To the Editor

We share the excitement on the launch of OCO-2 and appreciate the editorial on the importance of observations of atmospheric carbon dioxide (Nature Geosci. 7, 549; 2014). However, we would like to point out that OCO-2 is essentially an exploratory mission to figure out the extent to which carbon dioxide can be precisely and accurately measured from space. If OCO-2's carbon dioxide measurements improve our understanding of the carbon cycle, this will be a bonus.

From experience with the Japanese GOSAT satellite, which has been in orbit for five years, we know already that there will be significant challenges in interpreting the data, despite the fact that OCO-2 is expected to have higher performance. It is by no means a forgone conclusion that “Carbon dioxide sources and sinks can now be measured from space at high resolution”.

For this to occur, the signals received by OCO-2 first need to be attributed unambiguously to variations in atmospheric CO2 concentrations (via a radiative transfer model), as opposed to variations caused, for example, by aerosols, thin clouds or other artefacts. The stringent requirement for in situ measurements is that they are free of bias to 0.1 ppm against a global ambient level of 400 ppm. Second, even if unbiased CO2 retrievals are obtained, translating atmospheric CO2 abundance to sources and sinks requires the intermediary of an excellent model of horizontal and vertical atmospheric motion — no mean feat.

Extensive validation of both OCO-2's CO2 retrievals and atmospheric transport models will be needed before calculations of sources and sinks are credible enough to contribute in the policy arena. This will require significant investment to supplement existing networks of well-calibrated in situ CO2 measurements from surface and airborne platforms, such as NOAA's Global Greenhouse Gas Reference Network (www.esrl.noaa.gov/gmd/ccgg/ggrn.php) or the Japanese CONTRAIL aircraft observation programme (www.cger.nies.go.jp/contrail/index.html).

OCO-2 and similar satellite missions represent ambitious and necessary experiments of how to determine carbon sources and sinks on the surface. Although the potential of satellite-based CO2 sensors is great, in the long run it will likely be a combination of remotely sensed and in situ CO2 measurements that will produce the trusted answers that society needs.