Abstract/Summary

Observing Earth’s energy flows is vital for monitoring, predicting and explaining climate changes. The top-of-atmosphere outgoing energy flows, comprised of reflected-solar and emitted-thermal radiation, pose a particular observational challenge due to their rapid evolutions. Recent technological advances have presented an exciting opportunity to capture and understand outgoing radiation variability from a constellation of satellites, yet this potential has not been explored. Firstly, to evaluate the potential of the constellation concept, a new recovery method is developed and a series of simulation experiments are conducted. Using simple broadband radiometers as an example, a constellation of 36 satellites is found to be capable of monitoring global outgoing radiation to a spatial resolution of 1000 km at hourly time scales. The error in recovered daily mean reflected-solar and emitted-thermal irradiance is 0.16 and −0.13 W m−2, with estimated hourly uncertainty of 0.45 and 0.15 W m−2, respectively. Secondly, to gain insight into diurnal variability that a satellite constellation could reveal, dominant diurnal patterns are extracted from modelled global outgoing radiation. Diurnal patterns related to solar zenith angle (88.4 %) and both convective and marine stratocumulus cloud (6.7 %) are found to dominate the reflected-solar radiation, whereas surface heating (68.5 %) and convective cloud (16.0 %) dominate the emitted-thermal radiation. Strong coupling between radiation and other model variables support these processes, and reveal an intriguing time lag in the radiation response to cloud variations. Finally, to assess the additional value of spectrally resolved observations from the constellation, top-of-atmosphere short-wave reflectance spectra are computed. Clustering of the computed spectra reveals signatures related to distinct surface properties and cloud regimes. By assigning real observations to computed clusters, intra-annual and inter-annual variability associated with the West African monsoon is detected, including a 14.5 % annual range in relative frequency of a vertically-distributed cloud signature, demonstrating an alternative route forward for monitoring our climate system.