error reading online netcdf data

Dear All,

I am trying to read this file with Matlab, using the following as advised by Columbia website :

loaddap('http://iridl.ldeo.columbia.edu/SOURCES/.UCSB/.CHIRPS/.v2p0/.daily-improved/.global/.0p05/.prcp/Y/%281N%29%2817N%29RANGEEDGES/X/%2830E%29%2850E%29RANGEEDGES/dods')

loaddap is not working so I used this and it works :

filename =

    'http://iridl.ldeo.columbia.edu/SOURCES/.UCSB/.CHIRPS/.v2p0/.daily-improved/.global/.0p05/.prcp/Y/%281N%29%2817N%29RANGEEDGES/X/%2830E%29%2850E%29RANGEEDGES/dods'

and these are the variables in the netcdf file (latitude, longitude, time and precipitation) :

    ncdisp(filename)

Source:

http://iridl.ldeo.columbia.edu/SOURCES/.UCSB/.CHIRPS/.v2p0/.daily-improved/.global/.0p05/.prcp/Y/%281N%29%2817N%29RANGEEDGES/X/%2830E%29%2850E%29RANGEEDGES/dods

Format:

classic

Global Attributes:

Conventions = 'IRIDL'

Dimensions:

T = 14244

X = 400

Y = 320

Variables:

T

Size: 14244x1

Dimensions: T

Datatype: single

Attributes:

standard_name = 'time'

pointwidth = 1

calendar = 'standard'

expires = 1580515200

gridtype = 0

units = 'julian_day'

Y

Size: 320x1

Dimensions: Y

Datatype: single

Attributes:

standard_name = 'latitude'

pointwidth = 0.05

gridtype = 0

units = 'degree_north'

X

Size: 400x1

Dimensions: X

Datatype: single

Attributes:

standard_name = 'longitude'

pointwidth = 0.05

gridtype = 0

units = 'degree_east'

prcp

Size: 400x320x14244

Dimensions: X,Y,T

Datatype: single

Attributes:

pointwidth = 1

missing_value = -9999

standard_name = 'lwe_precipitation_rate'

units = 'mm/day'

long_name = 'precipitation'

expires = 1580515200

CE = 100

maxncolor = 254

scale_min = 0

CS = 0

colormap = '13882323

16777215

16777215

16777184

6

15453831

6

16748574

14

13434880

18

4026644

24

7451452

21

9419919

43

9234160

44

6333684

41

2237106

43

2237106'

colorscalename = 'prcp_dailyrate_max100_smooth'

scale_max = 100

ncolor = 254

Now if I read the variables latitude, longitude and time (X,Y and T), it works perfectly, but then when I arrive to read the precipitation variables I get the following error :

longitude=ncread (filename, 'X');
latitude=ncread (filename, 'Y');
time=ncread (filename, 'T');
prec=ncread (filename, 'prcp');

66 vardata = ncObj.read(varName, varargin{:});

Error using netcdflib

The NetCDF library encountered an error during execution of 'getVarFloat' function - 'Access failure (-77)'.

Error in netcdf.getVar (line 140)

data = netcdflib(funcstr,ncid,varid,varargin{:});

Error in internal.matlab.imagesci.nc/read (line 605)

data = netcdf.getVar(gid, varid);

Error in ncread (line 66)

vardata = ncObj.read(varName, varargin{:});

Can you please help me ?

Answer 1

[Andrew Whalen]

Hello,

I ran into a similar issue while trying to download another dataset from the same host webpage. The issue is that the command to read the data attempts to serve the entire dataset as one file which is too big, and I believe that explains the error you receive. Try instead to call the ncread function several times in a loop, and you may even have to save some of these data subsets to disk and remove from the workspace lest you run out of ram. Here is a short script to read in the data in smaller chunks and save:

filename = 'http://iridl.ldeo.columbia.edu/SOURCES/.UCSB/.CHIRPS/.v2p0/.daily-improved/.global/.0p05/.prcp/Y/%281N%29%2817N%29RANGEEDGES/X/%2830E%29%2850E%29RANGEEDGES/dods';
ncdisp(filename)

longitude=ncread(filename,'X');
latitude=ncread(filename,'Y');
time=ncread(filename,'T');

% Load data in 500 day increments:
incr=500; indVec=1:incr:length(time);
for i=1:length(indVec)
    if i==length(indVec)
        precip=ncread(filename,'prcp',[1 1 indVec(i)],[Inf Inf Inf]);
    else
        precip=ncread(filename,'prcp',[1 1 indVec(i)],[Inf Inf incr]);
    end
    save(['Prcp',num2str(i)],'precip'); clear precip;
end

Hope this helps!

-A

[le fou]

Hi Andrew,

I emailed Columbia university, and they told me exactly the same thing like one week ago, therefore I am accepting your answer, thanks for your help.