Problem when wrapping longitude dataset

Hello,

I have a global dataset that I would like to wrap. I want to change its longitude from [0 360] to [-180 180]. I can successfully wrapTo180 the longitude vector, but when I plot my data, there are lines sporadically across the map. There must be something wrong with the data that is making it do this.

Please help!!

I have attached my data so you can give it a whirl

Thanks,

Melissa

[Walter Roberson]

Duplicates http://uk.mathworks.com/matlabcentral/answers/216098-problems-with-wrapto180-map but I have let this one stand because it has the data attached and the other one does not.

Answer 1

[Chad Greene]

Does this work?

load lat
load lon 
load data

[lon,lat] = meshgrid(wrapTo180(lon_model),lat_model); 

worldmap('world')
pcolorm(lat,lon,double(first_model'))

And formatting with the help of rgb, borders, and brewermap,

setm(gca,'ffacecolor',rgb('ocean blue'))
borders('countries','k')
colormap(brewermap(1024,'OrRd'))

[Melissa]

This looks good! Except I would like to plot it using pcolor instead of worldmap, so it stays consistent with my other datasets. Is there a way to get this kind of output using that?

Thanks

[Chad Greene]

Strange! I see the striping you are referring to. Here's a manual workaround that also takes care of the vertical stripe of missing data at the prime meridian:

load lat
load lon 
load data

lon = wrapTo180(lon_model); 
[lon,lat] = meshgrid(linspace(min(lon),max(lon),length(lon)),...
    linspace(-90,90,192));
fm = double(first_model'); 
fm = fm(:,[145:end 1:144]);

% Not sure if you want to do this masking: 
greenland = landmask(lat,lon,'greenland'); 
antarctica = landmask(lat,lon,'antarctica'); 
fm(greenland) = 0; 
fm(antarctica) = 0; 

figure
pcolor(lon,lat,fm)
shading interp

set(gca,'color',rgb('ocean blue'))
borders('countries','k','nomap')
colormap(brewermap(1024,'OrRd'))

axis([-180 180 -82 86])

[Chad Greene]

...and in case you're wondering I used landmask for to turn data for Greenland and Antarctica into zeros.

[Melissa]

This works perfectly! Thank you so much.

However, if I wanted to keep the white vertical line there, how would I do that while keeping the rest of the blue lines out?

[Chad Greene]

I'm not sure why you want to keep that vertical line--it's an artifact of pcolor, which deletes a row and column of data. Before shifting fm with fm = fm(:,[145:end 1:144]);, pcolor would ignore the right-hand column of data, which was the column of data corresponding to the prime meridian. After the shift, pcolor discards the right-hand column of data east of New Zealand. If you want to arbitrarily discard a vertical line of data at the prime meridian, you can do so by

fm(:,144)=NaN;

which gives

[Chad Greene]

Or if you don't want to get rid of any data, you can use imagesc instead of pcolor:

load lat
load lon 
load data

fm = double(first_model'); 
fm = fm(:,[145:end 1:144]);

% Linear arrays of lat and lon:
lon = wrapTo180(lon_model); 
lon = linspace(min(lon),max(lon),length(lon)); 
lat = linspace(-90,90,192); 

% gridded lat/lon: 
[long,latg] = meshgrid(lon,lat); 

% landmask takes ~30 seconds: 
land = landmask(latg,long); 

% Set land NaNs to zero: 
fm(land & isnan(fm)) = 0; 

% Set ocean to -1: 
fm(isnan(fm))=-1; 

imagesc(lon,lat,fm); 
axis xy
set(gca,'color',rgb('ocean blue'))
borders('countries','k','nomap')
bm = brewermap(1024,'OrRd'); 
colormap([rgb('ocean blue');bm])
axis([-180 180 -60 86])