NCL Home> Application examples> Models || Data files for some examples

Example pages containing: tips | resources | functions/procedures

POP Vectors

POP data that have 2D lat/lon coordinates can be plotted directly in physical space. To do this, simply read in the coordinates and assign them the following attributes:
```
lat2d     = f->TLAT
lon2d     = f->TLONG
t@lon2d   = lon2d
t@lat2d   = lat2d
```
popvec_1.ncl: A basic vector plot.

UVEL and VVEL are NOT zonal and meridional velocities in the lat/lon sense. They are velocities with respect to the model grid, which is locally rotated by ANGLE radians with respect to lat/lon directions. So to convert to zonal & meridional velocities, you need to project UVEL and VVEL onto zonal & meridional directions. Here's how :
```
true_zonal_velocity = cos(ANGLE) * UVEL - sin(ANGLE) * VVEL
true_merid_velocity = sin(ANGLE) * UVEL + cos(ANGLE) * VVEL
```
If you wish to convert regular lat/lon data to the pop grid, the following is used:
```      uPop =  u*cos(rot) + v*sin(rot)
vPop = -u*sin(rot) + v*cos(rot)
```
popvec_2.ncl: Color vectors and use a simpler thinning scheme.

vcMonoLineArrowColor = False, Turns on the color vectors

vcMinDistanceF will thin the vectors just like the subsampling in example 1. This is a easier method.
popvec_3.ncl: A zoom plot

vcRefAnnoOrthogonalPosF = -1.0, Moves the reference vector up into the plot.
popvec_4.ncl: A different projection. You are not limited to cylindrical equidistant projections. There are several projections to choose from.

This is an orthographic projection.