Improving Climate Prediction By Climate Monitoring

Abstract:

Various climate agencies are pursuing concepts of space-based atmospheric monitoring based on ideas of empirically verifiable accuracy in observations. Anticipating that atmospheric monitoring systems based in observing the emitted longwave spectrum, the reflected shortwave spectrum, and radio occultation are implemented, we seek to discover how long-term records in these quantities might be used to improve our ability to predict climate change. This is a follow-up to a previous study that found that climate monitoring by remote sensing better informs climate prediction than does climate monitoring in situ. We have used the output of a CMIP5 historical scenario to hind-cast observation types being considered for space-based atmospheric monitoring to modify ensemble prediction of multi-decadal climate change produced by a CMIP5 future scenario. Specifically, we have considered spatial fingerprints of 1970­–2005 averages and trends in hind-cast observations to improve global average surface air temperature change from 2005 to 2100. Correlations between hind-cast observations at individual locations on the globe and multi-decadal change are generally consistent with a null-correlation distribution. We have found that the modes in inter-model differences in hind-casts are clearly identified with tropical clouds, but only Arctic warming as can be identified in radio occultation observations correlates with multi-decadal change, but only with 80% confidence. Understanding how long-term monitoring can be used to improve climate prediction remains an unsolved problem, but it is anticipated that improving climate prediction will depend strongly on an ability to distinguish between climate forcing and climate response in remotely sensed observables.