cdtdivergence documentation

cdtdivergence calculates the divergence of gridded vectors on the ellipsoidal Earth's surface.

See also: cdtgradient, cdtcurl and ekman.

Back to Climate Data Tools Contents.

Contents

Syntax

D = cdtdivergence(lat,lon,U,V)

Description

D = cdtdivergence(lat,lon,U,V) uses cdtdim to estimate the dimensions of each grid cell in the lat,lon grid, then computes the divergence of the gridded vectors U,V. Units of D are the units of U and V divided by meters.

Example 1: Theory

Here's a simple field of purely zonal wind with a uniform velocity of 1 m/s everywhere around the globe. Use cdtgrid to make a quarter-degree grid and then define the wind field:

[lat,lon] = cdtgrid(0.25); % quarter degree grid
u = ones(size(lat));       % purely zonal wind
v = zeros(size(lat));      % no meridional component.

Here's what the wind vectors and divergence look like for a purely zonal wind of uniform velocity:

D = cdtdivergence(lat,lon,u,v);

figure
imagescn(lon,lat,D)
hold on
borders('countries','color',rgb('gray'))
quiversc(lon,lat,u,v,'k')
cb = colorbar;
ylabel(cb,'divergence (s^{-1})')
cmocean('balance','pivot')

In the figure above, we see that the divergence is zero everywhere. That's because as a parcel of air moves from one grid cell eastward to the next grid cell over, the wind does not change speed or direction, and the neighboring grid cell is exactly the same size as the grid cell it just came from. Everything stays perfectly the same, so there is no divergence.

What about a purely meridional wind? Can we expect the same kind of zero divergence everywhere around the globe?

Using the same quarter degree lat,lon grid from above, define a purely meridional wind of uniform velocity everywhere:

u = zeros(size(lat)); % no zonal component
v = ones(size(lat));  % 1 m/s northward everywhere

% compute the divergence:
D = cdtdivergence(lat,lon,u,v);

figure
imagescn(lon,lat,D)
hold on
borders('countries','color',rgb('gray'))
quiversc(lon,lat,u,v,'k')
cb = colorbar;
ylabel(cb,'divergence (s^{-1})')
cmocean('balance','pivot')

Above, what we get is not zero divergence everywhere. Rather, there seems to be divergence in the northern hemisphere and convergence in the southern hemisphere. That's because the Earth is not a perfect sphere, but is an ellipsoid whose lines of latitude are spaced at near-but-not-quite-equal distances from each other. Play around with the earth_radius function, and you'll see that in a trip from the South Pole to the equator, lines of latitude get closer together. Keeping going, and from the equator to the North Pole the lines of latitude will begin to spread apart again. Check the scale on the colorbar above, and you'll see that while latitude spacing does affect divergence, its role here is still quite small.

Putting together the two examples above, what if the wind is uniform zonally and meridionally everywhere?

u = ones(size(lat));
v = ones(size(lat));

D = cdtdivergence(lat,lon,u,v);

figure
imagescn(lon,lat,D)
hold on
borders('countries','color',rgb('gray'))
quiversc(lon,lat,u,v,'k')
cb = colorbar;
ylabel(cb,'divergence (s^{-1})')
cmocean('balance','pivot')

Above we see that adding zero zonal divergence to a small amount of meridional divergence produces the same result as just looking at the meridional divergence. I suppose that's no big surprise.

Now consider a case where wind speed varies with longitude. Make it

u = lon.^2;
v = zeros(size(lat));

D = cdtdivergence(lat,lon,u,v);

figure
imagescn(lon,lat,D)
hold on
borders('countries','color',rgb('gray'))
quiversc(lon,lat,u,v,'k')
cb = colorbar;
ylabel(cb,'divergence (s^{-1})')
cmocean('balance','pivot')

caxis([-1 1]*1e-2)

Above, in the Western Hemisphere the wind converges as it goes from very strong at the international date line, to zero velocity at the prime meridian. There, at the prime meridian, the wind begins to pick up speed again, making it divergent as the wind begins to effectively pull itself apart.

Also note in the map above that the magnitude of convergence and divergene intensifies close to the poles because at high latitudes the wind accelerations from one grid cell to the next occur over a smaller distance. If you're curious about how grid cell spacing varies around the world, explore the effect with the cdtdim function.

Example 2: Reality and the ITCZ

For this example load some global surface wind data that comes with CDT. Load the 10-meter wind speeds u10 and v10, and keep it simple by just taking the mean surface winds for 2017:

filename = 'ERA_Interim_2017.nc';
u10 = mean(ncread(filename,'u10'),3);
v10 = mean(ncread(filename,'v10'),3);
lat = double(ncread(filename,'latitude'));
lon = double(ncread(filename,'longitude'));
[Lat,Lon] = meshgrid(lat,lon);

The raw data are on a grid that goes from 0 to 360 longitude. I'd rather put the prime meridian in the middle of the map, so let's recenter the grids. This step isn't necessary, but it's a preference:

[Lat,Lon,u10,v10] = recenter(Lat,Lon,u10,v10);

Calculating wind divergence is just as simple with real data as it is with the synthetic data we created in Example 1. The only adjustment we'll make here, for aesthetic purposes, is to mask out land grid cells using the island function:

% Calculate wind divergence:
D = cdtdivergence(Lat,Lon,u10,v10);

% Mask out land:
land = island(Lat,Lon);
D(land) = NaN;
u10(land) = NaN;
v10(land) = NaN;

Now we can plot wind vectors and their divergence just as in Example 1, but this time we'll start with an earthimage base map for context:

figure
earthimage;
hold on
pcolor(Lon,Lat,D)
shading interp
hold on
quiversc(Lon,Lat,u10,v10,'k','density',100)
cb = colorbar;
ylabel(cb,'surface wind divergence (s^{-1})')
caxis([-1 1]*1e-5)
cmocean balance

The map above shows how 10 meter wind vectors converge and diverge around the world. It shows a blue region of convergence just north of the Equator in the eastern Pacific, and that coincides with the Intertropical Convergence Zone (ITCZ), however, the ITCZ is seasonally migratory, so this whole-year average wind field may have the effect of reducing the true intensity of the ITCZ at any given time.

Author Info

This function is part of the Climate Data Toolbox for Matlab. The function and supporting documentation were written by Chad A. Greene of the University of Texas at Austin.