Comparison of published palaeoclimate records suitable for reconstructing annual to sub-decadal hydroclimatic variability in eastern Australia: implications for water resource management and planning

Abstract

Abstract. Knowledge of past, current and future hydroclimatic risk is of great importance. However, like many other countries, Australia's observed hydroclimate records are at best only ~ 120 years long (i.e. from ~ 1900 to present) but are typically less than ~ 50 years long. Therefore, recent research has focused on developing longer hydroclimate records based on palaeoclimate information from a variety of different sources. Here we review and compare the insights emerging from 11 published palaeoclimate records that are relevant to annual to sub-decadal hydroclimatic variability in eastern Australia over the last ~ 1000 years. The sources of palaeoclimate information include ice cores, tree rings, cave deposits and lake sediment deposits. The published palaeoclimate information was then analysed to determine when (and where) there was agreement (or uncertainty) about the timing of wet and dry epochs in the pre-instrumental period (1000–1899). The occurrence, frequency, duration and spatial extent of pre-instrumental wet and dry epochs was then compared to wet and dry epochs since 1900. The results show that instrumental records (~ 1900–present) underestimate (or at least misrepresent) the full range of rainfall variability that has occurred, and is possible, in eastern Australia. Even more disturbing is the suggestion, based on insights from the published palaeoclimate data analysed, that 71 % of the pre-instrumental period appears to have no equivalent in the instrumental period. This implies that the majority of the past 1000 years was unlike anything encountered in the period that informs water infrastructure, planning and policy in Australia. A case study, using a typical water storage reservoir in eastern Australia, demonstrates that current water resource infrastructure and management strategies would not cope under the range of pre-instrumental conditions that this study suggests has occurred. When coupled with projected impacts of climate change and growing demands, these results highlight some major challenges for water resource management and infrastructure. Though our case study location is eastern Australia, these challenges, and the limitations associated with current methods that depend on instrumental records that are too short to realistically characterise interannual to multidecadal variability, also apply globally.