Hotspots in the grid: Avian sensitivity and vulnerability to collision risk from energy infrastructure interactions in Europe and North Africa

The transition to zero carbon energy is essential avoid runaway climate change (IPCC, 2018). However, the expansion of renewable infrastructure (EI) required achieve this poses a challenge wildlife conservation due collision and electrocution risks, particularly for birds other aerial taxa (Bernardino et al., 2018a; Kiesecker 2019; Marques 2014). European, onshore wind capacity projected grow from approximately 169 GW in 2018 between 262 760 by 2050 with enough economically viable turbine locations (approximately 3.4 million) up 13.4 TW (Ryberg 2019). Countries Middle East North Africa also have targets increase share electricity supply Morocco Tunisia aiming 100% (Timmerberg Huge investment transmission network will accompany renewables, an estimated fivefold 2010 (Mckinsey & Company, 2010). when poorly designed or situated, farms power lines can result increased mortality susceptible such as large water birds, gulls, ibis, storks, owls, vultures raptors (Janss, 2000; Oppel 2021; Thaxter 2017). Organisations, companies, charged supporting rollout generation are obliged national, European legal (2009/147/EC, 2010) pan-flyway voluntary (Horns ?ekercio?lu, 2018) frameworks mitigate risks (Gyimesi Prinsen, 2015). Methods evaluate these impacts relatively well understood at project-specific local scales (Schaub 2020; Serrano 2020). assessments often occur after development site has already been selected because initial feasibility studies projects tend focus on economic viability over factors. scale pace new requires greater integration high-level potential cumulative impact regional flyway into highlight areas where additional EI likely significantly risk bird populations (Eichhorn 2017; Loss This important migratory species who may experience multiple developments operation within key migration routes, stopover sites, wintering grounds breeding sites Gove 2013). Bird sensitivity maps be developed illustrate relative associated sensitive (Vasilakis 2016; Warwick-Evans distribution behaviour inferred GPS tracking individuals used create spatio-temporal measure developments, identifying would most exposed (Ross-Smith For sufficient data, combining inputs, available resources, help planners optimise farm line avoiding high during selection stage process (Kiesecker This, turn, reduce mitigation costs produce better outcomes compared site-based alone (Bradbury 2014; Bright 2008). Sensitivity mapping useful assessing negative interactions level flyways. example, map created Red Sea estimates soaring (BirdLife International, tool enables preliminary assessment viewing protected raw tracks species. it does not account all dimensions related risk, height which fly, turn vary depending landscape, meteorological, seasonal species-specific factors (Kleyheeg-hartman 2018; In terrestrial context, largely rely trait-based analysis relation population densities (D'Amico context tracked could collide if turbines powerlines present (Thaxter We calculated susceptibility traits location altitude data 27 species, including resident Europe Northern Africa, describe EI. Within we obtained spatial resolution data; allows us identify hotspots future should discouraged. our work cannot reveal ‘safe’ encouraged. then overlay surface onto density existing vulnerability presence powerlines. Similar approaches using applied assess proposed offshore windfarm survey logistics more challenging Cleasby 2015; Lees Ross-Smith Our highlights variation tagging effort availability helps priority need sharing via online platforms Movebank fill gaps currently feasible publicly data. An overview diagram methods presented Supporting Information S1, section 1. sourced movement repository, web-based platform animal (Movebank, 2019), view maximising coverage Africa. November 2018, identified 254 Europe, Mediterranean A literature search undertaken October March 2019 was whether were summarised 2. did request less than five unless same studies. Data managers contacted January access their datasets response deadline end April 2019. Studies ARGOS Doppler tags (insufficient accuracy; Thomson 2017), captive laboratory-based tests devices, lacking predominantly pelagic included. total, permission use 65 suitable (Figure 1), representing 1,454 individual birds. included some hosted University Amsterdam Bird-tracking system database (UvA-BiTS, Bouten, 2018), offered inclusion requested datasets. To knowledge, fieldwork study relevant licensing authority, further details each dataset provided References paper 1, 3. earliest tag deployment any 2006, while latest date mean duration 2.7 ± 1.8 SD years. Infrastructure terrain processed QGIS ArcGIS (ESRI, Development Team, open project (Garret, OpenStreetMap, OpenStreetMap defines >50 kV (OpenStreetMap, 2019a). Flights 10 60 m above-ground here taken being danger 2) (Harker, Infante Peris, 2003). intersected fishnet grid consisting 554,993 5 × km cells total area 13.9 million square kilometres. Power length kilometres per cell, normalised 0–100 scale. size downloaded windpower.net (Pierrot, contained centroid coordinates information 18,681 cell. Density highest towards West Europe. From hub heights blade lengths dataset, derived (sometimes known rotor swept zone) 15–135 2), S1 4. collected guidance requirements national ethical committees. Measures accuracy uniformly indicated across Where number satellites provided, only ?5 visible (Morris Conner, Duplicate removed. raster package (R version 4.0.5, Hijmans, 2019) append elevation two 30 horizontal vertical digital models (DSMs), STRM-GL1 STRM-GL1-Ellipsoidal OpenTopography Portal (National Science Foundation, NGA NASA, 2000) location. small latitudes 60° latitude ALOS DSM instead (JAXA, 2016). Further surfaces 5. Height metres subtracting ground bird. above ellipsoid, ellipsoidal land used. sea level, orthometric (Péron datasets, Lesser Black-backed Gull studies, bespoke correction obtain had classified non-breeding (including period) season plotting week against latitude, (Supporting 6). Because instantaneous speed second (m/s) time distance subsequent anytime geosphere r packages (Hijmans Eddelbuettel, All 95% confidence interval speeds equal 1.39 m/s (~5 km/hr) flight. approach accounts error given many devices excludes stationary ground. position typically range 1.5 but 31 combined device (Marques categorised flying band 2). Some bursts frequency measurements (up 1hz). recorded independent, filtered remove source bias ensuring minimum 1 min resulting sample 18.0 locations. proportion flight (6.6 observed cell dataset. Due nature heterogeneous region hence considered reliable Silva accounted Wilson's score (Reichensdörfer 2017) whereby lower bound Wilson (WCI), binconf function Hmisc (Harrell, place percentage (Lewis Sauro, 2006; Lott Reiter, Compared proportion, low sizes, effect reducing value assigned uncertainty higher. three locations, (WS) tended WCI around central point estimate height, (n > 50) converged zero, WS became comparable (Cao, See 7. ensure classification driven (Gouhier Pillai, 2020), defined categories 25th, 75th 97.5th percentiles (or vulnerability) zero. Grid scoring 25th percentile ‘Low’, scores ‘Moderate’, ‘High’ top 2.5% observations ‘Very High’. Low’ there none ‘No Data’ lacking. emphasise, therefore, that method exists, absence high-vulnerability interpreted indicative bat (for no study) affected. mapped movements 1). travel continent converge along routes. As expected, flux through bottlenecks European–African Flyway, Southern Iberia, Sinai, Gulf Iskenderun Bosporus Turkey. Important existed Spain, Scotland, Scandinavia, Italy, Eastern median 21, White Stork Ciconia ciconia = 491; 3). 99,641 5x5km (18%) least one 54.9% 54,703) 3a). 13.57% 13,516, 337,900 km2) sensitivity, is, they upper quartile (>0.11). There significant variability (ANOVA F26,59,592 432.4, p < 0.001) Eurasian eagle owl Bubo bubo, whooper swan Cygnus cygnus, eurasian spoonbill Platalea leucorodia, common crane Grus grus white-fronted goose Anser albifrons exhibiting greatest (Table (10–60 band) 37.64% 37,509) 3b). Across 9.41% 9375, 234,375 (>0.14). Iberian imperial Aquila adalberti white stork exhibited determined separately seasons 2 Although proportions similar seasons, fewer overall clustered pattern product sampling smaller tagged season, centred move away observe coastlines major Notable include Western coast France Romania, Moroccan Coast, Sinai Peninsula Baltic Germany. Taxon specific S2 2, compare previous (see discussion) taxon Movebank. cranes, represented how long distances altitudes unlikely highly localised hotspots. plotted sum 4a). degree 28.2% 28,051) 7.0% 7,013) values (>1.13) 1.7% 702) very fall 2.5 (>9.03). Fewer 483, Figure 4b) 6,861, 5c). suggests ubiquitous turbines. High-vulnerability distributed evenly 4a): just countries (Germany, France, Turkey Poland) 50.5% 3,539) cells. Measuring country, Liechtenstein (14.2%, n Germany (7.2%, 1,028), Israel (5.8%, 48), Lebanon (5.4%, 22) Portugal (5.0%, 176; S2, Section must noted ranking partly influenced case Turkey, Israel, Portugal, heights. On hand, Central 55.2% 267) 483 4b). marked differences ranging 0.042 0.081 western marsh harrier Circus pygargus 2.14 3.62 (currently 18% area), identifies different bands These While replacing environmental site-specific relevance populations, successfully identified, powerline minimised protect integrity flyway. coincide bottlenecks, coasts either side Strait Gibraltar (Martín Strait, southern (Buechley supports idea fly exacerbate anthropogenic risks. Rigorous ecological assessment, planning needed bottleneck areas, highlighted (De Pascalis Martín Comparing results Laridae (lesser black-backed gull Larus fuscus) 2019, differed methodology utilised L. fuscus reveals patterns validity (Supplementary type It beyond scope provide explanations ongoing topic research itself, however, morphological style (flapping vs. soaring), behaviour, habitat preference foraging strategy influences Martin Shaw, 2010; Despite efforts complete possible, acknowledge northern Algeria Libya. reflect geographical telemetry (Bouten such, occurs, indicate occur. includes subset larger body mass 350 g more, leading sampling-related evident 4). biases issue ecology, (Brotons 2004). limitations, used, based accounting records provides simple way available. become available, updated code supplementary material 3 benefits expect near becomes affordable miniaturisation fitted Advances sensor technology soon allow detected real-time (O'Donoghue Rutz, One aid improving UvA-BiTS, promotion under-represented taxonomic groups, improved standardisation biologging loggers outwith (Sequeira 2021). Other address model relationship weather, year, topography cover. paper. Overlaying vulnerable effectiveness options, suggest EI, marking visibility, burying cables altering routes (Jenkis turbines, options repowering (Arnett May, 2016), blades (May temporary shutdown periods peak (de Lucas 2012) requirement Jordan (Tomé Another option retrofit radar camera-based systems monitor automatically shut down (McClure Future analyses official, multi-country, composed made researchers. enable consideration voltage open-source electrocution, study, cause injury Hernández-Lambraño Bald Ibises 45% losses caused (Fritz collision, several retrofitting insulators perches (Dixon problem entirely avoided constructing safe poles eliminate first (AEWA, 2012). kind readily transferable do preclude detailed post-construction monitoring disturbance, loss, ; inform constructed early developments. Moreover, effectively target surveys implemented. race tackle crisis, vital neglect biodiversity crisis transitions carbon. thank those granted assisted collection. full summary grateful people helped collect field. Many thanks to: Carlos Carrapato, Instituto da Conservação Natureza e das Florestas, Lisbon, permitting Imperial Eagle 28407489. Peter Desmet (ORCID 0000-0002-8442-8025), Research Institute Nature Forest (INBO), Brussels, Belgium, granting Marsh Harrier (Anselin Vladimir Dobrev, Bulgarian Society Protection Birds Bulgaria), Sofia, Bulgaria, his assistance Neophron percnopterus Bulgaria/Greece Egyptian vulture Study ID 15869951. Steffen Hollerbach, Storkenkate Preten (The Foundation), Preten/Amt Neuhaus, Germany, Sudewiesen 4502577. Elzbieta Kret, WWF Greece, Athens, her Jose Manuel Lopez Vazquez, Ministry Environment, Junta de Andalucía, Sevilla, Proyecto Eremita Geronticus eremita Reintroduction Andalusia (Spain) 463673774. Pacheco Portugal. Emmanuel Pixner, Waldrappteam, Technology Centre (FTZ), Mutters, Austria Ibis Waldrappteam https://doi.org/10.1111/izy.12163. Shay Rotics, Department Ecology, Evolution Behaviour, Hebrew Jerusalem, spring adult storks (data Rotics 560041066. Robin Séchaud, Ecology Evolution, Universite Lausanne, Switzerland, Barn Tyto alba 2001600. Luís Venâncio invaluable ‘Black Storks Portugal’ project. acknowledge: Support training NEXUSS CDT. Tuscan Archipelago National Park, Maremma Regional Diaccia Botrona Reserve, Orbetello Lagoon Orti-Bottagone Reserve Italy rest ‘Progetto Falco pescatore’ team, support activities Ospreys Italy. BTO collaboration Amsterdam; held jointly BTO, funders project, agreement. UvA-BiTS facilitated infrastructures e-Science, NLeSC (http://www.esciencecenter.com/) LifeWatch, carried out Dutch e-infrastructure SURF Foundation. Barnacle Geese Honey Buzzards E-Science, Netherlands eScience Center (NLeSC) conducted e-Infrastructure SURFfoundation. Jethro G. Gauld receives funding NERC part CDT (UKRI) capital grant devices; ‘Rough-legged buzzard’ ‘Peregrine falcon funded Max-Planck Animal German Air Space Administration (DLR). Barents geese FlySafe (https://www.flysafe-birdtam.eu/), initiated Integrated Applications Promotion (IAP) Programme Agency. cooperation Avian Research, Field Ornithology (SOVON) Amsterdam. Tracking Gulls Niedersächsische Wattenmeerstiftung (NWS 04/09); frame LIFE+ ‘Reason Hope’ (LIFE+12-BIO_AT_000143). supported UK Business, Energy Industrial Strategy (BEIS) Offshore Strategic Environmental Assessment (OESEA) programme, Marine Renewable Environment (MaREE) (funded Highlands Islands Enterprise, Fund, Scottish Funding Council), BAE Systems Natural England. Portuguese financed FEDER Funds Operational Competitiveness Factors Program — COMPETE FCT – Foundation ‘POCI-01-0145-FEDER-028176’, InBIO (UID/BIA/50027/2013 POCI-01-0145-FEDER-006821) FCT/MCTES. Env-East DPT, (grant NE/N012070/1), British Ornithologists Union Small Grant, Norwich Park Translational Anglia Innovation Earth Life Alliance funds. Black NERC, NE/N012070/1) Microwave Telemetry ‘Christiane Howey Rising Scholar Award’. Osprey financially (Italy). R.N. acknowledges grants BSF 255/2008, DIP-DFG NA 846/1-1, Minerva Movement Ecology. LIFE ‘The Return Neophron’ (LIFE10 NAT/BG/000152) ‘Egyptian Vulture New LIFE’ (LIFE16 NAT/BG/000874, www.LifeNeophron.eu) co-funded A. A.G. Leventis Balkans implemented ‘Recovery Populations Large Vultures Bulgaria’ (LIFE08 NAT/BG/000278), ‘Vultures back - (LIFE14 NAT/BG/000649) ‘Saving Balkans' last vultures: introducing Safe Areas scavenger Anthropocene’ Whitley Fund Nature. NWO (Netherlands Organisation Scientific Research) polar program (NPP 866.13.010). collection Max Planck Behaviour done under ICARUS directive Aerospace Wigeon Union's Horizon 2020 innovation programme agreement No 727922 (Delta-Flu). authors declare conflict interest content work. J.G.G. principal author work; A.M.A.F. suggested advice throughout writing stages P.W.A., J.P.S. P.R.; contributed critically collection, drafting manuscript gave final approval. Efforts foster entire region. brought together East; lack representation African Nations. by-product current engagement researchers afro-palaearctic download https://www.movebank.org/cms/movebank-main relate therefore prior download. paper, shapefiles infrastructure, Dryad Digital Repository: https://doi.org/10.5061/dryad.jm63xsjcw (Gauld 2022). supinfo Please note: publisher responsible functionality supplied authors. Any queries (other missing content) directed corresponding article.