International cooperation in space can be a very messy business. Just ask Alvaro Gimenez, who recently took over as director of science and robotic exploration at the European Space Agency (ESA) only to learn that two major efforts under his purview must be restructured because NASA will have far less funding to contribute than previously thought.

One of those projects, ExoMars, originated as a European mission but grew into a joint effort with NASA as part of a broader campaign expected to culminate in a collaborative mission to bring martian soil samples to Earth. The joint ExoMars mission was to include two rovers — one built by ESA and one by NASA — to be launched in 2018 and preceded by the 2016 launch of a Mars orbiter and test lander. ESA was forced to stop work on its mission elements after being told that NASA could not afford to do its own rover; tentative plans now call for a jointly built rover to launch in 2018.

The other project is ESA’s next Large, or L-class, science mission. ESA was considering three concepts, each featuring substantial participation by an outside agency; in two instances that outside agency was NASA. Upon getting word that NASA could not take as large a role as envisioned, ESA, which had hoped to move one or two concepts to a more advanced development stage this year, asked its design teams to scale each mission back to something Europe could undertake on its own, if need be.

These examples underscore the inherent risk of international collaboration: It is impossible to predict when fiscal or other circumstances for one partner will change to the extent that the entire effort must be reconfigured, or even canceled altogether.

Mr. Gimenez suggests the best way to mitigate this risk is to eschew missions that would be unaffordable without substantial contributions from an outside agency. Citing ESA’s Herschel and Planck space telescopes, he noted that ESA has the resources and capability to conduct large, groundbreaking science missions on its own.

True enough. But Mr. Gimenez’s suggested risk-mitigation strategy, while sound, has a troubling side: With NASA having made clear it will not be starting any new flagship-class science missions for the foreseeable future, it becomes difficult now to envision any undertakings — collaborative or solo — on the scale of, say, the NASA-led Cassini Saturn mission, featuring a large orbiter and the European-built Huygens probe that descended into the atmosphere of Saturn’s moon Titan. Never mind something like the $100 billion international space station, featuring collaboration by five of the world’s major space agencies. It’s also important to remember that ESA had a significant, unresolved funding shortfall on ExoMars before NASA came aboard.

Yet, having been stung by the experience on ExoMars and L-class, ESA has every reason to be wary of future collaborative efforts. If there’s a surefire way for an agency to inoculate itself against this type of thing it’s to avoid putting international partners on the so-called critical path to executing large, complex missions.

But as Ed Weiler, NASA associate administrator for space science, warned recently, single-agency flagship-class missions are not in the cards right now barring unforeseen changes in the fiscal outlook. He was speaking for NASA, of course, but if the world’s biggest civil space agency cannot go it alone on a grand scale, it is hard to imagine any others doing so.

That would mean no future missions like Voyager — twin probes launched in 1977 that have reached the edge of the solar system — or the Hubble Space Telescope, which has yielded a bonanza of scientific discoveries and is still going strong. ESA, NASA and other civil space agencies must find ways to mitigate the risks of collaboration without eliminating what arguably is its biggest benefit: enabling missions that otherwise would not be possible.

Interestingly, the space station partners have begun looking at ways to use the facility as a test bed for common technical standards that would facilitate international cooperation on other projects. The station is finally complete after more than two decades of cooperative effort and yet, as one ESA official recently pointed out, the computers used on the U.S. portion of the space station are different from those used on the European side. That was just one example.

That the station partners are focusing more on interoperability, and common technical standards and interfaces, is a very good sign; it shows a recognition that the most ambitious future space exploration endeavors will be international from the start.

International cooperation poses unique and frustrating challenges, to be sure. But prospective partners should respond not with defensive strategies that encourage duplication and limit what ultimately can be accomplished. Rather, they should choose major cooperative missions carefully, and develop mechanisms to accommodate the funding and programmatic hiccups that will inevitably arise among one or more individual partners.