Concerns for water resources have inspired research developments to determine the ecological effects of water withdrawals from rivers and flow regulation below dams, and to advance tools for determining the flows required to sustain healthy riverine ecosystems. This paper reviews the advances of this environmental flows science over the past 30 years since the introduction of the Instream Flow Incremental Methodology. Its central component, Physical HABitat SIMulation, has had a global impact, internationalizing the e-flows agenda and promoting new science. A global imperative to set e-flows, including an emerging trend to set standards at the regional scale, has led to developments of hydrological and hydraulic approaches but expert judgment remains a critical element of the complex decision-making process around water allocations. It is widely accepted that river ecosystems are dependent upon the natural variability of flow (the flow regime) that is typical of each hydro-climatic region and upon the range of habitats found within each channel type within each region. But as the sophistication of physical (hydrological and hydraulic) models has advanced emerging biological evidence to support those assumptions has been limited. Empirical studies have been important to validate instream flow recommendations but they have not generated transferable relationships because of the complex nature of biological responses to hydrological change that must be evaluated over decadal time-scales. New models are needed to incorporate our evolving knowledge of climate cycles and morphological sequences of channel development but most importantly we need long-term research involving both physical scientists and biologists to develop new models of population dynamics that will advance the biological basis for 21st Century e-flow science.