Thermal Quality Explains Shift in Habitat Association from Forest to Clearings for Terrestrial-Breeding Frogs Along

John Carroll University Carroll Collected

Masters Theses Master's Theses and Essays

Summer 2019 THERMAL QUALITY EXPLAINS SHIFT IN HABITAT ASSOCIATION FROM FOREST TO CLEARINGS FOR TERRESTRIAL-BREEDING FROGS ALONG AN ELEVATION GRADIENT IN COLOMBIA Zachary Lange John Carroll University, [email protected]

Follow this and additional works at: https://collected.jcu.edu/masterstheses Part of the Biology Commons

Recommended Citation Lange, Zachary, "THERMAL QUALITY EXPLAINS SHIFT IN HABITAT ASSOCIATION FROM FOREST TO CLEARINGS FOR TERRESTRIAL-BREEDING FROGS ALONG AN ELEVATION GRADIENT IN COLOMBIA" (2019). Masters Theses. 42. https://collected.jcu.edu/masterstheses/42

This Thesis is brought to you for free and open access by the Master's Theses and Essays at Carroll Collected. It has been accepted for inclusion in Masters Theses by an authorized administrator of Carroll Collected. For more information, please contact [email protected]. THERMAL QUALITY EXPLAINS SHIFT IN HABITAT ASSOCIATION FROM FOREST TO CLEARINGS FOR TERRESTRIAL-BREEDING FROGS ALONG AN ELEVATION GRADIENT IN COLOMBIA

A Thesis Submitted to the Office of Graduate Studies College of Arts & Sciences of John Carroll University in Partial Fulfillment of the Requirements for the Degree of Master of Science

By Zachary Kurtis Lange 2019

Table of Contents ...... 1

Abstract ...... 3

1. Introduction ...... 4

2. Materials and Methods ...... 7

2.1 Study area and species ...... 7

2.2 Field sites ...... 10

2.3 Environmental variables ...... 12

2.4 Mark-recapture surveys ...... 13

2.5 Detection probability ...... 13

2.6 Thermal assays ...... 14

2.7 Statistical analysis ...... 16

2.8 Analysis of thermal quality ...... 17

3. Results ...... 18

3.1 Environmental data ...... 18

3.2 Mark-recapture ...... 20

3.3 Detection probability ...... 22

3.4 Habitat associations ...... 22

3.5 Thermal preference ...... 25

3.6 Thermal quality ...... 25

3.7 Critical thermal maxima ...... 28

4. Discussion ...... 28

Acknowledgements ...... 34

References ...... 35

Supplemental Materials ...... 48

Figure S1 ...... 48

Figure S2 ...... 49

Figure S3 ...... 50

Figure S4 ...... 51

Figure S5 ...... 52

Figure S6 ...... 53

Figure S7 ...... 54

Table S1 ...... 55

Table S2 ...... 56

Table S3 ...... 57

Thermal quality explains shift in habitat association from forest to clearings for terrestrial-breeding frogs along an elevation gradient in Colombia

Abstract

Tropical ectotherms are considered particularly sensitive to changes in the thermal environment from climate change and habitat alteration. Understanding how such species’ thermal physiology relates to their habitat associations in thermally heterogeneous landscapes may help us predict responses and develop sound conservation strategies for the future. We conducted a mark-recapture study of three terrestrial- breeding anuran species (Pristimantis medemi, P. savagei, P. frater) in adjacent forest and anthropogenic clearings at field sites spread across seven elevations (415-1350 m asl) in the Colombian Andes. We also performed thermal preference and critical thermal maximum assays in the lab to investigate the relationship between Pristimantis thermal physiology and habitat associations at each elevation. Thermal traits did not differ based on the habitat or elevation occupied; yet as elevation increased, two of three species became more abundant in clearings than in forest. We found that some species track their thermal preference to warmer clearings with increasing elevation. In essence, upland anthropogenic clearings are perceived as higher quality thermal habitat by two species in our study. Our findings suggest the importance of thermal heterogeneity and human- disturbed environments for the conservation of thermally sensitive species.

1. Introduction

Organisms worldwide are experiencing novel thermal conditions brought about by anthropogenic forces, namely climate change (Parmesan & Yohe, 2003) and habitat alteration (Tuff et al., 2016). Although the threat of climate change is global in scale, habitat alteration has a more localized impact. However both threats typically lead to increasing environmental temperatures, and there is growing evidence of their synergistic negative effects on biodiversity (Hof et al., 2011; Nowakowski et al., 2016; Nowakowski et al., 2018a; Newbold, 2018). Across geographic regions and taxonomic groups, rising temperatures resulting from climate change have already induced a wide range of responses, including changes in behavior (Sinervo et al., 2010), distribution (Bustamante et al., 2005; Feeley et al., 2011), phenology (Parmesan & Yohe, 2003), and even local extinction (Sinervo et al., 2010). Moreover, warmer temperatures caused by habitat loss and degradation have resulted in changes to species assemblages, especially loss of thermally sensitive species from affected localities (Nowakowski et al., 2017; Frishkoff et al., 2019).

Ectotherms, whose basic physiological functions are largely governed by environmental temperature, are especially vulnerable to rising temperatures from climate change (Deutsch et al., 2008; Newbold, 2018). Because ectotherms in the tropics often live close to their thermal limits (Sunday et al., 2014; Chan et al., 2016), even small changes in temperature can impact their distribution and survival (Colwell et al., 2008; Deutsch et al.,

2008; Tewksbury et al., 2008). Furthermore, high levels of thermal specialization

(Laurance et al., 2011), caused by exposure to relatively invariable temperatures (Janzen

1967), suggests that many tropical ectotherms have a limited capacity to physiologically 4 acclimate or adapt to increasing temperatures associated with climate change (Ghalambor et al., 2006; Shah et al., 2017). The lack of thermal variation in the tropics is one reason tropical mountains harbor high levels of biodiversity and endemism (Janzen, 1967; Graham et al., 2014). Locations close in proximity but differing in elevation may see very little overlap in environmental temperatures, resulting in the range-restricted, thermally constrained species seen in tropical montane areas (Janzen, 1967).

As climate change progresses, tropical mountain ranges, such as the Andes, may become important refugia for lowland species. Temperature lapse rates along tropical elevation gradients (5.2-6.5°C per 1000 m elevation) are much steeper than latitudinal temperature gradients (Colwell et al., 2008), which means range shifts upslope are more likely than poleward shifts for lowland taxa found near and within tropical mountain ranges

(Bush, 2002). In the tropical Andes, terrestrial-breeding anurans (Terrarana: Hedges et al.,

2008) of the genus Pristimantis (Craugastoridae) are a diverse group, with many species occupying small ranges of elevation (see Bernal & Lynch, 2008; Pintanel et al., 2019).

While highly abundant and diverse (Mendoza et al., 2015), many Pristimantis species are of conservation concern (39% of genus members listed as vulnerable or worse: IUCN

2008). Many studies have found that Pristimantis and their close relatives respond negatively to disturbance of forest habitat (Marsh & Pearman, 1997; Toral et al., 2002;

Pineda et al., 2005; Isaacs-Cubides & Urbina-Cardona, 2011; Cole et al., 2014), indicating their close association with intact forest. As terrestrial-breeders, Pristimantis rely on cool, moist microclimates where they deposit clutches of direct-developing eggs (Duellman &

Lehr, 2009; Scheffers et al., 2013), which may explain their tendency to avoid disturbed habitats. Recent studies have also linked low critical thermal maxima (CTmax) to low 5 abundance in small forest fragments and the surrounding matrix (Nowakowski et al., 2017;

Nowakowski et al., 2018b). Given that the tropical Andes is a region where the impacts of both climate change and land-use are expected to be most severe on amphibians (Hof et al., 2011), it is critical that we attempt to understand how species will respond before a large proportion of biodiversity is lost.

While much research focuses on tropical ectotherms that occupy a narrow range of elevations, some species are distributed more broadly, from lowland rainforest to montane forest (see Bernal & Lynch, 2008). Because populations at different elevations experience drastically different thermal regimes (Navas et al., 2013), wide-ranging species may provide evidence of more rapid adaptation or plasticity of thermal limits than is typically assumed in tropical ectotherms (Sunday et al., 2011; Hoffmann et al., 2013; Gunderson &

Stillman, 2015).

However, the thermal conditions at a given elevation are not always uniform, and so populations at different elevations may not experience dramatically disparate temperature regimes. As a result of human activity, there may be a patchwork of warmer microhabitats interspersed among cooler forest habitats throughout an elevation gradient.

In Costa Rica, two frog species were found to follow a surprising pattern, wherein they occupy cooler forest in the lowlands, but shift to warmer deforested pastures upslope

(Frishkoff, et al. 2015). Rather than being uniquely adapted to distinct thermal conditions at different elevations, these species were tracking their thermal niche, as temperatures in high-elevation pastures were more comparable to temperatures in low-elevation forest

(Frishkoff et al., 2015). While thermal niche tracking and habitat switching may be a common response to thermal heterogeneity along elevation gradients, it has gone largely 6 undocumented (but see Ashton et al., 2009; Frishkoff et al., 2019). Studying the thermal biology of wide-ranging species along elevation gradients may provide crucial insight into nuanced responses to climate change (Miles, 1994; Donnelly & Crump, 1998), habitat alteration, and their synergistic effects.

Here we present a study of the thermal biology of an assemblage of terrestrial- breeding frogs of the genus Pristimantis along an elevation gradient in the Colombian

Andes. Our aim is to 1) determine affinity of individual frogs for forest or clearing habitats through mark-recapture methods; 2) test for local adaptation by assaying thermal traits

(thermal preference, critical thermal maximum); and 3) understand how thermal biology influences occupancy of a particular habitat at different elevations for all three species in the study. We also assess how our findings can be applied to understanding species responses to climate change in a thermally heterogeneous landscape of different macroclimates (elevations) and microclimates (habitats).

2. Materials and methods

2.1 Study area and species

Research was conducted at seven field sites along an elevation gradient (415-1350 m asl) on the eastern slope of the Cordillera Oriental in Colombia (Fig. 1). Individual sites were located at 415 m (Balmoral), 550 m (Coburgos), 650 m (Jardin Botanico), 820 m (La

Cumbre), 930 m (La Arabia), 1180 m (Buena Vista), and 1350 m (Manzanares). The region is characterized by a monomodal climate, with a rainy season from April-October, and the

7 rainiest period being from May to June. The average yearly precipitation is 4500 mm, while the mean annual temperature is 25.6°C (IDEAM, 2013).

Fig. 1 Idealized survey plot setup with three separate forest-clearing plot pairings of equal area per elevational site.

We selected field sites where forest abutted anthropogenic clearings, with a sharp ecotone between them. Survey plot areas were centered on the ecotone, with surveys stretching 50 meters into each habitat from the edge. Total survey area was equal per field site (6,000 m2) and was arranged as either three small (20 m x 100 m) survey plots (n = 3 sites), a small and a medium (40 m x 100 m) survey plot (n = 1 site), or a single large (60 m x 100 m) survey plot (n = 3 sites). We sampled at each site on 3-6 occasions between

June and July, 2018.

The focal species of this study were Pristimantis medemi (Lynch, 1994), P. savagei

(Pyburn & Lynch, 1981), and P. frater (Werner 1899), family Craugastoridae (Pyron &

Wiens, 2011). As terrestrial-breeding frogs (Terrarana), Pristimantis deposit their egg clutches in leaf litter, where they undergo direct development to hatch into tiny froglets

(Hedges et al., 2008). Their reproductive strategy, along with low thermal tolerances, means many Pristimantis are restricted to forest habitat throughout their lives. Adults spend most of their time on the forest floor or perched on low vegetation (Cáceres-Andrade &

Urbina-Cardona, 2009; Hedges et al., 2008). Of the three species, adult P. frater is the smallest, followed by P. savagei, and P. medemi (Pyburn & Lynch, 1981; Lynch, 1994).

P. frater are known to occur between 900-3000 m asl, P. savagei from 600-3000 m asl, and P. medemi from 450-2400 m asl (Bernal & Lynch, 2008).

2.2 Field sites

Balmoral (BA) was our lowest-elevation field site at 415 m asl (4.099193°N,

73.615577°W). The clearing habitat was a combination of short mowed grass and gravel pathways, while the forest was part of a municipal reserve, Reserva Forestal Balmoral. We 10 established three equal-sized survey plots split between two separate forest patches, both part of the reserve.

Coburgos (CO) was private property that included a large grassland for grazing cattle and contiguous upland forest at 550 m asl (4.061044°N, 73.737852°W). We established three equal-sized survey plots that were separated from one another by 25 meters.

Jardin Botanico (JB) was a botanical garden that does not allow public access and is located in the Andean foothills at 650 m asl, on the outskirts of Villavicencio

(4.152099°N, 73.654320°W). The clearing habitat was a deforested area with a low density of remnant trees, and was adjacent to various human-made structures. The forest was one contiguous block. We established a single large survey plot because of spatial constraints.

La Cumbre (LC) was a small, privately owned organic farm at 820 m asl

(4.080955°N, 73.745673°W). The clearing habitat was a combination of grassland that is occasionally mowed and used by grazing cattle, as well as part of a home garden. The forest habitat had two sections: one was on an east-facing downward slope and the other was on flat ground alongside a small stream. We established one small survey plot in the flat streamside forest, which led to the home garden. We created a larger survey plot to encapsulate the aforementioned grassland and downslope forest habitat.

La Arabia (LA) was a privately owned coffee farm at 930 m asl (4.172096°N,

73.665159°W). Clearing habitat was composed of coffee and other crops, some of which had been cleared or recently planted. Forest was all contiguous. We established small survey plots in three separate locations.

Buena Vista (BV) was a privately owned farm at 1180 m asl (4.176101°N,

73.671177°W). Clearing habitat included some remnant trees and a vegetable garden. The forest was one contiguous area. We created a single large survey plot because of spatial constraints, as at JB.

Manzanares (MZ) was a private property in a remote area outside Villavicencio at

1350 m (4.142619°N, 73.813947°W). Clearing habitat was cleared forest in an early successional stage, while the forested area was contiguous. We established a single large survey plot because of spatial constraints.

2.3 Environmental variables

At each site a Kestrel DROP or iButton data logger (Maxim hygrochron DS1923) was deployed 50 m from the ecotone in both forest and clearing habitat to record temperature once per hour. Data loggers were placed out of direct sunlight at the base of vegetation or under leaf litter to mimic operative temperatures of amphibians in the field.

To assess structural characteristics of forest and clearing habitat we measured the following at each site: leaf litter depth (mm), leaf quantity (# of leaf layers above soil), canopy cover

(%), and stem density. Stem density accounted for the number of vertical stems (< 10 cm 12 diameter) within a 1-meter radius (Fig. S7). Habitat data sampling occurred in three separate locations per habitat, each along the back edge of a survey plot, 50 m from the forest-clearing edge. Due to logistical constraints we did not record habitat data from our

1180-m site, Buena Vista. All other sites were sampled between 26 July and 4 August,

2018.

2.4 Mark-recapture surveys

Timed visual encounter surveys (VES) were conducted at a single field site each night, beginning at sunset and ending after 135 total minutes (67.5 minutes per habitat) using 3-5 observers. At sites with multiple survey plots, order of surveys was randomized per visit, and at all sites the starting habitat was randomized as well (forest vs. clearing).

The highest field site, Manzanares, was surveyed three consecutive nights (24-26 July,

2018) and all other sites were surveyed between five and seven times from 2 June to 22

July, 2018.

GPS coordinates were taken for all encountered Pristimantis and the following were recorded per individual: habitat (forest or clearing), resting height (cm), substrate type

(leaf litter, stem, leaf, trunk, etc.), body temperature (°C), SVL (mm), mass (g), and sex or life stage when possible. Pristimantis were also marked with a unique toe-clip pattern

(unless SVL < 10 mm) to allow for identification of recaptures in subsequent encounters.

2.5 Detection probability

To account for possible differences in detection between forest and clearings for P. medemi, we ran N-mixture models (Royle, 2004; Mazzotti et al., 2019) for the count data 13 from each habitat in program PRESENCE (Hines, 2006). This type of model simultaneously estimates detection probability (p) per replicate survey and estimated relative density (λ) per site. Manzanares, our highest-elevation site (1350 m), was not included in this analysis because it is the only site in which mark-recapture was not conducted. We ran multiple models that incorporated all of the following covariates: elevation, mean environmental temperature, mean daily maximum temperature, canopy cover (%), leaf litter depth (mm), stem density (m-2), and number of observers (Table S1).

AIC was used to rank candidate models. N-mixture models were not used for P. savagei or P. frater due to low sample sizes.

2.6 Thermal assays

Pristimantis used in thermal assays were collected after mark-recapture surveys from forest and clearing habitat at least 50 meters from survey plots, thus avoiding potential interference with the mark-recapture study. All animals used in thermal assays were housed in an ambient lab at approximately 650 m asl, near the midpoint of our elevation gradient.

Each frog was acclimated to lab conditions for 1-5 days before testing, and was released back to its original habitat the following week. We conducted two types of thermal assays: critical thermal maximum (CTmax) and thermal preference (Tpref). An organism’s CTmax is one way to delineate what is too hot to allow long-term persistence in a given habitat.

However, ectotherms tend to avoid temperatures near their CTmax, instead selecting a range of temperatures well below that threshold, and typically below their thermal optimum (Topt) as well (Martin & Huey, 2008). Determining an organism’s Tpref provides information

14 about the suite of temperatures that will be chosen for survival and persistence in realistic or natural settings (Hutchinson & Maness, 1979; Hertz et al., 1993).

To create a thermal gradient for Tpref trials (approximately 12 °C to 40 °C), two strips of heat tape were placed under the ends of five 1.5-meter metal gutters, and ice packs were placed under the opposite end of each gutter, following the setup of Sauer et al.

(2016). Moist towels lined the inside of each gradient to maintain constant humidity

(Galindo et al., 2018), and semi-transparent plastic panels sat atop gradients to minimize escape during Tpref trials. At the start of each trial a single frog was placed in the center of a gradient, and body temperature readings were recorded (via non-contact laser thermometer) every 15 minutes over the course of a four-hour testing period. Tpref was then calculated as the mean of the central 50% of body temperature values recorded per individual, based on the methodology of previous studies (Hertz et al., 1993; Thompson et al., 2018). No animals were tested more than once, and Tpref trials were always performed before other thermal assays.

In our critical thermal maximum (CTmax) assays we used two different methodologies; however, we did not detect a significant difference in results between them

(Fig. S1). The first methodology used loss of righting response as the proxy for an individual’s CTmax. For the trial each frog was placed in a small Petri dish containing 10 mL of water with the lid secured to prevent escape. We operated a HOBO data logging unit with four thermocouples, each adhered to the inside of a Petri dish, which allowed us to measure the temperatures (per second) experienced by each individual frog throughout the trial. Up to four Petri dishes were placed in a shallow 300 mL water bath. The temperature of the water bath, and thus the Petri dishes, was raised approximately 0.3 °C min-1 by 15 incrementally adding boiling water to the water bath. When an increase in activity was observed in a frog, the righting response was tested by manually inverting it dorsoventrally.

If a righting response was exhibited, we placed the frog back in its Petri dish and continued the trial. At the point in which a righting response was no longer observed, we quickly recorded the CTmax of the animal by contacting the dorsum with a thermocouple. The final temperature of the Petri dish was also recorded.

The second CTmax methodology measured jumping performance, rather than righting response, as a proxy for CTmax. We used the same setup to raise frog temperatures over the course of jumping performance trials, including the same rate of temperature increase (0.3 °C min-1). However, rather than testing the righting response late in the trial, the frog was removed from its Petri dish every three minutes, placed on the floor, and gently prodded from behind to induce a jumping response. After two consecutive jumps the frog was quickly added back into the thermal bath to continue the trial. The longer of the two jumps was measured and recorded as the performance at a given temperature. The temperature at which a jumping response could not be induced was considered the CTmax.

2.7 Statistical analysis

To test for species-specific effects of extrinsic factors on thermal preference values, we ran two-factor ANOVAs with Tpref as our response variable and number of days prior to assay (acclimation period) and start time of Tpref assay (morning or evening) as fixed factors. ANOVAs were then performed to test for effects of habitat and elevation on Tpref.

ANOVAs were also conducted to test for effects of habitat and elevation on CTmax. All

16 data met assumptions of normality and equal variance. All analyses were conducted using package ‘lme4’ in RStudio 1.1.423.

2.8 Analysis of thermal quality

We compared the interquartile range (central 50%) of Tpref to operative temperatures collected by our dataloggers to describe the thermal quality of different habitats (forest and clearing) at different elevations. We calculated thermal quality (Tqual) by the following equation, where Tenv is the mean daily maximum temperature collected from our dataloggers.

푇푞푢푎푙 = 푇푝푟푒푓 − 푇푒푛푣

Each species was analyzed separately for thermal quality. To better capture temperature extremes that could affect ectotherm occupancy, mean daily high temperatures

(Tenv) collected from data loggers were incorporated into our thermal quality index

(Camacho et al., 2015; see Frishkoff et al., 2015). Since individual Tpref did not vary by habitat (p = 0.121) or elevation (p = 0.799), we considered the thermal preference of each species as the mean of all individual thermal preferences previously calculated.

3 Results

3.1 Environmental data

At each elevation, clearings were generally characterized as warmer and more thermally variable than forest habitats (Fig. 2, Fig. S2). High temperature anomalies were also greatest in the clearings at each elevation.

45 45

Clearing Forest

40 40

(°C)

35 35

temp. 30 30

25 25 20 20 Daily high Daily 15 15 6/3 6/17 7/1 7/15 7/29 6/3 6/17 7/1 7/15 7/29 Date Fig. 2 Daily high operative temperatures (°C) measured by dataloggers in clearings and forests at six of seven elevations, with the exception of MZ (1350 m). Each colored line represents a unique elevation/sites from lower (lighter) to higher (darker) elevations. Gaps in data indicate periods of sensor failure.

3.2 Mark-recapture

Among 276 Pristimantis marked with a unique toe-clip pattern, 64 individuals were recaptured at least once (48 P. medemi, 10 P. savagei, 6 P. frater) and 22 were recaptured two or more times (21 P. medemi, 1 P. savagei). Of the 64 Pristimantis that were recaptured, 58 (90.625%) were repeatedly found in the same habitat type, forest or clearing.

Among those frogs that did not switch habitats, the mean movement distance was 5.59

(±4.045) meters, whereas the six individuals that did cross from one habitat to another traveled a mean of 25.12 (±11.889) meters per movement (Fig. 3).

Clearing Forest Fig. 3 Pristimantis recaptures at Buena Vista (1180 m), showing rarity of ecotone crossing. Colored dots represent capture 60 m points, with each color representing a unique individual. Lines connecting dots show consecutive movements of each frog. 5 0 m

3.3 Detection probability

Our top model for each habitat showed that detection probability varied as a function of stem density and number of observers, and relative density varied as a function of leaf litter depth, elevation, and mean environmental temperature (Table S2, S3). The next best model for each set included canopy cover as a covariate (Open: ΔAIC = 2.53,

Forest: ΔAIC = 1.99).

3.4 Habitat associations

Using λ values from our top performing forest and clearing N-mixture models, we determined proportion of P. medemi in clearings at each elevation. We then compared the model-based proportions to the proportions from our raw count data. There was no correlation between our raw proportions and the modeled proportions (r = 0.6696, p =

0.1457, R2 = 0.4483). However, at our lowest-elevation site (Balmoral), the model estimated a higher proportional density P. medemi in the clearing, albeit with λ<1 in each habitat. P. medemi were not observed in the clearings at Balmoral, which suggests our model density estimate was erroneous. Thus, we excluded Balmoral in a subsequent comparison, wherein we found a strong correlation between the proportion in clearings based on λ, and the proportion in clearings based on count data (r = 0.9770, p = 0.0042, R2

= 0.9546). Regarding detection probability (p) alone, p was significantly higher in clearings than in forest at Balmoral. However, at every other site the detection probability was either significantly lower in clearings, or not significantly different from detection in forest. This suggests that our raw count data could be a conservative estimate of actual abundance in

22 clearings. Therefore, we report our raw count data proportions to assess habitat associations

(Fig. 4).

For P. medemi, the only species found at all seven sites, the proportion of individuals present in clearings increased with elevation (R2 = 0.6303). Both juveniles (R2

= 0.5411) and adults (R2 = 0.7741) followed this pattern (Fig. 4). A greater proportion of

P. frater adults were found in clearings at higher elevations (R2 = 0.6579), however no juveniles of this species were observed at any site. Neither adults nor juveniles of P. savagei exhibited an increase in proportional abundance in clearings with elevation (Fig.

4).

(a) (d) Forest Clearing 1 0.8

0.6 R² = 0.5411

0.4 R² = 0.7741

0.2

(b) 1 300 500 700 900 1100 1300 1500 (e)

0.8

0.6

0.4 R² = 0.6579

Elevation (m) Elevation

clearings(%) Proportion in 0.2

0 (c) 1 300 500 700 900 1100 1300 1500 (f)

0.8

0.6 R² = 0.3638 0.4

0.2 R² = 0.0007 0 300 500 700 900 1100 1300 1500 1 0.5 0 0.5 1 Elevation (m) Proportion observed by habitat Fig. 4 Relationship between habitat associations and elevation in Pristimantis medemi (a, d), P. frater (b, e), and P. savagei (c, f). (a-c) Proportion of adults (filled squares) and juveniles (empty squares) found in clearings for each species. (d-f) Proportion of all life stages observed in forests and clearings along elevation gradient. All proportions represent total number of unique individuals observed, rather than total number of observations, which would include duplicates (recaptures).

3.5 Thermal preference

Time of Tpref assay had a significant effect on Tpref for both P. medemi (F1,83 =

7.0543, p = 0.0095) and P. savagei (F1,23 = 17.1924, p = 0.0004), but no effect on P. frater

Tpref. Trials conducted in the morning yielded significantly higher Tpref values than those conducted in the evening in the case of P. medemi (T86 = 2.6354, p = 0.01) and P. savagei

(T26 = 4.1583, p = 0.0003). In contrast, acclimation period had no significant effect on the

Tpref of any species.

To address the effect of start time on Tpref in P. medemi and P. savagei, we calculated the mean differences between morning and evening Tpref at each elevation.

Considering all three species are primarily nocturnal, we adjusted the Tpref values for each individual frog tested in the morning by subtracting the corresponding morning-evening difference from their Tpref, thus approximating each frog’s nocturnal Tpref. We re-ran the two-factor ANOVA after adjusting Tpref and no longer observed a significant effect of start time, nor a new effect of acclimation period on Tpref for either P. medemi or P. savagei.

Using the adjusted Tpref as our response variable, we then ran two-factor ANOVAs to test for an effect of elevation or habitat type on Tpref. Habitat and elevation did not have main effects nor an interactive effect on the Tpref for any of our three study species.

3.6 Thermal quality

Because there was no significant effect of elevation or habitat type on Tpref, we used the mean Tpref of all P. medemi (24.47 °C), P. savagei (26.31 °C), and P. frater (21.42 °C) to conduct a species-by-species analysis of thermal quality (Fig. 5). For each species, we

25 found significant non-interacting effects of habitat type (F1,6 = 12.0659, p = 0.013) and elevation (F6,6 = 5.1131, p = 0.034) on thermal quality (Fig. 5).

P. medemi

P. frater

P. savagei

Fig. 5 Pristimantis thermal quality (°C) in forest (filled triangles) and clearing (empty circles) per elevation. Symbol sizes correspond to the proportion of individuals detected in each habitat at a given elevation. Background colors represent temperatures that fall within the interquartile range of Tpref measurements (green), or are cooler (blue) or warmer (red) than the interquartile range. Gray dashed line is the mean Tpref. 27

To see if each species’ relative abundance in clearings coincided with thermal quality, we performed a linear regression with habitat and elevation as predictors, and since we found an increase in proportion of P. medemi in clearings with elevation, we tested for its relationship with thermal quality. There was a significant effect of thermal quality of clearings on the proportion of P. medemi occupying them (F1,5 = 29.977, p = 0.003; Fig.

S3). For both P. frater (F1,5 = 0.8226, p = 0.406; Fig. S4) and P. savagei (F1,5 = 1.4675, p = 0.2799; Fig. S5) the relationship was not significant between thermal quality of clearings and proportion of frogs occupying them.

3.7 Critical thermal maxima

There was no significant effect of elevation or habitat on CTmax in P. medemi

(elevation: F5,31 = 1.5555, p = 0.2018; habitat: F1,31 = 0.4246, p = 0.5294), P. savagei

(elevation: N/A; habitat: F1,10 = 0.0693, p = 0.7977), or P. frater (elevation: F1,2 = 0.4355, p = 0.5771; habitat: F1,2 = 0.2065, p = 0.6941). Additionally, we found no significant main effect or interactive effect when sex was incorporated into our models. When comparing mean CTmax among species, we found that P. medemi had a significantly higher CTmax than both other species (P. frater: p = 0.002; P. savagei: p = 0.0002, whereas the CTmax of P. frater and P. savagei did not differ significantly (p = 0.3123).

4. Discussion

Recent studies have examined how habitat conversion and thermal biology determine species distributions. However, few studies have considered how this 28 relationship may vary across broad environmental gradients such as latitude or elevation in the tropics. Here we show that within an assemblage of tropical terrestrial-breeding frogs, thermal niche tracking can lead some species to switch habitat associations along an elevation gradient.

The climate variability hypothesis (CVH, Janzen, 1967) states that ectothermic species in thermally stable climates (montane tropical forests, Navas et al., 2013) have inherently lower acclimation capacity compared to those found in highly variable climates

(temperate forests). We found that among our three species, Pristimantis medemi, P. savagei, and P. frater, thermal traits do not vary with elevation or habitat type (forest or clearing). Even in vastly different macroclimates (elevations) and microclimates (habitats), the species in our study do not exhibit population-level acclimation or adaptation of thermal preferences or tolerances, similar to results for other ectotherms (Labra, 1998; Gvozdik &

Castilla, 2001; Yang et al., 2008; Enriquez-Urzelai et al., 2018). The conserved nature of their thermal traits suggests a limited capacity for acclimation to new conditions under climate change, as has been commonly cited in tropical ectotherms (Ghalambor, et al.,

2006; Sunday et al., 2011; Hoffmann et al., 2013; Gunderson & Stillman, 2015; Shah et al., 2017).

Typically, we think of species with constrained physiological acclimation capacity as sensitive to habitat alteration (Nowakowski et al., 2017). Pristimantis medemi is generally considered a forest-dwelling species (Lynch, 2006; Cáceres-Andrade & Urbina-

Cardona, 2009), as is the case for most members of Family Craugastoridae (Nowakowski et al., 2018b). However, we found that at higher elevations where clearings provide temperatures well within their preferred range, P. medemi are particularly abundant. In fact, 29 with increasing elevation, thermal quality of forest becomes colder than the preferred temperature range of this species, whereas open habitats transition from being too warm to being a closer match to their thermal niche. In response, P. medemi shift their habitat association along the elevation gradient. At lower elevations the species occurs exclusively in the forest, but with increasing elevation a greater proportion of all life stages occupy clearings. Essentially, P. medemi is able to track its thermal niche into disturbed habitats at higher elevations (Frishkoff et al., 2015). Although we lack a robust sample size of P. frater, they too appear to switch habitat affiliation along the elevation gradient, including breeding pairs in amplexus observed in open habitat at both 1180 m and 1350 m (Fig. S6).

Of note, P. savagei, does not appear to follow the same pattern of habitat switching, despite preferring a warmer range of temperatures than P. medemi or P. frater. Instead, P. savagei remains largely restricted to forest at the higher elevations where it occurs. Future surveys at higher elevations may show that this species switches habitats as well, just starting from a higher base elevation, as was seen in two closely related terrestrial-breeding anurans of

Family Craugastoridae (Frishkoff et al., 2015).

Regardless of its habitat associations at higher localities, P. savagei seems to be selecting habitat based on a metric other than thermal quality. Assuming arthropod prey assemblages are different in forests and clearings, the broader dietary niche of P. medemi

(Astwood-Romero et al., 2016) might allow them to take advantage of certain prey items found in higher thermal quality clearings, while a narrower range of preferred prey items may restrict P. savagei to forest habitat. But this seems unlikely, as Pristimantis tend to be characterized broadly as generalists (Arroyo et al., 2008). Instead, Pristimantis savagei may be an example of a species that is more sensitive to deforestation (independent of 30 temperature) than its close relative P. medemi (see also Toral et al., 2002; Garcia et al.,

2017), as P. savagei juveniles were never observed in clearings and a very small proportion of adults were found in clearings. Moreover, adult P. savagei observed in clearings tended to be in close proximity to the forest edge, as opposed to P. medemi and P. frater, which were often seen much farther into the clearing side of the boundary. It is also possible that the species differ in their thermoregulatory efficiency when presented with changes to the thermal environment, as has been documented in phylogenetically close Liolaemus lizards in Argentina (Stellatelli et al., 2013). However, one would expect these differences to have been evident during thermal preference assays in the lab.

In the lowlands, thermal quality and narrow thermal safety margins suggest that all three species in our study are at risk of extirpation from climate change. The lowland forest provides thermally suitable habitat for now, but as temperatures rise, individuals will have nowhere to retreat. Elevational range shifts have been suggested as a solution for species living adjacent to montane forest zones, yet there are many potential barriers to successful establishment at new elevations (Forero-Medina et al., 2011), as well as uncertainty about the speed with which species can move upslope (Buckley et al., 2013). At higher elevations, rising temperatures will also impact where species are able to persist. In the locations where P. medemi occupy disturbed habitats, the adjacent thermally buffered forest will provide them at least temporary reprieve from the deleterious effects of climate change. Of course, there likely exist barriers to the re-colonization of the forest by P. medemi, just as there may be non-thermal barriers preventing P. savagei and other forest specialists from occupying clearings in the present day.

We also acknowledge that our study merely reports a snapshot of anuran activity and habitat association. Previously, P. medemi has been observed more abundantly in clearings during the rainy season, after being seen in greater abundance in the forest during the dry season (Acosta-Galvis, 2017). This seasonal change coincides with peak breeding activity, which suggests that clearings may be important for reproduction. Because of the timing of most studies, we do not have a clear understanding of the prevalence of seasonal shifts in habitat occupancy for many taxa, especially in the tropics.

Based on the definition of edge habitat, it can be argued that the forest we surveyed in was all strictly edge. For amphibians, edge effects are known to infiltrate the forest approximately 250 m (Schneider-Maunoury et al., 2016), yet our results suggest that the severity of perceived edge effects vary widely, depending on species and environmental context (Toral et al., 2002; Isaacs-Cubides & Urbina-Cardona, 2011).

Along a steep environmental gradient, such as elevation, the distance that edge effects penetrate the forest seems to shift. In cooler, higher-elevation macroclimates, the boundary between forest and clearing habitats may be far less defined than it is in the lowlands, as thermal regimes tend to be more similar (Fig. 1). At Balmoral, our lowest field site of 415 m asl, we did not find any Pristimantis during our mark-recapture surveys in the forest, yet during incidental surveys we were able to find P. medemi much deeper in the forest interior near a stream. Previous work has suggested that forest- affiliated amphibians tend to occur near streams during the dry season, likely to take advantage of higher humidity and thermal buffering (Urbina-Cardona et al., 2006;

Riemann et al., 2015; Paoletti et al., 2018). In the case of P. medemi, there is evidence that they occur in close proximity to streams at lower elevations (Cáceres-Andrade & 32

Urbina-Cardona, 2009), as is the case with some other amphibians (see Gade &

Peterman, 2019). At our higher-elevation sites where there is a less pronounced contrast in forest and clearing temperatures, stream proximity may be a less important predictor of

P. medemi occupancy than it appears to be in the lowlands.

The pattern of increasing P. medemi abundance and comparatively better thermal quality for all Pristimantis is partially maintained by quality of the matrix (clearing) habitat itself. The presence of canopy and understory vegetation are critical to thermally buffering microhabitats throughout the landscape on hot and dry days. Adequate microhabitats can effectively reduce the frequency and intensity of exposure to extreme temperatures

(Scheffers et al., 2014). Even though Pristimantis are mostly inactive at mid-day, they experience daytime temperatures from the safety of their selected microhabitats. As such it is critical that those microhabitats adequately shield them from extreme high temperatures. Because our temperature dataloggers were placed within leaf litter where one would expect to find Pristimantis, we believe we have captured some portion of the temperatures to which Pristimantis are exposed during daylight hours. In the case of P. frater that were found in the clearing at LC (820 m asl), our dataloggers suggest that the clearing at that site is often inhospitable. But this observation might be an error in sensor placement because a small group of P. frater were regularly seen in a patch of shrubs in the clearing. This discrepancy further highlights the importance of remnant vegetation for persistence of wildlife in human-altered landscapes (Mendenhall et al., 2014). Even isolated trees can provide substantive refuge to species in cattle pastures (Robinson et al.,

2013). We want to highlight the critical importance of remnant trees and shrubs in human- dominated matrix landscapes, as they provide species with a means of cover, 33 thermoregulation, dispersal pathways, and possible reproduction in the case of terrestrial- breeding anurans like Pristimantis.

Future research should investigate novel ecological interactions, long-term population demographics, and matrix quality to better understand what is driving the pattern of differential habitat use along tropical elevation gradients. As a natural laboratory of variation and adaptation, these systems may help us understand how species distributions are limited by temperature and further uncover how environments shape the thermal biology of their inhabitants. The potential for identifying tradeoffs in plasticity and adaption is also crucial to understanding how single species, and the community as a whole, will respond to changing thermal landscapes under pressure from climate change and habitat alteration.

Acknowledgements

Many thanks to Jhon J. Amaya-Bustos, Harold A. Céspedes-Rodríguez, Heidy

Villanueva, M. Felipe Parra-Torres, Alvaro J. Velásquez-Suárez, Andrew Benos, Osmary

A. Medina-Báez, and many others for their dedicated assistance in the field. This research was funded by the American Society of Ichthyology and Herpetology Gaige Fund

Award, the Chicago Herpetological Society Graduate Research Grant, and John Carroll

University.

References

Acosta-Galvis, A. R. (2017). Batracofauna de los bosques de niebla y estribaciones del

piedemonte en el municipio de Yopal (Casanare), Orinoquia colombiana. Biota

Colombiana, 18, 282–315.

Arroyo, S. B., Serrano-Cardozo, V. H., & Ramírez-Pinilla, M. P. (2008). Diet,

microhabitat and time of activity in a Pristimantis (Anura, Strabomantidae)

assemblage. Phyllomedusa, 7, 109–119.

Ashton, S., Gutiérrez, D., & Wilson, R. J. (2009). Effects of temperature and elevation on

habitat use by a rare mountain butterfly: Implications for species responses to climate

change. Ecological Entomology, 34, 437–446.

Astwood-Romero, J. A., Alvarez-Perdomo, N., Parra-Torres, M. F., Rojas-Pena, J. I., &

Nieto-Vera, M. T. (2016). Stomach contents in anurans species from Natural

Reserves in the Villavicencio municipality, Meta, Colombia. Caldasia, 38, 165–181.

Bernal, J. M. H., & Lynch, J. D. (2008). Review and analysis of altitudinal distribution of

the Andean anurans in Colombia. Zootaxa, 1826, 1–25.

Buckley, L. B., Tewksbury, J. J., & Deutsch, C. A. (2013). Can terrestrial ectotherms

escape the heat of climate change by moving? Proceedings of the Royal Society B:

Biological Sciences, 280, 20131149.

Burnham, R. J. (2001). Is Conservation Biology a Paleontological Pursuit? Palaios, 16,

423–424.

Bush, M. B. (2002). Distributional Change and Conservation on the Andean Flank: A

Palaeoecological Perspective. Global Ecology and Biogeography, 11, 463–473.

Bustamante, R., Ron, S. R., & Coloma, L. A. (2005). Cambios en la Diversidad en Siete

Comunidades de Anuros en los Andes de Ecuador. Biotropica, 37, 180–189.

Cáceres-Andrade, S. P., & Urbina-Cardona, J. N. (2009). Ensamblajes de anuros de

sistemas protuctivos y bosques en el piedemonte llanero, departamento del Meta,

Colombia. Caldasia, 31, 175–194.

Camacho, A., Trefaut Rodrigues, M., & Navas, C. A. (2015). Extreme operative

temperatures are better descriptors of the thermal environment than mean

temperatures. Journal of Thermal Biology, 49–50, 106–111.

Chan, W.-P., Chen, I.-C., Colwell, R. K., Liu, W.-C., Huang, C.-Y., & Shen, S.-F.

(2016). Seasonal and daily climate variation have opposite effects on species

elevational range size. Science, 351, 1437–1439.

Cole, E. M., Bustamante, M. R., Almeida-Reinoso, D., & Funk, W. C. (2014). Spatial

and temporal variation in population dynamics of Andean frogs: Effects of forest

disturbance and evidence for declines. Global Ecology and Conservation, 1, 60–70.

Colwell, R. K., Brehm, G., Cardelus, C. L., Gilman, A. C., & Longino, J. T. (2008).

Global Warming, Elevational Range Shifts, and Lowland Biotic Attrition in the Wet

Tropics. Science, 322, 258–261.

Deutsch, C. A., Tewksbury, J. J., Huey, R. B., Sheldon, K. S., Ghalambor, C. K., Haak,

D. C., & Martin, P. R. (2008). Impacts of climate warming on terrestrial ectotherms

across latitude. Proceedings of the National Academy of Sciences, 105, 6668–6672.

Donnelly, M. A., & Crump, M. L. (1998). Potential effects of climate change on two

neotropical amphibian assesmblages. Climatic Change, 39, 541–561.

Duellman, W. E., & Lehr, E. (2009). Terrestrial-breeding frogs (Strabomantidae) in

Peru. Germany: Natur und Tier Verlag.

Enriquez-Urzelai, U., Palacio, A. S., Merino, N. M., & Sacco, M. (2018). Hindered and

constrained: limited potential for thermal adaptation in post‐metamorphic and adult

Rana temporaria along elevational gradients. Journal of Evolutionary Biology, 31,

1852–1862.

Feeley, K. J., Silman, M. R., Bush, M. B., Farfan, W., Cabrera, K. G., Malhi, Y., Meir,

P., Revilla, N. S., Quisiyupanqui, M. N. R., & Saatchi, S. (2011). Upslope migration

of Andean trees. Journal of Biogeography, 38, 783–791.

Forero-Medina, G., Joppa, L., & Pimm, S. L. (2011). Constraints to species’ elevational

range shifts as climate changes. Conservation Biology, 25, 163–171.

Frishkoff, L. O., Hadly, E. A., & Daily, G. C. (2015). Thermal niche predicts tolerance to

habitat conversion in tropical amphibians and reptiles. Global Change Biology, 21,

3901–3916.

Frishkoff, L. O., Gabot, E., Sandler, G., Marte, C., & Mahler, D. L. (2019). Elevation

shapes the reassembly of Anthropocene lizard communities. Nature Ecology &

Evolution, 3, 638–646.

Gade, M. R., & Peterman, W. E. (2019). Multiple environmental gradients influence the

distribution and abundance of a key forest-health indicator species in the Southern

Appalachian Mountains, USA. Landscape Ecology, 34, 569–582.

Galindo, C. A., Cruz, E. X., & Bernal, M. H. (2018). Evaluation of the combined

temperature and relative humidity preferences of the Colombian terrestrial

salamander Bolitoglossa ramosi (Amphibia: Plethodontidae). Canadian Journal of

Zoology, 96, 1230–1235.

Garcia, V. O. S., Ivy, C., & Fu, J. (2017). Syntopic frogs reveal different patterns of

interaction with the landscape: A comparative landscape genetic study of Pelophylax

nigromaculatus and Fejervarya limnocharis from central China. Ecology and

Evolution, 7, 9294–9306.

Ghalambor, C. K. (2006). Are mountain passes higher in the tropics? Janzen’s hypothesis

revisited. Integrative and Comparative Biology, 46, 5–17.

Graham, C. H., Carnaval, A. C., Cadena, C. D., Zamudio, K. R., Roberts, T. E., Parra, J.

L., McCain, C. M., Bowie, R. C. K., Moritz, C., Baines, S. B., Schneider, C. J.,

Vanderwal, J., Rahbek, C., Kozak, K. H., & Sanders, N. J. (2014). The origin and

maintenance of montane diversity: integrating evolutionary and ecological processes.

Ecography, 37, 711–719.

Gunderson, A. R., & Stillman, J. H. (2015). Plasticity in thermal tolerance has limited

potential to buffer ectotherms from global warming. Proceedings of the Royal Society

B: Biological Sciences, 282, 20150401.

Gvoždík, L., & Castilla, A. M. (2001). A Comparative Study of Preferred Body

Temperatures and Critical Thermal Tolerance Limits among Populations of Zootoca

vivipara (Squamata: Lacertidae) along an Altitudinal Gradient. Journal of

Herpetology, 35, 486–492.

Hedges, S. B., Duellman, W. E., & Heinicke, M. P. (2008). New World direct-developing

frogs (Anura: Terrarana). Zootaxa 1737. Auckland, New Zealand: Magnolia Press.

Hertz, P. E., Huey, R. B., & Stevenson, R. D. (1993). Evaluating Temperature Regulation

by Field-Active Ectotherms: The Fallacy of the Inappropriate Question. The

American Naturalist, 142, 796–818.

Hof, C., Araújo, M. B., Jetz, W., & Rahbek, C. (2011). Additive threats from pathogens,

climate and land-use change for global amphibian diversity. Nature, 480, 516–519.

Hoffmann, A. A., Chown, S. L., & Clusella-Trullas, S. (2013). Upper thermal limits in

terrestrial ectotherms: How constrained are they? Functional Ecology, 27, 934–949.

Hutchison, V. H., & Maness, J. D. (1979). The Role of Behavior in Temperature

Acclimation and Tolerance in Ectotherms. American Zoologist, 19, 367–384.

IDEAM. Instituto de Hidrología, Meteorología y Estudios Ambientales. (2013). Datos

climáticos históricos de Villavicencio-Estación Vanguardia. Sistema de Información:

Nacional Ambiental.

http://www.tutiempo.net/clima/Villavicencio_Vanguardia/802340.htm

Isaacs Cubides, P. J., & Urbina Cardona, J. N. (2011). Anthropogenic disturbance and

edge effects on anuran assemblages inhabiting cloud forest fragments in Colombia.

Natureza a Conservação, 9, 39–46.

IUCN. (2008). Amphibians on the IUCN Red List. http://www.iucnredlist.org/initiatives

/amphibians/analysis/red-list-status. International Union for Conservation of Nature

and Natural Resources.

Janzen, D. H. (1967). Why Mountain Passes are Higher in the Tropics. The American

Naturalist, 101, 233–249.

Laurance, W. F., Carolina Useche, D., Shoo, L. P., Herzog, S. K., Kessler, M., Escobar,

F., et al., & Thomas, C. D. (2011). Global warming, elevational ranges and the

vulnerability of tropical biota. Biological Conservation, 144, 548–557.

Lynch, J. D. (1994). Two New Species of the Eleutherodactylus conspicillatus Group

(Amphibia: Leptodactylidae) from the Cordillera Oriental of Colombia. Rev. Acad.

Colomb. Cienc., 19, 187–193.

Lynch, J. D. (2006). The Amphibian Fauna in the Villavicencio Region of Eastern

Colombia. Caldasia, 28, 135–155.

Marsh, D. M., & Pearman, P. B. (1997). Effects of Habitat Fragmentation on the

Abundance of Two Species of Leptodactylid Frogs in an Andean Montane Forest.

Conservation Biology, 11, 1323–1328.

Martin, T. L., & Huey, R. B. (2008). Why “Suboptimal” is Optimal: Jensen’s Inequality

and Ectotherm Thermal Preferences. The American Naturalist, 171, 102–118.

Mazzotti, F. J., Smith, B. J., Squires, M. A., Cherkiss, M. S., Farris, S. C., Hackett, C.,

Hart, K. M., Briggs-Gonzalez, V., & Brandt, L. A. (2019). Influence of salinity on

relative density of American crocodiles (Crocodylus acutus) in Everglades National

Park: Implications for restoration of Everglades ecosystems. Ecological Indicators,

102, 608–616.

Mendenhall, C. D., Frishkoff, L. O., Santos-Barrera, G., Pacheco, J., Mesfun, E.,

Quijano, F. M., Ehrlich, P. R., Ceballos, G., Daily, G. C., & Pringle, R. M. (2014).

Countryside biogeography of Neotropical reptiles and amphibians. Ecology, 95, 856–

870.

Mendoza, Á. M., Ospina, O. E., Cárdenas-Henao, H., & García-R, J. C. (2015). A

likelihood inference of historical biogeography in the world’s most diverse terrestrial

vertebrate genus: Diversification of direct-developing frogs (Craugastoridae:

Pristimantis) across the Neotropics. Molecular Phylogenetics and Evolution, 85, 50–

58.

Navas, C. A., Carvajalino-Fernández, J. M., Saboyá-Acosta, L. P., Rueda-Solano, L. A.,

& Carvajalino-Fernández, M. A. (2013). The body temperature of active amphibians

along a tropical elevation gradient: Patterns of mean and variance and inference from

environmental data. Functional Ecology, 27, 1145–1154.

Newbold, T. (2018). Future effects of climate and land-use change on terrestrial

vertebrate community diversity under different scenarios. Proceedings of the Royal

Society B, 285, 20180792.

Nowakowski, A. J., Frishkoff, L. O., Agha, M., Todd, B. D., & Scheffers, B. R. (2018a).

Changing Thermal Landscapes: Merging Climate Science and Landscape Ecology

through Thermal Biology. Current Landscape Ecology Reports, 3, 57–72.

Nowakowski, A. J., Frishkoff, L. O., Thompson, M. E., Smith, T. M., & Todd, B. D.

(2018b). Phylogenetic homogenization of amphibian assemblages in human-altered

habitats across the globe. Proceedings of the National Academy of Sciences, 115,

E3454–E3462.

Nowakowski, A. J., Watling, J. I., Thompson, M. E., Brusch, G. A., Catenazzi, A.,

Whitfield, S. M., Kurz, D. J., Suárez-Mayorga, Á., Aponte-Gutiérrez, A., Donnelly,

M. A., & Todd, B. D. (2017). Thermal biology mediates responses of amphibians and

reptiles to habitat modification. Ecology Letters, 21, 345–355.

Nowakowski, A. J., Watling, J. I., Whitfield, S. M., Todd, B. D., Kurz, D. J., & Donnelly,

M. A. (2016). Tropical amphibians in shifting thermal landscapes under land-use and

climate change. Conservation Biology, 31, 96–105.

Paoletti, A., Darras, K., Jayanto, H., Grass, I., Kusrini, M., & Tscharntke, T. (2018).

Amphibian and reptile communities of upland and riparian sites across Indonesian oil

palm, rubber and forest. Global Ecology and Conservation, 16, e00492.

Parmesan, C., & Yohe, G. (2003). A globally coherent fingerprint of climate change

impacts across natural systems. Nature, 421, 37–42.

Pineda, E., Moreno, C., Escobar, F., & Halffter, G. (2005). Frog, Bat, and Dung Beetle

Diversity in the Cloud Forest and Coffee Agroecosystems of Veracruz, Mexico.

Conservation Biology, 19, 400–410.

Pintanel, P., Tejedo, M., Ron, S. R., Llorente, G. A., & Merino‐Viteri, A. (2019).

Elevational and microclimatic drivers of thermal tolerance in Andean Pristimantis

frogs. Journal of Biogeography, 00, 1–12.

Pyburn, W. F., & Lynch, J. D. (1981). Two Little-Known Species Of Eleutherodactylus

(Amphibia: Leptodactylidae) from the Sierra De La Macarena - Colombia.

Proceedings of the Biological Society of Washington, 94, 404–412.

Pyron, R. A., & Wiens, J. J. (2011). A large-scale phylogeny of Amphibia including over

2800 species, and a revised classification of extant frogs, salamanders, and caecilians.

Molecular Phylogenetics and Evolution, 61, 543–583.

Raxworthy, C. J., Pearson, R. G., Rabibisoa, N., Rakotondrazafy, A. M.,

Ramanamanjato, J. B., Raselimanana, A. P., Wu, S., Nussbaum, R. A., & Stone, D.

A. (2008). Extinction vulnerability of tropical montane endemism from warming and

upslope displacement: A preliminary appraisal for the highest massif in Madagascar.

Global Change Biology, 14, 1703–1720.

Riemann, J. C., Ndriantsoa, S. H., Raminosoa, N. R., Rödel, M. O., & Glos, J. (2015).

The value of forest fragments for maintaining amphibian diversity in Madagascar.

Biological Conservation, 191, 707–715.

Robinson, D., Warmsley, A., Nowakowski, A. J., Reider, K. E., & Donnelly, M. A.

(2013). The value of remnant trees in pastures for a neotropical poison frog. Journal

of Tropical Ecology, 29, 345–352.

Royle, J. A. (2004). N-Mixture Models for Estimating Population Size from Spatially

Replicated Counts. Biometrics, 60, 108–115.

Sauer, E. L., Sperry, J. H., & Rohr, J. R. (2016). An efficient and inexpensive method for

measuring long-term thermoregulatory behavior. Journal of Thermal Biology, 60,

231–236.

Scheffers, B. R., Brunner, R. M., Ramirez, S. D., Shoo, L. P., Diesmos, A., & Williams,

S. E. (2013). Thermal buffering of microhabitats is a critical factor mediating

warming vulnerability of frogs in the Philippine biodiversity hotspot. Biotropica, 45,

628–635.

Scheffers, B. R., Edwards, D. P., Diesmos, A., Williams, S. E., & Evans, T. A. (2014).

Microhabitats reduce animal’s exposure to climate extremes. Global Change Biology,

20, 495–503.

Schneider-Maunoury, L., Lefebvre, V., Ewers, R. M., Medina-Rangel, G. F., Peres, C.

A., Somarriba, E.,Urbina-Cardona, N., & Pfeifer, M. (2016). Abundance signals of

amphibians and reptiles indicate strong edge effects in Neotropical fragmented forest

landscapes. Biological Conservation, 200, 207–215.

Shah, A. A., Funk, W. C., & Ghalambor, C. K. (2017). Thermal Acclimation Ability

Varies in Temperate and Tropical Aquatic Insects from Different Elevations.

Integrative and Comparative Biology, 57, 977–987.

Sinervo, B., Mendez-de-la-Cruz, F., Miles, D. B., Heulin, B., Bastiaans, E., Villagran-

Santa Cruz, M.,et al, & Sites, J. W. J. (2010). Erosion of lizard diversity by climate

change and altered thermal niches. Science, 328, 894–899.

Stellatelli, O. A., Vega, L. E., Block, C., & Cruz, F. B. (2013). Effects on the

thermoregulatory efficiency of two native lizards as a consequence of the habitat

modification by the introduction of the exotic tree Acacia longifolia. Journal of

Thermal Biology, 38, 135–142.

Sunday, J. M., Bates, A. E., & Dulvy, N. K. (2011). Global analysis of thermal tolerance

and latitude in ectotherms. Proceedings of the Royal Society B: Biological Sciences,

278, 1823–1830.

Sunday, J. M., Bates, A. E., Kearney, M. R., Colwell, R. K., Dulvy, N. K., Longino, J. T.,

& Huey, R. B. (2014). Thermal-safety margins and the necessity of thermoregulatory

behavior across latitude and elevation. Proceedings of the National Academy of

Sciences, 111, 5610–5615.

Tewksbury, J. J., Huey, R. B., & Deutsch, C. A. (2008). Putting the Heat on Tropical

Animals. Science, 320, 1296–1297.

Thompson, M. E., Halstead, B. J., & Donnelly, M. A. (2018). Thermal quality influences

habitat use of two anole species. Journal of Thermal Biology, 75, 54–61.

Toral, E. C., Feinsinger, P., & Crump, M. L. (2002). Frogs and a cloud-forest edge in

Ecuador. Conservation Biology, 16, 735–744.

Tuff, K. T., Tuff, T., & Davies, K. F. (2016). A framework for integrating thermal

biology into fragmentation research. Ecology Letters, 19, 361–374.

Urbina-Cardona, J. N., Olivares-Perez, M., & Reynoso, V. H. (2006). Herpetofauna

diversity and microenvironment correlates across a pasture–edge–interior ecotone in

tropical rainforest fragments in the Los Tuxtlas Biosphere Reserve of Veracruz,

Mexico. Biological Conservation, 132, 61–75.

Yang, J., Hong, Y. S., & Xiang, A. (2008). Northern grass lizards (Takydromus

septentrionalis) from different populations do not differ in thermal preference and

thermal tolerance when acclimated under identical thermal conditions. Journal of

Comparative Physiology B: Biochemical, Systemic, and Environmental Physiology,

178, 343–349.

Supplemental Materials

Fig. S1 Two-sample two-tailed t-test comparing mean CTmax determined from righting response assays to mean CTmax calculated from jumping performance trials (p = 0.72353) for Pristimantis medemi.

Fig. S2 Mean daily temperature range (mean daily maximum - mean daily minimum) in clearing (white squares) and forest (black dots) at corresponding elevations.

in clearings in

R2 = 0.8285

P. medemi medemi P. Proportion Proportion

Thermal Quality Fig. S3 Thermal quality and proportional abundance of P. medemi in clearings.

in clearings in

R2 = 0.1413

P. frater frater P. Proportion Proportion

Thermal Quality Fig. S4 Thermal quality and proportional abundance of P. frater in clearings.

in clearings in

R2 = 0.2269

savagei

P. Proportion Proportion

Thermal Quality Fig. S5 Thermal quality and proportional abundance of P. savagei in clearings.

Fig. S6 Pristimantis frater male and female in amplexus in the clearing at Buena Vista, 1180 m. P. frater amplectic pair from Manzanares, 1350 m not shown.

Fig. S7 Means of environmental variables collected at each field site in forest (blue) and clearing (orange) habitats. Error bars represent standard error. X-axis is elevation (meters). 54

Table S1 List of covariates used to create two sets of N-mixture models, to estimate detection probability (p) and relative density (λ) of Pristimantis medemi within forest and clearing habitat at each elevation. One model set was built for forest and a separate model set was built for clearings. Detection (p); Relative Variable (units) Abbrev. Type density (λ)

Number of observers Obs Continuous p Stem density (m-1) Stem Continuous p Elevation (meters asl) Elev Continuous λ Leaf litter depth (mm) LL Continuous λ Mean temperature (°C) Temp Continuous λ Canopy cover (%) CC Continuous λ

Table S2 AIC values of candidate models for clearings with AIC values below 200. All models assumed a Negative Binomial variance structure. Model AIC ΔAIC AIC Weight

LL, Elev, Temp, Obs, Stem 144.61 0.00 0.7585 LL, Elev, Temp, CC Obs, Stem 147.14 2.53 0.2141 LL, Elev, Obs, Stem 152.30 7.69 0.0162 LL, Obs, Stem 154.22 9.61 0.0062 LL, Elev, Stem 154.64 10.03 0.0050

Table S3 AIC values of candidate models for forest with AIC values below 200. All models assumed a Negative Binomial variance structure. Model AIC ΔAIC AIC Weight

LL, Elev, Temp, Obs, Stem 157.77 0.00 0.7238 LL, Elev, Temp, CC, Obs, Stem 159.76 1.99 0.2676 LL, Temp, Obs, Stem 166.68 8.91 0.0084 LL, Elev, Obs, Stem 173.72 15.95 0.0002