How can I plot values in worldmap

Hi. I am new to matlab and I want to plot different values in the arctic region.

I have a set of longitudinal and latitudinal degrees and valuees for each point. I want to plot the ice concentration in the arctic, and have several points with information. But how to plot it?

I get to se the north pole and the coast lines, but how to go from there to get a map with all the information in it?

( I could not find a region called arctic either, so I chose North Pole ) * worldmap('North Pole')

load coast

plotm(lat,long)*

This is all I have.

Hope to get some help here,

-Elizabeth

Answer 1

[Angus]

Hi,

you can check out axesm and try executing "maps" to see a list of available projections and IDs to use with axesm. Depending on the format of your data you should be able to find a projection that works well (maybe the universal polar stereographic).

load coast
axesm('ups')% include other axesm options as needed to define lat lon range

framem on;
axis off;    % Turns off surrounding box
tightmap

pcolorm(latd,lond,mapdata) % pseudocolor plot "stretched" to the grid
geoshow(lat,long,'Color','k') %add continental outlines

lat and long are from the 'load coast', latd and lond are 2d matrices of the same dimensions as mapdata.

Unfortunately I havn't worked with the worldmap function so I dont know the options there.

Best luck, Angus

[Angus]

Short answer: try pcolorm to display your data, it should work with worldmap.

[Elizabeth]

I found that using geoshow would help me abit. I get to plot the vector I am trying to plot, but not the corresponding value of the vector. In other words I get geoshow(LAT(1,:),LON(1,:),'marker','.','color','red')

I want to plot this area as a contour some how, and with a value to the contour ( a z value ) how can I do this?

[Angus]

I dont think geoshow is what you will want to use to display your actual data, it works well for drawing out the coastlines though.

If you haven't yet tried pcolorm give it a go and see what you think, but if contours are what you are after then contourm(lat,lon,ice) or contourfm(...) for filled contours are good options to look into.

Answer 2

[Chad Greene]

Elizabeth:

Here are a couple of suggestions:

1. Initialize your map with ncpolarm. http://www.mathworks.com/matlabcentral/fileexchange/30414-ncpolarm

2. Plot values with scatterm or pcolorm. If you use pcolorm, by default it sometimes includes a grid which can obscure your data. Get around this by following it up with shading interp or one of the other shading commands.

Hope this helps.

[Chad Greene]

Also, if your data points are not a big grid, but in the form of a vector of lats, a vector of longs, and a vector of ice concentration values, try

scatterm(lats,lons,5,iceconcentrations,'filled')

Explore the different options for scatterm. The 5 above is simply the marker size. 'filled' means they're filled circles. You may want a different marker or settings.

[Elizabeth]

how do i get the plot to distibute all the vectors? I have a matrix of [nxm] for both long and lat. and a vector of values for the concentration of [nx1]

when I try to put the plot into a for-loop where n= 1:length(the concentration matrix) I get a bunch of errors

this is what I wrote :

figure('Color','white')

worldmap('North Pole')
load coast 
plotm(lat,long)

for n=1:length(CT);

pcolorm(LAT(:,n),LON(:,n),CT(n));

end

and I get the following errors :

Error using expandScalarInputs (line 60) Non-scalar inputs must have NaN-separators in identical positions.

Error in parseDistAzInputs (line 95) [lat1, lon1, lat2, lon2, insize] = expandScalarInputs(lat1, lon1, lat2, lon2);

Error in distance (line 96) [useGeodesic, lat1, lon1, lat2, lon2, ellipsoid, ...

Error in eqdazim>applyForward (line 118) [rng, az] = distance('gc', lat0, lon0, lat, lon, radius, 'radians');

Error in doApplyProj (line 31) [x, y, savepts] = mproj.applyForward(mproj, mstruct, in1, in2, objectType);

Error in eqdazim>applyAzimuthalProjection (line 37) outputs = doApplyProj(mproj, varargin{:});

Error in eqdazim (line 25) varargout = applyAzimuthalProjection(mproj, varargin{:});

Error in mfwdtran (line 37) [x, y, z, savepts] = feval( ...

Error in surfacem (line 82) [x, y, z, savepts] = mfwdtran(mstruct, lat, lon, alt, 'surface');

Error in pcolorm (line 49) surfacem(lat, lon, Z, varargin{:});

Error in map (line 11) pcolorm(LAT(:,n),LON(:,n),CT(n));

The LAT and LON matrix will vary for each week I pull out information from my database. And I will have one CT ( concentration) value for each column in the LAT and LON matrices.

The column length of both LAT and LON will change for each column, where the matrix has the dimension of the longest column and the rest of the columns that are not as long are filled with 0.

[Chad Greene]

Elizabeth,

It's strange that your sea ice concentration matrix data is a vector while your lats and longs are [m x n] matrices. In what format is your raw data? From the NSIDC it looks like their data comes in a form where the size of lat, long, and CT are all [m x n] matrices. For example, we can get an arbitrary month from here: <ftp://sidads.colorado.edu/pub/DATASETS/NOAA/G02202/north/monthly/> Then open it up in Matlab via

lat=ncread('seaice_conc_monthly_nh_f13_200308_v01r00.nc','latitude');
lon=ncread('seaice_conc_monthly_nh_f13_200308_v01r00.nc','longitude');
CT = ncread('seaice_conc_monthly_nh_f13_200308_v01r00.nc','seaice_conc_monthly_cdr');

Plotting this should be straightforward with pcolorm. I don't have the mapping toolbox on the computer I'm using today, but I can take another look at this tomorrow if you're still trying to get it to work then.

[Rob Comer]

Elizabeth,

What you saw was actually one error, but one error very far down the call stack. Functions like pcolorm (and surfacem, which is right below it in this case) try to validate their input and throw helpful errors if it's not correct. But your input seemed to pass the validation in surfacem and then get into trouble much later (way down the stack). If you can send me LAT, LON, and CT, then I can investigate. Or open a support case with MathWorks Technical Support: http://www.mathworks.com/support/contact_us/index.html and ask them to escalate it to Development.

-- Rob

Answer 3

[Angus]

From your previous comments:

~~~~

"The LAT and LON matrix will vary for each week I pull out information from my database. And I will have one CT ( concentration) value for each column in the LAT and LON matrices."

~~~~

This seems like it will be hard to deal with no matter what plotting method you choose.

If you have access to the algorithm that pulls the information I would start with changing that; instead of ordering by CT values just pull a grid box area and fill any missing values with NaNs.

If not then I would aim to convert the matrices and arrays that you have before trying to plot them. Here are some ideas for the steps to take, if this seems like a reasonable approach then I can help as you go through it.

- find weekly range of lat and lon (max/min)

- construct a 2D grid (meshgrid) for lat and lon according to your range and resolution.

- create a 2D NaN matrix with dimensions of your lat/lon matrices.

- pull CT values from your array and insert them in your NaN matrix at the correct lat/lon.

- use pcolorm(lat,lon,CT) with your new matrices. (or contourm)

Cheers, Angus.