Agricultural climate metrics for the National Climate Risk Assessment

Agricultural climate metrics were derived for the National Climate Risk Assessment.\n\nThis data collection describes the agricultural climate metrics derived from projections of climate variables across Australia from 8 CMIP6 global climate models, as well as a historical baseline from the Bureau of Meteorology’s Atmospheric high-Resolution Reanalysis for Australia (BARRA-R2) data. Agricultural climate metrics focus on climate conditions that are important to Australian agricultural productivity, for example the annual number of heat stress days for livestock or total growing season rainfall for grain cropping. The metrics are evaluated for a historical baseline (2001-2020) and global warming levels (GWLs) 1.5, 2.0 and 3.0 degrees Celsius above pre-industrial levels. Data are provided in netCDF format, and each grid location gives the minimum, median, and maximum metric values across the ensemble of global climate models climate models for each GWL, as well as the mean of the historical period. Projection data are provided for values (e.g. total rainfall in mm) as well as change from the mean historical baseline (e.g. % change in rainfall from historical baseline). \n\nMore information can be found in the technical report¹. \n\n1. Darbyshire R, Blamey L, Bolt A, Brodie S, Buhagiar L, Bustamante RH, Greenwood S, Holzworth D, Hopper M, Huth N, Keogh T, Mayberry D, Nidumolu U, Paroz A, Rich J, Rosauer D, Roxburgh S, Thomas I and van Herwaarden A (2025). Primary Industries Technical Report. A technical report prepared for the Australian Climate Service as part of the National Climate Risk Assessment. CSIRO, Australia. https://doi.org/10.25919/g1b2-fj19.\nLineage: DATASETS\n\nClimate data are from two sources. Historical data are from the Bureau of Meteorology using the Bureau of Meteorology Atmospheric High-Resolution Regional Reanalysis (BARRA-R2) dataset (http://www.bom.gov.au/research/publications/researchreports/BRR-067.pdf and project ob53 on National Computational Infrastructure). A baseline of 2001-2020 is used in line with projection data. Application ready CMIP6 climate projection data created using ‘quantile delta method’ were used. The climate projection data are described in Irving and Macadam² noting:\n\n-Temperature (tasmin, tasmax): No difference\n-Precipitation (pr): The version used uses 1000 quantiles with no monthly time grouping. The final published version used monthly grouping with 100 quantiles per month. The adjustment factors were capped at 5.0 and a common multiplicative mean scaling was applied to the data after the fact to ensure the annual mean change matched the original model data.\n-Relative humidity (hurs): The version used additive scaling whilst the final version used multiplicative scaling.\n\nAn ensemble of 8 global climate models were used for three global warming levels (GWLs), 1.5, 2.0 and 3.0°C above the pre-industrial mean for 1850 to 1900. The global climate models used are ACCESS-CM2, ACCESS-ESM1-5, CMCC-ESM2, CNRM-ESM2-1, EC-Earth3, MPI-ESM1-2, NorESM2, UKESM1.\nBoth the historical and projection climate data were provided on an 0.11 degree grid. Climate data are summarized by regions of interest using polygon shape files using:\n\n1. Australia\n(https://www.abs.gov.au/statistics/standards/australian-statistical-geography-standard-asgs-edition-3/jul2021-jun2026/access-and-downloads/digital-boundary-files) \n2. NCRA regions (https://www.acs.gov.au/datasets/4f8b960fd3694fc28d6ff3d1278e9e75_0/about) \n3. Spatially combined NCRA and AEG polygons (https://data.csiro.au/collection/csiro:64133) \n4. Forestry (Australia’s State of the Forests Report - 2018)\n5. Wine Regions (WineAustralia - OpenData, v. 4-Mar-2022)\n\nMETHODS\n\nA full description is in Darbyshire et al. (2025). To assess the impact of climate change in the agricultural setting agricultural climate metrics were evaluated under projected conditions and relative to a historical baseline. The historical baseline uses BARRA-R2 data over the years 2001-2020. The projected timing when each GWL is reached differs by global climate model. For this assessment, a GWL is defined as met when the mean of a 20-year period is exceeded. This period is used to represent the climate for that model.\n\nThese datasets contain daily climate variables with those used in the analysis being maximum temperature, minimum temperature, precipitation and mean relative humidity. Agricultural metrics are used to interpret daily climate variables into indicators that can be used to gauge agricultural impact. As an example "heatwave" is a metric that counts the number of heatwave events - periods of 3 days with maximum temperatures above 35 deg Celsius - from 1-Jul to 30-June. A list of metrics and descriptions are found in NCRA_climate_indices.docx.\n\nThe value for each metric, for each GWL, is determined by averaging over the range of years in each global climate model. Then, the minimum, median and maximum values for the metric across these mean values models are calculated.\n\nData are provided in netCDF format. The change in conditions from the baseline are also included in absolute terms ("_change") and in relative terms ("_percentage_change").\n\nA number of regions of interest (ROIs) are assessed throughout the project. NCRA specific masks are included in this collection. To characterise the change in a metric with GWL the spatial mean of the historical baseline was compared against minimum, median and maximum spatial means of each climate model at each GWL.\n\n2. Irving D and Macadam I. (2024). Application-ready climate projections from CMIP6 using the Quantile Delta Change method. CSIRO, Australia. https://doi.org/10.25919/03by-9y62