|
The original trajectory calculation started at the beginning of the tracer release, but conditions can change with time and space. We can investigate both of these options to see if it improves the simulation with respect to the ground-level concentration pattern.
- Start by configuring the simulation for the mid-boundary layer (750 m AGL) trajectory by retrieving traj_fwrd_control.txt into the trajectory / setup run menu.
- Then go to the advanced / trajectory / configuration tab, press the Reset button, and then open menu #3 multiple trajectories in time. Set the restart interval to 1 hour, which will cause the model to start a new trajectory every hour from the starting point and time. The second entry, trajectory duration, should be set to 4 hours, which will cause the model to only start new trajectories for four hours regardless of the duration of the run.
- Save all the changes and then run the model, open the trajectory display menu to view the result. The four trajectories, one hour apart, span the tracer release period from 17 to 20 UTC. Each successive trajectory is further south and west, indicating that changes to the flow field during this short 4 hour period could account for at least some of the tracer observed over Pennslyvania.
- Because we already know that a single trajectory level and time is insufficient to capture the plume structure, we can combine the multiple starting heights with starting trajectories in-time. Start with the same trajectory setup used for the three mixed layer trajectories (10, 750, 1500 m AGL), and then save the changes.
- Now open the configuration menu #3 again and change the interval to 6, the duration to 24, and the number of levels to 3, consistent with the three starting heights. What happens in this variation is not intuitive. The input parameters have the same meaning as before, but instead of starting 3 new trajectories every 6 hours at the initial starting location, new trajectories are started at the end of each of all the previous trajectories. The calculation result illustrates how changing wind direction with height within the mixed layer can account for much of the observed tracer dispersion.
- Another way of analyzing the flow field important for the tracer transport is to compute simultaneous trajectories over the whole domain. Open the trajectory setup menu and enter three starting locations, all at 750 m AGL: 38N 84W, 44N 74W, and 39N 83W. The first two points represent the lower left and upper right corner of the domain over which trajectories will be started. The difference between the first point and the third point represents the spacing of a virtual grid from which trajectories will be started. Also reduce the total run time to 6 hours. Save the changes and run the trajectory. If the message pops up that the SETUP.CFG namelist file found!, select Delete file then Run. The namelist file was leftover from the previous calculations and is not needed for this illustration.
- Only three trajectories are shown when the standard model run menu tab is selected. To create a matrix grid of trajectories go to the Trajectory / Special Runs / Matrix tab and you will get a prompt asking if the input was properly setup with three sources defining the grid domain. When the model run is finished, open the trajectory display, and you should see an illustration of the flow field over the domain.
- One element to remember from this section is that when the setup menu was configured for three locations the resulting control file only showed those three points. However, when the model was run through the special runs tab, it invoked a pre-processor step, that converted the three starting locations into a control file with 77 starting locations. Complex control files can also be created manually with any text editor.
The primary result shown in this section was that the surface tracer concentration distribution could be explained in large part by the changing wind directions with height and time within the mixed layer.
|