Extreme monthly rainfall over Australia in a changing climate
Ian G. Watterson, Zhi-Weng Chua and Pandora K. Hope
Journal of Southern Hemisphere Earth Systems Science
66(4) 402 - 423
Published: 2016
Abstract
Motivated by the important impacts of extreme rainfall, this study extends the CSIRO and BoM (2015) analyses and projections of 20-year means and daily extremes to rainfall on the monthly timescale. Frequency distributions for monthly rainfall rates simulated by 40 CMIP5 models for the 1986-2005 period are compared with those from the AWAP 0.25° gridded observational data. Distributions spatially-averaged over Australian regions provide a signature of seasonal rainfall. Composites of months in the top and lowest deciles for each grid point and each of the four seasons are then evaluated, along with the frequency of rainfall rates exceeding thresholds ranging from 0.5 mm d-1 to 8 mm d-1.The simulated changes by 2080-2099 under the RCP8.5 scenario for the various rainfall statistics are assessed. Maps of the ensemble mean of changes of the lowest and top deciles, as a percentage of the 1986-2005 base, partly reflect the tendency for increased mean rain in summer and autumn, with decreases in winter and spring. There is also a change in the frequency distribution, with the top decile rainfall tending to increase and the lowest decile to decrease. Bar graphs are used to represent the range of change across the models, for each of four seasons and four regions. In most cases the bars for each statistic cover both declines and increases, but there is again a shift towards the positive in the progression from lowest decile to top decile. The changes are consistent with a broadening of the distribution of monthly amounts. Model spatial resolution is not a major influence on the changes. These projections for monthly rainfall statistics should be applicable to a range of climate impacts.https://doi.org/10.1071/ES16025
© Commonwealth of Australia represented by the Bureau of Meterology 2016. This is an open access article distributed under the Creative Commons Attribution-NonCommerical-NoDerivatives 4.0 International License (CC BY-NC-ND).