Analysis data derived from bias corrected Australian climate projections

Overview: Novel analysis methods were applied to processed climate model data, including based on outputs previously made available from the Energy Sector for Climate Information (ESCI) project (e.g., ISBN: 978-1-925738-32-2; https://www.climatechangeinaustralia.gov.au/en/projects/esci/). The analysis methods applied here include using expanded ensembles that combine projections per degree warming from different emissions pathways and time periods, thereby increasing sample size for enhanced confidence in results. The resultant analysis data are referred to here as Expanded Ensemble Analysis (EEA) data, provided for mean and extreme values of four variables: daily maximum temperature (tasmax), daily minimum temperature (tasmin), daily precipitation (pr) and the Forest Fire Danger Index (FFDI). Further details on methods, data, uncertainties and guidance for interpretation are in the journal paper doi:10.3389/fclim.2024.1492228 and references therein.

Analysis data details: The EEA data files contain values for the modelled change per degree (Celsius; C) global warming, with individual files provided for the 10th percentile, 90th percentile and mean of the ensemble (also noting potential for changes outside of the 10th to 90th percentile range). Those data are provided for locations where the modelled change is significant at a statistical confidence level above 90%, with a value of -999999 used to fill other regions. For rainfall, some inland regions are also filled with -999999 due to uncertainties associated with very sparse observations data used for the bias correction (estimated here based on where annual average rainfall is less than 140 mm per year during the historical baseline period 1980 to 2005). Data are also provided for the historical baseline period (1980-2005) that these changes are calculated with respect to. Analysis was also done for other time periods during the period of available data from 1980 to 2099, including for comparison with observations-based analysis data for past decades. Data for extremes include analysis of occurrence frequencies for exceeding fixed magnitude thresholds (e.g., tasmax > 40°C, 'gt_40C'; tasmin < 0°C, 'lt_0C'; pr > 50 mm, 'gt_50mm'; FFDI > 50, 'gt_50'), as well as for percentile-based thresholds such as values that may be expected to be exceeded only once per year on average (1-yr average recurrence interval: ARI) at a given location during the historical baseline period (i.e., the occurrence frequency of values greater than the 1-yr ARI: 'gt_1yr_ARI'). Data files are provided here in NetCDF format with zip compression.

Input data: The input data used for the analysis methods applied here include processed climate model data previously made available from the ESCI project (which included bias corrected data from all available dynamical downscaling approaches throughout Australia, e.g., ISBN: 978-1-925738-32-2; https://www.climatechangeinaustralia.gov.au/en/projects/esci/), with further details documented in the journal paper that accompanies this dataset (doi:10.3389/fclim.2024.1492228). The ESCI projections data used a bias correction method known as Quantile Matching for Extremes (QME), trained using analysis data based on observations on a 0.05-degree (approximately 5 km) latitude and longitude grid throughout Australian land locations, with documentation on QME available (ISBN: 978-1-925738-75-9) and Python code adaptation https://github.com/AusClimateService/QME. The analysis data provided here are on a 0.05-degree grid, smoothed with a ±0.25-degree moving average in latitude and longitude. The model data are based on the CMIP5 set of projections including outputs that were available from three different dynamical downscaling approaches applied for the Australian region, for a moderate (RCP4.5) and high (RCP8.5) emissions pathway, noting potential for similar methods to be applied to bias corrected CMIP6 projections and downscaling for the Australian region as data may become available in the future.

General guidance: Aspects of this analysis were selected somewhat arbitrarily, as well as due to the availability of data, noting that similar analysis could also be done using other selections (such as other thresholds for extremes, other time periods for the baseline, etc.). The study detailing these climate analysis methods (doi:10.3389/fclim.2024.1492228) considered a range of different metrics including some details for climate analogues (e.g., finding different locations that have similar past, current or future climate characteristics to each other) and metrics such as the Probability Ratio (e.g., based on the occurrence frequency of extremes in a current climate period compared to a past climate period with less anthropogenic climate change), noting that about 1.1°C of anthropogenic global warming has already occurred. Although data are not directly provided for such metrics, various aspects relating to these types of metrics can be calculated from the data provided here. It is also noted that a wide range of climate projections data and analyses are available including based on outputs from various state and federal agencies in Australia and globally, with benefits often obtained through considering a broad range of lines of evidence. For example, confidence assessment for projected changes can often benefit from drawing on the combination of multiple datasets of observations and modelling as well as physical process understanding (e.g., ISBN:978-1-925738-32-2; doi:10.5194/hess-28-1251-2024; www.publish.csiro.au/mo/pdf/es15008). None of the authors give any representation or warranty of any kind in relation to the information (including data) provided from this page (as well as from pages connecting to this page for this site), including in relation to the currency, completeness, quality, accuracy or suitability for any purpose of this information. You should consider obtaining expert advice before relying on or using this information in any way.

Funding: This research received support from the ARC Centre of Excellence for Climate Extremes (CLEX), as well as the Melbourne Energy Institute (MEI), through the University of Melbourne.