-based transparency. This method greatly improves overall cloud cover estimation, but is limited to daylight hours.

Two new algorithms are now being used by the University of Wisconsin-Madison for combining real-time satellite imagery with basemaps. The first algorithm addresses the problems of the traditional method by utilizing surface temperatures (T_sfc) from observations or numerical models. Making the simple assumption that T_ir<T_sfc over cloudy regions and T_ir=T_sfc over cloud-free regions, one can create a map of transparency using T_sfc-T_ir, where the opacity of the cloud increases as the value T_sfc-T_ir increases. Comparisons of the traditional and new algorithm will be shown, illustrating the improved estimation of cloud coverage in our global products.

The second algorithm utilizes date/time information to blend visible and infrared imagery. Weighting functions based on location and time are applied to the IR brightness temperature and visible reflectance to calculate a transparency to combine the blended satellite imagery with a basemap.

Examples of these algorithms can be viewed at the following web locations:

http://www.ssec.wisc.edu/~rickk/temp_diff.html

http://www.ssec.wisc.edu/~rickk/vis_ir.html