Evaluating the Adequacy of Wind Turbine Setback Distances for Limiting Shadow Flicker Impacts

Wind energy is a crucial renewable energy source for reducing greenhouse gas emissions. In Europe, there is still significant onshore wind energy potential. At the same time, most countries use setback distance regulations for wind energy planning, and only a few have introduced shadow flicker (SF) rules. This research aims to examine whether the setback distance rule is sufficient to protect citizens affected by SF in reality. Below, for Denmark, we compare setback-distance regulations with the SF guideline.

For the methodology, we chose Denmark as the case study, grouped the existing 4775 turbines into 604 windfarms based on location, height, and commission date, calculated the SF-affected areas for these windfarms, and overlaid them with the population grid map to estimate the SF-affected population. Then we calculated the population covered by the setback-distance regulation-affected area around wind turbines and aggregated them by windfarms. In the end, comparisons are made between the two affected areas and populations. The model being used here to calculate SF impacts is the WIMBY_SF tool, an open-source SF simulation model written in Python that takes into consideration complex terrain and estimated turbine operation times.  The JRC-CENSUS population1 grid 2021 (JRC, 2024) with a 100 m x 100 m resolution is used to estimate the affected populations.

As a result, the total affected population due to physical impacts from>30 hours/year SF exposure is 16,514. At the same time, by regulation, the suggested distance from residences is four times the turbine tip height, resulting in a population of 16,334. The 30 hours/year is chosen because many EU countries have regulations or guidelines that follow the German guideline, which sets a shadow flicker limit of 30 hours per year for the astronomical maximum possible shadow duration (worst-case scenario). Despite similar affected population sizes, the areas affected by the two assessments vary considerably. The overlapping affected population from the two assessments is 8,939 (36.8% of the union-affected population), and the non-overlapping affected population from both methods is 24,278.

The Intersection over Union (IoU= Area(Model∪Reg)/Area(Model∩Reg)) shows spatial agreement from the setback distance and SF assessments. Across the studied 604 wind farms in Denmark, the IoU distribution is intensely concentrated between 0.50 and 0.70 (56.46%), indicating moderate spatial agreement between the modelling-based and regulation-based affected areas. Only 24 wind farms (3.97%) achieved high agreement (IoU ≥ 0.70), while a notable 85 wind farms (14.07%) exhibit near-zero overlap (IoU ≤ 0.05), implying mismatches in affected area alignment for certain farms.

Furthermore, 319 windfarms show that the physical SF affected population is smaller than the population that lives in the area defined by regulation, 258 windfarms with physical SF affected population larger than the regulation concerned, and only 27 windfarms show 0 affected population for both assessments.

In conclusion, the physically affected population and the regulation-affected population are of similar sizes but differ in geography. A similar analysis will be extended for further European countries