Analysis of low-level clouds and precipitation

Low-level clouds are partitioned into a large number of sub-classes, each depending on CH1 reflectance and the brightness temperature difference CH3-CH4. By use of the resulting sub-classes, a specific fog/stratus product is created which in addition to detection also indicates the thickness (qualitatively) of the clouds.

The display of precipitating clouds are treated in a similar way to infer the probability of three precipitation intensity categories (light, moderate and heavy precipitation). This specific precipitation analysis is based on threshold tests in channels 1, 3, 4 and 5. An example is given below. Remember, it is the same classification algorithm that produced the cloud classified image shown above but now presented with a special look up table dedicated for precipitation detection.

Click in the image to compare with the RGB image using channel 1, 2 and 4.