Abstract:
The South African wind energy sector is developing rapidly with numerous wind energy facilities currently being commissioned. According to the Intergovernmental Panel on Climate Change (IPCC), some risks and opportunities for wind power generation as a result of climate change could be anticipated in future.
The objectives of this study were therefore to:
a. determine whether seasonal near-surface winds over South Africa, as generated by a Regional Climate Model (RCM) using boundary conditions supplied from coupled Global Circulation Models (GCMs), during a reference period of 1981 to 2005, are realistically represented;
b. establish whether differences exist between seasonal near-surface winds calculated for the reference period (1981-2005) versus a projected period of 2051 to 2075, incorporating two future Representative Concentration Pathways (RCP4.5 and RCP8.5);
c. determine the projected impact of climate change on wind power density.
Wind output from sophisticated atmospheric models (GCMs) provides valuable information on projected changes in wind patterns as a result of climate change. Through the CORDEX-Africa (COordinated Regional Downscaling EXperiment) project, the so-called RCA4 RCM has, by dynamically downscaling eight GCMs, produced a substantial collection of regional climate simulations. RCA4 RCM data were employed in this study to determine the impacts of climate change on South African winds and wind power resources.
Mean seasonal winds speeds were calculated for 1981 to 2005 for observed (ERA-Interim reanalysis) and RCA4 RCM output. The Root Mean Square Error (RMSE) between ERA-Interim and RCA4 RCM simulations was calculated. Wind speed frequencies were then simulated from ERA-Interim and RCA4 RCM data for each season and for different speed categories. RCA4 RCM data were also verified independently against weather station data. The RCA4 RCM was found to perform well, but a positive bias in the simulations of winds was detected.
Mean seasonal winds were calculated for the future period using RCA4 RCM output for the two pathways. Anomalies between RCA4 RCM output in the historical and future periods were then calculated and expressed as percentage changes in mean seasonal wind speeds. Wind speed frequencies of different categories were also simulated for the projected period under the two pathways. Anomalies between the historical and reference periods were also calculated for frequencies. Future projections indicate that parts of the country not typically considered as having substantial wind energy resources may become useful, such as north-eastern South Africa. As for the areas in which wind farms are currently being developed, mean wind speeds are projected to decrease by only 2% in two of the seasons, and to increase in the other two.
RCA4 RCM data were corrected for biases. Corrected mean wind speeds were then used as input to the calculation of wind power density in the projected period. Wind power density is projected to remain fairly similar in the future period as the historical period, as wind speeds have been projected to change by a maximum of 9%, which is a very small change when considered in terms of wind power density calculations.