Risk to European birds from collisions with wind-energy facilities

This dataset contains outputs related to the following preprint:

 

 

We investigated the relative reported risk to European bird species from collisions with onshore turbines, for species that were known to fatally collide with wind-energy infrastructure. 

We employed a customised risk approach, based on two dimensions:

(1) the reported impacts, i.e., the total estimated number of fatalities at the species-level, accounting for current exposure to wind turbines (all turbines intersecting with the estimated suitable areas for each species).

(2) the likely vulnerability to these impacts (the degree to which species may be affected by collision mortality).

These two indices (impact index and vulnerability index) were rescaled between 0 and 1 across species (using a min-max rescaling), with 0 indicating lowest relative impacts or vulnerability and 1 indicating highest relative impacts or vulnerability.

 

We assumed that species that were both relatively more impacted and relatively more vulnerable would be at higher risk. Overlapping vulnerability with estimated impacts, we classified species into 4 different risk categories, and we assessed where species falling into the risk categories might occur, deriving pseudo-species richness maps for each risk category. We used two thresholds to define values of the impact index and of the vulnerability index above which species were considered to be at higher risk: the 75% and 50% quantiles of the distributions of impact index and vulnerability index. We classified species into the different risk categories for both thresholds.

 

In more details:

 

We used a quantitative synthesis of fatality numbers at wind-energy facilities to quantify collision-mortality rates at the species-level. We fitted a statistical model to predict collision rate (per year-turbine) as a function of species identity and turbine capacity. 

We derived (1) - the impact index - by combining this model with information on species suitable areas and on current wind-turbine deployment (location of turbines and turbine capacities). From the model, we then estimated average fatality numbers for all turbines intersecting the areas of suitable habitat for each species. Summing these numbers at the species level, we obtained the avaregee number of fatalities per year across each species suitable areas.

 

We estimated species' vulnerability from ecological characteristics (generation length, clutch size, and estimates of European suitable area) assumed to reflect species' ability to cope with disturbances.

 

##################################################################################################

 

The dataset contains a .csv with the species-level outputs, which also contains the classification into the different risk categories, for the two thresholds used in the study (75% and 50% quantiles). 

The dataset provides pseudo-species richness maps for each risk category and for the two threshold values (gridded raster tif files).

 

##################################################################################################

 

The "Species_risk_categories.csv" file notably contains:

- the cumulative collision-mortality estimations for 108 European bird species (that is, the average estimated number of fatalities for each species across their modelled suitable habitats, per year). This is calculated by intersecting each species modelled suitable habitat with the locations of the current wind turbines in Europe, and applying a statistical model that estimates the number of fatalities for each turbine as a function of species identity and turbine capacity. Then, the estimations at each wind turbine location are summed within species, across their suitable habitat (the sum is weighted by the relative habitat suitability of each grid cell where a species might occur).

- the species estimated vulnerability. This is estimated from additional characteristics: the littler or clutch size of the species, their area of suitable habitat, and their generation length.

Column descriptors: 

Order: species taxonomic Order

 

Family: species taxonomic Family

 

Scientific_name: species binomial scientific name

 

Suitable_Area: estimated area of suitability for each species (summed area across grid cells, weighted by the habitat suitability of each grid cell)Litter_clutch_size: specie-level average for number of offspring

 

Clutch_size: species-level average of number of offspring

 

Generation_length_d_BirdLife: species-level average of generation length

 

TotalTurbines_N: total number of turbine intersecting suitable grid cells for each species

 

Cumulative_Turbine_Capacity: total turbine capacity across all turbines intersecting suitable grid cells for each species

 

Impact_index_rescaled: Cumulative number of fatalities for each species (within suitable habitats, given current wind turbines intersecting with each species range), rescaled between 0 and 1 (using a min-max normalisation). 

 

Vulnerability_rescaled: vulnerability of the species, calculated from geographical rarity (using suitable area as a proxy), generation length, and litter/clutch size; rescaled between 0 and 1 across species (using a min-max normalisation).

 

Cumulative_mortality_rescaled: Total number of estimated fatalities at the species level. This accounted for all turbines currently intersecting the suitable habitats of a species, as well as turbine capacities. Summed estimations of average number of fatalities at each turbine location, weighted by the habitat suitability of the grid cells where the turbines are located.

 

Risk_category_0.75: risk category of the species, calculated using the 75% quantiles of the distributions of vulnerability and of cumulative mortality impacts as thresholds. 
One of: "high_risk"; "high_latent_risk"; "possible_adapter"; "possible_persister"

 

Risk_category_0.5: risk category of the species, calculated using the 50% quantiles of the distributions of vulnerability and of cumulative mortality impacts as thresholds. 
One of: "high_risk"; "high_latent_risk"; "possible_adapter"; "possible_persister"