<p>To better understand the variability of precipitation over the Asia-Pacific region, atmospheric moisture anomalies associated with two Indo-Pacific indices of variability, NINO3.4 and Pacific-Indian Dipole, are analysed for the ERA5 reanalysis and for simulations by 35 CMIP6 climate models, including ten that provide vertically integrated moisture flux data. The anomaly patterns associated with the indices are determined using linear regression of detrended seasonal means for the period 1980–2019. The simulated patterns of 12 variables are validated with those from ERA5 over the Indo-Pacific domain. This study shows that the skill for precipitation is closely related to those in the moisture flux and its convergence across the ten models. The overall skill for the 35-model means of temperature, pressure, and precipitation exceeds that of any individual model. The indices and associated patterns for the period 2040–2079 under the very high emissions scenario (SSP5-8.5) are scaled to provide a projection of change for a global warming of 2 °C. The ensemble-mean anomaly patterns are mostly similar to those for 1980–2019 over the domain but with enhanced flux and integrated water, indicating a thermodynamic change. Changes in moisture convergence largely account for those in precipitation. The inter-model spread of change in rainfall can be large for both variability and climatological means. For some regions, the changes for precipitation are well correlated across the 35 models with the changes in one or both indices. In particular, changes in all-Australia mean rainfall are linked to those in the Pacific-Indian Dipole index.</p>