Species observation data contain variation due to differences in observers, whether these data were collected through structured protocols by experts or opportunistically by volunteer observers. Analyses of wildlife population data generated by participatory science projects require accounting for observer effects to reduce bias and heterogeneity introduced by observer differences. Species detections may vary not only among observers, but also within individual observers over time due to learning, senescence, or adoption of new tools, for example. Existing models of observer effects in participatory science data assume that species detection rates for each individual observer either do not change or increase uniformly as the observer gains experience. Here, we develop a new index to capture inter- and intra-observer effects in species detection rates in data from the global, decades-long participatory science project eBird. We test the response of this index to simulated within-observer changes in the numbers of species reported on eBird checklists. We then compare its performance in species distribution modeling to a more restrictive method of measuring observer effects in eBird data. The index flexibly captures diverse, nonlinear inter-annual changes in species detection rates of individual observers, while conferring better predictive performance when included in occurrence models for most species. As participatory science projects expand in scope and lifespan, it is increasingly critical to deploy flexible, data-driven approaches to account for complex observer effects in the scientific use of these rich, yet nuanced, datasets. center dot Scientific data contain variation due to differences in data collectors (observers).center dot Participatory science projects often recruit many observers, some of whom contribute for years or decades.center dot Long-lived projects experience shifts in observer populations and technological landscapes, injecting more variation over time.center dot Methodological advances have accounted for variance among observers, but they impose a simplistic form of within-observer change in species detection rates.center dot Building on prior work that devised an index of observer species detections, we devise a more flexible index to capture inter-annual changes in the number of species d ' etected by individual eBird observers.center dot Using this new index, we reveal both common and idiopathic patterns of observer change.center dot Quantifying observer change allows models of bird populations to adjust for complex dynamics in observer behavior. Further, it motivates future research into the drivers and consequences of observer-level variation in long-running participatory science projects. Los datos de observaci & oacute;n de especies contienen variaci & oacute;n debida a diferencias entre observadores, ya sea que estos datos hayan sido recolectados mediante protocolos estructurados por expertos o de manera oportunista por observadores voluntarios. Los an & aacute;lisis de datos poblacionales de fauna silvestre generados por proyectos de ciencia participativa requieren considerar los efectos del observador para reducir el sesgo y la heterogeneidad introducidos por las diferencias entre observadores. Las detecciones de especies pueden variar no solo entre observadores, sino tambi & eacute;n para un mismo observador a lo largo del tiempo debido, por ejemplo, al aprendizaje, la senescencia o la adopci & oacute;n de nuevas herramientas. Los modelos existentes sobre los efectos del observador en datos de ciencia participativa asumen que las tasas de detecci & oacute;n de especies para cada observador individual no cambian o aumentan de manera uniforme a medida que el observador adquiere experiencia. Aqu & iacute; desarrollamos un nuevo & iacute;ndice para capturar los efectos inter- e intra-observador en las tasas de detecci & oacute;n de especies utilizando datos del proyecto global de ciencia participativa eBird, con varias d & eacute;cadas de duraci & oacute;n. Evaluamos la respuesta de este & iacute;ndice a cambios simulados para un mismo observador en el n & uacute;mero de especies reportadas en listas de eBird. Luego comparamos su desempe & ntilde;o en modelos de distribuci & oacute;n de especies con un m & eacute;todo m & aacute;s restrictivo para medir los efectos del observador en los datos de eBird. El & iacute;ndice permite analizar de manera flexible los cambios interanuales diversos y no lineales en las tasas de detecci & oacute;n de especies para observadores individuales, y adem & aacute;s mejora el desempe & ntilde;o predictivo cuando se incluye en modelos de ocurrencia para la mayor & iacute;a de las especies. A medida que los proyectos de ciencia participativa se expanden en alcance y duraci & oacute;n, resulta cada vez m & aacute;s cr & iacute;tico implementar enfoques flexibles y basados en datos que permitan considerar los complejos efectos del observador en el uso cient & iacute;fico de estos sets de datos ricos, pero que tienen sus matices.

ABSTRACT Accurately quantifying wildlife population change is essential for assessing ecosystem health and guiding conservation action. Model‐based integration of structured and opportunistic survey data can expand the spatial scope and improve the precision of trend estimates but requires methods that reconcile differences in observation processes, sampling coverage, and data volume across surveys. We present the integrated R‐learner trend model, a modular framework for estimating spatially explicit trends from multiple surveys. In Stage 1, the double machine learning (DML) trend model is applied separately to each survey, isolating ecological signals from intra‐ and interannual confounding. In Stage 2, survey‐specific trends are combined using the R‐learner framework, with weights based on sampling coverage probabilities estimated via a spatially explicit mixture‐distribution sub‐model. This two‐stage design enables bias mitigation, reconciles differences in observation processes and coverage, handles high‐dimensional feature sets, and balances influence across datasets of differing size and quality. Using simulations, we show that incorporating mixture‐distribution weighting improves accuracy and statistical power relative to independent survey‐specific models. We then apply our model to data from the North American Breeding Bird Survey and the eBird participatory science project, demonstrating how complementary coverage from structured and opportunistic surveys can be leveraged to produce high‐resolution, spatially explicit bird population trends.

Fire regimes are context-dependent, as are the ways that animals respond. However, most information on animal responses to fire comes from short-term local field studies, which are hard to extrapolate across large areas for fire management while also capturing spatial variation. To address this challenge, we modeled data from eBird to map the direction, magnitude, and importance of fire regime associations at 27-km resolution across the ranges of six bird species used to guide management decisions in the US: red-cockaded woodpecker (Leuconotopicus borealis), Bachman's sparrow (Peucaea aestivalis), greater sage-grouse (Centrocercus urophasianus), pinyon jay (Gymnorhinus cyanocephalus), American goshawk (Astur atricapillus), and olive-sided flycatcher (Contopus cooperi). Our findings revealed previously undocumented landscape-scale variation in fire impacts on birds. Critically, the strength of fire regime associations varied widely in magnitude even when the direction of those associations (positive, neutral, or negative) remained constant. This analytical workflow provides not only a flexible approach for assessing macroecological fire impacts but also finer-scale information sufficient for resource prioritization and decision-making.

Efforts to address declines of North American birds have been constrained by limited availability of fine-scale information about population change. By using participatory science data from eBird, we estimated continental population change and relative abundance at 27-kilometer resolution for 495 bird species from 2007 to 2021. Results revealed high and previously undetected spatial heterogeneity in trends; although 75% of species were declining, 97% of species showed separate areas of significantly increasing and decreasing populations. Populations tended to decline most steeply in strongholds where species were most abundant, yet they fared better where species were least abundant. These high-resolution trends improve our ability to understand population dynamics, prioritize recovery efforts, and guide conservation at a time when action is urgently needed.

AimHalting widespread biodiversity loss will require detailed information on species' trends and the habitat conditions correlated with population declines. However, constraints on conventional monitoring programs and commonplace approaches for trend estimation can make it difficult to obtain such information across species' ranges. Here, we demonstrate how recent developments in machine learning and model interpretation, combined with data sources derived from participatory science, enable landscape-scale inferences on the habitat correlates of population trends across broad spatial extents.LocationWorldwide, with a case study in the western United States.MethodsWe used interpretable machine learning to understand the relationships between land cover and spatially explicit bird population trends. Using a case study with three passerine birds in the western U.S. and spatially explicit trends derived from eBird data, we explore the potential impacts of simulated land cover modification while evaluating potential co-benefits among species.ResultsOur analysis revealed complex, non-linear relationships between land cover variables and species' population trends as well as substantial interspecific variation in those relationships. Areas with the most positive impacts from a simulated land cover modification overlapped for two species, but these changes had little effect on the third species.Main ConclusionsThis framework can help conservation practitioners identify important relationships between species trends and habitat while also highlighting areas where potential modifications to the landscape could bring the biggest benefits. The analysis is transferable to hundreds of species worldwide with spatially explicit trend estimates, allowing inference across multiple species at scales that are tractable for management to combat species declines.

Confidently estimating population trends is of vital importance for a wide range of ecological, conservation, and management applications. North America has 2 major data sources for estimating population trends of breeding birds—the North American Breeding Bird Survey (BBS) and the global participatory science project eBird. Because the surveys differ in protocols, coverage, and data analysis, their trend estimates are expected to vary in magnitude, direction, and/or precision for at least some species and regions. Here, we compare independently derived estimates of population change between 2012 and 2022 for every combination of species and bird conservation region (BCR) covered by both surveys (n = 5,577 combinations) as well as aggregated across entire ranges or within the U.S. or Canada. Uncertainty was substantial for both surveys, though more prevalent for BBS (81% of credibility intervals for estimates included zero) than eBird (34% of confidence intervals overlapped zero). We found agreement of trend directions between the 2 surveys. Only 1.3% of estimated trends were significant in opposite directions between the 2 surveys for all species/BCR combinations, with the median difference in trend magnitude being –0.02% (BBS minus eBird trend). Correlations between the 2 were strongest for estimates that were graded as being high credibility compared to estimates judged to have medium or low credibility. Both surveys were subject to species, taxonomic, and regional effects that influenced agreement. Overall, we show where trend estimates derived from BBS and eBird agree, explore where they diverge, present several comparisons to assist in interpreting results from both surveys, and inform efforts to integrate information from each.

1. Citizen and community-science (CS) datasets have great potential for estimating interannual patterns of population change given the large volumes of data collected globally every year. Yet, the flexible protocols that enable many CS projects to collect large volumes of data typically lack the structure necessary to keep consistent sampling across years. This leads to interannual confounding, as changes to the observation process over time are confounded with changes in species population sizes. 2. Here we describe a novel modeling approach designed to estimate species population trends while controlling for the interannual confounding common in citizen science data. The approach is based on Double Machine Learning, a statistical framework that uses machine learning methods to estimate population change and the propensity scores used to adjust for confounding discovered in the data. Additionally, we develop a simulation method to identify and adjust for residual confounding missed by the propensity scores. Using this new method, we can produce spatially detailed trend estimates from citizen science data. 3. To illustrate the approach, we estimated species trends using data from the CS project eBird. We used a simulation study to assess the ability of the method to estimate spatially varying trends in the face of real-world confounding. Results showed that the trend estimates distinguished between spatially constant and spatially varying trends at a 27km resolution. There were low error rates on the estimated direction of population change (increasing/decreasing) and high correlations on the estimated magnitude. 4. The ability to estimate spatially explicit trends while accounting for confounding in citizen science data has the potential to fill important information gaps, helping to estimate population trends for species, regions, or seasons without rigorous monitoring data.

Artificial light at night (ALAN) and roads are known threats to nocturnally migrating birds. How associations with ALAN and roads are defined in combination for these species at the population level across the full annual cycle has not been explored.

Orin J. Robinson , Jacob B. Socolar , Erica F. Stuber , Tom Auer, Alex J. Berryman , Philipp H. Boersch-Supan , Donald J. Brightsmith , Allan H. Burbidge , Stuart H. M. Butchart , Courtney L. Davis , Adriaan M. Dokter , Adrian S. Di Giacomo, Andrew Farnsworth , Daniel Fink , Wesley M. Hochachka , Paige E. Howell , Frank A. La Sorte , Alexander C. Lees , Stuart Marsden, Robert Martin, Rowan O. Martin , Juan F. Masello , Eliot T. Miller, Yoshan Moodley , Andy Musgrove, David G. Noble, Valeria Ojeda , Petra Quillfeldt , J. Andrew Royle , Viviana Ruiz-Gutierrez, Jos e L. Tella , Pablo Yorio, Casey Youngflesh , and Alison Johnston

Aim Animal migration is often explained as the result of resource tracking in seasonally dynamic environments. Therefore, resource availability should influence both the distributions of migratory animals and their seasonal abundance. We examined the relationship between primary productivity and the spatio-temporal distributions of migratory birds to assess the role of energy availability in avian migration. Location North America. Time period Full annual cycle, 2011-2016. Major taxa studied Nocturnally migrating landbirds. Methods We used observations of nocturnally migrating landbirds from the eBird community-science programme to estimate weekly spatial distributions of total biomass, abundance and species richness. We related these patterns to primary productivity and seasonal productivity surplus estimated using a remotely sensed measure of vegetation greenness. Results All three avian metrics showed positive spatial associations with primary productivity, and this was more pronounced with seasonal productivity surplus. Surprisingly, biomass showed a weaker association than did abundance and richness, despite being a better indicator of energetic requirements. The strength of associations varied across seasons, being the weakest during migration. During spring migration, avian biomass increased ahead of vegetation green-up in temperate regions, a pattern also previously described for herbivorous waterfowl. In the south-eastern USA, spring green-up was instead associated with a net decrease in biomass, and winter biomass greatly exceeded that of summer, highlighting the region as a winter refuge for short-distance migrants. Main conclusions Although instantaneous energy availability is important in shaping the distribution of migratory birds, the stronger association of productivity with abundance and richness than with biomass suggests the role of additional drivers unrelated to energetic requirements that are nonetheless correlated with productivity. Given recent reports of widespread North American avifaunal declines, including many common species that winter in the south-eastern USA, understanding how anthropogenic activities are impacting winter bird populations in the region should be a research priority.

Aim Ecological data collected by the general public are valuable for addressing a wide range of ecological research and conservation planning, and there has been a rapid increase in the scope and volume of data available. However, data from eBird or other large-scale projects with volunteer observers typically present several challenges that can impede robust ecological inferences. These challenges include spatial bias, variation in effort and species reporting bias. Innovation We use the example of estimating species distributions with data from eBird, a community science or citizen science (CS) project. We estimate two widely used metrics of species distributions: encounter rate and occupancy probability. For each metric, we critically assess the impact of data processing steps that either degrade or refine the data used in the analyses. CS data density varies widely across the globe, so we also test whether differences in model performance are robust to sample size. Main conclusions Model performance improved when data processing and analytical methods addressed the challenges arising from CS data; however, the degree of improvement varied with species and data density. The largest gains we observed in model performance were achieved with 1) the use of complete checklists (where observers report all the species they detect and identify, allowing non-detections to be inferred) and 2) the use of covariates describing variation in effort and detectability for each checklist. Occupancy models were more robust to a lack of complete checklists. Improvements in model performance with data refinement were more evident with larger sample sizes. In general, we found that the value of each refinement varied by situation and we encourage researchers to assess the benefits in other scenarios. These approaches will enable researchers to more effectively harness the vast ecological knowledge that exists within CS data for conservation and basic research.

Maps are essential tools for communicating information about wildlife distributions in space and time. As observational datasets grow and enable description of distributions at broader spatiotemporal extents and finer spatiotemporal resolutions, new opportunities arise for visualizing when and where animals occur. Here we present a package for the r statistical computing environment, colorist, that facilitates visualization of animal distributions in space and time using raster inputs. In addition to enabling display of sequential change in distributions through the use of small multiples, colorist provides functions for extracting several features of interest from a sequence of distributions and for visualizing those features within an HCL (hue–chroma–luminance) colour space. Resulting maps allow for ‘fair’ visual comparison of occurrence, abundance or probability density values across space and time and can be used to address questions about where, when and how consistently a species or individual is likely to be found. Functions can also be harnessed to visualize distributions of multiple species or individuals within a single time period when research questions are focused on understanding the degree to which species or individuals partition space. By colouring temporal features of distributions, while simultaneously controlling their perceptual weight, we expand the set of tools available for exploring wildlife movements through space–time and for communicating research results. We expect colorist functions will prove useful for visualizing other types of multivariate data that we have yet to consider.

Information on species’ distributions and abundances, and how these change over time are central to the study of the ecology and conservation of animal populations. This information is challenging to obtain at relevant scales across range-wide extents for two main reasons. First, local and regional processes that affect populations vary throughout the year and across species’ ranges, requiring fine-scale, year-round information across broad — sometimes hemispheric — spatial extents. Second, while citizen science projects can collect data at these scales, using these data requires appropriate analysis to address known sources of bias. Here we present an analytical framework to address these challenges and generate year-round, range-wide distributional information using citizen science data. To illustrate this approach, we apply the framework to Wood Thrush ( Hylocichla mustelina ), a long-distance Neotropical migrant and species of conservation concern, using data from the citizen science project eBird. We estimate occurrence and relative abundance with enough spatiotemporal resolution to support inference across a range of spatial scales throughout the annual cycle. Additionally, we generate intra-annual estimates of the range, intra-annual estimates of the associations between species and the local environment, and inter-annual trends in relative abundance. This is the first example of an analysis to capture intra- and inter-annual distributional dynamics across the entire range of a broadly distributed, highly mobile species.

Limited knowledge of the distribution, abundance, and habitat associations of migratory species hinders effective conservation actions. We use Neotropical migratory birds as a model group to compare approaches to prioritize land conservation needed to support ≥30% of the global abundances of 117 species. Specifically, we compare scenarios from spatial optimization models to achieve conservation targets by: 1) area requirements for conserving >30% abundance of each species for each week of the year independently vs. combined; 2) including vs. ignoring spatial clustering of species abundance; and 3) incorporating vs. avoiding human-dominated landscapes. Solutions integrating information across the year require 56% less area than those integrating weekly abundances, with additional reductions when shared-use landscapes are included. Although incorporating spatial population structure requires more area, geographical representation among priority sites improves substantially. These findings illustrate that globally-sourced citizen science data can elucidate key trade-offs among opportunity costs and spatiotemporal representation of conservation efforts.

Recent studies have highlighted the threat that climate change poses to species, as areas of climatic suitability contract or shift across the landscape. North American Neotropical long-distant migrant bird species present a unique problem compared to sedentary species because climate change may differ significantly across their breeding and wintering grounds. Studying the potential future distributions of these birds is challenging on many levels, including the fact that our understanding of the wintering grounds of these species is quite poor. To address this issue, we analyse available eBird data during the winter season in the Western Hemisphere in an effort to further promote and direct citizen science efforts to focus on areas that are climatically undersampled. We used Mobility-Oriented Parity (MOP) to understand the areas where climates are most dissimilar from climates sampled by existing eBird checklists, creating a map that ranks the western hemisphere at a 10 km resolution for climatic sampling during the winter season. We found that parts of Mexico and Central America, areas of Colombia, almost the entire Amazon Basin, coastal Peru and Chile, and northern Argentina are climatically undersampled. As a test case, we then used the map of survey priorities to simulate additional sampling in Colombia and recalculated the rankings. Guiding additional sampling with the priorities reduced climate dissimilarities between sampled and unsampled grid cells more than when additional sampling expanded in proportion to current sampling efforts or based on geographic undersampling. Analyses of sampling coverage in environmental space, such as this, will be a useful tool for targeting monitoring effort for bird species.

Spatial prioritizations are essential tools for conserving biodiversity in the face of accelerating climate change. Uncertainty about species' responses to changing climates can complicate prioritization efforts, however, and delay conservation investment. In an effort to facilitate decision-making, we identified three hypotheses about species' potential responses to climate change based on distinct biological assumptions related to niche flexibility and colonization ability. Using 314 species of North American birds as a test case, we tuned separate spatial prioritizations to each hypothesis and assessed the degree to which assumptions about biological responses affected the perceived conservation value of the landscape and prospects for individual taxa. We also developed a bet-hedging prioritization to minimize the chance that incorrect assumptions would lead to valuable landscapes and species being overlooked in multispecies prioritizations. Collectively, these analyses help to quantify the sensitivity of spatial prioritizations to different assumptions about species' responses to climate change and provide a framework for enabling efficient conservation investment despite substantial biological uncertainty.

Diporeia, formerly the dominant benthic macroinvertebrate in the Great Lakes, remains a keystone species in Lake Superior. Little is known, however, about fine scale amphipod distributions, especially as influenced by the production, transport and transformation of energy resources. Here, we document the distribution and abundance of Diporeia along 19 transects around the lake's perimeter. Regions of elevated density, averaging 958±408Diporeia/m2 (mean±S.D.) were observed along all transects, typically within slope habitat (depth of 30–125m). Waters shoreward (shelf habitat, <30m) and lakeward (profundal habitat, >125m) of these regions supported significantly lower densities, averaging 239±178/m2 and 106±59/m2, respectively. Amphipods within regions of elevated density, termed here the Ring of Fire, account for two-thirds of the lakewide population while occupying only one-quarter of the benthic habitat. The Ring of Fire, observed lakewide as a band averaging 14.2±9.4km in width, is characterized as a region of transitional sediment deposition with gentle slope, proximate to nearshore locations of elevated primary production. Within the Ring of Fire exceptionally high densities are found in the south central region, where the Keweenaw Current and slope bathymetries serve to funnel production from adjoining regions of high production. Density measurements for the 173 stations sampled here are used to estimate lakewide Diporeia standing stock (22.5–37.7trillion individuals, 4.4–7.4Gg dry weight, 2.1–3.5Gg C), individual and biomass density (274–460/m2, 0.05–0.09gDW/m2, 0.03–0.04gC/m2) and areal (0.02–0.03gC/m2/yr) and total (1.6–2.6GgC/yr) production.