They are two scientists with more than 60 years of experience between them in the world of space-based meteorological remote sensing systems: Paul Menzel, senior scientist with the Space Science and Engineering Center (SSEC), University of Wisconsin–Madison and Johannes Schmetz, chief scientist with the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT).
They have not only witnessed, but have advocated for and driven, many of the advances in imaging and sounding techniques so critical to the global observing system. Many of these advances bear their fingerprints.
Schmetz and Menzel recently spoke at SSEC’s 50th Anniversary celebration in Madison. The event corresponded to the early online release of their paper, “A Look at the Evolution of Meteorological Satellites: Advancing Capabilities and Meeting User Requirements,” to be published in Weather, Climate, and Society.
We spoke about the state of meteorological satellites and the global observing system: the challenges of developing applications, balancing requirements and capabilities with user needs, managing collaborations between public and private sector organizations, and creating a model for international collaboration into the future.
Their conversation follows.
On the needs of research versus the needs of operational systems:
Paul Menzel (PM): As we considered the article, Jo had a view of the big picture and I had a narrative of different things that had gone well and things that hadn’t gone well in the preparation and design of new satellite systems in the United States. I could also see progress in Europe because I visited there often enough. In the U.S. we were gradually pursuing a very different path because we were principal investigator (PI)-led — people like Verner Suomi and Bill Smith [at the University of Wisconsin–Madison] were demonstrating instruments that morphed into operational instruments. That was a model that probably could not work forever because a PI really isn’t interested in operational responsibility — he or she wants to keep testing new things, whereas operational really has to be sustained for 5 or 10 years, or whatever the long-term means.
Johannes Schmetz (JS): This is exactly the right model for development agencies, such as NASA or the European Space Agency (ESA) both of which include an approach that is partly PI-driven and where the goal is demonstrating novel things. But for operational or sustained observations from a satellite where the budget is relatively lean — at least in Europe — the program needs to be very focused on the things we have to do.
PM: There has to be a transition. Louis [Uccellini, National Weather Service Director] mentioned recently that what really worked for a long time in the U.S. was the Operational Satellite Improvement Program (OSIP). NASA would introduce a variation of an existing instrument or a new instrument and they would demonstrate it with NOAA already on board to evaluate and test it in order to see if this was something that they wanted to carry into operations. There was a budget on both sides — NASA and NOAA — there would be a planned transfer of monies in the national budget so that after the demonstration phase, you planned for success and then operations would take it on. The NASA budget would dwindle and the NOAA budget would accelerate up. This approach was phased out in the 1980s.
Ever since then, anything that is new must be built for an operational satellite with a plan to use it for 20-30 years. The problem with this approach is that if it’s new, you don’t know if you’ll want it; or whether it will succeed. It’s a question. We, as scientists, really needed this demonstration phase that we lost.
In Europe, ESA, the research agency, was typically on board with EUMETSAT, the operational agency.
JS: As teams, ESA and EUMETSAT, work well together. And what Paul says, when it comes to operational satellites, the roles are very clear. Operational agencies select proven technology that can, however, be far-reaching. For example, with the Infrared Atmospheric Sounding Interferometer (IASI), EUMETSAT flew the first hyperspectral interferometer as an operational instrument. However, the technology as such was known, so CNES, the French space agency that built it, was confident and delivered a great instrument.
PM: The issue, in a way, is that you can’t have the very old blocking the very new, so the question becomes: Can you have planned obsolescence? That is, can you plan for something to last a certain amount of time? We always want to get the most out of an instrument especially when it is still operating. It’s like driving an old car for 20 years knowing there are safety risks and a new car would have improved safety features, but you want to get as much mileage out of the old one as you possibly can.
It’s a similar situation with a rocket launch today. The launch is so expensive that you want to be sure that you get your money’s worth and sometimes that means the instrument needs to last for 15 years. Fifteen years in technology terms, however, is almost a lifetime. It’s too long.
We are still using satellite technology from 30 years ago — how many things that you plug into the wall in your house were built 30 years ago? Very few. If a computer is more than five years old you think about replacing it because of increased risks of information loss or vulnerabilities to hacking.
On balancing the need for innovation with the need for certainty of performance:
JS: How do you balance the need for innovation with the requirement that something has to last for a decade or more? It’s never risk free, but what we select for operational satellites has to be proven and affordable technology. Lead engineers, scientists, and users must agree that the system development risk is relatively low and that the instruments will provide the basis for continued and enhanced services in weather forecasting and environmental and climate monitoring.
In terms of innovation, we don’t have to invent a technology that does not exist. Using sensor technology as an example, scientists want to improve the signal to noise measurement for very good scientific reasons and for the benefit of the user. Immediately, though, the question becomes: Do those detectors already exist? If they don’t exist now, but they will exist 3-5 years from now, it’s probably already too risky for an operational program. For a research program, which has a complementary role, it’s probably the right way to go.
PM: Jo mentioned the key word, risk. In the U.S., for budget reasons and for management reasons, we’ve become so risk averse that we’ve forced impossible guarantees. We’ve told industry that they must guarantee success — 100% — and their response is that this sort of guarantee will cost a lot of money.
However, if both parties agree upon a 5% risk, for example, then you are working together to make that decision, but if you make success a requirement of industry, you’re placing all responsibility on them. In the end, you never escape the risk. If a satellite fails — fails to send a weather picture down — it comes back to NOAA or the National Weather Service. And the question of NOAA becomes, “how come you don’t have this advancement or that instrument?”
Whether industry failed is lost on the public. As NOAA became more and more risk averse, we ended up with more management people running programs without an understanding of what constituted a balanced approach. That is, how much new could we combine with the old to improve service but still provide continuity of established services?
Continuity became the watchword.
JS: We have many stories that we did not include in our paper. Paul, for example, included the water vapor channel, and that was a good example. Our Meteosat First Generation had a water vapor channel — it was an activity led by Pierre Morel, from CNES, after the development of the satellite had already started. It was a visionary — and successful — idea.
On relationships with industry, government, and academia:
PM: When you are embarking on something new, it’s difficult to write a set of requirements because what you are proposing has never been done before. More difficult yet is to fix the design before the instrument or sensor is actually built. That’s not how things used to be done — with Suomi and Parent, scientist and engineer, who modified designs as they went along. That’s not how it’s done now because there is a distrust between buyer and vendor where the government is the buyer and industry is the vendor. Added to the mix are universities, who are often accused of letting requirements creep — it’s simply human nature to want more and to keep adding capabilities. But, ultimately, we want this instrument to operate, come in on budget, and be new when it launches.
JS: We want it to be better than what we had. Basically, when I take the user perspective, what I want is a continuation of good services that I currently have and at the level to which I’ve been accustomed. Then, if possible, we want to improve the existing good services. That in turn leads to the typical requests for higher spatial resolution, higher radiometric accuracy, more spectral channels, etc. As an additional element, we consider innovation; that is, new observations enabling new services. This is the basis of what users want.
We think this approach works best when a few leading scientists and visionary engineers are asked for input and write the requirements together. This is an important ingredient in the process — and, by the way, we think that Vern Suomi, for example, would have been an excellent, pragmatic scientist to include in the process.
PM: With regard to the cooperation between key players, it is important to remember that it is in government’s interest and university’s interest to make sure that industry makes a good profit, because our industry partners are developing some things where there is zero profit. You want to sustain these relationships — and relationships with more than one provider.
We want them to succeed at their game and we want to succeed at our game — the government has a responsibility to its citizens and if any one of the players gets too little or too much it results in an imbalance that isn’t good for the whole system. Not only is mutual trust needed in this cooperative game, it is absolutely imperative to have technical and scientific knowledge and judgment on the government side.
JS: Balance is an important word here. The rules of the game have not always been clearly defined among the major players of industry, government, and science. It’s important to get the rules of the game straight from the beginning so that everyone knows their roles.
On the importance of global cooperation among satellite agencies:
PM: Global cooperation is happening through the World Meteorological Organization and some of these collaborations are magnificent. Some of this began because EUMETSAT and NOAA NESDIS shaped a bilateral agreement. The directors from EUMETSAT, John Morgan and Tillman Mohr, and from NOAA — really, extending back to Dave Johnson (NESS) and then, Tom Pike (NESDIS) — encouraged partnerships, inviting first the Japanese and then the Russians, to the table.
EUMETSAT and NOAA have periodically repositioned one of their satellites to provide coverage for the other agency in order to avoid gaps in coverage. This exercise really set the stage for future agreements between satellite agencies.
The Coordination Group for Meteorological Satellites (CGMS) provides a forum for all of the satellite providers for civilian environmental satellites to gather and talk about satellite programs and how to fill the observing gaps in the global observing system [and to facilitate access to and quality of data, help with contingency planning, and provide support to users]. Currently, the parties have arranged for satellites to be placed in orbits where they fill observing system gaps. You have the geostationary satellites spaced appropriately and most importantly, in the polar-orbiting area, we now have the Chinese willing to take on the early morning orbit which has sunlight challenges.
What we don’t have is one national provider offering to use their budget to develop a specific instrument and another national provider offering to develop a different instrument, with the agreement — and trust — that copies or data files will be shared. This remains a challenge.
JS: We have several things that we feature in the paper and this is the last issue that we discuss. It’s a difficult issue and a longshot because it may not be compatible with current industrial interests and policies if each country or international organization does not build all of their own instruments. This is also a political issue, but we characterize it as a vision for the future.
In fact, the idea is probably not so far-fetched. For instance, a pertinent question arose during the special seminar given by Stephen Volz, NOAA Assistant Administrator for Satellite Services for the SSEC 50th Anniversary on this very topic. Following the presentation, someone asked him, why don’t national agencies agree that “if you do that, and they do that, then we will do this.”
It’s an interesting question to consider and is exactly what we are talking about.
On the importance of data sharing:
PM: I was in the satellite business during what I consider to be the golden era; when the United States had a very strong, robust program. When I first came on, it was almost exclusively U.S. satellites and NOAA was very open with the data sharing. If you traveled anywhere in the world and told scientists what you were doing with the data, you would learn from them what they were doing with the same data. These scientists would also express tremendous gratitude that these data were available to them. Radiosondes can give you a bit of information, but satellites really provided blanket coverage of the planet and over the continents.
We were doing interesting work and were well-received wherever we went. As a result, we gained international colleagues who were also giving something back to us. Very soon, Europe came, Japan came, and other countries started launching instruments and that collaborative environment and spirit continued.
It has to be free and open access to the data — it’s for the planet; it’s for the people.
It’s not for profit.
JS: We’ve said it many times — at meetings and elsewhere — that the world really needs to be thankful to the United States. They provided the capability and set the tone and shared their observations with the rest of the world. No boundaries. It was fantastic. We should be reminded of this from time-to-time and perhaps it can be a model for other geo-political relationships.
PM: Looking back over the last 30 years, there was clearly a camaraderie that developed among the leaders of the national satellite services. That was the excellent thing about the Coordination Group for Meteorological Satellites — we would say that we needed some scientists from the U.S. to work on a program with the Europeans, and the administrators would make it happen in short order. People were reassigned or resources set aside so that these things could happen. If you wrote a proposal from the bottom up, it would take months, maybe years, whereas this arrangement — resulting from the positive relationships built over time — would make it happen almost immediately.
A year later, you would meet to discuss what you’d done. A year is a relatively short time for some of these programs to move forward.
Jo and I were lucky enough to be at meetings where we were given a fair amount of latitude to suggest things — not everything we suggested got done, but quite a few of our ideas were embraced and became reality. The inter-calibration of all the satellite sensors is just one example. We pushed very hard at these meetings and it happened.
JS: Right. If you look back at the report from the CGMS meeting in St. Petersburg, Russia in 1997, there are two or three related actions which were very explicit. That’s when it started. One action said that all satellite operators should calibrate their infrared geostationary window channels against the AVHRR window channels. The concept evolved from there.
I would even take a step further back — when Paul visited us, not at EUMETSAT, but at ESA, when I was there, the topic we spent the most time discussing was inter- calibration between GOES and METEOSAT. We wrote a report about this concept. When you start talking about the history of global inter-calibration, I think that was a beginning. It was the first time this was discussed and done for an operational satellite.
PM: For me, it became clear that you could not pursue the study of weather and climate without international cooperation on inter-calibration. That cooperation is in very good shape right now.
In the U.S. we’ve sort of alluded to the fact that we haven’t progressed with our satellite systems as far as I think — and this is my opinion — that we could have or should have. The rest of the world is contributing remote sensing capabilities that are filling that void. So it’s not an entirely U.S.-driven system. And this is very good. Hopefully, the U.S. can come back and demonstrate some new things in the future, but right now, the lead is elsewhere.
JS: In some areas, that’s true, but in many areas, when it comes to true innovation, it’s still the U.S. Take MODIS or CLOUDSAT — both of them were NASA principal investigator-driven missions and both have had an enormous impact on Earth observation and science.
PM: The Earth Observing System was extremely informative and is still providing new insight.
JS: And sometimes a realization takes time. For instance, the idea to fly the first wind lidar was already presented at the very first International Winds Workshop in 1991. We were both there in Washington. Implicitly, it was said that atmospheric motion vectors were kind of outdated and we would fly Doppler lasers to measure the wind directly. But now, nearly 25 years later, this mission is still not without difficulty. This is just one corroboration of the fact that new technology needs to be proven before it can be operational.
PM: This is a paper where two old guys are sitting back and saying, “Do you remember back in the old days?” It was good.
JS: And it’s still good. We wanted this paper to be a very personal perspective highlighting the positive. What we didn’t want was a report format. Basically, it’s a paper from Paul and Jo.
PM: We also need to recognize — and remember — that there are many things that were done in ways that I would not have recommended, but those engaged in the activities were doing the best that they could. This was a best effort — no one was looking to cut back the system or looking for it to fail. As a result, we made some mistakes and some things could have been planned better, but in hindsight, most of it was very, very good.
Now we have the opportunity, with new players, to bring new ideas to the international agreement. The hardest thing will be to convince the globalization of industry that different countries will have different specialties in the remote sensing game.
JS: That last point is going to be very tough yet potentially very rewarding.
PM: The challenge is there — we need observations that we don’t have right now. Things are moving very fast with the planet. Our challenge is to move fast enough to keep up with its changes.
By Jean Phillips