Blogging BIOSPACE25!

28th February, 2025

Hello world, this is my first Jekyll blog post.

With that formality out the way… a couple weeks ago I headed off to the Biospace conference at the ESA-ESRIN Observation Center in Frascati, Italy. While I was only there for 2 days, there was a lot to be excited about.

Just past the entrance to the ESA-ESRIN Observation Center, and also past some fairly intense security!

My big takeaway from the opening speeches was that this is the first year that the ESA is spending more on building out its data science capabilities than it is on putting satellites into space. To me, this is indicative of the fact that the marginal benefit from putting effort into effectively wrangling huge amounts of data is now greater than that from collecting huge amounts of data at a faster pace.

There was a lot of discussion around the Kunming-Montreal Global Biodiversity Framework, which introduces a new set of Essential Biodiversity Variables (EBVs). A lot of the discussion here was frankly making the case for much of my PhD research for me. Some quotes of note:

“Why create even more indicators when we can’t even measure the ones we already have?”

“We want to ensure that the mistakes that were made with the SDG indicators are not made again. These mistakes you only begin to learn about as you dive into the data.”

And yet, even with those mistakes, the SDGs are the targets that countries broadly have agreed upon.

Don't mind the satellites hanging from the ceiling

A key point multiple speakers made note of (there were a dozen or so speakers talking for perhaps ~10 minutes each) was that introducing frameworks and methodologies to give countries national ownership of their data and the ability to independently generate compatible statistics was the priority, not introducing new data products. If we can move towards all countries using the same standards, we can enable the aggregation of statistics up in a reliable manner.

That’s not to say that the ESA isn’t introducing new data products: they noted that future missions include Biomass and Flex, which are designed to capture forest biomass and vegetation florescence respectively, both very on point for the theme of the conference.

There was also a palpable frustration around the lack of people who exist in the zones between science, economics, and policy (a point that Simon Sharpe repeatedly makes in his fantastic book, Five Times Faster, which I should be reading at least two times faster given that I’m only halfway through it…). “We can’t drive impact without more of these people popping up over the next few years.”

Jilian Campbell of the Convention on Biological Diversity very astutely noted that even once we do have standarised EO-derived indicators, we still need robust mechanisms of connecting these to on-the-ground data for validation and verification purposes.

Ilaria Dimatteo of the UN Statistical Commission explained how even though they carefully map out both environmental and economic circumstances for policy-making, “when decisions are take at the nation level, the environment does not really come into play.” In 2021, the commission adopted the System of Environmental-Economic Accounting Ecosystem Accounting (SEEA EA, not a typo) to essentially force the connection between these two domains in a spatially explicit manner. “From a statistical perspective, we really want international compatability. Methods to ensure that information generated is reliable, replicable, and widely known.”

The conference also featured a great talk from my fellow Cambridge PhD Andrés Zúñiga González focusing on on-device scalable learning to aid urban tree management with a hardware focus.

I particularly enjoyed his hedgehog networking analogies!

At my own poster presentation, it turned out that almost everyone at the conference had an opinion on causal relationships between indicators and were either incredibly glad that someone was trying to tackle it or incredibly frustrated with the concept as a whole having bumped their own heads against it. The most valuable feedback I received involved mentions of additive and generalised statistical models which can better capture characteristics of causality in specific sorts of systems.

My last day at the conference also involved a workshop focusing on the EBVs (I forgive you for forgetting what this stands for - Essential Biodiversity Variables), in which we went through the entire list of EBVs thinking through whether these were observable from satellite based data products and how feasible accurate measurement of them was at scale. It turned out that many people were only just beginning to gain familiarity with these, so it was a good learning exercise for us all.

A wonderful view and delicious food to wrap up my trip. It is Italy after all!

Ciao