Extended Data Figure 4: Contributions of water vapour change to model spread in temperature-mediated SWA response.
From: An observational radiative constraint on hydrologic cycle intensification
a, The Gregory method (Methods) is applied to anomalies of globally averaged specific humidity (q) at standard atmospheric levels, total column precipitable water (PW), and upper tropospheric precipitable water (PW(500–200), computed by vertically integrating q between 500 and 200 mbar) to quantify the temperature-mediated response of atmospheric water vapour for each model. The natural log was taken before computing annual anomalies. For each quantity, the symbols (circle, diamond, square) represent the model mean and the whiskers represent the full model spread. The globally averaged annual-mean clear-sky shortwave atmospheric q kernel (Methods) is overlaid (blue curve). b, The cross-model correlation between the responses of water vapour in a and the temperature-mediated clear-sky SWA response (clr-dSWA/dT) computed with model-produced fluxes (black) or radiative kernels (blue). Filled symbols are statistically significant at the 5% level based on a two-tailed t-test33, with degrees of freedom corresponding to the number of participating modelling institutions (14) within the 25 model ensemble.
