Predictions and assessments of water availability across large regions are required so that water managers and policy makers can adequately plan for current and future water availability and water use. A major issue in making such predictions is that, in general, only some catchments are gauged and thus estimates of discharge must also be made for large ungauged areas. Numerous techniques have been developed to make these estimates, with most relying on some sort of hydrologic regionalisation. The regional calibration approach where all catchments, within a hydrologically and climatologically similar region, are calibrated together to yield one set of optimum parameters for the region is getting increasingly used to overcome issues associated with the classical regional approaches. For undertaking water assessment and accounting in regulated and unregulated river systems, spatio-temporal estimates of streamflow across the whole basin are required for river system modelling, which considers all the important hydrological processes (e.g. river routing, overbank flow, surface water – groundwater interactions, etc.) and anthropogenic water demands (e.g. irrigation diversions, storages, urban water and environmental water etc). This session invites papers which (i) examine how best to undertake regional scale water balance modelling and regionalise hydrologic response across large gauged and ungauged areas, with a particular focus on assessing/quantifying the uncertainty in the regionalised predictions so that they can be used with confidence to undertake catchment management and river basin planning, (ii) examine the impacts of regionalisation on streamflow characteristics (e.g. high flows, low flows, number of dry days/flow below a certain threshold, sequencing of high and low flows, etc.) for use in aquatic ecosystems and flood risk management, and (iii) demonstrate large scale applications of river system models for water assessment and accounting including any innovative modelling approaches in process representation, calibration, integration with landscape and landcover models.