NCL: Deviations from the Mean

dev_1.ncl: Creates panel plot of TS data and TS deviations from a zonal mean.

TS is dimensioned (/"time","lat","lon"/)

dim_rmvmean is the built-in NCL function that will calculate deviations from a mean. This function first calculates the mean of the data using dim_avg, and then subtracts that mean from the original data. Since dim_avg operates on the rightmost dimension, dim_rmvmean results in the deviation from the zonal mean.

A Python version of this projection is available here.

dev_2.ncl: Reorders the TS data so that time is the rightmost dimension, which produces a deviation from time plot. In version 5.1.1, you will be able to use dim_rmvmean_n (dim_rmvmean_n_Wrap to preserve metadata) to avoid having to reorder the data.

The NCL syntax |: is called named subscripting, and it can be used to efficiently reorder a data array:
anom = dim_rmvmean(ts(lat|:,lon|:,time|:)). In version 5.1.1, you can use anom = dim_rmvmean_n(ts,0).

The zonal average plot is modified by using gsnZonalMeanXMinF, gsnZonalMeanXMaxF, and gsnZonalMeanYRefLine to set the minimum X-axis value, the maximum X-axis value, and the X-axis value where the reference line will be drawn.

A Python version of this projection is available here.