Reanalysis models are commonly used to model renewable energy for power systems of different scales.
Several studies have compared their data with measurements and
found different results regarding quality.
Since shares of renewables are growing in power system worldwide, reliable weather data is a
necessity to integrate ever larger fluctuating shares of wind and solar power.
In this work, we compare wind speeds from global and regional reanalyses with measurements and with each other for different regions such as Europe or Vietnam
and investigate their impact on a cost-optimised highly renewable power system.
We show the following:
i) differences between models are relatively large with RMSE of wind speeds at hub heights reaching 3 m/s and more
ii) the impact of the specific weather dataset used on power systems with high shares of renewables is tremendous
iii) many degrees of freedom in power systems are, largely due to computational limitations, not exploited in power system models of the present
iv) In some regions such as Vietnam the uncertainty from model differences makes systematic long term planning of the power system almost impossible
We conclude that further research is necessary to understand the sensitivity of power system models to input data, to validate this input data and, in consequence, to increase the results of power system models.