We have a West declination in West Virginia, which means that a magnetic compass points to the West of True North. The declination amount is subtracted from a magnetic compass reading in order to get the true bearing.
To specify declination in survex, you give the reading that the compass would read when pointing to True North. In our case, that means a positive value (because it is subtracted from the reading). So our "negative", or West, declination would be specified in survex with a positive number like this:
*calibrate declination 8.983
Despite claiming to use a similar data format, therion calculates true bearing the "right way", the offset of Magnetic North from True North, which is opposite the way that survex does. The practical consequence of this is that if you use the declination value from .svx files in .th files, you must change the sign in therion. Specifying the same West declination in therion is:
declination -8.983 deg
You can calculate declination for any point at any given date in time according to the IGRF geomagnetic model at this NOAA website.
Walls is smart enough that so long as you provide a location for the cave entrance and a date for every survey, it uses this geomagnetic model to automatically apply declination to every survey shot. Update: According to the therion book (PDF), therion (0.5.0 and higher) can do this also, provided that you specify a location, date, and don't manually specify any declination - though I have been unable to get this feature to work properly.
Interestingly, Bob Thrun analyzed a month's worth of actual measured magnetic declination taken at one-hour intervals and came to the following conclusions about short-term declination fluctuations (ie. not predicted by the geomagnetic model):
- Declination can change up to a quarter degree over the course of one day.
- There is some daily periodicity to declination change.
- Some days are more "noisy" or "quiet" than others, ditto for some locations.
- Declination is also affected by sunspot activity.
Since Bob's investigation in 1997, we now have declination data available in one-minute intervals, which could be studied in a similar way.