Fog or fog enhanced with smoke has been shown to reduce visibility to levels where diving is unsafe. Numerous pre-dawn accidents attest to this fact. The solution is to be able to slow or redirect traffic to avoid unsafe speeds in low visibility conditions. Even more desirable would to be able to anticipate those emerging conditions. It is impractical to have instrumentation at all the possible fog locations. Thus, some constantly vigilant mechanism would be desirable. Satellites offer such a capability. To do this, the instrument must be able to at least recognize fog from space. Low earth orbiting satellites offer the best opportunity because of their greater proximity and therefor better spatial resolution. But they are orbiting the earth and are only able to view a given location a few times in a 24 our cycle. However the geostationary (GOES) satellites can see up to a hemisphere constantly. Their drawback is that they are about 23,000 miles away so resolution is about spot about 8 km on a side. An area of fog is usually much smaller than that so the ability to detect a local foggy spot is reduced. The method employed is to measure the upwelling radiation for two different infrared channels. The most useful is Channel 4, which is centered at 11 microns, and channel 2, which is centered at 3.9 microns. For a clear night and a given surface temperature, they will each measure the a different frequency in the Plank function resulting from the upwelling radiation, both inferring the same surface temperature. However, if there is intervening cloud layer, it will intercept, absorb and reradiate at the lower temperature of the cloud, and the signal received will be weaker. Since Channel 4 is less sensitive to the absorption by water, it will more correctly measure the surface temperature while channel 2 will typically infer a colder temperature. Often for the case of fog, this is less than two degrees Kelvin. First the procedure had to be calibrated with know temperatures and fog conditions. From this it was determined the range of brightness temperature difference that indicated the presence of fog. Unfortunately, many times the same brightness temperature difference was found when there was no fog. In addition, it was found that the pixels of the two channels were not geographically colocated. This required development and implementing an algorithm of correcting channel 2 so the images and data between the channels were collocated. These results are quantified and corrections applied which enabled fog climatology with higher resolution for the State of Florida than any previous fog climatology. However, it is shown that the next generation of imager will bring at least 8 times more resolution and has the potential of truly making an improvement in the climatology and fog forecasting from the use of satellites. (Author/publisher)
Abstract