Add further BCO isotope product

[observingClouds]

• fix BCO PICARRO unsupported time dimension upstream
• change datatype of flag variable in BCO precipitation isotope product or change access protocol

[observingClouds]

@leonie-villiger I encountered an issue with the PICARRO dataset. The time dimension has an unsupported unit leading to an error in the CI but also when opening the dataset with common tools like panoply as the time dimension cannot be interpreted. Could you change that and release a new version of the dataset?

netcdf EUREC4A_BCO_PICARRO_WVISOS_L1.nc {
  dimensions:
    time = 51838;
  variables:
    double time(time=51838);
      :units = "seconds since 19700101_00";

seconds since 1970-01-01 00:00:00 would work for example if the time is given in UTC. Please have a look at the time dimension information given in CF-Conventions which are widely supported.

[observingClouds]

@leonie-villiger I also encountered an issue with the rainwater isotope product. The variable flag has the type long (int64). This is generally fine but when accessing the file via OPeNDAP this causes an issue because the datatype is not supported. AERIS provides this access to your dataset via this protocol (see Download | Data access section at https://doi.org/10.25326/242) and we generally try to access these datasets the same way as it allows us to load the datasets lazily, i.e. do not need to download the entire dataset.

Looking at the dataset it seems the flag has only the values [0,1,2] so you could just use short (int16) instead. @d70-t has put together a nice overview about problematic data types at https://howto.eurec4a.eu/netcdf_datatypes.html

Could you look into this as well a release a version of the dataset with a different datatype? In this particular case, we might also use a different access protocol to circumvent this issue because the dataset is in total only 21KB 🤣 and downloading the entire dataset does not hurt.

[leonie-villiger]

@observingClouds: working on both issues. I will let you know as soon as we succeeded releasing new versions of the datasets

[leonie-villiger]

Suggested change

	urlpath: https://observations.ipsl.fr/thredds/dodsC/EUREC4A/BARBADOS/BCO/PICARRO/EUREC4A_BCO_PICARRO_WVISOS_L1.nc
	urlpath: https://observations.ipsl.fr/thredds/dodsC/EUREC4A/BARBADOS/BCO/vapour/EUREC4A_BCO_PICARRO_WVISOS_L3.nc

That is the official version (i.e., the one linked from https://doi.org/10.25326/245). But it still has the issue with the non-standard time unit. Working on it...

[observingClouds]

Hi @leonie-villiger, just wanted to check if you had a chance to look into this again 😬

[leonie-villiger]

• A new version of the BCO precipitation isotope product with corrected datatype of flag variable has been uploaded and is now accessible via OPeNDAP.
• Franziska will take care of the unsupported time dimension in the BCO PICARRO dataset.