By invoking the time, the predicted depth of contamination
layer is first calculated from some TBD coefficients, stored in a depth
extension. *Ad interim* we assume a linear growth from a reference
point of form:

Depth = Growth . (Time - Reference Time)

The `attenuation` extension then has linear attenuation coefficients versus
energy appropriate for the witches brew building up on the CCDs. From the
depth value, a transmission factor is thus obtained.

Michael Smith 2011-09-20