Development of a computationally efficient floodplain ecological response model for large-scale, data-sparse riparian environments

Abstract

Quantitative assessment of floodplain ecological response to flow regimes is challenging but essential for setting targets and estimating impacts for environmental water management. This paper proposes a model that takes long-term (90 years) and large-scale (9 million grid cells) flood maps as input to estimate the response of floodplain vegetation using infinitely differentiable functions. The model, named Floodplain Ecological Response Model (FERM), is calibrated against 1-D temporal Leaf Area Index (LAI) data from the WAVES energy and water balance model at a daily timestep, and validated on the entire floodplain using condition data of the Icon Sites of the Murray River aggregated to a yearly timestep. Results show that FERM can adequately simulate the response of different types of vegetation on the floodplain, while reducing the data requirements and runtime drastically compared to other approaches. The FERM modeling approach is a first step towards a quantitative modeling of floodplain forest ecosystems at large scale with realistic data and computation requirements. It is intended to indicate the potential of such an approach in semi-arid systems where data availability is limited, and to encourage the further research needed to improve our understanding of floodplain forests and our capacity to model the impact of floods on their ecological response.