Climate change vulnerability of global hydropower generation

Abstract:

This paper explores the vulnerability of global hydropower generation to the variability in seasonal averages as well as changes in extreme conditions of precipitation, surface runoff, and temperature. A statistical model is used to estimate the elasticity of hydroelectricity generation to the historical variation (1962-2010) in precipitation or runoff, while controlling for potential confounding factors and temperature changes. The estimated elasticities, which informs about hydropower sensitivity to meteorological variations, are combined with changes in future exposure around 2050 in different warming scenarios as simulated by an ensemble of GCMs participating in the CMIP5 project (Taylor et al., 2012).

We use a panel regression model to estimate the parameters characterizing a reduced-form relationship between hydropower electricity generation at country level, a set of meteorological indicators, and number of other covariates that control for time-invariant country-specific heterogeneity (country effect), unspecified exogenous influences affecting all countries and units (time effects), and other confounding factors such the electricity generation mix.

The estimated model shows that total annual runoff has a significant impact on the annual generation from the small and medium-sized units, whereas large-sized units do not appear to be sensitive to the inter-annual variation in runoff. This finding is reasonably explained by the greater buffer effect of reservoir capacity, which sensibly increases the resilience of these plants to inter-annual runoff variability. In medium-sized units an increase in total runoff by 1% increases electricity generation by 0.028%. Small-sized units are more sensitivity to inter-annual variations in runoff, and the same change in total runoff (1%) increases electricity generation by 0.037%. Seasonal temperature has also a significant impact. A 1% increase in spring temperature reduces electricity generation by 1.63%, while a 1% increase in summer temperature reduces electricity generation by 1.58%. While an increased frequency of warm days has a positive coefficient. Including the SPI indicators reduces marginal effects of the inter-annual variation in total runoff from 0.028 to 0.022 for the medium units and from 0.037 to 0.031 for the small units.