You are on page 1of 11

| |

Received: 15 May 2017    Revised: 13 September 2017    Accepted: 28 September 2017

DOI: 10.1002/ece3.3520

ORIGINAL RESEARCH

Have bird distributions shifted along an elevational gradient on
a tropical mountain?

Marconi Campos-Cerqueira1  | Wayne J. Arendt2 | Joseph M. Wunderle Jr2 | 
T. Mitchell Aide1

1
University of Puerto Rico-Río Piedras, San
Juan, Puerto Rico Abstract
2
International Institute of Tropical An upward shift in elevation is one of the most conspicuous species responses to cli-
Forestry, USDA Forest Service, Sabana Field
mate change. Nevertheless, downward shifts and, apparently, the absences of
Research Station, Luquillo, Puerto Rico
response have also been recently reported. Given the growing evidence of multiple
Correspondence
responses of species distributions due to climate change and the paucity of studies in
Marconi Campos-Cerqueira, University of
Puerto Rico-Río Piedras, San Juan, Puerto the tropics, we evaluated the response of a montane bird community to climate
Rico.
change, without the confounding effects of land-­use change. To test for elevational
Email: marconi.campos.cerqueira@gmail.com
shifts, we compared the distribution of 21 avian species in 1998 and 2015 using
Funding Information
MCC was supported by the fellowship occupancy models. The historical data set was based on point counts, whereas the
“Science Without Borders” from the contemporary data set was based on acoustic monitoring. We detected a similar num-
Coordenação de Aperfeiçoamento de Pessoal
de Nível Superior (CAPES) from Brazil ber of species in historical (36) and contemporary data sets (33). We show an overall
(8933/13-8). pattern of no significant change in range limits for most species, although there was a
significant shift in the range limit of eight species (38%). Elevation limits shifted mostly
upward, and this pattern was more common for upper than lower limits. Our results
highlight the variability of species responses to climate change and illustrate how
acoustic monitoring provides an easy and powerful way to monitor animal populations
along elevational gradients.

KEYWORDS
acoustic monitoring, ARBIMON, climate change, occupancy

1 |  INTRODUCTION for the tropics (Lenoir & Svenning, 2014). The paucity of studies in
the tropics may compromise our predictions of species extinctions risk
There is now ample evidence that climate change is altering species (Maclean & Wilson, 2011) and undermine our ability to manage these
distributions throughout the globe (Chen, Hill, Ohlemüller, Roy, & ecosystems. As climate change is expected to push species up the
Thomas, 2011; Hughes, 2000; Lenoir, Gégout, Marquet, de Ruffray, & mountains, and given that tropical species are already showing higher
Brisse, 2008; Parmesan, 2006; Parmesan & Yohe, 2003; Walther et al., levels of local extinctions due to climate changes (Wiens, 2016), there
2002). In montane environments, the most commonly documented is an urgent need to monitor species responses through the entire el-
spatial pattern associated with climate change is a general upward evational gradient, especially in the tropics, to support conservation
shift in species distribution. Species are shifting to higher elevation planning.
areas at a median rate of 11 meters per decade (Chen, Hill, Ohlemüller, Even though many montane species have already shifted their distri-
et al., 2011), and although upward shifts have been observed both in bution upward, there are also an increasing number of studies document-
temperate and tropical mountains, much less documentation exists ing different patterns of changes in species distributions. For instance,

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,
provided the original work is properly cited.
© 2017 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd.

Ecology and Evolution. 2017;1–11. www.ecolevol.org |  1
  

and land-­use change (Archaux. 2012) demonstrating that individual species respond tions. In the French Alps. & Beissinger. The black circles represent sites sampled in 2015 and yellow circles sites with historical data. USA (Rowe et al. Different colors represent differences in elevation (m a. 1995. 2011). The historical data set was based on also observed for plants species in California.3 1.s. 2008. which is the largest protected area (115 km2) with primary Legend Current sites Historical sites Altitudinal range (m a. Puerto Rico.. 2004. A downward shift in optimum elevation was 21 species in 1998 and in 2015. Abatzoglou. where many other factors.l. 2007). point counts. a 2. Tingley. 2004). 2010). | 2       CAMPOS-­CERQUEIRA et al. probably following changes in vegetation composition To test our hypothesis. ences related with the sampling process between the two time periods. in the Luquillo Mountains.1 | Study site der Putten. Wilcove. Macel. Dobrowski. we compared the elevational distribution of (Deluca & King. such as biotic interactions (Hättenschwiler & Körner. Forest. affect species elevational distribution. 2016). & monitoring.s.65 1. El Yunque National Forest. Merrill et al. These diverse responses reflect 2 | METHODS a more complex phenomenon.3°C increase in spring tem. eastern Puerto Rico (Figure 1). where species were track.. Jetz.070 ¯ 0 0.95 km F I G U R E   1   Map of the Luquillo Mountains and its location in NE Puerto Rico (black area in the inserted map). downward. Koo. Greenberg. to test for shifts in bird distribu- Rush. Most of the LM are protected by the Given the high conservation priority of tropical montane ecosystems. 2010). The study was conducted in the Luquillo Mountains (LM) in north-­ & Dobson. 2015. peratures did not result in a significant upward shift in the distribution of the bird community (Archaux. Van 2.000 – 1. also known as the Luquillo Experimental and the growing evidence of multiple responses in species distributions. 2017). Downward shifts tion upward relative to their historical elevational distribution given that in elevation were also observed in the Appalachian Mountains where 9 the average temperature in Puerto Rico has increased by approximately of 11 high-­elevation species have shifted their lower elevational range 2.l) 0 – 200 200 – 400 400 – 600 600 – 800 800 – 1. stochastic population fluctuations (Lenoir et al. spatially heterogeneous shifts of birds and small mammals were detected we compared past and contemporary elevational distributions of birds in Sierra Nevada Mountains.24°C from 1950 to 2014 (Méndez-­Tejeda. Moritz. Occupancy models were used to take into account differ- Mynsberge. whereas the contemporary data set was based on acoustic ing water availability (Crimmins. & Visser. We predicted that species would shift their elevational distribu- differently across geographically replicated landscapes..) .000 1.

This protected site is ideal for testing for elevational and grouped observations by 0–10. Therefore. and wind velocity (Briscoe. 2007. 2013).24°C from 1950 to 2014. (3) The time period. which recorded sounds between 0 and 22 kHz. period between 1932 and 1994 in the lowlands (100–450 m) of the In this way. Pico del Yunque—1. Recorders collected data at each site for approximately from 1970 to 2000 (Burrowes. Campos-­Cerqueira & census. com/store) were used to collect the audio recording. (2) the LM comprises three count. and we have The historical distribution data were based on the long-­term forest bird found that most bird species were detected up to approximately 50 m.. 2002). Surveyors re- shifts for three reasons: (1) there are no direct effects of land-­use corded the number of individuals of all species detected during the change during the last 100 years in the LM. Studies conducted at the local three main mountain peaks of the LM (Pico del Este. Comarazamy. the frequency range of most bird species. 1966. and are available in the ARBIMON To eliminate double counting of vociferous species such as parrots project (arbimon. ~20 sampling 2.1994) and spans an elevational range from and thrashers. Civco. hurricanes) during this The average temperature in Puerto Rico has increased by approxi.000 mm/year at higher elevation sites (Waide et al.sieve-­analytics. This sampling and were distributed along an elevational gradient (151–1. Gould. For instance.google. 1990). 1996. Warner. 1 week and were then moved to another elevation transect.051 m) (Lasso & Ackerman. Joglar.074 m. & Rivera. humidity.. February to June. as Gould et al. but recorders were placed more than 200 m from any road.1 kHz/16 bit. Scatena.02°C/year recorders. & Lugo. 2013.CAMPOS-­CERQUEIRA et al. the distribution of plants and animals is also strongly affected 1998 survey closely coincided with the March to May 2015 acoustic by this elevation gradient (Brokaw et al. The recording system used a sampling rate of 44. 2016. that is.047 m. of the dition. two recorders. we truncated data to detections with distances ≤40 m 100 to 1. 2010). 1996). 2013). 11–20. and 21–40. Gonzalez. In this way.011 m) area is similar to the sampling area from the point counts. runoff. minimizing observer bias. sampling period. ducted most of the surveys over the 17-­year period. and Pico del Toro). and canopy height (Wang. & Although the historical data include 17 years of bird monitoring. cloud cover. were analyzed. each time period (an essential assumption in occupancy models). least three surveys for each sampling site). (2) About annual rainfall ranges from 2.022°C/year (Méndez-­Tejeda. Willig et al. birds were sampled visually and aurally for band of point counts terminated at a distance of 40 m. Recorders were perature (~0. The contemporary distribution data derive from 60 sampling sites in the LM along three elevational transects (85–1. three field biologists con- has a positive relationship with rainfall. 2000.051 m. & Mineard. with an increase in minimum tem. population study (1989–2006) led by WJA (Arendt. all sampling sites from each transect (~20 sampling sites LM (Greenland & Kittel. & Wu.02°C/year) across the elevation gradient (Van Beusekom. Weaver & Gould. a site is defined as a three-­dimensional hemisphere with 2013).. and a approximately 26. Scatena. González.074 m) providing several elevational gradients. placed on trees at the height of 1. There Luquillo Mountains were no cyclonic storms (i. González et al.050 m. 2011). All recordings 10 min in fixed-­radii concentric circles between 05:00 and 10:30 hr. & Green. Weaver 2015 acoustic survey.2 | Climate change in Puerto Rico and the sites per elevational transect) between March and May of 2015.. these independent sources pro. The main reasons for using the 1998 data were: & Murphy. trails. In terproof case with an external connector linked to a Monoprice micro- addition. (3) there birds were not recorded unless they perched in vegetation during the is extensive research documenting abiotic and biotic changes along 10-­min count.. 2003. .450 mm/year at lower elevations to over 81% of the detections from this year were made by only one observer. To estimate the detection area within the study area. a short-­term study from 2001 to 2013 along the elevational phone) running the ARBIMON Touch application (https://play. Along each elevational transect. García-­Martinó. 1-­min recordings per day. of audio every 10 min for a total of 144. Weaver & Gould. most birds breed from March to June (Reagan & Waide. The elevational transects took advantage of roads and mately 2. gradient in El Yunque showed an increase in the daily minimum tem. Weaver. Pico del Yunque and regional scale encompassing the LM have shown a significant in. Surveys were conducted approximately monthly from these elevational gradients.007°C/year) during the 62 year arated by at least 200 m. we 2. 2003). an increasing trend in the minimum per elevational transect) were sampled simultaneously using 20 audio daily temperature at El Verde Station (350–450 m) of 0. groups of two or more (Thomas et al. Pico cluster. In ad..048°C/year and an increase in maximum temperature elevational transect started in the lowlands and reached one of the 0. 2004).. 4. 2012. Weaver & Gould. Hall.5°C in the lowlands to approximately 20°C at the few months before Hurricane Georges. as the outer in the LM. Aerial del Toro—1.com/project/elevation/dashboard). we used the data from February to June 1998 (with at temperature. Qian. forest growth. 2013) and a negative relationship with for this study.e. sep- crease in annual mean temperature (0. the LM elevational gradient 1989 to 2006. 2006. Average temperature declines with elevation from (1) This sampling period occurred 10 years after Hurricane Hugo. thus reducing any extreme vari- summit (Waide. estimated distance to detected birds or to the center of the main peaks (Pico del Este—1. vide strong evidence for an increase in temperature in our study area. Each perature of 0. for comparison with the Fetcher. To minimize observer bias. Together. Average ation in the elevational distributions due to the hurricanes.5 m and programed to record 1 min González. 2017). At each site. and a signifi. were deployed at 100-­m elevation intervals. Hall. cant increase in the mean minimum temperature between 1970 and ARBIMON portable recorders (LG smartphone enclosed in a wa- 2000 on the East Peak summit (1. & Rivera.3 | Survey and resurvey data played bird songs at different distances from the recorder. permanently stored. The sampling sites (n = 105) were spaced at least 100 m apart a radius of approximately 50 m around the recorder. |       3 forest in Puerto Rico (Lugo. tropical storms. it is easier to assume population closure within Aide. 2013. 2008. 2015).

we focused Occupancy-­elevation profiles were created for each species and can be our comparisons on species that had at least two detections in each found in Fig. the Bridled Quail-­Dove (Geotrygon mystacea) and logit(ψ) = β0 + β1 era + β2 elevation + β3 elevation2 the Puerto Rican Vireo (Vireo latimeri) were detected at just one site. pfa < .. To estimate temporal shifts in the lower and upper range limits for each species. the Puerto Rican Sharp-­shinned logit(p) = β4 + β5 era Hawk (Accipiter striatus venator) and the Puerto Rican Broad-­winged . One advantage of this ap. in which the most widespread species were the Scaly-­naped Pigeon (Patagioenas squamosa. The detection of the species in the audio recordings followed two To estimate elevational profiles of occupancy (ψ) for each species. Then. we used an 3 | RESULTS unpaired-­site model (Tingley & Beissinger. To esti- tectability among two time periods (eras) were used to compare the mate pfa. Vocalizations that could not exploratory analyses indicated problems with model convergence.g. is a standardized continuous covariate represented by a linear (elev) and and (2) with a preliminary bird list from the first survey day from each a quadratic (elev + elev2) function. 2011).0. (2002). | 4       CAMPOS-­CERQUEIRA et al. Thirty-­seven species were observed during the paired occupancy model can be described as: historical surveys. includ- Sampling level—Detectability (p) ing two rare and endangered species. one recording every hour (e. we compared the naïve elevation profile of each species 2. This resulted in the analyses of the elevational dis. 12:00. Following species was detected in one era. thus an addi. each covariate separately in the detection function resulting in a total of tional approximately 41. sites and escaped detection (probability of false absence—p fa). and an underlying biological level describing the probability (ψ) that a site is occupied. in which single-­ estimated the probability that the species was present at all of those season occupancy models that take into account differences in de. whereas the exclusion of the hummingbirds was tween eras to determine which era had the highest average detection mainly due to difficulties in species identification of vocalizations dur. so be identified by MCC were reviewed by the other authors. Differently by MacKenzie et al. 2008). An example equation of the un. probability. whereas elevation (n = 35. but not the other. from Moritz et al.0. 33 species were detected in the acoustic monitoring. n = 101/105). 11:00. The Red-­tailed hawk (Buteo jamaicensis) was ods based on model-­averaging of models with ΔAIC <2. S1 along with cumulative AIC weights and model-­averaged sampling periods. Because the contemporary surveys were not con- ducted at the same sampling sites of the historical surveys. and then. ing the acoustic monitoring. 2002). 2009) where data from all sites from both periods are used in a single-­season occupancy model. with interaction terms were considered for the detection function. similar (37 historical species. (2008).000 1 min recordings were evaluated. com- and both “era” and “elevation” are used as covariates influencing both prising 42% of the reported avifauna inhabiting the LM (Wunderle occupancy and detection parameters. Combinations of the covariates along this process. The eight occupancy models were each run with 2013). model (intercept-­only). MCC visually scanned all spectrograms of recordings (Campos-­ (ψ) by combining these covariates along with interaction terms and a null Cerqueira & Aide. All models Although the number of species detected in both time periods was were fitted using the package Unmarked in R (Fiske & Chandler. n = 103/105 sites) and Biological level—Occupancy (ψ) the Bananaquit (Coereba flaveola. Truskinger. We tested eight parameterizations of site..4 | Species data set tions sampled by model-­averaging all models with ΔAIC <2. we steps: (1) First author (MCC) listened to all recordings from 05:30 constructed a set of competing hypothesis of how occupancy changed to 10:00 hr. but ously unidentified vocalization for the site. 1-­min recordings/site) for a total of 2.5 | Analyses to determine how many sites. we followed the approach developed by Moritz et al. Johnson. During 32 models per species (Table S1). and range. we first estimated detection probabilities (p*) for each site by elevational profiles of the bird species. for example. We then calculated pfa by multiplying 1 − p* to all sites where a that could confound our comparisons (Moritz et al. (2008). Forty-­two bird species were detected in this study (Table S2). we considered all range limit shifts with scribing the probability of detection conditioned on occupancy (p). 2011). the Wilcoxon signed-­rank test to compare detection probabilities be- sisted of flying birds. we further developed by Tingley and Beissinger (2009). 33 contemporary species). & Roe.05 as statistically significant. at both the upper and lower limits of the We followed the approach proposed by Moritz et al. 2016. (2008). over era and elevation (Table S1). and we estimated occupancy profiles across the range of eleva- 2. MCC listened to every recording that included a previ. We also used excluded from the analyses because many of the vocalizations con. Era is a covariate factor indicating the 13:00 hr) until 18:00 hr on the first survey day for each sampling site time period of surveys (historical and contemporary). a species was detected in one era. our models contain a sampling level de. occupancy-­elevation profiles for historic (blue) and current (gray) peri- tributions of 21 species. Cottman-­Fields. We compared models using AIC. but not in the other. we opted to include covariates in the simplest form to reduce model set complexity (Burnham & Anderson.0. & Arendt. and then obtained the standard maximum likelihood hierarchical approach introduced the product of these values across these sites (Figure 2). First. model-­averaging model-­specific p estimated from models with ΔAIC proach is that it may also control for differences in methodologies <2.100 validated recordings. Similarly. from 05:30 to 18:00 hr for all subsequent days. whereas six spe- cies.

Species detect- detected at just one site. Shown are occupied (dark gray) and unoccupied (empty circles) sites. probability of false absence (pfa).0 1. W = 4. six expanded and . whereas six species. The covariate era was present in the best occupancy models Caribbean region Lesser Antillean Pewee (Contopus latirostris blancoi) for most species (n = 16/20) providing quantitative evidence of dis- not detected in the March–June 1998 sampling period.8 800 Elevation (m) p 0.4 400 0. we did not create an occupancy cies (38%) that had a significant shift in their range. p = .CAMPOS-­CERQUEIRA et al. We observed an overall pattern of no significant change in the pancy analyses (Table 1).6 600 0. Of the eight spe- because it occurred in all but one site.0 Historical Current 100 200 300 400 500 600 700 800 900 1.4 400 0. were of high bird occurrence at middle elevations (n = 8).8 800 Elevation (m) 0. the widespread species in the contemporary data set were the Bananaquit covariate elevation was included in the most parsimonious models for (59/60 sites) followed by the Puerto Rican Bullfinch (Loxigilla portoric. Furthermore.000 Historical Current Occurrence probability 0.0 1. montana).4 400 0. ability was. 57/60 sites) and the Puerto Rican Spindalis (Spindalis portoricen. and model-­averaged distribution-­elevation profiles Hawk (Buteo platypterus brunnescens). range shifts in the 20 species that had sufficient data for the occu.000 Elevation (m) F I G U R E   2   Example of elevation profiles of upward range expansion (Coccyzus vieillot and Vireo altiloquus) and downward range contraction (Margarops fuscatus).0 1. higher for the contemporary period than for the Of the 26 species detected in both time periods. for example. A decrease in bird occurrence with elevation was the most sis. The Bananaquit was included in Table 1.8 800 p Elevation (m) 0. 57/60 sites). we tested for historical period (Wilcoxon test. |       5 1.6 600 0. and one rare endemic of the model.2 200 0.2 200 0.000 Historical Current Occurrence probability 0.2 200 0.0 1.5. in general. all species in both eras except for the Ruddy Quail-­Dove (Geotrygon ensis. but range limits for most species (n = 13/21) (Figure 3).007).6 p 600 0. Adelaide’s Warbler common pattern in both times periods (n = 11) followed by a pattern (Setophaga adelaidae) and the Puerto Rican Broad-­winged Hawk. The most tributional changes between the two time periods.0 1.000 Historical Current Occurrence probability 0.

28%).12 181–190 +55U Elev 0. mean = 296 m).17 .25 .67 2 BWVI Vireo altiloquus . 2003). Of the six species that ex.34 20 GRAK Tyrannus dominicensis .17 152–776 −480 U Era × Elev 0. 80%.30 .65 No change 16 PRBU Loxigilla portoricensis .51 151–990 +47 U Elev2 0. contracted.011 No change Era × Elev2 0.58 . (4) changes in upper (U) and lower (L) range limit (bold text indicates significant shifts resulting from the pfa tests). these studies vary in the extent of temperature change.04 . Overall. Figures 3 and 4). many species (Table S3). The elevational distribution of eight of 21 species changed sig. these range shifts concur with other studies that nificant decrease in two species (9.58 151–763 +9 U Era × Elev2 0.53 3 EWWA Setophaga angelae .30 152–763 +223 U Era × Elev2 0.70 152–968 +78 U Era × Elev2 0. two contracted (Table 1. 52%. and differences in the elevational gra- ­response by these species is driven by increasing temperature in dient.51 647–684 +67 U Era × Elev 1. NA = not analyzed). the most common pattern was an increase in their study site has been protected for >100 years.55 +121 L 4 PRSO Megascops nudipes .011 No change Era × Elev 0.16 .25 Range contraction 14 PETH Margarops fuscatus .33 .16 .38 .54 2 15 ZEND Zenaida aurita .70 . The elevation distribu- upper range limit (n = 5).35 200–776 +133 U Era × Elev 0. mean = 148 m) and a sig. (2) average detectability per site for historical [p (H)] and contemporary [p (C)] periods.46 + 66 L 6 PRTA Nesospingus speculiferus .011 No change NA NA .35 10 ANEU Euphonia musica . A significant have revealed a common pattern of an upward shift in elevation for increase in lower range limit was detected in three species (14.31 +83L 13 PROP Icterus portoricensis .11 .59 2 8 AMRE Setophaga ruticilla .06 . | 6       CAMPOS-­CERQUEIRA et al.23 18 SNPI Patagioenas squamosa . Such variation results in disparate humidity and precipitation T A B L E   1   Summary of elevation changes for 21 species of forest birds inhabiting the Luquillo Mountains. Puerto Rico.00 +408L 9 PRWO Melanerpes portoricensis .73 +523 L 11 PRSP Spindalis portoricensis .49 .68 152–993 +4 U Era + Elev2 0. Comparison of avian elevational distributions changes for birds 4 |  DISCUSSION across sites and studies is challenging because there are relatively few studies on bird range shifts.32 .15 151–564 +197 U Era × Elev 0.15 .93 151–755 +45 U Era + Elev 0. We detected a significant increase in the tions of six bird species expanded.57 7 PRFL Myiarchus antillarum . especially in the tropics (Table S3).83 19 RLTH Turdus plumbeus . Displayed are as follows: (1) American Ornithologists’ Union four-­letters alpha codes for bird species. upper range limit as a result of the effect of climate change (Parmesan & Yohe.15 .05 641–684 +116 U Era × Elev2 0. (6) the cumulative Akaike’s Information Criterion weight for all models with those terms (weight) Historical elevation Range limit Best occupancy ID Code Species p (H) p (C) range (m) change (m) model Weight Range expansion 1 PRLC Coccyzus vieillot .76 151–1.68 .06 153–183 +239 Era + Elev 0.70 152–968 −112 U Era × Elev2 0.86 152–979 +93U Era × Elev 0. nificantly over a span of 17 years.51 12 MACU Coccyzus minor .98 21 BANA Coereba flaveola NA NA 151–1.19 151–999 No change Era × Elev2 0.28 +105 L 5 RUQD Geotrygon montana . Furthermore.76 . whereas the ranges of two species upper range limit in five species (23.16 184–660 +329 U Era 0. usually associated with the expansion of the of which all increased (mean = 332 m).19 .37 17 PRTO Todus mexicanus .89 151–1. (3) historical elevation range. rather than anthropogenic disturbances given that the panded their range.36 . We hypothesize that the rapid the time scale of each study.82 660–980 +46 U Elev2 0. (5) the best supported form of the occupancy model (Elev = elevation.

Waide. diet.CAMPOS-­CERQUEIRA et al. 1995). which in turn affect vegetation associa.. our study along with these previous studies provides essential baseline information on the F I G U R E   4   Number and proportion of upper and lower range response of species distributions to climate change in tropical mon- shifts for 21 species of forest birds inhabiting the Luquillo Mountains. 2006. The only other study that has tested for elevational distributions cal sample of March–June 1998. whereas nonsignificant shifts are colored green (Table 1) Species Lower range limit Upper range limit detect significant range shifts in birds. to minimize extreme variations in bird distribution associated with there are no clear patterns among these variables and the ability to these hurricanes. contractions (red). Nevertheless.66% only included two points in time. Raxworthy et al. or no significant change (gray). avian resources such as shelter. Nevertheless. whereas range contractions are 100 colored in red for expansions and blue for contractions. survival rates. therefore. Significant (pfa < . 2008). regimes. the LM have not suffered deforestation for more than 100 years. 1998.28% Using occupancy models that consider differences in detectabil- 23. For instance. and 18 14 2 other studies on the effects of climate change on species distribu- Contract tions (Chen et al. have been exposed to at least two severe hurricanes during the last tion. Expand Expand 14.000 * * * 800 * * * Elevation (m) 600 F I G U R E   3   Elevational range changes 400 for 21 species of forest birds inhabiting the Luquillo Mountains (see Table 1 for species’ alpha codes). The extent of the elevational gradient in our study is rela. which was more than 10 years after changes in birds in tropical island ecosystem found a strong pattern Hurricane Hugo and more than 2 months before Hurricane Georges. Another important consideration for our study is that although tions and. but both significant and non. Pie charts display the proportions of range limits that exhibited they do indicate a significant and rapid shift in the ranges of eight spe- significant expansions (green). tane forests.05) 200 * shifts are marked with (*) at the range limit. Shiu et al. Chen. they thereby dictating consequential variants of species richness. 2014). and edaphic effects. Hugo. (Table S3). and distribution of birds have been Changes in temperatures reported for Puerto Rico and the LM documented to change (Arendt. Furthermore. Hill. Body mass. food. . and water. nificant and nonsignificant changes in bird distributions (Table S3). 2011. abundance.71% No change 66. 2011. composi. of upslope shift in birds (Freeman & Class Freeman. Because of the significant changes were found over larger elevational gradients (Table long-­term impact caused by hurricanes. reproduction. Although our results should be considered with caution. proximately 2 years’ posthurricanes (Arendt. abundance. Socolar. Numbers represent the number of individual shifts cies in a protected tropical mountain reserve. immediately after the Hurricane Hugo. we provide observed in each category a quantitative documentation of the distributions of 21 species that can guide future research on the impacts of climate change. One limitation of our study.52% Terborgh. both significant and nonsignificant changes were found matic decrease in abundance of nectarivorous and frugivorous bird during short time scales (<20 years). there was a dra- Similarly. Therefore. and more studies are needed to understand the drivers of range shifts in birds.. levels in the next breeding season (Waide..80% 3 5 ity associated with the sampling process in the two time periods. Wunderle. similar to the period of our study species and an increase in abundance of insectivores and omnivores. tion.200 Expansion Contraction No change * 1. we provided an unbiased comparison of distributional changes along a tropical mountain elevation gradient. the abundance of all species returned to prehurricane tively small compared to other studies. Range expansions are colored in * * green. 50 years (Georges. Forero-­Medina. we chose to use the histori- S3). 2009. even after ap- fall within the temperature ranges of studies that detected both sig. 1991). reproduc- ical parameters. |       7 1. and many other biological and ecolog. 1991). 9. is that it No change 85. & Pimm. 1989). 2000.

2006: fig. range are essential to improving predictions for species vulnerability A similar pattern was also found in Jamaica where the spatial dis. Campos-­Cerqueira & 13 individuals (1975). 2001). has a restricted elevational distribution now oc- Kepler. Lesser logical consequences for the dynamics of the LM. acoustic monitoring has many that thrasher abundance is decreasing at mid-­ and lower elevations in other advantages: (1) Recordings can be permanently stored. which are now overgrown with wood ponent of the diurnal and nocturnal biomass of all vertebrates in vegetation. It feeds mainly on mis- ing effects and disruption of trophic interactions. lizards. one important top predator in Puerto sionally feed on invertebrates (Pérez-­Rivera. unpublished data). p. 1. acoustic monitoring provides a more robust method fluence the biological and ecological interactions between these two to detect species (Celis-­Murillo. Schowalter. The relatively rapid distributional change over 17 years in one of Euphonia (Euphonia musica). This top dance.. Because elevational distribution affects spe- tribution. as recordings can be ac- pansion. because birds are tletoe but as are many Caribbean birds. potential seed-­dispersers species may influence vegetation’s abun- Campos-­Cerqueira and T. 2013). 2014). our results indicate that two invertivorous mi. along with the Pearly-­eyed Thrasher (Margarops fuscatus— other species that exhibited range contraction was the Zenaida Dove range contraction) and the Puerto Rican Tanager (Nesospingus spe. evidence suggests Besides an increase in detectability. it is a dietary generalist and the main top predators and seed dispersers on the island (Reagan & forages on a wide variety of plant material. and not abundance of predators and resident competitors or with therefore. it is the first study to use occupancy mod- known predator on the eggs. Mitchell Aide. func- the LM (Arendt. Symonds. suggested a strong association with elfin woodland above 800 m and historically occurred along the entire elevational gradient (Snyder (Kepler & Parkes. distribution of invertebrate biomass may also be changing. 1996. distribution. Wimmer. Arendt. minimizing temporal or spatial bias in the and is known to feed on fruits. 1995). 2016). is needed to assess their vulnerability to climate change. 2015. & (Setophaga angelae). within the Palo Colorado International (2016) suffered a population decline of 2000 (1937) to forest (Anadón-­Irizarry. although this is the second study to quantita- vation implications is an elevational contraction of the Pearly-­eyed tively assess forest bird distribution in the LM along an elevational Thrasher. & Ward. & Williamson. itat. 1987). constitute the largest com. which feeds mainly on seeds (especially of grasses and culiferus—range expansion) is important in the ecological dynamics herbs) and is a typical lowland species that may have taken advan- of the LM as they exert top–down pressure on the Anolid lizards tage of hurricane-­induced openings and clearings along many of the and Eleutherodactylus frogs that in turn. as does the Antillean sampling process. and assemblages in high-­elevation areas. Sekercioglu. 2006) is a well-­ gradient (see Pagán. any distributional changes by the thrasher may in. Brooke. Snyder. shifts in bird species in a tropical mountain. the range contraction of this species may alleviate predation LM. 1987). including fruits but occa- Waide. In addition. 2006. For instance. Dinsmore. (3) sampling can occur simultaneously the lowlands of the Neotropics (although occurring up to 2. abandoned roads inside the LM. In addition. Stewart. The range shifts of these two birds may suggest that the cies extinction risk (Blake & Loiselle. Fay. The critically endangered Puerto Rican Parrot BirdLife curring mainly between 600 and 900 m. tioning as “museum vouchers.50 below 600 m). (2) observer bias is minimized. Roe. young. and is currently ex. The only predator. Antilles (Arendt. and a quantitative base- grants the Black-­wiskered Vireo (Vireo altiloquus) and the American line for future studies of these and other species. pressure on other species inhabiting high-­elevation areas. had an estimated population of 16–20 birds by Aide. not only in the Redstart (Setophaga ruticilla) may be changing their elevational dis. Our study provides a quantitative description of rapid elevational 1995). 2017). which inhabits at least 16 islands in the the best-­protected forests in Puerto Rico may have significant eco. This supertramp species (Arendt.18). and frogs panding its distribution to high-­elevation areas. negatively affecting both species’ reproductive Towsey. Deppe. which include the (Pérez-­Rivera. to climate and land-­use changes. Gage. This shift & Wege. which can now be used to improve conservation management. The Puerto Rican Rico. servation perspective. Reagan & Waide. along the entire gradient. Schneider. but especially that for the parrot in the lowlands. ants. the Ruddy quail-­Dove (Geotrygon Montana) is widespread in cessed by different experts. et al. For instance. & Presley. 2000. in litt. | 8       CAMPOS-­CERQUEIRA et al. 2004. 1991). and (4) acoustic monitoring can be used to better . Collazo. & Llerandi-­Roman. the Puerto Rican Lizard-­Cuckoo (Coccyzus vieilloti) is found Tanager feeds mostly on palm fruits. while de. the range expansion of these two elevation zone of the most endangered frog species in the LM (M. Thus. given general higher detection success. probabilities and lower uncertainty in the occupancy estimates. 375). and even adults of other birds eling to provide a quantitative elevational profile of the birds in the and thus. (Zenaida aurita). 2006: appendix 1. 1995). Caribbean from Hispaniola in the Greater Antilles to Grenada. analyses. Puerto Rico (Pérez-­Rivera. our study corroborates three previous studies docu- the thrasher is known to compete for cavity nests with the endangered menting that the endemic and now threatened Elfin-­woods Warbler Puerto Rican Parrot (Amazona vittata) (Arendt. 2008). 2012. Together. studies in invertebrate biomass was recently observed by other researchers that provide a detailed quantitative description of species elevational in the LM (Brad Lister. Overall. & Loarie. such as cascad. insects and seeds.” allowing future comparisons and Among the other bird species that showed significant range ex. almost half (49%) of tribution of migratory warblers was correlated with insect biomass the resident birds detected in this study are endemic species and. 1996). However.600 m). In addition. Willig. From a con- Another significant range shift that may have important conser. Indeed. Cecropia. tropics but worldwide. 1995. and species from the genus primarily in lowland (ψ > 0. nest-­site competitors. they occasionally eat spiders. creasing competition at high elevations. Wieley. even though the original description of the species’ hab- 2013 in the LM (White. 2006. 2004. detailed information about changes in their distributions vegetation characteristics (Johnson & Sherry. 1972).

Orlando Acevedo. and life history of the pearly-eyed thrasher (Margarops fuscatus). 11. with comments on the potential im- timating species abundance than the acoustic monitoring and may plications for Puerto Rican Parrot recovery.S. A.. mist nettings) allow an easier means of es. https://doi.1655%2F03-50 Campos-Cerqueira.1073/pnas. MCC wrote first 9263. I. H. W. & Loiselle.en Blake.org/10. K.. H. org/10.. All authors contributed to revisions of the Chen. J. traditional methods Arendt. e1177. Qian. J. (2013). & Anderson. Benedick. especially in regions Tech. https://doi..1466-8238.org/10. J. Open-­and cavity-­nesting traits merge in the evolution ment of active incubation patches. https://doi. Methods in Ecology and Evolution. V. (2011).. D. & Thomas. https://doi. Ornitologia Neotropical. S.7717/peerj. C. W.hljournals. BirdLife International. that is. A Caribbean forest tapestry: The multidimensional na- ture of disturbance and response. … Thomas. and Andrés Hernández Burnham. of high species richness such as the tropics.2305/iucn. Puerto Rico: U. In contrast. J. MCC and TMA. (1966).2010. Brokaw. Greenberg.org/10.org/perlserv/?request=get- abstract&doi=10.. & Aide. (2006). R. Oxford. S.1111/j.T22686239A93104104. Service Research Paper ITF.. the Pearly-­eyed Thrasher (Margarops fuscatus). (8933/13-­8). increased org/10.. (2016). T.rlts. presented the Elfin-­woods Warbler Setophaga angelae in eastern Puerto Rico. D.x Chen. Asymmetric boundary shifts of trop- CO NFLI CT OF I NTE RE ST ical montane Lepidoptera over four decades of climate warm- ing.2016-3. (2006) Distribution. G. Elevation increases in moth assemblages O RCI D over 42 years on a tropical mountain. (2004). (2000). J. NY: for assistance with data collection. https://doi.. W.. Retrieved from monitor animal populations along elevational gradients. 13–63. J. Science. S. & Mineard. S. M. (2011). (2012). (USDA Forest Service). F. R. corded in the historic population surveys in LM. P. 3. Shiu. Changes in climatic water balance drive . acoustic monitoring Publication location: San Juan. J. K. None declared. Forest tions given the threat of climate change.. Holloway. Barlow. N. commis. 3.. asites on Pearly-­eyed Thrashers. Weather in the Luquillo Mountains of Puerto Rico. T. habitat occupancy and population nal of reproduction. C. P. I. J. R. V. Felipe Caño and Pedro Rios University Press. & Green. K. D.12599 Celis-Murillo. 23(2). 455–465. Dobrowski. F. & Ward. Department of Agriculture. Puerto Rico: U. 1340–1348. Although these latter two parameters were re. A. and corrobo- S0959270913000166 rate the level of threat for particular species. the combination of different techniques may be the best Arendt. JMW.. Amazona vittata.1126/science. & ACKNOWLE DG ME NTS Willig. 146. Roy.00366. S. |       9 infer patterns of arrival and departure of migrants and temporal REFERENCES changes in avian reproduction as vocalizations are often a good sig.. https://doi. JMW. Population decline of elevational shifts in bird species along a tropical mountain. V. 34–45. WJA identified bird species in the recordings. Puerto Rico: University of Puerto Rico. 333. San Juan. B. Breeding upwards when climate is becoming warmer: No bird response in the French Alps.. Potential causes for am- Aperfeiçoamento de Pessoal de Nível Superior (CAPES) from Brazil phibian declines in Puerto Rico..). A. Rapid range shifts of species associated with high levels of climate warming. manuscript.uk. Rep. K. … Thomas. 1479–1483. all of which aide in sexing and aging birds in the hand. J. Enigmatic declines in bird numbers in age studies that will test potential mechanisms driving these rapid bird lowland forest of eastern Ecuador may be a consequence of climate shifts. 106(5). 136–146. D. A. 27. Retrieved from http://www. Hill..2012. Adaptations of an avian supertramp: Distribution. Improving distribution data of AUT HOR CONTRI B UTI O N S threatened species by combining acoustic monitoring and occupancy modelling..00246. D. Crowl... provide other important biological information such as the assess- Arendt. Global Ecology and Biogeography.. Lugo.1111/j. A. (2012). Río Piedras. In this study. 166–179. WJA performed point utility of acoustic recordings for surveying tropical birds. cloacal protuberances. B. (2009). W. (2002). H. I. D. New York.1557- MCC. Hill.. increases the precision of the empirical data because of the units’ Archaux. J. Briscoe. Scatena. J. Benedick. A. 404 pp. 60(2). B. Holloway. density of the Elfin-woods Warbler (Dendroica angelae) in Puerto Rico. Effectiveness and recordings and performed modeling work.. K. A.1177 projects around the globe to improve information of species distribu. Institute of Tropical Forestry. The IUCN Red List of trate how acoustic monitoring provides an easy and powerful way to Threatened Species 2016: e.1206432 Chen. & org/0000-0001-6561-5864 Mynsberge.x draft of the manuscript.CAMPOS-­CERQUEIRA et al. of an avian supertramp.t22686239a93104104. San Juan. Ornitologia Neotropical. S. C.. https://doi. 20. Impact of nest predators.. Forest Service. (2004). Moreover.. https://doi. Journal of counts. 1024–1026. MCC was supported by the fel- Springer. Crimmins.. and ectopar- (e. M. C. Brokaw (Ed. Deppe. 7.org/10. distance sampling. J. designed the study. UK: Oxford We thank Paul Furumo. (2011). Serge Aucoin. N. P.. S. Ibis. (2004). PNAS. We encour. Our results further illus. high-­quality recordings and cost-­effectiveness. D. R. Joglar. Bird Conservation International.org/10. Waide. ecol- ogy. J. S. (2015). C. Ohlemüller.. D.. M. sures (gapes).2004.. M. McDowell. together.. L.. Chey. W.g. C.1111/j.1017/ data to confirm general impressions of bird distribution. J. International Institute of Tropical Forestry. 83(2). K. Shiu.org/10. MCC processed audio org/10. Field Ornithology. and we encourage the establishment of long-­term monitoring change. Anadón-Irizarry. Abatzoglou.x low-­cost spatiotemporal coverage. WJA and TMA analyzed output data. 141–154. Z. R. we have provided a quantitative description of rapid Arendt. Model selection and multimodel inference: A pratical information-theoretic approach. T. competitors. (2016). Chey.1474-919X. MCC. C.00594.0809320106 Marconi Campos-Cerqueira  http://orcid. https://doi.. 138–144. 15... Herpetologica. PeerJ. Department of Agriculture. M. lowship “Science Without Borders” from the Coordenação de Burrowes.1111/2041-210X. Gen. solution to sampling the entire bird community.

org/10. https://doi. C. Journal of Animal Ecology. Impact of a century of climate change on small-­mammal Agriculture. 62. Gutiérrez.. K. Unmarked: An R package for fitting Lenoir. Journal of Field Ornithology.1318190111 & Langtimm.1046/j. Gonzalez. Change in diet and foraging behavior of the https://doi. Annual Review of Ecology. J.1146/annurev.org/doi/abs/10. & Civco. 6.0. L.x in Puerto Rico: An Indicator of climate change? Selbyana. Climate-­related range shifts—A global hind warming temperatures.5539/enrr. S. J. vertebrate species distributions. 261–264. Tech. Vittoz. Gégout. https://doi.ejsobi. Gégout.. https://doi. J. 322(5899)..01303. https://doi. https://doi. & Ackerman. a bromeliad from the dwarf montane forest doi..1365-2699. H. N. (2007).org/10.-B. J. and land stewardship. & Pimm. (2014). Patton. E. M. L. Warner. Nichols. 68(2). International Institute of Tropical Forestry.Y. (2008). B. Gutiérrez. IL and London. T.2010.1111/j. Jetz. A. Department of S.. M. D. M. science. Pérez-Rivera. 111(12). 5(6). E.0050157 Popy.. https://doi. https://doi. A. B.1994. Raxworthy. 178–182.2007. Nature. la dieta y la conducta de forrajeo de Euphonia musica en Puerto Rico nal already.. & Yohe. R. 1st 95–104.1658/1100-9233(2006)17[653:sacova]2. org/10. 16–26. P. 24(1). Flowering phenology of sif in Madagascar. JSTOR. 19. Bordignon. Marquet... Animal Conservation.. Guisan. F. https:// Ecological correlates of the threat of extinction in Neotropical bird doi. D. 437–441. (2000). de Ruffray. de. 413–424. org/10.0. Lenoir. Proceedings of the National Academy of org/10. R. P. Puerto Rico: U.. 320. IITF-GTR-39.org/10. runoff and elevation relationships in the Luquillo Mountains of 108(30). M. R. Trends in Ecology & Evolution. I.org/10. 1–18.. G. A. S. T. A. Martinuzzi.. J. Potts. R. W.1111/j. Rakotondrazafy. M. communities in Yosemite National Park. https://doi.x Ramanamanjato. A. J. https://doi. D. (1996). … Svenning. R.org/10.bioone.. K.2307/3236235 Antillean euphonia in Puerto Rico after Hurricane Hugo (Cambio en Hughes. … Stone. (2010). 357–368.00967 Freeman.-C. 4490–4494. Evolution. Rainfall. Combined effects of climate and biotic interactions on the Gould.. Puerto Rico. A. C. Ecology. Puerto Rico: University of Puerto Rico.2307/3072056 species. D. org/10. (2008). H. Responses to recent climatewarm. Science. R. (1991).v7n3p16 Northeastern Puerto Rico. (2011).x Sciences of the United States of America. Ecography. Kepler. 421(6918).org/10.. C. Boyce. 331. B.1111/ecog..1017/ Maclean. M. https://doi.1038/nature01286 vational range changes of birds on Mount Kinabalu.. A. 32(4).. & Chandler.org/10.1156831 Elevational ranges of birds on a tropical montane gradient lag be.. J. Rep. después del Huracán Hugo). G. & Kittel. C. E. Science (New York. C. J. Symonds. Earthworm communities along an elevation gradient in 7(3). B. e157. M. Río Rico. Habitat relationships of avian communities on an ele- position of vegetation along an elevational gradient in Puerto vational gradient in the Luquillo Experimental Forest of Puerto Rico.co. 37(1).. Caribbean Journal of Science.37. J. & King. https://doi. D. change impacts across natural systems. & Rivera.. & Wilson. 653–664.. E. A new species of warbler (Parulidae) from Extinction vulnerability of tropical montane endemism from warming Puerto Rico. Zalamea. of the National Academy of Sciences of the United States of America. Méndez-Tejeda. hierarchical models of wildlife occurrence and abundance. Río Piedras...08. org/10. H. https://doi..org/10. W.. Brooke. https://doi. G.1523-1739. Parra.1163428 Gould. S. (2004). V.org/10. D.. (2006). Ecography. S. Piedras. Borges. B. & Wege. 17. G. E.2007. W. (2006). O. J. Journal of Vegetation Science. Malaysian Borneo. https://doi.044 R..).1111/j.0028535 38(1). (1995). Rapid upslope shifts in New Lugo. R.. 1703–1720.org/10. 43(10). C. Montane birds shift downslope despite Potential mechanisms for unexpected downslope range shifts in a recent warming in the northern Appalachian Mountains. J. 83(8).2307/4084056 and upslope displacement: A preliminary appraisal for the highest mas- Lasso. S. García. The Condor. 2005(January). ity on the distribution of migratory warblers among habitats in 37. C. e28535. Recent ecological responses to S1367943004001246 climate change support predictions of high extinction risk. R.00522.org/10. Lachman. 324–327. Pagán. J. 161–168. Science. PLoS Biology. https://doi. https:// MacKenzie. … Wei.. 56–61. (2016).1016/j. 7..x Jamaica. Yong. J.10 the 20th century. C.. S.org/10. bution tropical birds in Southeast Asia. Journal of Field Ornithology..x Fiske. & Sherry.1007/s10336-016-1414-7 org/10. (2011). M. F. The Puerto Rico Gap Analysis Project volume 1: 77(1). (2012). in Puerto Rico 1950–2014. M. UK: University of Chicago Press. tion probabilities are less than one.. the distributions of breeding birds in the Italian Alps. Journal of warming climate. J. B. Terborgh. gers.org/10... E. 25). Hättenschwiler. Conservation Biology. Journal of Biogeography.. C. Wilcove.1365-2656.1600-0587.. and Systematics.. (2002). Going against the flow: Deluca. J. Fevold.. A weak upward elevational shift in Johnson.S.org/10. 197–247.CO%3B2 tion zone in the Swiss Alps.org/10. D. D. T. https://doi. The importance of vertebrates in the diet of tana- mate and land-­use change on the global diversity of birds. Droege. & Prodon. Socolar. elevational range of a phytophagous insect. D.1016/s0169-5347(99)01764-4 474–478. (2017). https://doi.01596. C. Scatena. S. A. M..1371/journal.1073/pnas. & Wilson.. The food web of a tropical rainforest. 637– US Long-­Term Ecological Research (LTER) sites..110100 213–231.. A. Temporal variability of climate at the change.. Climate Research. https:// Werauhia sintenisii. G.org/10.08041122..org/10. 165 pp.2 Parmesan.. S. J. Global Change Biology. The Auk. Wohlgemuth. & Waide...2008. 1–13. P.. https://doi. J. 14(8). Projected impacts of cli. J. D. L. 295–303.. White. & Dobson. Andrew Royle. tane species to global warming. https://doi. W. J. 15(2).. W. C. Cruz.06279.org/10. B.. J. G. L.. 1–23. Jimenez.. (1995). (2008). B.1126/science. Structure and com.1371/journal. 37–42. N. https://doi.. I.. C.02197.. Alarcon. 8(4). R. D. Conroy. Lenoir. (2010). multidimensional synthesis and new research directions. Guinean birds illustrate strong distributional responses of tropical mon. Biological consequences of global warming: Is the sig. doi. (2014).. L..1073/pnas. A. R. . (2011). 12337–12342.org/10.1111/j. Using diverse data sources to detect ele.. (1994). Pearson. & Brisse. (2008). Sheldon. 43. Díez.1365-2656. I.1199040 Zimmermann. Reagan. (2003). S. 158. (2003). … Ventosa. Journal of A significant upward shift in plant species optimum elevation during Statistical Software. N. Ecological and evolutionary responses to recent climate Greenland. Raselimanana.. Bernard. V. G. PLoS ONE.. (1972). European Journal of Soil Biology. D.. A. G. & Rivera. Effects of food availabil.1017352108 Puerto Rico.1126/ Forero-Medina. https://doi. & Parkes.. USA. 109. 1768–1771.1365-2486. Retrieved from http://www. & Körner. D. Gen. Potential effects of climate change on elevational distri- Raffles Bulletin of Zoology(Suppl. (1995). S24– Merrill. J. Chicago. 57–67. S. N. & Class Freeman. (2007). (2001). L. https://doi. C. R.org/10. Preservation of primary forests in the Luquillo Mountains. S. L.. H.org/10. 2007)109%5B437:PEOCCO%5D2... Increase in the number of hot days for decades González. T. G. J. Moritz. G. 89(1). Eaton.1046/j. Estimating site occupancy rates when detec- Gage. 15–28.1126/science.3354/cr019213 Parmesan.ecolsys... https://doi. (2002).. G. S32. & Svenning. Environment and Natural Resources Research. J. C. & Beissinger. 70. R. R. I. Journal of Vegetation Science.pbio. M. F. M. Peh. Ornithology. (1996).. Journal of Animal Ecology. 2248–2255. L.091305. C.1650/0010-5422( ing of Pinus sylvestris and Pinus cembra within their montane transi.pone. 669. P.-C. E. Lewis. Rabibisoa. (2007). P.2001. 6(12). | 10       CAMPOS-­CERQUEIRA et al.2009. Proceedings García-Martinó. Pérez-Rivera. 1122–1131. A. G.1002/wics. J. P.x Land cover. ed. A globally coherent fingerprint of climate Harris. A. 145–155. 546–560. A. A.. Forest Service. downhill shifts in plant species’ optimum elevations.. (2008). D.

Predicting species distribution and abundance responses to climate change: Why it is es. (2000). A. a tropical rainforest. The effect of hurricane Hugo on bird populations in the Luquillo Experimental Forest.org/10. 54. Los Angeles. Forest vegetation along environmental Sekercioglu. D.. M. … Moritz. Buckland. China: 896. https://doi. New York. R. J. H. H.us/pubs/47292 22(1). Castro-Arellano.. I. 1497–1508. 22(1).0037 porting information tab for this article. Wieley. (1990). Roe. Trends in Ecology nities along elevational gradients: Effects of forest type and other & Evolution. Series B. N.org/10. S. G..1016/ org/10. Beebee. Waide. W. D. F... C. J. T.. pp. E. & Williamson. S. www.. & Rivera. Additional Supporting Information may be found online in the sup- 365(1549). E. Biological Sciences. (2010). Biotropica. A. S. https://doi. Hall. L.. A.1 Truskinger. 14. (2002).2013. T. P.1098/rspb. A. W.1111/j.. R. 18(11). Post. J. Menzel. A. elevational range shifts. P. (2009). G.1098/rstb. mental acoustic recordings to determine bird species richness.00:1–11. & Loarie. A. Schneider. & Klingbeil. J. 97. (2013).CAMPOS-­CERQUEIRA et al. M. Environmental Entomology. 254–259. S.02784. C. 270–277. A..org/10. northeastern Puerto Rico. Sampling environ- avian elevational ranges. White. org/10.. Bloch... Philosophical Transactions of the Royal Society of London. https://doi.org/10. Forest structure and productivity Snyder. Post-­hurricane suc. M. https://doi..2001104 org/10. 11–20. J.1365-2486..1002/ece3. J. C. F.interciencia. 140–150. 25(5). 3–33.. and bird extinctions. M. B. H.035 org/10. 475–480. In B.. C.. C. 262. P. 625–633. & Beissinger.. Spatially heterogeneous impact of climate Modeling the spatial and temporal variability in climate and pri- change on small mammals of montane California.). org/10. https://doi. (2017). F.. 2025–2034. & Wu. https://doi.treesearch. 1–18. S. (2013). T..1111/j. M.1523-1739. 3279–3290. 47(1). R. 23(4). Hedley. C. Beijing.. Forest Royal Society B: Biological Sciences. Avian studies and research oppor- Van Beusekom. 19(3)..2009.org/10. Tingley. M.. 69–82. Puerto Rico. (2011).. Ecological https://doi. Ecological Bulletins. H.tree. https://doi. P. Environmental gradients affect forest structure cessional dynamics in abundance and diversity of canopy arthropods in in Puerto Rico’s Luquillo Mountains.1038/416389a .fs. 1419–1428.. Comarazamy... NY: Thomas.1016/j.2010. (2015). Ecology and Management. J. S. (2008).2009. L. 21–41. J. Waide. P. Proceedings of the mary productivity across the Luquillo Mountains. M.org/10. (2014). E. 33–48.org/10.... & Roe. Dinsmore. Cisneros. P. Tingley. Moritz.jstor.org/10. C.2307/1369528 IEEE. 879– big data. Koo. org/10.. Responses of bird populations in a Puerto Rican ning of spectrograms for efficient identification of bioacoustic events in forest to hurricane Hugo: The first 18 months. https://doi. https://doi. T. & Presley.. The Condor. https://doi. I. H.01737. M. R.2307/2388269 How to cite this article: Campos-Cerqueira M. 46(1). Aide TM. (2013).org/10. 20141857. L. (2013). 29(3).2012.pbio..07.1093/ee/nvw155 Weaver. R. P. Ecological responses to recent climate change.pdf doi. W. Global Change Biology. Rexstad.foreco. B. Fetcher. |       11 Rowe. L. (1995).org/stable/10. (2016). Evol.1109/escience.2307/1564995 impacts on biodiersity and reduction strategies (pp.1016/j. Ecological Bulletins. 69–94. https://doi. P. 179(1–3). R. Ecol Walther. W. Weaver.. Rush. Johnson. E. Parmesan.1111/j. Arendt WJ. 1–83). L.. (2011). 24(11). J. Convey. among plant and animal species. 437–446. C. Habitat loss: Causes. https://doi. Oikos. 282(1799). P.3520 … Bairlein. C. González.2010. González. Biotropica. Fay. Willig.1365-2664. R. M. M.1175/EI-D-14-0023. J.. (2012). along an elevational gradient on a tropical mountain?.1371/journal. E.1 Van der Putten. S. Nature. S. 2017.00852. The parrots of Luquillo: in Puerto Rico’s Luquillo Mountains. W.. Retrieved from http://www. J. N.2014.x Applications. W. J.. S..1600-0706. K. Conservation Biology. https://doi.. Retrieved from http:// case of the Puerto Rican parrot. J. M.. change. M. 416. A. L. JSTOR. & Murphy. M. & Arendt. M.. C. (1987). Interciencia. Cottman-Fields. PLoS Biology. (2015).. Rapid scan. Forest Ecology and Management. https:// Retrieved from http://www. B. D. P. L. M.. C.2011. 43–65... https://doi.. Stewart.. (2010). K. Puerto Rico.1111/j. Conroy.x Willig. A.org/10. (2003). Detecting range shifts from histor. T. & Gould.x Weaver.2307/2388721 CA: Wester Foundation of Vertevrate Zoology. 23(6)..x The push and pull of climate change causes heterogeneous shifts in Wimmer. & Visser. Have bird distributions shifted Luquillo Mountains... S. Short-­term tunities in the Luquillo Experimental Forest: A tropical rain forest in precipitation and temperature trends along an elevation gradient in Puerto Rico. 5–14. & Llerandi-Roman. B. Rowe. & Kepler. Strindberg. Wunderle. S. K. Scatena.. Macel. M. Tropical metacommu- ical species occurrences: New perspectives on old data. N.fed. Collazo. Nova Science Publishers Inc. R. Wang. Patton. J. S. & Beissinger.org/10.05.. 389–395. ulation size. Tingley. Climate gradients in northeastern Puerto Rico. Climate-­related local extinctions are already widespread and analysis of distance sampling surveys for estimating pop. J. M. W. Climate variability at multiple spatial and temporal scales in the Wunderle JM Jr. Devore (Ed.. https://doi.25 Wunderle. S.2007.1857 S0378-1127(02)00489-9 Schowalter. Koo. Laake.. G. W. I.. C. In IEEE 9th International Conference on eScience. R. Climate driven population fluctuations in rain forest Niche restriction and conservatism in a Neotropical psittacine: The frogs’..19218. Hall.org/10. E. Distance software: Design Wiens.. & Lugo. Journal of Applied Ecology. https://doi. L. E. https://doi. Towsey.org/v25_05/weaver.. … Burnham. M.009 environmental factors. S. Earth Interactions. Natural history and conservation of the Puerto Rican parrot. Higgins. S U P P O RT I NG I NFO R M AT I O N sential to include biotic interactions across trophic levels. J. (1995). 120(10).1890/12-2088. Presley. Puerto Rico. 54. Journal of Herpetology. (1991).org/10.