<p>South east Australia experienced periods of multi-year droughts particularly within the last 2 millennia. However, given the limited evidence from smaller number of sites and scarcity of  quantitative, high-resolution climate records, it is largely unknown whether these droughts are a feature of climate through the Holocene and the extent to which they are experienced throughout the region. Where conditions are suitable, oxygen isotopes preserved in lake sediments are a useful tool for reconstructing past climate and environmental conditions. Here, we present preliminary results of a Holocene length record from Lake Surprise in western Victoria, from which we analysed δ<sup>18</sup>O of aquatic cellulose as a proxy for lake-water δ<sup>18</sup>O, complemented by organic carbon/nitrogen ratios, organic carbon isotopes and XRF (ITRAX) inferred elemental composition. Our interpretation of the palaeo-data is supported by ~3 monthly monitoring of water and sediment geochemistry to track the modern hydrology of the lake. Our preliminary results show a strong positive correlation between precipitation and sedimentary calcium (carbonate deposition) over the last 150 years, likely linked to changes in primary productivity. The aquatic cellulose δ<sup>18</sup>O record through Holocene is also correlated with carbonate concentration, reinforcing our interpretation of CaCO<sub>3</sub> deposition in the lake during wet periods. The cellulose δ<sup>18</sup>O record indicates a trend of gradually increasing aridity from early to late Holocene, with a notable extremely dry phase over the last 2 ka. Comparison of the cellulose δ<sup>18</sup>O record with high-resolution Holocene climate records indicates that multiple climate drivers such as ENSO intensification and Antarctic warming are strongly linked to increasing aridity of the region. Further work will focus on both increasing the resolution of the record to better identify the frequency and duration of key events and on quantifying natural hydroclimate variability, particularly via lake hydrologic modelling to better constrain the paleoclimate record.</p>