{"paper_id":"09d7ddc2-aa2b-4ea4-8e19-2c8a223cd48f","body_text":"Assessing the impact of renewable energy\ninstallations on biodiversity and identifying\nsustainable trade-o!s\nMarie-Ange Dahito 1,2*, David Yang Shu 3, Gabriel Wiest 3,\nStefano Moret 3,5, Tobias Wechsler 1,4, Lo ¨ ıc Pellissier1,2\n1Swiss Federal Institute for Forest, Snow and Landscape Research\n(WSL), Z¨ urcherstrasse 111, 8903, Birmensdorf, Switzerland.\n2Department of Environmental Systems Science, ETH Z¨ urich,\nUniversit¨ atstrasse 16, 8092, Zurich, Switzerland.\n3Department of Mechanical and Process Engineering, ETH Z¨ urich,\nTannenstrasse 3, 8092, Zurich, Switzerland.\n4Research Unit RECOVER, INRAE Aix-Marseille University, 3275\nRoute C´ ezanne, 13100, Aix-en-Provence, France.\n5Department of Management, Economics and Industrial Engineering,\nPolitecnico di Milano, Via Lambuschini 4/b, 20156, Milan, Italy.\n*Corresponding author. E-mail: marie-ange.dahito@polymtl.ca.\nAbstract\nRenewable energy is crucial to achieve climate neutrality, but its rapid expansion\ncan threaten biodiversity through habitat loss or fragmentation and ecological\ndisruption. We present a spatially explicit assessment framework that quanti-\nfies biodiversity impacts from land use change associated with renewable energy\ninfrastructure across a broad range of species groups, and identifies siting configu-\nrations that balance energy provision and conservation goals. Drawing on metrics\nfrom life cycle assessment, combined with species distribution models and sit-\ning strategies, we evaluate alternative deployment strategies. Using Switzerland\nas a case study, we compare three siting strategies (maximizing energy output,\nminimizing biodiversity impact, and a trade-o! approach) for photovoltaic sys-\ntems, run-of-river hydropower, and wind turbines. For solar and hydropower\n1\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\ninstallations, prioritizing energy e”ciency yields the highest cumulative biodi-\nversity losses. However, these impacts can be substantially reduced with only a\nslight increase in land use by favouring biodiversity protection. For wind installa-\ntions, strict avoidance of sensitive ecosystems may increase total impacts, as less\ne”cient and therefore additional sites are required to achieve the same annual\nenergy yield. Overall, our results show that trade-o!-based siting strategies can\ne!ectively balance performance and biodiversity protection, highlighting that\nrenewable energy can be provided without sacrificing sensitive ecosystems.\nKeywords: biodiversity, renewable energy, hydropower, wind power, photovoltaic,\nenergy systems planning, characterization factors\n1 Introduction\nClimate change and biodiversity loss are interconnected environmental challenges.\nClimate change disrupts species distributions, alters ecosystem dynamics, and erodes\necosystem services [IPBES et al., 2019, IPCC, 2023]. The resulting loss of biodiversity\nreduces ecosystem resilience and weakens forests, wetlands, and other natural carbon\nsinks, exacerbating climate change. Thus, both crises must be addressed at once.\nWhile habitat destruction, pollution, and over-exploitation of resources have his-\ntorically been the primary drivers of biodiversity erosion [Millennium Ecosystem\nAssessment, 2005], recent increase in the rate of climate change and the frequency of\nextreme events create additional pressures on biodiversity that will rival habitat loss\nin the second half of the century [Newbold , 2018]. Of nearly 170,000 species globally\nassessed by the International Union for Conservation of Nature (IUCN) Red List,\nover a quarter face extinction [IUCN, 2025] and more than 10,000 are under direct\nthreat from climate change [IUCN, 2019]. Hence, greenhouse gas emissions must be cut\nrapidly as the main driver of climate change, including an energy transition based on\nrenewable technologies such as hydro, solar, and wind power [ IPCC, 2023, Bogdanov\net al., 2019].\n2\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nHowever, while renewable technologies can dramatically reduce the impacts of\nclimate change, they may harm biodiversity if poorly sited [Katzner et al., 2013, Gas-\nparatos et al., 2017]. Hydropower installations fragment river systems [Kuriqi et al.,\n2021, He et al., 2024] and alter flow regimes [Wechsler et al., 2023, Brunner and\nNaveau, 2023, B¨ atz et al.,2025] as well as the transport of sediments and organic mat-\nter [ Arroita et al., 2015, Yu et al., 2019]; ground-mounted solar panels may compete\nwith natural or agricultural areas [Hernandez et al., 2015, Zhang et al., 2024]; and\nwind energy installations a!ect flying species [Thaxter et al., 2017, Tolvanen et al.,\n2023]. Therefore, careful spatial planning based on ecological assessments is critical\nto ensuring a sustainable and biodiversity-friendly energy transition.\nMeasures have been taken to mitigate biodiversity loss, as is the case in the Euro-\npean Union with the adoption of the EU Birds and Habitats Directives. These two\nkey laws led to the creation of Natura 2000, the largest network of protected areas\nworldwide [Evans, 2012]. Nonetheless, many ecologically valuable areas remain outside\nprotected zones, due to incomplete national inventories, weak ecological connectivity,\nor inadequate management [Dubos et al., 2022, Castillo et al., 2020, Oberosler et al.,\n2020].\nSiting decisions can be supported by spatial conservation planning tools such as\nMarxan [Ball and Possingham, 2000, Ball et al., 2009], Zonation [Moilanen et al.,\n2005], and ConsNet [Ciarleglio et al., 2009], which are designed to identify priority\nconservation areas. These tools rely on biodiversity data such as species distribution\nmodels (SDMs), habitat suitability indices, or range maps [Moilanen et al., 2009,\nPressey et al., 2007]. However, even when socio-economic layers are included [Salak\net al., 2024], these tools are not designed to directly provide implementable siting deci-\nsions for infrastructure. Their outputs represent relative spatial priorities, typically\nat the scale of planning units, and therefore do not incorporate fine-scale constraints\nrequired for infrastructure deployment. Besides, these methods do not quantify the\n3\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nspecific biodiversity impacts of proposed infrastructure projects. Thus, comparisons\nbetween di!erent siting options are not possible for supporting transparent decision\nmaking [McIntosh et al., 2018].\nLife cycle assessment (LCA) is a methodology commonly used to quantify environ-\nmental impacts of human activities over the full life cycle. LCA methods specific to\nbiodiversity impacts have been developed [Damiani et al., 2023], and can be used to\nsupport planning decisions. However, they commonly have low spatial resolution due\nto data limitations [De Baan et al., 2013, Verones et al., 2013, May et al., 2020], their\nunderlying datasets tend to be biased toward vertebrates [Damiani et al., 2023], and\nthey cannot be applied for local infrastructure siting decisions. Chaudhary et al. [2015]\nand Scherer et al. [2023] studied the biodiversity impacts of di!erent land-use types\nand intensities on amphibians, birds, mammals, reptiles, and plants, with the latter\nstudy also considering land fragmentation. They both consider species vulnerability\non a global level.\nSpecialized methodologies applied in LCA with higher spatial resolution are com-\nmonly limited to specific geographic regions (e.g., [Zelm et al., 2011, Geyer et al.,\n2010, de Baan et al., 2015, May et al., 2021]). Few studies quantified the biodiversity\nimpacts of renewable energy installations on limited taxonomic groups. For example,\nMay et al. [2021] investigated the e!ects of 39 existing onshore WT on the species\nrichness of 13 bird species groups in Norway, and Dorber et al. [2020] evaluate the\nprojected biodiversity impact per kWh for 1933 possible future reservoirs worldwide,\nconsidering five species groups, with fish being the only representative for aquatic\nbiodiversity.\nNeither of these studies explores siting strategies or proposes trade-o! site selection\nframeworks. Furthermore, they consider only a limited number of taxonomic groups\nand rely mostly on low-resolution data, such as species–area relationship (SAR) mod-\nels derived from global-scale studies, rather than on locally calibrated biodiversity\n4\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\ninformation. To our knowledge, no existing work jointly evaluates land-use change\nand biodiversity impacts of renewable infrastructure expansion across a wide range\nof species groups, while providing a framework to identify siting configurations that\nbalance electricity generation and conservation priorities.\nHere, we address these limitations by developing a spatially explicit renewable\nenergy siting framework that integrates LCA metrics with fine-resolution biodiversity\nmodelling to evaluate biodiversity impacts associated with land use change across a\nwide range of taxonomic groups. We apply this framework to investigate the trade-o!s\nbetween maximizing energy conversion and preserving biodiversity, considering three\nsiting strategies. We demonstrate its capabilities by assessing the e!ects of the siting\nstrategies for renewable installations on biodiversity impacts in Switzerland. In this\ncase study, we consider the e!ects of projected expansions of photovoltaic (PV) panels,\nwind turbines (WT), and run-of-river (ROR) hydroelectric plants in 2050. We use\nhigh-resolution locally calibrated datasets, such as detailed species distribution and\nrange data for 20 taxonomic groups, SARs tailored by species group and calculated at\nbioregion [FOEN, 2020] and country levels, region-specific extinction probabilities, and\nrenewable energy potential maps. Thus, we account for both local and system-wide\nbiodiversity impacts, as well as local renewable potentials and system-wide electricity\ndemands. Our assessments consider installation-specific impacted species groups and\nextinction rates based on an extensive literature review (see Table 3), ensuring that\nsiting decisions reflect technology-specific ecological pressures.\n2 Methods\nThe proposed infrastructure siting framework integrates local biodiversity impact\nassessment through a spatially explicit data-driven approach. It requires spatial data\non the state of biodiversity in the geography under investigation, spatial estimates of\n5\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nthe expected energy conversion potential of renewable technologies, and the identifica-\ntion of exclusion areas where installation is undesirable or not possible. Based on these\ninputs, the framework quantifies local, regional, and global biodiversity impacts asso-\nciated with land use change due to infrastructure deployment. Within the framework,\nalternative siting strategies can be developed and systematically assessed to identify\noptions that balance energy provision potential with biodiversity conservation.\nThis framework is broadly applicable for assessing the biodiversity impacts\nof infrastructure siting in any geographic region, provided that suitable data are\navailable.\n2.1 Metrics for biodiversity impact due to land use change\nTo quantify the biodiversity impacts of land use change associated with infrastruc-\nture installation, we combine two metrics that are regional extinction probabilities\n(REPs)[Adde et al., 2025a] and characterization factors (CFs). The resulting metric\nmeasuring the overall biodiversity impact at the scale of the entire study region is\nhereafter referred to as the global CF.\nThe REP is based on the global extinction probability (GEP)[Kuipers et al., 2019],\nwhich assesses the risk of global extinction for a species group resulting from local\nextirpations and is computed for each region and species group based on species range\nsizes, threat levels, and spatial distributions.\nAdde et al. [2025a] introduced the REP to preserve the high resolution of regional\ndatasets and assess extirpation risks more accurately. The REP adapts the GEP to\nthe spatial scale of the studied area.\n6\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nFor a species group G, the REP, whose value lies between 0 and 1, is computed\nper pixel as:\nREP (p)\nG =\n∑\ns→G\nω(p)\ns\n∑\np→→P\nω(p→)\ns\n· εs\n∑\ns→G\nεs\n, →p ↑P , (1)\nwhere ω(p)\ns ↑ [0, 1] is the occurrence value of species s at pixel p based on species\ndistributions (see Table 1), P is the set of all pixels of the considered region, and\nεs ↓ 0 is the weighting factor representing the threat level of species s (see Table 2).\nCFs are used in LCA to score environmental damage, allowing di!erent pressures\non ecosystems to be consistently assessed and compared. To translate biodiversity\nimpacts associated with new infrastructure installations into measurable values, we\nadapt the methodology proposed by Chaudhary et al. [2015] and compute CFs using\nthe pre-existing land use as the reference condition, instead of a natural reference\nstate. This choice enables a specific focus on land-use change impacts (rather than\nland-use). We compute the CFs for spatial units defined as sets of pixels within the\nstudy region (e.g. a single pixel or all pixels within an installation area).\nThe loss of reference area due to land use change at spatial unit u ↔P is:\nA\n(u)\nlost = A(u)\nref\n↗ A(u)\nnew, (2)\nwhere A(u)\nref is the area under the reference land-use state and A(u)\nnew the area that\nremains unchanged after potential land-use modification. The regional loss in species\nrichness due to new cumulative land use at spatial unit u is:\nS(u)\nlost,G = S(u)\nref,G ↗ S(u)\nnew,G , (3)\nwhere S(u)\nref,G and S(u)\nnew,G are the species richness of G at the reference state and\nat the new state, respectively. S(u)\nref,G is derived from the species distribution data\n7\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n(see Section 2.3 and Supporting Information), while S(u)\nnew,G is obtained using the\ninstallation-specific extinction rate of G, which is represented by the local CF, denoted\nCF (u)\nloc,G , that is,\nS(u)\nnew,G =( 1 ↗ CF (u)\nloc,G ) · S(u)\nref,G . (4)\nThe derivative of the loss in species richness [Chaudhary et al., 2015] can be\nexpressed as:\nϑS (u)\nlost,G\nϑA(u)\nlost,G\n=\nS(u)\nref,G\nA(u)\nref\n· z ·\n(\n(A(u)\nref\n↗ A(u)\nnew)+h (u)\nG · A(u)\nnew\nA(u)\nref\n) z↑1\n, (5)\nwhere h(u)\nG is the a”nity of G to the new land-use type, calculated as:\nh(u)\nG\n=\n(\nS(u)\nnew,G\nS(u)\nref,G\n) 1/z\n, (6)\nand z is the power parameter of the classic SAR of G.\nThe regional damages resulting from land occupation and transformation are\nexpressed via a regional CF as the equivalent species loss throughout the operating\nlifetime of the installation (species-eq lost·years). The regional CFs for land occupa-\ntion CF\n(u)\nreg,occ,G and land transformation CF (u)\nreg,tra,G at spatial unit u ↔P can be\nwritten as follows:\nCF (u)\nreg,occ,G =\nϑS (u)\nlost,G\nϑA(u)\nlost,G\n· tocc · A(u)\nlost\n, (7)\nand\nCF (u)\nreg,tra,G =0 .5 ·\nϑS (u)\nlost,G\nϑA(u)\nlost,G\n· t(u)\nre,G · A(u)\nlost, (8)\nwhere tocc is the duration for which the land is occupied by the installation, and t(u)\nre,G\nis the regeneration time needed to recover the biodiversity of the species group G once\nall installations are removed and the land is not managed.\n8\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nThe overall regional impact CF (u)\nreg,G on group G at spatial unit u is:\nCF (u)\nreg,G = CF (u)\nreg,occ,G + CF (u)\nreg,tra,G . (9)\nAt the scale of the study region P, we characterize the global impact as the regional\nimpact adjusted to reflect regional species vulnerabilities. The global impact on species\ngroup G at spatial unit u can be written as follows:\nCF (u)\nglo,G = CF (u)\nreg,G · REP (u)\nG . (10)\nNotably, the probability of extinction at a spatial unit u is simply the sum of the\nextinction probabilities of its constituent pixels, that is,\nREP (u)\nG =\n∑\np→u\nREP (p)\nG . (11)\nMultiplying the regional CF by the REP yields the global CF, expressed in species-\neq lost·years, which accounts for the species ranges and vulnerabilities.\n2.2 Approaches for biodiversity impact calculation\nWe use two complementary approaches to calculate biodiversity impacts (i.e., global\nCFs Equation (10)) from infrastructure deployment: a fine-scale pixel-level assessment\nand a global infrastructure-level assessment.\nIn the fine-scale approach, the equations described in Section 2.1 are applied at\nthe pixel scale, that is, u ↘ p in Equations (2)t o( 11), with the parameter z specific to\nthe bioregion in which each pixel is located (see SAR model calculation described in\nSection 2.3 and the Supporting Information). Therefore, CFs are calculated for every\n9\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\npixel. This method captures fine-scale spatial variation, allowing for a detailed under-\nstanding of how impacts di!er across regions. It is particularly useful for identifying\nbiodiversity hotspots and areas most a!ected by development (see the Supporting\nInformation). However, to obtain a total global impact value, the impacts from all\nindividual pixels of the study region must be aggregated. The choice of aggregation\nmethod does not alter the ranking of stressors, but a!ects the resulting values [ Verones\net al., 2015], which may influence the interpretation of the results and introduce\nadditional complexity.\nIn contrast, to avoid aggregation e!ects, the global-level impact calculation treats\nthe entire impacted zone as a single region of interest. In this case, the spatial unit\nu considered in Equations (2)t o( 11) corresponds to the set of all pixels impacted\nby new installations, i.e. the pixels containing the infrastructure together with their\nassociated impact areas (see Supporting Information).\nWhile it still incorporates local biodiversity information, such as the spatial dis-\ntribution of selected pixels and combined SDM maps, this approach directly yields a\nsingle global CF which captures the overall e!ect of the infrastructure on the consid-\nered species group. Here, the z parameter is taken as the national scale estimate, and\na single regeneration time is used, computed as the mean regeneration time across all\nimpacted pixels. Finally, the REP is simply the sum of the REPs of the a!ected pixels.\nThe global-level approach does not provide insights into the spatial distribution\nof impacts, but avoids the need for aggregating pixel-scale values. As such, it is well\nsuited for comparing overall biodiversity outcomes across di!erent infrastructure sit-\ning scenarios. Therefore, we use the global-level impact calculation in Section 3 to\ngenerate biodiversity impact curves as a function of national energy output in our\ncase study. Analyzing the impacts of renewable energy installations at a pixel scale\nenables to identify areas most vulnerable to infrastructure installation. The results\n10\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nfor the pixel-level approach and additional Switzerland-specific analyses are provided\nin the Supporting Information.\n2.3 Case study\nWe apply our framework to Switzerland, assessing the biodiversity impacts of siting\nfour renewable energy technologies to meet installation targets required in the base\nscenario of the Energy Perspectives 2050+ [Kemmler et al., 2021] for a net-zero energy\nsystem in Switzerland: rooftop and ground-mounted PV systems, WT, and small ROR\nhydropower plants.\nCalculation of the extinction probabilities\nFor each species group G, we compute the REP values for 25m ≃ 25m-resolution pixels,\nconsidering only species ranges inside Switzerland.\nWe start from presence probabilities 0 ⇐ v ⇐ 100 (see Supporting Information)\nthat are given for each pixel p and species s, obtained from species habitat suitability\nmaps. A species is assumed to be possibly extant only when its presence probability\nin the pixel exceeds a presence threshold ϖ\ns (see Supporting Information). To make\nthe probabilities comparable across species, we follow the approach of Adde et al.\n[2025a] and scale the probabilities so that a value of 40 represents the lower bound\nfor a species to be considered “possibly extant”, regardless of its original threshold:\n¯v =\n\n\n\n40 · v\nϖs\nif 0 ⇐ v ⇐ ϖs\n40 + (100 ↗ 40) · v ↗ ϖs\n100 ↗ ϖs\nif ϖs <v ⇐ 100.\nThe species occurrence values ω(p)\ns are obtained using the weighting scheme of\nKuipers et al. [2019], which is based on Montesino Pouzols et al. [2014]. Table 1 shows\nthe mapping of the scaled presence probabilities ¯ v to the species occurrence values\nω(p)\ns .\n11\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nT able 1: Weighting scheme for species\noccurrence ω(p)\ns .\nCondition Presence ω(p)\ns\n80 < ¯v → 100 extant 1\n60 < ¯v → 80 probably extant 0.5\n40 < ¯v → 60 possibly extant 0.5\n20 < ¯v → 40 possibly extinct 0.1\n10 < ¯v → 20 presence uncertain 0\n0 → ¯v → 10 extinct 0\nKuipers et al. [2019] compare three threat-level quantification schemes (linear,\ncategorical, and logarithmic). Of the three schemes, we chose the linear categorization\napproach. Using Equation (1), we calculate the REP values.\nT able 2: Linear weighting scheme associated to\nthe IUCN threat level of each species s.\nIUCN conservation status εs\nextinct, extinct in the wild, or regionally extinct 0\ncritically endangered 1\nendangered 0.8\nvulnerable 0.6\nlower risk or near threatened 0.4\nleast concern, data deficient or not evaluated 0.2\nCalculation of the regional characterization factors\nIn line with the REPs, we calculate CFs at a spatial resolution of 25m ≃ 25m.\nWe use the most recent land cover data (see Supporting Information) to define\nthe reference state prior to the installation of new energy infrastructure. The a”nity\nof each species group G (Equation (6)) is assumed to be equal to 0 if no species of G\nis present at the reference state.\nWe assume the land occupation for the full lifetimes of the installations. We set\nthe occupation times to 20 years for WT, 30 years for both rooftop and ground-\nmounted PV panels, and 80 years for ROR hydropower plants, based on average\nlifetimes reported by the life-cycle inventory database ecoinvent 3.11. These values\n12\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nare based on data for 2 MW onshore WT (at global scale), 3 kWp multi-Si PV flat-\nroof installation in Switzerland, 570 kWp multi-Si PV plant on open ground (at global\nscale), and ROR hydropower plants in Switzerland.\nThe regeneration times of species groups are based on [Curran et al., 2014]w h e r e\naverage recovery times were computed for di!erent families of species under passive\nand active restoration, distinguishing forest and non-forest biomes. We specifically use\nthe predicted average recovery times relative to species richness for passive restoration\nin palearctic realm of latitude 45°. For the recovery of fish species, for which no data\nwas available, we assume a recovery time of one year as these species can rapidly\nrecolonize an aquatic area once an installation is removed. Details on the species\ngroups considered are provided in the Supporting Information.\nThe parameter z (Equations (5) and (6)) was estimated by fitting SAR models\nto species distribution data using the R package sars [Matthews et al., 2019], with\nspecies data aggregated at the catchment level. Models were built for each bioregion\nand at the national scale. Details on the SAR models are provided in the Supporting\nInformation.\nData collection\nThe case study assesses biodiversity impacts of renewable energy installations across\nSwitzerland using national-scale geospatial datasets at a 25 m ≃25 m resolution,\ncovering both terrestrial and aquatic ecosystems.\nBiodiversity patterns are represented using fine-resolution SDM maps derived from\nSDMapCH [Adde et al., 2025c,b] integrated with a range mapping algorithm [Fopp ,\n2025] based on validated occurrence records (1980 – 2021) [D« epraz et al., 2025]. These\ndata are used to generate binary presence–absence maps and species richness layers\nfor the considered taxonomic groups (see Supporting Information).\nSpecies groups potentially a!ected by land use change were identified for ground-\nmounted PV systems, WT, and small ROR hydropower based on the literature (see\n13\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nTable 3). Where available, published estimates of species richness reduction were used\nas local CFs; otherwise, assumptions were informed by expert judgment. Rooftop PV\nsystems were assumed to have negligible biodiversity impacts, as they rely on existing\ninfrastructure [Katzner et al., 2013].\n14\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nCommon\ngroup\nTaxonomic\nclass PVG WT ROR\nPlants\nAngiospermae\n0.6\n[Armstrong et al. , 2016]\n[Moscatelli et al. , 2022]\n[Tanner et al. , 2021]\n0.5 [Urziceanu et al. , 2021]\n[Zhao et al. , 2025] 0.4 [Bejarano et al. , 2020]\n[Anderson et al. , 2015]\nGnetophyta\nLycopodiophyta\nPteridophyta\nTrees Pinophyta\nFishes Actinopterygii 0.05 [European Commission et al. , 2020] - 0.3 [Benejam et al. , 2016]\n[Kuriqi et al. , 2021]Cephalaspidomorphi\nArthropods\nEPT (Ephemeroptera,\nPlecoptera, Trichoptera) 0.3\n[Horv´ ath et al., 2010]\n[European Commission et al. , 2020]\n[G´ omez-Catas´ us et al., 2024]\n0.2\n[Long et al. , 2011]\n[Voigt, 2021]\n[Weschler and Tronstad , 2024]\n0.2 [Anderson et al. , 2015]\n[Wang et al. , 2016]Odonata\nInsecta, except\nEPT and Odonata\nArachnida\n0.3 [European Commission et al. , 2020]\n[G´ omez-Catas´ us et al., 2024] 0.2 [Weschler and Tronstad , 2024] 0.2 [Anderson et al. , 2015]\n[Wang et al. , 2016]Malacostraca\nDiplopoda -\nMolluscs Bivalvia -- 0.2 [Anderson et al. , 2015]\n[Wang et al. , 2016]Gastropoda\nBirds A ves 0.25\n[European Commission et al. , 2020]\n[Kosciuch et al. , 2020]\n[Visser et al. , 2019]\n[Haga et al. , 2020]\n[G´ omez-Catas´ us et al., 2024]\n0.42\n[Tolvanen et al. , 2023]\n[Perold et al. , 2020]\n[Marques et al. , 2014]\n[Drewitt and Langston , 2006]\n-\nBats Chiroptera 0.25\n[European Commission et al. , 2020]\n[Montag et al. , 2016]\n[Szabadi et al. , 2023]\n[G´ omez-Catas´ us et al., 2024]\n0.42\n[Tolvanen et al. , 2023]\n[Long et al. , 2010]\n[Voigt et al. , 2022]\n[Kunz et al. , 2007]\n-\nMammals,\nexcept bats\nMammalia,\nexcept chiroptera 0.2 [European Commission et al. , 2020]\n[Lafitte et al. , 2023] 0.05 [Tolvanen et al. , 2023]\n[!L o p u c k i e t a l ., 2017] -\nReptiles Reptilia 0.3\n[European Commission et al. , 2020]\n[G´ omez-Catas´ us et al., 2024]\n[Lafitte et al. , 2023]\n-0 .25 [Crnobrnja-Isailovi´ c et al., 2021]\n[B´ arcenas-Garc ´ ıa et al., 2022]\nAmphibians Amphibia 0.3 [European Commission et al. , 2020]\n[Lafitte et al. , 2023] -0 .25\n[Crnobrnja-Isailovi´ c et al., 2021]\n[Jiang et al. , 2022]\n[Guzy et al. , 2018]\nT able 3: Species groups and associated extinction rates for ground-mounted photovoltaic, wind, and run-of-river power plants.\nAbbreviations: Ephemeroptera, Plecoptera, Trichoptera (EPT); ground-mounted photovoltaic (PV G); wind turbines (WT); run-of-river\nhydropower (ROR).\n15\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nImpacted areas were defined according to technology type and species mobility. For\nhydropower, impacted river sections were represented as polygons extended laterally\nto account for both aquatic and riparian e!ects. For PV and WT, minimum impacted\nareas correspond to ground coverage, with additional bu !ers applied for mobile species\nto reflect disturbance e!ects during operation. Bu !er distances and their ecological\njustification are provided in the Supporting Information.\nProduction potentials of renewable energy technologies determine the number and\nspatial distribution of installations. We use data of the Swiss Federal O”ce of Energy\n[Klauser et al., 2022] to get the annual rooftop PV potential in Switzerland. Ground-\nmounted PV potential was derived from national solar irradiation datasets under\ncurrent climate conditions [Sharma, 2023], accounting for panel e”ciency, ground\ncoverage, topography, and land-use exclusions. Based on [Dujardin and Lehning,\n2022], wind energy potential was calculated on regional capacity factors and repre-\nsentative turbine models, while small hydropower potential relied on the national\nhydropower inventory [Hertach, 2012, SFOE, 2012] and the hydraulic potential model\nHYDROpot\nintegral [Hirschi et al., 2013, Laub et al., 2022]. Detailed assumptions,\nexclusion criteria, and parameter values are documented in the Supporting Informa-\ntion. Across all technologies, areas of biodiversity importance were excluded in line\nwith conservation guidelines. A complete list of protected-area datasets and exclusion\nbu!ers, as well as the treatment of border e!ects, is provided in the Supporting Infor-\nmation. The maps showing the production potential for ground-mounted PV, WT\nand ROR hydroelectric plants, along with the corresponding ERI computed at pixel\nlevel, are presented in Figure A1.\n2.4 Siting strategies\nWe consider three siting strategies for selecting pixels where energy infrastructure\ncan be installed. Pixels in excluded area or without energy generation potentials for\n16\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nthe considered technology are not considered for siting. For ROR plants, the selected\nlocations are computed as the centroids of the river section polygon geometries.\nFor a given energy infrastructure, we define the boolean variable x =\n[x1,x 2,...,x n]↓ ,w h e r en is the total number of suitable pixels. The elements of x,\nwhich are relative to each pixel, are only active (i.e., equal to 1) if the corresponding\npixels are selected for the installation. The production potentials of pixels are given\nby V =[ V\n1,V 2,...,V n]↓ .\nFollowing Adde et al. [2025a], an extinction risk indicator (ERI) is derived by\nnormalizing the logarithm of the REP:\nERI (p)\nG =\nlogREP (p)\nG ↗ min\np→→P\nlogREP (p→)\nG\nmax\np→→P\nlogREP (p→)\nG ↗ min\np→→P\nlogREP (p→)\nG\n, →p ↑P , (12)\nwhere logREP (p)\nG is the logarithmic transformation of REP (p)\nG where zero values have\nbeen replaced before the transformation.\nFor all pixels, we determine the ERI E =[ ERI1,E R I2,...,E R I n] for a com-\nbined extinction risk raster that accounts for all impacted groups. Therefore, G in\nEquation (12) stands for the group consisting of all species impacted by the renew-\nable energy considered. This normalized metric is particularly suited for trade-o!\nstrategies based on convex combinations to balance energy provision and extinction\nrisk. The ERI is independent of the new infrastructure location, as it considers only\nbiodiversity data in the reference state.\nFinally, the demand D corresponds to the minimum production to be reached by\nthe considered technology.\nMaximizing production\nThe first strategy minimizes the surface transformed for the new installations, by\nchoosing locations with the highest potential of production. We consider this approach\n17\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nas a proxy of the minimum cost strategy, since it reduces the number of required\ninstallations. The first siting approach is formalized as:\nmin\nx→{0,1}n\nn∑\ni=1\nxi\ns.t.\nn∑\ni=1\n(xi · Vi) ↓ D.\n(maxprod )\nMinimizing extinction risks\nThe second strategy selects pixels with minimum risk of extinction for potentially\nimpacted species groups, thus avoiding areas with high conservation values in terms\nof species richness of these groups. The ERI serves here as a biodiversity impact\nindicator, and the strategy can be formulated as:\nmin\nx→{0,1}n\nn∑\ni=1\nxi · Ei\ns.t.\nn∑\ni=1\n(xi · Vi) ↓ D.\n(minERI )\nHere, we apply the ERI as the biodiversity impact indicator to reduce computa-\ntion time, as the global CF can only be computed after all siting decisions have\nbeen made (Section 2), thus resulting in a computationally expensive combinatorial\nproblem. To solve Equation (min ERI ), the pixels are prioritized according to the lex-\nicographic order (E i, ↗Vi), that is, the strategy selects regions with the lowest ERI\nvalues, prioritizing higher-production areas when extinction risks are equal.\n18\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nIdentifying trade-o! s\nFinally, we consider a trade-o! strategy that balances the surface area used and\nbiodiversity impact. To that end, we introduce a weighting parameter ϱ ↑ [0, 1]:\nmin\nx→{0,1}n\nn∑\ni=1\n(\nϱ · Ei ↗ (1 ↗ ϱ) · ˜Vi\n)\n· xi\ns.t.\nn∑\ni=1\n(xi · Vi) ↓ D,\n(trade-o! )\nwhere ˜V is the normalization of V in [0, 1].\nWe assess site selections, varying ϱ between 0 and 1. For each value, available\npixels are prioritized based on the objective function defined in Equation (trade-o! ),\nresulting in unique siting configurations for each ϱ. As with the minERI approach,\nwe apply a lexicographic sorting (ϱ · Ei ↗ (1 ↗ ϱ) · ˜Vi, ↗ ˜Vi): when multiple pixels share\nthe same objective value from Equation (trade-o! ), those with higher production\npotential are prioritized.\nBecause the ERI serves only as a proxy for biodiversity damage in this selection\nstrategy, we also compute the global CFs associated with each value of ϱ and its\ncorresponding selected pixels. These are used to estimate a Pareto front describing\njoint trade-o! between biodiversity impact and infrastructure size.\nIn the results, we apply the least-distance-to-ideal method to identify a represen-\ntative trade-o! siting. For this purpose, the global CFs and infrastructure sizes of the\nnon-dominated solutions are normalized between 0 and 1. In this normalized two-\ndimensional space, the ideal siting corresponds to the minimum impact and minimum\ninstallation size, that is, the point (0, 0). We then select the Pareto solution with the\nsmallest Euclidean distance to this ideal point, which we report as the trade-o! siting\nconfiguration.\n19\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n3 Results\nThe biodiversity impacts of rooftop PV, ground-mounted PV, WT and ROR in\nSwitzerland are calculated on a global national level and for the local level on a pixel\nbasis, as introduced in Section 2.2. In this section, we compare the results across\nvarious site-selection scenarios on a national level.\nIn the baseline scenario of Energy Perspectives 2050+ [Kemmler et al., 2021] for\na net-zero Swiss energy system, PV energy conversion in Switzerland increases by\n29.34 TW h between 2025 and 2050, and the projected expansion of production from\nWT of 4 TW h. In our calculations of energy infrastructure deployment, we use these\nexpansions as production goals for both rooftop and ground-mounted PV systems, as\nwell as WT, accounting for local solar irradiation and wind availabilities. For small\nhydroelectric plants, an expansion of 0.77 TW h is assumed in the baseline scenario,\nmainly involving micro-hydropower plants, each with an installed capacity of less\nthan 300 kW. To estimate the additional hydroelectric capacity required, we used a\nmedian number of full load hours of 4397 h per year, based on Swiss hydropower\nplant statistics [Dasen and Hertach , 2012]. Using this metric, an additional 182 MW\nof capacity is required to produce the target 0.77 TW h per year. A summary of the\nresults presented in this section is provided in Table 4, covering all energy types and\nsiting strategies.\nMaximizing production\nFirst, we apply the max\nprod siting strategy maximizing electricity production of\neach renewable technology. For rooftop PV, this strategy leads to the selection of\napproximately 236 thousand roofs (Figure B2a), covering a rooftop area of approxi-\nmately 152 km2. The majority of these installations are situated in low-altitude regions\n20\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n(Figure B2b), with approximately 60 % concentrated between 400 m and 600 m in ele-\nvation and less than 2 % located above 1500 m. The installations are spread out across\nthe country.\nGround-mounted PV installations leads to the selection of pixels covering a total\nland area of approximately 152 km 2 (Figure B3a), and results in a national biodi-\nversity impact of 7.42 species-eq lost·years. About 38 % of these installations are\nsituated above 1500 m (Figure B3b). The panels are mostly located in agricultural\nand wooded areas, with about 59 % and 35 %, respectively, and the remaining 6 %\ncover unproductive land.\nWhen siting WT to maximize the electricity production (Figure B3a), 279 WT\nare built resulting in a national biodiversity impact of approximately 6.42 ≃ 10\n↑2\nspecies-eq lost·years (Figure B7a). The vast majority (90 %) is sited below 1500 m\n(Figure B7b). The installations are located in wooded (59 %) and agricultural (41 %)\nareas.\nFavouring river sections with higher production potentials, 616 watercourse\nsections are selected to each host a ROR plant, resulting in a national biodiversity\nimpact of approximately 6.74 ≃ 10\n↑2 species-eq lost ·years. The hydropower sites are\nspread across the country Figure B11a, with half of the installations in low-altitude\nriver reaches (49 % under 1500 m) Figure B11b.\nMinimizing extinction risks\nThe second siting strategy applied minimizes the extinction risk (Equation (min ERI )).\nThis strategy is applied for the siting of all renewable technologies apart from\nrooftop PV, which we assume to have no biodiversity impact and thus an ERI of zero.\nBy prioritizing sites with the lowest ERI for the impacted species groups, the total\narea for ground-mounted PV increases by approximately 15 % to 176 km 2 compared\nto the strategy maximizing production, but with a total biodiversity impact reduced\nby over 96 % to approximately 2.70 ≃ 10↑1 species-eq lost·years. The vast majority\n21\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n(96 %) of the sites are located below 1500 m (Figure B4a, Figure B4b). Further, a\nlarger share of installations is sited on agricultural land (76 %), while wooded and\nunproductive areas are selected for only 23 % and 1 %, respectively.\nApplied to WT, the siting strategy minimizing ERI fails to protect biodiversity and\nincreases national impacts to 4.32 ≃ 10↑1 species-eq lost·years, that is almost 7 times\nhigher than the solution siting of Equation (max prod ). This is due to the selection of\nlow productivity sites, thus increasing the total number of installed WT to meet the\nelectricity generation target to 1476 WT, that is 5 times more installations compared\nto Equation (max\nprod ). The WT are almost evenly located in both altitudes below\n(49 %) and above (51 %) 1500 m on agricultural, wooded, and unproductive land,\naccounting for 53 %, 38 %, and 9 %, respectively (Figure B8a, Figure B8b).\nWhen applied to small ROR hydroelectric plants, the strategy minimizing\nERI selects 639 sites, which corresponds to an increase by 4 % compared to\nEquation (max\nprod ). However, the global impact is reduced of approximately 58 %,\nwith 2.86 ≃ 10↑2 species-eq lost·years. Two thirds of the plants lie in high altitudes\nareas above 1500 m (Figures B12a and B12b).\nTrade-o! strategy\nFinally, we analyze siting decisions that balance electricity production and ERI by\napplying the trade-o! strategy (Equation (trade-o! )) to all renewable technologies\nexcept rooftop PV, as this technology is assumed to have no impact on biodiversity.\nThe trade-o! strategy uses a weighting factor ϱ to balance electricity production\nand ERI, generating Pareto-optimal siting configurations that meet the electricity\nproduction target. The selected trade-o! solution is the non-dominated configuration\nclosest to the ideal point (see Section 2.4).\nThe total land footprint of the renewable installations and the global CFs are shown\nfor varying values of ϱ in Figure B5a for ground-mounted PV. As ϱ increases, the\narea requirement monotonically increases, whereas the biodiversity impact decreases\n22\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nsharply. An estimate of the Pareto front for the trade-o! between land footprint and\nbiodiversity impacts is estimated and represented in Figure B5b.\nUsing the least-distance-to-ideal method (see Section 2.4), we identify the siting\nconfiguration associated with the optimal trade-o! weighting parameter ϱPVG =0 .34\nfor ground-mounted PV. The spatial distribution of selected sites for the trade-o!\nstrategy is shown in Figure B6a and the associated altitude distribution is presented\nin Figure B6b. The tradeo! strategy results in a total land use of 157 km 2,w h i c hi s\n3 % higher than maxprod but 11 % lower than minERI . The corresponding national\nbiodiversity impact is reduced by 75 % compared to the maxprod strategy to 1.89\nspecies-eq lost·years, but is 7 times higher than that of minERI . Low-altitude locations\n(below 1500 m) account for 65 % of the selected sites, while the remaining are situated\nabove 1500 m elevation. The installations are predominantly located in agricultural\nareas, representing 77 % of the land selected, with the remaining share being located\nin wooded (19 %) and unproductive (4 %) areas.\nThe number of WT to deploy and the corresponding biodiversity impacts generated\nare evaluated for varying values of ϱ and shown in Figure B9a. As ϱ increases, the\nnumber of WT rises monotonically. For small ϱ values, the number of WT remains\nnearly constant, increasing only slightly from 279 at ϱ = 0 to 302 at ϱ =0 .6. Beyond\nthis range, the number of WT increases sharply, reaching 1476 units at ϱ = 1. In\nparallel, for ϱ ⇐ 0.8, the biodiversity impacts show limited variation, from 5.36 ≃\n10\n↑2 to 6.51 ≃ 10↑2 species-eq lost·years, followed by a pronounced rise at higher\nϱ values, peaking at 4.33 ≃ 10↑1 species-eq lost·years for ϱ = 1. When priority is\ngiven to minimizing the extinction risk, turbine locations shift to less productive\nareas, requiring more WT to meet the production target. Lower extinction risks on\na local level are overcompensated by higher infrastructure needs, ultimately resulting\nin increased ecological costs.\n23\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nAs with ground-mounted PV, we exclude dominated solutions to estimate the\nPareto front (Figure B9b). In this case, we identify ϱWT =0 .48 as the trade-o!\nstrategy for WT, balancing infrastructure deployment and ecological impact. The cor-\nresponding selected pixels for wind energy installations are shown in Figure B10a,\nand their altitude distribution is presented in Figure B10b. In total, 287 WT are\nplanned under this trade-o! configuration, that is a slight increase of 3 % compared\nto maxprod and a substantial 81 % decrease compared to minERI . The trade-o! con-\nfiguration results in a global impact of 5.69 ≃ 10↑2 species-eq lost·years, which is lower\nthan the impacts generated by both the maxprod and minERI siting strategies with\n11 % and 87 % decreases, respectively. A total of 89 % of the WT are located below\n1500 m. Wooded areas account for 63 % of the selected land, and the remaining 37 %\nare located in agricultural zones.\nThe number of small ROR plants and the associated biodiversity impacts are\nshown in Figure B13a for varying values of ϱ. The number of ROR hydroelectric\nplants increases with ϱ, while the global biodiversity impact decreases. The estimated\nPareto front is shown in Figure B13b.\nIn this case, ϱ\nROR =0 .7 is the trade-o! strategy for ROR plants siting. The strat-\negy results in the selection of 623 river reaches for small hydropower installations,\nwhich is slightly higher (1 %) compared to the max\nprod strategy, and slightly lower\n(↗3 %) than the infrastructure requirement of the minERI strategy. The trade-o! con-\nfiguration induces a global impact of 3.89 ≃ 10↑2 species-eq lost·years, 42 % lower and\n36 % higher compared to maxprod and minERI , respectively. The spatial distribution\nof these river reaches is shown in Figure B14a, displaying the centroids of the reach\npolygons, and their altitude distribution is presented in Figure B14b. Installations\nlocated above 1500 m account for 62 % of the total.\n24\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nThe spatial distribution of the trade-o! siting solutions of all renewable technolo-\ngies, together with the relative shares of land cover types for solar and wind energy,\nare shown in Figure 1.\nBiodiversity impact curves\nIn this section, the impacts on biodiversity of the di!erent renewable energy\ntechnologies are assessed for increasing production targets (see Figure 2).\nUsing the three siting strategies defined in Section 2.4, and selecting ϱPVG =\n0.34 for the trade-o! approach, the biodiversity impact curves corresponding to the\nstrategies prioritizing areas of maximum energy potential (Equation (max prod )), areas\nwith lower extinction risk for impacted species groups (Equation (min ERI )), and the\ntrade-o! between minimal land use and biodiversity impact are shown in Figure C15a\nfor ground-mounted PV, and the associated areas used are depicted in Figure C15b.\nAcross all levels of solar energy conversion, the three siting strategies are ranked\nconsistently in terms of biodiversity impact. The maxprod strategy yields the highest\nimpacts, followed by the trade-o! strategy, whereas the minERI strategy systemati-\ncally results in the lowest impacts. Notably, the impacts associated with the maxprod\nstrategy increase at a substantially steeper rate, highlighting its disproportionately\nstronger e!ect on biodiversity as production levels rise. When considering the area\nrequired for ground-mounted PV installations, the ranking of the three strategies\nremains consistent but becomes inverted: the max\nprod strategy requires the smallest\nsurface area, followed by the trade-o! strategy, while the minERI strategy requires\nthe largest area. The latter strategy also exhibits the steepest increase in required\nsurface area as electricity production rises.\nSimilarly, we investigate the impacts and required infrastructure sizes of wind\npower for a production target varying between 0 up to 4 TW h (Figures C16a\nand C16b), which corresponds to the planned wind power production expansion\nbetween 2025 and 2050, as well as for small ROR hydropower plants up to the national\n25\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nexpansion target of 0.77 TW h (Figures C17a and C17b). We use the optimal trade-o!\nparameters ϱWT =0 .48 and ϱROR =0 .7 for WT and ROR installations, respectively.\nFigure 2 shows the cumulative biodiversity impacts measured as the global CFs,\nas a function of electricity production for all the technologies considered, under the\ndi!erent siting strategies. For reference, rooftop PV is included as a baseline scenario,\nand because this technology is assumed to have negligible land-use–related biodiversity\ne!ects [Katzner et al., 2013], its curve remains e!ectively at zero across all production\nlevels. Comparing curves across the other energy types considered, ROR systems\nshow the steepest impact gradients, with the highest biodiversity impacts over the\nentire production range considered. Their impacts increase rapidly at low electricity\noutputs and continues to rise steadily with increasing production. Ground-mounted\nPV plants globally yield the lowest biodiversity impacts compared to ROR and WT\ntechnologies, with an exception when considering the max\nprod strategy. The latter\nused for ground-mounted PV, results in an impact that increases faster than WT\ncurves for the min\nERI and trade-o! approaches. Finally, WT globally exhibit the\nsecond-highest impact levels. Although their curves start with relatively similar values\nto ROR at low production, the rate of increase is less pronounced.\n26\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nT able 4: Summary of electricity production, biodiversity impact (CF glo), and infrastructure\nsize (in area or number of units) for rooftop photovoltaic (PV R), ground-mounted photo-\nvoltaic (PV G), wind turbines (WT), and small run-of-river hydropower (ROR) under the\nsiting strategies maxprod , minERI , and trade-o!.\nEnergy Strategy Production [TWh] CFglo [species-eq lost·years] Land use/number of installations\nPVR maxprod 2.9340 · 101 01 .5175 · 102 km2\nPVG\nmaxprod 2.9341 · 101 7.4233 1.5246 · 102 km2\nminERI 2.9341 · 101 2.6979 · 10→1 1.7603 · 102 km2\ntrade-o! 2.9343 · 101 1.8889 1.5655 · 102 km2\nWT\nmaxprod 4.0017 6.4215 · 10→2 279\nminERI 4.0062 4.3235 · 10→1 1476\ntrade-o! 4.0031 5.6874 · 10→2 287\nROR\nmaxprod 7.7036 6.7444 · 10→2 616\nminERI 7.7011 2.8646 · 10→2 639\ntrade-o! 7.7028 3.8904 · 10→2 623\n27\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n(a) Selected locations\n(b) Land cover distribution\nFig. 1: Locations of the trade-o! siting solutions for ground-mounted photovoltaic\n(PVG), wind turbines (WT), and small run-of-river hydropower (ROR) (a), and their\nrelative shares of land cover types for PV G and WT (b). Unproductive areas corre-\nspond to unwooded, non-built areas that are unsuitable for cultivation.\n28\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nFig. 2: Biodiversity impacts of rooftop photovoltaic (PV R), ground-mounted photo-\nvoltaic (PV G), wind turbines (WT), and small run-of-river hydropower (ROR) as a\nfunction of electricity production under the siting strategies maxprod , minERI , and\ntrade-o!.\n29\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n4 Discussion\nIn this study, we developed and applied a framework for siting infrastructure that\nincorporates biodiversity protection by using fine-scale datasets and accounting for a\nbroad range of aquatic and terrestrial species groups. Rather than representing bio-\ndiversity as a fixed spatial constraint independent of infrastructure placement, the\nframework quantifies biodiversity impacts as a function of the selected siting configu-\nration. Therefore, it captures how alternative siting strategies redistribute ecological\npressure. In contrast, traditional spatial conservation planning tools [Ball and Poss-\ningham, 2000, Ball et al., 2009, Moilanen et al., 2005, Ciarleglio et al., 2009] generate\nrelative conservation priorities but do not provide implementable infrastructure sit-\ning guidance [McIntosh et al., 2018]. Specifically, we assessed how deployment choices\ninfluence the ecological impacts of rooftop and ground-mounted PV, WT, and small\nROR hydroelectric plants. Three siting strategies (Section 2.4) were evaluated and\ncompared in terms of their biodiversity impacts and land footprint. For each renew-\nable energy technology, the relevant impacted species groups were identified (Table 3),\nand detailed biodiversity maps were used to estimate site-specific extinction probabil-\nities. The latter were then combined with CFs to quantify national level biodiversity\nimpacts of renewable energy infrastructure (Section 2.1). This integration of high-\nresolution data extends LCA-based biodiversity metrics, which are typically applied\nat coarse spatial scales or global averages [De Baan et al., 2013, Verones et al., 2013,\nMay et al., 2020, Chaudhary et al., 2015, Scherer et al., 2023], toward a spatially\nexplicit decision-support context suitable for infrastructure planning. Furthermore, to\nour knowledge, the inclusion of such a broad range of taxonomic groups represents a\nnovel contribution, as previous LCA-based studies generally focus on a limited set of\nspecies groups [Damiani et al., 2023]. Although the empirical results presented here\nare specific to Switzerland, the methodological framework is transferable to other\nregions provided that su ”ciently resolved data are available.\n30\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nThe scalability of rooftop PV is constrained by the limited availability of suitable\nrooftop area and its lower production potential compared to ground-mounted PV.\nHowever, given its negligible biodiversity impact, it serves as a reference baseline in\nour study. For a given energy type, we observe a consistent ordering of the three siting\nstrategies for increasing electricity production targets, both in terms of infrastructure\nrequirements and biodiversity impacts (Figure 2). The persistence of these patterns\nhighlights that biodiversity-aware siting leads to systematic and predictable changes\nin infrastructure distribution and biodiversity impacts, supporting transparent eval-\nuation of planning trade-o!s that would not be captured by coarse-scale approaches\nor limited-taxon studies [Zelm et al., 2011, Geyer et al., 2010, de Baan et al., 2015,\nVerones et al., 2013]. Importantly, this stable ranking of strategies for a given tech-\nnology likely reflects structural trade-o!s between energy potential and ecological\nsensitivity, and may therefore extend beyond the Swiss context.\nThe biodiversity impacts of ground-mounted PV and ROR power plants are\nhighly sensitive to siting decisions. Notably, reductions in biodiversity impact are\nachieved at the cost of increasing area demand. The max\nprod strategy, which priori-\ntizes pixels with the highest energy potential, yields the smallest installation footprint\n(Figures C15b and C17b) but the highest biodiversity impacts for ground-mounted PV\nand ROR power plants (Figures C15a and C17a). In contrast, strategies that incor-\nporate extinction risk information of impacted species groups can yield significant\nreductions in biodiversity impact, with only modest increases in land use or infras-\ntructure requirements. Explicitly accounting for species-group vulnerabilities reshapes\noptimal deployment patterns, revealing trade-o!s that are not apparent when energy\nperformance alone guides siting [May et al., 2021, Dorber et al., 2020]. This observation\nunderscores the importance of integrating biodiversity considerations into renewable\nenergy system planning. However, for wind power, prioritizing areas of low extinc-\ntion risk results in the selection of areas with low availability, dramatically increasing\n31\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nthe number of turbines required to meet production targets (Figure C16b). Conse-\nquently, prioritizing low local extinction risk paradoxically increases the overall impact\n(Figure C16a) due to substantially higher infrastructure requirements. This demon-\nstrates that siting strategies based solely on local biodiversity considerations can\nbackfire at the system scale. Instead, e!ective siting strategies must balance availabil-\nity and biodiversity impacts on a system level, with trade-o! configurations o!ering\nlower impacts for only marginally higher infrastructure needs.\nAcross all technologies, clear spatial patterns emerge when comparing the three\nsiting strategies on our case study. The max\nprod approach systematically concentrates\ninstallations in areas with high energy potential but also higher environmental sen-\nsitivity: low-elevation forested zones for wind, and substantial shares of wooded land\nfor ground-mounted PV. In contrast, the min\nERI strategy shifts deployments toward\nless sensitive landscapes, typically agricultural areas and, for wind and ROR hydro,\nhigher elevations. Trade-o! solutions generally fall between these extremes but tend\nto resemble one or the other depending on the technology: for wind, the compromise\nremains close to max\nprod , whereas for ground-mounted PV it aligns more strongly\nwith the ERI-oriented pattern, strongly favouring agricultural land. These spatial con-\ntrasts illustrate how biodiversity-informed siting can redirect infrastructure toward\nlandscapes where ecological pressure is lower without eliminating energy potential, a\ncapability that prior studies could not o!er [May et al., 2021, Dorber et al., 2020].\nIn real-world siting decisions, additional criteria beyond plant availability and\nbiodiversity impact must be considered, including technical constraints related to\naccessibility and grid connection as well as issues of social acceptance [Huber et al.,\n2017]. In the maximum production and trade-o! strategies, installations tend to be\nheavily clustered in a limited number of regions, where a higher production potential\nof renewable plants is expected. Specifically, southern regions with higher solar irra-\ndiation account for a substantial share of the ground-mounted PV installations, while\n32\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nmost additional small ROR hydroelectric plants are concentrated in mountainous\nregions with suitable hydrological conditions. Such spatial clustering of infrastructure\ncan exacerbate social acceptance challenges, which remain a critical non-technical\nbarrier to the deployment of renewable energy [Bogdanov et al., 2019] and therefore\nmust be included in siting decisions. The modular structure of the framework allows\nadditional spatial and socio-technical criteria to be incorporated, supporting more\ncomprehensive planning analyses [Moilanen et al., 2009, Pressey et al., 2007, Kienast\net al., 2017].\nWhen interpreting the results of this study, several limitations should be consid-\nered. First, impacts on biodiversity are highly region-specific and require detailed\necological surveys that are hard to generalize. Here, we calculate the ERI using\nhigh-resolution species distribution data for Switzerland. This indicator does not\naccount for species migration within or across borders. However, species that disappear\nin Switzerland may persist elsewhere and potentially recolonize from neighbouring\nregions. Consequently, our extinction metrics should not be interpreted as measures\nof irreversible global loss.\nAdditionally, our analysis assumes static species distributions, neglecting ecological\ndynamics such as dispersal or climate-driven range shifts. This introduces uncertainty\nin the estimated biodiversity impacts, depending on species mobility and habitat\nconnectivity.\nAlthough species richness is a widely used metric for measuring biodiversity, it\ncaptures only one facet of biological diversity. Other dimensions, such as genetic,\nfunctional, or ecosystem diversity, are not represented in our indicators, which may\noversimplify biodiversity responses to energy development. While this study focuses\non species richness, future multi-criteria siting approaches should incorporate these\nadditional components.\n33\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nFinally, our assessment focuses on local land-use–related impacts occurring during\nthe operational phase of renewable energy infrastructure. Upstream and downstream\nprocesses of the life cycle, such as construction, manufacturing, and waste treatment\nare intentionally excluded as may occur in other geographies with di!erent ecolog-\nical sensitivities. These life cycle stages can also pose risks to biodiversity in those\nlocations.\nThese limitations suggest that our results represent a partial yet informative esti-\nmate of the overall biodiversity impacts associated with renewable energy expansion.\nWhile our dataset provides a detailed and spatially explicit assessment within Switzer-\nland, comparable data are often unavailable at similar resolution in other countries,\nlimiting the direct transferability of our approach. Therefore, the numerical impact\nlevels and optimal parameter values reported here should be interpreted as context-\nspecific rather than globally applicable. Nevertheless, the framework presented here\nprovides a foundation for evaluating trade-o!s between energy infrastructure deploy-\nment and local biodiversity conservation, and o!ers a basis for future refinement as\nmore detailed ecological and life-cycle data become available, especially at broader\ngeographical scales.\n5 Conclusion\nOne essential strategy to reach net-zero emissions is to shift away from fossil fuels in\nfavour of a substantial expansion of renewable energy capacity. Yet, this transition\nmust be designed to avoid adding pressure on ecosystems and accelerating biodiversity\nloss, which would further degrade the benefits humans get from nature. This study\npresents a transferable framework for assessing and mitigating biodiversity impacts\nassociated with infrastructure deployment. Although tailored here to renewable energy\ntechnologies, namely rooftop and ground-mounted PV, WT, and ROR hydroelectric\n34\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nplants, and demonstrated using Switzerland as a case study, the approach is designed\nto be applicable to other infrastructure types and geographical contexts.\nBuilding on metrics initially designed for LCA integration, the framework inte-\ngrates high-resolution species distribution data, installation-specific impacted species\ngroups, and region-specific species vulnerabilities, considering 20 impacted species\ngroups based on an extensive literature review. Combining this with spatially explicit\nproduction potential data of the considered infrastructure, we quantify biodiversity\nimpacts under alternative siting strategies for projected renewable energy deployment\nin 2050.\nOur results show that siting choices greatly influence biodiversity outcomes. A\nstrategy that maximizes electricity production concentrates infrastructure in areas of\nhigh potential but also potentially high ecological sensitivity, leading to dispropor-\ntionately high biodiversity impacts, particularly for ground-mounted PV and ROR\nplants. Conversely, strategies that minimize local extinction risk may shift deploy-\nment toward areas with lower generation potential, increasing installation areas. For\nwind energy, such strategies paradoxically substantially increase overall biodiversity\nimpact. Trade-o! strategies consistently o!er more balanced outcomes, reducing eco-\nlogical impacts relative to pure production maximization while avoiding the excessive\ninfrastructure requirements of strict extinction risk minimization.\nThe e!ectiveness of these strategies is technology-dependent, demonstrating that\na single cross-technology siting approach is not su ”cient. Instead, the decision\ncriteria should be technology-specific and explicitly account for biodiversity consider-\nations. More importantly, our findings show that integrating biodiversity information\ninto early-stage planning enables energy systems to be designed in ways that can\nsignificantly reduce harm to ecosystems.\n35\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nOverall, the proposed framework provides a robust foundation for sustainable\ninfrastructure planning and can be further refined or adapted to other regions. More-\nover, the biodiversity impact functions can be directly integrated into energy system\nmodels, enabling dynamic and scenario based biodiversity-friendly planning. Future\nresearch could further reduce ecological impacts by exploring the large-scale potential\nof hybrid systems, integrating di!erent renewable energy sources, such as leveraging\nexisting hydropower infrastructure to exploit PV potential.\n6 Associated content\nData and code availability . The data and code generated in this study will be\npublicly available upon publication.\nSupporting Information. The Supporting Information will be available upon\npublication.\n7 Acknowledgements\nThe authors thank J« erˆ ome Dujardin for providing valuable maps of capacity factors\nfor di!erent types of wind turbines in Switzerland. They also thank Stefanie Hellweg,\nStephan Pfister, and their team for their constructive feedback at an earlier stage of\nthis work. The authors gratefully acknowledge funding from the Joint ETH-Initiative\nSPEED2ZERO. The project conducts research, develops tools, creates action plans,\nand implements technologies to support a sustainable transformation in Switzerland.\nA transformation that meets international and national climate targets ensures a\nresilient energy supply, allowing biodiversity to regain its richness. SPEED2ZERO\nreceived support from the ETH-Board under the Joint Initiatives scheme. S.M. and\nG.W. acknowledge support from the Swiss National Science Foundation (SNSF) under\ngrant no. PZ00P2 202117.\n36\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n8 Author information\nAuthor contributions. Marie-Ange Dahito: Conceptualization, Methodology,\nSoftware, Formal analysis, Investigation, Data curation, Visualization, Writing - orig-\ninal draft, Writing - reviewing & editing.\nDavid Yang Shu: Conceptualization, Writing - reviewing & editing.\nGabriel Wiest: Conceptualization, Writing - reviewing & editing.\nStefano Moret: Conceptualization, Writing - reviewing & editing.\nTobias Wechsler: Data curation, Writing - reviewing & editing.\nLo ¨ ıc Pellissier: Conceptualization, Funding acquisition, Supervision, Writing - review-\ning & editing.\nCompeting interests. The authors declare that they have no competing interests.\n37\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nAppendix A Production potentials and extinction\nrisks\n(a) PV G potential\n (b) ERI for PV\n(c) WT potential\n (d) ERI for WT\n(e) ROR potential\n (f) ERI for ROR\nFig. A1: Maps of production potential for ground-mounted photovoltaic (PV G), wind\nturbines (WT), and small run-of-river hydropower (ROR) alongside the extinction\nrisk indicator (ERI) values of species a!ected by the respective energy systems.\n38\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nAppendix B Selected renewable energy sites and the\nassociated national biodiversity impacts\nB.1 Rooftop photovoltaic\n(a) Selected roofs\n (b) Altitude distribution\nFig. B2: Selected roofs and their altitude distribution for rooftop photovoltaic pro-\nduction using strategy maxprod .\nB.2 Ground-mounted photovoltaic\n(a) Selected pixels\n (b) Altitude distribution\nFig. B3: Selected pixels and their altitude distribution for ground-mounted photo-\nvoltaic production using strategy max\nprod .\n39\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n(a) Selected pixels\n (b) Altitude distribution\nFig. B4: Selected pixels and their altitude distribution for ground-mounted photo-\nvoltaic production using strategy minERI .\n(a) Areas and biodiversity impacts\n (b) Estimated Pareto front\nFig. B5: Areas required for ground-mounted photovoltaic panels and associated bio-\ndiversity impacts (a), and estimated Pareto front (b) across varying ϱ values under\nthe strategy trade-o!.\n40\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n(a) Selected pixels\n (b) Altitude distribution\nFig. B6: Selected pixels and their altitude distribution for ground-mounted photo-\nvoltaic production using strategy trade-o!.\nB.3 Wind power\n(a) Selected pixels\n (b) Altitude distribution\nFig. B7: Selected pixels and their altitude distribution for wind power production\nusing strategy max\nprod .\n41\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n(a) Selected pixels\n (b) Altitude distribution\nFig. B8: Selected pixels and their altitude distribution for wind power production\nusing strategy minERI .\n(a) Number of wind turbines and biodiver-\nsity impacts\n (b) Estimated Pareto front\nFig. B9: Number of wind turbines and associated biodiversity impacts (a), and esti-\nmated Pareto front (b) across varying ϱ values under the strategy trade-o!.\n42\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n(a) Selected pixels\n (b) Altitude distribution\nFig. B10: Selected pixels and their altitude distribution for wind power production\nusing strategy trade-o!.\nB.4 Small hydropower\n(a) Selected locations\n (b) Altitude distribution\nFig. B11: Selected locations (computed as the centroids of the river section polygon\ngeometries) (a) and their altitude distribution (b) for small run-of-river hydropower\nproduction using strategy max\nprod .\n43\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n(a) Selected locations\n (b) Altitude distribution\nFig. B12: Selected locations (computed as the centroids of the river section polygon\ngeometries) (a) and their altitude distribution (b) for small run-of-river hydropower\nproduction using strategy min\nERI .\n(a) Number of run-of-river hydroelectric\nplants and biodiversity impacts\n (b) Estimated Pareto front\nFig. B13: Number of small run-of-river plants and associated biodiversity impacts\n(a), and estimated Pareto front (b) across varying ϱ values under the strategy trade-\no!.\n44\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n(a) Selected locations\n (b) Altitude distribution\nFig. B14: Selected locations (computed as the centroids of the river section polygon\ngeometries) (a) and their altitude distribution (b) for small run-of-river hydropower\nproduction using strategy trade-o!.\nAppendix C Biodiversity impact curves\nC.1 Ground-mounted photovoltaic\n(a) Biodiversity impact\n (b) Selected area\nFig. C15: Biodiversity impact (a) and total selected area (b) for ground-mounted\nphotovoltaic (PV\nG) installations, as a function of the production target, under strate-\ngies maxprod , minERI , and trade-o!.\n45\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nC.2 Wind power\n(a) Biodiversity impact\n (b) Selected number of wind turbines\nFig. C16: Biodiversity impact (a) and total number of installations (b) for wind\nturbines (WT), as a function of the production target, under strategies maxprod ,\nminERI , and trade-o!.\nC.3 Small hydropower\n(a) Biodiversity impact\n (b) Selected number of run-of-river plants\nFig. C17: Biodiversity impact (a) and total number of installations (b) for small run-\nof-river hydropower (ROR), as a function of the production target, under strategies\nmax\nprod , minERI , and trade-o!.\n46\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nReferences\nAntoine Adde, Victor Boussange, Yohann Chauvier, Marie-Ange Dahito, Johan Fruh,\nAndrin Gross, Silvia Stofer, Emmanuel Rey, Petra Sieber, Fabian Fopp, et al. Spa-\ntial biodiversity indicators and a composite index for conservation prioritization in\nSwitzerland. bioRxiv, June 2025a. doi: 10.1101/2025.06.10.657334.\nAntoine Adde, Pierre-Louis Rey, Nathan K¨ ulling, Yohann Chauvier-Mendes, Fabian\nFopp, Manuel R. Popp, Olivier Broennimann, Blaise Petitpierre, Nicolas Strebel,\nAndrin Gross, Silvia Stofer, Anthony Lehmann, Niklaus E. Zimmermann, Lo ¨ ıc Pel-\nlissier, Antoine Guisan, and Florian Altermatt. SDMapCH (v1. 3): a Comprehensive\ndatabase of modelled species habitat suitability maps for Switzerland, 2025b.\nAntoine Adde, Pierre-Louis Rey, Nathan K¨ ulling, Yohann Chauvier-Mendes, Fabian\nFopp, Manuel R Popp, Olivier Broennimann, Blaise Petitpierre, Nicolas Strebel,\nAndrin Gross, et al. SDMapCH: a Comprehensive database of >7,500 modelled\nspecies habitat suitability maps for Switzerland. Scientific Data, 12(1):1752, 2025c.\ndoi: 10.1038/s41597-025-06037-x.\nDavid Anderson, Helen Moggridge, Philip Warren, and James Shucksmith. The\nimpacts of ‘run-of-river’hydropower on the physical and ecological condition of\nrivers. Water and Environment Journal , 29(2):268–276, 2015. doi: 10.1111/wej.\n12101.\nAlona Armstrong, Nicholas J Ostle, and Jeanette Whitaker. Solar park microclimate\nand vegetation management e!ects on grassland carbon cycling. Environmental\nResearch Letters, 11(7):074016, 2016. doi: 10.1088/1748-9326/11/7/074016.\nMaite Arroita, Ibon Aristi, Joserra D « ıez, Miren Martinez, Gorka Oyarzun, and Arturo\nElosegi. Impact of water abstraction on storage and breakdown of coarse organic\nmatter in mountain streams. Science of the Total Environment, 503:233–240, 2015.\ndoi: 10.1016/j.scitotenv.2014.06.124.\n47\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nIan R Ball, Hugh P Possingham, and Matthew Watts. Marxan and relatives: soft-\nware for spatial conservation prioritisation. Spatial conservation prioritisation:\nQuantitative methods and computational tools, 14:185–196, 2009.\nIR Ball and HP Possingham. Marxan (v1. 8.2). Marine reserve design using spatially\nexplicit annealing, a manual, 2000.\nAndrea B« arcenas-Garc « ıa, Fernanda Michalski, William H Morgan, Rebecca K Smith,\nWilliam J Sutherland, James P Gibbs, and Darren Norris. Impacts of dams on\nfreshwater turtles: a global review to identify conservation solutions. Tropical Con-\nservation Science, 15:19400829221103709, 2022. doi: 10.1177/19400829221103709.\nNico B¨ atz, Clarisse Judes, Davide Vanzo, Nicolas Lamouroux, Herv« e Capra, Jan\nBaumgartner, Benjamin Berger, and Christine Weber. Patch-scale habitat dynam-\nics: three metrics to assess ecological impacts of frequent hydropeaking. Journal of\nEcohydraulics, 10(1):79–106, 2025. doi: 10.1080/24705357.2024.2426790.\nMar « ıa D Bejarano, Judith Sarneel, Xiaolei Su, and Alvaro Sordo-Ward. Shifts in\nriparian plant life forms following flow regulation. Forests, 11(5):518, 2020. doi:\n/10.3390/f11050518.\nLlu « ıs Benejam, Sandra Saura-Mas, M` onica Bardina, Carolina Sola, Antoni Munn« e,\nand Emili Garc « ıa-Berthou. Ecological impacts of small hydropower plants on head-\nwater stream fish: from individual to community e!ects. Ecology of Freshwater\nFish, 25(2):295–306, 2016. doi: 10.1111/e!.12210.\nDmitrii Bogdanov, Javier Farfan, Kristina Sadovskaia, Arman Aghahosseini, Michael\nChild, Ashish Gulagi, Ayobami Solomon Oyewo, Larissa de Souza Noel Simas Bar-\nbosa, and Christian Breyer. Radical transformation pathway towards sustainable\nelectricity via evolutionary steps. Nature communications, 10(1):1–16, 2019. doi:\n10.1038/s41467-019-08855-1.\nManuela Irene Brunner and Philippe Naveau. Spatial variability in alpine reservoir\nregulation: deriving reservoir operations from streamflow using generalized additive\n48\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nmodels. Hydrology and Earth System Sciences, 27(3):673–687, 2023. doi: 10.5194/\nhess-27-673-2023.\nLuis Santiago Castillo, Camilo Andres Correa Ayram, Clara L Matallana Tobon,\nGerm« an Corzo, Alexandra Areiza, Roy Gonz« alez-M, Felipe Serrano, Luis Cha-\nlan Briceno, Felipe Sanchez Puertas, Alexander More, et al. Connectivity of\nprotected areas: e!ect of human pressure and subnational contributions in the\necoregions of tropical andean countries. Land, 9(8):239, 2020. doi: 10.3390/\nland9080239.\nAbhishek Chaudhary, Francesca Verones, Laura De Baan, and Stefanie Hellweg.\nQuantifying land use impacts on biodiversity: combining species–area models and\nvulnerability indicators. Environmental science & technology , 49(16):9987–9995,\n2015. doi: 10.1021/acs.est.5b02507.\nMichael Ciarleglio, J Wesley Barnes, and Sahotra Sarkar. Consnet: New software for\nthe selection of conservation area networks with spatial and multi-criteria analyses.\nEcography, 32(2):205–209, 2009. doi: 10.1111/j.1600-0587.2008.05721.x.\nJelka Crnobrnja-Isailovi« c, Bogdan Jovanovi« c, Marija Ili« c, Jelena «Corovi« c, Tijana\nˇCubri« c, Dragana Stojadinovi« c, and Nada«Cosi« c. Small hydropower plants’ prolifera-\ntion would negatively a!ect local herpetofauna. Frontiers in Ecology and Evolution,\n9:610325, 2021. doi: 10.3389/fevo.2021.610325.\nMichael Curran, Stefanie Hellweg, and Jan Beck. Is there any empirical support\nfor biodiversity o!set policy? Ecological Applications, 24(4):617–632, 2014. doi:\n10.1890/13-0243.1.\nMattia Damiani, Taija Sinkko, Carla Caldeira, Davide Tosches, Marine Robuchon, and\nSerenella Sala. Critical review of methods and models for biodiversity impact assess-\nment and their applicability in the LCA context. Environmental Impact Assessment\nReview, 101:107134, 2023. ISSN 0195-9255. doi: 10.1016/j.eiar.2023.107134. URL\nhttps://www.sciencedirect.com/science/article/pii/S0195925523001002.\n49\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nGerhard Dasen and Martin Hertach. Dokumentation «minimales Geodatenmodell» -\nStatistik der Wasserkraftanlagen (W ASTA). Technical report, Swiss Federal O”ce\nof Energy SFOE, May 2012.\nLaura De Baan, Christopher L Mutel, Michael Curran, Stefanie Hellweg, and Thomas\nKoellner. Land use in life cycle assessment: global characterization factors based\non regional and global potential species extinction. Environmental science &\ntechnology, 47(16):9281–9290, 2013. doi: 10.1021/es400592q.\nLaura de Baan, Michael Curran, Carlo Rondinini, Piero Visconti, Stefanie Hellweg,\nand Thomas Koellner. High-resolution assessment of land use impacts on biodiver-\nsity in life cycle assessment using species habitat suitability models. Environmental\nscience & technology, 49(4):2237–2244, 2015. doi: 10.1021/es504380t.\nMartin Dorber, Anders Arvesen, David Gernaat, and Francesca Verones. Control-\nling biodiversity impacts of future global hydropower reservoirs by strategic site\nselection. Scientific reports, 10(1):21777, 2020. doi: 10.1038/s41598-020-78444-6.\nAllan L Drewitt and Rowena HW Langston. Assessing the impacts of wind farms on\nbirds. Ibis, 148:29–42, 2006. doi: 10.1111/j.1474-919X.2006.00516.x.\nNicolas Dubos, Maxime Lenormand, Leandro Castello, Thierry Oberdor!, Antoine\nGuisan, and Sandra Luque. Protection gaps in Amazon floodplains will increase\nwith climate change: Insight from the world’s largest scaled freshwater fish. Aquatic\nConservation: Marine and Freshwater Ecosystems, 32(11):1830–1841, 2022. doi:\n10.1002/aqc.3877.\nJ« erˆ ome Dujardin and Michael Lehning. Wind-Topo: Downscaling near-surface wind\nfields to high-resolution topography in highly complex terrain with deep learning.\nQuarterly Journal of the Royal Meteorological Society , 148(744):1368–1388, 2022.\nISSN 1477-870X. doi: 10.1002/qj.4265.\nAline D« epraz, Stefan Eggenberg, Heike Hofmann, Hubert Kr¨ attli, Glenn Litsios,\nAndrin Gross, Silvia Stofer, Samuel Wechsler, and Pascal Tschudin. Swiss species\n50\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\noccurrence data for controlled access by speed2zero.ch, v1. Dataset published via\nGBIF.org, 2025.\nEuropean Commission, Directorate-General for Environment, Niras, Stella, Birdlife,\nEcosystems Ltd, IEEP, and Arcadis. Potential impacts of solar, geothermal and\nocean energy on habitats and species protected under the habitats and birds directives\n– Final report. Publications O”ce of the European Union, 2020. doi: doi/10.2779/\n784760.\nDouglas Evans. Building the European Union’s Natura 2000 network. Nature\nconservation, 1:11–26, 2012. doi: 10.3897/natureconservation.1.1808.\nFOEN. Biogeographical regions of Switzerland (CH). geocat.ch metadata cat-\nalogue, 2020. URL https://www.geocat.ch/geonetwork/srv/eng/catalog.search#\n/metadata/fc110d43-cfbc-4029-a73c-f24bd67c9689.\nFabian Fopp. FFopp/Swiss-Catchment: range mapping - 06/2025 (v0.9), 2025.\nAlexandros Gasparatos, Christopher NH Doll, Miguel Esteban, Abubakari Ahmed,\nand Tabitha A Olang. Renewable energy and biodiversity: Implications for tran-\nsitioning to a green economy. Renewable and Sustainable Energy Reviews, 70:\n161–184, 2017. ISSN 1364-0321. doi: 10.1016/j.rser.2016.08.030. URL https:\n//www.sciencedirect.com/science/article/pii/S1364032116304622.\nRoland Geyer, David M Stoms, Jan P Lindner, Frank W Davis, and Bastian Witt-\nstock. Coupling GIS and LCA for biodiversity assessments of land use. The\ninternational journal of life cycle assessment, 15:454–467, 2010. doi: 10.1007/\ns11367-010-0170-9.\nJulia G« omez-Catas« us, Manuel B Morales, David Giralt, David Gonz« alez del Portillo,\nRobert Manzano-Rubio, Laura Sol« e-Bujalance, Francesc Sard` a-Palomera, Juan\nTraba, and Gerard Bota. Solar photovoltaic energy development and biodiversity\nconservation: Current knowledge and research gaps. Conservation Letters, 17(4):\ne13025, 2024. doi: 10.1111/conl.13025.\n51\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nJacquelyn C Guzy, Evan A Eskew, Brian J Halstead, and Steven J Price. Influ-\nence of damming on anuran species richness in riparian areas: A test of the\nserial discontinuity concept. Ecology and Evolution, 8(4):2268–2279, 2018. doi:\n10.1002/ece3.3750.\nChihiro Haga, Marimi Maeda, Wataru Hotta, Takahiro Inoue, Takanori Matsui,\nTakashi Machimura, Masahiro Nakaoka, Junko Morimoto, Hideaki Shibata, Shizuka\nHashimoto, et al. Scenario analysis of renewable energy–biodiversity nexuses using\na forest landscape model. Frontiers in Ecology and Evolution , 8:155, 2020. doi:\n10.3389/fevo.2020.00155.\nFengzhi He, Christiane Zarfl, Klement Tockner, Julian D Olden, Zilca Campos, Fabio\nMuniz, Jens-Christian Svenning, and Sonja C J¨ ahnig. Hydropower impacts on\nriverine biodiversity. Nature Reviews Earth & Environment, 5(11):755–772, 2024.\ndoi: 10.1038/s43017-024-00596-0.\nRebecca R. Hernandez, Madison K. Ho!acker, Michelle L. Murphy-Mariscal, Grace C.\nWu, and Michael F. Allen. Solar energy development impacts on land cover change\nand protected areas. Proceedings of the National Academy of Sciences, 112(44):\n13579–13584, 2015. doi: 10.1073/pnas.1517656112. URL https://www.pnas.org/\ndoi/abs/10.1073/pnas.1517656112.\nMartin Hertach. Dokumentation Geodatenmodell Kleinwasserkraftpotentiale der\nSchweizer Gew¨ asser. Technical report, Swiss Federal O”ce of Energy SFOE, June\n2012.\nJo¨ elle Hirschi, Tobias Wechsler, Emmanuel Rey, and Rolf Weingartner. Ganzheitliche\nBeurteilung des Wasserkraftpotentials schweizerischer Fliessgew¨ asser: Handbuch\nzur GIS-gest¨ utzten Anwendung der Methode HYDROpot\nintegral. 2013. doi:\n10.48350/47942.\nG« abor Horv« ath, Mikl« os Blah« o,«Ad« am Egri, Gy¨ orgy Kriska, Istv« an Seres, and Bruce\nRobertson. Reducing the maladaptive attractiveness of solar panels to polarotactic\n52\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\ninsects. Conservation Biology , 24(6):1644–1653, 2010. doi: 10.1111/j.1523-1739.\n2010.01518.x.\nNica Huber, Rico Hergert, Bronwyn Price, Christian Z¨ ach, Anna M Hersperger, Marco\nP¨ utz, Felix Kienast, and Janine Bolliger. Renewable energy sources: conflicts and\nopportunities in a changing landscape. Regional Environmental Change, 17(4):\n1241–1255, 2017. doi: 10.1007/s10113-016-1098-9.\nBrondizio ES IPBES et al. Global assessment report on biodiversity and ecosys-\ntem services of the intergovernmental science-policy platform on biodiversity and\necosystem services. IPBES secretariat, 1148, 2019.\nIPCC. Summary for policymakers. In Core Writing Team, H. Lee and J. Romero,\neditor, Climate Change 2023: Synthesis Report. Contribution of Working Groups\nI, II and III to the Sixth Assessment Report of the Intergovernmental Panel on\nClimate Change, pages 1–34. IPCC, Geneva, Switzerland, 2023. doi: 10.59327/\nIPCC/AR6-9789291691647.001.\nIUCN. Species and climate change, December 2019. URL https://iucn.org/sites/\ndefault/files/2022-04/species\nand climate change issues brief-2019-12.pdf .\nIUCN. The iucn red list of threatened species. version 2025-1. https://www.\niucnredlist.org, 2025. Accessed: 2025-08-17.\nYing Jiang, Wen Bo Liao, and Alexander Kotrschal. Small-scale dams deplete frogs\nand toads. Conservation Science and Practice, 4(5), 2022. doi: 10.1111/csp2.12669.\nTA Katzner, JA Johnson, DM Evans, TWJ Garner, ME Gompper, R Altwegg,\nTA Branch, IJ Gordon, and N Pettorelli. Challenges and opportunities for ani-\nmal conservation from renewable energy development. Animal Conservation. 16:\n367-369.[editorial]., 16:367–369, 2013. doi: 10.1111/acv.12067.\nAndreas Kemmler, Almut Kirchner, Sven Kreidelmeyer, Alexander Pi« egsa, Thorsten\nSpillmann, Hans Dambeck, Hanno Falkenberg, Sebastian L¨ ubbers, Andreas\n53\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nBrutsche, Dina Tschumi, Julia Thurau, Florian Ess, Christoph Thormeyer, Hans-\nJ¨ org Althaus, Brian Cox, Benedikt Notter, Martin Jakob, Ulrich Reiter, Giacomo\nCatenazzi, Benjamin Sunarjo, Lia Weinberg, Jonas M¨ uller, and Larissa Lien-\nhard. Energieperspektiven 2050+: Technischer Bericht, Gesamtdokumentation Der\nArbeit. Technical report, Swiss Federal O”ce of Energy SFOE, Bern, December\n2021.\nFelix Kienast, Nica Huber, Rico Hergert, Janine Bolliger, Lorena Segura Moran, and\nAnna M. Hersperger. Conflicts between decentralized renewable electricity pro-\nduction and landscape services – A spatially-explicit quantitative assessment for\nSwitzerland. Renewable and Sustainable Energy Reviews, 67:397–407, 2017. ISSN\n1364-0321. doi: 10.1016/j.rser.2016.09.045.\nDaniel Klauser, Simon Albrecht-Widler, and Eric Matti. Solarpotentialanalyse f¨ ur\nSonnendach.ch, Schlussbericht. Technical report, Bern, 2022.\nKarl Kosciuch, Daniel Riser-Espinoza, Michael Gerringer, and Wallace Erickson.\nA summary of bird mortality at photovoltaic utility scale solar facilities in the\nSouthwestern US. PloS one , 15(4):e0232034, 2020. doi: 10.1371/journal.pone.\n0232034.\nKoen J. J. Kuipers, Stefanie Hellweg, and Francesca Verones. Potential consequences\nof regional species loss for global species richness: A quantitative approach for esti-\nmating global extinction probabilities. Environmental Science & Technology, 53(9):\n4728–4738, 2019. doi: 10.1021/acs.est.8b06173. URL https://doi.org/10.1021/acs.\nest.8b06173. PMID: 30995027.\nThomas H Kunz, Edward B Arnett, Wallace P Erickson, Alexander R Hoar, Gregory D\nJohnson, Ronald P Larkin, M Dale Strickland, Robert W Thresher, and Merlin D\nTuttle. Ecological impacts of wind energy development on bats: questions, research\nneeds, and hypotheses. Frontiers in Ecology and the Environment, 5(6):315–324,\n2007. doi: 10.1890/1540-9295(2007)5[315:EIOWED]2.0.CO;2.\n54\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nAlban Kuriqi, Ant« onio N. Pinheiro, Alvaro Sordo-Ward, Mar « ıa D. Bejarano, and Luis\nGarrote. Ecological impacts of run-of-river hydropower plants—current status and\nfuture prospects on the brink of energy transition. Renewable and Sustainable\nEnergy Reviews, 142:110833, 2021. ISSN 1364-0321. doi: 10.1016/j.rser.2021.110833.\nURL https://www.sciencedirect.com/science/article/pii/S1364032121001271.\nAlix Lafitte, Romain Sordello, Dakis-Yaoba Ou« edraogo, Chlo« e Thierry, Geo!roy Marx,\nJ« er« emy Froidevaux, Bertrand Schatz, Christian Kerbiriou, Philippe Gourdain, and\nYorick Reyjol. Existing evidence on the e!ects of photovoltaic panels on biodi-\nversity: a systematic map with critical appraisal of study validity. Environmental\nEvidence, 12(1):25, 2023. doi: 10.1186/s13750-023-00318-x.\nCarla Laub, Tobias Wechsler, Dorothea Hug Peter, Massimiliano Zappa, and Rolf\nWeingartner. HYDROpot\nintegral, 2022.\nChloe V Long, James A Flint, and Paul A Lepper. Wind turbines and bat mortality:\nDoppler shift profiles and ultrasonic bat-like pulse reflection from moving turbine\nblades. The Journal of the Acoustical Society of America, 128(4):2238–2245, 2010.\ndoi: 10.1121/1.3479540.\nCV Long, JA Flint, and PA Lepper. Insect attraction to wind turbines: does colour\nplay a role? European Journal of Wildlife Research, 57:323–331, 2011. doi: 10.1007/\ns10344-010-0432-7.\nRafa#l #Lopucki, Daniel Klich, and Sylwia Gielarek. Do terrestrial animals avoid areas\nclose to turbines in functioning wind farms in agricultural landscapes? Environmen-\ntal Monitoring and Assessment, 189:1–11, 2017. doi: 10.1007/s10661-017-6018-z.\nAna Teresa Marques, Helena Batalha, Sandra Rodrigues, Hugo Costa, Maria\nJo˜ ao Ramos Pereira, Carlos Fonseca, Miguel Mascarenhas, and Joana Bernardino.\nUnderstanding bird collisions at wind farms: An updated review on the causes\nand possible mitigation strategies. Biological Conservation, 179:40–52, 2014. ISSN\n0006-3207. doi: 10.1016/j.biocon.2014.08.017. URL https://www.sciencedirect.\n55\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\ncom/science/article/pii/S000632071400305X.\nThomas J Matthews, Kostas A Triantis, Robert J Whittaker, and Fran¸ cois Guil-\nhaumon. sars: an r package for fitting, evaluating and comparing species–area\nrelationship models. Ecography, 42(8):1446–1455, 2019.\nRoel May, Heleen Middel, B ˚ ard G. Stokke, Craig Jackson, and Francesca Verones.\nGlobal life-cycle impacts of onshore wind-power plants on bird richness. Envi-\nronmental and Sustainability Indicators, 8:100080, 2020. ISSN 2665-9727. doi:\n10.1016/j.indic.2020.100080. URL https://www.sciencedirect.com/science/article/\npii/S2665972720300647.\nRoel May, Craig R. Jackson, Heleen Middel, B ˚ ard G. Stokke, and Francesca Verones.\nLife-cycle impacts of wind energy development on bird diversity in norway. Envi-\nronmental Impact Assessment Review, 90:106635, 2021. ISSN 0195-9255. doi:\n10.1016/j.eiar.2021.106635. URL https://www.sciencedirect.com/science/article/\npii/S0195925521000858.\nEmma J McIntosh, Sarah Chapman, Stephen G Kearney, Brooke Williams, Glenn\nAlthor, Jessica PR Thorn, Robert L Pressey, Madeleine C McKinnon, and Richard\nGrenyer. Absence of evidence for the conservation outcomes of systematic conser-\nvation planning around the globe: a systematic map. Environmental Evidence,7\n(1):22, 2018. doi: 10.1186/s13750-018-0134-2.\nMillennium Ecosystem Assessment. Ecosystems and Human Well-Being: Synthesis.\nIsland Press, 2005.\nAtte Moilanen, Aldina MA Franco, Regan I Early, Richard Fox, Brendan Wintle, and\nChris D Thomas. Prioritizing multiple-use landscapes for conservation: methods for\nlarge multi-species planning problems. Proceedings of the Royal Society B: Biological\nSciences, 272(1575):1885–1891, 2005. doi: 10.1098/rspb.2005.3164.\nAtte Moilanen, Hugh P Possingham, and Kerrie A Wilson. Spatial conservation\nprioritization: Past, present, and future. In Spatial Conservation Prioritization:\n56\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nQuantitative Methods and Computational Tools. Oxford University Press, May 2009.\nISBN 9780199547760. doi: 10.1093/oso/9780199547760.003.0019.\nHannah Montag, Guy Parker, and Tom Clarkson. The e!ects of solar farms on local\nbiodiversity: a comparative study. Clarkson and Woods and Wychwood Biodiversity,\n2016.\nFederico Montesino Pouzols, Tuuli Toivonen, Enrico Di Minin, Aija S Kukkala,\nPeter Kullberg, Johanna Kuuster¨ a, Joona Lehtom¨ aki, Henrikki Tenkanen, Peter H\nVerburg, and Atte Moilanen. Global protected area expansion is compromised\nby projected land-use and parochialism. Nature, 516(7531):383–386, 2014. doi:\n10.1038/nature14032.\nMaria Cristina Moscatelli, Rosita Marabottini, Luisa Massaccesi, and Sara Marinari.\nSoil properties changes after seven years of ground mounted photovoltaic panels in\nCentral Italy coastal area. Geoderma Regional, 29:e00500, 2022. doi: 10.1016/j.\ngeodrs.2022.e00500.\nTim Newbold. Future e!ects of climate and land-use change on terrestrial vertebrate\ncommunity diversity under di!erent scenarios. Proceedings of the Royal Society b,\n285(1881):20180792, 2018. doi: 10.1098/rspb.2018.0792.\nV Oberosler, S Tenan, EF Zipkin, and F Rovero. Poor management in protected areas\nis associated with lowered tropical mammal diversity. Animal Conservation, 23(2):\n171–181, 2020. doi: 10.1111/acv.12525.\nV Perold, S Ralston-Paton, and P Ryan. On a collision course? The large diversity\nof birds killed by wind turbines in South Africa. Ostrich, 91(3):228–239, 2020. doi:\n10.2989/00306525.2020.1770889.\nRobert L Pressey, Mar Cabeza, Matthew E Watts, Richard M Cowling, and Kerrie A\nWilson. Conservation planning in a changing world. Trends in ecology & evolution,\n22(11):583–592, 2007.\n57\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nBoris Salak, Marcel Hunziker, Adrienne Grˆ et-Regamey, Reto Spielhofer, Ulrike Wis-\nsen Hayek, and Felix Kienast. Shifting from techno-economic to socio-ecological\npriorities: Incorporating landscape preferences and ecosystem services into the sit-\ning of renewable energy infrastructure. Plos one , 19(4):e0298430, 2024. doi:\n10.1371/journal.pone.0298430.\nLaura Scherer, Francesca Rosa, Zhongxiao Sun, Ottar Michelsen, Valeria De Lau-\nrentiis, Alexandra Marques, Stephan Pfister, Francesca Verones, and Koen JJ\nKuipers. Biodiversity impact assessment considering land use intensities and frag-\nmentation. Environmental Science & Technology, 57(48):19612–19623, 2023. doi:\n10.1021/acs.est.3c04191.\nSFOE. Potential of small hydropower plants in Switzerland, June 2012.\nVarun Sharma. Dataset for ”Geospatial segmentation of high- resolution photovoltaic\nproduction maps for Switzerland”, Frontiers Energy, August 2023. URL https:\n//doi.org/10.5281/zenodo.8296864.\nKriszta Lilla Szabadi, Anik« o Kurali, Nor Amira Abdul Rahman, J« er« emy SP Froide-\nvaux, Elizabeth Tinsley, Gareth Jones, Tam« as G¨ orf¨ ol, P« eter Est« ok, and S« andor\nZseb˝ ok. The use of solar farms by bats in mosaic landscapes: Implications for con-\nservation. Global Ecology and Conservation, 44:e02481, 2023. doi: 10.1016/j.gecco.\n2023.e02481.\nKaren E Tanner, Kara A Moore-O’Leary, Ingrid M Parker, Bruce M Pavlik, Sophia\nHaji, and Rebecca R Hernandez. Microhabitats associated with solar energy devel-\nopment alter demography of two desert annuals. Ecological Applications, 31(6):\ne02349, 2021. doi: 10.1002/eap.2349.\nChris B. Thaxter, Graeme M. Buchanan, Jamie Carr, Stuart H. M. Butchart, Tim\nNewbold, Rhys E. Green, Joseph A. Tobias, Wendy B. Foden, Sue O’Brien, and\nJames W. Pearce-Higgins. Bird and bat species’ global vulnerability to collision\nmortality at wind farms revealed through a trait-based assessment. Proceedings\n58\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nof the Royal Society B: Biological Sciences, 284(1862):20170829, September 2017.\nISSN 0962-8452. doi: 10.1098/rspb.2017.0829. URL https://doi.org/10.1098/rspb.\n2017.0829.\nAnne Tolvanen, Henri Routavaara, Mika Jokikokko, and Parvez Rana. How far are\nbirds, bats, and terrestrial mammals displaced from onshore wind power develop-\nment? – a systematic review. Biological Conservation, 288:110382, 2023. ISSN\n0006-3207. doi: 10.1016/j.biocon.2023.110382. URL https://www.sciencedirect.\ncom/science/article/pii/S0006320723004834.\nMihaela Urziceanu, Paulina Anastasiu, Laurentiu Rozylowicz, and Tatiana Eugenia\nSesan. Local-scale impact of wind energy farms on rare, endemic, and threatened\nplant species. PeerJ, 9:e11390, 2021. doi: 10.7717/peerj.11390.\nFrancesca Verones, Dominik Saner, Stephan Pfister, Daniele Baisero, Carlo Rondinini,\nand Stefanie Hellweg. E!ects of consumptive water use on biodiversity in wetlands of\ninternational importance. Environmental science & technology, 47(21):12248–12257,\n2013.\nFrancesca Verones, Mark AJ Huijbregts, Abhishek Chaudhary, Laura de Baan,\nThomas Koellner, and Stefanie Hellweg. Harmonizing the assessment of biodiversity\ne!ects from land and water use within LCA. Environmental science & technology,\n49(6):3584–3592, 2015. doi: 10.1021/es504995r.\nElke Visser, Vonica Perold, Samantha Ralston-Paton, Alvaro C Cardenal, and Peter G\nRyan. Assessing the impacts of a utility-scale photovoltaic solar energy facility on\nbirds in the Northern Cape, South Africa. Renewable energy, 133:1285–1294, 2019.\ndoi: 10.1016/j.renene.2018.08.106.\nChristian C Voigt. Insect fatalities at wind turbines as biodiversity sinks. Conservation\nScience and Practice, 3(5):e366, 2021. ISSN 2351-9894. doi: 10.1111/csp2.366. URL\nhttps://www.sciencedirect.com/science/article/pii/S2351989422001512.\n59\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\nChristian C Voigt, Klara Kaiser, Samantha Look, Kristin Scharnweber, and Carolin\nScholz. Wind turbines without curtailment produce large numbers of bat fatalities\nthroughout their lifetime: A call against ignorance and neglect. Global Ecology and\nConservation, 37:e02149, 2022. ISSN 2351-9894. doi: 10.1016/j.gecco.2022.e02149.\nURL https://www.sciencedirect.com/science/article/pii/S2351989422001512.\nHaoran Wang, Yongcan Chen, Zhaowei Liu, and Dejun Zhu. E !ects of the “Run-\nof-River” hydro scheme on macroinvertebrate communities and habitat conditions\nin a Mountain River of Northeastern China. Water, 8(1):31, 2016. doi: 10.3390/\nw8010031.\nTobias Wechsler, Manfred St¨ ahli, Klaus Jorde, Massimiliano Zappa, and Bettina\nSchaefli. The future of alpine run-of-river hydropower production: Climate change,\nenvironmental flow requirements, and technical production potential. Science of\nthe Total Environment, 890:163934, 2023. doi: 10.1016/j.scitotenv.2023.163934.\nMichelle Weschler and Lusha Tronstad. Wind energy and insects: reviewing the state\nof knowledge and identifying potential interactions. PeerJ, 12:e18153, 2024. doi:\n10.7717/peerj.18153.\nXuezhong Yu, Daming He, and Phouvin Phousavanh. Transboundary Environmental\nE!ects of Hydropower: Sediment Transport and Geomorphology, pages 161–182.\nSpringer Singapore, Singapore, 2019. doi: 10.1007/978-981-13-1565-7\n6.\nRosalie van Zelm, Aafke M Schipper, Michiel Rombouts, Judith Snepvangers, and\nMark AJ Huijbregts. Implementing groundwater extraction in life cycle impact\nassessment: characterization factors based on plant species richness for the nether-\nlands. Environmental science & technology , 45(2):629–635, 2011. doi: 10.1021/\nes102383v.\nPengyi Zhang, Chao Yue, Yu Li, Xianhui Tang, Binbin Liu, Mengyang Xu, Mengyu\nWang, and Lele Wang. Revisiting the land use conflicts between forests and\nsolar farms through energy e”ciency. Journal of Cleaner Production, 434:139958,\n60\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint \n\n2024. ISSN 0959-6526. doi: 10.1016/j.jclepro.2023.139958. URL https://www.\nsciencedirect.com/science/article/pii/S0959652623041161.\nXuancheng Zhao, Fengshi Li, Yuan Yuan, Guna Ari, Yongzhi Yan, Qing Zhang,\nAruhan Olhnuud, and Pengtao Liu. Wind farms reduce grassland plant community\ndiversity and lead to plant community convergence. BMC Ecology and Evolution,\n25(1):10, 2025. doi: 10.1186/s12862-025-02350-6.\n61\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted February 26, 2026. ; https://doi.org/10.64898/2026.02.24.707751doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}