Can a seasonal prediction system warn of upcoming compound hot–dry extremes in Australia?

Abstract Skilful subseasonal forecasting of hot, dry, and compounding hot–dry events has the potential to support effective planning and risk mitigation. This study evaluates the performance of the Australian Bureau of Meteorology's ACCESS‐S2 seasonal prediction system for hot, dry, and compound hot–dry events across Australia using hindcast data from 1981 to 2018, focusing on lead time‐0 (the first forecast month following initialization), and validation against gridded daily observations. Across Australia, the ACCESS‐S2 ensemble mean shows good skill in capturing interannual variability (IAV) and good fidelity in reproducing interseasonal variability of hot, dry, and compound hot–dry event frequencies, with regionally averaged correlation scores of 0.75, 0.68, and 0.66, and 0.55, 0.94, and 0.67, respectively. Model skill in capturing IAV is highest using the first week of the forecast horizon, in particular for hot extremes, but declines rapidly for longer horizons (weeks 2–4). For compound events, the system displays moderate skill in capturing both timing and frequency. However, event magnitudes are often underestimated, even when the full range of ensemble members is examined. These biases in compound events largely stem from biases in the underlying hot and dry univariate events, but also likely limited model skill in capturing the co‐occurrence of hot and dry conditions. ACCESS‐S2 reproduces the positive relationship between hot and dry conditions evident in observational data but generally underestimates the regional strength of this relationship. The prediction system is also capable of categorical prediction (predicting whether event frequencies are below or above climatological averages), with compound hot–dry events showing higher match rates compared with univariate events. Overall, we provide encouraging evidence of the ability of the ACCESS‐S2 system to forecast compound hot–dry events at lead time‐0 for the month ahead in certain regions across Australia, particularly those with concentrated agricultural activity.

Authors

Institutions

Publication Details

Journal
Quarterly Journal of the Royal Meteorological Society
Published
2026-11-17
DOI
https://doi.org/10.1002/qj.70355
Primary Topic
Climate variability and models
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Can a seasonal prediction system warn of upcoming compound hot–dry extremes in Australia?

Anna Ukkola, A. J. Pitman, Doug Richardson, Mahmood Fooladi
2 links
Quarterly Journal of the Royal Meteorological Society
Climate variability and models
article

Can a seasonal prediction system warn of upcoming compound hot–dry extremes in Australia?

Anna Ukkola, A. J. Pitman, Doug Richardson, Mahmood Fooladi
article en

Abstract

Abstract Skilful subseasonal forecasting of hot, dry, and compounding hot–dry events has the potential to support effective planning and risk mitigation. This study evaluates the performance of the Australian Bureau of Meteorology's ACCESS‐S2 seasonal prediction system for hot, dry, and compound hot–dry events across Australia using hindcast data from 1981 to 2018, focusing on lead time‐0 (the first forecast month following initialization), and validation against gridded daily observations. Across Australia, the ACCESS‐S2 ensemble mean shows good skill in capturing interannual variability (IAV) and good fidelity in reproducing interseasonal variability of hot, dry, and compound hot–dry event frequencies, with regionally averaged correlation scores of 0.75, 0.68, and 0.66, and 0.55, 0.94, and 0.67, respectively. Model skill in capturing IAV is highest using the first week of the forecast horizon, in particular for hot extremes, but declines rapidly for longer horizons (weeks 2–4). For compound events, the system displays moderate skill in capturing both timing and frequency. However, event magnitudes are often underestimated, even when the full range of ensemble members is examined. These biases in compound events largely stem from biases in the underlying hot and dry univariate events, but also likely limited model skill in capturing the co‐occurrence of hot and dry conditions. ACCESS‐S2 reproduces the positive relationship between hot and dry conditions evident in observational data but generally underestimates the regional strength of this relationship. The prediction system is also capable of categorical prediction (predicting whether event frequencies are below or above climatological averages), with compound hot–dry events showing higher match rates compared with univariate events. Overall, we provide encouraging evidence of the ability of the ACCESS‐S2 system to forecast compound hot–dry events at lead time‐0 for the month ahead in certain regions across Australia, particularly those with concentrated agricultural activity.

Quarterly Journal of the Royal Meteorological Society
Australian Research Council (AU), UNSW Sydney (AU), ARC Centre of Excellence for Climate Extremes (AU)
National Computational Infrastructure, Australian Government, University of New South Wales, Australian Research Council
Climate action
Climate variability and models
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon