[daip] Re: how AIPS calculates opacity

[Eric Greisen]

Bryan Butler writes:
 > 
 > all,
 > 
 > some notes on how CLCOR and FILLM calculate opacity.  i don't know what the 
 > 'right' answer is - probably neither of these (CLCOR is an outdated 
 > implementation of what used to be in CALC *long* ago, and the FILLM one is too 
 > simple).  they only really diverge for _very_ low elevations, where this all 
 > breaks down anyway.
 > 
 > at the very least, i would suggest changing CLCOR to do the full
 > exponential of opacity (instead of only up to 2nd order terms in the 
 > polynomial expansion), and change the calculation of sine(elevation) [as noted 
 > below - change:
 >        DARG = SINLAT * SINDEC + COSLAT * COSDEC * COS (HRANG)
 > to
 >        DARG = SIN(ELV)
 > ...]
 > 
 >     terms only to 2nd order in the exponential expansion are used.  why?    
 > it's not like it's extremely expensive to just calculate the    exponential.  
 > this is likely a historical artifact...
 > 
 >     where
 > 
 >                    tau
 >       a =  ------------------------
 >                         0.00143
 >            sin(e) + ---------------
 >                     tan(e) + 0.0045
 > 
 >     for opacity tau and elevation e.
 > 

 >        f = sqrt { exp(a) }
 > 

Made these two changes - finally - to CLCOR.  (I am doing archaeology
on my Inbox).

Eric Greisen