The Bureau of Meteorology is in the process of producing a reanalysis that allows us to look back at Australia’s weather for the past three decades. Built using Australia’s national weather forecast model (Australian Community Climate and Earth-System Simulator, ACCESS), this dataset will give businesses, emergency services and farmers a consistent set of past weather data to guide their future decisions.
The Bureau of Meteorology high-resolution atmospheric Regional Reanalysis for Australia, or BARRA, will be a hugely valuable resource: weather datasets for the entire country going back to 1990. It will be especially useful for anyone interested in the kinds of environmental variables that are difficult to measure directly, like wind speeds at different heights and soil wetness. The software optimisation expertise, data management and high-performance computing capacity of NCI are central to BARRA’s success.
NCI works alongside Bureau staff to improve the weather modelling code that is at the core of the reanalysis. This includes decreasing the code execution time by up to 30%, leading to more efficient use of compute resources. In particular, NCI’s expertise provides the supercomputing power and know-how to properly implement all of the improvements at a large scale.
Dr Dörte Jakob, Manager of Advanced Climate Services at the Bureau of Meteorology, says “Running BARRA effectively required us to make use of all of NCI’s capabilities, from high performance computing for this high-resolution reanalysis code, all the way through to the excellent data management systems, data archives and visualisation services.”
The close partnership between NCI and the Bureau of Meteorology allows for important advancements in some of our most critical national infrastructure. The BARRA Reanalysis will benefit a range of sectors, including emergency services, energy companies and agricultural businesses, and through them, the Australian community.
This research highlight was originally published in the 2017-2018 NCI Annual Report.
