Contour plot of spatial distribution of temperature

i have 5 block data i only need 3

Lat      lon     month   year     temp
22.25000 -98.25,    1,  2000, 0.1061891
22.25000, -97.75,    1,  2000, 0.2013340
22.25000, -97.25,    1,  2000, -999.9000
22.25000, -96.75,    1,  2000, -999.9000
22.75000, -100.75,   1,  2000, 0.4650421
22.75000, -100.25,   1,  2000, 0.3308848
22.75000, -99.75,    1,  2000, 0.2076520
22.75000, -99.25,    1,  2000, 0.1700439
22.75000, -98.75,    1,  2000, 0.1031799
22.75000, -98.25,    1,  2000, 0.1318403
22.75000, -97.75,    1,  2000, 0.3397914
22.75000, -97.25,    1,  2000, -999.9000
22.75000, -96.75,    1,  2000, -999.9000
23.25000, -100.75,   1,  2000, 0.4653015
23.25000, -100.25,   1,  2000, 0.3465271

i have develop a contour plot of spatial distribution of temperature

even some values have -999.999 some way to interpolate these values and replace it in the vector of temp.

i already started with this

archivo=load('escenarioA2.txt');
lat=archivo(:,1);
lon=archivo(:,2);
month=archivo(:,3);
year=archivo(:,4);
inc=archivo(:,5);
[m,n]=size(archivo);
nx = length(lon);
ny = length(lat);
[X,Y]=meshgrid(lon,lat);

Answer 1

[Geoff]

I would start by removing the invalid readings:

archivo( archivo(:,5) < -999, : ) = [];

You need to have a higher (and uniform) granularity for your grid... How about this:

latmin = min(lat); latmax = max(lat);
lonmin = min(lon); lonmax = max(lon);
glat = linspace( latmin, latmax, ceil((latmax-latmin)/0.01) );
glon = linspace( lonmin, lonmax, ceil((lonmax-lonmin)/0.01) );
[X,Y] = meshgrid(glon, glat);

Now, you want to interpolate your data into that grid. Look at the following function:

doc interp2

[David Lopez]

and for plot a spatial distribution which command do you recommend me?

Answer 2

[Chad Greene]

The regular quarter-degree spacing of the data suggests that no interpolation is necessary for this dataset--it just needs to be restructured. An easy way to do that is with a function of mine called xyz2grid.

For your matrix M whose columns correspond to [lat, lon, month, year, temp], use longitude as the "x" data and latitude as the "y" data, then grid it up like this:

[Lon,Lat,Temp] = xyz2grid(M(:,2),M(:,1),M(:,5)); 

Then NaN-out the no-data values like this:

Temp(Temp==-999.9) = nan;