I want to thank NASA IMPACT for supporting this work and the NASA's PO.DAAC team (especially Ed Armstrong and Dean Henze) for collaborating with me on the MUR SST dataset!

3) This demonstrates the transformative potential of the virtual approach enabled by icechunk and virtualizarr. While adopting new technology presents challenges, innovation is necessary to make large Earth datasets accessible and to democratize access across a broader range of users and services.

I would like to thank NASA IMPACT for supporting this work and the team at NASA's PO.DAAC for supporting the development of the MUR SST dataset!

3) This demonstrates the transformative potential of the virtual approach enabled by icechunk and virtualizarr. While adopting new technology presents challenges, these innovations are necessary to make large datasets accessible and democratize access across a broader range of users and services.

2) With icechunk.io + virtualizarr, we built a cloud-optimized ARCO interface to NASA’s MUR SST dataset—20+ years of daily sea surface temps, accessible in seconds for less than a cup of coffee. All from original NetCDFs. 🌊🚀
🔗 lnkd.in/gHhXVcda

1) Earth data producers face the monumental task of managing vast satellite and sensor collections. NASA, NOAA, and others create tremendous public value by releasing data to public clouds—but cloud access alone doesn't solve the core limitations of traditional file-based methods.

@rabernat.bsky.social (Earthmover), Tom Nicholas ([C]Worthy), Moriah Cesaretti, Camille Teicheira (both Sofar Ocean), Hannah Kerner (NASA Harvest), Ivor Bosloper (Taylor Geospatial Institute), Denis Willett (NC Institute for Climate Studies), @kylebarron.dev and Sean Harkins (Development Seed)

And if we should, what are the challenges to doing so?