Introduction to Python – Argo Float Data
This summer, as part of our virtual REU workshop, we introduced students to the basics of using Python to analyze oceanographic data. When we frantically designed the course back in May (during the early even more frantic days of the pandemic), we originally intended to introduce students to data from both NDBC moorings and Argo profilers. In the end, we only used NDBC, as that dataset was rich enough for students to work on a number of interesting research questions, while also learning python. Keeping things simple is a good mantra in both design and education.
But the Argo dataset, would have added another dimension to their learning, literally.
I love the NDBC dataset, because it provides timeseries measurements of a few basic atmosphere/ocean variables at hundreds of locations all over the globe. It’s also relatively easy to access, and somewhat intuitive for students to understand, especially when they investigate data from stations near where they live.
But almost all of the data from NDBC is measured at or near the ocean surface.
What’s missing is depth, again literally.
That is why the Argo dataset is immensely useful. The Argo network now includes almost 4,000 real-time floats that students can use to investigate oceanic profiles from all over the world. They can look at timeseries of those profiles, or they can look at how profiles in a given area have changed over the past 20 years that the network has been active. Not only that, but a new set of Biogeochemical Argo floats is coming online, allowing students to study the subsurface structure of the ocean from several new instruments, with data that might help you enhance topics in chemical and biological oceanography courses.
Ultimately, this summer we decided that the NDBC dataset was more than enough to introduce students to the basics of python and timeseries data analysis, before they moved onto data from the OOI and other observatories for their research projects. The Argo dataset is a bit more complex to interpret, and that’s after you’ve figured out how to access that data you want. Because the network is a global partnership, there are a number of different data centers, each with its own subset of profilers and data format that takes some time to figure out.
But if you’re interested in in creating activities delve into the Argo dataset, hopefully you’ll find this notebook helpful as a first step.
This post is part of our 2020 Summer REU Intro to Python series. See also Part 1, Part 2, and Part 3.
Hi Sage: The part 4 of your phython programming does not work. I think the URL for the Argo’s data site is somehow mis-labelled. Thanks Ajoy
Hi Ajoy, so it looks like NCEI has finally moved their data server for the data that used to be hosted by NODC. For whatever reason, the Argo data files aren’t on the new server, but hopefully they’ll add them back soon.
Alternatively, I recommend checking out the new argopy library, which I discovered after I wrote this tutorial. It makes searching and accessing Argo data in python far easier. You can also check out this recent paper by Maze and Balem in JOSS.
Hope that helps!