Robust and actionable information on climate change and extreme rainfall events in South America

Obtaining robust and actionable information on regional precipitation change to enable adaptation planning and decision-making is a matter of great concern, since there are multiple sources of information.  Projections from large CMIP6 model ensembles (e.g., IPCC Interactive Atlas) show weak signal of climate change in total annual and seasonal precipitation over most of South America (SA), with low agreement between models. Besides, information from smaller ensembles is frequently discrepant. A dynamic framework for climate change in SA is necessary to achieve robust and actionable changes.

Even though they are weak and not robust, the precipitation changes produced over SA by large model ensembles suggest that their main driver is the ENSO increased variability in eastern Pacific, especially intensified El Niño events, produced by transient greenhouse-gas-induced warming. This is consistent with the large impact of ENSO on precipitation in SA. This dynamical framework requires that models used for climate projections in SA demonstrate good simulation not only of the climatology, but also of ENSO and its teleconnections with SA. The assessment of 31 models that provided at least three runs from the present (1979-2014) to the future climate (2065-2100), based on both criteria, selected five best-performing models. This reduced set accurately reproduces the observed seasonal impact of ENSO on precipitation in SA and produces strong and robust patterns of climate change with seasonal variation dynamically consistent with more intense future ENSO in a more El Niño-like mean state.

Since the most dramatic impacts of climate change are produced by changes in the frequency and intensity of extreme precipitation events, it is essential that robust and actionable information is also provided on changes of these events, defined as above the 90th percentile. The analysis is based on the same dynamic framework of the changes in total seasonal/monthly rainfall, since ENSO also exerts a large impact on the extreme events in SA, and the selected set of models shows good simulation of the observed seasonal/monthly impact of ENSO on the frequency and intensity of extreme events. The available information usually shows changes of annual extreme indices. We adopt a seasonal/monthly resolution, which is very useful, especially in a monsoon regime with pronounced annual precipitation cycle. The future changes in extreme events is obtained for SA with monthly temporal resolution and 1 degree spatial resolution. The patterns of change in frequency and intensity of extreme events do not coincide, as changes in frequency depend on dynamic changes, while changes in intensity also depend on thermodynamic changes that determine the precipitable water vapor. Patterns of change in the frequency of extreme events in future are similar to the patterns of El Niño impact on the frequency of extreme events in the present. Changes in the average intensity of precipitation in future extreme events are generally positive and predominate in southeastern South America, where the frequency also generally increases, maximizing impacts on densely populated areas of great importance for agricultural and energy production. The provided information contributes to increase societal preparedness to extreme precipitation in SA.