Plotting Station Data using pcolor

I have over 6000 stations with precipitation data (some with "junk", others with realistic values). I have loaded the data and have computed the trends in the precipitation dataset. I have three vectors:

1. ptrend - Contains the trends in precipitation for select stations (2428 x 1).
2. lon - Contains the longitude values for the stations (2428 x 1).
3. lat - Contains the latitude values for the stations (2428 x 1).

Note: Lat and lon are not evenly spaced.

I am trying to find a way to plot these values on a map. From what I understand, pcolor requires that the input data be a matrix and not a vector. So, I tried repmat:

ptrend_rep = repmat(p_trend,[1 length(p_trend)]);

figure(1);
pcolor(lon,lat,ptrend_rep);

However, the figure comes out looking very strange (striations and "garbage"). I also tried meshgrid for the lat and lon vectors.

[lat2,lon2] = meshgrid(lat,lon);

figure(2)
pcolor(lon2,lat2,ptrend_rep);

Still, the figure looks terrible.

Any ideas how to get this plotted correctly and "nicely"?

Answer 1

[Patrick Kalita]

You're on the right track that you need to convert your scattered data into gridded data. A good tool for that job is the TriScatteredInterp class.

% Some data
lat = rand(300, 1);
lon = rand(300, 1);
ptrend = peaks(lat, lon);

% Make a TriScatteredInterp object
F = TriScatteredInterp(lat,lon,ptrend);

% Make a grid of lats and lons, based on the min and max of the original vectors
[lat_grid, lon_grid] = meshgrid(linspace(min(lat), max(lat), 50), linspace(min(lon), max(lon), 50));

% Do the interpolation
ptrend_grid = F(lat_grid, lon_grid);

% Now PCOLOR will accept the gridded data
pcolor(lat_grid, lon_grid, ptrend_grid);