Constraining different facets of global water cycle projections with observations*

For the first time in the IPCC history, the AR6 WG1 assessed the future global warming by combining multi-model projections with observational constraints. Yet, it also recognized that such robust methods did not yet exist to constrain the projections for most other variables, so that projected changes in the global water cycle were not fully consistent with the assessed future global warming level. Here, we will illustrate how observations and a Bayesian statistical method can be used to constrain various facets of global water cycle projections, consistently with the observed historical global warming and more robustly than using empirical emergent constraints. A wide range of applications will be covered, including changes in precipitable water, near-surface relative humidity or runoff over land, but also sea surface salinity as a proxy for the water cycle intensity.