This repository contains the data processing and modelling scripts used to quantify how projected renewable energy growth (hydropower, wind, solar and associated grid infrastructure) translates into land pressure on Norwegian biodiversity. The core idea is to link technology-specific land transformation (direct footprint and surrounding disturbance/rights-of-way) to species loss risk using habitat affiliations and a countryside species-area relationship (cSAR), and then explore how impacts accumulate under alternative energy-system scenarios and key modelling uncertainties.
Renewable energy growth amplifies land pressure on Norwegian biodiversity
DOI: https://doi.org/10.1016/j.cles.2026.100238
The workflow combines species occurrence and habitat information with spatial infrastructure data to estimate biodiversity impacts from land conversion. Species data are derived from the Norwegian Red List (2021) and filtered to include assessed categories for Norway. Species are associated with counties and then linked to landscape subregions through polygon intersection. Habitat information is parsed and expanded so each species can be connected to one or more main habitats, enabling habitat-specific modelling. To focus the terrestrial land-pressure component, strictly marine and strictly freshwater species are excluded from the main terrestrial impact calculations.
Infrastructure layers for wind, hydropower, solar, reservoirs, and grid components are processed as spatial features and intersected with land-cover data within each landscape subregion. The analysis estimates affected area by land-cover class, including both direct footprints and surrounding influence zones represented through buffer distances (evaluated across multiple radii). For grid infrastructure, both rights-of-way and pylon footprints are included.
Land-cover classes are mapped to habitat categories using a conversion table, and habitat “quality” is represented with affinity values (h) that describe how suitable each habitat/land-cover type is for each focal taxonomic group (Amphibians & reptiles, Birds, Vascular plants, Mammals). These affinities are then used in a countryside species-area relationship framework to translate effective habitat area change into biodiversity impact expressed as PDF (Potentially Disappeared Fraction of species). Species-group specific z-values (region-dependent) control how strongly persistence responds to area change, and alternative z assumptions are explicitly explored.
Uncertainty is propagated through Monte Carlo sampling of key parameters (e.g., footprint intensity per capacity, distributions used to fill missing values such as reservoir inundation fractions). Each iteration generates a complete set of technology-specific habitat losses and resulting PDFs.
In different scenarios we then construct alternative expansion pathways by combining existing (operational) power plants with sampled planned power plants and assigning commissioning dates within scenario-specific time windows. Power plants are added until scenario electricity-generation targets are met for each technology at milestone years (2030, 2040, 2050). Cumulative PDFs are tracked through time, producing scenario envelopes (mean and uncertainty intervals) across many runs. Additional sensitivity analyses quantify how results change with buffer distance and z-values.
Estimated biodiversity impacts for six different energy production scenarios at milestones 2030, 2040 and 2050.
data/– prepared spatial layers, conversion tables, species tables and supporting datasets\functions/– helper functions used across scripts (e.g., affected land cover, loss accounting, cSAR calculations)\results/– intermediate outputs and exported tables/figures\scripts/– code used for preparing and analyzing the data
- Results quantify land pressure on biodiversity expressed as potentially disappeared fraction of species (PDF) using a countryside SAR approach.
- Scenario pathways are constructed by sampling from available planned assets and are intended to be representative of trajectories consistent with scenario-level generation targets.