Skip to content

Extract Lightning Flash Time series data

4 messages · Michael Sumner, sownal chand, Edzer Pebesma

Original
sownal chand
# 
Hello Sir/Madam,

I have downloaded a TRMM_LIS.nc File with lightning time series data and
was trying to extract it into a CSV file but I am not sure on how to do it.

I am also attaching the TRMM_LIS file via this link:
https://drive.google.com/drive/u/0/folders/19nJNwPKlJaPlkHn_NGzmjMHMaoUevH04

Request if someone can guide me and thanking you in advance
sownalc
Michael Sumner
# 
I don't understand this file structure yet, but you might try tidync:

library(tidync)
## open the file and get a summary of all vars and the dimensions they use
(nc <- tidync::tidync("TRMM_LIS.nc"))

See how viewtime_lat, viewtime_lon (and others) exist on the grid
identified by "D0" (other grids include multiple dims and so have
identifiers like "D1,D2" - this concept doesn't exist in NetCDF itself but
tidync wanted to group variables by the dims they use).

The grid "D0" is not active by default, so you can do

nc |> activate("D0")

(everything is lazy up to this point, no data has been read)

Convert to dataframe with

nc |> activate("D0") |> hyper_tibble()

You can use hyper_array() to get in native array format with some informal
metadata on the output.

HTH, otherwise package stars has another approach and some similar
facilities.  I often use tidync just to understand what's there, but I
might craft a solution with a different tool after I've explored. RNetCDF
will give you direct low level access to this format, for example.

Cheers, Mike




On Sun, Sep 11, 2022 at 1:22 PM sownal chand <sownalchandfms at gmail.com>
wrote:

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      sownal chand
    

    

      
      #
    

  


  

      
Thanks Michael


But i will try out your approach and let see how it goes.

Kind regards
Sownalc

        
On Sun, Sep 11, 2022, 16:51 Michael Sumner <mdsumner at gmail.com> wrote:

        

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

  
  


  


      

    

    
1 day later

    

      
      

        


  

        

      Edzer Pebesma
    

    

      
      #
    

  


  

      
Someone tried to pack a lot of different things in that netcdf file! Is 
there one particular variable that is of main interest to you? AFAICT 
the primary observations seem like (long,lat,time,something_measured)
sets, as if along a trajectory. Finding the variables that depend on 
dimension viewtime_dim give you 2097 records of them.

Package stars is not so well equiped for one-dimensional arrays - sf 
would be more suitable in terms of data structure but GDAL doesn't seem 
to understand the file to contain vector (POINT) geometries.

I think I would go for using packages like RNetCDF or ncdf4 and 
extracting the one-dimensional variables of interest, then cbind-ing them.

        
On 11/09/2022 07:16, sownal chand wrote: