Climate change driven persistence changes in Australian snowpatches

Snowpatches play a key role in alpine and high latitude environments and are highly vulnerable to climate change, particularly where there is no nival zone where they and dependent vegetation communities can retreat. We used a mix-methods approach, centred around satellite and snow course data, to identify and understand long-term trends and changes in persistence for 43 snowpatches in South-eastern Australia. Additionally, an eXtreme Gradient Boosting (XGBoost) model with SHapley Additive exPlanations (SHAP) analysis was applied, proving highly effective in identifying and ranking key predictors of snowpatch persistence. Results demonstrate the relative importance of spatial rather than temporal factors, and of the importance of fetch, elevation, peak snow depth, snow metre days, and ablation season temperature. We attribute the long-term decline in snowpatch persistence to global warming induced changes in the regional climate, with fluctuations in inter-annual variability in persistence caused by complex interactions between climate drivers. Applying changes in the mean day of year ablation date for surveyed snowpatches to the preferred snow cover ranges of snowpatch dependent vegetation communities is shown as an effective means to show potential changes in these communities over time, and through imputation of snowpatch melt date into the future, of probable impending impacts.