Climate influence on compound solar and wind droughts in Australia

Australia’s rapid transition to renewables increases exposure of the electricity sector to weather and climate variability. The study targets the Australian Energy Market Operator’s (AEMO) National Electricity Market and its Renewable Energy Zones (REZs), asking how often and under what conditions widespread reductions in solar irradiance and wind speed (“droughts”) occur, whether solar and wind droughts co-occur, and how synoptic weather systems and large-scale climate modes (ENSO, IOD, SAM) modulate these risks. The purpose is to quantify grid-wide vulnerability, identify characteristic weather patterns associated with widespread events, and evaluate the predictive value of climate modes for risk assessment and seasonal planning.

Prior work shows renewable supply is sensitive to meteorological variability and that spatial diversification can reduce variability. Studies outside Australia link energy output to weather regimes, aiding system design and forecasting. In Australia, ENSO can reduce solar energy in La Niña-like conditions by up to ~10% in parts of the east and north; negative IOD phases reduce winter solar radiation; wind power is generally negatively correlated with ENSO across eastern and western Australia, suggesting La Niña may enhance wind but reduce solar. The SAM, reflecting storm-track latitude, may indicate wind variability. Despite global progress, systematic mapping of Australian grid-wide energy-relevant weather patterns and climate-mode influences has been limited, though climate modes are known to modulate concurrent regional climate anomalies, motivating a comprehensive assessment.

Study domain: 36 of AEMO’s 39 Renewable Energy Zones (REZs) with existing or planned solar and/or wind capacity (7 solar-only, 10 wind-only, 19 both). Data: ERA5 reanalysis at 0.25°; hourly variables aggregated to daily means defined over 24 h from 14:00 UTC (≈ local midnight); period 1959–2021. Variables: 100 m wind speed W100 = sqrt(u100^2 + v100^2); surface downward short-wave radiation flux for solar resource; 2 m temperature; mean sea-level pressure (MSLP); 500 hPa geopotential height; daytime mean fraction of total cloud cover (12 h from 20:00 UTC). SST: HadISST monthly fields. Anomalies computed relative to full-period day-of-year (daily) or calendar-month (monthly) climatologies. Drought definitions: A solar (wind) drought occurs when daily mean solar irradiance (100 m wind speed) is below the all-REZ, full-period 25th percentile threshold (solar: 133 W m−2; wind: 4.2 m s−1). Compound droughts occur when both solar and wind droughts co-occur on the same day in the same grid cell or REZ. Spatial assignment: grid cells belong to a REZ if their center lies within the REZ boundary; REZ-mean series computed by averaging member grid cells. Widespread-event metrics: (1) Daily simultaneity—count of REZs in drought each day for solar (max 26), wind (max 29), compound (max 19); widespread days are the top 5% by affected-REZ count, used for synoptic composites. (2) Seasonal severity—per-season counts of REZs in drought; top 10% seasons (seven seasons) used for large-scale SST–MSLP composites. Climate modes: ENSO via Niño3.4 (5°N–5°S, 120°–170°W) SST anomalies; IOD via Dipole Mode Index (west minus southeast Indian Ocean SST anomalies); SAM from MSLP anomalies (normalized) as 40°S minus 65°S. Diagnostics: composites of cloud fraction, near-surface winds, temperatures, MSLP, 500 hPa heights for widespread days by season; seasonal composites of SST and MSLP during top-10% seasons; phase-stratified differences in grid-cell drought frequencies (positive minus negative phase: index beyond ±1 SD).

• Drought prevalence and co-occurrence: On 50% of days, ≥15 REZs experience some drought (solar, wind, or compound); on ~20% of days, ≥28 REZs do. If compound droughts are counted as double, 10% of days feature >39 droughts across REZs, underscoring potential grid-wide vulnerability. • Seasonality: Compound solar–wind droughts occur most frequently in winter; at least five REZs simultaneously experience a compound drought on ~10% of winter days. In summer, 95% of days have ≤1 REZ in compound drought. • Spatial patterns: Solar drought risk follows the seasonal cycle (southern REZs higher risk; northern REZs lower), while wind drought risk is lower in western/southern regions (South Australia, Victoria) exposed to frequent fronts, and higher in the east; offshore wind REZs near Sydney show comparatively lower wind-drought risk than nearby onshore REZs. • Synoptic drivers: Widespread solar droughts are characterized by extensive positive cloud anomalies (up to ~55% above normal) tied to moist onshore easterlies and Tasman Sea anticyclones. Widespread wind droughts (winter, spring, autumn) show anticyclonic circulation centered over Victoria/NSW with elevated mid-tropospheric heights south of Tasmania; summer wind-drought circulation differs aloft. Compound droughts in winter/autumn resemble wind-drought circulation but with substantially greater cloudiness; in spring and summer, compound patterns differ, with summer showing twin surface highs southwest of Australia and east of New Zealand. • Temperature anomalies: During widespread solar and summer compound droughts, 2 m temperature anomalies suggest reduced heating/cooling demand (warmer in winter solar droughts due to cloud insulation; cooler in summer due to reduced insolation). • Large-scale climate: Widespread solar drought seasons exhibit La Niña-like SSTs in spring and summer; winter composites suggest SAM involvement (positive phase-like MSLP anomalies over Antarctica). Widespread wind droughts associate with El Niño-like SSTs except in winter (neutral ENSO); MSLP patterns indicate enhanced anticyclonicity/southerly storm-track shifts. Compound droughts show La Niña-like SSTs and SAM-like MSLP signatures in most seasons (except autumn), plus positive MSLP south of New Zealand. • Predictive value of indices: Despite composite patterns, mode indices during the most widespread seasons typically lie within ±1 SD of the mean, limiting their standalone predictive power for grid-wide droughts. Exceptions include strong negative Niño3.4 in several widespread summer solar-drought seasons and avoidance of winter compound droughts during El Niño or negative SAM. • Teleconnections and regionality: Phase differences (positive minus negative) in grid-cell drought frequencies reveal strong regional contrasts. Solar droughts increase during La Niña, negative IOD, and positive SAM across large areas of NSW and southern QLD (often >10 days per winter). Wind and compound drought teleconnections are more heterogeneous, with phase contrasts exceeding 10 days per season in parts of southern and southeastern Australia; some regions show opposing signals, implying potential for spatial balancing.

The study shows that widespread reductions in solar irradiance and wind speed can co-occur across multiple REZs, especially in winter, posing a non-negligible risk to a renewables-dominated grid. Synoptic analyses clarify that solar and wind droughts arise from distinct yet sometimes overlapping weather regimes, explaining when compound events emerge (wind-drought-like circulation plus enhanced cloudiness). Temperature anomalies during widespread events often mitigate energy demand, lessening net system stress. While composites implicate ENSO and SAM, the lack of consistently extreme mode-index values during the most widespread seasons reflects spatially diverse teleconnections across the large AEMO footprint. Consequently, climate modes alone are weak predictors of grid-wide droughts but remain informative at regional scales. The spatial variability of teleconnections highlights opportunities to increase resilience by strategically deploying assets in regions with opposing responses to modes and by leveraging interconnection (including potential links to southwest Australia) to offset regional deficits. The identified weather patterns and climate–drought relationships can support subseasonal-to-seasonal risk assessment via statistical bridging from predictable large-scale modes to energy-relevant variables.

This work systematically links compound solar and wind droughts in Australia to characteristic synoptic regimes and large-scale climate variability, quantifying their spatial extent, seasonality, and drivers across AEMO REZs over 1959–2021. It demonstrates frequent multi-REZ drought co-occurrence, wintertime dominance of compound events, and distinct synoptic fingerprints for solar versus wind droughts. Although ENSO and SAM imprint the large-scale climate during widespread seasons, mode indices alone generally lack predictive power for grid-wide events due to spatially opposing teleconnections. The findings underscore opportunities to minimize production variability through spatial diversification, targeted siting in regions with anticorrelated teleconnections, and enhanced grid connectivity. Future research should translate physical-variable droughts to energy metrics (e.g., capacity factors), incorporate sub-daily dynamics relevant to operational risks, integrate demand-side measures (e.g., degree days, population distribution), and develop predictive tools that bridge climate-mode forecasts to renewable resource outlooks.

The analysis uses physical climate proxies (solar irradiance, 100 m wind speed) rather than generation-based metrics (e.g., capacity factors), omits turbine cut-off and PV temperature effects, and applies a uniform all-REZ 25th-percentile threshold that may not reflect regional installed capacities or technology mixes (onshore/offshore). Daily averaging overlooks sub-daily extremes that can drive price shocks and reliability events. REZ capacity weightings and evolving infrastructure are not accounted for. These choices limit direct translation to operational impacts and may underrepresent short-term risks.