Validate CRU and ClimateNA time series

[cmahony]

There are big differences between CRU and ClimateNA time series. we need to provide users with an understanding of the pros, cons, and pitfalls of these two data sources. also, we need to evaluate our optimal approach for calculating the 1961-1990 reference surfaces from the combination of these time series with the 1981-2010 normals, as there are potential artefacts here that would cascade through the rest of climr.

CRU has a more credible pattern of change for temperature; the patchiness of ClimateNA is likely mostly station artefacts. relative credibility of GPCC and ClimateNA are difficult to assess; both products are different though.

Note that there is a lot of variability in the trends calculated from individual stations.

The differences between CRU and ClimateNA aren't necessarily large relative to natural variability, but they can amount to a difference in the direction of the trend.

[cmahony]

need to investigate

• station density of CRU over time (it used to be terrible in canada, but may have improved)

[cmahony]

@aseemsharma1 here is the plot that i would like to make as a starting point.

[cmahony]

@aseemsharma1 the code for the plots in this issue is https://github.com/bcgov/climr/blob/devl/tests/manual_tests/check_TimeSeriesVsNormals.R
[we do all our work in the devl branch, or sub-branches.]