refactor using 1981-2010 as the reference climatology.

[cmahony]

Currently the reference climatology is nominally 1961-1990. However, the PRISM and climr mosaic are 1981-2010. this introduces errors into the reference climatology during translation to 1961-1990. once the climr mosaic is ready, we would ideally change the reference climatology to 1981-2010, with functionality to calculate normals for other periods.
However, that's not a small task since it will require recalculating and re-uploading all GCM anomalies. an alternative is to maintain the 1961-1990 reference period and document the errors.

[cmahony]

the approach we likely will take is to calculate a reference period delta raster for each GCM simulation and then apply those to the time series and periods of that simulation. This could be done in the database, rather than having to replace the data.

[cmahony]

@cmahony to investigate the scale of the problem (look at the transfer anomalies and see if there are obvious artefacts).