A number of satellite related issues have come up this weekend: The NSIDC reminded us that satellite sensors are (like all kinds of data) not perfectly reliable and do not last forever. Two satellites collided by accident last week, littering the orbit with dangerous amounts of debris. In San Diego this weekend, I was fortunate enough to attend a meeting with some of the Apollo astronauts and some of the scientists involved in Cassini and the Mars Phoenix missions. And yesterday morning we heard that the Orbiting Carbon Observatory mission launch failed to insert the satellite into orbit, and it is presumably measuring carbon dioxide somewhere at the bottom of the Southern Ocean. Coincidentally, when it came up on the news, I was in a meeting with one of the scientists who had been working on setting up a climate model to assimilate the OCO data in order to pin down the carbon sinks.
All of these events have served to remind me at least, that although the space age is 50 years old, we are a long way from the point where we can take our ability to launch and control off-planet machines for granted. Getting into space was, and remains, a tremendous challenge. This makes the successes we’ve had all the more incredible, and a testament to the hard work the engineers and scientists do over many years before a launch to give the missions the best chance of success.
For the climate-related satellites/instruments – SSM/I for the sea ice, OCO for high-precision CO2 concentrations – there is some redundancy with other existing missions. The JAXA AMSR-E sensor can still be used for sea ice extent (and indeed, SSM/I is still sending enough data back to construct 3 day mean pictures). For CO2, the substitutes are slightly orthogonal – the Japanese Ibuki satellite launched last month will measure CO2 but with a very different footprint than OCO would have used, and the AIRS instrument on Aqua has recently been used to produce a timeseries of mid-troposphere CO2 concentrations since 2004. Nonetheless, both of these other missions should provide some of the information that was anticipated from OCO – though not at the spatial resolution envisaged.
It’s worth discussing a little what OCO was going to be useful for. It wasn’t because we don’t know the average amount of CO2 in the atmosphere and how much it’s increasing – that is actually pretty well characterised by the current station network (around 386 ppm growing at ~2ppm/year). However, the variations about the mean (tens of ppm) have a lot of extra information about the carbon cycle that are only coarsely resolved. The measurements would have been from nearer the surface than the AIRS data, and so closer to the sources and sinks. You would have been able to see point sources quite clearly and this would have been a good check on the national inventories of fossil fuel use, and may have been useful at constraining the rather uncertain deforestation contribution to the anthropogenic carbon dioxide sources. More importantly, the OCO data combined with inverse modelling might have helped with constraining the terrestrial sinks. We know they exist from residual calculations (what’s left over from knowing how much we are adding, and seeing how much is in the air and what is in the ocean), and they’ve mainly been associated with boreal ecosystems from the inverse modelling done so far, but there are quite large uncertainties (see 7.3.2 and fig 7.7 in AR4 Chp. 7). The Ibuki and AIRS data will help with this same issue, but OCO data would have been somewhat orthogonal.
Another important consequence perhaps, is that the upcoming Glory mission may be further delayed since it is slated to use the same launch vehicle as the one that malfunctioned for OCO. Glory has had a troubled past, surviving a number of bouts with near-cancellation, but was basically all set to go in June. This is a big deal because Glory will carry one new instrument (an aerosol polarimeter) that has the unique ability (among sensors flying today) to distinguish between aerosol types in the atmosphere. Currently, aerosol remote sensing can retrieve the total aerosol optical depth, with some ability to discriminate between fine particles and more massive ones, but it can’t tell the difference between sea salt and sulphates, dust or soot. This has been a huge problem for aerosol modellers since it is hard to evaluate simulations of each individual aerosol type (and which consequently are all over the place (AEROCOM)). A polarimeter detects the changes in polarisation associated with the aerosols which differs greatly between the different types. The second instrument on Glory is a total irradiance monitor (TIM) which is needed to prevent a gap from forming in the satellite observations of the sun should the current (6 year old) TIM on the SORCE satellite start to falter.
Ironically, space on satellites is at a huge premium. There are always dozens of possible candidate instruments that could be flown and ensuring that the right mix of monitoring and experimental measurements get made is very hard. For instance, the group behind the polarimeter on Glory were trying to find space on a suitable satellite for years before the Glory mission was resurrected.
All this to say, that while the OCO failure will be devastating for the teams that worked on the mission, the relatively high chances of a complete failure are part of the price to be paid for working on satellite missions. Fortunately, OCO was a relatively cheap proof-of-concept mission and so it might someday get another day in the sun.
(*For the reference in the title, see here).