Aligning conservation status, vulnerability factors, and ecological and evolutionary uniqueness to produce integrated assessments of the world’s birds

Biodiversity monitoring, Birds, Ecological integrity metrics, Conservation 8 prioritization, Phylogenetic and functional uniqueness, Global biodiversity assessments 9 Word count: 5,115 10 Author affiliations: 11 1 American Bird Conservancy, The Plains, V A 20198, USA 12 2 Department of Biology, Indiana University of Pennsylvania, Indiana, PA 15705, USA 13 14 Corresponding author: Eliot T. Miller, American Bird Conservancy, The Plains, V A 20198, USA, 15 [email protected] 16 17 .CC-BY-NC-ND 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint

18 19 A growing awareness, now enshrined in the Kunming-Montreal Global Biodiversity Framework, 20 of the need to monitor biodiversity effectively at scale has led to a proliferation of novel 21 solutions for doing so. Although global biodiversity encompasses all life, from tiny nitrogen-22 fixing bacteria to emergent rainforest trees, birds have several characteristics that make them a 23 frequent focus of such monitoring efforts. In particular, birds frequently give diagnostic, species-24 specific vocalizations that simplify monitoring, they perform a number of critical ecosystem 25 services, they are widely distributed in most ecosystems with strong representation on all 26 continents, and the basic ecology, conservation status, populations, and distributions of many 27 species is well known; birds thus provide a window into the underlying health and habitats of the 28 systems under study. How best to summarize biodiversity monitoring results is a research 29 question that has led to the development of approaches that incorporate species’ IUCN Red List 30 threat assessments into site-level biodiversity scores. Notably, birds’ vocal behavior means that 31 they can be effectively surveyed at scale with passive acoustic monitoring, and the potential to 32 link such monitoring with automated identification and therefore quickly generate site-level 33 biodiversity scores is an appealing approach to implement rigorous evaluations of global 34 biodiversity. Yet, while many of the world’s birds are suffering worrisome population declines, 35 the vast majority of species (78%) are still ranked “Least Concern” by the Red List. In an effort 36 to develop a species scoring system that would be more conducive to such site-level valuations, 37 we integrated key databases of species’ population status assessments, exposure to known 38 vulnerability factors, and their functional and phylogenetic uniqueness to provide quantitative 39 summaries of their conservation significance. We augmented these databases with two novel data 40 sets available for most of the world’s birds: quantitative measurements of migration distances, 41 and species-level phylogenetic and functional uniqueness values comparing each species to those 42 it co-occurs with throughout its range. While the resulting BirdsPlus species scores also 43 inherently reflect our own scientific expertise and judgement, our approach is transparent, 44 dynamic, easily updated, and readily modified by users with different goals or values. 45 46

47 48 The IUCN (International Union for the Conservation of Nature) Red List (IUCN 2022) has, for 49 decades, helped to identify species and geographic areas in need of urgent conservation action. 50 First released in 1966 (Simon 1966), the frequency with which the Red List is used in 51 conservation decisions, the regular updates, and the global attention focused on the list attest to 52 the critical need for such population assessments (Butchart et al. 2025). The Red List can help 53 prioritize conservation of the most endangered species, or on sites that host such species. Setting 54 aside and managing protected areas in this way is essential for biodiversity conservation (Luther 55 et al. 2021), and underpins major global conservation initiatives such as Target 3 of the 56 Kunming-Montreal Global Biodiversity Framework (the “30x30 target” (Convention on 57 Biological Diversity 2022)). However, conserving nature outside protected areas through the 58 implementation of conservation practices on working lands is also critical (Ciuzio et al. 2013; 59 Bennett et al. 2018; Michel et al. 2020; Dertien & Baldwin 2022). 60 One good example of scalable working lands conservation efforts that can be utilized to 61 benefit wildlife and biodiversity are programs and initiatives offered by the United States 62 Department of Agriculture-Natural Resources Conservation Service. For example, the agency’s 63 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint Working Lands for Wildlife and Regional Conservation Partnership Programs focus on restoring 64 and enhancing habitat for target species such as the Golden-winged Warbler, Vermivora 65 chrysoptera (McNeil et al. 2020), Cerulean Warbler, Setophaga cerulea (Shaffer et al. 2025), 66 Greater Sage Grouse, Centrocercus urophasianus (Naugle et al. 2024), and Northern Bobwhite, 67 Colinus virginianus (Burger Jr. et al. 2006). Several successful state-led and partnership-led 68 wildlife conservation programs that mimic aspects of the NRCS-Environmental Quality 69 Incentive Program have also been developed in the past 10-15 years, such as the Oaks and 70 Prairies Joint Venture’s Grassland Restoration Incentive Program (GRIP) which has been 71 adopted by several Migratory Bird Joint Ventures (Giocomo et al. 2017). Additionally, many 72 conservation groups, including American Bird Conservancy (“Bird-Friendly Working Lands” 73 n.d.), National Audubon Society (“Conservation Ranching | Audubon” 2025), BirdLife 74 International (“Biodiversity - Grassland Alliance” n.d.), and the Smithsonian Institution 75 (“Virginia Working Landscapes” 2025) have promoted conservation practices on farms and 76 ranches to enhance habitat for birds. One obstacle to scaling this work is the need for metrics that 77 can quantify the conservation benefits after management intervention. Monitoring is critical to 78 gauging program effectiveness and modifying practice implementation. Easy-to-interpret metrics 79 that move beyond simple estimates of species occurrence are particularly desirable, since they 80 could catalyze further conservation investment, including from non-philanthropic sources 81 (Rodewald et al. 2020). While there is a rich suite of literature and countless cases employing 82 traditional sources of monitoring such as distance sampling (Buckland et al. 2001), many 83 researchers, companies and NGOs are working to develop tools to provide biodiversity metrics 84 that cost-effectively link monitoring at scale with emerging technologies such as LiDAR 85 (McNeil et al. 2023), eDNA (Beng & Corlett 2020), camera traps (Cordier et al. 2022), and 86 passive acoustic monitoring paired with machine learning classifiers like BirdNET (Wood et al. 87 2022; Larkin et al. 2024). There is great potential in these technologies, and while such 88 approaches often incorporate an element of the conservation significance of detected species, the 89 default has been to use coarse threat assessments to guide these valuations (Mair et al. 2021). 90 From a quantitative and practical perspective, however, coarse, ordinal assessments like 91 the Red List neither offer the resolution, nor express the diversity of relevant conservation 92 considerations that some practitioners desire. Often, more nuanced assessments are available, but 93 their regionally limited nature can hamper efforts to extend these approaches to larger spatial 94 scales (Michel et al. 2020). Species-level factors that have emerged as valuable indicators of 95 conservation priority, beyond threat status, include: ecological function (Díaz et al. 2013; Asner 96 et al. 2017), phylogenetic uniqueness (Faith 1992), and the identification of characteristics that 97 can predispose species to adverse outcomes, such as long-distance migration (Saunders et al. 98 2025) or human exploitation (Ripple et al. 2016). While some indices have endeavored to 99 incorporate elements of these factors, e.g., phylogenetic uniqueness (Isaac et al. 2007), none 100 have sought to derive a comprehensive, integrated conservation score that incorporates a wide 101 variety of such factors. Moreover, while the IUCN Red List is globally authoritative, other, 102 regionally or taxonomically limited population assessments are available that provide a more 103 fine-grained understanding of population status and trends (Fink et al. 2021; SoIB 2023; 104 Kittelberger et al. 2023; Partners in Flight 2024). Harmonizing these disparate data sources has 105 the potential to produce a comprehensive conservation score at the global level that incorporates 106 not only other key factors of potential relevance, but more detailed population trends when 107 available. 108 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint Critical features to include in generating a conservation valuation score are that it should 109 be data-driven, easy to update, and values must be available for all species in question (Burgess 110 et al. 2024). Moreover, it should incorporate species’ characteristics that the research community 111 has identified as important, such as phylogenetic and functional uniqueness (Eisenhauer et al. 112 2023). While no score will ever be perfect, and there is inherent subjectivity in such an exercise, 113 data-driven approaches can be more objective, easier to modify and update, and greatly facilitate 114 comparison between sites. Birds, in particular, are well-suited to such an approach, owing to 115 their rapid responses to environmental change and habitat specificity (Stillman et al. 2025), well-116 resolved taxonomy (Rheindt et al. 2025) and phylogeny (McTavish et al. 2025), and the long-117 standing, substantial contributions of citizen scientists to understanding their distributions at a 118 global scale (Sullivan et al. 2014). We used these convenient properties to develop a 119 conservation score for the world’s birds that is intended to value more than just species richness. 120 The factors used to generate these scores include three broad categories that are combined to 121 produce a final value: conservation status, vulnerability to known threats, and ecological and 122 phylogenetic uniqueness. We call these values BirdsPlus species scores. The name is meant to 123 indicate that biodiversity is multifaceted; we focus on birds as indicators, and we include 124 elements beyond species richness, with the intention to help quantify ecological integrity and 125 restoration progress, amongst other things. Here we outline the methods taken to create these 126 scores, and we make the scores and associated computer code freely available to the 127 conservation community. 128 129

130 131 Source of phylogeny 132 A key resource we used to create BirdsPlus scores is a comprehensive global avian phylogeny. 133 The phylogeny and its underlying taxonomy served three critical functions: facilitating the 134 alignment and integration of disparate datasets, calculation of phylogenetic uniqueness indices, 135 and phylogenetic imputation of missing data. We used the clootl package 136 (https://github.com/eliotmiller/clootl) to access the most recently released global avian 137 phylogeny (McTavish et al. 2025). This is v1.4 of the phylogeny, and we output it in the 2023 138 Clements taxonomy, which includes 11,017 species-level bird taxa. 139 140 Source of conservation assessments 141 We incorporated three sources of conservation assessment data. First, we used the IUCN Red 142 List, which uses BirdLife taxonomy (Burfield et al. 2017). Thus, to assemble these scores, we 143 first used IUCN scores that had previously been mapped to the 2023 Clements taxonomy (Birds 144 of the World, downloaded on 31 May 2024). Due to taxonomic mismatches, 810 species in 145 Clements are considered as unassessed by the IUCN. These mismatches tend to be fairly subtle 146 issues related to taxon concepts (Lepage et al. 2014). For example, American Robin (Turdus 147 migratorius, including subspecies confinis) does not have an IUCN status in Clements taxonomy 148 because BirdLife recognizes San Lucas Robin (T. confinis) as a separate species, and has 149 assessed each of these two taxonomic concepts independently. To recover some of these IUCN 150 assessments (which, undisputedly relate to imperfectly matched taxon concepts), we downloaded 151 the most recent version of IUCN Red List on 28 November 2024, and incorporated the 152 assessments of 520 species with perfectly matching scientific names in both BirdLife and 153 Clements taxonomies. IUCN assessments are ordinal, ranging from “Least Concern” to 154 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint “Critically Endangered”. We converted these to a numeric system as outlined in Table 1, and the 155 missing species were imputed following methods detailed below. 156 The second source of conservation assessment data was the Avian Conservation 157 Assessment Database (ACAD) (Partners in Flight 2024). These are expert-assessed threat levels 158 (Combined Conservation Score, or CCS) that range from 4 to 20 across species, and that are 159 assessed separately for breeding and non-breeding seasons for migratory birds. Of these, we 160 extracted the CCS-max scores from the database, which are, per-species, the higher of the two 161 scores. We used a version of the database we obtained from the Birds of the World on 15 162 November 2024, which had already been updated to Clements 2024 taxonomy. We converted 163 this to 2023 Clements taxonomy, which resulted in 1,610 species being assigned ACAD scores. 164 The third and final source of conservation assessment data we used were eBird trend 165 estimates. We obtained these using the ebirdst (Strimas-Mackey et al. 2023) package on 24 166 November 2024. Specifically, we identified all species with available trend estimates, 167 downloaded the data, and derived species-level average, abundance-weighted trends. This was 168 calculated as the sum of the abundance estimate for each 27km2 grid cell multiplied by its 169 percent per year trend, divided by the total abundance of the species across all grid cells ((grid 170 cell abundance * ppy)/total abundance across all cells). 171 172 Source of vulnerability factors 173 We included six separate measures of species’ vulnerability factors, albeit with some redundancy 174 amongst these. Species with small geographic ranges are frequently considered at increased risk 175 of extinction, and it was therefore important to us to include measures of range size. 176 Furthermore, this measure doubles as a proxy for endemism, e.g., to small islands, which is 177 another factor frequently considered in conservation decision making. We included two measures 178 of range size. The first was the area of the BirdLife range maps, i.e., range polygons (Murali et 179 al. 2021). This dataset was available in an older BirdLife taxonomy (10,372 species total in the 180 dataset), and after matching with taxonomic crosswalks available through the A VONET project 181 (Tobias et al. 2022), we were able to retain 9,686 of these range estimates. The second source of 182 range size data we used was based on modeled species ranges from eBird data, again obtained 183 via the ebirdst package. Here, we first identified all species with available population status 184 estimates, then downloaded the data at 27km2 resolution. For migratory species, we focused only 185 on their breeding season ranges. Then we removed any 27km2 cells where the species was not 186 inferred to occur at all, and took the remaining number of cells as an estimate of range size. 187 Species that are locally rare are also likely at increased risk of extinction, and thus we 188 included a measure of average local abundance as another vulnerability factor. Here, we took the 189 species-level median of the modeled abundance from the eBird breeding and resident grid cells 190 described above (Strimas-Mackey et al. 2023). 191 It is well documented that some species show a higher tolerance for anthropogenic 192 impacts than others, and we wanted to include a measure of human tolerance in the BirdsPlus 193 species score. A recently published dataset (Marjakangas et al. 2024) provides such an estimate, 194 derived from eBird data and modeled species’ occurrences. We used species’ median 195 “conservative HTI” (human-tolerance index) scores, retaining only those estimates where the 196 scientific name from the dataset directly matched that in Clements 2023 taxonomy (n = 5,985). 197 Migration has been shown to be the most dangerous part of many species’ life cycles 198 (Rushing et al. 2017; Saunders et al. 2025), and there is mounting evidence that long-distance 199 migrants are faring particularly poorly in today’s world (Rosenberg et al. 2019). As such, it was 200 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint critical that we include estimates of migration distance. We used two datasets to derive these 201 measures. The first came from (Dufour et al. 2020), specifically the distance_quanti_ALL 202 column, and after taxonomic matching, was available for 10,228 species. Like the first of our 203 range size estimates, these values were derived directly from BirdLife range maps, with added 204 information from reference handbooks (del Hoyo et al. 2017); the column we used corresponds 205 to a migration distance in kilometers for all species, including nominal residents. We further 206 derived our own estimates of migration distance from eBird data. To do so, on 5 May 2023, we 207 downloaded all available eBird data from complete checklists that were neither longer than 12 208 hours in duration nor 20 km in length. We then grouped these checklists into h3 grid cells 209 (https://h3geo.org/) by month of the year, using a resolution of 6, which corresponds to cells of 210 36km2. Then, per cell-month combination, we found the proportion of total individuals from 211 each of the detected species, and we used these proportions as weights in the following step. 212 Here, per species per month, we identified its geographic centroid as the weighted average of 213 latitude and longitude, after converting to Cartesian coordinates. Finally, we calculated the 214 Haversine distances between each centroid in successive months from January through January 215 of the following year, resulting in 12 distances (usually), which we summed for a final migration 216 distance estimate. We included species which had been detected in at least 10 months of the year, 217 and did not account for any potential movements during these unobserved months (most such 218 data-limited species are tropical and show limited seasonal movements). This resulted in 9,559 219 species with available eBird-based migration distance estimates. 220 221 Source of phylogenetic and functional uniqueness factors 222 A species’ independent evolutionary history—the amount of time it has been evolving 223 independently from extant relatives—is another factor that has been suggested as a prioritization 224 tool for conservation (Isaac et al. 2007). We used the phylogeny to calculate species’ edge scores 225 (but without incorporating IUCN status in the calculation) as our measure of “global” 226 phylogenetic uniqueness (n = 11,017). In truth, phylogeny is often incorporated in metrics such 227 as this because it can be a proxy for actual functional diversity and divergence. In our case, 228 however, we were also able to leverage the A VONET dataset (Tobias et al. 2022), a 229 comprehensive morphological database of the worlds’ birds. To derive a measure of global 230 functional uniqueness from morphology, we log-transformed species-level values for: beak 231 length (both from the mechanical hinge as well as from the nares), width, and depth; tarsus 232 length; primary wing length; secondary wing length; tail length; and mass. We then ran a 233 principal components analysis after scaling and centering. Finally, we found the Mahalanobis 234 distance between each species’ point in multivariate space and the centroid of that total space, 235 after accounting for the covariance among the points therein (n = 10,349 species). 236 We also derived measures of phylogenetic and functional uniqueness that reflect a 237 species’ role in the communities it occurs in. This can be thought of as a coarse measure of 238 phylogenetic or ecological redundancy, where species with large values are unique in comparison 239 to those around them. To derive these measures, we lumped the raw eBird data we downloaded 240 for migration distance calculations into h3 grid cells of 36km2. For each species i, we identified 241 all grid cells j in which it was observed. We excluded grid cells with fewer than three species 242 reported. In each remaining grid cell, we computed the weighted mean distance (phylogenetic or 243 morphological) between species i and all co-occurring species k ≠ i, where the weights were the 244 number of checklists in which each species k was reported within that grid cell. Then, for species 245 i, we computed a final weighted average across all included grid cells, where the weights were 246 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint the proportion of checklists in each grid cell j in which species i was reported. Weights at both 247 steps reflected local species presence (for co-occurring species k) and focal species frequency 248 (for species i across cells), which ensured that both community context and reporting effort were 249 accounted for. While to our knowledge this is an entirely novel approach, it shares close 250 similarities both to cell-level mean pairwise phylogenetic and functional distance approaches 251 (Miller et al. 2016), and to species-level “fields” (Villalobos et al. 2013). We used these final 252 species-level weighted values as measures of regional phylogenetic and functional uniqueness (n 253 = 10,564 species). 254 255 Normalizing data 256 To reduce the impact of outliers on the BirdsPlus species scores, and particularly to improve the 257 performance of data imputation procedures described below, we log transformed range size, 258 abundance, migration distance, global phylogenetic and functional uniqueness scores, and 259 regional morphological uniqueness to better approximate normal distributions. 260 261 Phylogenetic imputation 262 We used Rphylopars (Goolsby et al. 2017) to impute missing data. Prior to imputation, every 263 species had at least one known value, and the 25%, 50%, and 75% quantiles for missing columns 264 per species were 4, 4, and 5, respectively (out of 22 columns total). We used an Ornstein-265 Uhlenbeck model of evolution for the data imputation procedure. This model allows variance to 266 accumulate with random drift, but includes a term to pull the imputed traits back towards the 267 mean. The structure of the phylogeny (specifically the variance-covariance matrix) is used to 268 determine how closely missing species’ traits match their known ancestors’. 269 270 Aggregation into a final BirdsPlus species score 271 After imputing missing data, we scaled all measures to range from 0 to 1, and then inverted 272 scores for trend, range size, human tolerance, and abundance, such that after inversion species 273 with scores of 1 for these measures were of the greatest conservation concern. Then we 274 established weights for each factor (conservation status, vulnerability factors and uniqueness), 275 summarized in Table 2, and used these to derive a species-level weighted average for each factor. 276 Importantly, however, for measures which were imputed, we multiplied the nominal weight by 277 0.05, such that it only contributed 5% of its “ideal” weight from Table 2 towards a given factor. 278 Finally, per-species, we summed each of the three factor scores to arrive at a total BirdsPlus 279 species score. 280 281

282 283 We derived complete BirdsPlus scores for all 11,017 species of birds in the 2023 Clements 284 taxonomy (Fig. 1, Table S1). These scores are normally distributed and range between 0.58 and 285 2.72. Depending on the application, users may wish to transform these scores, for example by 286 cubic transformation, to place particular emphasis on species of conservation concern. The code, 287 raw data, and full, imputed dataset are all freely available at the links indicated at the end of the 288 manuscript. 289 While most of the constituent measures have been published elsewhere, we developed 290 and used three entirely novel measures. Our new, data-driven method of calculating migration 291 distance is available for 9,559 species. The longest distance migrants are all seabirds, while the 292 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint vast majority of species migrate very little if at all. There is a strong correspondence between our 293 new measure of migration distance and a previously published measure (r2 = 0.48, p < 0.001 294 (Dufour et al. 2020)). 295 Our regional phylogenetic and functional uniqueness values revealed that species which 296 are closely related to those they co-occur with include members of the families Pellorneidae, 297 Leiothrichidae, and Timaliidae, as well as some seabirds like Pycroft's Petrel (Pterodroma 298 pycrofti). In contrast, those which are phylogenetically unlike others they co-occur with tend to 299 be ratites such as kiwis (Apteryx spp.) and tinamous (Tinamidae). Similarly, species which are 300 morphologically akin to others they co-occur with tend to be seabirds, whereas the most 301 functionally irreplaceable species are such things as kiwis, hornbills (Bucerotidae), Shoebill 302 (Balaeniceps rex), pelicans (Pelecanidae), large storks (Ciconiidae), and unique island endemics 303 such as Mayr’s Swiftlet (Aerodramus orientalis), Inagua Woodstar (Nesophlox lyrura) and other 304 Caribbean Trochilinae, and the Vampire Ground-Finch (Geospiza septentrionalis). These 305 regional uniqueness values, as well as the h3 grid cell bird assemblages used to derive them, are 306 included as supplemental materials. 307 308

309 310 Here, we developed an approach for integrating a broad set of factors, from ecological function 311 to current population status, into a unified, continuous score intended to summarize a species’ 312 conservation value in a single number. Conservation practitioners, researchers, policy makers, 313 and funders alike have a need for such assessments. Decades of use of the IUCN Red List for 314 everything from reserve prioritization (Mair et al. 2021) to incorporation in emerging global 315 biodiversity frameworks (Burgess et al. 2024) corroborate this need for additional data-driven 316

to guide conservation. Yet, Red List assessments are coarse and ordinal (as opposed to 317 continuous), which limits some downstream uses. Regionally focused and more detailed 318 conservation assessments have been produced, and a variety of other factors of conservation 319 importance can be readily assembled from available databases, but a global project of this scale 320 is currently lacking. Our approach is transparent and dynamic, and while BirdsPlus species 321 scores also reflect our own scientific expertise and judgement, scores are open to easy 322 modification by others; they can be easily recalculated as new information becomes available 323 and taxonomy changes over time. 324 Our primary purpose in developing these scores has been to provide a quantitative means 325 of valuing site-level biodiversity in a more meaningful way than simple species richness. For 326 example, while the number of species detected from a single point in secondary tropical forest 327 can rival that of primary forests, the species inhabiting these secondary habitats tend to be 328 ecological generalists (Hughes et al. 2020), tolerant of human disturbance (Powell et al. 2015), 329 and to have more secure populations than those inhabiting primary forests. As a result, in order to 330 understand the improved outcomes for biodiversity after conservation management, practitioners 331 often focus on indicator or target species (Kessler et al. 2011). This is certainly a viable 332 approach, but ours is intended to provide an alternative, more comprehensive, and more granular 333 data-driven means of quantifying improved outcomes such as those seen in shade-grown versus 334 sun-grown coffee plantations, for example. 335 To illustrate how BirdsPlus species scores behave across species with differing ecological 336 roles and a range of conservation statuses, we explore some comparative examples here. 337 Cerulean Warbler (Setophaga cerulea), a declining, long-distance Neotropical migrant, has a 338 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint final BirdsPlus score of 1.47, which is in the 81st percentile of all birds globally. In comparison, 339 the closely related Tropical Parula (S. pitiayumi), a largely resident tropical species with a much 340 larger breeding range that co-occurs broadly in the non-breeding season with Cerulean Warbler, 341 has a final score of 0.97, which is in the 10th percentile of all species globally. Cubing these 342 scores, which emphasizes the significance of species of conservation concern (i.e. intentionally 343 introducing some skewness to an otherwise normal distribution), yields scores of 3.15 and 0.91, 344 respectively. Such transformations may be desirable in many cases. At any rate, these differences 345 in final scores are driven in particular by Cerulean Warbler’s increased vulnerability to human 346 impact—the species ranks in the 92nd percentile of all species in terms of migration distance. 347 Cerulean Warbler also has an elevated conservation status, and it is considered near threatened 348 by the IUCN, has an ACAD score of 14 (as compared to Tropical Parula’s 10), and is declining 349 across its range, whereas Tropical Parula is generally increasing throughout its range, particularly 350 in Mexico (Fink et al. 2023; Strimas-Mackey et al. 2023). As another comparison, while there 351 are many Neotropical species with higher BirdsPlus scores (Tachira Antpitta, Grallaria chthonia, 352 for example), when focusing only on those with little to no imputed inputs, the highest ranked is 353 Montezuma Quail (Cyrtonyx montezumae, 1.87), while the lowest ranked are Yellow-winged 354 Tanager (Thraupis abbas, 0.71) and Palm Tanager (T. palmarum, 0.76). Although classified as 355 least concern by the IUCN, Montezuma Quail is a locally uncommon species native to perennial 356 grasslands and oak woodlands of southwestern United States and northern Mexico—plant 357 communities that are routinely overgrazed and also subject to increasing drought—and where it 358 is often hunted and is steeply declining (Stromberg et al. 2020; Fink et al. 2021). In contrast, 359 Palm Tanger (and the closely related Yellow-winged Tanager) is widespread, often found in 360 heavily disturbed environments, and accordingly increasing throughout most of its range (Hilty 361 2020; Fink et al. 2021); it has a low ACAD score and is ecologically and phylogenetically 362 similar to many other species it co-occurs with. 363 There is a reasonably close correlation between our new measure of migration distance 364 and a previously published measure (Dufour et al. 2020), as well as tracking data for specific 365 species, e.g., Cerulean Warbler (Raybuck et al. 2022), although neither is a real measure of 366 individual-based migration distance as can be derived from more direct tracking approaches. 367 Notably, our new measure, which does not treat a species’ range as an evenly filled polygon and 368 which uses monthly measurements as opposed to coarse seasonal shifts, appears to address some 369 shortcomings in the range map approach. For example, the three species with the largest positive 370 residuals from a model of the new migration measure as a function of the previous are Peregrine 371 Falcon (Falco peregrinus), Blue Seedeater (Amaurospiza concolor), and Salmon-crested 372 Cockatoo (Cacatua moluccensis), all of which have a migration distance of 0 in the previous 373 measure, but which are inferred to migrate fairly large distances with our new measure. Because 374 some populations of Peregrine Falcon are well known to migrate large distances, and the Blue 375 Seedeater is a nomadic bamboo specialist assumed to move seasonally (García et al. 2023), these 376 shifts suggest increased resolution with our new measure. Salmon-crested Cockatoo is an IUCN 377 vulnerable species restricted in its native range to the South Moluccas, but escaped cage birds are 378 present in Singapore and Hawaii, amongst other regions, and recent changes in eBird now allow 379 these to be reported; this change presumably affected our results, and in the future, we will 380 exclude escapees from the analysis. The three species with the largest negative residuals from the 381 model are Coiba Spinetail (Cranioleuca dissita), Arfak Catbird (Ailuroedus arfakianus), and 382 Perija Antpitta (Grallaria saltuensis), all of which have limited geographic ranges and do not 383 migrate any substantial distances according either to the literature (Billerman et al. 2020) or to 384 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint our new measure, but which have fairly large migration distances according to the previous 385 measure. Again, these shifts suggest that our new migration values are a significant improvement 386 on previously available measures. 387 Potential future developments with BirdsPlus scores that warrant additional investigation 388 include incorporation of measures of ecological and cultural significance. Certain species like the 389 Greater Adjutant (Leptoptilos dubius, (Barman et al. 2020)) and Philippine Eagle (Pithecophaga 390 jefferyi, (Panopio et al. 2021)), for example, carry tremendous cultural significance, and it may 391 be worthwhile to provide a quantitative boost to convey such significance. That said, in these 392 specific cases at least, both species already have BirdsPlus scores in the 96th-99th percentile of all 393 species globally. Relatedly, certain species, such as those that excavate nesting cavities (Bednarz 394 et al. 2004) or lead mixed flocks (Zou et al. 2018), offer ecosystem services that may go above 395 and beyond the values currently captured by the evolutionary and functional uniqueness aspects 396 of BirdsPlus species scores, and it might be desirable to capture this significance in the scores in 397 the future. Comprehensive trait databases that convey this significance in a quantitative and 398 continuous way, however, are generally lacking, though some progress has been made (Schuetz 399 & Johnston 2019; Mittermeier et al. 2021). Other planned future developments include the 400 incorporation of additional regional assessments, such as the State of India’s Birds (SoIB 2023) 401 and, ideally, regional variation in species scores, such that invasive species could be 402 downweighted outside their native ranges, and key target species could be upweighted where 403 restoration efforts have been implemented to help conserve them. Last, continuing to shift from 404 broad-stroke IUCN range maps, which frequently overestimate range size (Ramesh et al. 2017), 405 to those that account for actual area of habitat (Brooks et al. 2019), or modeled occurrence when 406 data permits (Fink et al. 2021), is a central goal of our approach going forward. 407 BirdsPlus species scores (Table S1) provide a simple and intuitive way of conveying the 408 conservation and ecological significance of each of the world’s birds. While there is no doubt 409 conservation is a multidimensional process, and no single score will ever perfectly capture the 410 variety of relevant considerations, there is value in synthesizing existing databases to more 411 accurately reflect the breadth of factors that go into prioritizing species and regions. Simply put, 412 global assessments such as the IUCN Red List have a proven demand, and are rapidly finding 413 their way into globally significant monitoring and verification methodologies (Burgess et al. 414 2024), but in practice there is extremely limited resolution in these rankings, such as in 415 applications on working lands that might not harbor globally threatened species but provide 416 critical habitat for regionally important species. BirdsPlus scores build upon these assessments 417 by incorporating additional, more detailed regional assessments and factors such as ecological 418 uniqueness and vulnerability factors to arrive at a single score that should be of use both in 419 similar applications to Red List assessments, and particularly to emerging metrics intended to 420 measure progress towards global biodiversity targets. 421 422

423 Asner GP, Martin RE, Knapp DE, Tupayachi R, Anderson CB, Sinca F, Vaughn NR, Llactayo W. 424 2017. Airborne laser-guided imaging spectroscopy to map forest trait diversity and guide 425 conservation. Science 355:385–389. 426 Barman PD, Sharma DK, Cockrem JF, Malakar M, Kakati B, Melvin T. 2020. Saving the Greater 427 Adjutant Stork by Changing Perceptions and Linking to Assamese Traditions in India. 428 Ethnobiology Letters 11:20–29. 429 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint Bednarz JC, Ripper D, Radley PM. 2004. Emerging Concepts and Research Directions in the 430 Study of Cavity-Nesting Birds: Keystone Ecological Processes. The Condor 106:1–4. 431 Beng KC, Corlett RT. 2020. Applications of environmental DNA (eDNA) in ecology and 432 conservation: opportunities, challenges and prospects. Biodiversity and Conservation 433 29:2089–2121. 434 Bennett RE, Leuenberger W, Bosarreyes Leja BB, Sagone Cáceres A, Johnson K, Larkin J. 2018. 435 Conservation of Neotropical migratory birds in tropical hardwood and oil palm 436 plantations. PLoS ONE 13:e0210293. 437 Billerman SM, Keeney BK, Rodewald PG, Schulenberg TS. 2020. Birds of the World. Cornell 438 Laboratory of Ornithology, Ithaca, NY , USA. 439 Biodiversity - Grassland Alliance. (n.d.). Available from 440 https://alianzadelpastizal.birdlife.org/en/biodiversity/ (accessed April 22, 2025). 441 Bird-Friendly Working Lands. (n.d.). Available from https://abcbirds.org/program/latin-america-442 caribbean/bird-friendly-working-lands/ (accessed April 22, 2025). 443 Brooks TM et al. 2019. Measuring Terrestrial Area of Habitat (AOH) and Its Utility for the 444 IUCN Red List. Trends in Ecology & Evolution 34:977–986. Elsevier. 445 Buckland ST, Anderson DR, Burnham KP, Laake JL, Borchers DL, Thomas L. 2001. 446

to distance sampling: estimating abundance of biological populations. 447 Oxford University Press, Oxford. Available from https://academic.oup.com/book/53746 448 (accessed April 24, 2025). 449 Burfield IJ, Butchart SHM, Collar NJ. 2017. BirdLife, conservation and taxonomy. Bird 450 Conservation International 27:1–5. 451 Burger Jr. LW, Mckenzie D, Thackston R, Demaso SJ. 2006. The Role of Farm Policy in 452 Achieving Large-Scale Conservation: Bobwhite and Buffers. Wildlife Society Bulletin 453 34:986–993. 454 Burgess ND et al. 2024. Global Metrics for Terrestrial Biodiversity. Annual Review of 455 Environment and Resources 49:673–709. 456 Butchart SHM et al. 2025. Measuring trends in extinction risk: a review of two decades of 457 development and application of the Red List Index. Philosophical Transactions of the 458 Royal Society B: Biological Sciences 380:20230206. 459 Ciuzio E, Hohman WL, Martin B, Smith MD, Stephens S, Strong AM, Vercauteren T. 2013. 460 Opportunities and challenges to implementing bird conservation on private lands. 461 Wildlife Society Bulletin 37:267–277. 462 Conservation Ranching | Audubon. 2025, April 23. Available from https://www.audubon.org/our-463 work/prairies-and-forests/ranching (accessed April 22, 2025). 464 Convention on Biological Diversity. 2022. Kunming–Montreal Global Biodiversity Framework. 465 Page CBD/COP/DEC/15/4. Secretariat of the Convention on Biological Diversity, 466 Montreal, Canada. 467 Cordier CP, Ehlers Smith DA, Ehlers Smith Y , Downs CT. 2022. Camera trap research in Africa: 468 A systematic review to show trends in wildlife monitoring and its value as a research tool. 469 Global Ecology and Conservation 40:e02326. 470 del Hoyo J, Elliott A, Sargatal J, Christie DA, Kirwan G, editors. 2017. Handbook of the Birds of 471 the World Alive. Lynx Editions, Barcelona. Available from http://www.hbw.com 472 (accessed November 1, 2017). 473 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint Dertien JS, Baldwin RF. 2022. Importance of private working lands for the conservation of 474 neotropical birds on an urban-wilderness gradient. Global Ecology and Conservation 475 38:e02200. 476 Díaz S, Purvis A, Cornelissen JHC, Mace GM, Donoghue MJ, Ewers RM, Jordano P, Pearse 477 WD. 2013. Functional traits, the phylogeny of function, and ecosystem service 478 vulnerability. Ecology and Evolution 3:2958–2975. 479 Dufour P, Descamps S, Chantepie S, Renaud J, Guéguen M, Schiffers K, Thuiller W, Lavergne S. 480 2020. Reconstructing the geographic and climatic origins of long-distance bird 481 migrations. Journal of Biogeography 47:155–166. 482 Eisenhauer N, Hines J, Maestre FT, Rillig MC. 2023. Reconsidering functional redundancy in 483 biodiversity research. npj Biodiversity 2:1–4. 484 Faith DP. 1992. Conservation evaluation and phylogenetic diversity. Biological Conservation 485 61:1–10. 486 Fink D et al. 2021. eBird Status and Trends. Cornell Lab of Ornithology, Ithaca, NY . Available 487 from https://doi.org/10.2173/ebirdst.2020. 488 Fink D et al. 2023. A Double machine learning trend model for citizen science data. Methods in 489 Ecology and Evolution 14:2435–2448. 490 García NC, Ramos-Ordoñez MF, Rodríguez-Flores CI, Soberanes-González CA, Arizmendi M 491 del C, Boesman PFD. 2023. Blue Seedeater (Amaurospiza concolor), version 1.1. Birds 492 of the WorldDOI: 10.2173/bow.blusee1.01.1species_shared.bow.project_name. Cornell 493 Lab of Ornithology, Ithaca, NY , USA. Available from 494 https://birdsoftheworld.org/bow/species/blusee1/cur/introduction (accessed January 6, 495 2025). 496 Giocomo J, Hayes J, Gee K, Raasch J, Perez R. 2017. Strategic Habitat Conservation for 497 Declining Grassland Wildlife Populations in the Oaks and Prairies Joint Venture. National 498 Quail Symposium Proceedings 8. Available from 499 https://trace.tennessee.edu/nqsp/vol8/iss1/21. 500 Goolsby EW, Bruggeman J, Ané C. 2017. Rphylopars: fast multivariate phylogenetic 501 comparative methods for missing data and within-species variation. Methods in Ecology 502 and Evolution 8:22–27. 503 Hilty S. 2020. Palm Tanager (Thraupis palmarum), version 1.0. Page in del Hoyo J, Elliott A, 504 Sargatal J, Christie DA, de Juana E, editors. Birds of the World. Cornell Lab of 505 Ornithology, Ithaca, NY . Available from https://doi.org/10.2173/bow.paltan1.01. 506 Hughes EC, Edwards DP, Sayer CA, Martin PA, Thomas GH. 2020. The effects of tropical 507 secondary forest regeneration on avian phylogenetic diversity. Journal of Applied 508 Ecology 57:1351–1362. 509 Isaac NJB, Turvey ST, Collen B, Waterman C, Baillie JEM. 2007. Mammals on the EDGE: 510 Conservation Priorities Based on Threat and Phylogeny. PLOS ONE 2:e296. 511 IUCN. 2022. The IUCN Red List of Threatened Species. https://www.iucnredlist.org. 512 Kessler M et al. 2011. Cost-effectiveness of plant and animal biodiversity indicators in tropical 513 forest and agroforest habitats. Journal of Applied Ecology 48:330–339. 514 Kittelberger KD, Tanner CJ, Orton ND, Şekercioğlu ÇH. 2023. The value of community science 515 data to analyze long-term avian trends in understudied regions: The state of birds in 516 Türkiye. Avian Research 14:100140. 517 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint Larkin JT et al. 2024. A large-scale assessment of eastern whip-poor-will (Antrostomus 518 vociferus) occupancy across a gradient of forest management intensity using autonomous 519 recording units. Journal of Environmental Management 366:121786. 520 Lepage D, Vaidya G, Guralnick R. 2014. Avibase – a database system for managing and 521 organizing taxonomic concepts. ZooKeys:117–135. 522 Luther D et al. 2021. Conservation actions benefit the most threatened species: A 13-year 523 assessment of Alliance for Zero Extinction species. Conservation Science and Practice 524 3:e510. 525 Mair L et al. 2021. A metric for spatially explicit contributions to science-based species targets. 526 Nature Ecology & Evolution 5:836–844. 527 Marjakangas E-L, Johnston A, Santangeli A, Lehikoinen A. 2024. Bird species’ tolerance to 528 human pressures and associations with population change. Global Ecology and 529 Biogeography 33:e13816. 530 McNeil DJ, Fisher G, Fiss CJ, Elmore AJ, Fitzpatrick MC, Atkins JW, Cohen J, Larkin JL. 2023. 531 Using aerial LiDAR to assess regional availability of potential habitat for a conservation 532 dependent forest bird. Forest Ecology and Management 540:121002. 533 McNeil DJ, Rodewald AD, Ruiz-Gutierrez V , Johnson KE, Strimas-Mackey M, Petzinger S, 534 Robinson OJ, Soto GE, Dhondt AA, Larkin JL. 2020. Multiscale drivers of restoration 535 outcomes for an imperiled songbird. Restoration Ecology 28:880–891. 536 McTavish EJ, Gerbracht JA, Holder MT, Iliff MJ, Lepage D, Rasmussen PC, Redelings BD, 537 Sánchez Reyes LL, Miller ET. 2025. A complete and dynamic tree of birds. Proceedings 538 of the National Academy of Sciences 122:e2409658122. 539 Michel NL, Burkhalter C, Wilsey CB, Holloran M, Holloran A, Langham GM. 2020. Metrics for 540 conservation success: Using the “Bird-Friendliness Index” to evaluate grassland and 541 aridland bird community resilience across the Northern Great Plains ecosystem. Diversity 542 and Distributions 26:1687–1702. 543 Miller ET, Farine DR, Trisos CH. 2016. Phylogenetic community structure metrics and null 544 models: a review with new methods and software. Ecography 40:461–477. 545 Mittermeier JC, Correia R, Grenyer R, Toivonen T, Roll U. 2021. Using Wikipedia to measure 546 public interest in biodiversity and conservation. Conservation Biology 35:412–423. 547 Murali G, Gumbs R, Meiri S, Roll U. 2021. Global determinants and conservation of 548 evolutionary and geographic rarity in land vertebrates. Science Advances 7:eabe5582. 549 Naugle DE, Maestas JD, Morford SL, Smith JT, Mueller KR, Griffiths T, Heater T. 2024. From a 550 Bird to a Biome: Exploring the Sage Grouse Initiative’s Role in Defending and Growing 551 Sagebrush Core Areas. Rangeland Ecology & Management 97:115–122. 552 Panopio JK, Pajaro M, Grande JM, Dela Torre M, Raquino M, Watts P. 2021. Conservation 553 Letter: Deforestation—The Philippine Eagle as a Case Study in Developing Local 554 Management Partnerships with Indigenous Peoples. Journal of Raptor Research 55:460–555 467. 556 Partners in Flight. 2024. Avian Conservation Assessment Database, version 2024. Available from 557 http://pif.birdconservancy.org/ACAD (accessed January 4, 2025). 558 Powell LL, Wolfe JD, Johnson EI, Hines JE, Nichols JD, Stouffer PC. 2015. Heterogeneous 559 movement of insectivorous Amazonian birds through primary and secondary forest: A 560 case study using multistate models with radiotelemetry data. Biological Conservation 561 188:100–108. 562 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint Ramesh V , Gopalakrishna T, Barve S, Melnick DJ. 2017. IUCN greatly underestimates threat 563 levels of endemic birds in the Western Ghats. Biological Conservation 210:205–221. 564 Raybuck DW et al. 2022. Cerulean Warblers exhibit parallel migration patterns and multiple 565 migratory stopovers within the Central American Isthmus. Ornithological Applications 566 124:duac031. 567 Rheindt FE et al. 2025. AviList: a unified global bird checklist. Biodiversity and Conservation:1–568 18. 569 Ripple WJ et al. 2016. Bushmeat hunting and extinction risk to the world’s mammals. Royal 570 Society Open Science 3:160498. 571 Rodewald A, Arcese P, Sarra J, Tobin-de la Puenta J, Sayer J, Hawkins F, Martin T, Guy B, 572 Wachowicz K. 2020. Innovative Finance for Conservation Roles for Ecologists and 573 Practitioners. Issues in Ecology 22. 574 Rosenberg KV et al. 2019. Decline of the North American avifauna. Science 366:120–124. 575 Rushing CS, Hostetler JA, Sillett TS, Marra PP, Rotenberg JA, Ryder TB. 2017. Spatial and 576 temporal drivers of avian population dynamics across the annual cycle. Ecology 577 98:2837–2850. 578 Saunders SP et al. 2025. Multispecies migratory connectivity indicates hemispheric-scale risk to 579 bird populations from global change. Nature Ecology & Evolution:1–14. 580 Schuetz JG, Johnston A. 2019. Characterizing the cultural niches of North American birds. 581 Proceedings of the National Academy of Sciences 116:10868–10873. Proceedings of the 582 National Academy of Sciences. 583 Shaffer DR, McNeil DJ, Gray LF, Duchamp JE, Larkin JL. 2025. Factors driving minimal focal 584 species response to the implementation of habitat guidelines on private lands. Avian 585 Conservation and Ecology 20:9. 586 Simon N. 1966. Red Data Book. V olume 2. Aves. International Union for the Conservation of 587 Nature. 588 SoIB. 2023. State of India’s Birds, 2023: Range, trends, and conservation status. The SoIB 589 Partnership. 590 Stillman AN, Davis CL, Dunham KD, Ruiz-Gutierrez V , Rodewald AD, Johnston A, Auer T, 591 Strimas-Mackey M, Ligocki S, Fink D. 2025. A Framework for Assessing the Habitat 592 Correlates of Spatially Explicit Population Trends. Diversity and Distributions 593 31:e70030. 594 Strimas-Mackey M, Ligocki S, Auer T, Fink, Daniel. 2023. ebirdst: Access and Analyze eBird 595 Satus and Trends Data Products. Available from https://ebird.github.io/ebirdst/. 596 Stromberg MR, Montoya AB, Holdermann D. 2020. Montezuma Quail (Cyrtonyx montezumae), 597 version 1.0. Page in Schulenberg TS, editor. Birds of the World. Cornell Lab of 598 Ornithology, Ithaca, NY . Available from https://doi.org/10.2173/bow.higtin1.01. 599 Sullivan BL et al. 2014. The eBird enterprise: An integrated approach to development and 600 application of citizen science. Biological Conservation 169:31–40. 601 Tobias JA et al. 2022. A VONET: morphological, ecological and geographical data for all birds. 602 Ecology Letters 25:581–597. 603 Villalobos F, Rangel TF, Diniz-Filho JAF. 2013. Phylogenetic fields of species: cross-species 604 patterns of phylogenetic structure and geographical coexistence. Proceedings of the 605 Royal Society B: Biological Sciences 280:20122570. 606 Virginia Working Landscapes. 2025, April 14. Available from 607 https://www.vaworkinglandscapes.org/ (accessed April 22, 2025). 608 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint Wood CM, Kahl S, Rahaman A, Klinck H. 2022. The machine learning–powered BirdNET App 609 reduces barriers to global bird research by enabling citizen science participation. PLOS 610 Biology 20:e3001670. 611 Zou F, Jones H, Colorado Z. GJ, Jiang D, Lee T-M, Martínez A, Sieving K, Zhang M, Zhang Q, 612 Goodale E. 2018. The conservation implications of mixed-species flocking in terrestrial 613 birds, a globally-distributed species interaction network. Biological Conservation 614 224:267–276. 615 616 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint FIGURE AND TABLE LEGENDS Table 1. Framework for converting qualitative, ordinal IUCN Red List assessments into quantitative values for downstream analysis. To emphasize the large conceptual leap between species of least concern, which compose 78% of the global avifauna, and species with a more worrisome Red List category, we skip from zero to four when progressing from Least Concern to Vulnerable. Had we not done so, then species considered Vulnerable by the Red List would have been, counterintuitively, ranked closer to those considered Least Concern than those considered Endangered. IUCN category Ordinal score Least Concern 0 Vulnerable 4 Near Threatened 5 Endangered 6 Extinct 7 Critically Endangered 7 Extinct in the Wild 7 Data Deficient 5 Table 2. Values used in the weighted averaging step taken to arrive at final conservation, vulnerability, and uniqueness factor scores for each species. These weights are used to defer to preferred databases when the species in question is included in the database. For example, if an eBird trend is available for a given species, 50% of its final conservation score will be driven by its trend as quantified there. However, if the species in question is not included in the database, these weights are multiplied by 0.05, such that the phylogenetically imputed value only imparts 5% of its ideal weight to the final average. Factor Component Weight Percent of factor Conservation IUCN 1 16.6666667 Conservation ACAD 2 33.3333333 Conservation eBird Trend 3 50 Vulnerability BirdLife Range Size 1 6.25 Vulnerability eBird Range Size 3 18.75 Vulnerability eBird Abundance 4 25 Vulnerability Human Tolerance 4 25 Vulnerability BirdLife Migration Distance 1 6.25 Vulnerability eBird Migration Distance 3 18.75 Uniqueness Global phylogenetic uniqueness 1 0.25 Uniqueness Global functional uniqueness 1 0.25 .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint Uniqueness Regional phylogenetic uniqueness 1 0.25 Uniqueness Regional functional uniqueness 1 0.25 Figure 1. Genus-level average BirdsPlus scores illustrated on a phylogeny of the world’s birds, where genera of greater than average conservation concern, with larger BirdsPlus scores, are colored in red, and those with smaller scores are colored in blue. Clades of well-known conservation concern (including worrisome population statuses, exposure to vulnerability factors, or high ecological and evolutionary uniqueness) are clearly discernible here. Starting at .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint three o’clock, examples include: some grebes, such as Junin Grebe, Podiceps taczanowskii (copyright Andrew Spencer, Macaulay Library 382046181); many shorebirds, including Long- billed Curlew, Numenius americanus (copyright Jeff Dyck, Macaulay Library 561412711); many Galliformes, including Montezuma Quail, Cyrtonyx montezumae (copyright Doug Gochfeld, Macaulay Library 112164941); a number of hummingbirds, including Black-breasted Puffleg and many other Eriocnemis, E. nigrivestris (copyright Jay McGowan, Macaulay Library 611304791); most hornbills, including the Helmeted Hornbill, Rhinoplax vigil (copyright Ayuwat Jearwattanakanok, Macaulay Library 172593861); some raptors, such as the Crested Eagle, Morphnus guianensis (copyright Carlos Bran); a great many parrots (a clade that has also suffered more than its share of extinctions already), particularly macaws like the Hyacinth Macaw, Anodorhynchus hyacinthinus (copyright Greg Homel); a fair number of suboscine passerines, including Araripe Manakin, Antilophia bokermanni (copyright Ciro Albano), Long- wattled Umbrellabird, Cephalopterus penduliger (copyright Greg Homel), Sharp-tailed Tyrant, Culicivora caudacuta (copyright Thiago del Toledo e Silva), Recurve-billed Bushbird, Clytoctantes alixii (copyright Fundación ProAves), Cock-tailed Tyrant, Alectrurus tricolor (copyright Jose Carlos Motta-Junior); some bowerbirds, such as Golden Bowerbird, Prionodura newtoniana (copyright Jenny Stiles, Macaulay Library 493305361); a number of odd-looking, species-poor Old World families, including Picathartidae like the White-necked Rockfowl, Picathartes gymnocephalus (copyright Daniel López-Velasco | Ornis Birding Expeditions, Macaulay Library 425480471); the two species of enigmatic river martins (one presumed extinct), including African River Martin, Pseudochelidon eurystomina (copyright Lionel Sineux, Macaulay Library 627667036); the genus Laniellus, which includes Gray-crowned Crocias, Laniellus langbianis (copyright Ngoc Sam Thuong Dang, Macaulay Library 399374851); the genus Sholicola, including Nilgiri Sholikili, Sholicola major (copyright Honza Grünwald, Macaulay Library 620583993); nearly all of the Hawaiian honeycreepers (most already extinct), including Akohekohe, Palmeria dolei (copyright Jim Denny) and Anianiau, Magumma parva (copyright Peter LaTourrette); and some nine-primaried oscine passerines, especially those targeted for the cage bird trade like Yellow Cardinal, Gubernatrix cristata (copyright Dubi Shapiro, Macaulay Library 620559958). SUPPLEMENTARY MATERIALS Table S1. Complete BirdsPlus scores, including all constituent values, for the world’s bird species. These include both real and phylogenetically imputed values. As described in the main text, all values are scaled from 0 to 1 before taking a weighted average using the weights provided in Table 2. Those values which have been imputed can be determined from Table S2. Table S2. Raw, untransformed values used to derive species’ final BirdsPlus scores. If a cell contains an NA, then the relevant database did not contain a value for that taxon, and the final value in Table S1 was derived via phylogenetic imputation. Table S3. The final weights used, per species, to derive weighted average conservation status, vulnerability factor, and ecological and evolutionary uniqueness values. Ideal weights are provided in Table 2, but if values were unavailable (see Table S2), these weights have been multiplied by 0.05 to decrease their influence on the final factor-level averages. .CC-BY-NC-ND 4.0 International licenseavailable under a 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 The copyright holder for this preprint (whichthis version posted July 18, 2025. ; https://doi.org/10.1101/2025.07.15.664934doi: bioRxiv preprint