****************************************************************

04/21/2005:

Changes from sami2-0.95a to sami2-0.97:

1. The subroutine 'etemp' has been substantially rewritten.
   The photoelectron heating calculation in sami-0.95a could
   lead to a 'choppy' latitudinal behavior in Te and Ti, and 
   the constant 'cqe' was fixed at 3e-14. The subroutine in 
   sami2-0.97 significantly reduces this 'choppy' behavior, and 
   the constant 'cqe' is now in the namelist so it can be changed. 

2. The grid routine 'grid-0.97.f' has been improved so that
   the calculation of the mesh along the field line is more
   robust. The calculation in 'grid-0.95a' used an iterative
   technique to set up the mesh; the calculation in 'grid-0.97'
   uses a direct solve method.

   Additionally, the transformations between the magnetic 
   coordinate system and geographic coordinate system 
   [subroutines 'gtob' and 'btog'] have been modified. 
   The subroutines used in sami2-0.95a did not translate
   the grid (e.g., via x0, y0, and z0) correctly. This has
   been corrected in sami2-0.97. 

3. The sign of u(i,nfl) in the definition of tvn(i) in the
   subroutine 'transprt' has been corrected; it has been
   changed to '+' in sami2-0.97.

4. The calling sequence of 'photprod,' 'chemrate,' 'chempl,'
   and 'recorate' in the subroutine 'transprt' has been changed
   to improve the computational speed.

5. The do loop ordering in the subroutine 'update' for the
   calculation of ion-neutral collision frequencies has
   been changed from 'do i ... do nn' to 'do nn ... do i'
   to improve computational speed.

6. The constant 1.00 used in the variable 'fac' used in the 
   calculation of ion-N2 in 'update' has been changed to
   1.04.

7. The definition of 'kapi(i)' in the ion temperarture 
   subrountines has been corrected.

8. The 'if' statement in the subroutines 'htemp' and 
   'hetemp' has been corrected. In sami2-0.95a it
   was 'if ( ni .ne. ptop)' in both 'htemp' and 'hetemp';
   in sami2-0.97 it is now 'if ( ni .ne. pthp)' in
   'htemp' and 'if (ni .ne. pthep)' in 'hetemp.'

9. The calculation of the photoionization rates in
   'photprod' has been improved to include zenith angles
   greater than 90 degrees.

10. The parameters 'rexbmid' and 'rexbdel' have been
    removed from the namelist and are not used. Instead,
    the altitude dependence of the E x B drift at the
    magnetic equator is proportional to L^2, i.e., the
    E x B drift increases with altitude.

11. Two parameters have been added to the namelist:
    alt_crit and cqe.

    alt_crit:     The E x B drift is exponentially decreased below
                  this altitude with a scale length 20 km. The
                  default value is 150 km. [This is done to allow
                  rmin to be less than 150 km without using an
                  extremely small time step.]
    cqe:          Constant used in the subroutine 'etemp' associated
                  with photoelectron heating. The typical range is
                  3e-14 -- 8e-14. The higher this value, the lower
                  the electron temperature above 300 km.

****************************************************************