EROEI depends on resource class - depends on mean wind speed optimistic 18 pessitimistic 5 \cite{King_2018}
All wind resources \cite{Teske_2019} - nearing limits except  wind ans solar and geo and ocean
ds
solar demand for transitions \cite{Breyer_2017}
Storage critical \cite{Koskinen_2016}

Results

Methods

Conclusions

Since it is not mean wind speed that we are interested in but capacity factors I tried to find a way to model this. I found a couple of papers were good in doing this analysis, and one had a fairly simple approach which I followed to estimate the CF for German onshore and then compare it with historical values.
You can take a look at the attached modified spreadsheet. The worksheet is DEonCalc (the CF methodology is also presented in pictures). The sheet GermanData has the historical values from the ministry. I considered three types of turbines: optimized for low, medium and high speeds.
A few stray observations: 
1 the CF is difficult to calculate and highly dependent on primarily two parameters – the standard deviation observation for winds and the parameter n that characterizes the turbine (it is not efficiency!). 
2 In any case, the simulated weighted average CF is much higher for reasonable values (~35%) than the observed one (~15%) for 1995. I can only reach the same levels if I increase std dev to 15m/s which is clearly unreasonable for mean speeds of 6-9 m/s.
3 The CF is actually REDUCING the variability that wind speeds demonstrate (compare the two graphs which are chronologically ordered)
4. For onshore Germany, the total amount of installed capacity is missing about 7GW from the actual one in 2016. If we exclude the projects that we don’t have year data for – the difference is 10GW
5. The historical CF evolution in Germany is rather flat. It is influenced by the annual variability of the winds, the technology maturity, and the location depletion but it is nearly impossible to disentangle these effects