Regions along the edges of the tropics host vast populations and ecosystems which are sensitive to climate change. Here we examine the extent of tropical climate land areas in the ERA5 and MERRA-2 reanalyses in high-emission scenarios of 45 models participating in phases 5 and 6 of the Coupled Model Intercomparison Project (CMIP5/6). Based on the definition of tropical climate land areas as regions where the diurnal temperature range exceeds the seasonal temperature range, we find a net reduction of tropical land area with global warming. This change is primarily due to an increased seasonal temperature range, driven by enhanced summer warming. The reduction in tropical land area is consistent with the expansion of the subtropical descending zones and with the expansion of drylands with global warming. However, the particular contributions of dynamic and thermodynamic processes are not clear.

Heavy precipitation events (HPEs) can lead to deadly and costly natural disasters and are critical to the hydrological budget in regions where rainfall variability is high and water resources depend on individual storms. Thus, reliable projections of such events in the future are needed. To provide high-resolution projections under the RCP8.5 scenario for HPEs at the end of the 21st century and to understand the changes in sub-hourly to daily rainfall patterns, weather research and forecasting (WRF) model simulations of 41 historic HPEs in the eastern Mediterranean are compared with “pseudo global warming” simulations of the same events. This paper presents the changes in rainfall patterns in future storms, decomposed into storms’ mean conditional rain rate, duration, and area. A major decrease in rainfall accumulation (-30% averaged across events) is found throughout future HPEs. This decrease results from a substantial reduction of the rain area of storms (-40%) and occurs despite an increase in the mean conditional rain intensity (+15%). The duration of the HPEs decreases (-9%) in future simulations. Regionally maximal 10-min rain rates increase (+22%), whereas over most of the region, long-duration rain rates decrease. The consistency of results across events, driven by varying synoptic conditions, suggests that these changes have low sensitivity to the specific large-scale flow during the events. Future HPEs in the eastern Mediterranean will therefore likely be drier and more spatiotemporally concentrated, with substantial implications on hydrological outcomes of storms.

Projections of extreme precipitation based on modern climate models suffer from large uncertainties. Specifically, unresolved physics and natural variability limit the ability of climate models to provide actionable information on impacts and risks at the regional, watershed and city scales relevant for practical applications. Here we show that the interaction of precipitating systems with local features can constrain the statistical description of extreme precipitation. These observational constraints can be used to project local extremes of low yearly exceedance probability (e.g., 100-year events) using synoptic-scale information from climate models, which is generally represented more accurately than the local-scales, and without requiring climate models to explicitly resolve extremes. The novel approach offers a path for improving the predictability of local statistics of extremes in a changing climate, independent of pending improvements in climate models at regional and local scales.

End of century projections from Coupled Model Intercomparison Project (CMIP) models show a decrease in precipitation over subtropical oceans that often extends into surrounding land areas, but with substantial intermodel spread. Changes in precipitation are controlled by both thermodynamical and dynamical processes, though the importance of these processes for regional scales and for intermodel spread is not well understood. The contribution of dynamic and thermodynamic processes to the model spread in regional precipitation minus evaporation (P-E) is computed for 48 CMIP models. The intermodel spread is dominated essentially everywhere by the change of the dynamic term, including in most regions where thermodynamic changes dominate the multi-model mean response. The dominant role of dynamic changes is insensitive to zonal averaging which removes any influence of stationary wave changes, and is also evident in subtropical oceanic regions. Relatedly, intermodel spread in P-E is generally unrelated to climate sensitivity.

A basin-wide significant shallow bias is found in the southern tropical Pacific thermocline in an ensemble of models from the coupled model intercomparison project phases 5 and 6. In contrast to observations, where the southern thermocline is far deeper than its northern counterpart, models have a hemispherically symmetric tropical thermocline. The shallow thermocline bias is closely related to the well known double intertropical convergence zone (ITCZ) bias, as shown by ensemble partitioning. The physical thermocline (i.e., depth of maximal vertical thermal gradient) is more strongly linked to the double ITCZ bias than the commonly used 20$^\circ$C isotherm thermocline proxy. A shallow thermocline bias is further found to be associated with wider separation of double ITCZ peaks, stronger southern precipitation, a stronger cold tongue, and a spurious south equatorial counter current. Climatic implications and feedback mechanisms between the biases are discussed.