What if the rains do not come?

• SPEI/SPI-12 are better predictors of hydrological stores like groundwater (GW). • Influence of global climate teleconnections is better diagnosed with SPEI/SPI-12. • GW loss in Murray Darling Basin occurred about 44% of the time during 2002–2017. • Extreme events (droughts and floods) affect the distribution of surface water extent. • SPEI/SPI-3 are excellent metrics that capture variability in annual crop production. Risk-management solutions that increase opportunities for adaptation plans and resilience to climate change require an understanding of climate change processes and the trickle down effects of droughts and floods on hydrological and socio-economic systems. To guide future expectations about droughts, this study aims to assess large-scale variability of hydrological stores (groundwater, soil moisture, and surface water) and their responses to drought intensities over large semi-arid domains in Australia. Multi-scaled indicators synthesised from in-situ rainfall (SPI-standardised precipitation index), water budget (SPEI-standardised evapotranspiration precipitation index), model (soil moisture-SM), and satellite observations (groundwater storage-GWS, surface water extent-SWE, and terrestrial water storage-TWS) are employed to assess this variability in climatic hotspots identified through statistical rotation. The link between these hotspots and climate modes are diagnosed using gaussian kernel-based SVMR (support vector machine regression), and quantile function of storage (QFS) is used to assess the response of hydrological stores to climate variability. The capability of these indicators to capture impacts of water deficit on agricultural systems is explored by implementing a PLSR (partial least square regression). Results show that (i) drought characteristics and intensities vary greatly across Australia but hotspots of higher drought duration and intensity are predominant in regions below latitude 25 ° S, (ii) hydrological drought indicators (SPEI/SPI-12) are better predictors of hydrological stores such as TWS and GWS, though the latter shows strong but opposite phase relationship with climate data, (iii) influence of climate teleconnections is better diagnosed and predicted with SPEI/SPI-12, and the Murray Darling basin-MDB (r = 0.84 , α = 0.05) and central/east coast (r = 0.80 , 0.79 , α = 0.05) are major teleconnection hotspots, (iv) SM and GWS losses in the MDB occurred 60% and about 44% of the time, respectively, during the 2002 - 2017 period, (v) extreme events (droughts and floods) affect the distribution of SWE, and (vi) meteorological drought indicators (SPEI/SPI-3) are excellent metrics that capture variability in crop production. Choosing appropriate indicators that reflect the response of freshwater ecosystems (including groundwater) to climatic and anthropogenic constraints is part of a key process to guide investments in drought resilience and sustainable management of water resources. [ABSTRACT FROM AUTHOR]