Bird extirpations and community dynamics in an Andean cloud forest over 100 years of land‐use change

Abstract

Long‐term studies to understand biodiversity changes remain scarce—especially so for tropical mountains. We examined changes from 1911 to 2016 in the bird community of the cloud forest of San Antonio, a mountain ridge in the Colombian Andes. We evaluated the effects of past land‐use change and assessed species vulnerability to climate disruption. Forest cover decreased from 95% to 50% by 1959, and 33 forest species were extirpated. From 1959 to 1990, forest cover remained stable, and an additional 15 species were lost—a total of 29% of the forest bird community. Thereafter, forest cover increased by 26% and 17 species recolonized the area. The main cause of extirpations was the loss of connections to adjacent forests. Of the 31 (19%) extirpated birds, 25 have ranges peripheral to San Antonio, mostly in the lowlands. Most still occurred regionally, but broken forest connections limited their recolonization. Other causes of extirpation were hunting, wildlife trade, and water diversion. Bird community changes included a shift from predominantly common species to rare species; forest generalists replaced forest specialists that require old growth, and functional groups, such as large‐body frugivores and nectarivores, declined disproportionally. All water‐dependent birds were extirpated. Of the remaining 122 forest species, 19 are vulnerable to climate disruption, 10 have declined in abundance, and 4 are threatened. Our results show unequivocal species losses and changes in community structure and abundance at the local scale. We found species were extirpated after habitat loss and fragmentation, but forest recovery stopped extirpations and helped species repopulate. Land‐use changes increased species vulnerability to climate change, and we suggest reversing landscape transformation may restore biodiversity and improve resistance to future threats.

Full text

Bird extirpations and community dynamics in an Andean cloud forest over 100
years of land-use change
Ruben D. Palacio1,2, Gustavo H. Kattan3 and Stuart L. Pimm.1
1 Nicholas School of the Environment, Duke University, Durham, NC 27708.
2 Fundación Ecotonos, Cra 72 No. 13A-56, Cali, Colombia.
3 Pontificia Universidad Javeriana Cali, Departamento de Ciencias Naturales y Matemáticas, Cali, Colombia.
---------------------------------------------------------------------------------------------------------------------
NOTE: This is the “Researcher’s cut” of our article published in the journal Conservation Biology. It has the
supplementary methods and results combined plus more text and figures — free of the editorial constrains of
word limits and other journal impositions.
For citation and the journal’s version of the article follow the DOI: https://doi.org/10.1111/cobi.13423
Abstract
Long-term studies to understand biodiversity changes remain scarce— especially so for tropical mountains.
Here, we examined changes from 1911 to 2016 in the bird community of the cloud forest of San Antonio, a
mountain ridge in the Colombian Andes. We evaluated the effects of past land-use change and assessed species
vulnerability to climate disruption. Forest cover decreased from 95% to 50% by 1959 and 33 forest species were
extirpated. From 1959 to 1990, forest cover remained stable, and an additional 15 species were lost—a total of
29% of the forest bird community. Thereafter, forest cover increased by 26% and 17 species recolonized the
area. The main cause of extirpations was the loss of connections to adjacent forests. Of the 31 (19%) extirpated
birds, 25 have ranges peripheral to San Antonio, mostly in the lowlands. Most still occurred regionally, but
broken forest connections limited their recolonization. Other causes of extirpation were hunting, wildlife trade,
and water diversion. Bird community changes included: (1) a shift from predominantly common species to a
prevalence of rare ones; (2) forest generalists replaced forest specialists that require old-growth, and (3)
functional groups, such as large-body frugivores and nectarivores, declined disproportionally. All water
dependent birds were extirpated. Of the remaining 122 forest species, 19 are vulnerable to climate disruption
and 10 have declined in abundance, including four threatened birds. Our results show unequivocal species losses
and changes in community structure and abundance at the local scale. In general, species were extirpated after
habitat loss and fragmentation, but forest recovery stopped extirpations and helped some species repopulate.
Land-use changes increased species vulnerability to climate change, and we suggest reversing landscape
transformation may restore biodiversity and improve resistance to future threats.
Keywords: local extinctions, community shifts, population extirpations, Andean cloud forests.

Introduction
Assessing the effects of drivers of change on local biotic assemblages is a major challenge (Brook
et al. 2008). In the tropics, habitat loss and fragmentation have extirpated species and caused
declines of functional groups (Sodhi et al. 2004; Sekercioglu et al. 2004). These historical legacies
may influence future species responses to environmental change and have far-reaching effects for
the provisioning of ecosystem services (Bregman et al. 2014; Essl et al. 2015). Furthermore, under
climate change, extirpations are predicted to occur mainly in tropical mountains —through upward
range shifts due to rising temperatures or disruption of habitat conditions (Sekercioglu et al. 2008;
Anderson et al. 2013; Helmer et al. 2019). Such places hold many endemic species (Hazzi et al.
2018) and the interactions with further land-use change could worsen declining trends (Mantyka-
Pringle et al. 2012; Oliver & Morecroft 2014; Sirami et al. 2017).
Long-term studies provide evidence for temporal changes in community composition and structure
that one can relate to their underlying causes. Such studies remain rare for neotropical bird
communities, and the drivers of change may not always be apparent (Blake & Loiselle 2015;
Rosselli et al. 2017). Moreover, most work has been in lowlands e.g., Laurance et al. (2011). Here,
we documented changes over a century to the bird community in the cloud forest of San Antonio,
a mountain ridge in the western Andes of Colombia. We used data from bird surveys conducted
from 1911 to 1990 and surveyed the site in 2016. We extended our survey into the nearby buffer
zone of the Farallones de Cali National Natural Park where the forest is continuous, to evaluate
changes at a regional level.
Our unusually long-term dataset allowed an assessment of the conservation significance of species
losses and gains in response to land-use changes. Species losses occur after habitat loss and may
continue after forest fragmentation, according to an “extinction debt” phenomenon (Kuussaari et
al. 2009; Haddad et al. 2015; Halley et al. 2016). But what happens when forest cover recovers?
Studies have documented delayed extinctions for tropical birds, years after forest loss stopped or
increased (Shaw et al. 2013; Boyle & Sigel 2015). Alternatively, the avifauna may remain stable
(Willrich et al. 2016) or hold a “species credit” (Latta et al. 2017) that allows extirpated species to
recolonize. The dynamics of the bird community at San Antonio provided some answers.
Furthermore, local dynamics depend on habitat configuration at landscape and regional scales
(Opdam & Wascher 2004). We asked whether species with peripheral populations were more
likely to be extirpated with the loss of connectivity from San Antonio to adjacent forests (Terborgh
& Winter 1983; Nathan et al. 1996).
Although biodiversity loss at a global scale is well-understood (Pimm et al. 2014), there is
controversy at the local level. Recent studies contend that species losses can be compensated for
by increasing numbers of other species (Vellend et al. 2013; Dornelas et al. 2014). Gonzalez et al.
(2016) disputed the generality of that conclusion because most of these studies lack appropriate
baselines and only considered species richness (Hillebrand et al. 2018; Cardinale et al. 2018). We
evaluated how community structure and abundance changed over time, assessed trends for
functional groups, and provided measures of functional homogenization (Clavel et al. 2011).
Finally, it is critical to understand how habitat loss and climate disruption intersect (Pimm 2008,
2009). Consider that a species cannot be extirpated twice — so if habit loss exterminates species,
they cannot then succumb to climate change (or they may remain in a perilous state and be

extirpated later). Because mountains tend to have been spared the more extensive habitat
destruction of adjacent lowlands (Pimm 2008) the impacts of land use change and climate will be
additive (Brook et al. 2008). The identities of extirpation-prone species may be predicted using
vulnerability analysis at the local scale (Foden et al. 2013; Fortini & Schubert 2017). Nevertheless,
studies have mostly used models at large spatial scales, and the effect that climate change may
have on the composition and structure of specific bird assemblages has rarely been addressed
(Pearce-Higgins & Green 2014).
In sum, our work had two general objectives: to evaluate the impact of land use change over one
hundred years and to assess the potential future impacts of climate disruption in a tropical mountain
bird community. Specifically, we addressed three key topics for conservation that transcend our
case-study. (1) Whether local species losses are offset by species additions, such that there is no
net biodiversity loss (2) Whether regional habitat loss influences species extirpations; and (3)
species vulnerability to future climate change.
Methods
Study area
San Antonio lies in the Western Andes of Colombia, west of the city of Cali. It is a mountain ridge
with maximum elevations of 2250 m, extending 25 km northeast from the Farallones de Cali massif
(4100 m elevation) near the sources of Dagua and Felidia river. San Antonio is at the intersection
between the humid Pacific lowlands, the drier Cauca Valley, and the high-Andean forests of
Farallones de Cali (Fig. 1). The region encompasses 8,795 ha and is considered a Key Biodiversity
Area KBA (Birdlife International 2019) with cloud forests at elevations over 1700 m. Chapman
(1917) wrote: “At an altitude of 5700 feet (1737 m) we entered the clouds and, at the same time,
the lower border of the cloud forest (...) The cloud-line coincided with the tree-line”. Climate data
from La Teresita meteorological station show the mean temperature was 16.7 °C for the period
1989-2014. The mean rainfall was 1897 mm for the period 1965-2014 and bimodal with peaks of
precipitation occurring in March to May and September to November.

Figure 1. Study area and historical species richness of forest birds. (a) San Antonio is a mountain ridge extending
north-northeast of the Farallones de Cali massif (>4000 m) in the western Andes of Colombia. The visualization
was obtained with a 1 arc-second (30 m) resolution Digital Elevation Model from the Shuttle Radar Topography
Mission (SRTM) led by NASA. (b) the higher number of forest birds of the historical core avifauna concentrates
towards the foothills in the western slope of the western Andes.
Historical data sources
We assembled historical data from five sources of the 20th century. The first source is a
predisturbance baseline of 1911, from an expedition led by Frank M. Chapman to Colombia
between 1911 and 1915 (Chapman 1917). Collecting efforts were unique in that they aimed to
obtain a complete inventory of birds present at each site to evaluate species biogeographical limits
(Kattan et al. 2016). Chapman and his team collected 766 bird skins of 168 species from January
4 to February 21, and from March 30 to April 7. They regarded San Antonio as “one of our most
important stations” (Chapman 1917). A few years earlier, Mervin G. Palmer collected 104 species
from 1907 to 1908 (data obtained from http://www.vertnet.org). Chapman (1917) wrote “The
collections made by us here in connection with those of Palmer, are believed to contain a large
proportion of the birds which occur in this locality.”
The third data source corresponds to a mid-century peak disturbance period after construction of
the Cali to Buenaventura road from 1920 to 1940. It consisted of one-year study in 1959 by Alden
H. Miller, who for 25 years was the director of the Museum of Vertebrate Zoology at the University
of California, Berkeley. Miller provided detailed natural history accounts and qualitative
abundance estimates for most of 167 species recorded (Miller 1963). The latter two sources are
from a period after long-term forest fragmentation. M. Giraldo surveyed in 1985 and Kattan et al.

(1994) conducted an inventory from 1989 to 1990, providing an analysis on the extent of bird
extinctions for the locality since Chapman (1911).
We updated the taxonomy of all sources following the South American Classification Committee
(Remsen et al. 2017). We traced the nomenclature of species over the years using several resources
including de Schauensee (1948), Hilty & Brown (1986), McMullan & Donegan (2014), the
AviBase database (Lepage et al. 2014), and del Hoyo et al. (2017). During this process, we
performed distributional error checking to make sure no geographical outliers were in the data
(Lees et al. 2014). For instance, the records of Chapman (1911) and Miller (1963) suggest the
occurrence of White-crowned tapaculo (Scytalopus atratus) in the early 20th century, rarely present
in the western slope of the western Andes. In contrast, contemporary records list only Nariño
tapaculo (Scytalopus vicinior). This would mean a possible species replacement, but email
communications with the American Museum of Natural History (AMNH) suggest the alternative
and more parsimonious hypothesis of misidentification (Fig. 2), as species distributions and limits
for tapaculos where poorly known at that time (Krabbe & Schulenberg 1997)
Figure 2. Example of distributional error checking. The AMNH skin 108907 labeled as Scytalopus femoralis
confusus will now be considered as White-crowned Tapaculo (Scytalopus atratus). However, expert notes from
T.S Schulenberg and N. Krabbe, indicate that the species may be in fact Nariño Tapaculo (Scytalopus vicinior),
the only species of the genera found in the San Antonio region. Pictures courtesy of Thomas J. Trombone, Data
Manager at the AMNH.
Contemporary bird surveys

We resurveyed birds in two fragmented-forest areas at San Antonio divided by the Cali-
Buenaventura road: Cerro de la Horqueta, and the area known as Km-18. To evaluate the changes
at the regional scale, we also surveyed the locality of La Teresita at the buffer zone of the
Farallones de Cali National Natural Park. Here, the forest is extensive (>100,000 ha) with
elevational continuity (at least to upper elevations) reaching 4100 m.
We visited the three localities four times from May 2015 to February 2016. A sampling session
consisted of three consecutive days and two observers sampled each site simultaneously (Forcey
et al. 2006). Observations were made 13 point-count stations separated >150 m covering edge and
forest interior at each of the three sites. Counts lasted 10 minutes and had a fixed radius of 25 m
— that is where observers detect most species in the dense cloud forest (Ocampo-Peñuela & Pimm
2015). We complemented our efforts with additional data sources to avoid classifying birds as
extirpated due to insufficient sampling (Boakes et al. 2016): two contemporary surveys at San
Antonio (Orejuela-Gartner et al. 2002; Montealegre-Talero 2009), an exploratory survey at the
locality of Monteguadua in the buffer zone of the Farallones de Cali National Park, and eBird data
for San Antonio (eBird Basic Dataset 2017).
Species classification
We classified birds as forest specialists, forest generalists, and non-forest species. Specialists live
only in old-growth forest interior and edges. Generalists are associated with early and late second-
growth forest and edges. Non-forest species inhabit open areas with scattered trees. We assigned
functional groups which incorporate species roles at the ecosystem level (Sekercioglu et al. 2004;
Bregman et al. 2014). Functional groups incorporate information such as foraging strata, diet, and
body mass (e.g., large canopy frugivores), two of them (e.g., Ground Omnivore) or just the trophic
group (e.g., carnivores). We used an a priori classification based on Kattan et al. (1994) and
Renjifo et al. (1999) with some modifications.
We followed Stotz et al. (1996) to assign birds to foraging strata as Aerial, Canopy, Midstory,
Understory, Ground, Water or Multiple strata; and to assign elevational center of abundance as
Lowland, 0-1000 m; Lower Montane, 1000-1700; Middle Montane, 1700-2500 m, Upper
Montane, 2500 m or higher. Species lower and maximum elevational limits were obtained using
the Field Guide to the Birds of Colombia (McMullan & Donegan 2014) and HBW Alive (del Hoyo
et al. 2017), selecting for values that give the widest range for each species within their distribution
in Colombia (but excluding obvious extreme values), to account for the phenotypic plasticity of
birds and be conservative in our estimates on the possible effects of climate change.
Bird body masses were from the CRC Handbook of Avian Body Masses (Dunning 2008) and
HBW Alive (del Hoyo et al. 2017). We grouped birds into four size classes: ≤25 g (I), 25-50 g (II),
50-100 g (III), >100 g (IV). We classified diet types according to the number of food categories
consumed and their estimated proportions (Lopes et al. 2016). Species with an estimated >90% of
food ingested in a single category were classified as Carnivore, Insectivore, Frugivore, Nectarivore
or Granivore. A second category was added if it comprised >10% of the diet (e.g., Frugivore-
Insectivore). Species with three or more food categories in about the same proportion and that
consumed both plant and animal material were considered omnivores (Lopes et al. 2016).

We also classified birds into core or peripheral species based on their biogeographical distribution
(Andes, Pacific, or Cauca Valley) and their elevational center of abundance (Stotz et al. 1996;
Renjifo 1999). Core species had Lower Montane (1000 to 1700 m) to Middle Montane (1700 to
2500 m) Andean distribution. Peripheral species had Lowland (0-1000 m) and Upper Montane
(>2500 m) centers of abundance. We considered birds that occur in the foothills of the Pacific and
Cauca Valley regions as peripheral species.
Changes in community composition
We first assembled the historical core avifauna, i.e., birds that regularly breed or migrate to the
area excluding vagrants (Remsen 1994). To do so, we selected species that had been recorded in
at least one of the three earliest datasets (between 1911-1959) to account for incomplete sampling
(Boakes et al 2016).
We compared community composition in each survey against the historical baseline to document
species extirpations and colonizations. If a species was not recorded in a historical survey but it
was recorded in a later survey, we considered the species to have been present. In case of doubt
(e.g., low detectability), as occurred with Barred Forest-Falcon (Micrastur ruficollis), we also
considered the species as present. Our criteria may underestimate the number of recolonizations
that occurred but kept low our estimates of species extirpations. We classified species as extirpated
up to 2016 if there were fewer than five records the previous ten years, meaning that there are no
viable populations. For example, the Chestnut-crowned Antpitta (Grallaria ruficapilla) is common
and easily detected (Kattan & Beltran 2002), but in the San Antonio region it is seldom recorded.
We calculated the richness-based species-exchange ratio (SERr) (Hillebrand et al. 2018) as a
measure of how many species are exchanged between surveys: SERr = (Simm + Sext)/ Stot, where
Simm = newly immigrated species, Sext = extirpated species from the previous survey, and Stot =
total number of species across both surveys. We calculated the proportion of forest specialists to
forest generalists as a measure of functional homogenization (Clavel et al. 2011).
Species relative abundances
To estimate population trends we used five categories of relative abundance to reconcile the results
of different sampling methods used over the years (Curtis & Robinson 2015; Rosselli et al. 2017;
Freeman et al. 2018): abundant, common, fairly common, uncommon, and rare.
We assigned relative abundances to the different surveys as follows. For Chapman (1917), we used
the distribution of his collected number of individuals. We were able to do so because chapman´s
collecting efforts were unique in that he sought to secure all possible individuals on a given locality
(Kattan et al. 2016), and we caution against using this method for other datasets based on specimen
collections. The distribution was right skewed with a median of 3 and a 75th percentile of 6
collected birds. We then assigned species to relative abundance categories as follows: 1 specimen
- rare or uncommon; 2- uncommon or fairly common; 3 to 6 common; > 6- abundant. We
complemented this criterion with Chapman’s (1917) description of abundances when given.
Because collecting is not an efficient way to survey birds, the relative abundances we estimated
are conservative.

Miller (1963) provided qualitative estimates of bird abundances on most species that we could
assign to our five categories. For the third period, observational data from 1989 to 1990 (Kattan et
al. 1994) were organized in 10-species lists (MacLeod et al. 2011). For the contemporary survey
in 2016, we obtained the abundances from standard point-count analysis using the number of
individuals per hectare. For the two last datasets, we used percentiles to assign species relative
abundances: abundant (>90), common (75 to 90), fairly common (50 to 75), uncommon (25 to 50)
and rare (1 to 25).
Changes in community abundance
We obtained population trends for the years 1911-1990 (forest loss and species extirpations), 1990-
2016 (forest recovery and species recolonizations), and the historical trend for each bird species
after 100 years. Our criterion for a difference in relative abundance was a shift in at least two
categories from the historical to the contemporary (e.g., from fairly common to rare). Otherwise,
the population was considered stable (e.g., from fairly common to uncommon). This makes our
estimates conservative and less likely to reflect minor changes in abundances that may arise from
the natural variation in populations and from methodological artifacts.
We also compared changes in relative categories for the forest bird community by summarizing
information in the historical surveys (from 1911 to 1959, representing an estimate for the early to
mid 20th century) and contemporary surveys (from 2007 to 2016, an estimate for the 21st century).
We did so to make our estimates more robust because few species had relative abundance
categories in all surveys. We conducted a Chi-square test to evaluate whether the abundance
distribution differed between both periods.
Specifically, to obtain the relative abundance categories of the historical avifauna, we used the
relative abundances estimated for Chapman (1917) and Miller (1963), using a survey in a nearby
locality in cases of doubt (Gniadek 1973). We sought for a best estimate based on our expert
knowledge and mostly used the highest available categories because early surveys likely
underestimated species abundances. Moreover, population trends were mostly declining based on
known processes of land use change (see discussion). For the contemporary surveys we used our
data from 2016 complemented with a survey conducted by the regional environmental authority
(CVC 2007). The quantitative survey by Kattan et al. 1994 also helped us clear doubts for assigning
relative abundances in both periods.
Forest Cover changes
We combined multiple sources to estimate forest cover changes between 1911 and 2016. We
assumed a pre-disturbance baseline of ~95% forest cover in 1911, following Chapman´s (1917)
forest descriptions before the construction of the Cali-Buenaventura road. For the mid-century
peak-disturbance time, we digitized a 1977 land-cover and land-use map at scale 1: 10,000 (plate
299-II-B-I) of the Agustin Codazzi Geographical Institute of Colombia, based on aerial
photographs. This map covered the southern portion of San Antonio (28%) where the historical
and contemporary surveys were conducted. It is representative of Miller´s survey in 1959 because
forest cover remained stable through the mid 20th century (Kattan et al. 1994).

To obtain current forest cover, we performed a supervised forest versus non-forest classification,
mosaicking two RapidEye Ortho – Level 3A imagery of 5-m resolution from 22 November 2016
(tile_1840707) and 25 July 2015 (tile_1840607). For this, we performed a top-of-atmosphere
reflectance correction on each image and obtained a mosaic with the most recent image on top.
We masked the clouds and city of Cali polygons and generated regions of interest of pure forest
and non-forest classes using a visual interpretation of the image given our familiarity with the area.
We then randomly selected 70% of each category as training sites and 30% for validation to obtain
the overall accuracy of the classification. We run the supervised classification with Maximum
Likelihood on all five spectral bands available. We then performed a post-classification
generalization using a minimum size of 200 pixels that correspond to a minimum mapping unit of
0.5 ha. We performed all analyses in ENVI 5.3.2. We then determined forest-cover changes by
comparing the 1911 baseline to the forest values of 1977 and 2016 maps.
Landscape and regional analysis
We detected core forest areas of at least 100 ha (Bregman et al. 2014) in our 2016 forest cover
map with the Core Mapper tool from Gnarly Landscape Utilities (Shirk & McRae 2013). It makes
use of a circular moving window with a user-defined area that centers through each pixel. We set
it to assign cores only if all the other pixels within this 100-ha window are classified as forest. We
analyzed connectivity between the core areas using a model that analogizes the movement of
organisms in the landscape to current flow in an electric circuit. Core areas connected through
multiple forested pathways possess less resistance to movement (i.e., electric flow) than others
with few available routes, meaning there are fewer barriers for dispersal.
We used the forest versus non-forest cover as a layer resistance raster, where we set forest cells as
1 (low resistance) and non-forest cells to infinite resistance (no data) because we were interested
in the dispersal of forest birds that only use forested areas for movement. We considered each core
area as a node and calculated the current flow between all pairs (pairwise model) of core areas
where arbitrarily one of them connects to a current source while the other connects to the ground
to identify high voltage gradients. We then generated least cost paths through the forest between
core areas to find optimal routes and evaluated their spatial match with landscape barriers
(deforested areas). We performed these computations with Circuitscape 4.0 (McRae et al. 2008)
and the spatial analysis with ArcGIS 10.4.1.
To investigate geographical patterns of change in species richness we overlapped species ranges
of the historical forest avifauna refined by their elevational amplitude (the difference between the
maximum and minimum elevations). Distribution maps were from obtained from Birdlife
International (2016) and rasterized to a resolution of 30 m (900 m2) per cell size. We also created
richness maps for persisting and extirpated species and divided their values on a per-pixel basis by
the richness values of the historical forest avifauna. These ratios were used to visualize
geographical patterns of change. To analyze differences in current community composition and
structure among the two fragmented (Cerro de la Horqueta, Km-18) and continuous forest (La
Teresita) sites, we performed an analysis of similarity (ANOSIM; Clarke & Warwick 2001), which
generates an R statistic that varies between 0 and 1. High values of R indicate high dissimilarity
among groups. In addition, we used similarity percentage analysis (SIMPER; Clarke & Warwick
2001) to evaluate the quantitative contribution of individual species to community structure among
sites.

Vulnerability to climate disruption
We assessed whether the 122 forest birds that remained are vulnerable to climate disruption based
on three factors that underlie species responses to environmental change (Foden et al. 2013; Fortini
& Schubert 2017): species population trends, habitat specialization, and narrow elevational range
(<1500 m) (Newbold et al. 2012; White & Bennett 2015; Keinath et al. 2017).
We grouped species into six combinations of “declining” versus “stable abundance” and three
degrees of specialization — either habitat specialist or with a narrow elevational range, both factors
together, and none of them. We used a Chi-square test to determine whether there were significant
differences among numbers observed. Furthermore, we tested whether body mass and size of the
elevational range affected population density with an unweighted least squares multiple regression.
We did not model the elevational shifts to predict whether species must move to remain within
existing thermal limits due to the uncertainties associated with these models (McCain & Colwell
2011)
Results
Bird community dynamics: 1911-2016.
In 1911, San Antonio was 95% forested and had 201 species — 105 forest specialists, 62 forest
generalists, and 34 open-area birds (Supplementary material). Forest declined to 48.6% after the
construction of the Cali to Buenaventura road from 1920 to1940. In 1959, 33 forest birds were
extirpated — 26 specialists and seven generalists. Six species from open areas colonized. Forest
cover remained stable through 1990, yet 14 more birds were lost for a total of 47 species or 28%
of the historical forest community. Moreover, 37 species decreased in abundance. There were nine
colonizations of forest birds and four into open areas. From 1911 to 1990, of the 123 remaining
forest birds.
Forest cover increased 26.1% after the late 1990s — a gain of 1270 ha of naturally regenerated
forests to a current extent of 6,065 ha or 57.6% (Classification accuracy was 98.1%, Fig S1).
Recent surveys (2000 to 2016) showed no further extirpations. Instead, 17 species re-established
populations (Fig 3., Appendix, Table S1). Ten were already missing by 1959 but surveys recorded
them again in the early 2000s. An example is the Colombian Chachalaca (Ortalis columbiana),
with a continuous presence in the area since 2003 (eBird Dataset 2016). Nine new forest birds
colonized. Three were forest specialist birds within their elevational range (Drymophila
striaticeps, Xiphorhynchus erythropygius and Lophotriccus pileatus) and the others were forest
generalists. We recorded four new open area species in our records. From 1990 to 2016, 7 species
declined but 24 increased populations and most remained stable (Supplementary material).

Figure 3. Trends in historic number of species in the San Antonio cloud forest (western Andes of Colombia)
relative to forest cover. Bars: forest cover for intervals with data. Circles: bird surveys fitted to a relaxation
formula for the decay in species richness at forest fragments (Halley et al. 2016).
Community changes after one hundred years.
Contemporary surveys showed 31 extirpated forest birds, 24 specialists and seven generalists, for
a total of 19% of the historical avifauna, 23 of them lost since 1959 (Appendix, Table S2). The
richness-based species-exchange ratio (SERr) between surveys ranged from 0.12 to 0.19. Between
the historical and contemporary bird community was 0.24 (24% species exchange). The forest
community became more functionally homogenous. The proportion of specialists to generalists
declined from 63% to 43% (Z= 3.60, p < 0.001). There was a shift from a preponderance of
abundant and common species to a community dominated by uncommon and rare species (X2 =
35.34, p < 0.001; Fig. 4). Compared with historical levels, 43 species had lower relative
abundances, four increased, and 105 were stable (supplementary material).

Figure 4. - Changes in relative abundance categories for birds in the cloud forest of San Antonio, Colombia, in
the historical (early to mid 20th century;1911-1959) and contemporary (21th century; 2007-2016) periods.
We identified 17 functional groups formed by between 3-19 species (Table 1). Eight of them had
50% or more of their species as extirpated or declining in abundance. Large canopy frugivores
declined substantially, whereas half of all nectarivores declined in abundance. All four water
dependent birds were extirpated.
Table 1. Functional groups of the historical bird community of San Antonio (excluding migrants) and the
long-term impacts of land-use change after 100 years.
Functional group
No. species
Extirpated (%)
Declining (%)
Stable (%)
Water dependent*
4
4 (100) *
0 (0)
0 (0)
Large canopy frugivores*
5
3 (60.0) *
2 (40)
0 (0)
Canopy omnivores*
6
3 (50.0) *
2 (33.3)
1 (16.7)
Ground omnivores*
7
1 (14.2)
4 (57.1) *
2 (28.6)
Ground insectivores*
6
2 (33.3)
2 (33.3)
2 (33.3)
Nectarivores*
19
2 (10.5)
10 (52.6) *
7 (36.8)
Canopy granivores*
6
0 (0)
4 (66.7.0) *
2 (33.3)
Carnivores*
4
1 (25.0)
1 (25.0)
2 (50.0)
Midhigh insectivores
19
5 (26.3)
4 (21.1)
10 (52.6)
Medium canopy frugivores
9
3 (33.3)
1 (11.1)
5 (55.6)
Understory omnivores
5
0 (0)
2 (40.0)
3 (60.0)
Large midhigh frugivores
3
1 (33.3)
0 (0)
2 (66.7)
Trunk insectivores
6
1 (16.7)
1 (16.7)
4 (66.7)
Understory insectivores
18
2 (11.1)
4 (22.2)
12 (66.4)
Small canopy insectivores
10
1 (10.0)
2 (20.0)
6 (60.0)
Small canopy frugivores
18
2 (11.1)
3 (16.7)
13 (72.2)
Canopy insectivores
9
0 (0)
1 (11.1)
8 (88.9)
*Functional groups with more than 50% of species extirpated or with declining abundances.

Patterns of landscape change.
We detected three core forest areas of >100-ha in the San Antonio region: Cerro de la Horqueta
(227 ha), the large forest patch called Km-18 (1041 ha), and Chicoral-Dapa (333 ha) to the north
(Fig. 5). The Cali-Buenaventura road divides the first two core areas, and we observed only a
minor difference in species composition that resulted from 4 species not being found in Cerro de
la Horqueta and one not being detected in Km-18.These five species (Table 2) were recorded
recently in the core area of Chicoral-Dapa, according to the eBird basic dataset (2017) and a 2016
checklist (M. Gable, pers. Comm.).
Figure 5. Loss of forest connectivity at the landscape level. (a) forest cover for 2016 based on a supervised
forest/non-forest classification of 5-m resolution satellite imagery. A-F, core forest areas of at least 100 ha; A-
Chicoral-Dapa; B- Km-18; C- Cerro de la Horqueta (San Antonio summit); D- Monteguadua; E- National
Natural Park Farallones de Cali; circles, point-count surveys; star, 1-day inventory at Monteguadua. (b) close-
up of satellite imagery showing least-cost paths through forest between core areas C and E in a largely
fragmented area (c) movement barrier arising from limited forest pathways between core areas C and E, based
on a model of landscape connectivity that analogizes dispersal to current flow in an electric circuit (see methods).

In contrast, seven extirpated species in the San Antonio region were present at the buffer zone of
the Farallones de Cali National Natural Park (Table 2).Four of these species were recorded at La
Teresita (Pyroderus scutatus, Habia cristata, Psarocolius angustifrons, Rupicola peruvianus) and
had been present since the mid 20th century (S. Gniadek 1973, unpublished data). At the locality
of Monteguadua, we recorded three additional extirpated species (Hapaloptila castanea,
Synallaxis unirufa and Grallaria ruficapilla) and one that was absent from the core area of Cerro
de la Horqueta (Cloropipo flavicapilla).
An analysis of similarity (ANOSIM) revealed a significant difference between our control site of
La Teresita, and the localities surveyed at San Antonio mainly because of the absences (R = 0.46,
P = 0.001). A SIMPER analysis also revealed differences in community structure but were mostly
related with variation in the local abundance of common species among localities, where between
10 and 13 species represented 80% of the cumulative dissimilarity in pairwise comparisons.
Table 2. Compositional differences detected between the three main core areas of the San Antonio region. A-
Chicoral-Dapa; B- Km-18; C- Cerro de la Horqueta (San Antonio summit); D- PNN Farallones de Cali Buffer
Zone (localities of La Teresita, Monteguadua). Some species are extirpated throughout the whole region.
Family
Scientific name
English name
Found
at
Absent from
Trogonidae
Pharomachrus antisianus
Crested Quetzal
A, B
C
Bucconidae
Hapaloptila castanea
White-faced nunbird
D
Extirpated
Grallaridae
Grallaria ruficapilla
Chestnut-crowned antpitta
D
Extirpated
Furnariidae
Sclerurus mexicanus
Tawny-throated Leaftosser
A, B
C
Furnariidae
Synallaxis unirufa
Rufous spinetail
D
Extirpated
Tyrannidae
Phyllomyias plumbeiceps
Plumbeous-crowned Tyrannulet
A, B
C
Cotingidae
Rupicola peruvianus
Andean Cock-of-the-rock
D
Extirpated
Cotingidae
Pyroderus scutatus
Red-ruffed fruitcrow
D
Extirpated
Pipridae
Chloropipo flavicapilla
Yellow-headed Manakin
D
C
Polioptilidae
Ramphocaenus melanurus
Long-billed gnatwren
A, C
B
Cardinalidae
Habia cristata
Crested Ant-tanager
D
Extirpated
Patterns of regional change
Eighty-one percent of extirpated birds were peripheral species with source populations in adjacent
biotic regions. The highest ratios of extirpated species concentrated in the Pacific lowlands,
whereas stable species were widespread (Fig. 6). We attributed this difference to a <5 km
deforestation gap. It was present in our 1977 map and remained a prevalent feature confirmed by
visual inspection using the Timelapse of Google Earth Engine from the years 1984 to 2016. This
gap was in a critical connecting point for both the Pacific lowlands and Andean highlands, where
the highest species richness of the historical avifauna occurred (Fig. 1). Our connectivity analysis
showed this gap is the major barrier for dispersal between the two regions due to limited available
pathways to species movement through forests.

Figure 6. (a) Extirpated species concentrated in the Pacific lowlands as a percentage of expected species
richness and (b) stable species without a detectable abundance change over time as a percentage of expected
species richness.
Species vulnerability to climate disruption
Of the 19 species that are habitat specialists and also have narrow elevational ranges, 10 (53%) are
declining compared to 35% overall. They include four species that Colombia’s bird Red List
considers to be vulnerable (Renjifo et al. 2014): the endemic Multicoloured Tanager
(Chlorochrysa nitidissima), the near-endemic Yellow-headed Manakin (Choropipo flavicapilla),
Cloud-forest Pygmy-Owl (Glaucidium nubicola) and Rufous-crested Tanager (Creurgops
verticalis). For the entire bird community, there was no effect of body mass or elevational range
on population density (F2,135 = 0.8, P = 0.44; Fig. 7) meaning that large-bodied and species with
narrow vertical ranges are not less abundant.
Figure 7. Elevational range and body mass relationships to population density (ind/ha)

Discussion
“… and that everything written on them was unrepeatable since time immemorial and forever
more, because lineages condemned to one hundred years of solitude did not have a second
opportunity on earth”. Gabriel García Márquez, One hundred Years of Solitude (1967).
We revealed the impacts of land-use change over a century (1911-2016) on the dynamics of the
bird community in the cloud forest of San Antonio. Species were extirpated when forest cover
declined to ~50% after the construction of a major road from 1920 to 1940. This a well-known
pattern where habit loss follows development projects (Rudel & Roper 1997; Etter et al. 2006).
Forest cover remained stable up to 1990, but extirpations continued (Kattan et al. 1994). Extinction
debt i.e., the temporal lag of species losses after land-use change occurred, is likely the best
explanation to our findings (Brooks et al. 1999; Ferraz et al. 2003; Halley et al. 2016). Forest cover
recovered starting in the late 1990s, a general trend observed in the Colombian Andes (Sanchez-
Cuervo et al. 2012; Aide et al. 2013). There were no further species losses and several species
recolonized. More extirpations would surely have occurred had there been less forest remaining
and it had been more extensively fragmented (Rybicki & Hanski 2013). Nevertheless, our results
are in line with recent estimates that suggest loss of biodiversity when forest cover values drop
below 50% (Martensen et al. 2012; Morante-Filho et al. 2015).
Other factors help explain species losses. The toucan barbet (Semnornis ramphastinus) was heavily
traded by the 1950s (Lehmann 1957) and hunting affected species such as the Pale-vented Pigeon
(Patagioenas cayennensis). But given the deforestation patterns in the region, land-use changes
are the most parsimonious explanation for species extirpations and colonizations, with these other
factors having added impacts (Brook et al. 2008). Could climate change have explained the loss
of species up to 1990? Even though studies have found climate change effects in tropical birds
(Blake & Loiselle 2015), we found that in San Antonio the extirpated species were not the most
vulnerable to climate disruption, and population declines were linked to habitat loss. Moreover, it
would be surprising if species extirpations stopped or reversed thereafter if climate change was the
cause.
We argue that our results on species extirpations are conservative. Collecting efforts of the early
20th century yielded more species than later surveys, even though sampling methods have
improved, and survey effort increased. Thus, other extirpations may have gone unnoticed. Several
species also recolonized. Were they really extirpated? The evidence suggests so. Most species have
been lost in other well-studied fragments in the central Andes (Renjifo 1999; Castaño-Villa &
Patiño-Zabala 2008). Others are highly vocal and detectable when present but were not recorded
since the middle 20th century (Appendix, Table S1). However, some species might have survived
in the agriculture-dominated mosaics between remnant forests. Yet claims of “substantial
biodiversity” in these mosaics touted by those who stress “countryside conservation”
e.g.,Mendenhall et al. (2012) are hard to reconcile with the overwhelmingly dominant role of forest
loss in driving species extirpations.

Our work shed light on 3 important conservation topics that transcended our case-study. First, our
findings factor in the debate on whether local species losses are offset by colonizations, such that
there might be no net biodiversity loss (Cardinale et al. 2018). Compared with the historical bird
community in 1911, San Antonio lost 31 forest birds — 24 specialists and 7 generalists. It gained
only 18 forest birds — 7 specialists and 11 generalists. Simply, there was a decline in species
richness and there were far fewer forest specialists as well. At least 14 non-forest species have
colonized due to the increase of clearings, but it is the change in forest birds that addresses the
question at hand, not the overall species richness (Primack et al. 2018).
We also found that species have declined in abundance over time (Fig. 4). This trend may be an
artifact of different sampling methods, but again, our methods were conservative. We required
changes in two classes of relative abundance as evidence of decline or increase. There were many
more declines than increases. Moreover, the raw numbers of collected individuals also indicated
the depletion in abundance for most of the forest community. For instance, Chapman (1917)
secured eight individuals of the Golden-headed quetzal (Pharomachrus auriceps), the Red-ruffed
fruitcrow (Pyroderus scutatus) and six individuals of the Black-billed Mountain-Toucan
(Andigena nigrirostris). The former is now uncommon and the latter two are extirpated. Other
changes mirror well-known declines elsewhere. Our estimates are consistent with the documented
trend of the Canada Warbler (Cardellina canadensis), which has declined 66% in the last 50 years
(Sauer et al. 2017). In our classification, this migrant went from abundant in Chapman’s time
(1911), to fairly common in 1990, and to uncommon in our contemporary survey (2015-2016).
Second, we investigated whether regional habitat loss influences local species extirpations. This
involves different processes at a range of geographical scales, and we chose to study the spatial
arrangement of the extirpated species ranges. A key result of our study is that the loss of forest
continuity with adjacent regions is a major cause of extinction. Because San Antonio is a mid-
elevation mountain ridge in the Andes, the loss of forest continuity with the adjacent regions
(Andean highlands, Pacific and Cauca Valley lowlands) resulted in a differential loss of peripheral
species. Several of these species are still present regionally in nearby localities of the Farallones
de Cali National Park, and further explorations may find others (Palacio et al. 2017). Although
populations in the periphery of their ranges may not be intrinsically vulnerable (Channell &
Lomolino 2000), a decrease in connectivity limits the dispersal of species whose source
populations are found in other regions or located at lower or higher elevations. Thus, peripheral
species whose ranges only partially overlap with a particular region may be more prone to
extirpation (Terborgh & Winter 1983; Nathan et al. 1996).
Third, even assuming no future loss of forest cover at San Antonio, species losses from climate
disruption are likely and their causes complex (Pimm 2009). Rising temperatures will force species
to higher elevations as they have done elsewhere in the Andes (Forero-Medina et al. 2011). To
complicate matters, an analysis of recent trends for the region shows an increase in the number of
dry days (Cardona-Guerrero et al. 2014). In the late 1950s, the vegetation was constantly wet and
dripping (Miller 1963), but this is no longer the case (GHK, pers. obs.). The diversion of water for
urbanization (Kattan et al. 1994) has already contributed to the extirpation of water-specialist birds
such as the Crested Ant-Tanager (Habia cristata), a bird found near streams, and others that
depend on humid conditions such as the Fulvous-dotted Treerunner (Margarornis stellatus).
Furthermore, there are projected precipitation changes across the tropics that may increase
extirpation risks in mountain sites (McCain & Colwell 2011).

Our vulnerability analysis showed that most species that are intrinsically vulnerable to climate
change, based on their ecological and biological traits, have declined in abundance (Table 1).
Although the numbers are too small for statistical significance, they do suggest that land-use
changes have left the remaining bird species in a perilous state, rather than hardened it against
future losses. For instance, many species with narrow elevational ranges are also forest specialists.
They are currently below their historical abundances and their populations were declining. We
found no relationship between elevational range and population density, meaning that these past
changes are independent of future threats. Also, there were other forest birds that were not as
sensitive to climatic disruptions, but contemporary results showed they are at immediate risk of
extirpation because they are rare and decreasing. The overall effect is increasing the vulnerability
of the community to new drivers of change.
The vulnerability of the cloud forest community in San Antonio was most apparent when in the
analysis of functional groups. Eight out of 17 functional groups had 50% or more of their species
extirpated or declining in abundance. Canopy association, large body size, and ecological
specialization were the main traits shared. Large-bodied frugivores have lost several species, yet
they are critical to maintaining seed-dispersal processes (Palacio et al. 2016). Moreover,
nectarivores (hummingbirds and flowerpiercers) emerged in our study as particularly vulnerable.
More than 50% have declined in abundance, and two were extirpated — White-tipped Sicklebill
(Eutoxeres aquila) and Green-fronted Lancebill (Doryfera ludovicae). Bird nectarivores are
essential pollinators in Andean mountain forests (Renjifo et al. 1997). Therefore, future species
losses, if not avoided, may profoundly alter forest dynamics and capacity for provisioning of
ecosystem services (Sekercioglu et al. 2004).
Our study revealed the impacts of one hundred years of land-use change in a local cloud forest
faunal assemblage. There were important losses that included extirpations, population declines,
and increased vulnerability to future threats. But there were also important gains as the forest
recovered, including recolonizations and population increases. Our results suggest that an effective
way to conserve biodiversity during climate disruption will be to counter the past effects of
landscape transformation by increasing the extent and connectivity of remaining habitat patches at
local and landscape scales. These actions may give disappearing cloud forest lineages a second
opportunity on Earth.
Acknowledgments
We thank Jose Luna-Solarte, Giancarlo Ventolini, John Restrepo, Anderson Muñoz and Jenny
Muñoz for field assistance, and Eduardo Carvajal and Corporación Autónoma Regional del Valle
del Cauca (CVC) for logistical support. The Pontificia Universidad Javeriana Cali funded
fieldwork.

Literature Cited
Aide TM, Clark ML, Grau HR, López-Carr D, Levy MA, Redo D, Bonilla-Moheno M, Riner G,
Andrade-Núñez MJ, Muñiz M. 2013. Deforestation and reforestation of Latin America
and the Caribbean (2001-2010). Biotropica 45:262–271.
Anderson AS, Storlie CJ, Shoo LP, Pearson RG, Williams SE. 2013. Current analogues of future
climate indicate the likely response of a sensitive montane tropical avifauna to a warming
world. PLoS ONE 8:e69393.
Blake JG, Loiselle BA. 2015. Enigmatic declines in bird numbers in lowland forest of eastern
Ecuador may be a consequence of climate change. PeerJ 3:e1177.
Boakes EH, Fuller RA, McGowan PJK, Mace GM. 2016. Uncertainty in identifying local
extinctions: the distribution of missing data and its effects on biodiversity measures.
Biology Letters 12:20150824.
Boyle WA, Sigel BJ. 2015. Ongoing changes in the avifauna of La Selva Biological Station,
Costa Rica: Twenty-three years of Christmas Bird Counts. Biological Conservation
188:11–21.
Bregman TP, Sekercioglu CH, Tobias JA. 2014. Global patterns and predictors of bird species
responses to forest fragmentation: Implications for ecosystem function and conservation.
Biological Conservation 169:372–383.
Brook B, Sodhi N, Bradshaw C. 2008. Synergies among extinction drivers under global change.
Trends in Ecology & Evolution 23:453–460.
Brooks TM, Pimm SL, Oyugi JO. 1999. Time Lag between Deforestation and Bird Extinction in
Tropical Forest Fragments. Conservation Biology 13:1140–1150.
Cardinale BJ, Gonzalez A, Allington GRH, Loreau M. 2018. Is local biodiversity declining or
not? A summary of the debate over analysis of species richness time trends. Biological
Conservation 219:175–183.
Castaño-Villa GJ, Patiño-Zabala JC. 2008. Local extinctions of birds in forest fragments in the
Santa Elena region, central Andes, Colombia. Hornero 23:23–34.
Channell R, Lomolino MV. 2000. Trajectories to extinction: spatial dynamics of the contraction
of geographical ranges. Journal of Biogeography 27:169–179.
Chapman FM. 1917. The distribution of bird-life in Colombia : a contribution to a biological
survey of South America. Bulletin of the American Museum of Natural History 36:1–
729.
Clarke KR, Warwick RM. 2001. Change in marine communities: an approach to statistical
analysis and interpretation. PRIMER-E:Plymouth.
Clavel J, Julliard R, Devictor V. 2011. Worldwide decline of specialist species: toward a global
functional homogenization? Frontiers in Ecology and the Environment 9:222–228.
Curtis JR, Robinson WD. 2015. Sixty years of change in avian communities of the Pacific
Northwest. PeerJ 3:e1152.
CVC. 2007. Evaluar el estado de la fauna y flora del bosque de San Antonio, área de interés
ambiental de la cuenca del río Cali, municipio de Cali. Informe Final. Convenio CVC
051 de 2007. Corporación Autónoma Regional del Valle del Cauca-CVC, Dirección
Técnica Ambiental, Grupo Biodiversidad y Asociación Río Cali-ASORÍOCALI.
Santiago de Cali, 96 pp.
de Schauensee RM. 1948. The Birds of the Republic of Colombia: Their Distribution and Keys
for Their Identification. Instituto de Ciencias Naturales de la Universidad Nacional de
Colombia. Available from https://books.google.com/books?id=jNmmjwEACAAJ.

del Hoyo J, Elliot A, Sargatal J, Christie DA, de Juana E, editors. 2017. Handbook of the Birds
of the World Alive. Lynx Edicions, Barcelona. Available from <http://www.hbw.com/>.
Dornelas M, Gotelli NJ, McGill B, Shimadzu H, Moyes F, Sievers C, Magurran AE. 2014.
Assemblage time series reveal biodiversity change but not systematic loss. Science
344:296–299.
Dunning JB, editor. 2008. CRC handbook of avian body masses2nd ed. CRC Press, Boca Raton.
eBird Basic Dataset. 2017. Version: EBD_relMay-2017. Cornell Lab of Ornithology, Ithaca,
New York. May 2017.
Essl F, Dullinger S, Rabitsch W, Hulme PE, Pyšek P, Wilson JRU, Richardson DM. 2015.
Historical legacies accumulate to shape future biodiversity in an era of rapid global
change. Diversity and Distributions 21:534–547.
Etter A, McAlpine C, Pullar D, Possingham H. 2006. Modelling the conversion of Colombian
lowland ecosystems since 1940: Drivers, patterns and rates. Journal of Environmental
Management 79:74–87.
Ferraz G, Russell GJ, Stouffer PC, Bierregaard RO, Pimm SL, Lovejoy TE. 2003. Rates of
species loss from Amazonian forest fragments. Proceedings of the National Academy of
Sciences 100:14069–14073.
Foden WB et al. 2013. Identifying the world’s most climate change vulnerable species: A
systematic trait-based assessment of all birds, amphibians and corals. PLoS ONE
8:e65427.
Forcey GM, Anderson JT, Ammer FK, Whitmore RC. 2006. Comparison of two double-observer
point-count approaches for estimating breeding bird abundance. Journal of Wildlife
Management 70:1674–1681.
Fortini L, Schubert O. 2017. Beyond exposure, sensitivity and adaptive capacity: a response
based ecological framework to assess species climate change vulnerability. Climate
Change Responses 4:1–7.
Freeman BG, Scholer MN, Ruiz-Gutierrez V, Fitzpatrick JW. 2018. Climate change causes
upslope shifts and mountaintop extirpations in a tropical bird community. Proceedings of
the National Academy of Sciences:201804224.
Gniadek S. 1973. Las aves de “La Margarita”, Peñas Blancas, río Pichindé, Cali, Valle del
Cauca. Page in S. Sarria-Salas, editor. Parque Nacional Natural Farallones de Cali.
Monografía. Corporación Autónoma Regional del Valle del Cauca-CVC y Fundación
Protectora de las Cuencas-PROCUENCAS, Santiago de Cali.
Gonzalez A, Cardinale BJ, Allington GRH, Byrnes J, Arthur Endsley K, Brown DG, Hooper
DU, Isbell F, O’Connor MI, Loreau M. 2016. Estimating local biodiversity change: a
critique of papers claiming no net loss of local diversity. Ecology 97:1949–1960.
Haddad NM et al. 2015. Habitat fragmentation and its lasting impact on Earth’s ecosystems.
Science Advances 1:e1500052–e1500052.
Halley JM, Monokrousos N, Mazaris AD, Newmark WD, Vokou D. 2016. Dynamics of
extinction debt across five taxonomic groups. Nature Communications 7:12283.
Hazzi NA, Moreno JS, Ortiz-Movliav C, Palacio RD. 2018. Biogeographic regions and events of
isolation and diversification of the endemic biota of the tropical Andes. Proceedings of
the National Academy of Sciences 115:7985–7990.
Helmer EH, Gerson EA, Baggett LS, Bird BJ, Ruzycki TS, Voggesser SM. 2019. Neotropical
cloud forests and páramo to contract and dry from declines in cloud immersion and frost.
PLOS ONE 14:e0213155.

Hillebrand H et al. 2018. Biodiversity change is uncoupled from species richness trends:
Consequences for conservation and monitoring. Journal of Applied Ecology 55:169–184.
Hilty SL, Brown WL. 1986. A guide to the birds of Colombia. Princeton University Press,
Princeton, N.J.
Kattan GH, Alvarez-Lopez H, Giraldo M. 1994. Forest fragmentation and bird extinctions: San
Antonio eighty years later. Conservation Biology 8:138–146.
Kattan GH, Beltran JW. 2002. Rarity in antpittas: territory size and population density of five
Grallaria spp. in a regenerating habitat mosaic in the Andes of Colombia. Bird
Conservation International 12.
Kattan GH, Tello SA, Giraldo M, Cadena CD. 2016. Neotropical bird evolution and 100 years of
the enduring ideas of Frank M. Chapman. Biological Journal of the Linnean Society
117:407–413.
Keinath DA, Doak DF, Hodges KE, Prugh LR, Fagan W, Sekercioglu CH, Buchart SHM,
Kauffman M. 2017. A global analysis of traits predicting species sensitivity to habitat
fragmentation: Species sensitivity. Global Ecology and Biogeography 26:115–127.
Krabbe N, Schulenberg TS. 1997. Species Limits and Natural History of Scytalopus tapaculos
(Rhinocryptidae), with Descriptions of the Ecuadorian Taxa, including Three New
Species. Ornithological Monographs:46–88.
Kuussaari M et al. 2009. Extinction debt: a challenge for biodiversity conservation. Trends in
Ecology & Evolution 24:564–571.
Latta SC, Brouwer NL, Olivieri A, Girard-Woolley J, Richardson JF. 2017. Long-term
monitoring reveals an avian species credit in secondary forest patches of Costa Rica.
PeerJ 5:e3539.
Laurance WF, Camargo JLC, Luizão RCC, Laurance SG, Pimm SL, Bruna EM, Stouffer PC,
Bruce Williamson G, Benítez-Malvido J, Vasconcelos HL. 2011. The fate of Amazonian
forest fragments: A 32-year investigation. Biological Conservation 144:56–67.
Lees AC, Naka LN, Aleixo A, Cohn-Haft M, Piacentini V de Q, Santos MPD, Silveira LF. 2014.
Conducting rigorous avian inventories: Amazonian case studies and a roadmap for
improvement. Revista Brasileira de Ornitologia-Brazilian Journal of Ornithology 22:14.
Lehmann FC. 1957. Contribuciones al estudio de la fauna de Colombia XII. Novedades
Colombianas 3:101–156.
Lepage D, Vaidya G, Guralnick R. 2014. Avibase – a database system for managing and
organizing taxonomic concepts. ZooKeys 420:117–135.
Lopes LE, Fernandes AM, Medeiros MCI, Marini MÂ. 2016. A classification scheme for avian
diet types. Journal of Field Ornithology 87:309–322.
MacLeod R, Herzog SK, Maccormick A, Ewing SR, Bryce R, Evans KL. 2011. Rapid
monitoring of species abundance for biodiversity conservation: Consistency and
reliability of the MacKinnon lists technique. Biological Conservation 144:1374–1381.
Mantyka-Pringle CS, Martin TG, Rhodes JR. 2012. Interactions between climate and habitat loss
effects on biodiversity: a systematic review and meta-analysis. Global Change Biology
18:1239–1252.
Martensen AC, Ribeiro MC, Banks-Leite C, Prado PI, Metzger JP. 2012. Associations of Forest
Cover, Fragment Area, and Connectivity with Neotropical Understory Bird Species
Richness and Abundance. Conservation Biology 26:1100–1111.

McCain CM, Colwell RK. 2011. Assessing the threat to montane biodiversity from discordant
shifts in temperature and precipitation in a changing climate: Climate change risk for
montane vertebrates. Ecology Letters 14:1236–1245.
McMullan M, Donegan T. 2014. Field guide to the birds of Colombia.2nd Edition. Fundación
Proaves de Colombia, Bogotá, Colombia.
McRae BH, Dickson BG, Keitt TH, Shah VB. 2008. Using circuit theory to model connectivity
in ecology, evolution, and conservation. Ecology 89:2712–2724.
Mendenhall CD, Daily GC, Ehrlich PR. 2012. Improving estimates of biodiversity loss.
Biological Conservation 151:32–34.
Miller AH. 1963. Seasonal activity and ecology of the avifauna of an American equatorial cloud
forest. University of California Publications in Zoology 66:1–74.
Montealegre-Talero C. 2009. Patrones de diversidad en ensamblajes de aves en dos bosques de
niebla de la Cordillera Occidental de Colombia. Trabajo de Grado. Universidad de los
Andes. Facultad de Ciencias. Departamento de Ciencias Biológicas., Bogotá D.C.,
Colombia.
Morante-Filho JC, Faria D, Mariano-Neto E, Rhodes J. 2015. Birds in Anthropogenic
Landscapes: The Responses of Ecological Groups to Forest Loss in the Brazilian Atlantic
Forest. PLOS ONE 10:e0128923.
Nathan R, Safriel UN, Hadoram S. 1996. Extinction and vulnerability to extinction at
distribution peripheries: an analysis of the israeli breeding avifauna. Israel Journal of
Zoology 42:361–383.
Newbold T, Scharlemann JPW, Butchart SHM, Sekercioglu CH, Alkemade R, Booth H, Purves
DW. 2012. Ecological traits affect the response of tropical forest bird species to land-use
intensity. Proceedings of the Royal Society B: Biological Sciences 280:20122131–
20122131.
Ocampo-Peñuela N, Pimm SL. 2015. Elevational ranges of montane birds and deforestation in
the Western Andes of Colombia. PLOS ONE 10:e0143311.
Oliver TH, Morecroft MD. 2014. Interactions between climate change and land use change on
biodiversity: attribution problems, risks, and opportunities. Wiley Interdisciplinary
Reviews: Climate Change 5:317–335.
Opdam P, Wascher D. 2004. Climate change meets habitat fragmentation: linking landscape and
biogeographical scale levels in research and conservation. Biological Conservation
117:285–297.
Orejuela-Gartner JE, Malagon LI, Amezquita DF, Bustamante BE, Piedrahita ID, Soto IC,
Restrepo S. 2002. Uso de la avifauna como indicadora de diversidad biológica en
fragmentos de bosque de la zona andina del Valle del Cauca. Informe Final de
Investigación. Corporación Universitaria Autónoma de Occidente CEADES y Programa
Administración del Medio Ambiente. Cali, Colombia.
Palacio RD, Valderrama-Ardila C, Kattan GH. 2016. Generalist species have a central role in a
highly diverse plant-frugivore network. Biotropica 48:349–355.
Palacio RD, Vidal-Astudillo V, Cárdenas G, Luna-Solarte J. 2017. Aves del municipio de
Santiago de Cali, Valle del Cauca, Colombia. Biota Colombiana 18:228–238.
Pearce-Higgins JW, Green R. 2014. Birds and climate change: impacts and conservation
responses. Cambridge University Press, Cambridge, United Kingdom.
Pimm SL. 2008. Biodiversity: Climate change or habitat loss — Which will kill more species?
Current Biology 18:R117–R119.

Pimm SL. 2009. Climate disruption and biodiversity. Current Biology 19:R595–R601.
Pimm SL, Jenkins CN, Abell R, Brooks TM, Gittleman JL, Joppa LN, Raven PH, Roberts CM,
Sexton JO. 2014. The biodiversity of species and their rates of extinction, distribution,
and protection. Science 344:1246752–1246752.
Primack RB, Miller-Rushing AJ, Corlett RT, Devictor V, Johns DM, Loyola R, Maas B,
Pakeman RJ, Pejchar L. 2018. Biodiversity gains? The debate on changes in local- vs
global-scale species richness. Biological Conservation 219:A1–A3.
Remsen JV. 1994. Use and Misuse of Bird Lists in Community Ecology and Conservation. The
Auk 111:225–227.
Remsen JV et al. 2017. Version 22 April 2017. A classification of the bird species of South
America. American Ornithologists’ Union.
http://www.museum.lsu.edu/~Remsen/SACCBaseline.htm.
Renjifo LM. 1999. Composition changes in a subandean avifauna after long-term forest
fragmentation. Conservation Biology 13:1124–1139.
Renjifo LM, Servat GP, Goerck JM, Loiselle BA, Blake JG. 1997. Patterns of species
composition and endemism in the northern Neotropics: A case for conservation of
montane avifaunas. Ornithological Monographs:577–594.
Rosselli L, Stiles FG, Camargo PA. 2017. Changes in the avifauna in a high Andean cloud forest
in Colombia over a 24-year period. Journal of Field Ornithology 88:211–228.
Rudel T, Roper J. 1997. The paths to rain forest destruction: Crossnational patterns of tropical
deforestation, 1975–1990. World Development 25:53–65.
Rybicki J, Hanski I. 2013. Species-area relationships and extinctions caused by habitat loss and
fragmentation. Ecology Letters 16:27–38.
Sanchez-Cuervo AM, Aide TM, Clark ML, Etter A. 2012. Land cover change in Colombia:
Surprising forest recovery trends between 2001 and 2010. PLoS ONE 7:e43943.
Sauer JR, Niven DK, Hines JE, Ziolkowski, Jr DJ, Pardieck KL, Fallon JE, Link WA. 2017. he
North American Breeding Bird Survey, Results and Analysis 1966 - 2015. Version
2.07.2017. USGS Patuxent Wildlife Research Center, Laurel, MD.
Sekercioglu CH, Daily GC, Ehrlich PR. 2004. Ecosystem consequences of bird declines.
Proceedings of the National Academy of Sciences 101:18042–18047.
Sekercioglu CH, Schneider SH, Fay JP, Loarie SR. 2008. Climate Change, Elevational Range
Shifts, and Bird Extinctions. Conservation Biology 22:140–150.
Shaw DW, Escalante P, Rappole JH, Ramos MA, Oehlenschlager RJ, Warner DW, Winker K.
2013. Decadal changes and delayed avian species losses due to deforestation in the
northern Neotropics. PeerJ 1:e179.
Shirk AJ, McRae BH. 2013. Gnarly Landscape Utilities: Core Mapper User Guide. The Nature
Conservancy, Fort Collins, CO. Available at: http://www.circuitscape.org/gnarly-
landscape-utilities.
Sirami C, Caplat P, Popy S, Clamens A, Arlettaz R, Jiguet F, Brotons L, Martin J-L. 2017.
Impacts of global change on species distributions: obstacles and solutions to integrate
climate and land use: Land-use and climate change integration. Global Ecology and
Biogeography 26:385–394.
Sodhi NS, Liow LH, Bazzaz FA. 2004. Avian extinctions from tropical and subtropical forests.
Annual Review of Ecology, Evolution, and Systematics 35:323–345.
Stotz DF, Conservation International, Field Museum of Natural History, editors. 1996.
Neotropical birds: ecology and conservation. University of Chicago Press, Chicago.

Terborgh J, Winter B. 1983. A method for siting parks and reserves with special reference to
Colombia and Ecuador. Biological Conservation 27:45–58.
Vellend M, Baeten L, Myers-Smith IH, Elmendorf SC, Beausejour R, Brown CD, De Frenne P,
Verheyen K, Wipf S. 2013. Global meta-analysis reveals no net change in local-scale
plant biodiversity over time. Proceedings of the National Academy of Sciences
110:19456–19459.
White RL, Bennett PM. 2015. Elevational Distribution and Extinction Risk in Birds. PLOS ONE
10:e0121849.
Willrich G, Calsavara LC, Lima MR, de Oliveira RC, Bochio GM, Medina-Rosa GL, Muz VC,
dos Anjos L. 2016. Twenty-three years of bird monitoring reveal low extinction and
colonization of species in a reserve surrounded by an extremely fragmented landscape in
southern Brazil. Revista Brasileira de Ornitologia 24:235–259.

APPENDIX
Table S1. Seventeen species that were extirpated until 1959 and recovered populations after the 21st century.
Family
Scientific name
English name
Observations
Tinamidae
Nothocercus bonapartei
Highland Tinamou
Lost to Kattan et al. 1994. Not recorded by Gniadek (1973). Very
vocal; high detectability when present, despite seldom seen.
Cracidae
Ortalis columbiana
Colombian Chachalaca
Lost to Miller 1963. Not recorded by Gniadek (1973). Noisy and
conspicuous; extirpated due to hunting. Increasing populations -
now Common to Abundant. However, few remaining habitat.
Should be considered as Endangered (Ocampo-peñuela & Pimm
2017)
Columbidae Patagioenas cayennensis Pale-vented Pigeon Lost to Miller 1963. Extirpated due to hunting. Lehman (1970)
considered extirpated.
Trochilidae Calliphlox mitchellii Purple-throated Woodstar Not recorded by Gniadek (1973). Lost to Kattan et al. 1994. EF**
(C.mulsant). Should be considered as Vulnerable (Ocampo-peñuela
& Pimm 2017)
Trochilidae Thalurania colombica Crowned Woodnymph Lost to Miller 1963. Presumed altitudinal migrant (HBW Alive)
Strigidae Glaucidium nubicola Cloud-forest Pygmy-Owl Lost to Kattan et al. 1994. Was declared in 2010 as extirpated
despite targeted efforts to find it (Fierro-Calderón & Montealegre
2010). New records since 2016 (eBird)
Trogonidae Pharomachrus antisianus Crested Quetzal Lost to Miller 1963. He states the other sympatric quetzal, P.
auriceps, was “becoming very scarce with the cutting and
disturbance of the heavy forests.” Kattan et al. 1994 estimated to
only 5 pairs in 600 ha.
Picidae Melanerpes formicivorus Acorn Woodpecker Lost to Kattan et al. 1994. Distribution in Northen Andes limited
by oaks (Freeman & Mason 2015), which have been extensively
logged in Colombia.
Psittacidae Bolborhynchus lineola Barred Parakeet Lost to Kattan et al. 1994. Requires altitudinal connectivity.
Psittacidae Amazona mercenarius Scaly-naped Parrot Lost to Miller 1963. Lehmann (1957) reports extirpation of the sp.
in the nearby regions. Also was lost from another locality in the
Farallones area (Bermúdez-Vera et al. 2013)
Furnariidae Sclerurus mexicanus Tawny-throated Leaftosser Lost to Miller 1963. Not recorded by Gniadek (1973). “Disappears
from fragmented or selectively logged forests” (HBW Alive).
Depends on large trees, uses roots as dormitories (Campos-Silva
2013)
Tyrannidae Phyllomyias plumbeiceps Plumbeous-crowned Tyrannulet Lost to Miller 1963. Not recorded by Gniadek (1973). EF*, ER**
(P.nigrocapillus and P. cinereiceps)
Troglodytidae Pheugopedius mystacalis Whiskered Wren Lost to Kattan et al. 1994. EF** Also was lost from another
locality in the Farallones area (Bermúdez-Vera et al. 2013)
Polioptilidae
Ramphocaenus melanurus
Long-billed Gnatwren
Lost to Kattan et al. 1994. ER*, ER**
Thraupidae
Creurgops verticalis
Rufous-crested Tanager
Lost to Miller 1963. EF*, ER**
Thraupidae
Sphenopsis frontalis
Oleaginous Hemispingus
Lost to Miller 1963. EF**
Icteridae
Icterus chrysater
Yellow-backed Oriole
Lost to Miller 1963. Was affected by Illegal Wild Trade (cage bird)

Table S2. Thirty-one extirpated species at San Antonio after 100 years of surveys and comparison (when given) with two studies from the central Andes of
Colombia. An asterisk next to the scientific name indicates species with potential to recover populations if connectivity to other regions is restored.
Family Scientific name English name Last records Observations
Tinamidae Tinamus tao* Gray Tinamou 5< records
(2006-2016)
Vulnerable (IUCN 2017). Already extirpated
from the breeding avifauna by Miller (1963)
due to hunting.
Cracidae Penelope perspicax Cauca Guan 1911 Extirpated from most localities in the
Western Andes. Endangered (Renjifo et al.
2014)
Columbidae Patagioenas subvinacea* Ruddy Pigeon 1911 Extirpated In Fragments (Renjifo 1999)
Trochilidae Eutoxeres aquila* White-tipped Sicklebill 5< records
(2006-2016)
Trochilidae Doryfera ludovicae* Green-fronted Lancebill 5< records
(2006-2016)
Rallidae Laterallus albigularis White-throated Crake 1911
Breeds in humid sites and is associated with
wet cloudforest (HBW Alive)
Rallidae Mustelirallus colombianus Colombian Crake 1959 Recorded in Farallones by Gniadek (1973)
but not recorded since. Data Deficient
(IUCN 2017)
Accipitridae Accipiter striatus* Sharp-shinned Hawk 5< records
(2006-2016)
Bucconidae
Hapaloptila castanea*
White-faced Nunbird
1959
Capitonidae
Semnornis ramphastinus*
Toucan Barbet
1959
Ramphastidae
Ramphastos ambiguus
Yellow-throated Toucan
1911
Ramphastidae
Andigena nigrirostris*
Black-billed Mountain-Toucan
1911
Picidae Campephilus melanoleucus* Crimson-crested Woodpecker 5< records
(2006-2016)
Grallaridae Grallaria ruficapilla* Chestnut-crowned Antpitta 5< records
(2006-2016)
Already declining by mid 20TH century in
the Cauca Valley (Lehmann 1970)
Near Threatened (Renjifo et al. 2014).
Lehman (1957) warned they could go
extirpated in the region due to trade. In
Farallones recorded by Gniadek (1973). Lost
from another locality in the Farallones area
(Bermúdez-Vera et al. 2013)
Extirpated in the Cauca Valley (HBW Alive).
Extirpated Regionally (Renjifo 1999,
Castaño-Villa & Patiño-Zabala 2008).
Habitat is “Moist to wet cloudforest” (HBW
Alive). Near Threateaned (Renjifo et al.
2014)
Requires large dead trees (HBW Alive).
Extirpated in Fragments (Castaño-Villa &
Patiño-Zabala 2008).

Table S2. (continued)
Family Scientific name English name Last records Observations
Grallaridae Grallaricula flavirostris* Ochre-breasted Antpitta 1911 Extirpated Regionally (G.nana &
G.cucullata). (Castaño-Villa &
Patiño-Zabala 2008).Should be considered as
Near Threateaned (Ocampo-peñuela & Pimm
2017)
Formicariidae Formicarius rufipectus Rufous-breasted Antthrush 1911 – seen once in
1989 (eBird 2017)
Furnariidae Glyphorynchus spirurus* Wedge-billed Woodcreeper 5< records
(2006-2016)
Extirpated Regionally (Castaño-Villa &
Patiño-Zabala 2008) Habitat is “Humid to
Wet Primary forest” (HBW Alive)
Extirpated In Fragments (Renjifo 1999)
Furnariidae Campylorhamphus pusillus Brown-billed Scythebill 1911 Extirpated In Fragments (Renjifo 1999)
Furnariidae Margarornis stellatus* Fulvous-dotted Treerunner 1911 Habitat is “Montane evergreen forest heavily
laden with moss and epiphytes” (HBW
Alive). Near Threatened (IUCN 2017)
Furnariidae Synallaxis unirufa* Rufous Spinetail 1911 Habitat is "undergrowth or edge of humid
cloudforest" (HBW Alive)
Tyrannidae Elaenia pallatangae Sierran Elaenia 1959 Recent records in the area of Farallones de
Cali National Natural Park (eBird)
Tyrannidae Legatus leucophaius* Piratic Flycatcher 1959
Tyrannidae Myiarchus apicalis* Apical Flycatcher 5< records
(2006-2016)
Cotingidae Pipreola aureopectus Golden-breasted Fruiteater 1911
Cotingidae Ampelion rufaxilla* Chestnut-crested Cotinga 1911 – one record in
2015. (eBird)
Cotingidae Rupicola peruvianus Andean Cock-of-the-rock 5< records
(2006-2016)
Should be considered as Endangered
(Ocampo-peñuela & Pimm 2017)
Extirpated Regionally (Castaño-Villa &
Patiño-Zabala 2008)
Extirpated In Fragments (Renjifo 1999)
Cotingidae
Pyroderus scutatus*
Red-ruffed Fruitcrow
1911
Extirpated Regionally (Castaño-Villa &
Patiño-Zabala 2008). Vulnerable (Renjifo et
al. 2014)
Turdidae
Turdus obsoletus
Pale-vented Thrush
5< records
Lehman (1957) could not find the species in
(2006-2016).
San Antonio despite targeted searches.
Thraupidae
Tersina viridis
Swallow Tanager
5< records
Nomadic.
(2006-2016)
Cardinalidae
Habia cristata*
Crested Ant-Tanager
1911
Found only near streams. Should be
considered as Vulnerable (Ocampo-peñuela &
Pimm 2017)
Icteridae
Psarocolius angustifrons*
Russet-backed Oropendola
1911
Extirpated Regionally (Renjifo 1999)

Table S3. Birds that colonized the San Antonio region over one hundred years.
Family
Scientific name
English name
Habitat
First known record
Columbidae
Zenaida auriculata
Eared Dove
Non-forest
1985 (eBird)
Cuculidae
Crotophaga ani
Smooth-billed Ani
Non-forest
1959 (Miller 1963)
Trochilidae
Florisuga mellivora
White-necked Jacobin
Forest Generalist
1990 (Kattan et al. 1994)
Trochilidae
Colibri delphinae
Brown Violetear
Forest Generalist
1990 (Kattan et al. 1994)
Trochilidae
Colibri coruscans
Sparkling Violetear
Forest Generalist
2004 (eBird)
Ardeidae
Bubulcus ibis
Cattle Egret
Non-forest
1989 (eBird)
Accipitridae
Spizaetus ornatus
Ornate Hawk-Eagle
Forest Specialist
1985 (Giraldo 1985; eBird)
Accipitridae
Rupornis magnirostris
Roadside Hawk
Non-forest
1959 (Miller 1963)
Psittacidae
Pionus menstruus
Blue-headed Parrot
Forest Generalist
Early 2000s (eBird)
Psittacidae
Pionus chalcopterus
Bronze-winged Parrot
Forest Generalist
2012 (eBird)
Thamnophilidae
Drymophila striaticeps
Streak-headed Antbird
Forest Specialist
2008 (Montealegre-Talero 2009)
Furnariidae
Xiphorhynchus erythropygius
Spotted Woodcreeper
Forest Specialist
2010 (eBird)
Furnariidae
Synallaxis brachyura
Slaty Spinetail
Forest Generalist
1985 (Giraldo 1985)
Tyrannidae
Lophotriccus pileatus
Scale-crested Pygmy-Tyrant
Forest Specialist
Early 2000s (eBird)
Tyrannidae
Todirostrum cinereum
Common Tody-Flycatcher
Non-forest
1959 (Miller 1963)
Tyrannidae
Pitangus sulphuratus
Great Kiskadee
Non-forest
1985 (Giraldo 1985)
Tyrannidae
Myiarchus cephalotes
Pale-edged Flycatcher
Forest Generalist
1985 (eBird)
Corvidae
Cyanocorax yncas
Green Jay
Forest Generalist
Early 2000s (eBird)
Hirundinidae
Stelgidopteryx ruficollis
Southern Rough-winged Swallow
Non-forest
1959 (Miller 1963)
Troglodytidae
Troglodytes solstitialis
Mountain Wren
Forest Specialist
1990 (Kattan et al. 1994)
Turdidae
Turdus fuscater
Great Thrush
Non-forest
1988 (eBird)
Thraupidae
Sporophila nigricollis
Yellow-bellied Seedeater
Non-forest
1959 (Miller 1963)
Thraupidae
Thraupis palmarum
Palm Tanager
Non-forest
1959 (Miller 1963)
Thraupidae
Thraupis cyanocephala
Blue-capped Tanager
Forest Generalist
1985 (Giraldo 1985)
Emberizidae
Chlorospingus semifuscus
Dusky Chlorospingus
Forest Specialist
1989 (eBird)
Emberizidae
Atlapetes latinuchus
Yellow-breasted Brushfinch
Forest Generalist
Early 2000s (eBird)
Parulidae
Basileuterus culicivorus
Golden-crowned Warbler
Forest Specialist
Early 2000s (eBird)
Fringillidae
Spinus spinescens
Andean Siskin
Forest Generalist
1990 (Kattan et al. 1994)
Late non-forest colonisers (early 2000s): Anthracothorax nigricollis ,Amazilia tzacatl, Dryocopus lineatus, Ramphocelus dimidiatus. We note our survey
focused on forest areas so the number of late colonisers from open areas is likely to be bigger.

Figure S1. Comparison of forest cover values between the 1977 and 2016 maps of the southern portion of the San Antonio region, comprising 28% of its area.