Review calculation of average values

[gjr80]

gauge-data.txt includes a number of average values; some are (defined) 10 minute averages otehrs are not defined. The approach taken in calculating these averages is not consistent. Some are taken from archive data only and some use values derived from the archive and until the loop buffer is full later at which time it reverts to loop data derived values.

This is largely an issue that only affects wind related obs.

There is a compromise to be made between complexity, accuracy, consistency and speed of calculation. Consistency should be maintained and speed is probaly not such an issue.