Horne et al. (2025) Stress testing water allocations across large river basins

Identification

• Journal: Climatic Change
• Year: 2025
• Date: 2025-12-01
• Authors: Avril Horne, Andrew John, Keirnan Fowler, Ziqi Zhang, Rory Nathan
• DOI: 10.1007/s10584-025-04051-x

Research Groups

• Managing climate change and variability lab, The University of Melbourne, Melbourne, Australia

Short Summary

This study develops a novel, rapid-assessment method to stress test water allocations in large, complex river basins, demonstrating substantial spatial variation in climate sensitivity and highlighting the critical role of water allocation policy in mediating climate impacts.

Objective

• To explore how climate, hydrological responses, and water allocation policies shape future water scarcity and its spatial distribution across a large river basin.
• To assess how different the projections of changes in precipitation and temperature are across the Southern Connected Murray-Darling Basin (SCMDB).
• To understand how differences in rainfall-runoff relationships influence the sensitivity of streamflow responses to changes in rainfall.
• To determine how sensitive water allocations are to changes in streamflow.
• To evaluate the overall vulnerability of water allocations to climate change across the SCMDB.

Study Configuration

• Spatial Scale: Southern Connected Murray-Darling Basin (SCMDB) in Australia, covering approximately 0.5 million square kilometres, divided into 25 sub-areas across 6 major river catchments (Goulburn, Murray, Murrumbidgee).
• Temporal Scale:
  • Climate projections: Future period of 2045–2074 (centred around 2060).
  • Stochastic climate data generation: Informed by the period 1889–2022, with a baseline matched to 1976–2005.
  • Allocation model calibration: Historic data from 1897–2009 (NSW Murray, Murrumbidgee, VIC Murray) and 1943–2021 (Goulburn), including dry climate scenarios where available.

Methodology and Data

• Models used:
  • Rainfall-runoff model: WAPABA model (monthly, updated to include snowmelt processes), calibrated using a multi-objective approach (AMALGAM Pareto optimiser) emphasizing robustness over dry periods.
  • Allocation model: Low-fidelity statistical linear regression model, trained on outputs from more complex jurisdictional water resource models (Bigmod, SMM, SGEFM).
  • Climate data generation: Lag-one multi-site method (Matalas 1967) for stochastic timeseries, incorporating a module for low-frequency climate behaviour.
  • Calibration of statistical allocation model: Differential Evolution Algorithm minimizing a two-part objective function (Success Index and Normalised Nash-Sutcliffe Efficiency).
• Data sources:
  • Climate data: Stochastic climate sequences for precipitation, potential evapotranspiration, and temperature; projections from 37 CMIP6 Global Climate Models (GCMs) using SSP5-8.5 scenario.
  • Rainfall-runoff calibration: Gauged data, CAMELS-AUS v2 database, and a spatial raster dataset of interpolated mean annual streamflow information.
  • Allocation model training: Long time-series outputs from detailed jurisdictional water resource models (Bigmod, SMM, SGEFM) and future dry climate scenarios.

Main Results

• The study developed a novel, rapid-assessment method for stress testing water allocations in large river basins, overcoming computational challenges.
• Climate projections show spatial differences: precipitation decreases are greater in the Goulburn, temperature projections are similar across the three rivers (slightly higher in Murray and Murrumbidgee), and seasonality impacts are less pronounced in the Goulburn.
• Streamflow is most sensitive to changes in precipitation across all systems, with a given change in rainfall eliciting approximately 2.5 times that change in streamflow. The Murray system's streamflow is more sensitive to temperature increases, potentially due to snowmelt contributions.
• Under baseline conditions, average end-of-season allocations are approximately 80% for NSW General Security and 96% for Victorian High Reliability.
• Water allocation policy significantly mediates climate impacts: Victorian "high reliability" allocations are less sensitive to small streamflow changes (±5%) but exhibit a sharp decline in performance for streamflow reductions greater than approximately 25%, indicating a "tipping point" behaviour. NSW allocations show a more gradual decline.
• Overall vulnerability assessment indicates the Goulburn system is the least vulnerable, while NSW systems (Murray and Murrumbidgee) show consistent, gradual declines, with Murrumbidgee allocations potentially dropping to 50% under projected climate change.
• Climate change is projected to have a much greater impact on allocation reliability (percentage of years full allocation is received) than on average end-of-season allocation.

Contributions

• Presents a novel, rapid-assessment method for stress testing water allocations in large, complex river basins, addressing computational tractability challenges of bottom-up climate risk assessment.
• Offers a scalable framework for evaluating climate risks to water resources in regulated river systems globally.
• Provides insights into the spatial distribution of climate change impacts, identifying key drivers of water allocation changes and regions most affected.
• Demonstrates the critical role of water allocation policy in mediating climate impacts and shaping the robustness profiles of water allocations, particularly highlighting differences between NSW and Victorian systems.
• Emphasizes the utility of low-fidelity models within a bottom-up framework for rapid assessment, identifying system vulnerabilities, non-linear dynamics, and thresholds, which can inform more targeted investigations with detailed models.
• Highlights the importance of considering different metrics (e.g., average allocation vs. reliability) for understanding climate risk and tailoring them to specific water management and adaptation decisions.

Funding

• OneBasin CRC Limited

Citation

@article{Horne2025Stress,
  author = {Horne, Avril and John, Andrew and Fowler, Keirnan and Zhang, Ziqi and Nathan, Rory},
  title = {Stress testing water allocations across large river basins},
  journal = {Climatic Change},
  year = {2025},
  doi = {10.1007/s10584-025-04051-x},
  url = {https://doi.org/10.1007/s10584-025-04051-x}
}