NCL Website header
NCL Home> Application examples> Data Analysis || Data files for some examples

Example pages containing: tips | resources | functions/procedures

Correlations

In addition to the functions demonstrated on this page, there are two more that might be of interest: esacr computes an auto correlation and escorc computes an cross correlation at zero lag
corel_1.ncl: Calculates a cross correlation.

esccr: Calculates the cross correlation between two variables.

dimsizes: Calculates the size of the dimensions of a variable.

corel_2.ncl: The intrinsic NCL correlation function esccr only calculates the positive side of a correlation. This script demonstrates how a pos and neg lag can be created. NCL coordinate variable syntax [ {...} ] is used to specify locations. NCL index syntax [ ::-1 ] is used to reverse the array order.

corel_3.ncl: Demonstrates using the same functions to do a 2D correlation in time.

The data must be reordered to put time as the rightmost dimension.

This example computes the cross-correlation at lags 0,1 and 2. If the cross-correlation at 0-lag only were desired, then it would be more efficient to use escorc.

Since the correlation values will always span -1 to 1, the two end labelbar boxes representing values above 1 and below -1 were removed by setting res@cnLabelBarEndStyle to "ExcludeOuterBoxes".

indices_soi_2.ncl: Read gridded sea level pressure from the 20th Century Reanalysis; use proxy grid points near Tahiti and Darwin to construct an SOI time series spanning 1950-2010; perform lag-0 correlations between the SOI and SLP; SOI and temperature; and, SOI and preciptation. The later uses the GPCP data which spans 1979-2010. To more clearly delineate the main pattern structure correlations between, -0.1 and +0.1 were set to _FillValue.

FYI: The linear correlation between the station based SOI (previous example) and the SOI derived from the 20th Century Reanalysis for the 1950-2010 period is 0.96.