Get started
Work

Spectrally transformed CMIP6 decadal projections improve interannual rainfall forecasts

Year: 2025
Type: article
Abstract: Forecasting rainfall over long lead times (from interannual to decadal scales) is essential for informing water resource management and adaptation planning, especially for drought-prone countries such as Australia. The latest suite of Coupled Model Intercomparison Project Phase 6 (CMIP6) models provides simulations of decadal prediction experiments, although their skill in simulating rainfall remains limited. This study investigates the merit of using spectrally transformed simulations of sea surface temperature anomaly (SSTA) from CMIP6 decadal experiments for interannual rainfall forecasting over Australia. SSTA indices representative of large-scale modes of climate variability are transformed to enhance the prediction of rainfall at specific locations using Wavelet System Prediction (WASP) method. These spectrally transformed SSTA indices are then used as inputs into a hierarchical linear combination (HLC) model to forecast rainfall over interannual timescales. Results show that this hybrid HLC-WASP model significantly improves the predictability of interannual rainfall compared to the existing approach where no spectral transformation is used, up to lead times of five years ahead. This spectral transformation technique has the potential to enhance the quality of inputs for other rainfall forecasting models, particularly over regions with significant teleconnections between climate variability and regional rainfall. Improved rainfall forecasts have significant implications for water resource planning and management, enabling better drought mitigation, efficient water allocation, informed infrastructure planning, and more effective climate adaptation strategies.
DOI: 10.1016/j.jhydrol.2025.132870
Language: English
Open Access status: hybrid
Estimated APC paid: $4,030
OpenAlex ID: W4407367185
FWCI: 1.595
Cites: 58
Cited by: 6
Related to: 10