mix/mix.f90 - CLaMS

The program mix

The options for the execution of mix are set in the configuration file mix.inp:

1)  traj_EM7_2d_97022312_97022412.nc ! trajectory file
2)  ../example        ! its directory
3)  y                 ! one/more than 1 tracer n/y
4)  1                 ! /0/1/ = /nothing/implicit mixing
5)  1.5               ! critical Ljapunov exponent (1/day)
6)  2. 2. 350. 1300.  ! fac1,fac2,th1,th2 (r_limit=fac*r_max) fac1/fac2 outside/inside (th1,th2)
7)  50.               ! maximal vert. dist betw aps (in K)
8)  1  0.0            ! number of adaptation loops
9)  i                 ! rdfpv (i-initialisation time/s-start time of traj/e-end time of traj)
10) 0                 ! 0/1/2 = grid_switch for 3d runs, original grid/shifted grid/old reference
11) 1                 ! number of sublayers

Remarks:

Trajectory file with or without chemistry. Using this file the positions of the APs at the end of the advection step together with their mixing ratios are read. The positions of the APs at the beginning of the advection step are derived from the init-file.
The directory of the traj/chem-file.
n - only 1 tracer in the variable TRACER
y - n species + variable TRACER
0 - no mixing (note that in the 3d-version the boundaries will be updated)
1 - 2D or 3D mixing dependent on the traj/chem-name (2d or 3d)
critical Lyapunov exponent, for 2d (3d) runs use $\lambda_c\Delta t\approx$ 0.8 ( $\lambda_c\Delta t\approx$ 1.2) with 6< $\Delta t$ <24 hours.
This factor is important if the initial distribution of APs is cut by lat_min or/and lat_max. In such a case, "empty regions" from outside of the considered domain can be advected into the region between lat_min and lat_max. To avoid that such "holes" will be filled with APs, the adaptation procedure is limited to distances between NNs with r<r_limit. Of course, after sufficiently long time these regions will be stretched into thin filaments with the thickness of the order r_limit and, consequently, filled with the APs.
fac1 is used outside the given range (<th1 or >th2) and fac2 is used inside this area (>th1 and <th2)
Vertical limitation of mixing, i.e. APs which are separated by a vertical distance larger than this critical value (in K) are not mixed (recommended value: larger then the thickest layer).
Number of adaptation loops replacing the NNs by a new one, if r<r_c^-. Sensitivity studies show that 1 adaptation loop removes about 95% of all APs which match this criterion (recommended value 1).
i - PV as a passive tracer. The initialized PV fields are advected and mixed (useful only for 2d applications)
e - use analyzed PV fields which are read from the end points of the trajectories and are derived meteor. data
Intern parameter defining the vertical grid (0=original, 1=shifted, 2=shifted (old version, only if #sublayers=1))
Number of sublayers

Furthermore the variable STATE describes the details, how the mixing procedure affected the particular APs:
0 - this AP was not affected by the mixing procedure,
1 - NN to the considered AP was eliminated,
2 - the considered AP was included because the distance between the NNs was too large,
3 - next neighbor to the considered AP was eliminated because there were too many APs after the the new APs were included (after they were set to STATE=2).