Login / Signup

Episodic deluges in simulated hothouse climates.

Jacob T SeeleyRobin D Wordsworth
Published in: Nature (2021)
Earth's distant past and potentially its future include extremely warm 'hothouse'1 climate states, but little is known about how the atmosphere behaves in such states. One distinguishing characteristic of hothouse climates is that they feature lower-tropospheric radiative heating, rather than cooling, due to the closing of the water vapour infrared window regions2. Previous work has suggested that this could lead to temperature inversions and substantial changes in cloud cover3-6, but no previous modelling of the hothouse regime has resolved convective-scale turbulent air motions and cloud cover directly, thus leaving many questions about hothouse radiative heating unanswered. Here we conduct simulations that explicitly resolve convection and find that lower-tropospheric radiative heating in hothouse climates causes the hydrologic cycle to shift from a quasi-steady regime to a 'relaxation oscillator' regime, in which precipitation occurs in short and intense outbursts separated by multi-day dry spells. The transition to the oscillatory regime is accompanied by strongly enhanced local precipitation fluxes, a substantial increase in cloud cover, and a transiently positive (unstable) climate feedback parameter. Our results indicate that hothouse climates may feature a novel form of 'temporal' convective self-organization, with implications for both cloud coverage and erosion processes.
Keyphrases
  • climate change
  • machine learning
  • deep learning
  • high frequency
  • healthcare
  • lymph node
  • molecular dynamics
  • single molecule
  • neural network