How to process NC file in matlab

Question

Muhammad Usman Saleem 2021-12-11

2
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1608670-how-to-process-nc-file-in-matlab

评论： Walter Roberson 2022-1-6

I've monthly soil monisture NC file. This file contains monthly soil monisture data since 1948. I want to extract soil data within my shapefile in text file (Shapefile has also attached). I google alot to proccess NetCDF file in matlab find not solution, The ncdisp of my nc file is given below:

Format:
           netcdf4_classic
Global Attributes:
           _NCProperties = 'version=1|netcdflibversion=4.4.1.1|hdf5libversion=1.10.1'
           Conventions   = 'CF-1.0'
           title         = 'CPC Soil Moisture'
           institution   = 'NOAA/ESRL PSD'
           dataset_title = 'CPC Soil Moisture'
           history       = 'Wed Oct 18 15:13:37 2017: ncks -d time,,-2 soilw.mon.mean.x.nc soilw.mon.mean.xx.nc
                           Wed Oct 18 15:12:08 2017: ncks -d time,,-3 soilw.mon.mean.nc soilw.mon.mean.x.nc
                           CPC Soil Moisture
                            Obtained on Nov 2004 from CPC's website
                            and written to netCDF by Cathy Smith 12/2004.
                            he CPC Global monthly soil moisture dataset is a 1/2 degree resolution grid from 1948 to the present.
                            The file is written in COARDS and CF compliant netCDF at NOAA ESRL/PSD https://www.esrl.noaa.gov/psd/ 
                            Converted to chunked, deflated non-packed NetCDF4 Jul 2014'
           NCO           = '4.6.9'
           References    = 'https://www.psl.noaa.gov/data/gridded/data.cpcsoil.html'
Dimensions:
           lat  = 360
           lon  = 720
           time = 886   (UNLIMITED)
Variables:
    lat  
           Size:       360x1
           Dimensions: lat
           Datatype:   single
           Attributes:
                       long_name          = 'Latitude'
                       units              = 'degrees_north'
                       actual_range       = [8.98e+01 -8.98e+01]
                       standard_name      = 'latitude'
                       axis               = 'Y'
                       coordinate_defines = 'point'
    lon  
           Size:       720x1
           Dimensions: lon
           Datatype:   single
           Attributes:
                       long_name          = 'Longitude'
                       units              = 'degrees_east'
                       actual_range       = [2.50e-01 3.60e+02]
                       standard_name      = 'longitude'
                       axis               = 'X'
                       coordinate_defines = 'point'
    soilw
           Size:       720x360x886
           Dimensions: lon,lat,time
           Datatype:   single
           Attributes:
                       long_name     = 'Model-Calculated Monthly Mean Soil Moisture'
                       missing_value = -9.97e+36
                       units         = 'mm'
                       valid_range   = [0.00e+00 1.00e+03]
                       dataset       = 'CPC Monthly Soil Moisture'
                       var_desc      = 'Soil Moisture'
                       level_desc    = 'Surface'
                       statistic     = 'Monthly Mean'
                       parent_stat   = 'Other'
                       standard_name = 'lwe_thickness_of_soil_moisture_content'
                       cell_methods  = 'time: mean (monthly from values)'
                       actual_range  = [0.00e+00 1.00e+30]
    time 
           Size:       886x1
           Dimensions: time
           Datatype:   double
           Attributes:
                       long_name          = 'Time'
                       units              = 'days since 1800-01-01 00:00:0.0'
                       delta_t            = '0000-01-00 00:00:00'
                       avg_period         = '0000-01-00 00:00:00'
                       standard_name      = 'time'
                       axis               = 'T'
                       bounds             = 'time_bnds'
                       coordinate_defines = 'start'
                       prev_avg_period    = '0000-00-01 00:00:00'
                       actual_range       = [5.41e+04 8.10e+04]
                       
 

I've tied

  
  file='data.nc';
ncdisp(file)
long = ncread(file,'lon');
latt = ncread(file,'lat');
time = ncread(file,'time');

I want output of Soilw in this format in text file

Date        Soilw
01/01/2015   10
01/02/2015    20
01/03/2015    0.5

Please help me to convert monthly netcdf datainto text file using matlab please?

7 个评论
显示 5更早的评论隐藏 5更早的评论

Walter Roberson 2021-12-12

@Muhammad Usman Saleem

I do not seem to find you in my list of customers of my company who have signed 4-Hour Response Time Contracts with us ?

Muhammad Usman Saleem 2021-12-13

@Walter Roberson Hahaha, you can signed with your company 4 Hour Response Time Contract. Actually Time is importantance for me, I'm end user of the NC files and all time of the research spent in proccessing the raw dataset.. I would like to thank you very much for your timely response..

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Walter Roberson 2021-12-12

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1608670-how-to-process-nc-file-in-matlab#answer_853610

在 MATLAB Online 中打开

file = 'data.nc';
shapefile = 'lahore.shp';
long = ncread(file, 'lon');
latt = ncread(file, 'lat');
time = ncread(file, 'time');
soilw = ncread(file, 'soilw');
S = shaperead(shapefile);
%ncread sometimes returns the transpose of what we expect
if size(soilw, 1) ~= length(lat)
  soilw = permute(soilw, [2 1 3]);
end
Ntime = length(time);
Date = datetime('1800-01-01 00:00:00') + days(time(:));
mask = inpolygon(lat, lon, S.Y, S.X);  %caution, lat is Y not X !
Soilw = zeros(Ntime, 1);
for K = 1 : Ntime
  sl = soilw(:,:,K);
  Soilw(K) = mean( sl(mask) );
end
output = table(Date, Soilw);

17 个评论
显示 15更早的评论隐藏 15更早的评论

Walter Roberson 2021-12-20

在 MATLAB Online 中打开

Shape files define shapes in terms of coordinates. There is no inherent coordinate system -- not inherently degrees or kilometers or pixels, just relative units.

Each coordinate other than the first and last is an endpoint of a line from the previous coordinates to the current coordiantes, and in turn is the start point of a line from the current coordinates to whatever is next in the file.

It would be perfectly valid, for example, for a shape to have coordinates such as

which would describe a closed right triangle with sides of length 4 (horizontal) and 3 (vertical).

This shape would be intended to enclose an area -- it would include coordinates [3 1] for example.

Now you want to apply that shape to a data file that uses the same units. But the data file does not necessarily have a data resolution that is 1:1 with the coordinates of the shape. The data file does not necessarily have data stored [3 1] or [3 0] -- it might, for example, only have data stored at even coordinates such as [0 0], [2 0], [4 0]: it might be lower resolution than the data coordinates. Or the data file could be higher resolution: it might have data at (for example) every 1/2 coordinate unit.

Because of these factors, the space implied by the coordinates in the shape file might include data coordinates that are not right "at" the coordinates given in the shape. There could even happen to be a data at a coordinate that was completely interior to the shape -- for example the data file might happen to have data at coordinates [0 1] and [3 1] -- the [0 1] is outside of the triangle shape, and the [3 1] is inside the triangle shape but neither 3 nor 1 are "mentioned" in the shape. As coodinates, [3 1] is enclosed by the shape.

The code I posted does not try to map out the shape to find data-file coordinates that might be internal but not "mentioned" by the shape.

The reason that I did not both to try to map out internal coordinates is that the shape spans quite a small portion of the data file. A really careful mask might even find that the exact location of one of the data points is "just outside" the shape, if you think of the data coordinates as being point-like objects. But at least for this file, the data coordinates are not point-like objects: they represent amount of precipitation over the entire grid square, so knowing that the shape coordinates discretized to the same grid location is good enough for this purpose. Likewise, none of the edges of this shape file are very long at all.

With higher resolution data sets or lower resolution shapes, the differences could turn out to be important. But as it is, the shape is just a very small smudge against the data file.

Muhammad Usman Saleem 2022-1-6

@Walter Roberson dear sir, whether the above code elimainating missing values (-9.97e+36) from the mean of dataset?

Walter Roberson 2022-1-6

在 MATLAB Online 中打开

@Muhammad Usman Saleem

file = 'soilw.mon.mean.v2.nc';
shapefile = 'Shapefile/lahore.shp';
long = ncread(file, 'lon');
latt = ncread(file, 'lat');
time = ncread(file, 'time');
soilw = ncread(file, 'soilw');
info = ncinfo(file, 'soilw');
soil_fill = info.FillValue;
[found, idx] = ismember('missing_value', {info.Attributes.Name});
if found
  soil_fill(end+1) = info.Attributes(idx).Value;
end
mask = ismember(soilw, soil_fill);
soilw(mask) = nan;

Then near the bottom

for K = 1 : Ntime
  sl = soilw(:,:,K);
  Soilw(K) = mean( sl(ind), 'omitnan' );
end

请先登录，再进行评论。

Answer 2

KSSV 2021-12-12

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1608670-how-to-process-nc-file-in-matlab#answer_853350

Refer this: https://in.mathworks.com/matlabcentral/answers/574741-create-a-box-around-points-and-use-in-inpolygon

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Muhammad Usman Saleem 2021-12-12

Many thanks for your reply @KSSV. I've read your number of answers on NC file proccessing in Matlab but I not able to form code according to my requirement. I spent one week fully to solve my this problem under your replies on different questions but not able to solve that why I post here a question. I not expert in coding will you please write code dear? Thanks for this

请先登录，再进行评论。

How to process NC file in matlab

7 个评论
显示 5更早的评论隐藏 5更早的评论

采纳的回答

17 个评论
显示 15更早的评论隐藏 15更早的评论

更多回答（1 个）

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

另请参阅

类别

标签

产品

版本

Community Treasure Hunt

How to process NC file in matlab

7 个评论 显示 5更早的评论隐藏 5更早的评论

采纳的回答

17 个评论 显示 15更早的评论隐藏 15更早的评论

更多回答（1 个）

1 个评论 显示 -1更早的评论隐藏 -1更早的评论

另请参阅

类别

标签

产品

版本

Community Treasure Hunt

7 个评论
显示 5更早的评论隐藏 5更早的评论

17 个评论
显示 15更早的评论隐藏 15更早的评论

1 个评论
显示 -1更早的评论隐藏 -1更早的评论