For an efficiently managed wind farm and wind power generation under adverseweather, knowledge of meteorological parameters influencing wind speed is ofcrucial importance for optimized and improved forecasts. We investigate temporaleffects of wind speed related processes such as wakes within the wind farmusing the Heterogeneous Graphical Granger model. The ERA5 meteorologicalreanalysis was used to generate wind farm power production data in Eastern Austria.We evaluated six different scenarios for the hydrological half-year period,based on moderate wind speed and varying temporal intervals of low or high extremewind speed This allows to carry out causal reasoning about possible causesof extreme wind speed in a wind farm. A set of causal parameters for each ofthe scenarios was discovered enabling future early warning and for taking managementmeasures for wind farm power generation management under adverseweather conditions.