Issue Accessing OPeNDAP Server via R and Broken URL on Website

[Chloe Gao]

Dear OPeNDAP Support Team,

I hope you’re doing well. I am an MD/PhD student examining climate change and pediatric critical care admissions, and am writing to report an issue with an OPeNDAP server and a broken link on the associated website.

I have been trying to access data via OPeNDAP in R, but the connection repeatedly fails. 

When I try to open this same URL in a web browser, it also leads to a broken link / error page instead of the expected dataset interface (screenshot below).

Of note, when I accessed the server and website link during this past summer (in July 2025), both were functional. 

Could you please let me know whether this OPeNDAP endpoint is currently down and/or if there is an updated URL or recommended workaround?

Thank you very much for your time and assistance.

Chloe Gao

MD/PhD Candidate at The University of British Columbia
Vanier Scholar
Canadian Queen Elizabeth II Diamond Jubilee Scholar  

LinkedIn: https://www.linkedin.com/in/chloe-gao-5a220521a/ 

Website: https://mdprogram.med.ubc.ca/mdphd/students/

[Miguel Jimenez]

Dear Chloe Gao, 

Thanks for reaching out, and thanks for telling us about your work. Sounds pretty fascinating!

NASA is in the process of migrating their data to the cloud, resulting in some on-premises servers also being migrated to the Cloud. This is likely the root of your ongoing issues, with experiencing broken links. 

Currently, the most reliable way to find OPeNDAP links associated with any of NASA’s data archives is to query these URLs regarding your data of interest directly to the CMR (Common Metadata Repository) or via Eart data Search. The CMR is nice because is can be done programmatically and there are already several tools that enable users to search for NASA's OPeNDAP URLs. For example PyDAP, written in Python, enables users to find updated OPeNDAP URLs from NASA’s CMR. The information you will need at a minimum is the Concept Collection ID or the DOI of your collection of interest.

You can find a Python example in the following documentation: https://pydap.github.io/pydap/en/notebooks/GetCMRURLS.html

Thanks again for reaching out, and do not hesitate to reach out if your problems persist.

Miguel Jimenez-Urias, PhD

[Chloe Gao]

Hi Dr. Jimenez-Urias, 

Thank you so much for all your help and support with this! Earlier this year, I was accessing the Daymet data by using an OPeNDAP connection through a URL with the following format: 

https://thredds.daac.ornl.gov/thredds/dodsC/ornldaac/2131/daymet_v4_*_monavg_na_YYYY.nc

I’m now getting an error that the file is unavailable, and the link labelled as “Daymet THREDDS Data Server” on the page below now gives me the “bad request” page that I previously sent.

When I looked into the CMR, I didn’t see Daymet available in that. Since neither my original approach nor the method you suggested is working, do you have any suggestions for accessing the Daymet data programmatically in R, as I did earlier this year? 

I am able to find Daymet in the Earth data page, but it looks like I have to download very large netCDF files for each variable and year, whereas with R, I could extract only the small spatial subset of data that I needed for my research. 

Thank you kindly, 

Chloe 

Website: https://mdprogram.med.ubc.ca/mdphd/students/<Screenshot 2025-12-05 at 9.08.29 PM.png>

[Miguel Jimenez]

Hi Chloe, 

I meant to write yesterday, my apologies. That URL you listed in your email will no longer work for cloud hosted data on NASA. When data is moved to S3, the OPeNDAP urls get updated to have the following format:

  https://opendap.earthdata.nasa.gov/collections/<CollectionConceptID>/granules/<GranuleID>

Where 

<CollectionConceptID> ——>  Data product you are interested in, for example DAYMET has a CollectionConcept ID  that is unique.

<GranuleID>     ——>  Specific granule or file name

The above listed format is a typical Cloud OPeNDAP url. Since DAYMET data has been migrated to the cloud (along with updated clouds access points), that came with the unfortunate changes to the OPeNDAP url. We are aware that these changes broke changes to many people's workflows, but unfortunately these changes were out of our control.

How does this apply to your data of interest?

So for example, for DAYMET V4 (monthly averages) you have:

<CollectionConceptID> = C2532007210-ORNL_CLOUD

<GranuleID> = Daymet_Monthly_V4R1.daymet_v4_tmax_monavg_na_2024.nc  # an example

These result in the following cloud OPeNDAP URL:

https://opendap.earthdata.nasa.gov/collections/C2532007210-ORNL_CLOUD/granules/Daymet_Monthly_V4R1.daymet_v4_tmax_monavg_na_2024.nc

You can access the DAP2 metadata by appending a dds to the above URL, or the DAP4 metadata by appending .dmr to the url above. And request subsets of data using DAP2 or DAP4 protocols as you have in the past.

And as mentioned in my previous email, you can search for N opendap urls using pydap (directly querying it from the CMR) whether these are Cloud instances or not. Following the example in the following resource:

 https://pydap.github.io/pydap/en/notebooks/GetCMRURLS.html#discoverying-daily-4km-cholophyll-data-from-pace-level-3 

That is,

from pydap.client import get_cmr_urls

urls = get_cmr_urls(ccid=PACE_ccid, limit=1000) # limit by default = 50

But replae PACE_CCID with the collection concept ID from DAYMET’s v4 R1 monthly averages data provided above. 

As with the CMR, you can add filters such as time_range if you are interest in a specific month and year. The documentation shows an example of that scenario.

Hope this helps, and do reach out if you continue to experience issues accessing data.

Cheers, 

Miguel

[Chloe Gao]

Hi Miguel, 

No worries at all! My apologies for my delay in getting back to you as well! 

This was so incredibly helpful, and I sincerely appreciate your detailed guidance and support on this project! 

Thank you kindly, and I wish you a happy holiday season! 

Best, 

Chloe