QUESTIONS

1. Select for isotopes when reading HITRAN?
   (Currently I am not.  Fix =  rewrite sread and hitran)

RY Don't select for isotopes

2. What frequencies to use for CH4, CH4a, CH4b, CH4c, C2H2, C2H2a, C2H6,
   HCNb, HCNa? 
   (Currently I am using the limits from the HITRAN paper.
    Fix = Edit arrays in sread)

RY For Jupiter, Roger used
   CH4-A  1200-1400 cm-1
   CH4-B  2000-3500 cm-1
   CH4-C  3500-5000 cm-1
   CH4-D  5000-6660 cm-1
 
   C2H2    700- 800 cm-1


3. Why does the Yelle 1991 paper say that Appleby ignored interactions
   between vibrational levels?  I just read that paper for the
   Uranus paper, and he spends pages on vibration-vibration transitions.
   (Currently just wondering if I'm missing something)

RY Appleby did not allow the vibrational levels to interact.  He treats
   only transitions between the excited vibration levels and the
   ground state.  As a consequence his expression for the source
   function for the nu-4 band does not include a term that represents
   cascading from higher vibrational levels.

LY This answer does not make everything clear for me.  Besides just
   wanting to recognize right from wrong (since VV terms are important
   all over the outer solar system), I'd want to understand the differences
   in your approaches if I'm ever involved in modifying the cascade heating
   terms in the GCM.

   Appleby specifically discusses transitions from, eg, modes in the
   nu3 group going to 2 nu 4, followed by 2 nu 4 to nu4, and nu4 to ground.
   He specifically states that "collisional relaxations from, say, n nu4 
   directly to ground, are assumed to be negligible compared to 
   n nu4 -> (n-1) nu4." (p. 360).

   Although you include the cascade term, theta, in equation 10,
   and emphasize it's importance at the end of section 3.1,
   it seems not to be included in the appendix.  If it were, the
   vector K would have an additional term

       theta B_i^d/B^d (1+eps^d)

   I have not yet looked at your cascade code to see if there
   are differences between formulation and implementation.

4. Is there a way to tell the pre-compiler to skip over large chunks of code?
   (Currently using imicro to skip over chunks of code)

RY Not that I know of.

5. Is there a way to define a boolean in Fortran, so you can ask
   if (itest)
   instead of
   if (itest .eq. true) ?
   (Currently using the second form)

RY Not that I know of.

LY Found it.  Variable type is "logical".

[ ---------------- end of first email exchange with RVY, 26jul99 ------- ]

6. For timing test, what is a way to get the system time in 
   Fortran?  If there isn't a function to do this, I can
   write an ugly one if there is a way to call unix commands from
   inside a Fortran program.

7. For accuracy tests, what's the target accuracy on the cooling rates?

8. I believe eq 33 should read B0 = 1(2 k0) * (the integral), and
   that the upper limit on the integral in eqn 32 and 36 supposed to be
   +infinity.  This is what the code does, I believe, and also what
   I derive.