Can a Global Earth Observation Accord Work, Even If It Is Not Global?

Housekeeping: I'm sending this essay today mainly as a reaction to the EO-related declarations from the International Space Summit in Paris. Your regular TerraWatch Space Pro & Essentials newsletter with analysis and insights will land in your inbox on Wednesday, September 16.

This is another follow-up to my December essay, Towards a Global Earth Observation Accord, and the more detailed piece I wrote in June, Why Earth Observation Needs Its Own Artemis Accords.

I did not expect to come back to this topic and write another essay so quickly. But this week, at the International Space Summit in Paris on September 9-10 (which I attended), governments signed two declarations that are directly relevant to the argument I have been making around open Earth observation (EO), global EO governance and the need for political commitments around the data the world increasingly depends on.

So, here's me breaking those two declarations down and sharing some thoughts on what it means for EO and whether it moves us towards a political framework for EO commitment globally. As always, hit reply if you have any comments.

Context

Let's start with some quick context – the two declarations signed in Paris.

The first declaration is on space-based scientific data, covering observations of both the Earth and the wider Universe. It sets out four broad principles: accessibility, openness, integrity and international cooperation. It supports open standards, open licences where appropriate, traceability and reciprocal data sharing. This was signed by 49 countries and the European Space Agency.

The second declaration is specifically on the Copernicus programme. It calls the European programme a "global public good whose continuity is essential", explicitly linking the value to its free, full and open data policy, and calls for the future missions and services of the programme to be sustained. This was signed by 27 countries and five other organisations (ESA, ECMWF, EUMETSAT, Mercator Ocean International and the Brazilian Space Agency).

In a conference, which almost entirely focused on exploration, spaceflight and European sovereignty and autonomy, these two declarations barely made the news – I mean, why would they? Most people don't think about EO until it goes missing, take the benefits it provides for granted and assume that the data will keep flowing forever, even if there is no global political framework to protect it.

What These Declarations Mean

First, let's be clear: neither declaration is proposing to create a new institution, nor are the principles providing any commitments around specific datasets. So, to give you a spoiler, they do not solve the EO governance problem I have articulated in both of my earlier essays – "the world relies on open EO data, but there is no shared political framework defining whether it will remain open, which countries are responsible for what, what is the global EO baseline dataset and how that baseline is sustained when national priorities change."

Nevertheless, I think, these declarations are still quite significant. Why?

The current arrangements for open data are still largely national or regional – The US decides what happens to Landsat (and we already know things are changing there) and Europe (the EU to be precise), decides what happens to Copernicus (and we already know it had a small hiccup after Brexit).

And, in general, individual governments decide which national datasets they make available and under what conditions. Some international cooperation and bilateral partnerships connects a few those national programmes, but the decisions about keeping them running remain with their funders, the politicians.

So, what I think these two declarations do, is really bring a political commitment to the principles of openness and continuity on the table as matters of shared international responsibility. These are not new ideas—the Group on Earth Observations (GEO) already promotes open data and international cooperation—but having governments publicly reaffirm them at this summit attended by heads of state gives a renewed political weight.

The second declaration calling Copernicus a global public good and asking specific governments to keep it running, is a pretty big deal. Whether or not anything follows from it, that is a different kind of statement to make.

As I have articulated before, what has been missing is a shared political framework for maintaining a global EO baseline, and these declarations are a step towards that framework, not the framework itself.

Open EO Data Is Still A Political Choice

One of the main arguments in my June essay was that open EO data is a political choice, not a natural law. That sounds obvious, but it is easy to forget because Landsat and Copernicus have existed for so long.

Someone in Kenya can download decades of Landsat data without paying the US government. The same person can access Sentinel data paid for overwhelmingly by European governments. Researchers can build long-term records on top of it. Governments without major space programmes can monitor their territory and their environment using infrastructure funded elsewhere. Companies can now train EO foundation models, without having to buy or license petabytes of satellite imagery.

We have become used to this. But, in the times we are living in, I am inclined to point out again that there is nothing inherent keeping EO data open for all. It is important to remember that governments choose to make Landsat and Copernicus open, they chose to keep funding the satellites, ground segments, archives, processing infrastructure and dissemination systems behind them. And those choices can change (and I expect them to change in some form or the other).

I discussed the Switzerland example I discussed in my June essay. A country could decide against being a member of the Copernicus programme but retain access to most of the data. The more countries can enjoy the public good without contributing to the programme, the more the long-term funding model becomes critical.

The Paris declarations make this political nature of open data unusually explicit. The scientific-data declaration says data should be accessible to all, as open as possible, interoperable and protected for integrity. More interestingly, it describes the production, acquisition and sharing of this data as a collective responsibility and explicitly refers to reciprocity.

The Copernicus declaration goes even further in another direction. It says, quite plainly, that Copernicus is a global public good. I think that phrase matters – but once you describe Copernicus as a global public good, the next question becomes uncomfortable quickly:

Who is responsible for maintaining that public good?

EO Is A Global Public Good, But Who Is Responsible?

There is one line at the end of the Copernicus declaration that I found particularly fascinating.

After recognising Copernicus as a global public good, supporting its continuity, calling for future missions to be deployed and encouraging wider international data exchange, the declaration states that it creates "no additional financial obligations for the signatory states."

Understandable? Yes, a summit declaration is not call for a blank cheque. But, it is also the core of the structural problem I wrote about in June.

Everyone wants Copernicus to continue, everyone benefits from the data being open, countries appreciate it, want greater access, more exchange and more services built around it. But ... somehow Europe remains responsible for keeping the programme running.

That does not mean the other signatories contribute nothing to EO overall – several countries have launched important EO missions and some have bilateral partnerships with ESA or the EU for data sharing. It means that calling Copernicus a global public good has not, through this declaration, created a new way of sharing responsibility for sustaining it.

This is not just about Copernicus. The US has funded Landsat for decades while users around the world benefit from its archive. Weather observations are exchanged internationally because forecasts in one country depend on observations from many others. And, in the June essay, I discussed how at least sixty countries are launching or announcing plans for national EO systems. So, the question is really about how those investments contribute to something beyond each country’s own programme.

The benefits of open EO data keep growing, but the political and financial responsibility remains concentrated in the same couple of programmes – even as dozens of countries build their own national EO capabilities without any of it flowing into a shared baseline.

So perhaps the next phase of the open data discussion is not simply whether open EO is valuable – we already have decades of evidence that it is. The harder question is who contributes to the global EO baseline, and what that contribution actually means.

And, no, I don't necessarily mean sending money to Brussels to continue funding Copernicus (although that might be an option).

One country might contribute an open national dataset, another might commit to maintaining a long-term measurement, another might provide calibration infrastructure, another could make specific high-resolution observations available during disasters, and others could commit to common standards, provenance requirements or processing infrastructure.

This was always one of the ideas behind the EO Accord I have been proposing: not that every country pays equally, but that countries with growing EO capabilities start treating at least some of those capabilities as contributions to a shared system rather than purely national assets – for the global good.

As I wrote in June:

Sixty countries launching more EO satellites could mean 60 contributions to a global Earth observation system or just 60 isolated national EO programs that do not add up.

The Paris declarations do not give clarity on which direction we are actually heading towards.

How Close Is This to a Global EO Accord?

I want to be careful here because it would be very easy for me to point at the two declarations and say: look, this is exactly what I proposed – it is not!

My argument since December has been that the gap in EO governance is not another technical coordination mechanism. The Committee on Earth Observation Satellites (CEOS) already coordinates missions and standards, GEO already works on open data, interoperability and the broader Global Earth Observation System, and the World Meteorological Organization provides the strongest existing example of international data sharing architecture in weather.

The missing commitment, I think, is more specific: governments saying which observations they will contribute to a global baseline, on what terms, and how they will work with others to sustain it.

The scientific data declaration provides many of the principles those commitments would need – accessibility, openness, interoperability, integrity & provenance, reciprocity and international cooperation. The Copernicus declaration adds something equally important: continuity.

Those are very close to the principles I proposed in June for a global EO baseline. But, the declarations stop before getting to the difficult part.

They do not say which datasets a country is actually committing; they do not define a minimum global EO baseline; they do not say what happens when an EO mission is cancelled, a satellite fails or a government changes its data policy; they do not define what reciprocity means in practice; and, they do not create a process for another country or programme to pick it up when continuity is threatened.

The scientific data declaration contains the exceptions for national security, public safety and defence. These are reasonable, but also leave the hardest EO question unresolved: what happens when the same dataset has both civilian and military value?

This is not a hypothetical problem. In July, the EU delayed the release of Sentinel-1 and Sentinel-2 products over the Sea of Oman at the request of the US. Two months later, the Copernicus declaration now reaffirms free, full and open access. The principle and the exception now sit inside the same programme, and nothing tells a user when it can be restricted or how long such a measure lasts.

This is why, in June, I moved away from trying to define a clean category called "civilian EO". I am now convinced that it does not really exist at the sensor level. The same image that supports agriculture, insurance, journalism, humanitarian response also supports intelligence and military analysis – as we have seen from the Washington Post use of Copernicus data for analysis about the war in Iran, especially as commercial imagery became restricted.

A more realistic approach, I assumed, was to ask governments to designate specific datasets that they are willing to put inside the global EO baseline, and then make some commitments around those datasets. Obviously, that is harder than just saying EO data should be "as open as possible".

In other words, I was expecting something like: these EO datasets are part of the global baseline, this is how it can be accessed by anyone, this is how long we intend to maintain it, and these are the conditions under which that commitment could change.

So, Paris gives us many of the principles, but the more important step is really about moving from principles to contributions.

Who Did Not Sign and Why That Matters

This is where the declarations get even more interesting.

The United States, China, India, Russia and Japan – arguably five of the world's major space/EO powers – do not appear on either signature list.

Canada and the Brazilian Space Agency appear on the Copernicus declaration only, while South Korea appears on the scientific data declaration only.

I don't think this is a minor thing at all. This means that these countries, whose capabilities could make substantial contributions to a global EO baseline, have not endorsed either declaration. And even among those who have signed, the two declarations do not have the same membership.

I want to avoid over-interpreting what this tells us about the choices of any individual country. But, not appearing on a signature list is certainly not the same as opposing the principles of open EO data or others that were signed – Landsat alone makes that interpretation too simplistic in the US case.

So, my broader point stands: Agreeing that open EO is useful is not the same as agreeing on the commitments and the political framework that should govern it.

And, that for me, is likely to be the biggest constraint on any global EO Accord we will work on in the coming years.

Are We Going To End Up With An EO Coalition Instead?

In my last essay, I made a useful but frankly, incorrect comparison to the Artemis Accords – the set of non-binding international principles to guide civil exploration and use of the Moon, Mars, comets, and asteroids.

The Artemis Accords now have 71 signatories, soon becoming 72, with Djibouti scheduled to sign on 14 September. Obviously, China is not one of them, nor is Russia.

The important thing is that Artemis did not wait for universal agreement among every major space power. The US, along with an initial group of seven other countries agreed on a set of principles, started signing countries up and built political momentum around them.

Whatever one thinks about the Accords as a governance model, they have clearly succeeded in attracting countries: the coalition has grown from eight founding signatories in 2020 to 71 today (and soon 72) – which is formidable.

The optimist in me thinks that we may be seeing the beginnings of a political coalition around EO. The two declarations have different countries signed up, but, they do bring groups of governments together around principles that reach beyond their own programmes – on a national level, not just their space agencies.

This could evolve in a similar direction: a group of countries agree principles around open EO data, integrity, interoperability, continuity and reciprocity, then gradually make those principles more operational.

In one sense, that is probably more realistic than waiting for the US, China, Europe, India, Russia and everyone else to agree on one framework from day one. But this is where an EO Accord is different from the Artemis Accords.

An EO Accord is supposed to sustain a global observational EO baseline, irrespective of political alignment – similar to how weather data is shared today between Russia, China, Europe, US and other nations. As we all know, climate, weather and disasters do not divide neatly into geopolitical blocs.

In addition, the value of EO for disaster response, agricultural monitoring, ocean observation and many other applications increases when observations from different systems and geographies can be combined. China, US, India, Europe, Japan and others are not just political actors I would ideally like in a global EO framework because I am naive and utopian. I also know they operate a large share of the EO satellites that could make the global EO baseline stronger.

And this is not entirely hypothetical. India, China and Russia already cooperate with Brazil and South Africa through the BRICS Remote Sensing Satellite Constellation, an arrangement for sharing data from existing satellites. In June, their space agencies discussed expanding participation to newer BRICS members and in the BRICS Summit, held, last week, reaffirmed their interest in growing it.

So, some of the countries absent from the Paris declarations are not waiting around for an invitation to cooperate on EO – they already have an arrangement of their own. That is not the same as committing those observations to a globally open baseline. But it makes me wonder: can these different arrangements contribute to the same global EO baseline, or will cooperation deepen within each group, leaving us with separate EO blocs rather than a shared global baseline?

This brings me to a difficult but pragmatic question: What if a global consensus on EO governance is just simply not possible?

I can see three broad ways this could go:

1. The Global Consensus

This means to continue (somewhat naively) working towards a framework that includes the US, China, Europe, India and the other major EO powers. This would give us the opportunity to have a global baseline that includes most of the world's EO capacity and thereby, distribute responsibility more widely, and benefit more.

But requiring universal adoption from the start could mean waiting indefinitely or ending up with commitments so weak that they do not change very much. A declaration everyone can sign is not necessarily a baseline anyone can depend on.

2. A Coalition of the Willing

This means a group of countries could agree commitments around specific datasets, continuity, interoperability, provenance and crisis access, and then try to expand that coalition over time. This is closer to how the Artemis Accords evolved and is probably much easier politically.

The risk is that EO governance starts to mirror geopolitical blocs – which may not be compatible with the fact that EO is a global public good, and as we just discussed, a global EO baseline needs to have as much data as possible in it.

3. A Layered Architecture

The third option separates two things the first two treat as one: membership and contribution. A country can supply observations to the global EO baseline without signing up to all the principles that govern it.

Members would designate specific datasets, commit to keeping them open, and agree in advance what happens to those datasets when a mission ends or a policy changes. Countries outside the coalition could still feed observations into the same baseline, through bilateral or programme-level data-sharing agreements, without endorsing the wider framework.

Members would jointly decide what the baseline should provide and accept shared responsibility for maintaining it. Non-members could agree to supply particular datasets without taking on that wider responsibility. Everyone globally would benefit from the global open data, but the difference is that members help govern and sustain the whole system, while non-members make more limited contributions under separate agreements.

--

I increasingly think that third option may be the most realistic one. It accepts that we may not get one political agreement covering every major country, but we can get something while trying to avoid turning the underlying EO system itself into competing blocs.

I know that is not perfect global governance, but it may be better than making universal adoption a precondition for doing anything or completely accepting a fragmented global EO system.

What the Paris Declarations Changed and What They Did Not

So, to conclude this, I reckon this is what changed for me after reading the Paris declarations.

In December last year, I was asking whether EO needed a political framework at all and made a case for one.

In June, I tried to make that idea more concrete: a principle-based Accord around a designated global baseline, with commitments on transparency, interoperability, crisis sharing, provenance and continuity.

But, now, in September, some governments have done something that gets us surprisingly close to the first step – they have collectively stated that openness, accessibility, interoperability and integrity of EO data matter, they have described data sharing as a collective responsibility and separately, they have explicitly called Copernicus a global public good whose continuity should be protected.

This is meaningful progress, but it also exposes the next set of questions much more clearly:

• If Copernicus is a global public good, why should its continuity remain mainly Europe's responsibility?
• If space-based scientific data is a collective responsibility, what is each country actually contributing?
• If reciprocity matters, what does reciprocity mean between a country with EO satellites that shares its data openly and one with EO satellites but shares nothing to the global baseline?
• Lastly, if some of the largest EO powers do not join the framework, can it still become a global baseline? And if the answer is no, do we wait for global consensus or start with a coalition and build outward?

I still do not have all the answers to those questions.

But, getting back to the question in the title of this essay: can a global EO Accord work without every major country signing it? I think it can—provided that countries outside the Accord can still contribute observations and everyone can use the resulting baseline.

In any case, in just 9 months, since I first wrote about this subject, the discussion seems to have evolved. The geopolitical situation has shifted so much so that it feels foolish to be writing about collaboration and cooperation.

The two Paris declarations have given us a more concrete starting point for building a political framework for EO, one through which governments are beginning to discuss a global EO baseline.

The harder job now is turning that endorsement into actual commitments — who provides them, on what terms, and what happens when their continuity is threatened.

If not, these declarations aren't an EO governance framework, they are just pages on the French presidential website.

Recognising EO as a global public good is one thing, but building a global EO governance framework that actually keeps that public good alive is another. And, we are still very far away from that.

Until next time,

Aravind