This is a brief description of the format of the adjustments files. All of the files are
formatted identically. They provide the results of our adjustments procedure (for
further details of this procedure see Thorne et al., 2005). Each line has the following
ASCII code format

'(I7,I5,I8,I8,I8,13(F8.3),2X,A)'

In column order these variables are:

1. WMO station identifier
2. Pressure level (hPa)
3. Time assigned to the break point (YYYYMM) - where MM is first month of season
	i.e. 196603 is March (spring) 1966.
	
Note: Columns 4 and 5 are empty if a deletion is assigned rather than an adjustment. 

4. Last point in time series used to create the adjustment factor estimate (YYYYMM).
5. First point in time series used to create the adjustment factor estimate (YYYYMM).

If deletion is being applied:
6. Missing data identifier of -999.

Otherwise:
6. Median estimate of all adjustment factors calculated - this is the adjustment applied.
7. Mean estimate of all adjustment factors calculated.

8-18. percentiles of the adjustment factors calculated under our approach:
	(1,2.5,5,10,25,50,75,90,95,97.5, and 99)

19. Metadata type associated with the assigned breakpoint. Only static metadata
	events (known timing and type) are considered. Cases where no static
	metadata exist to support the break-point are assigned "none". In reality
	metadata is imperfect and it is likely that more are associated with metadata
	than is indicated by this strict association criteria.

Peter Thorne 31/01/05