Assessment of environmental flow scenarios using state‐and‐transition models. (21st January 2018)

• Record Type: Journal Article
• Authors: Bond, Nick R.; Grigg, Nicola; Roberts, Jane; McGinness, Heather; Nielsen, Daryl; O'Brien, Matthew; Overton, Ian; Pollino, Carmel; Reid, Julian R. W.; Stratford, Danial
• Other Names: Kennen Jonathan G. guestEditor.; Stein Eric D. guestEditor.; Webb J. Angus guestEditor.
• Is Part Of: Freshwater biology. Volume 63:Number 8(2018)
• Page Start: 804
• Page End: 816
• Publication Date: 2018-01-21
• Subjects: drought -- floodplain -- hydro‐ecology -- inundation -- vegetation dynamics; Freshwater biology -- Periodicals; Biologie d'eau douce -- Périodiques; 577.605
• DOI: 10.1111/fwb.13060
• Languages: English
• ISSNs: 0046-5070

Abstract

Abstract: Numerous methods have been developed to support the assessment of environmental flow requirements for rivers. Most methods are based around models of hydrologic time series rather than models of the ecological endpoints of interest. Important limitations that arise from this include (1) an inability to represent the state dependency of response to future conditions (i.e. the effects of current ecosystem condition on future condition), (2) the inability to predict ecological states through time under alternative flow regimes and (3) limited sensitivity to compare the differences between flow regimes with similar return intervals of ecologically important events, but different sequencing of those events. Here we outline a simple state‐and‐transition modelling approach to assess differences in ecological responses to alternative sequences of floodplain inundation events in a lowland river system. Our approach explicitly incorporates the state dependency of biotic response to flooding, thereby representing the influences of both antecedent conditions and current condition (in this case population state; good > medium > poor > critical). Our approach thus captures the influence of the entire historical sequence of flow events via a first‐order Markov chain process. We use prior data and expert opinion to determine state transitions for a broad suite of ecological indicators. Despite being implemented with deterministic transitions, and drawing heavily on expert opinion, this approach greatly improves on existing methods used in environmental flows planning, particularly when comparing scenarios with the different sequencing of ecologically relevant flow events. The outputs from these models are testable, and the approach is readily extensible to incorporate probabilistic state transitions and uncertainty, mechanistic links (via increased model complexity) and quantitative measures of population state (e.g. measures of abundance or tree condition). Most importantly, the adoption of such a framework represents a fundamental shift to modelling ecological endpoints rather than relying on just quantifying hydrologic surrogates to compare environmental flow scenarios.