https://orcid.org/0000-0001-8435-4516
ABSTRACT
Animals endemic to tropical mountains are known to be especially vulnerable to climate change. The Cordillera de Talamanca (Costa Rica and Panama) is a geographically isolated mountain chain and global biodiversity hotspot, home to more than 50 endemic bird species. We used eBird community science observations to predict the distributions of a suite of 48 of these endemic birds in 2006–2015, and in 2070, under four climate change scenarios. Species distributions were predicted using program Maxent, incorporating elevation, satellite derived habitat data, and WorldClim climate variables. Model fit, as assessed by Area under the Receiver Operator Curve (AUC), was very high for most species, ranging from 0.877 to 0.992 (mean of 0.94). We found that most species are predicted to undergo range contractions by 2070, with a mean of 15% under modest climate change (RCP 2.6) up to a mean of 40% under more severe climate change (RCP 8.5). Most of the current ranges of these species are within existing protected areas (average of 59% in 2006–2015), and with prospective range contractions, the importance of these protected areas is forecast to increase. We suggest that these predicted range declines should elevate conservation concerns for this suite of species, and vigilance, in the form of better population monitoring, is urgently needed.
Introduction
The global average air temperature has risen by more than 1°C since pre-industrial times [Citation1], changing at a rate faster than many species can adapt [Citation2]. By the year 2100, the surface of the earth could have warmed by 3.5°C, which could result in 600 to 900 extinctions of land bird species, with 89% of those in the tropics [Citation3]. Species found at elevations of more than 500 m are more sensitive to warming temperatures because they have low basal metabolic rates and experience limited temperature variation [Citation4]. Upward shifts could result in an “escalator to extinction” for some high elevation species, which simply run out of room as the area of suitable habitat diminishes [Citation5].
On average, birds have been documented to be shifting their ranges upslope by 11 m per decade [Citation6], but shifts are larger for birds found in the tropics than those from temperate regions [Citation7]. The causal mechanism for the stronger response of tropical montane bird species to climate change is not well understood, but it could be due to changes in prey abundance and adaptation to narrow temperature regimes [Citation5]. There are various drivers of these potential range shifts, including internal species traits (e.g. time delay in species’ response, individualistic physiological constraints) and external drivers of change (e.g. multispecies interactions, non-climatic factors) [Citation6]. Furthermore, the effects of climate change on tropical birds will be exacerbated by the synergies of climate change and other threats such as diseases, invasive species, hunting, and habitat loss [Citation3].
Despite these predictions, there are some uncertainties. Birds have been shown to be capable of closely tracking warming temperatures [Citation8] and because they are mobile, it is relatively easy for them to move upslope compared with less mobile organisms. Why then would they still face the possibility of going extinct? While birds could potentially shift their range, their food resources and habitat might not be able to shift within a short period of time. For example, synchronizing their reproductive cycle with the emergence of food has been identified as a serious challenge for insectivorous birds in the face of climate change [Citation9]. The effects of climate change on some ecosystems are very specific. One example is the cloud forests of Central America. Cloud forests are montane forests that have permanently moist climates due to frequent low-cloud and are among the most biodiverse habitats globally [Citation10]. According to a global survey of elevation-restricted birds, over 20% of species are found primarily in cloud forest, and many of these species are endangered [Citation11].
Southern Central America is one of the most biodiverse places in the world [Citation12], with more breeding bird species than the United States and Canada combined [Citation13,Citation14]. Many bird species found in Central America are endemic to the region [Citation13–15]. Of those endemic species, a majority are found at high elevations in the Cordillera de Talamanca, or on isolated volcanoes in the north of Costa Rica, within the geographical area called the Talamanca Montane Forests [Citation13,Citation14]. These species, with small geographic ranges and sometimes narrow elevational ranges, are vulnerable to future decline, as climate change reduces their prospective habitats. Latitudinal shifts away from the equator are not an adaptation option for these species, because forested areas to the immediate north and south of this region are at low elevation for several hundred kilometers [Citation16,Citation17].
In this study, we predict the ranges of 48 elevation-restricted birds that are endemic to Cordillera de Talamanca and surrounding areas, when faced with predicted climate change. Our baseline bird data are for c. 2010 (2006–2015), and future predictions are for the year 2070, approximately 50 years hence. While our study region focusses on the Cordillera de Talamanca, which reaches a maximum elevation of 3,819 m at Cerro Chirripó, we include the whole of Costa Rica and Panama west of 70° west (), to encompass the entire geographic ranges of the focal species [Citation19]. Almost all these species are currently listed as of “Least Concern” on the Red List of International Union for Conservation of Nature, with the only exception being the Red-fronted Parrotlet Touit costaricensis, which is listed as “Vulnerable” (). All species have global range sizes that average less than 10,000 km2 [Citation20]. Most of the study species are forest obligates, and hence largely restricted to the Talamancan montane forest ecosystem, but some of the most elevationally restricted are found in the páramo shrub and high elevation grasslands that are found above 3,000 m [Citation13,Citation14,Citation17,Citation19].
Figure 1. Elevation within the study area of Costa Rica and western Panama (top), and forested (MODIS land cover) and protected areas (bottom) from the world database on protected areas [Citation18].
![Figure 1. Elevation within the study area of Costa Rica and western Panama (top), and forested (MODIS land cover) and protected areas (bottom) from the world database on protected areas [Citation18].](/cms/asset/4918e697-13dc-4122-bc70-0e7d8dc460f2/tneo_a_2261196_f0001_oc.jpg)
Table 1. List of species included in the analysis, with IUCN Red list categorization.
Methods
We used program Maxent [Citation21] to predict baseline (2006–2015), and future (2070) species distributions under the four climate change scenarios of the IPCC’s AR5 Greenhouse Gas Concentration Pathways [Citation22].
Data sources
To predict baseline species distributions, we used data from the Cornell Lab of Ornithology’s eBird program – a community science database of bird observations compiled from checklists submitted through an online portal [Citation23]. eBird data are verified by both model-based automated filters and a network of voluntary reviewers [Citation23]. We used bird observations for the 10-year period 2006–2015 to estimate baseline distributions. This time period was carefully chosen to ensure that sample sizes were sufficient for as many species as possible, while also ensuring that it was as close as possible to the baseline climate data from 1970 to 2000. Because eBird began in 2002, and data submissions were sparse for the first few years, it was not possible to more closely align the bird data and climate data. To reduce the effects of sampling bias we thinned locations, eliminating duplicate locations less than 2 km apart, using R package SPthin [Citation24]. We choose 2 km for thinning because many of the species have small geographic distributions and thinning by larger distances would result in a greatly diminished sample for some species. Following thinning, there were between 35 and 436 locations for the 48 study species (), with a mean of 193 locations per species. The Talamanca Hummingbird Eugenes spectabilis was added to the analysis following its split from the Rivoli’s Hummingbird Eugenes fulgens in 2017 [Citation25].
Environment data layers included elevation (SRTM DEM [Citation26]), land cover, ecosystem type, and climate. We used the Model Builder in ArcMap 10 [Citation27] to clip each environmental raster to the study area to standardized projection, extent and cell size (90 × 90 m) and convert them to ASCII format for use in Maxent.
Land cover type was derived from Moderate Resolution Imaging Spectroradiometer (MODIS) satellite imagery obtained in 2001–2012 [Citation28]. These images are based on International Geosphere–Biosphere Programme (IGBP) ecosystem classes [Citation29], which do not differentiate between types of broadleaf evergreen forest, which is the dominant forest type in our study area, comprising 50.5% of the land surface. To provide a more nuanced representation of forest types, we also included an ecosystem typing based on the Central American Ecosystems Map from year 2000 [Citation30] reclassified into 18 ecosystem types, based on expert opinion (Appendix 1).
We obtained raster layers of monthly climate data from Version 2 of the WorldClim database [Citation31]. Future climate projections were from the HadGEM2-ES global climate model. Based on previous studies [Citation32–35] we selected seven climate variables for inclusion in the models, representing both temperature and precipitation metrics: BIO1 = Annual Mean Temperature; BIO4 = Temperature Seasonality (standard deviation × 100); BIO5 = Max Temperature of Warmest Month; BIO6 = Min Temperature of Coldest Month; BIO7 = Temperature Annual Range (BIO5-BIO6); BIO12 = Annual Precipitation; and BIO15 = Precipitation Seasonality (Coefficient of Variation).
Climate data were for a baseline period of 1970 to 2000, and the four future scenarios for the year 2070: RCP 2.6, RCP 4.5, RCP 6.0, and RCP 8.5. RCP 2.6 assumes that global greenhouse gas emissions peak between 2010 and 2020, and then decline; RCP 4.5 and 6.0 assume that emissions continue to rise, with peak emissions occurring in 2040 and 2080, respectively, while RCP 8.5 assumes emissions continue to rise throughout the 21st Century. Hence, climate change would be modest under RCP 2.6, more severe under 4.5 and 6.0, and most severe under RCP 8.5 [Citation36]. As suggested by other authors [Citation37,Citation38], we assume that the RCP 8.5 is the most likely scenario, and that RCP 6.0 would be a more moderate scenario that could occur if the global community takes major steps to reduce emissions in coming decades. All temperatures were in x10 degree Celsius. Climate data were aligned and downscaled to 90 m × 90 m cells to ensure compatibility with landscape data.
Species distribution models
We ran Maxent with a regularization multiplier of 2, to avoid overfitting [Citation39]. We tested model performance using Area under the Receiver Operator Curve (AUC) based on a reserved random selection of 20% of sample locations as the test points, and ran a total of 10 replicates, with up to 1500 iterations for each species. We used the Maxent clamping procedure to constrain the range of projected values to the range of the background values used to train the model [Citation21].
To produce maps of estimated area of occupancy in 2006–2015 and for each climate change scenario in 2070, we classified the model output (probability of suitability in each cell, ranging from 0 to 1) into two classes, suitable (0) and unsuitability (1). There are many methods available for classifying probability data into two discrete classes [Citation40]. We used the “Equal test sensitivity and specificity” (ETSS) complementary log–log threshold, which balances errors of omission and commission [Citation21]. The threshold was applied to the Maxent output maps in ArcGIS 10, and the areas with probabilities above the ETSS threshold were assumed suitable, i.e. to be within the species’ baseline range.
Protected areas outlook
Based on the predicted area of occupancy, we calculated how much of the 2006–2015 and 2070 ranges were within protected areas. We overlaid a map of currently protected areas, taken from the World Database on Protected Areas (WDPA, November 2019 update [Citation41]), with raster data of the projected species ranges in Program R. From these data we were able to calculate the percentage of range for each species that remained in protected areas under each projected climate outcome [Citation42].
Results
We found that model fit was excellent for all 48 endemic species analyzed, with a mean AUC of 0.941 (). Only one species had an AUC of less than 0.9, the Red-fronted Parrotlet (0.877; ). These high AUC values indicate that the current distributions of these species can be accurately predicted based on the seven environmental variables selected for our study.
Table 2. Area under the Curve (AUC) scores for our Maxent models for all 48 bird species, baseline range size estimates from BirdLife International maps and Maxent models based on eBird data from 2006 to 15, and predicted future range sizes and range changes under 4 climate change scenarios.
Estimated baseline ranges and range sizes were similar to species distribution maps curated by BirdLife international for all of our study species (e.g. Volcano Junco Junco vulcani; ). However, there was a general pattern of over-estimation of range size for species with the most restricted ranges, but under-estimation for the more widespread species (). Our baseline range sizes were larger than birdlife ranges by a mean of 8%, varying from 59% smaller for the Black-faced Solitaire Myadestes melanops, up to 175% larger for the Sooty Thrush Turdus nigrescens ().
Figure 2. Current distribution [Citation20,Citation20], locations of eBird records (post-thinning) and our baseline (2006–2015) predicted range of the Volcano Junco Junco vulcani. Bird illustration by A. Wilson.
![Figure 2. Current distribution [Citation20,Citation20], locations of eBird records (post-thinning) and our baseline (2006–2015) predicted range of the Volcano Junco Junco vulcani. Bird illustration by A. Wilson.](/cms/asset/cf35b562-1e45-4498-b45d-a9202dcdbf5e/tneo_a_2261196_f0002_oc.jpg)
Figure 3. Comparison between estimated range sizes from birdlife (BirdLife International 2019) and our baseline (2006–2015) estimate.
Our models predict that under RCP 2.6, 39 of the 48 species would undergo range contractions, with a mean range decline of 14.4% (), but no species were predicted to undergo range declines of more than 50%. Under the intermediate scenarios, the number of species predicted to undergo range declines was 44 and 43 under RCP 4.5 and RCP 6.0, respectively. However, the number of species predicted to undergo range reductions of more than 50% was 9 and 11 under RCP 4.5 and RCP 6.0, respectively. Under the more extreme RCP 8.5 scenarios, 45 species were predicted to undergo range declines, 16 of them showing predicted range declines of more than 50%.
Figure 4. Predicted percent changes in range size by 2070 for the 48 species under the four IPCC AR5 greenhouse gas concentration pathways [Citation22]. Circles indicate outliers.
![Figure 4. Predicted percent changes in range size by 2070 for the 48 species under the four IPCC AR5 greenhouse gas concentration pathways [Citation22]. Circles indicate outliers.](/cms/asset/be9edf24-e7c1-4718-ac49-9a1f408ec5f2/tneo_a_2261196_f0004_oc.jpg)
Upslope range contractions could result in a loss of these endemic birds from some mid-elevation protected areas. The Collared Redstart Myioborus torquatus, for example – a species whose predicted range contractions were close to the average among the 48 species – is predicted to be lost from protected areas, such as the cloud forests of Monteverde and Arenal Volcano National Park (). We estimate that on average 58.9% of species’ ranges were in protected areas in 2006–2015 (). Overall, the average percentages of ranges in protected areas is predicted to increase slightly – to 60.9% under RCP 2.6, 64.0% under RCP 4.5, 63.9% under RCP 6.0, and 67.2% under RCP 8.5. For most species, more than 50% of their range is predicted to be protected now and in the future, exceptions being White-tailed Emerald Microchera chionura (39.6% baseline, 37.4–40.4% future), Coppery-headed Emerald Microchera cupreiceps (41.8% baseline, 40.0–46.7% future), and Red-fronted Parrotlet (44.5% baseline, 45.4–53.1% future). At the other extreme, species with the highest percentage of ranges protected include páramo or montane forest specialists, such as Volcano Junco (83.7% baseline, 84.8–85.0% future), Timberline Wren (80.4% now, 84.4–85.0% future), and Silvery-throated Jay (78.5% baseline, 86.7–89.3% future).
Figure 5. Predicted changes of Collared Redstart Myioborus torquatus distribution in central Costa Rica between 206–2015 and 2070 under severe climate change (RCP 8.5), and how these changes would affect the species’ distribution within protected areas (major protected areas labelled). Bird illustration by A. Wilson.
Figure 6. Anticipated percentage of suitable habitat within protected areas in 2006–2015 and under the four IPCC AR5 greenhouse gas concentration pathways [Citation22] for 48 bird species. Circles indicate outliers.
![Figure 6. Anticipated percentage of suitable habitat within protected areas in 2006–2015 and under the four IPCC AR5 greenhouse gas concentration pathways [Citation22] for 48 bird species. Circles indicate outliers.](/cms/asset/3cf5f6f4-630c-433e-b16d-d6019bac41ab/tneo_a_2261196_f0006_oc.jpg)
Discussion
Our study showed that birds endemic to the Cordillera de Talamanca are likely to undergo significant range declines over the next 50 years due to climate change. The extent of those range declines could be modest, if the international community rapidly and aggressively addresses the causes of climate change (RCP 2.6) as has been proposed in the Kyoto Protocol and Paris Agreement [Citation43,Citation44]. However, delay and inaction could push many of these species into more imperiled circumstances (other scenarios). In addition to the as yet unknown trajectory of climate action by humans, there is great uncertainty about how ecosystems will respond to climate change. If ecosystem changes are rapid, high-elevation coniferous forests and páramo shrub and high elevation grasslands could be replaced by montane forest [Citation45]. The Costa Rican páramo is already small and localized, extending over just 150 km2, almost two-thirds of which is in one near-contiguous block around the country’s highest mountain, Cerro Chirripó [Citation45]. Ecosystem replacement is therefore of particular concern for species that are endemic to páramo, such as Timberline Wren Thryorchilus browni, Volcano Hummingbird Selasphorus flammula, and Volcano Junco Junco vulcani, or habitat specialist as Peg-billed Finch Acanthidops bairdii in bamboos (Chusquea sp.) or the Nightingale-thrush Catharus gracilirostris, Sooty Thrush Turdus nigrescens, and Large-footed Finch Pezopetes capitalis in forest edges.
In addition to possible loss of habitat extent, biotic factors such as intraspecific and interspecific interactions as well as the availability of food sources may ultimately play a significant role in species’ responses to climate change [Citation46]. For example, the Sooty Thrush is a high-elevation endemic that could be affected by the upslope spread of close relatives the Mountain Thrush T. plebejus and the Clay-colored Thrush T. grayi, which could use the same food resources and nesting sites. The extent to which high elevation bird species in the Cordillera de Talamanca are vulnerable to such factors is not known.
While our predicted average range losses were similar to those for other regions in the tropics, including Colombia [Citation34], Bolivia and Perú [Citation33], it is important to note that unlike those studies, we did not predict that any species will lose their climactically suitable range completely. Further, unlike some other parts of the world, this region of Central America has a robust network of protected areas, including the transboundary La Amistad International Park, which forms the core of the species ranges for most of the 48 species in our study [Citation19]. We show that the proportion of each species’ range found within these protected areas is likely to increase, because populations are predicted to retract from the less protected lower-elevation forests. The network of protected areas in the Talamancan Mountains has been identified to overlap strongly with priority areas for endemic bird conservation [Citation15]; that is, there is relatively little potential to increase the area of protected habitats for these species. An ongoing commitment to maintaining the protected areas in Costa Rica and Panama is therefore critical to the long-term prospects of the 48 bird species in our study. Some 15, 938 ha for the period 2000–2008 were lost in the Bosque Protector de Palo Seco protected area in Panama. Such habitat loss adds to the pressures that these species face from climate change, it is quite possible that the conservation outlook for these species will be much bleaker if such deforestation continues.
Of all the species included in our study, the status Red-fronted Parrotlet is of most conservation concern. Its population size is thought to be fewer than 10,000 individuals with long-term population declines, and it is classified as Vulnerable by IUCN [Citation47], although this assessment was based on expert opinion in the absence of data. The parrotlet’s decline is thought to be due to habitat loss on the Caribbean slope of Costa Rica, where this species is found during altitudinal migrations from 3,000 down to 500 m [Citation14]. Our study suggests that this species’ range loss by 2070 in the highlands will be between 30.6% and 86.5% (). Therefore, without efforts to protect the altitudinal corridor used for this species, the future of this species in the wild looks gloomy. However, this is not the only species analyzed that conducts altitudinal migrations (), and the needs of other species throughout their annual cycle need further research.
In addition to the species included in our study which are predicted to undergo rapid range contractions, there are two endemic species that we did not include in our study, due to a paucity of eBird observations. The Glow-throated Hummingbird Selasphorus ardens and Yellow-green Brushfinch Atlapetes luteoviridis are endemic to mid-elevations of the eastern Cordillera de Talamanca in Panama and listed as Endangered and Vulnerable by the IUCN, respectively [Citation47]. Both species are thought to be vulnerable to habitat loss and fragmentation due to forest clearing in its distribution that occurred outside of protected areas [Citation48,Citation49].
The Cordillera de Talamanca is the core range for at least 70 bird species [Citation50], even if they are not endemic to it. This list includes near endemics that have relict populations in the Darién Gap of eastern Panama and western Columbia, including the Bare-shanked Screech-Owl Megascops clarkia [Citation51], and Ochraceous Wren Troglodytes ochraceus [Citation52]; or endemic subspecies as Red-tailed Hawk Buteo jamaicensis costaricensis [Citation53] and Unspotted Saw-whet Owl Aegolius ridgwayi ridgwayi [Citation54]. Therefore, the habitat changes and range contractions due to climate change will affect many species beyond the 48 species included in our study, and the implications for genetic diversity of a much wider array of species merits attention.
Despite patchy occurrence data, the predicted area of occupancy in 2006–2015 closely matched the species distribution maps of BirdLife International for all of our study species, which reflects the fact that this suite of species is strongly constrained by elevation and climate [Citation17]. This is supported by the very high AUC values of our models, which indicate that predictions made from eBird data have high sensitivity (true negative rate) and high specificity (true negative rate). This is perhaps not surprising, because the suite of species chosen for our study is known to be climatically or elevationally restricted. However, it is known that current species range sizes are over-estimated for many species, which results in a significant under-estimation of extinction risks [Citation55]. Our study indicates that this large suite of endemic bird species will undergo range contractions, and likely, concomitant population declines in the future, but for all of them, robust population size and trend data are lacking. We recommend that a population monitoring program possibly using a community science approach (e.g. via eBird [Citation56]) be established to monitor trends of these species.
Study limitations
There were some limitations in this study. The eBird data were geographically biased because there are large tracts of the Talamancan forests that are difficult to access, such as the Caribbean slope of the range. Hence, we have no way of verifying our predictions for those areas in which observations are lacking. In addition, there was a temporal mismatch between the bird data and our baseline climate data: eBird data were for 2006 to 2015, while the current climate variables were out-of-date, dating from the period 1970 to 2000. The climate data in the period of 2006–2015 were likely different from those in the period of 1970–2000 since a general warming trend was identified in the region of Central America [Citation57]. Moreover, our assumption which served as the foundation of the whole research was that environmental variables other than the climate are held constant until 2070, whereas, in reality, this is unlikely, for example it is yet to be seen if and how vegetation will respond to climate change.
While the majority of species included in our study are sedentary, six of them conduct altitudinal migrations (). We did not incorporate seasonality in our analysis, but this might be a useful future analysis, since habitat conditions in non-breeding area ranges at lower elevations could be as critical as that in their higher-elevation breeding sites. Additionally, it would be useful to widen this study to include all the non-endemic highland species that undergo elevational migration and to expand on our understanding of the effects of climate change on bird species in this region.
Conclusions
Our analysis predicts that due to climate change, the geographic ranges of elevation-restricted birds in Costa Rica and Panama will decline substantially over the next 50 years. If this were to happen, almost all species would be candidates for upgrading from “Least Concern” to “Vulnerable”, “Endangered” or “Critically Endangered” on the IUCN Red List. However, the predicted future ranges for the majority of endemic species will be inside protected areas, which should help to lessen the likelihood of species extinctions.
Acknowledgments
Bird data were provided by the Cornell Lab of Ornithology and are collated from the observations of hundreds of citizen scientists. We thank Kenneth Boyle, Cody Kiefer, and Jennifer Gilmore for help with data analysis. Funding was provided by the Gettysburg College Provost’s Office. LS was supported by Vicerrectoría de Investigación, Universidad de Costa Rica under Project number B9123.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Data availability statement
Output maps are available at: https://github.com/AndyWilson-GettysburgCollege/Birds-in-Talamancan-Mountains
eBird data are available at: https://ebird.org/science/use-ebird-data/download-ebird-data-products
References
- Schurer AP, Mann ME, Hawkins E, et al. Importance of the pre-industrial baseline for likelihood of exceeding Paris goals. Nat Clim Chang. 2017;7(8):563–567. doi: 10.1038/nclimate3345
- Abolafya M, Onmuş O, Şekercioğlu ÇH, et al. Using citizen science data to model the distributions of common songbirds of Turkey under different global climatic change scenarios. In: Magar V, editor. PLoS one [internet] Vol. 8. 2013. p. e68037. [Available from 2016 Sep 5]]: http://dx.plos.org/10.1371/journal.pone.0068037
- Şekercioğlu ÇH, Primack RB, Wormworth J. The effects of climate change on tropical birds. Biol Conserv [Internet]. 2012;148(1):1–18. 10.1016/j.biocon.2011.10.019
- Şekercioğlu CH, Schneider SH, Fay JP, et al. Climate change, elevational range shifts, and bird extinctions. Conserv Biol. 2008;22(1):140–150. doi: 10.1111/j.1523-1739.2007.00852.x
- Freeman BG, Scholer MN, Ruiz-Gutierrez V, et al. Climate change causes upslope shifts and mountaintop extirpations in a tropical bird community. Proc Natl Acad Sci. 2018;115(47):11982–11987. doi: 10.1073/pnas.1804224115
- Chen I-C, Hill JK, Ohlemüller R, et al. Rapid range shifts of species Associated with high levels of climate warming. Science. 2011;333(6045):1024–1026. doi: 10.1126/science.1206432
- Freeman BG, Class Freeman AM. Rapid upslope shifts in New Guinean birds illustrate strong distributional responses of tropical montane species to global warming. Proc Natl Acad Sci U S A [Internet]. 2014;111(12):4490–4494. Available from 2020 Sep 10 http://www.ncbi.nlm.nih.gov/pubmed/24550460
- Devictor V, Julliard R, Couvet D, et al. Birds are tracking climate warming, but not fast enough. Proc Biol Sci [Internet]. 2008;275(1652):2743–2748. Available from: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2605823&tool=pmcentrez&rendertype=abstract
- Charmantier A, McCleery RH, Cole LR, et al. Adaptive phenotypic plasticity in response to climate change in a wild bird population. Science [Internet]. 2008;320(5877):800–803. Available from: http://science.sciencemag.org/content/320/5877/800.abstract
- Karmalkar AV, Bradley RS, Diaz HF. Climate change scenario for Costa Rican montane forests. Geophys Res Lett [Internet] 2008;35. 10.1029/2008GL033940.
- Foster P. The potential negative impacts of global climate change on tropical montane cloud forests. Earth-Sci Rev [Internet]. 2001;55(1–2):73–106 10.1016/S0012-8252(01)00056-3
- Mittermeier RA, Myers N, Thomsen JB, et al. Biodiversity hotspots and major tropical wilderness areas: approaches to setting conservation priorities. Conserv Biol [Internet]. 1998;12(3):516–520. doi: 10.1046/j.1523-1739.1998.012003516.x
- Ridgely RS, Gwynne JA. A Guide to the birds of Panama: with Costa Rica, Nicaragua, and Honduras. Princeton, NJ: Princeton University Press; 1992.
- Stiles F, Skutch A. A Guide to the birds of Costa Rica. Ithaca, New York: Cornell University Press,; 1989.
- Ramírez-Albores JE, Prieto-Torres DA, Gordillo-Martínez A, et al. Insights for protection of high species richness areas for the conservation of Mesoamerican endemic birds. Divers Distrib. 2021;27(1):18–33. doi: 10.1111/ddi.13153
- Chavarría-Pizarro T, Gutiérrez-Espeleta G, Fuchs EJ, et al. Genetic and morphological variation of the Sooty-capped Bush Tanager (Chlorospingus pileatus), a highland endemic species from Costa Rica and western Panama. Wilson J Ornithol. 2010;122(2):279–287. doi: 10.1676/09-111.1
- Barrantes G, Yglesias M, Fuchs EJ. The roles of history and habitat area in the distribution and composition of avian species assemblages in the highlands of Costa Rica. J Trop Ecol. 2011;27(4):421–428. doi: 10.1017/S0266467411000046
- IUCN. No title [Internet]. Prot. Planet World Database Prot. Areas (WDPA. 2017. https://www.iucn.org/theme/protected-areas/our-work/world-database-protected-areas.
- Barrantes G. The role of historical and local factors in determining species composition of the highland avifauna of Costa Rica and western Panamá. Rev Biol Trop. 2009;57:333–346.
- BirdLife International. No title [Internet]. IUCN Red List bird. 2019. Available from: http://www.birdlife.org.
- Phillips SJ, Anderson RP, Schapire RE. Maximum entropy modeling of species geographic distributions. Ecol Modell. 2006;190(3–4):231–259. doi: 10.1016/j.ecolmodel.2005.03.026
- IPCC. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel Clim Chang. Geneva, Switzerland: IPCC; 2014.
- Sullivan B, Wood C, Iliff M, et al. eBird: a citizen-based bird observation network in the biological sciences. Biol Conserv. 2009;142(10):2282–2292. doi: 10.1016/j.biocon.2009.05.006
- Aiello-Lammens ME, Boria RA, Radosavljevic A, et al. spThin: an R package for spatial thinning of species occurrence records for use in ecological niche models. Ecography (Cop). 2015;38(5):541–545. doi: 10.1111/ecog.01132
- Chesser RT, Burns KJ, Cicero C, et al. Fifty-eighth supplement to the American Ornithological Society’s check-list of north American birds. Auk. 2017;134(3):751–773. doi: 10.1642/AUK-17-72.1
- Jarvis A, Guevara E, Reuter HI, et al. Hole-filled SRTM for the globe: version 4: data grid [Internet]. 2008 [Available from 2020 Nov 9]: http://srtm.csi.cgiar.org/.
- ESRI. ArcGIS desktop: release 10. Redlands, CA: Environmental Systems Research Institute; 2011.
- Friedl M, Sulla-Menashe D, Tan B, et al. MODIS collection 5 global land cover: algorithm refinements and characterization of new datasets. Remote Sens Environ. 2010;114(1):168–182. doi: 10.1016/j.rse.2009.08.016
- Strahler AH, Muller JP MODIS BRDF albedo product: algorithm theoretical basis document. MODIS prod ID MOD43 [Internet]. 1999;Version 5.: 1–53. Available from: %5CBiblioteca_Digital_SPR%5CStrahler1999_ATBD.pdf.
- Vreugdenhil D, Meerman J, Meyrat A, et al. Map of the Ecosystems of Central America: Final Report. World Bank. 2002. Washington, D.C.
- Fick SE, Hijmans RJ. WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. Int J Climatol. 2017;37(12):4302–4315. doi: 10.1002/joc.5086
- Buermann W, Saatchi S, Smith TB, et al. Predicting species distributions across the Amazonian and Andean regions using remote sensing data. J Biogeogr [Internet]. 2008;35:1160–1176. doi: 10.1111/j.1365-2699.2007.01858.x
- Avalos V, Del R, Hernández J. Projected distribution shifts and protected area coverage of range-restricted Andean birds under climate change. Glob Ecol Conserv. 2015;4:459–469. doi: 10.1016/j.gecco.2015.08.004
- Velásquez-Tibatá J, Salaman P, Graham CH. Effects of climate change on species distribution, community structure, and conservation of birds in protected areas in Colombia. Reg Environ Chang [Internet]. 2013;13:235–248. doi: 10.1007/s10113-012-0329-y
- Ortega-Huerta MA, Vega-Rivera JH. Validating distribution models for twelve endemic bird species of tropical dry forest in western Mexico. Ecol Evol. 2017;7(19):7672–7686. 10.1002/ece3.3160
- Meinshausen M, Smith SJ, Calvin K, et al. The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Clim Change [Internet]. 2011;109:213. doi: 10.1007/s10584-011-0156-z
- McCollum D, NagaiI YU, Raiahi K, et al. Energy investments under climate policy: a comparison of global models. Clim Chang Econ. 2013;4(04):1340010. doi: 10.1142/S2010007813400101
- Lyra A, Imbach P, Rodriguez D, et al. Projections of climate change impacts on central America tropical rainforest. Clim Change. 2017;141(1):93–105. doi: 10.1007/s10584-016-1790-2
- Radosavljevic A, Anderson RP, Araújo M. Making better Maxent models of species distributions: complexity, overfitting and evaluation. J Biogeogr. 2014;41(4):629–643. doi: 10.1111/jbi.12227
- Escalante T, Rodríguez-Tapia G, Linaje M, et al. Identification of areas of endemism from species distribution models: threshold selection and Nearctic mammals. TIP. 2013;16(1):5–17. doi: 10.1016/S1405-888X(13)72073-4
- UNEP-WCMC and IUCN. Protected planet: the world database on protected areas (WDPA). Cambridge, UK: UNEP World Conservation Monitoring Centre; 2019.
- Swets JA. Measuring the accuracy of diagnostic systems. Science [Internet]. 1988;240(4857):1285–1293. 10.1126/science.3287615
- Oberthür S, Ott H. The Kyoto protocol : international climate policy for the 21st Century. New York, NY, USA: Springer Berlin Heidelberg; 1999. doi: 10.1007/978-3-662-03925-0
- Rogelj J, den Elzen M, Höhne N, et al. Paris agreement climate proposals need a boost to keep warming well below 2 °C. Nature [Internet]. 2016;534(7609):631–639. doi: https://doi.org/10.1038/nature18307
- Kappelle M, Horn SP. The Páramo Ecosystem of Costa Rica’s Highlands. Chapter 15. Costa Rican Ecosyst. Chicago, Illinois: The University of Chicago Press; 2016;492–523.
- Walther G-R, Post E, Convey P, et al. Ecological responses to recent climate change. Nature [Internet]. 2002;416(6879):389–395. doi: 10.1038/416389a
- BirdLife International. Touit costaricensis. the IUCN Red list of threatened species 2004: e.T22686017A36511305.No title [Internet]. IUCN Redlist. 2004; Available from 2020 Sep 11, https://www.iucnredlist.org/species/22686017/36511305#assessment-information.
- Stiles F, Sharpe C. Glow-throated Hummingbird (Selasphorus ardens), version 1.0. In: del Hoyo J, Elliott A Sargatal J eds. Birds of the world. Ithaca, New York:Cornell Lab of Ornithology. 2020. 10.2173/bow.glthum1.01
- Baldwin J. Yellow-green Brushfinch (Atlapetes luteoviridis), version 1.0. In: Schulenberg T, editor. Birds of the world. Ithaca, New York: Cornell Lab of Ornithology; 2020.
- Barrantes G. Aves de los páramos de Costa Ricae. In: Kappelle M Horn S, editors. Páramos Costa Rica. Heredia, CR: Editorial INBio; 2005. p. 767.
- Holt DW, Berkley R, Deppe C, et al.et al. Bare-shanked screech-Owl (megascops clarkii), version 1.0. In: del Hoyo J, Elliott A Sargatal J editors. Birds of the World [Internet], Ithaca, New York:Cornell Lab of Ornithology, 2020. 10.2173/bow.bssowl.01
- Kroodsma DE, Brewer D Ochraceous Wren (Troglodytes ochraceus), version 1.0. In: del Hoyo J, Elliott A Sargatal J eds., Birds of the World [Internet], Ithaca, New York: Cambridge University Press. 2020. 10.2173/bow.ochwre1.01
- Clark WS, Schmitt NJ. Raptors of Mexico and Central America. Princeton, NJ: Princeton University Press; 2017.
- König C, Weick F, Beckig JH. Owls of the world. 2nd ed. New Haven, CT: Yale University Press; 2009.
- Ocampo-Peñuela N, Jenkins CN, Vijay V, et al. Incorporating explicit geospatial data shows more species at risk of extinction than the current Red list. Sci Adv. 2016;2(11): doi: 10.1126/sciadv.1601367
- Kelling S, Johnston A, Bonn A, et al. Using semistructured surveys to improve citizen science data for monitoring biodiversity. BioScience. 2019;69(3):170–179. doi: 10.1093/biosci/biz010
- Aguilar E, Peterson TC, Obando PR, et al. Changes in precipitation and temperature extremes in Central America and northern South America, 1961–2003. J Geophys Res. 2005;110(D23):D23107. doi: 10.1029/2005JD006119
- Partida-Lara R, Enríquez PL. No TitleTalamanca Hummingbird (Eugenes spectabilis), version 1.0. In: Schulenberg T, editor. Birds of the world. Ithaca, NY, USA: Cornell Lab of Ornithology; 2020.