Trouble loading data from .nc file

Hi,

This is my first time trying to pull data from an netCDF file using MATLAB. I'm having difficulty pulling data from this file below:

url = 'http://www.esrl.noaa.gov/psd/thredds/dodsC/Datasets/ncep.reanalysis/pressure/hgt.1968.nc';

This is a file that basically should have the "thickness" (height) of the atmosphere at certain: levels (different from height), latitude, longitude and time. When I use netcdf.getVar(ncid, var) I can get data about each of those 4 variables but not the thickness data itself... does that make sense?

Anyways, any help is appreciated and sorry in advance if this seems vague b/c I don't know where to start.

Thanks!

Answer 1

[Kelly Kearney]

What version of Matlab are you running? They've introduced some nice higher-level reading and writing functions in more recent versions (R2010 and above, I think). You should be able to get the data quickly via ncread. Based on the file details:

ncdisp(url, '/', 'min')

 Source:
           http://www.esrl.noaa.gov/psd/thredds/dodsC/Datasets/ncep.reanalysis/pressure/hgt.1968.nc
Format:
           64bit
Variables:
    level
           Size:       17x1
           Dimensions: level
           Datatype:   single
    lat  
           Size:       73x1
           Dimensions: lat
           Datatype:   single
    lon  
           Size:       144x1
           Dimensions: lon
           Datatype:   single
    time 
           Size:       1464x1
           Dimensions: time
           Datatype:   double
    hgt  
           Size:       144x73x17x1464
           Dimensions: lon,lat,level,time
           Datatype:   int16

you might not want to read in that whole variable all at once (b/c it has a scale factor and offset, it will be read in as a double array). To get one slice at time 1 and level 1:

x = ncread(url, 'hgt', [1 1 1 1], [Inf Inf 1 1]);

[Ryan]

Thanks Kelly this helps a lot! I'm running an up-to-date MATLAB so it worked well. A couple questions though just so I understand:

1) What does [1 1 1 1], [Inf Inf 1 1] mean?

2) The variable (hgt) that I'm reading in is by definition a positive number however there is both positive and negative values. I'll look into if this is how that data is packaged or if these are anomalies but if there is a coding reason behind this let me know.

Thanks again!

[Kelly Kearney]

The ncread function allows you to specify the start, count, and stride values to read certain subsets along each dimension. hgt is a 4-dimensional lat x lon x level x time variable, so I requested that the read start on the first index of each of these 4 dimensions; for the first two dimensions, Inf indicates to read until the end of that dimension, and for the latter 2 I only chose to read a single value (resulting in a 144 x 73 x 1 x 1 array).

Regarding the values, according to the attributes, it looks like those values are as intended:

>> ncdisp(url, 'hgt')
Source:
         http://www.esrl.noaa.gov/psd/thredds/dodsC/Datasets/ncep.reanalysis/pressure/hgt.1968.nc
Format:
         64bit
Dimensions:
         lon   = 144
         lat   = 73
         level = 17
         time  = 1464  (UNLIMITED)
Variables:
  hgt
         Size:       144x73x17x1464
         Dimensions: lon,lat,level,time
         Datatype:   int16
         Attributes:
                     long_name               = '4xDaily Geopotential height'
                     actual_range            = [-5.55e+02  3.24e+04]
                     unpacked_valid_range    = [-7.00e+02  3.50e+04]
                     units                   = 'm'
                     add_offset              = 3.21e+04
                     scale_factor            = 1
                     missing_value           = 3.28e+04
                     precision               = 0
                     least_significant_digit = 0
                     GRIB_id                 = 7
                     GRIB_name               = 'HGT'
                     var_desc                = 'Geopotential height'
                     dataset                 = 'NMC Reanalysis'
                     level_desc              = 'Multiple levels'
                     statistic               = 'Individual Obs'
                     parent_stat             = 'Other'
                     valid_range             = [-3.28e+04  2.93e+03]

Perhaps these are geopotential height anomalies instead of absolute values?