Backward run won't go beyond first time step

[leonardozm]

I am attempting a backward trajectory run from a single point, tracking the trajectory for 744h. When I use the settings from these attached files (CONTROL and SETUP), I get the result in the file 97janfdp. The model reports a full run (Percent complete: 100.0 message), but it seems that the program doesn't go beyond the starting time step.

CONTROL.txt

Starting point and basic settings.

(351 Bytes) Downloaded 314 times

SETUP.txt

Advanced menu configurations.

(313 Bytes) Downloaded 310 times

97jan10_fdp.txt

The output from the model.

(461 Bytes) Downloaded 302 times

[ariel.stein]

Can you please send the MESSAGE file?

[leonardozm]

Hi there, here's the message file:

MESSAGE.txt

(32.97 KiB) Downloaded 319 times

I actually could get the program to work by setting TOUT = 60. For some reason, when TOUT = 180, it won't work. That was the only change I think I made.

[barbara.stunder]

In the MESSAGE file from the run that failed (TOUT=180) there are a lot of these lines
"NOTICE advrng: (kg ,xyr,xy1) - 1 10 10 131 10"
but if you compare it to the MESSAGE file from your successful run with TOUT=60, you can see with TOUT=180 the run failed. Notice also with TOUT=60 there is a line " NOTICE main: Initial time step (min) -60"
Negative because this is a back-trajectory. For TOUT=180 it is -90.

The subroutine 'advrng' (advection range) sets the meteorological subgrid size and location. I need to run some tests to see how the output interval (TOUT) relates to the subgrid and specifically why this run failed.

If you add this line to the SETUP file:
DELT = 60,
this forces a negative 60-minute time step, then the run works with TOUT=180, giving you 3-hourly output in the trajectory endpoints file.