A crucial shortcoming of modern climate research is the lack of a long-term benchmarks against which to test the rate of change of key climate indices. As part of an effort to design a system of small satellites to make benchmark observations of spectrally resolved infrared radiation emitted by the earth, we explore the impact of imperfect sampling on the ability of a suite of low earth orbiting satellites to reproduce mean radiance on a variety of spatial and temporal scales. Our aim is find those orbits which maximize the temporal and spatial resolution at which mean radiance can be measured to an accuracy of 0.1 K or better in brightness temperature. This level of accuracy is chosen to agree with the expected magnitude of decadal trends in temperature forced by changes in greenhouse gas concentration.

We model the sampling error of various potential satellite orbital combinations using the Salby Global Cloud Imagery (GCI) 11 micron brightness temperature data set to simulate real variations in radiance. We compare averages over space and time of the GCI data, and of the same data sampled along the paths traversed by a downlooking satellite footprint for each potential orbit, including the role of anticipated instrument error properties. We vary the inclination and altitude of orbits, as well as the orbital stability, to see how these affect retrieval accuracy. Maps of retrieval accuracy are formed for monthly mean and annual mean radiance.