Modelling future extreme weather

Extreme weather is the sharp end of climate change. Aotearoa New Zealand is highly vulnerable to the impacts of extreme weather like flooding from heavy rainfall. Homeowners, the government and insurers are just some of those eager to know what the future holds in store.

New Zealanders need information on climate change at scales they can usefully use for their particular needs. Climate change information e.g. changes in temperature, precipitation and wind are needed to drive a range of onward modelling in the chain that, ultimately, links increased greenhouse gas emissions (and other aspects of human influence) to the eventual impact of interest for example crop growth or flood extent. Many of these impacts require their inputs at very local scales.

Due to the nature of extreme events and the computational expense of running climate models, it is
very difficult to extract information on extreme weather using traditional methods, as the number of
available model runs is typically fairly small.

By crowdsourcing computational power across the world, this project is running and analysing thousands of climate model runs to better understand future extreme weather in Aotearoa New Zealand.

Key outcomes:

This project has produced a very large collection of climate model simulations for future +1.5°C, +2.0°C and +3.0°C worlds at a resolution of 50 km, an unprecedented dataset for New Zealand. The large number of model simulations means that the frequency of rare weather events can be counted directly, improving our understanding of the future frequency and intensity of extreme events.
Analysing these large model ensembles, this project has evaluated statistical models to better extract extreme rainfall information from one-off, or small ensembles of model runs, with improved understanding of the way model parameters scale with increases in temperature.
The current tool for understanding changes to extreme rainfall with climate change is HIRDSv4 (High Intensity Rainfall Design System). Currently, HIRDSv4 assumes that climate change will have the same effect on extreme rainfall everywhere across the country. This project is investigating regional differences in the climate change signal, and will offer insights and stakeholder-targeted guidance into how to interpret and use the HIRDSv4 data to take into account regional differences.

How this research is being used:

This knowledge will contribute to the next generation of the HIRDS tool for understanding extreme rainfall.
Where the results are robust, this project will offer New Zealanders good understanding of what we expect to happen in their region.
In regions where there is high uncertainty, for example, due to complex topography, this work will identify priorities for future research.

In the media:

Stages of research:

Future climate and weather extremes: 2022 – 2024

Budget: $379,782

Since the first version in 1992, the High Intensity Rainfall Design System (HIRDS) has been Aotearoa New Zealand’s primary tool for characterising current extreme precipitation hazard across the country, and since 2010, also for characterising future extreme precipitation hazard. The current version, HIRDSv4, accounts for how climate change will affect rainfall extremes of different durations and intensities, but it does so at a national scale, ignoring possible regional variations.

This project will examine output from computer models of the climate system that have been conducted over Aotearoa to assess regional differences in climate change-induced extreme precipitation, in order to provide spatially refined guidance on interpretation of HIRDSv4 projections of extreme rainfall hazard, and to inform development of the next version of HIRDS. The project team will examine the output of simulations from the ‘Weather@Home Australia-New Zealand’ modelling system, which provide an unmatched large collection of simulations for analysis under various levels of human-induced warming. We will repeat the analysis with output from other climate models, and synthesise results to characterise evidence and agreement with HIRDSv4 predictions, in order to provide guidance of confidence in and possible bias in HIRDSv4 regional predictions

Future extreme weather in Aotearoa: 2019 – 2022 and Weather@Home: 2015 – 2019

Budget: $575,000 and budget from the Near-term climate predictions for New Zealand project

This work has created new, very large datasets of the ‘Weather@Home Australia-New Zealand’ regional climate model for future warming levels of +1.5°C, +2.0°C and +3.0°C. Such large datasets mean that with so many years of model data we can count even very rare weather extremes directly with some confidence. We applied statistical fits to these datasets and tested how well these fits perform, then subsetted the large datasets to mimic the smaller amount of output that is typically obtained from other climate models (including the New Zealand Earth System Model). We extracted as much information as possible from these smaller datasets and compared different models to refine our methods for extracting extreme weather information from the smaller number of model runs that are typically available

Acknowledgements

This is an international collaboration effort and we would like to thank our colleagues for being part of this effort. The project has long-established links to the University of Oxford and the U.K. Met. Office, as it forms part of ‘Weather@Home’ and ‘climateprediction.net’, which are Oxford-based but have absorbed substantial amounts of expertise from the U.K. Met. Office. There are many well-established links and active collaborations among the project personnel here and world-leading researchers at these U.K.-based establishments. Similarly, the project has now been linked for several years to the Tasmanian Partnership for Advanced Computing at the University of Tasmania, where it still receives computing infrastructure support and the support of Prof. Nathan Bindoff