Energy flood: The overlooked surplus extreme in global wind-solar power systems

Dear Editor,

The ambitious goal of net-zero carbon emissions has driven an unprecedented expansion of global wind and solar power.1 Increasing reliance on these weather-dependent resources amplifies the variability in energy supply, exacerbating the supply-demand mismatches within power systems.2 To date, scientific and policy attention has focused disproportionately on the extreme energy deficit events, in which demand far exceeds supply, commonly termed “energy droughts.”3 However, the opposing extreme energy surplus events remain critically overlooked: periods of massive oversupply—energy floods.

In reality, the critical deficit in energy systems is often not in total electricity generation but in flexibility. Aggressive capacity expansion of wind and solar power aimed at mitigating energy droughts may paradoxically exacerbate the severity of energy floods, which pose systemic risks across three key dimensions. At the operational level, energy floods challenge electricity grid stability by overwhelming transmission lines and storage capacities, resulting in large-scale curtailment and resource waste.4 In 2024, for instance, wind and solar curtailment rates reached 10% in Chile and 8.5% in the United Kingdom,5 while in Tibet, the utilization rate for solar power was limited to merely 68.6%.6 At the market level, such massive oversupply frequently drives negative electricity prices, transforming generation from an asset into a liability.7 For example, the proportion of hours with negative electricity prices rose sharply across multiple regions from 2022 to 2024, including increases from 19% to 26% in South Australia, 1%–13% in Southern California, and <1%–8% in Finland.8 At the planning level, such revenue uncertainty creates a severe barrier to future investment in renewable projects,9 perversely stalling the capacity expansion required for carbon neutrality. While regional studies increasingly highlight the escalating risks posed by renewable energy surplus, a systematic global assessment of energy floods in wind-solar power systems remains lacking, particularly regarding their spatiotemporal regimes, dominant inputs, and mitigation strategies.

Here, we bridge this gap by developing a standardized framework to map global energy floods, utilizing the bias-adjusted ERA5 reanalysis (WFDE5, 1979–2019).10 We characterize distinct energy flood regimes across diverse regions and identify their dominant inputs. Furthermore, while spatial interconnection is widely recognized as an effective approach for managing wind and solar variability, we demonstrate its limitations in mitigating energy floods and outline region-specific planning implications. By shifting the analytical focus from scarcity to abundance, this study offers critical insights for reliable yet efficient decarbonized power systems.

Defining and mapping global energy floods

We develop a standardized framework to systematically identify global energy floods. Within each grid (1° × 1°), we assume a local energy system with a specific dependence on wind and solar power. This dependence is quantified as the proportion (here 50%) of days in which their aggregate supply meets or exceeds local demand. On the demand side, daily load is modeled as the sum of a constant baseline and temperature-sensitive components (i.e., heating/cooling loads during cold/hot days).2 On the supply side, daily wind and solar power output is derived from the capacity factor (CF) and installed capacity (IC). Specifically, CFs are calculated from surface wind speed, downwelling shortwave irradiance, and near-surface air temperature; ICs are then optimized to minimize surpluses while meeting the specified dependence level. Finally, the daily residual load (RL) series is calculated as the difference between energy demand and the combined wind and solar supply, where positive values indicate supply deficits while negative values imply surpluses (Figure 1A). An energy flood event is defined as one or more consecutive days during which RL falls below a specific threshold, set here as the mean of the negative RL series minus 1.5 standard deviations.