Atmosphere's emission fingerprint affected by how clouds are stacked
Clouds, which can absorb or reflect incoming radiation and affect the amount of radiation escaping from Earth's atmosphere, remain the greatest source of uncertainty in global climate modeling.
By combining space-based observations with climate models, researchers are able to derive baseline spectral signals, called spectral fingerprints, of how changes in the physical properties of the Earth's atmosphere, such as the concentration of carbon dioxide or the relative humidity, affect the amount of radiation escaping from the top of the atmosphere. Researchers can then use these spectral fingerprints to attribute changes in the observed top-of-atmosphere radiation to changes in individual atmospheric properties. However, recent research has shown that the way global climate models represent the interactions between clouds and radiation can complicate the process of making these spectral fingerprints. Researchers are finding that what matters is not only the presence or absence of clouds at each location represented in the model but also how the clouds are stacked vertically within each model grid.
Using a simulation experiment to mimic the future climate, Chen et al. tested how different approaches to parameterize cloud stacking affect the attributions of climate change signals in the longwave spectra recorded at the top of the atmosphere. The authors tested three approaches to parameterize cloud stacking and find that the differences in stacking assumptions affected the modeled global mean for outgoing longwave radiation by only a few watts per square meter. The global average for outgoing longwave radiation at the top of the atmosphere is around 240 watts per square meter. However, based on which parameterization is used, similar changes in the portion of the sky covered by clouds (especially the clouds in the middle and lower troposphere) can lead to spectral fingerprints that differ by up to a factor of two in the amplitude.