Keywords: climate change, species distributions, abundance shifts, non-linear species–environment models

Shifts in Species’ Latitudinal Optima Over Time.

The distribution ofp-values obtained from individual tests for a trend in latitudinal optimum ( $m$  ) over time for every species was clearly non-uniform (Fig. 1). Using a significance level of$\alpha$  = 0.05 (and assuming independent tests), we would expect 5% of thep-values to be less than 0.05 by chance alone. However, one-third of North American bird species (70 out of 209) demonstrated a significant shift in their latitudinal optimum (P < 0.05), with 61% of these shifting northward (Table 1; combined dataset). At a continental scale, the pooled estimated mean shift in latitudinal optimum across all species at the continental scale was 1.5 km northward per year (±0.58 SE), corresponding to a total distance moved, on average, of 82.5 km (±31.9 SE) over the last 55 y ( $m$  = 0.0134°, CI = {0.0033°, 0.0236°},P = 0.0095;Fig. 1A andSI Appendix, Table S1). In terms of absolute shift (i.e., regardless of the direction north or south), the estimated mean size of the latitudinal shift in absolute value was$\left|m\right|$ = 0.0463° per year (95% CI = {0.0309°, 0.0616°},P < 0.001;SI Appendix, Table S2).

For the subset of bird species in the combined dataset that shifted significantly northward, the overall estimate was$m_{\mathit{north}}$ = 0.1171° (CI = {0.0889°, 0.1452°},P < 0.001,Fig. 1A andSI Appendix, Table S3), corresponding to ca. 13 km/y (~715 km in 55 y). This was 1.6 times higher than the overall estimated shift for the subset of birds moving significantly southward ( $m_{\mathit{south}}$ =$-$0.0781°, CI = { $-$0.1432°,$-$0.0130°},P < 0.001,Fig. 1A andSI Appendix, Table S4), which was ca. 8.7 km/y (~478.5 km in 55 y).

At the regional scale, there was a higher proportion of significant northward shifts in latitudinal optima in the West and a higher proportion of significant southward shifts in the East (Table 1,${\chi }_1^2$ = 7.38,P < 0.01). For the West, the estimated average overall northward shift was$m_{\mathit{West}}$ = 0.0202° (CI = {0.0041°, 0.0362°},P = 0.0137,Fig. 1B andSI Appendix, Table S5), approximately 2.2 km/y. For the East, however, results were more variable across species, and the overall estimate did not differ significantly from zero (Fig. 1C andSI Appendix, Table S6).

For each species and data subset, we divided the rate of change for individual species ( $\beta$ ) by its SE, giving us a measure to identify species exhibiting stronger and more consistent latitudinal shifts (Dataset S3). Those with the highest$\beta /SE\left(\beta \right)$ ratios shifting northward were: Cassin’s Sparrow (Peucaea cassinii) in the East (60.02 ± 4.9 km/y), Brown Creeper (Certhia americana) (19.54 ± 2.50 km/y), and Eastern Wood-Pewee (C. virens) (6.93 ± 0.65 km/y). Those with the highest$\beta /SE\left(\beta \right)$ ratios shifting southward were Mississippi kite (Ictinia mississippiensis) (17.07 ± 2.19 km/y), northern harrier (Circus cyaneus) (12.68 ± 1.82 km/y), and Baltimore oriole (Icterus galbula) in the West (3.14 ± 0.40 km/y).

Significant shifts in latitudinal optima were detected for four out of the nine “near threatened” bird species examined in our study. In the East, the wood thrush (Hylocichla mustelina) displayed a significant southward shift of 5.54 ± 2.35 km/y, while significant northern shifts were detected for the common grackle (Quiscalus quiscula) (4.41 ± 1.13 km/y) and the loggerhead shrike (Lanius ludovicianus) (7.37 ± 1.12 km/y). In addition, significant northward shifts were observed for the eastern meadowlark (Sturnella magna) in both the West (6.72 ± 1.75 km/y) and in the East (8.93 ± 2.05 km/y).

Relationship between Shifts in Optima and Species’ Traits.

There was no qualitative difference between the model that included all bird species and the model excluding species that lacked precise estimates of thermal range. Therefore, the final continental-scale model was built using all predictor variables and all species (except for three species that had no values for SSI,Materials and Methods).

We found the absolute value of the shift in latitudinal optimum at the continental scale was significantly related to four traits: thermal optimum, territoriality, habitat specialization, and dispersal capability (SI Appendix, Table S7). However, the direction of the shift (northward vs. southward) was also significant, as were its interactions with species’ traits (SI Appendix, Table S7). Therefore, we conducted separate analyses for species that exhibited southward trends (SI Appendix, Table S8) and species that exhibited northward trends (SI Appendix, Table S9). For species showing a southward shift, the global test (across all predictors) was not statistically significant (seeSI Appendix, 3, although the individual predictor of the hand-wing index was marginally so, seeSI Appendix, Table S8). Therefore, we focus our attention here on the results of the model for species with northward trends (SI Appendix, Table S9).

At the continental scale, northward shifts in latitudinal optima ( $m$ ) per year were significantly related to (in order of importance): a bird species’ thermal optimum, its territoriality, and its degree of habitat specialization (SI Appendix, Table S9). Bird species with colder optimal temperatures had significantly higher northward annual shifts (Fig. 2B). More specifically, we estimated that the magnitude of their annual optimum shift decreased by −0.003 latitudinal degrees (CI = {−0.0043°, −0.0017°},P < 0.0001), or about 0.33 km, for every 1 °C increase in a species’ thermal optimum (Fig. 2B andSI Appendix, Table S9). Non-territorial species had significantly greater northward shifts in their latitudinal optima than either weakly territorial or strongly territorial species (Fig. 2A). In addition, habitat generalists shifted more, on average, than habitat specialists; that is, annual shifts were negatively linked to the degree of habitat specialization (SSI slope estimate = −0.0193; CI = {−0.0333, −0.0053},P = 0.0068,Fig. 2C andSI Appendix, Table S9).

At the regional scale, the relationships between shifts in the absolute value of latitudinal optima and the species-specific traits investigated here were only statistically significant for the West (SI Appendix, Table S10). (In the East, although thermal range appeared initially to be significant, it did not remain so once we removed species without precise thermal range estimates;SI Appendix, 3). Furthermore, unlike the continental-scale analysis, the direction of the shift (northward vs. southward) was not significant for either the West or East regional-scale models (SI Appendix, 3 and Table S10). As for the continental-scale models, no qualitative differences attended results obtained for the western region when species with less precise thermal tolerances were omitted, so the final model for the West was built using the full set of predictor variables and with all bird species. The absolute value of the shift in the latitudinal optimum per annum for birds in the West was negatively related to (in order of importance): estimated thermal optimum, preferred density of habitat, the degree of habitat specialization, and the degree of territoriality (Fig. 3 andSI Appendix, Table S10). In other words, in the West, a bird species showed a greater latitudinal shift, on average, if they i) prefer colder temperatures; ii) prefer denser habitats; iii) are not habitat specialists; iv) and/or are not territorial.

Phylogenetic relatedness among species accounted for very little of the inter-specific variation in shifts of latitudinal optima. The non-phylogenetic component of inter-specific variation was estimated to be 13 times the size of the phylogenetic component at the continental scale (SI Appendix, Table S9), and 15 times the size of the phylogenetic component for birds in the West (SI Appendix, Table S15).

Bird Dataset.

We used annual count data from the North American BBS over a period of 55 y, from 1966 to 2021 (84). Data are collected from thousands of roadside survey routes in 37 Bird Conservation Regions (BCRs) across the United States and Southern Canada. Each route is 40 km long and consists of 50 stops separated by 800 m. Every year, in a single day during the local bird breeding season, a single observer conducts a 3-min point-count at each stop (85,86). We only used data that fit the official BBS data quality criteria (see ref.84 for more details).

The complete BBS dataset includes more than 700 avian taxa. For our study, we removed i) taxa not identified to species level; ii) species with fewer than 50 occurrences over 10 or more years; and iii) primarily nocturnal or aquatic species. Sub-species were grouped under their main epithet. In addition, birds were filtered based on how much of their known range was covered by the BBS spatial extent. First, we used all routes to create a geographic concave hull representing the BBS coverage. Next, we obtained independent distributional data for each species fromBirdLife International andHandbook of the Birds of the World (87). We retained all species for whom the BBS coverage covered at least 70% of their total known breeding distribution (for migrant and partial migrant birds) or resident distribution (for resident birds). Some species are naturally rare, despite having most of their range sampled, so we also retained birds with at least 20 occurrences over 10 or more years for whom the BBS coverage included at least 90% of their breeding or resident distribution. Finally, four bird species were excluded from subsequent analyses because our models estimated either that their optimum lay outside the range of latitudes sampled by the BBS (Contopus cooperi,Junco hyemalis,Melospiza lincolnii) or their abundance pattern was multi-modal (i.e., showing multiple peaks in mean abundance vs. latitude,Cathartes aura).

Our final dataset therefore included counts in at least 10 y for each of 209 bird species from seven orders (Dataset S1): Passeriformes (174), Piciformes (12), Accipitriformes (10), Galliformes (8), Columbiformes (2), Cuculiformes (2), and Falconiformes (1). In terms of conservation status, two species are classified by the IUCN (88) as “vulnerable”: the chestnut-collared longspur (Calcarius ornatus) and the pinyon jay (Gymnorhinus cyanocephalus). Nine are classified as near threatened: the greater sage-grouse (Centrocercus urophasianus), northern bobwhite (Colinus virginianus), wood thrush (H. mustelina), loggerhead shrike (L. ludovicianus), Bachman’s sparrow (Peucaea aestivalis), common grackle (Q. quiscula), cerulean warbler (Setophaga cerulea), eastern meadowlark (S. magna), and golden-winged warbler (Vermivora chrysoptera). The remaining 198 (out of 209) are considered of “least concern” (SI Appendix, Table S3).

SeeDataset S1 for our final list of birds with their migration category based on Tobias et al. (89) and percentage of geographical range covered by the BBS sampling extent. The number of routes sampled by the BBS has increased over time from 1966 (400 routes) to 2021 (2,540 routes). The total number of observation rows (all route-by-year combinations) utilized in our study was 119,567.

Geographical Subsets.

Eastern and western regions of North America are divided by large-scale topographic barriers (90) and differ in their atmospheric circulation, moisture transport, and bio-physical climatic conditions (64). We expected eastern and western regions to have differing bird communities that would potentially have different responses to broad-scale climate change. To identify an appropriate east–west divide on the basis of bird community composition, we performed a spatially constrained cluster analysis on the basis of Bray-Curtis dissimilarities among BCR units. Dissimilarities were calculated from square-root transformed dispersion-weighted abundance values to decrease the relative importance of highly abundant and/or erratic or flocking bird species (91). We used regionalization with dynamically constrained agglomerative clustering and partitioning (REDCAP) (92,93) with group-average linkage. Following this, all BBS routes were classified as either eastern or western. Our resulting east-west divide roughly matches the 100th Meridian (SI Appendix, Fig. S1).

We classified each of the 209 bird species as being “eastern”, “western”, “widespread,” or “intermediate” in its distribution. A species was classified as either eastern (32 species) or western (56 species) if it occurred in 50 or more routes in only one of the two regions. A species was classified as widespread (88 species) if it occurred in at least 50 routes in both eastern and western regions. Species not meeting either of these criteria (e.g., with 20 occurrences in the east and 30 occurrences in the west) were classified as intermediate (33 species).

We performed separate analyses for each of three geographical subsets of data (Table 1): i) East (eastern birds and the eastern routes for widespread birds); ii) West (western birds and the western routes for widespread birds); and iii) Combined (all birds at the continental scale). One species (Passerina caerulea) was widespread, but was only analyzed for the combined dataset because its estimated optimum occurred outside the latitudinal range of BBS data in the west.

Statistical Models of Species’ Latitudinal Optima.

The mean abundance response of a species along a natural broad-scale spatio-environmental gradient (such as latitude, moisture, temperature, etc.) is generally expected to be unimodal (15,16), although not necessarily symmetric or Gaussian (48,94,95). We used species–environment non-linear models (senlm) (51) as a flexible parametric framework to characterize the unimodal mean response of each species vs. latitude within each year. Two important parameters from these models include the maximum estimated mean abundance ( $H$ , the height of the mode in the unimodal response), and the position along the gradient of that peak mean abundance ( $m$ , the modal latitudinal position). The estimated value of$m$ is directly interpretable as the optimal latitude for the species along the gradient. Note that eastern, western, and intermediate species only have one estimated modal position per year, whereas widespread birds have three per year: one for the east, one for the west, and one for the entire dataset.

Here, we extended thesenlm framework to create a non-linear mean function (calledmodskurt;SI Appendix, 1) that is modular, allowing increasing complexity in the shape of the unimodal response curve to include i) left/right asymmetry, ii) peakedness/flattening, and/or iii) exaggerated tails,via inclusion of relevant additional parameters. A bespoke model-selection procedure was then used to choose an appropriate level of complexity for the abundance-vs-latitude model (mean function + error distribution) for each species-by-year combination (SI Appendix, 2). Importantly, the meaning and interpretation of themodskurt function’s core parameters ( $H$   and/or$m$   ) is not altered by increasing (or decreasing) the complexity of the required mean response shape in individual cases. Thus, estimated values of$m$   can be validly compared through time and/or across species. In addition, the error component in oursenlm framework directly models the most vital and well-recognized statistical properties inherent in broad-scale observational species count data (SI Appendix, 1); namely, overdispersion (91,96), zero-inflation (37,50), mean-variance relationships (49,96) and occupancy-abundance relationships (37).Dataset S2 contains the individual estimated values of species’ latitudinal optima ( $m$ ) for all species-by-year combinations for each dataset (East, West, and Combined).

We consider oursenlm modskurt models to be particularly efficacious at drilling through the “noise” inherent in continental-scale observational data to quantify these key signals. Calculations of mean (or weighted mean) latitude values (i.e., a “center of abundance”) using raw counts or densities, or analyses of mean (or modeled) abundances for broad-scale spatial units will not accommodate the skewed/asymmetric nature of count data that is apparent both within a given latitude (handled here via the use of a zero-inflated negative binomial error distribution), and alsoacross latitudes (handled here by the flexible shape of themodskurt non-linear function for the mean response). Unfortunately, asymmetry and zero-inflation can cause even flexible empirical GAMs to yield biased estimates of species optima along gradients (51).

For each species, we plotted the estimated values of their latitudinal optimum ( $m$ ) vs. time (in years) and estimated the linear slope ( ${\beta }_m$ ) of this relationship (Fig. 4) using generalized least squares (GLS), with temporal autocorrelation being accommodated by using either first-order or second-order auto-regressive errors (i.e., AR(1) or AR(2), chosen by AIC). All GLS models were fitted using thegls function in the R packagenlme (97). Species classified as eastern, western, or intermediate had one fitted GLS model (hence, one estimated slope parameter), while those classified as widespread had three: one for each regional subset (East and West) and one for their entire distribution across all of North America (Combined).Dataset S3 contains detailed results from all GLS models (estimates of parameters, associated SE, andP-values) for each individual bird species and dataset. Each estimated slope parameter provides a direct estimate of the latitudinal shift (per annum) in a given species’ peak mean abundance (its optimal latitudinal position) for a particular dataset (East, West, or Combined), with a positive slope indicating a northward shift.

Meta-Analyses of Regional and Continental Trends.

Slopes (and their SE) from GLS models for each species were used as input into a multi-level random-effect meta-analysis to obtain an overall quantitative estimate of the shift in latitudinal optima across all bird species. A separate meta-analysis was done for each of the three datasets: East, West, and Combined (Table 1). We used therma.mv function in the R packagemetafor (98) and included two random factors associated with individual species’ identities: One was an independent species effect, and the other explicitly incorporated phylogenetic correlations among species. Variance components were estimated using restricted maximum likelihood (REML) (99). This approach allows species-level variance in slopes to include an evolutionary component and a component unexplained by phylogenetic relationships (100,101).

To create a phylogenetic correlation matrix among species, we first obtained 999 ultrametric phylogenetic trees from the BirdTree project (http://birdtree.org) (102) and computed a consensus majority rule phylogenetic tree (103) using theconsensus.edges function of thephytools R package (104). The BirdTree phylogeny was built using an older bird taxonomy (105), so we linked each species in our dataset to its appropriate corresponding position in the phylogenetic tree (Dataset S3). Two species in our dataset (Aphelocoma woodhouseii andTroglodytes pacificus) were absent from the phylogeny, so were added manually to their respective genera. The phylogenetic correlation matrix was calculated assuming a Brownian-motion model of evolution [(104); functionvcv], and with branch lengths computed using the method of Grafen [(106); functioncompute.brlen] in theape R package (107).

Linking the Magnitude of Latitudinal Shifts to Species’ Traits.

The modulus of individual species’ GLS slope estimates was used in a multi-level random-effect meta-regression to investigate and quantify the extent to which several potentially important species’ traits might explain inter-specific variation in the absolute sizes of shifts in latitudinal optima. The meta-regression included the same two random-effect components described above to account for phylogenetic relatedness among species. We focused on traits directly or indirectly related to i) temperature optima and/or tolerances, ii) preferred habitat density and/or the degree of habitat specialization, and iii) processes of emigration, vagility, proliferation (e.g., dispersal capability) and/or establishment (25). Information on the following traits: preferred habitat density (an ordered categorical variable with three levels: open, semi-open or dense), migration strategy (three levels: resident, partial migrant, and fully migratory), territoriality (three levels: non-territorial, weakly territorial and strongly territorial) and dispersal capability (the hand-wing index, in mm) were obtained from Sheard et al. (76) and Tobias et al. (89). In addition, we used our bird abundance dataset from the BBS to obtain three more (continuous and quantitative) species-specific traits: thermal optimum (°C), thermal tolerance (a range, also in °C), and a SSI (108,109) to measure the degree of habitat specialization. All categorical variables (migration strategy, territoriality, and preferred habitat density) were converted into ranks ranging from one to three and then were standardized along with the continuous variables to ensure comparability of model coefficients for assessing traits’ relative importance.

To estimate thermal optima and tolerances, we first obtained monthly mean air temperatures for each BBS route for all of the sampled years in our study by accessing the “GHCN_CAMS Gridded 2 m Temperature (Land)” dataset provided by the National Oceanic and Atmospheric Administration (NOAA) on the following website:https://psl.noaa.gov/data/gridded/index.html (110). BBS data are collected during the breeding season, and weather conditions during this period are critical to birds’ reproductive success (111,112). Thus, we filtered the temperature data to include only values obtained during the breeding season for each particular species according to Vuilleumier (113) at each route in each year. For each species, the abundance values vs. temperature values across all routes and years were then used to fit a singlemodskurt model to estimate i) its thermal optimum ( $m_{\mathit{temp}}$ ) along the temperature gradient (in °C); and ii) its thermal tolerance (a range in °C), calculated as the absolute difference between the lower and upper bounds along the temperature gradient generated by the intersection of the 90th quantile curve of the species’ unimodalmodskurt model with a horizontal line corresponding to one-half the height ( $H_{\mathit{temp}}$ ) of the species’ peak estimated abundance (SI Appendix, Fig. S2). For species whose estimated thermal ranges included temperatures outside the BBS sampling frame, we used the absolute difference between the maximum and minimum temperatures at which a non-zero abundance value was observed for that species within the BBS sampling frame.

To estimate SSI, we obtained measures of land cover for BBS routes using the 2015 land-cover map of North America provided by the Commission for Environmental Cooperation (CEC;http://www.cec.org/nalcms). We focused only on BBS routes sampled over the period from 2013 to 2017, to ensure temporal commensurability with the 2015 land-cover map. Given that the geographical coordinates provided in the BBS dataset correspond to the first observation stop of each route, we only used the abundance data from the first two stops to calculate the SSI for each bird species. The map identifies nineteen land-cover classes in accordance with the Land Cover Classification System (LCCS). We combined classes having the same dominant vegetation type (e.g., “temperate or sub-polar needle-leaf forest” and “sub-polar taiga needle-leaf forest” were combined into a single class: “needle-leaf forest”), and ignored five cases that were classified as water or were unclassified. This yielded nine land-cover classes: needle-leaf forest, broad-leaf forest, mixed forest, shrubland, grassland, wetland, cropland, barren, and urban. We estimated the density of each bird species per land-cover class by dividing the number of individuals found in each class by the number of routes in that class. The SSI was then obtained as the coefficient of variation (CV = SD/mean) in the estimated densities among land-cover classes (109). The higher the SSI, the higher the variation in abundance across land-cover classes, indicating a more habitat-specialized species (108). It was not possible to obtain SSI measures for sharp-shinned hawk (Accipiter striatus), red-tailed hawk (Buteo jamaicensis), or Lewis’ woodpecker (Melanerpes lewis) because they were not recorded in any of the first two stops for any of the routes sampled between 2013 and 2017. As a result, these species were omitted from any trait analyses that included the SSI measure.

A separate meta-regression analysis was performed for each data subset (East, West, and Combined). As some species’ thermal ranges were not precisely estimated (i.e., in cases where a species’ estimated thermal range landed outside the BBS sampling extent of temperature values), we completed a second set of meta-analyses that omitted these species. We retained the original model (including all species) when omission of these species yielded no qualitative difference in the results (to maximize the information content for the meta-regression); otherwise, we retained the second model to ensure sufficient precision of estimated thermal ranges.

We used the absolute value of each species’ shift estimate as the response variable in our continental and regional analyses. Our objective was to identify potential links between species’ traits and the magnitude of latitudinal shifts in a given direction. Thus, our full meta-regression model included i) the shift direction as a binary covariate (northward vs. southward), ii) all of the above-mentioned traits, and iii) the interactions between shift direction and every trait. The full model was then simplified using a series of sequential Wald-type tests via the “anova.rmv” function of the “metafor” R package (98). Starting with the term having the highest p-value, we sequentially removed non-significant terms until all remaining predictors had at least a marginally significant effect on the magnitude of latitudinal shift (i.e.,P-value ≤ 0.1). SeeSI Appendix, 3 for full details.

Author contributions

P.M.M. and M.J.A. designed research; P.M.M. and M.J.A. performed research; M.J.A., W.L.S., and A.J.P. contributed new reagents/analytic tools; P.M.M. and M.J.A. analyzed data; and P.M.M. and M.J.A. wrote the paper.

Competing interests

The authors declare no competing interest.

• 1.Bellard C., Bertelsmeier C., Leadley P., Thuiller W., Courchamp F., Impacts of climate change on the future of biodiversity: Biodiversity and climate change. Ecol. Lett.15, 365–377 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 2.IPCC, Climate Change 2022: Impacts, Adaptation and Vulnerability (IPCC, 2022). [Google Scholar]
• 3.Grinnell J., The niche-relationships of the California thrasher. Auk34, 427–433 (1917). [Google Scholar]
• 4.Brown J. H., On the relationship between abundance and distribution of species. Am. Nat.124, 255–279 (1984). [Google Scholar]
• 5.Parmesan C., Ecological and evolutionary responses to recent climate change. Annu. Rev. Ecol. Evol. Syst.37, 637–669 (2006). [Google Scholar]
• 6.Quintero I., Wiens J. J., Rates of projected climate change dramatically exceed past rates of climatic niche evolution among vertebrate species. Ecol. Lett.16, 1095–1103 (2013). [DOI] [PubMed] [Google Scholar]
• 7.Tingley M. W., Monahan W. B., Beissinger S. R., Moritz C., Birds track their Grinnellian niche through a century of climate change. Proc. Natl. Acad. Sci. U.S.A.106, 19637–19643 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 8.Parmesan C., Yohe G., A globally coherent fingerprint of climate change impacts across natural systems. Nature421, 37–42 (2003). [DOI] [PubMed] [Google Scholar]
• 9.Lenoir J., et al. , Species better track climate warming in the oceans than on land. Nat. Ecol. Evol.4, 1044–1059 (2020). [DOI] [PubMed] [Google Scholar]
• 10.Virkkala R., Lehikoinen A., Patterns of climate-induced density shifts of species: Poleward shifts faster in northern boreal birds than in southern birds. Global Change Biol.20, 2995–3003 (2014). [DOI] [PubMed] [Google Scholar]
• 11.Shoo L. P., Williams S. E., Hero J.-M., Detecting climate change induced range shifts: Where and how should we be looking?Aust. Ecol.31, 22–29 (2006). [Google Scholar]
• 12.Maggini R., et al. , Are Swiss birds tracking climate change?Ecol. Modell.222, 21–32 (2011). [Google Scholar]
• 13.Lenoir J., Svenning J.-C., Climate-related range shifts—A global multidimensional synthesis and new research directions. Ecography38, 15–28 (2015). [Google Scholar]
• 14.Spooner F. E. B., Pearson R. G., Freeman R., Rapid warming is associated with population decline among terrestrial birds and mammals globally. Global Change Biol.24, 4521–4531 (2018). [DOI] [PubMed] [Google Scholar]
• 15.Whittaker R. H., Vegetation of the Great Smoky Mountains. Ecol. Monogr.26, 1–80 (1956). [Google Scholar]
• 16.Whittaker R. H., Gradient analysis of vegetation. Biol. Rev.42, 207–264 (1967). [DOI] [PubMed] [Google Scholar]
• 17.Weber M. M., Stevens R. D., Diniz-Filho J. A. F., Grelle C. E. V., Is there a correlation between abundance and environmental suitability derived from ecological niche modelling? A meta-analysis. Ecography40, 817–828 (2017). [Google Scholar]
• 18.de la Fuente A., Hirsch B. T., Cernusak L. A., Williams S. E., Predicting species abundance by implementing the ecological niche theory. Ecography44, 1723–1730 (2021). [Google Scholar]
• 19.Lenoir J., Gégout J. C., Marquet P. A., de Ruffray P., Brisse H., A Significant upward shift in plant species optimum elevation during the 20th century. Science320, 1768–1771 (2008). [DOI] [PubMed] [Google Scholar]
• 20.Pacifici M., et al. , Species’ traits influenced their response to recent climate change. Nat. Clim. Change7, 205–208 (2017). [Google Scholar]
• 21.Auer S. K., King D. I., Ecological and life-history traits explain recent boundary shifts in elevation and latitude of western North American songbirds: Range boundary shifts in songbirds. Global Ecol. Biogeogr.23, 867–875 (2014). [Google Scholar]
• 22.La Sorte F. A., Butchart S. H. M., Jetz W., Böhning-Gaese K., Range-wide latitudinal and elevational temperature gradients for the world’s terrestrial birds: Implications under global climate change. PLoS ONE9, e98361 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 23.Roots E. F., Climate change: High-latitude regions. Clim. Change15, 223–253 (1989). [Google Scholar]
• 24.Peterson A. T., Projected climate change effects on Rocky Mountain and Great Plains birds: Generalities of biodiversity consequences. Global Change Biol.9, 647–655 (2003). [Google Scholar]
• 25.Estrada A., Morales-Castilla I., Caplat P., Early R., Usefulness of species traits in predicting range shifts. Trends Ecol. Evol.31, 190–203 (2016). [DOI] [PubMed] [Google Scholar]
• 26.Calosi P., Bilton D. T., Spicer J. I., Thermal tolerance, acclimatory capacity and vulnerability to global climate change. Biol. Lett.4, 99–102 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 27.Schloss C. A., Nuñez T. A., Lawler J. J., Dispersal will limit ability of mammals to track climate change in the Western Hemisphere. Proc. Natl. Acad. Sci. U.S.A.109, 8606–8611 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 28.MacLean S. A., Beissinger S. R., Species’ traits as predictors of range shifts under contemporary climate change: A review and meta-analysis. Global Change Biol.23, 4094–4105 (2017). [DOI] [PubMed] [Google Scholar]
• 29.Gaston K. J., Birds and ecosystem services. Curr. Biol.32, R1163–R1166 (2022). [DOI] [PubMed] [Google Scholar]
• 30.Rosenberg K. V., et al. , Decline of the North American avifauna. Science366, 120–124 (2019). [DOI] [PubMed] [Google Scholar]
• 31.Williams S. E., de la Fuente A., Long-term changes in populations of rainforest birds in the Australia Wet Tropics bioregion: A climate-driven biodiversity emergency. PLoS ONE16, e0254307 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 32.Lees A. C., et al. , State of the World’s birds. Annu. Rev. Environ. Resour.47, 231–260 (2022). [Google Scholar]
• 33.Zuckerberg B., Woods A. M., Porter W. F., Poleward shifts in breeding bird distributions in New York State. Global Change Biol.15, 1866–1883 (2009). [Google Scholar]
• 34.Hockey P. A. R., Sirami C., Ridley A. R., Midgley G. F., Babiker H. A., Interrogating recent range changes in South African birds: confounding signals from land use and climate change present a challenge for attribution: Range changes in South African birds. Diversity Distrib.17, 254–261 (2011). [Google Scholar]
• 35.Tayleur C., et al. , Swedish birds are tracking temperature but not rainfall: Evidence from a decade of abundance changes: Swedish birds on the move. Global Ecol. Biogeogr.24, 859–872 (2015). [Google Scholar]
• 36.Elith J., et al. , Presence-only and presence-absence data for comparing species distribution modeling methods. Biodiversity Inf.15, 69–80 (2020). [Google Scholar]
• 37.Smith A. N. H., Anderson M. J., Millar R. B., Incorporating the intraspecific occupancy–abundance relationship into zero-inflated models. Ecology93, 2526–2532 (2012). [DOI] [PubMed] [Google Scholar]
• 38.La Sorte F. A., Thompson F. R. III, Poleward shifts in winter ranges of North American birds. Ecology88, 1803–1812 (2007). [DOI] [PubMed] [Google Scholar]
• 39.Bateman B. L., et al. , The pace of past climate change vs. potential bird distributions and land use in the United States. Global Change Biol.22, 1130–1144 (2016). [DOI] [PubMed] [Google Scholar]
• 40.Currie D. J., Venne S., Climate change is not a major driver of shifts in the geographical distributions of North American birds. Global Ecol. Biogeogr.26, 333–346 (2017). [Google Scholar]
• 41.Huang Q., Sauer J. R., Dubayah R., Multidirectional abundance shifts among North American birds and the relative influence of multifaceted climate factors. Global Change Biol.23, 3610–3622 (2017). [DOI] [PubMed] [Google Scholar]
• 42.Brommer J. E., Lehikoinen A., Valkama J., The breeding ranges of Central European and Arctic bird species move poleward. PLoS ONE7, e43648 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 43.Lehikoinen A., Johnston A., Massimino D., Climate and land use changes: similarity in range and abundance changes of birds in Finland and Great Britain. Ornis Fenn.98, 1–15 (2021). [Google Scholar]
• 44.Massimino D., Johnston A., Pearce-Higgins J. W., The geographical range of British birds expands during 15 years of warming. Bird Study62, 523–534 (2015). [Google Scholar]
• 45.Välimäki K., Lindén A., Lehikoinen A., Velocity of density shifts in Finnish landbird species depends on their migration ecology and body mass. Oecologia181, 313–321 (2016). [DOI] [PubMed] [Google Scholar]
• 46.Viana D. S., Chase J. M., Ecological traits underlying interspecific variation in climate matching of birds. Global Ecol. Biogeogr.31, 1021–1034 (2022). [Google Scholar]
• 47.McCaslin H. M., Heath J. A., Patterns and mechanisms of heterogeneous breeding distribution shifts of North American migratory birds. J. Avian Biol.51, jav02237 (2020). [Google Scholar]
• 48.Oksanen J., Minchin P. R., Continuum theory revisited: What shape are species responses along ecological gradients?Ecol. Modell.157, 119–129 (2002). [Google Scholar]
• 49.Taylor L. R., Aggregation, variance and the mean. Nature189, 732–735 (1961). [Google Scholar]
• 50.Martin T. G., et al. , Zero tolerance ecology: Improving ecological inference by modelling the source of zero observations. Ecol. Lett.8, 1235–1246 (2005). [DOI] [PubMed] [Google Scholar]
• 51.Anderson M. J., Walsh D. C. I., Sweatman W. L., Punnett A. J., Non-linear models of species’ responses to environmental and spatial gradients. Ecol. Lett.25, 2739–2752 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 52.Lehikoinen A., Virkkala R., North by north-west: Climate change and directions of density shifts in birds. Glob Change Biol.22, 1121–1129 (2016). [DOI] [PubMed] [Google Scholar]
• 53.Hovick T. J., et al. , Informing conservation by identifying range shift patterns across breeding habitats and migration strategies. Biodiversity Conserv.25, 345–356 (2016). [Google Scholar]
• 54.Anderson M. J., Animal-sediment relationships re-visited: Characterising species’ distributions along an environmental gradient using canonical analysis and quantile regression splines. J. Exp. Mar. Biol. Ecol.366, 16–27 (2008). [Google Scholar]
• 55.Mason L. R., et al. , Population responses of bird populations to climate change on two continents vary with species’ ecological traits but not with direction of change in climate suitability. Clim. Change157, 337–354 (2019). [Google Scholar]
• 56.Antão L. H., et al. , Climate change reshuffles northern species within their niches. Nat. Clim. Change12, 587–592 (2022). [Google Scholar]
• 57.Chen I.-C., Hill J. K., Ohlemüller R., Roy D. B., Thomas C. D., Rapid range shifts of species associated with high levels of climate warming. Science333, 1024–1026 (2011). [DOI] [PubMed] [Google Scholar]
• 58.Lenoir J., et al. , Going against the flow: Potential mechanisms for unexpected downslope range shifts in a warming climate. Ecography33, 295–303 (2010). [Google Scholar]
• 59.Årevall J., Early R., Estrada A., Wennergren U., Eklöf A. C., Conditions for successful range shifts under climate change: The role of species dispersal and landscape configuration. Diversity Distrib.24, 1598–1611 (2018). [Google Scholar]
• 60.Platts P. J., et al. , Habitat availability explains variation in climate-driven range shifts across multiple taxonomic groups. Sci. Rep.9, 15039 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 61.Scheffers B. R., Edwards D. P., Diesmos A., Williams S. E., Evans T. A., Microhabitats reduce animal’s exposure to climate extremes. Global Change Biol.20, 495–503 (2014). [DOI] [PubMed] [Google Scholar]
• 62.Portmann R. W., Solomon S., Hegerl G. C., Spatial and seasonal patterns in climate change, temperatures, and precipitation across the United States. Proc. Natl. Acad. Sci. U.S.A.106, 7324–7329 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 63.Meehl G. A., Arblaster J. M., Branstator G., Mechanisms contributing to the warming hole and the consequent U.S. east-west differential of heat extremes. J. Clim.25, 6394–6408 (2012). [Google Scholar]
• 64.Seager R., et al., Whither the 100th Meridian? The once and future physical and human geography of America’s arid-humid divide. Part I: The story so far. Earth Interact.22, 1–22 (2018).31097909 [Google Scholar]
• 65.Curley S. R., Manne L. L., Ramírez-Garofalo J. R., Veit R. R., Evaluating compositional changes in the avian communities of eastern North America using temperature and precipitation indices. J. Biogeogr.49, 739–752 (2022), 10.1111/jbi.14340. [DOI] [Google Scholar]
• 66.Stegen J. C., et al. , Stochastic and deterministic drivers of spatial and temporal turnover in breeding bird communities. Global Ecol. Biogeogr.22, 202–212 (2013). [Google Scholar]
• 67.Marjakangas E.-L., et al. , Ecological barriers mediate spatiotemporal shifts of bird communities at a continental scale. Proc. Natl. Acad. Sci. U.S.A.120, e2213330120 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 68.Angert A. L., et al. , Do species’ traits predict recent shifts at expanding range edges?Ecol. Lett.14, 677–689 (2011). [DOI] [PubMed] [Google Scholar]
• 69.Schuetz J. G., et al. , Complex patterns of temperature sensitivity, not ecological traits, dictate diverse species responses to climate change. Ecography42, 111–124 (2019). [Google Scholar]
• 70.Beissinger S. R., Riddell E. A., Why are species’ traits weak predictors of range shifts?Annu. Rev. Ecol. Evol. Syst.52, 47–66 (2021). [Google Scholar]
• 71.Buckley L. B., Kingsolver J. G., Functional and phylogenetic approaches to forecasting species’ responses to climate change. Annu. Rev. Ecol. Evol. Syst.43, 205–226 (2012). [Google Scholar]
• 72.Kristensen T. N., et al. , Costs and benefits of cold acclimation in field-released Drosophila. Proc. Natl. Acad. Sci. U.S.A.105, 216–221 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 73.Wölfling M., Uhl B., Fiedler K., Multi-decadal surveys in a Mediterranean forest reserve—Do succession and isolation drive moth species richness?Nat. Conserv.35, 25–40 (2019). [Google Scholar]
• 74.Ducatez S., Clavel J., Lefebvre L., Ecological generalism and behavioural innovation in birds: Technical intelligence or the simple incorporation of new foods?J. Animal Ecol.84, 79–89 (2015). [DOI] [PubMed] [Google Scholar]
• 75.Tingley M. W., Koo M. S., Moritz C., Rush A. C., Beissinger S. R., The push and pull of climate change causes heterogeneous shifts in avian elevational ranges. Global Change Biol.18, 3279–3290 (2012). [Google Scholar]
• 76.Sheard C., et al. , Ecological drivers of global gradients in avian dispersal inferred from wing morphology. Nat. Commun.11, 2463 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 77.Jarzyna M. A., Zuckerberg B., Finley A. O., Porter W. F., Synergistic effects of climate and land cover: Grassland birds are more vulnerable to climate change. Landscape Ecol.31, 2275–2290 (2016). [Google Scholar]
• 78.Haesen S., et al. , Microclimate reveals the true thermal niche of forest plant species. Ecol. Lett.26, 2043–2055 (2023). [DOI] [PubMed] [Google Scholar]
• 79.Schoennagel T., et al. , Adapt to more wildfire in western North American forests as climate changes. Proc. Natl. Acad. Sci. U.S.A.114, 4582–4590 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 80.Gutierrez A. A., et al. , Wildfire response to changing daily temperature extremes in California’s Sierra Nevada. Sci. Adv.7, eabe6417 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 81.Guo F., Lenoir J., Bonebrake T. C., Land-use change interacts with climate to determine elevational species redistribution. Nat. Commun.9, 1315 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 82.Princé K., Zuckerberg B., Climate change in our backyards: The reshuffling of North America’s winter bird communities. Global Change Biol.21, 572–585 (2015). [DOI] [PubMed] [Google Scholar]
• 83.Van der Putten W. H., Macel M., Visser M. E., Predicting species distribution and abundance responses to climate change: Why it is essential to include biotic interactions across trophic levels. Philos. Trans. R. Soc. B365, 2025–2034 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
• 84.Ziolkowski D. J. Jr., Lutmerding M., Aponte V. I., Hudson M. A. R., North American breeding bird survey dataset 1966–2021: U.S. Geological Survey data release (2022), 10.5066/P97WAZE5. Accessed 17 July 2022. [DOI]
• 85.Sauer J. R., Niven D. K., Pardieck K. L., Ziolkowski D. J., Link W. A., Expanding the North American breeding bird survey analysis to include additional species and regions. J. Fish Wildlife Manage.8, 154–172 (2017). [Google Scholar]
• 86.Sauer J. R., et al. , The first 50 years of the North American Breeding Bird Survey. Condor119, 576–593 (2017). [Google Scholar]
• 87.BirdLife International and Handbook of the Birds of the World (2020) Bird species distribution maps of the world. Version 2020.1.http://datazone.birdlife.org/species/requestdis. Accessed 10 June 2022.
• 88.IUCN, The IUCN Red List of Threatened Species (2022). Version 2022-1.https://www.iucnredlist.org. Accessed 25 March 2023.
• 89.Tobias J. A., et al. , AVONET: Morphological, ecological and geographical data for all birds. Ecol. Lett.25, 581–597 (2022). [DOI] [PubMed] [Google Scholar]
• 90.Tiffney B. H., Manchester S. R., The use of geological and paleontological evidence in evaluating plant phylogeographic hypotheses in the Northern Hemisphere tertiary. Int. J. Plant Sci.162, S3–S17 (2001). [Google Scholar]
• 91.Clarke K. R., Chapman M. G., Somerfield P. J., Needham H. R., Dispersion-based weighting of species counts in assemblage analyses. Mar. Ecol. Progr. Ser.320, 11–27 (2006). [Google Scholar]
• 92.Guo D., Regionalization with dynamically constrained agglomerative clustering and partitioning (REDCAP). International Journal of Geographical Information Science22, 801–823 (2008). [Google Scholar]
• 93.Padgham M., spatialcluster: R port of redcap. R package version 0.1.0.006 (2022). Accessed 7 April 2022.
• 94.Austin M. P., On non-linear species response models in ordination. Vegetatio33, 33–41 (1976). [Google Scholar]
• 95.Austin M. P., Searching for a model for use in vegetation analysis. Vegetatio42, 11–21 (1980). [Google Scholar]
• 96.McArdle B. H., Anderson M. J., Variance heterogeneity, transformations, and models of species abundance: A cautionary tale. Can. J. Fish. Aquat. Sci.61, 1294–1302 (2004). [Google Scholar]
• 97.Pinheiro J., Bates D., R Core Team, nlme: Linear and Nonlinear Mixed Effects Models.R package version 3.1-163 (2023). Accessed 12 September 2023.
• 98.Viechtbauer W., Conducting meta-analyses in R with the metafor package. J. Stat. Software36, 1–48 (2010). [Google Scholar]
• 99.Corbeil R. R., Searle S. R., Restricted maximum likelihood (REML) estimation of variance components in the mixed model. Technometrics18, 31–38 (1976). [Google Scholar]
• 100.Hadfield J. D., Nakagawa S., General quantitative genetic methods for comparative biology: Phylogenies, taxonomies and multi-trait models for continuous and categorical characters. J. Evol. Biol.23, 494–508 (2010). [DOI] [PubMed] [Google Scholar]
• 101.Cinar O., Nakagawa S., Viechtbauer W., Phylogenetic multilevel meta-analysis: A simulation study on the importance of modelling the phylogeny. Methods Ecol. Evol.13, 383–395 (2022). [Google Scholar]
• 102.Jetz W., Thomas G. H., Joy J. B., Hartmann K., Mooers A. O., The global diversity of birds in space and time. Nature491, 444–448 (2012). [DOI] [PubMed] [Google Scholar]
• 103.Bryant D., “A classification of consensus methods for phylogenetics” in DIMACS Series in Discrete Mathematics and Theoretical Computer Science (2003), vol. 61, M. F. Janowitz, F.-J. Lapointe, F. R. McMorris, B. Mirkin, F. S. Roberts, Eds. (American Mathematical Society), pp. 163–184. [Google Scholar]
• 104.Revell L. J., Harmon L. J., Collar D. C., Phylogenetic signal, evolutionary process, and rate. Syst. Biol.57, 591–601 (2008). [DOI] [PubMed] [Google Scholar]
• 105.BirdLife International. The BirdLife checklist of the birds of the world, with conservation status and taxonomic sources. Version 3 (2010).http://www.birdlife.org/datazone/species/downloads/BirdLife_Checklist_Version_3.zip. Accessed 5 July 2022.
• 106.Grafen A., The phylogenetic regression. Philos. Trans. R. Soc. Lond. B326, 119–157 (1989). [DOI] [PubMed] [Google Scholar]
• 107.Paradis E., K., Schliep, ape 5.0: An environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics35, 526–528 (2019). [DOI] [PubMed] [Google Scholar]
• 108.Julliard R., Clavel J., Devictor V., Jiguet F., Couvet D., Spatial segregation of specialists and generalists in bird communities. Ecol. Lett.9, 1237–1244 (2006). [DOI] [PubMed] [Google Scholar]
• 109.Martin A. E., Fahrig L., Habitat specialist birds disperse farther and are more migratory than habitat generalist birds. Ecology99, 2058–2066 (2018). [DOI] [PubMed] [Google Scholar]
• 110.Fan Y., van den Dool H., A global monthly land surface air temperature analysis for 1948–present. J. Geophys. Res.: Atmos.113, D01103 (2008). [Google Scholar]
• 111.Selås V., et al. , Climate change in Norway: Warm summers limit grouse reproduction. Popul. Ecol.53, 361–371 (2011). [Google Scholar]
• 112.Coppes J., Kämmerle J.-L., Schroth K.-E., Braunisch V., Suchant R., Weather conditions explain reproductive success and advancement of the breeding season in Western Capercaillie (Tetrao urogallus). Ibis163, 990–1003 (2021). [Google Scholar]
• 113.Vuilleumier F., AMNH Birds of North America (DK Publishing, 2020). [Google Scholar]
• 114.Martins P. M., Supplementary data and code 1 for “Significant shifts in latitudinal optima of North American birds” (PNAS). Figshare. 10.6084/m9.figshare.24881544. Deposited 20 December 2023. [DOI]
• 115.Martins P. M., Supplementary code 2 for “Significant shifts in latitudinal optima of North American birds” (PNAS). Figshare. 10.6084/m9.figshare.24881547. Deposited 20 December 2023. [DOI]

Supplementary Materials

Data Availability Statement

The code used to filter the North American BBS dataset, assess the extent to which the BBS sampling covers each bird distribution (Dataset S1), and conduct the primary analysis has been archived on Figshare (114,115). All data for the primary analysis are provided as supplementary datasets. Bird Life´s (BL) geographic distributions for the birds of the world are available under request from BL website (https://datazone.birdlife.org/species/requestdis). Previously published data were used for this work (84,89).