How to Subset Data from Multiple NetCDF Files based on the desired Lat Long

Hello, I am working on the COSMIC-2 RO dataset that contains nearly 5k profiles for each day for 6 satellites. My goal is to find specific datasets that contains the nearest values to my desired coordinates i.e. the Tonga volcano. Please see the attached link to the dataset if you want to reproduce the problem at your end.

link to data = " https://data.cosmic.ucar.edu/gnss-ro/cosmic2/nrt/level2/2022/015/atmPrf_nrt_2022_015.tar.gz " beware it is nearly 2 gb.

my desired lat long values are: (20.5532 deg south, 175.3841 deg west)

the tar.gz archive has multiple files and I want to extract only files with nearest coordinates if not exact.

It would be best if someone just mention a method. Of course, I am not asking for a complete code but I have spent a lot of time and couldn't figured it out.

[Walter Roberson]

Note: the files are described at https://cdaac-www.cosmic.ucar.edu/cdaac/cgi_bin/fileFormats.cgi?type=atmPrf

[Imtiaz nabi]

Dear Walter, its not about the file type or the data structure in the file, I am looking for a way to actually find data near to my location of interest

[Walter Roberson]

In most cases of nc files, there is a 3d array in which one of the dimensions is latitude, one is longitude, and the third dimension is time. Sometimes though the dimensions are UTM — physical coordinates such as every 100 m, in which case you have to convert to latitude and longitude and deal with the fact that you will have an irregular grid in lat/long coordinates.

When we examine the description of the atmPrf files, we do not see anything like that. I can barely find any 2d arrays there, and it is not clear how those arrays are indexed. It kind of looks like you might need to convert the information about the tangent axes to figure out whether any particular coordinates are in view, but I can't tell what you would do then.

The documentation of the file structure is important for understanding how data might be extracted from it.

I think I would recommend that volunteers sit out on this one unless they just happen to have experience with these kinds of files.