These tests are implemented as part of the pruning process, after the candidate paths have been constructed. The tests are performed in the order that they are listed here.
Removes all tracks that leave a user-specified polygonal area (defined by a list of (lon,lat) coordinates in a file) at any time.
Removes all tracks that don't end within a user-specified distance tol_return of any destination airport (including the one the track started from). Decreasing tol_return will reduce the number of candidate tracks.
Removes all tracks with too much repetition (e.g., if a track makes a loop, returns to the starting airport, then repeats this process multiple times). Currently, a track is defined as not being repetitive if the number of unique waypoints in the track is at least 50% of the total number of waypoints (including the start and end points). The threshold is adjustable.
Removes all tracks that are over land more than a specified percentage of the time (currently set to 10%). The percentage overland is computed based on a subdivided track (adding in N points between each waypoint). Currently N=5. For speed, all points in all subdivided candidate tracks are tested for being inside the land polygons using the intermediate-resolution GSHHS coastline file; the full-resolution coast file is used for plotting. Because this test can be slow, it is done next-to-last.
Removes all tracks that traverse the same set of segments, but in a different order (e.g., a track leaves an airport, makes a clockwise loop and returns, vs. one that follows the same path but counterclockwise.