Copeia 2014, No. 3, 454–461

Litter Dynamics Regulate Population Densities in a Declining Terrestrial Herpetofauna Steven M. Whitfield1,2, Kelsey Reider1, Sasha Greenspan3, and Maureen A. Donnelly1 Loss of biodiversity within relatively pristine protected areas presents a major challenge for conservation. At La Selva Biological Station in the lowlands of Costa Rica, amphibians, reptiles, and understory birds have all declined over the past four decades, yet the factors contributing to these declines remain unclear. Here, we conduct two tests of the hypothesis that faunal declines are linked to shifting dynamics of leaf litter, a critical microhabitat for amphibians and reptiles and a major component of forest carbon cycles. First, we conduct a 16-month manipulation of leaf litter and measure response by terrestrial amphibians and reptiles. Second, we synthesize three year-long datasets collected over four decades to evaluate potential multi-decade change in standing litter depth. We show that litter depth regulates density of amphibians and reptiles, and that the strongest response to manipulations is in species that decline most rapidly based on long-term data. Our synthesis of litter depth data suggests considerable interannual variability in standing stocks of leaf litter with lowest quantity of leaf litter in the most recent sampling period. These tests are consistent with the hypothesis that these faunal declines may be in part driven by changes in forest litter dynamics, and ultimately to climate-sensitive carbon cycles.

A

MPHIBIAN assemblages across the world have suffered rapid, unexpected population declines and widespread extinctions (Alford and Richards, 1999; Young et al., 2001; Stuart et al., 2004). Conventional threats such as habitat loss and modification are directly responsible for many of these declines, yet much research attention has been focused on so-called ‘‘enigmatic declines’’ (Lips et al., 2006, 2008; Pounds et al., 2006)—those declines occurring in protected habitats that cannot be attributed to obvious local anthropogenic disturbances (Stuart et al., 2004). One of the best-documented examples of such declines is at La Selva Biological Station—a protected lowland (,150 m asl) reserve in Costa Rica where populations of terrestrial amphibians declined by ,75% between 1970 and 2005 (Whitfield et al., 2007). These declines cannot be attributed to direct anthropogenic effects (i.e., habitat loss, overexploitation) because La Selva is protected and has experienced minimal recent human disturbance. La Selva’s declines are largely inconsistent with other published accounts of amphibian declines because these declines occurred over decades rather than a few months, and because virtually all other sites in the Neotropics where widespread amphibian declines have been reported are from montane regions with cooler climates (generally .400 m asl; Whitfield et al., in press). A comprehensive understanding of causative agents of declines is critical for effective conservation management of amphibian biodiversity. Declines at La Selva may be a part of broader ecosystemwide shifts ultimately tied to climatic change (Clark et al., 2003, 2011; Sigel et al., 2006). Long-term meteorological data from La Selva indicates an average 0.25uC increase in air temperature per decade since the 1960s, primarily driven by increases in daily minimum temperatures (Clark and Clark, 2011). These temperature increases drive reductions in tree growth (Clark et al., 2003) and increases in tree mortality (Clark et al., 2011). Long-term vertebrate populations trends show major declines in density of not only terrestrial

1

amphibians, but also terrestrial reptiles (Whitfield et al., 2007) and understory insectivorous birds (Sigel et al., 2006). Leaf litter depth has been correlated with density of both terrestrial amphibians and terrestrial reptiles across seasons, elevations, and biogeographic regions (Fauth et al., 1989; Scott, 1976; Lieberman, 1986; Heinen, 1992; Whitfield and Pierce, 2005), and is the base of brown food webs upon which understory insectivorous birds rely (Sigel et al., 2006). Most biomass produced by trees eventually falls to the forest floor as leaf litter, thus climate-induced reductions in tree growth may be associated with reductions in litter production. Further, decomposition rates are climate-sensitive and increase with temperature (Raich et al., 2006). Thus, long-term shifts in vertebrate population densities may be the result of broader ecosystem-level shifts in quantity of standing leaf litter, and ultimately, carbon balance in tropical forests. Here, we present two tests of the hypothesis that longterm reductions in standing litter is a component of faunal declines at La Selva. First, we conduct a replicated field experiment to demonstrate that leaf litter amphibian and reptile populations are limited by quantity of standing leaf litter. Second, we use two year-long historic datasets of standing litter quantity with a two-year modern dataset using the same methodology to show potential long-term change in standing litter stocks. MATERIALS AND METHODS Study site.—La Selva Biological Station is a 16 km2 private biological reserve in the lowland wet forest of Sarapiquı´, Costa Rica (10u269N, 83u599W, elevation 35–137 asl). La Selva receives on average 4000 mm annual precipitation, with no month on average receiving ,100 mm rainfall. The majority of the site is primary forest, apparently with little recent direct human impact (McDade and Hartshorn, 1994). Study plots and experimental design.—We established nine 15 3 15 m permanent study plots within old-growth forest. We

Florida International University, Department of Biological Sciences, Miami, Florida. Present address: Gonzaga University, Biology Department, 502 E Boone Avenue, Spokane, Washington 99258; E-mail: [email protected]. Send reprint requests to this address. 3 Joseph W. Jones Ecological Research Center, 3988 Jones Center Drive, Newton, Georgia. Submitted: 28 May 2013. Accepted: 12 March 2014. Associate Editor: J. W. Snodgrass. DOI: 10.1643/CE-13-061 Published online: October 8, 2014 F 2014 by the American Society of Ichthyologists and Herpetologists 2

Whitfield et al.—Leaf litter and population declines

judged these plots to be of sufficient size for this study because the leaf litter amphibians and reptiles studied here are highly philopatric and have home ranges much less than 15 3 15 m (Donnelly, 1987, 1989), and because plots of this size have detected pronounced trends in population size for these species previously (Donnelly, 1987, 1989; Guyer, 1988). We arranged plots in three spatial blocks to maximize spatial extent of the study and to encompass variation in habitat characteristics, but to minimize spatial heterogeneity within replicates. We separated each plot within a spatial block by between 34 and 103 m (mean 5 59.8 m). Our basic experimental design involved a six-month pre-treatment study period (September 2006 through March 2007) and a ten-month post-treatment period (March 2007 through December 2007). Amphibian and reptile sampling.—We sampled amphibians and reptiles on each plot every other week from 1 September 2006 to 14 December 2007, for a total of 33 sampling ‘‘sessions’’ per plot. Each sampling session consisted of three samples of each plot, generally on three subsequent days, but in a few instances (35 of 890 total sampling sessions) sampling occurred over three days within a four-day period. The three plots within each sampling block were always sampled subsequently to control for variation in weather or amphibian and reptile activity, and the team of observers fixed among plots within a block. We randomized the sampling sequence of spatial blocks, and the sampling sequence of treatments within each block in advance of that sampling session. During each sampling event, between one and three observers searched each plot for ,30 min. Observers carefully scanned the ground and all vegetation to a height of ,2 m for any amphibians and reptiles, and gently agitated leaf litter with a probe to elicit movement by individuals of cryptic species. Litter manipulation.—Between 12 March 2007 and 31 March 2007, we randomly assigned plots to one of three treatments: litter addition (L+), litter removal (L2), and shamtreatment control (L0). At the initial manipulation of litter quantity, we removed virtually all standing leaf litter from the L2 plots, inspected leaf litter by hand to remove any amphibians or reptiles (which were released at the same point on the plot where they were originally located), and applied the entire quantity of leaf litter to the adjacent L+ plot in the same spatial block. For our sham treatment, we removed all leaf litter from the L0 plot and immediately replaced the leaf litter on the same plot. To maintain these treatments, we continued to remove recently senesced leaf litter every two weeks from all L2 plots, and applied this leaf litter to adjacent L+ plots. Litter removal and addition was conducted immediately after sampling from that two-week period. We stratified treatments by spatial blocks so that each block received one treatment plot of each type. Current litter dynamics.—Throughout the study, we measured the quantity of standing leaf litter on the study plots. At each standing litter measurement period, we calculated the amount of standing leaf litter at 20 random locations in each plot, for a total of 2965 measurement points. At each point, we measured leaf litter using three metrics: ‘‘measured depth’’—the distance between the top of the soil and the top of the leaf litter measured to the nearest 0.1 mm with dial calipers; ‘‘count depth’’—the number of leaves pierced by the probe on the dial calipers; and ‘‘litter cover’’—

455

presence or absence of leaf litter. While these three metrics are correlated, they provide different information about the quantity of standing leaf litter. In the post-treatment period, we measured standing litter quantity immediately before litter manipulation; our measurements provide a conservative estimate of litter quantity because they represent two weeks of relaxation following manipulation. Microclimate.—To determine whether standing leaf litter affects microhabitat characteristics that may regulate population densities of litter amphibians and reptiles, we deployed a network of small temperature dataloggers (Embedded Data Systems Hygrochron iButtons). These data loggers were deployed between 25 and 29 January 2007, during the pre-treatment phase, and collected data nearly continuously until the termination of the experiment. We programmed dataloggers to record temperature and humidity data every 30 minutes, and one data logger was placed under the leaf litter at a randomly selected point; a paired datalogger was placed approximately 30 cm above the soil at the same random location. Two pairs of dataloggers were placed in each plot. We analyzed microclimate data (temperature and relative humidity) with linear mixed models with treatment (L0, L2, L+), period (pre-treatment, post-treatment), and datalogger location (under leaf litter or in air) as fixed factors and with datalogger location as a random factor. Long-term litter depth.—To determine whether substantial changes have occurred in quantity of leaf litter on the forest floor, we compiled two historic datasets that had measured standing litter depth from previous decades (Lieberman, 1986; Whitfield et al., 2007). These historic studies (1973– 1974 and 1985–1986) measured depth of leaf litter at each corner of 8 3 8 m litter quadrats randomly located throughout primary forest within the La Selva reserve. Both studies used the same methodology, and each study collected monthly samples over a .12-month period. Each study estimated ‘‘measured depth,’’ ‘‘count depth,’’ and ‘‘litter cover’’ as described above. We replicated the methodology of these two studies between May 2006 and March 2008. Analysis.—We analyzed quantity of standing litter on plots with generalized linear mixed effects models using the lmer procedure in the lme4 package of R, specifying treatment and pre-treatment/post-treatment as fixed effects and block, plot, and sample as random effects to account for repeated measures on plots. We specified Gaussian errors for depth in mm, Poisson errors for count depth, and binomial errors for litter cover. We evaluated changes in number of encounters using a generalized linear mixed effects model with Poisson errors in lmer of the package lme4. We used number of encounters per sample session as a response variable, specified treatment and treatment period as fixed factors, and included block and plot as random factors. We conducted this analysis for all amphibians and reptiles, for pooled frogs, pooled lizards, and for each of the six most commonly encountered species in this study (the frogs Oophaga pumilio, Craugastor bransfordii, and Craugastor mimus; and the lizards Norops humils, Norops limifrons, and Sphenomorphus cherriei). To analyze change in measured depth we used a linear mixed effects model with decade (1970s, 1980s, 2000s) as factors and month and quadrat identity as random factors.

456

Fig. 1. Effects of litter manipulation on three metrics of litter depth in the pre- and post-treatment periods.

To analyze change in count depth and litter cover, we used generalized linear models using Poisson and binomial error distributions and the random factors specified as for measured depth. RESULTS Current litter dynamics.—In the pre-treatment period, count depth and litter cover did not differ among experimental treatments (all P . 0.1; Fig. 1), but measured depth was greater in both L+ (z 5 2.14, P 5 0.032) and L2 (z 5 2.19, P 5 0.028) plots than in L0 plots. Experimental manipulations of litter depth caused a dramatic reduction in quantity of standing litter on L2 plots (measured depth: z 5 27.83, P , 0.0001; count depth: z 5 210.506, P , 0.0001; litter cover: z 5 26.763, P , 0.0001; Fig. 1). Litter addition produced a dramatic increase in depth of standing leaf litter (measured depth: z 5 2.75, P 5 0.006; count depth: z 5 6.911, P , 0.0001; litter cover: z 5 4.516, P , 0.0001; Fig. 1). There was a much less severe increase in quantity of standing litter on L0 plots during the post-treatment period according to leaf counts and measured litter depth, but not litter cover (measured depth: z 5 2.88, P 5 0.004; count depth: z 5 2.564, P 5 0.010; litter cover: z 5 1.449, P 5 0.147; Fig. 1). Encounters.—We encountered a total of 4192 amphibians and reptiles, including 3020 frogs representing 20 species,

Copeia 2014, No. 3

1152 lizards representing 13 species, and 18 snakes among seven species (Table 1). For all amphibians and reptiles pooled, there were increases in total encounters in both L0 (z 5 6.694, P , 0.0001) and L+ plots in the post-treatment period (although no difference in size of increase between L0 and L+ (z 5 0.265, P , 0.791), and a strong decrease in total encounters in the L2 plots (z 5 23.758, P 5 0.002; Fig. 2A). For all frogs, there was a similar pattern, with more encounters in L0 (z 5 6.700, P , 0.0001) and L+ plots, and a smaller increase in number of encounters for L2 plots relative to L0 plots (z 5 23.522, P 5 0.0004), but no difference in size of increase between L0 and L+ plots (z 5 1.878, P 5 0.060; Fig. 2B). For all lizards, while there was an increase in number of encounters for L0 plots (z 5 2.680, P 5 0.007), there was a decrease in total number of encounters for both L2 (z 5 22.387, P 5 0.017) and L+ (z 5 22.578, P 5 0.0010; Fig. 2C) plots relative to L0 plots. Species showed individualistic responses to litter manipulations (Table 2, Fig. 2). Encounters of Oophaga pumilio did not change on L0 plots (z 5 1.326, P 5 0.185) or L2 plots (z 5 21.620, P 5 0.105), but increased on L+ plots (z 5 2.147, P , 0.032). Encounters of Craugastor bransfordii increased on L0 plots (z 5 6.289, P , 0.0001), but decreased on L2 plots relative to L0 plots (z 5 21.912, P 5 0.056) and the increase on L+ plots was comparable to increases for L0 plots (z 5 0.648, P 5 0.517). Encounters of Craugastor mimus increased on L0 (z 5 3.147, P 5 0.0017) and relative to L0 plots, decreased on L2 plots (z 5 22.242, P 5 0.0025) but response on L+ plots did not differ from L0 or L+ plots (z 5 0.543, P 5 0.586). Encounters of Norops humilis showed no change in L0 (z 5 0.668, P 5 0.504), and L+ treatments did not differ from L0 plots (z 5 21.780, P 5 0.075), but there was a decrease on L2 plots relative to L0 plots (z 5 22.489, P 5 0.013). Encounters of Norops limifrons showed no change in the post-treatment period in L0 plots (z 5 1.897, P 5 0.578), but there were fewer encounters relative to L0 plots on both L2 plots (z 5 20.608, P 5 0.5429) and L+ plots (z 5 23.606, P 5 0.0003). Encounters of Sphenomorphus cherriei increased on both L0 plots (z 5 2.372, P , 0.018) and L+, and although there was no difference between L0 plots and L+ plots (z 5 0.245, P 5 0.806), there were far fewer encounters on L2 plots relative to L0 plots (z 5 23.592, P 5 0.0003). Effect sizes for pooled and individual taxa are given in Table 2. Microclimate.—Average daily air temperatures in the pretreatment period (23.0uC) were slightly higher than average soil temperatures (22.8uC, Table 3A), maximum daily temperatures air were higher in air (26.0uC) than in soil (25.0uC, Table 3B), and minimum daily temperatures were lower in air (20.9uC) than in soil (22.2uC, Table 3C). In the pretreatment period, there was no effect of treatment on average daily temperature (Table 3A), maximum daily temperature (Table 3B), or minimum daily temperature (Table 3C). Both air and soil temperatures in the posttreatment period were warmer than in the pre-treatment period (Table 3). There was no interaction between treatment and treatment period for average daily temperature (Table 3A), but there were significant interactions between treatment and treatment period for maximum daily temperatures (Table 3B) and for minimum daily temperatures (Table 3C). These interactions between treatment and treatment period for air temperature are attributable to a smaller increase in average air temperatures in the L2 treatment compared to the L0 treatment, a larger increase

Whitfield et al.—Leaf litter and population declines

457

Table 1. Encounters by treatment period and manipulation type.

Pre-treatment Taxon Frogs Rhaebo haematiticus Incilius melanochlorus Craugastor bransfordii C. crassidigitus C. fitzingeri C. megacephalus C. mimus C. talamancae Pristimantis cerasinus P. cruentus Diaspora diastema P. ridens Lithobates warsczewitchii Oophaga pumilio Leptodactylus savagei Gastrophryne pictiventris Hyalinobatrachium fleishmanii Scinax boulengeri S. eleaochroa Smilisca baudinii Unidentified frog Lizards Corytophanes cristatus Norops biporcatus N. capito N. carpenteri N. humilis N. lemurinus N. limifrons Ameiva festiva Lepidophyma flavimaculata Lepidoblepharis xanthostigma Gonatodes humeralis Thecadactylus rapicaudus Sphenomorphus cherriei Unidentified lizard Snakes Coniophanes fissidens Imantodes cenchoa Leptophis mexicanus Rhadinea decorata Pseustes poecilochilonotus Bothrops asper Porthidium nasutum Unidentified snake Unidentified Total

Post-treatment

L0

L2

L+

109

1 3 90

2 1 136

1 13 5 3

8 2 5

L0 3

1

299 1 5 2 31 2 21

1 4 7

L2

L+

Total

2 5 178

1 411 3

13 11 15

2

10

14

10

11

177

201

169

1 273

1 249 6

61 4 11 1 14 1 1 348

1 1 1 5 1 4

5

1 2 60 36 12

5 1

7 9

10

16

5

1

1 2

2

1 91 4 72 18

7 2 3 1 55 5 76 27

22 7

1 2 1

67 1 45 9

12 2

83 2 103 13 1 1 3 12 1

1 84 5 83 32 1 2 61 5 2

2 1

1 1

1 1

1 2 1

436

466

in maximum air temperatures in the L2 treatment than in L0 or L+ treatments, and a smaller increase in minimum air temperatures in the L2 treatment than in L 0 or L + treatments. For soil temperatures, there was a higher increase in average soil temperatures in the post-treatment period in L2 treatment than in L0 or L+ treatments and a higher increase in maximum daily soil temperatures in the

568

1 1 1 876

1 2 1 683

2 1165

8 10 1223 1 10 32 123 11 53 1 61 1 3 1417 6 1 1 1 12 1 44 14 4 7 4 440 17 415 111 2 1 5 1 114 17 2 2 1 3 1 1 4 4 0 4 4194

L2 treatment than in L0 or L+ treatments, but no difference in the rate of change for minimum temperatures. Long-term litter depth.—There was no change in measured litter depth between the three study periods (1973–1974, 1994–1995, 2006–2008; F2,989 5 0.40331, P 5 0.668; Fig. 3A). There was a difference in count depth between

458

Copeia 2014, No. 3

sampling period, but no change between the 1973–1974 and 1995–1996 sampling periods (Fig. 3C). DISCUSSION

Fig. 2. Number of encounters per sampling session in response to manipulations and controls for (A) total amphibian and reptile encounters, (B) encounters for frogs alone, and (C) encounters for lizards alone.

study periods (x25 5 25.913, P , 0.0001; Fig. 3B) attributable to a sharp decrease in number of leaves between the 1995– 1996 sampling period and the 2006–2008 sampling period, but no change between the 1973–1974 and 1995–1996 sampling periods. There was also a change in litter cover (x25 5 43.563, P , 0.0001) attributable to a sharp decrease between the 1995–1996 sampling period and the 2006–2008

Our study demonstrated that approximately 50% reductions in standing litter mass have dramatic effects on number of encounters for both amphibians and reptiles, confirming correlations detected previously (Scott, 1976; Lieberman, 1986; Fauth et al., 1989; Heinen, 1992). For all sampled species, effect sizes for litter removal treatments were negative, consistent with our expectations. Generally, there were increases in number of encounters in the posttreatment period in both control and litter addition treatments. Further, we found that increasing litter depth produced a range of responses for focal species, including either increases or decreases in number of encounters. Only one species, Oophaga pumilio, significantly increased in litter addition treatments relative to controls, although nonsignificant effect sizes for common species were positive with the exception of two species of Norops (Table 1). Differences in detection probability among our manipulations likely result in a severe underestimation of effect sizes reported here. We expect that because litter is a common refuge from predators, greater litter depth reduced detection probability for most or all species of leaf litter amphibians and reptiles. Consequently, we expect detection probability to be highest on litter removal treatments, intermediate on control treatments, and lowest on litter addition treatments. Such biases in detection probability likely cause considerable underestimations in effect size, and our estimates of effect size are likely highly conservative. The rates at which frog and lizard species have declined at La Selva vary considerably among species (Whitfield et al., 2007). If reductions in standing litter quantity are at least in part responsible for long-term declines in amphibian and reptile populations at La Selva, we may expect effect sizes for litter manipulation treatments to correlate with decline rate for sampled taxa. The small number of common species complicates efforts to correlate decline rate with effect size of experimental manipulations. However, the two species with the largest effect size in litter removal treatments were Craugastor mimus and Sphenomorphus cherriei—the frog and lizard species with the highest decline rate for those amphibians and reptiles sampled in this study (Whitfield et al., 2007). Litter may be important for terrestrial vertebrates because it provides a buffer of stable temperature and high humidity, or because leaf litter is the base of brown food webs upon which litter arthropods are trophically derived. Reductions

Table 2. Estimates of effect size (and 95% CIs in parentheses) for litter removal and litter addition on encounters of amphibians and reptiles.

Taxon All amphibians and reptiles All frogs All lizards Oophaga pumilio Craugastor bransfordii Craugastor mimus Norops humilis Norops limifrons Sphenomorphus cherriei

L2 20.3154 20.3520 20.3830 20.2192 20.3271 21.5061 20.6139 20.1691 23.273

(20.48, 20.15) (20.55, 20.16) (20.7, 20.07) (20.48, 0.05) (20.66, 0.01) (22.82, 20.19) (21.1, 20.13) (20.071, 0.38) (25.06, 21.49)

L+ 0.0206 0.1771 20.3772 0.2890 0.0969 0.3403 20.4046 20.9091 0.1443

(20.13, 0.17) (20.01, 0.36) (20.66, 20.09) (0.03, 0.55) (20.2, 0.39) (20.89, 1.57) (20.85, 0.04) (21.4, 20.41) (21.01, 1.3)

Whitfield et al.—Leaf litter and population declines

459

Table 3. GLM results for microclimate data. (A) Average temperature, (B) Daily maximum temperature, (C) Daily minimum temperature.

(A) Average temperature (Intercept) Sensor location (Air vs. Soil) Treatment period (Pre vs. Post) Treatment type (L0, L +, L2) Location x Treatment period Location x Treatment Treatment period x Treatment Location x Treatment period x Treatment

df 1, 1, 1, 2, 1, 2, 2, 2,

(B) Maximum temperature (Intercept) Sensor location (Air vs. Soil) Treatment period (Pre vs. Post) Treatment type (L0, L +, L2) Location x Treatment period Location x Treatment Treatment period x Treatment Location x Treatment period x Treatment

df 1, 1, 1, 2, 1, 2, 2, 2,

(C) Minimum temperature (Intercept) Sensor location (Air vs. Soil) Treatment period (Pre vs. Post) Treatment type (L0, L +, L2) Location x Treatment period Location x Treatment Treatment period x Treatment Location x Treatment period x Treatment

7429 7429 7429 63 7429 7429 7429 7429

7429 7429 7429 63 7429 7429 7429 7429 df

1, 1, 1, 2, 1, 2, 2, 2,

7429 7429 7429 63 7429 7429 7429 7429

in standing litter quantity could result in fewer refugia from periodic drought, to which both amphibians and the eggs of terrestrial lizards are highly sensitive (Schlaepfer, 2003; Socci et al., 2005). Holistic efforts to account for amphibian, reptile, and understory insectivorous bird population declines may require a more sophisticated understanding of how availability of arthropods for prey is related to fluctuating litter stocks (McGlynn et al., 2009). Our replication of historic measurements of litter depth provides the first evaluation of multi-decade fluctuations in quantity of standing leaf litter. While the measured depth of leaf litter did not change, the depth in number of leaves and litter cover were both significantly lower in 2006–2008 than in 1973–1974 and 1995–1996. The difference in response metrics is likely due to the high variability in measured depth on short timescales: dry litter may be rather voluminous, yet measured litter depth may be greatly reduced at the same point after compaction of litter by heavy rains. Because of this variation, the depth in number of leaves represents a more stable metric for assessing litter quantity with considerably less variation attributable to weather or observer effects. These data suggest reductions in standing litter over time, but we would refrain from interpreting these data as strong evidence of a sustained directional trend until more complete historical datasets become available. We urge researchers to give more detailed attention to long-term changes in litter dynamics in lowland tropical forests.

F

P

135873.4 19.34 995.52 0.12 0.26 4.81 1.32 2.45

,0.0001 ,0.0001 ,0.0001 0.8881 0.6131 0.0081 0.2669 0.0867

F

P

61091.43 613.19 170.94 0.3 1.8 35.47 10.97 0.95

,0.0001 ,0.0001 ,0.0001 0.7405 0.1799 ,0.0001 ,0.0001 0.3862

F

P

201192.09 1236.31 800.19 0.09 19.42 21.47 7.76 7.95

,0.0001 ,0.0001 ,0.0001 0.9183 ,0.0001 ,0.0001 0.0004 0.0004

At least three non-mutually exclusive hypotheses may contribute to interannual variability or sustained directional changes in quantity of leaf litter. Whitfield et al. (2007) suggested that climatic shifts at La Selva trend toward a more consistently wet climate, potentially reducing litterfall which is triggered by drought (Frankie et al., 1974) or accelerating decomposition which is presumably constrained by dry-season moisture limitation (Wieder and Wright, 1995). Further, periodic climatic events such as El Nino/La Nina events may have profound impacts on either litterfall or litter decomposition. Alternatively, apparent and rapid population growth of collared peccaries at La Selva has been suggested to impact litter quantity though mechanical trampling and rooting in leaf litter (Reider et al., 2013). Further, non-native earthworms have become established at La Selva in recent decades, and are well-known to accelerate decomposition in other sites. Our long-term litter data indicate that there was no substantial difference in litter depth between 1973–1974 and 1994–1995, though there was considerably less litter from 2006–2008 than in past sampling. This timescale corresponds both to apparent increases in population size of collared peccaries, as well as increased frequency of ENSO events. However, Whitfield et al. (2007) report that declines in populations of amphibians and reptiles were well underway by the mid-1990s, which is somewhat inconsistent with the timing of population increases of peccaries, and experimental exclusion of peccaries from forest plots did not increase densities of leaf

460

Copeia 2014, No. 3

provided funding to SMW. FIU IACUC and MINAET provided permits. T. Doan provided valuable comments. LITERATURE CITED

Fig. 3. Comparison of litter depth among three studies over three decades.

litter amphibians and reptiles (Reider et al., 2013). In any case, while changes to standing litter quantity may be one factor contributing to declines in populations of amphibians and reptiles, it is at best one of a suite of long-term changes at La Selva which may be affecting populations of frogs, lizards, and other vertebrates (Whitfield et al., 2012a, 2012b, 2013, in press; Young et al., 2008). Tropical ecosystems are poorly studied, and have been described as ‘‘data vacuums’’(Gardner et al., 2007), yet La Selva is one of a very small number of tropical forest sites for which detailed and long-term data exist for a broad number of ecological metrics (Clark, 1990). By linking independent studies of long-term change in vertebrate populations to climate-sensitive carbon cycles, this study underscores the critical importance of biological stations as ecological observatories for monitoring biodiversity and ecosystem function. The results of this study underline the critical need for a more comprehensive understanding of linkages between climate, carbon cycling, and biodiversity. ACKNOWLEDGMENTS L. Gentry, N. Ratledge, S. Miller, M. Smith, and many others helped in the field. FIU Presidential Fellowship, FIU Dissertation Year Fellowship, EPA GRO fellowship, FIU LACC Tinker Grant, AMNH Roosevelt Grant, OTS Graduate Research Fellowship, APS Lewis and Clark Fund Grant

Alford, R. A., and S. J. Richards. 1999. Global amphibian declines: a problem in applied ecology. Annual Review of Ecology and Systematics 30:133–160. Clark, D. A., and D. B. Clark. 2011. Assessing tropical forests’ climatic sensitivities with long-term data. Biotropica 43:31–40. Clark, D. A., S. C. Piper, C. D. Keeling, and D. B. Clark. 2003. Tropical rain forest tree growth and atmospheric carbon dynamics linked to interannual temperature variation during 1984–2000. Proceedings of the National Academy of Sciences of the United States of America 100:5852–5857. Clark, D. B. 1990. La Selva Biological Station: a blueprint for stimulating tropical research, p. 9–27. In: Four Neotropical Forests. A. H. Gentry (ed.). Yale University Press, New Haven, Connecticut. Clark, D. B., D. A. Clark, and S. F. Oberbauer. 2011. Annual wood production in a tropical rain forest in NE Costa Rica linked to climatic variation but not to increasing CO2. Global Change Biology 16:747–759. Donnelly, M. A. 1987. Territoriality in the poison-dart frog Dendrobates pumilio (Anura: Dendrobatidae). Unpubl. Ph.D. diss., University of Miami, Miami. Donnelly, M. A. 1989. Effects of reproductive resource supplementation on space-use patterns in Dendrobates pumilio. Oecologia 81:212–218. Fauth, J. E., B. I. Crother, and J. B. Slowinski. 1989. Elevational patterns of species richness, evenness, and abundance of the Costa Rican leaf-litter herpetofauna. Biotropica 21:178–185. Frankie, G. W., H. G. Baker, and P. A. Opler. 1974. Comparative phenological studies of trees in tropical wet and dry forests in the lowlands of Costa Rica. Journal of Ecology 62:881–919. Gardner, T. A., J. Barlow, L. W. Parry, and C. A. Peres. 2007. Predicting the uncertain future of tropical forest species in a data vacuum. Biotropica 39:25–30. Guyer, C. 1988. Food supplementation in a tropical mainland anole, Norops humilis: effects on individuals. Ecology 69:362–369. Heinen, J. T. 1992. Comparisons of the leaf-litter herpetofauna in abandoned cacao plantations and primary rain forest in Costa Rica: some implications for faunal restoration. Biotropica 24:431–439. Lieberman, S. S. 1986. Ecology of the leaf litter herpetofauna of a Neotropical rain forest: La Selva, Costa Rica. Acta Zoologica Mexicana (ns) 15:1–72. Lips, K. R., F. Brem, R. Brenes, J. D. Reeve, R. A. Alford, J. Voyles, C. Carey, L. Livo, A. P. Pessier, and J. P. Collins. 2006. Emerging infectious disease and the loss of biodiversity in a Neotropical amphibian community. Proceedings of the National Academy of Sciences of the United States of America 103:3165–3170. Lips, K. R., J. Diffendorfer, J. R. Mendelson, and M. W. Sears. 2008. Riding the wave: reconciling the roles of disease and climate change in amphibian declines. Plos Biology 6:441–454. McDade, L. A., and G. S. Hartshorn. 1994. La Selva Biological Station, p. 6–14. In: La Selva: Ecology and Natural History of a Neotropical Rain Forest. L. A.

Whitfield et al.—Leaf litter and population declines

McDade, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn (eds.). University of Chicago Press, Chicago. McGlynn, T. P., R. M. Fawcett, and D. A. Clark. 2009. Litter biomass and nutrient determinants of ant density, nest size, and growth in a Costa Rican tropical wet forest. Biotropica 41:234–240. Pounds, J. A., M. R. Bustamante, L. A. Coloma, J. A. Consuegra, M. P. L. Fogden, P. N. Foster, E. La Marca, K. L. Masters, A. Merino-Viteri, R. Puschendorf, S. R. Ron, G. A. Sanchez-Azofeifa, C. J. Still, and B. E. Young. 2006. Widespread amphibian extinctions from epidemic disease driven by global warming. Nature 439:161–167. Raich, J. W., A. E. Russell, K. Kitayama, W. J. Parton, and P. M. Vitousek. 2006. Temperature influences carbon accumulation in moist tropical forests. Ecology 87:76–87. Reider, K. E., W. P. Carson, and M. A. Donnelly. 2013. Effects of collared peccary (Pecari tajacu) exclusion on leaf litter amphibians and reptiles in a Neotropical wet forest, Costa Rica. Biological Conservation 163:90–98. Schlaepfer, M. A. 2003. Successful lizard eggs in a humandisturbed habitat. Oecologia 137:304–311. Scott, N. J. 1976. The abundance and diversity of the herpetofaunas of tropical forest litter. Biotropica 8:41–58. Sigel, B. J., T. W. Sherry, and B. E. Young. 2006. Avian community response to lowland tropical rainforest isolation: 40 years of change at La Selva Biological Station, Costa Rica. Conservation Biology 20:111–121. Socci, A. M., M. A. Schlaepfer, and T. A. Gavin. 2005. The importance of soil moisture and leaf cover in a female lizard’s (Norops polylepis) evaluation of potential oviposition sites. Herpetologica 61:233–240. Stuart, S. N., J. S. Chanson, N. A. Cox, B. E. Young, A. S. L. Rodrigues, D. L. Fischman, and R. W. Waller. 2004. Status and trends of amphibian declines and extinctions worldwide. Science 306:1783–1786. Whitfield, S. M., K. E. Bell, T. Philippi, M. Sasa, F. Bolanos, G. Chaves, J. M. Savage, and M. A. Donnelly. 2007. Amphibian and reptile declines over 35 years at La Selva,

461

Costa Rica. Proceedings of the National Academy of Sciences of the United States of America 104:8352–8356. Whitfield, S. M., M. A. Donnelly, E. Geerdes, and J. L. Kerby. 2012a. Ranavirus infection in native amphibians at La Selva Biological Station, Costa Rica: the first report of ranavirus in Central America. Herpetological Review 43:425–427. Whitfield, S. M., E. Geerdes, I. Chacon, E. Ballestero Rodriguez, R. R. Jimenez, M. A. Donnelly, and J. L. Kerby. 2013. Infection and co-infection by ranavirus and the amphibian chytrid fungus in wild Costa Rican frogs. Diseases of Aquatic Organisms 104:173–178. Whitfield, S. M., J. Kerby, L. R. Gentry, and M. A. Donnelly. 2012b. Temporal variation in infection prevalence by the amphibian chytrid fungus in three species of frogs at La Selva, Costa Rica. Biotropica 44:779–784. Whitfield, S. M., K. R. Lips, and M. A. Donnelly. In press. Decline and conservation of amphibians in Central America. In: Status of Conservation and Decline of Amphibians: Western Hemisphere. Vol. 8B. H. H. Heatwole, C. Barrio-Amoros, and J. W. Wilkenson (eds.). Surrey Beatty and Sons, Pty. Ltd., Sydney. Whitfield, S. M., and M. S. F. Pierce. 2005. Tree buttress microhabitat use by a neotropical leaf-litter herpetofauna. Journal of Herpetology 39:192–198. Wieder, R. K., and S. J. Wright. 1995. Tropical forest litter dynamics and dry season irrigation on Barro Colorado Island, Panama. Ecology 76:1971–1979. Young, B. E., K. R. Lips, J. K. Reaser, R. D. Iba´n ˜ ez, A. W. Salas, J. R. Ceden ˜ o, L. A. Coloma, S. Ron, E. La Marca, J. R. Meyer, A. Mun ˜ oz, F. Bolan ˜ os, G. Chaves, and D. Romo. 2001. Population declines and priorities for amphibian conservation in Latin America. Conservation Biology 15:1213–1223. Young, B. E., T. W. Sherry, B. J. Sigel, and S. Woltmann. 2008. Nesting success of Costa Rican lowland rain forest birds in response to edge and isolation effects. Biotropica 40:615–622.