Evaluation and Future Projections of Hourly Precipitation Extremes from ERA5- and CMIP6-Driven Regional Climate Simulations of over Australia
This study evaluates ERA5- and CMIP6-driven regional climate simulations against observational data to demonstrate that while future projections of Australia's hourly precipitation extremes vary by emission scenario and model, robust increases in the frequency and magnitude of these events are expected under high-emission conditions, underscoring the critical need for emission reductions and multi-model ensembles to improve flood risk assessments.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Rain is a fundamental part of life on Earth, but when it falls too fast, it becomes a destructive force. While scientists have long understood how daily rainfall totals change with a warming climate, the most dangerous floods often happen on a much shorter timescale. These are flash floods, caused by storms that dump massive amounts of water in just an hour. To understand and predict these events, researchers must look at the intensity of rain falling in single hours, not just the total for the day. This is a difficult task because the atmosphere is complex, and the tools used to simulate it—powerful computer models—have historically struggled to capture the violent, short-lived bursts of rain that drive urban flooding and infrastructure failure. As the planet warms, the air holds more moisture, which can fuel these intense downpours, but the exact pattern of how this will play out across a vast and varied continent like Australia has remained unclear.
A team of researchers set out to solve this puzzle by testing the latest generation of climate simulations against real-world data and projecting how hourly rain extremes will change in the future. They focused on Australia, a continent where floods are a recurring and costly hazard, affecting communities from the tropical north to the temperate south. The team utilized five different regional climate models, which are sophisticated computer programs that zoom in on specific areas to provide detailed weather forecasts. These models were run using two different types of input: one set used historical weather data to see how well the models could recreate the past, and another set used future climate scenarios to predict what might happen by the end of the century. The researchers compared the model outputs against a high-quality global dataset that combines satellite, radar, and ground station observations to create a reliable picture of what actually happened.
The evaluation of the past revealed that the models are quite good at capturing the general patterns of rainfall across Australia. They successfully reproduced where rain falls most often and how it changes with the seasons. However, when it came to the most extreme events—the rare, one-hour bursts of rain that cause the most damage—the models showed significant weaknesses. While they could mimic the average behavior of rain, they often underestimated the sheer intensity of the most severe storms. Some models were better than others; a few specific configurations managed to replicate the observed extremes with much greater accuracy, while others consistently missed the mark, particularly in the tropical north where the most intense convection occurs. This finding highlighted that while the broad strokes of climate are understood, the fine details of extreme weather remain a challenge for current technology.
When the researchers turned their attention to the future, the results depended heavily on the path humanity chooses regarding greenhouse gas emissions. Under a scenario where emissions are drastically reduced, the changes in hourly rainfall extremes were weak and inconsistent, with different models disagreeing on whether rain would get slightly wetter or drier. However, under a scenario where emissions continue to rise, the picture became much clearer and more alarming. In this high-emission future, all the models agreed that the intensity of hourly rainfall would increase significantly across the continent. The most extreme storms, those that dump the heaviest amounts of water in the shortest time, were projected to become much more frequent and much stronger. This intensification was not uniform; the northern and eastern parts of Australia were expected to see the largest absolute increases in rainfall intensity.
Perhaps the most striking finding was how the rarity of an event influenced its future risk. The researchers calculated how much the frequency of these storms would increase for every degree the planet warms. They found that while common heavy rain events would become more frequent, the rarest and most violent storms would increase at a much faster rate. For the most extreme hourly rainfall events, the frequency was projected to rise by more than thirty percent for every degree of warming. This means that the most dangerous flash floods, which are currently rare occurrences, could become significantly more common in a warmer world. The study suggests that the risk of catastrophic flooding is not just increasing, but accelerating disproportionately as the climate warms.
The study also emphasized that the uncertainty in these projections comes largely from the differences between the models themselves. Because the models use different mathematical approaches to simulate the atmosphere, they produce different results, especially for the most intense events. However, despite these differences, the direction of the change is consistent: a warmer world will bring more frequent and more intense hourly rainfall extremes. The researchers concluded that to manage future flood risks, planners and policymakers must rely on these ensemble projections, which account for a range of possibilities, rather than a single model. They also noted that future research should focus on even higher-resolution models that can explicitly simulate the physics of storms, which could further improve our ability to predict these dangerous events. The work underscores a critical reality: while we cannot yet predict the exact magnitude of every future storm, the trend is clear, and the most severe impacts are likely to be felt most acutely in the regions already prone to intense rainfall.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.