Species diversity and habitat fragmentation: frogs in a tropical montane landscape in Mexico

Species diversity and habitat fragmentation: frogs in a tropical montane landscape in Mexico

Biological Conservation 117 (2004) 499–508 www.elsevier.com/locate/biocon Species diversity and habitat fragmentation: frogs in a tropical montane la...

317KB Sizes 3 Downloads 37 Views

Biological Conservation 117 (2004) 499–508 www.elsevier.com/locate/biocon

Species diversity and habitat fragmentation: frogs in a tropical montane landscape in Mexico Eduardo Pineda*, Gonzalo Halffter Departamento de Ecologı´a y Comportamiento Animal, Instituto de Ecologı´a, A.C., Apartado Postal 63, C.P. 91000, Xalapa, Veracruz, Mexico Received 17 June 2003; accepted 22 August 2003

Abstract We evaluate the alpha (within patch species richness), beta (spatial turnover among patches) and gamma (landscape) diversity of frogs in a tropical montane cloud forest (TMCF) in central Veracruz, Mexico in order to assess (1) the influence of forest fragmentation on frog assemblages, (2) the importance to diversity of the various elements of the landscape matrix, including the shaded coffee plantations and cattle pastures that surround TMCF and (3) to identify the frog guilds most affected by habitat transformation. We sampled ten sites between May 1998 and November 2000: five TMCF fragments and five anthropogenic habitats. For the entire landscape, we registered 21 species belonging to six families. 100% of these were found in the TMCF fragments and 62% in the surrounding mosaic of anthropogenic habitats. Gamma diversity (g) is determined to a greater extent by species exchange (b) than by local species richness (a). Elevational variation, the degree of conservation of the vegetation canopy and fragment size appear to determine the species diversity of this landscape. Large species, terrestrial species, those whose eggs develop outside water, and those whose larvae develop in the water seemed to be most affected by habitat transformation. On its own, even the largest and most species-rich cloud forest fragment is not capable of preserving the current anuran diversity. Neither are the shaded coffee plantations that are interspersed among and link the patches of TMCF. However together they form a diverse system of habitats crucial to species conservation in this landscape. # 2003 Elsevier Ltd. All rights reserved. Keywords: Alpha, beta and gamma diversity; Cloud forest; Shaded coffee plantation; Frog community; Mexico

1. Introduction Habitat fragmentation is recognized as one of the most important threats to species diversity, particularly in tropical zones where diversity is high and forests are being transformed at ever-increasing rates. Tropical montane cloud forest (TMCF) is recognized as one of the ecosystems most threatened by deforestation in the world, and in some regions, the deforestation rate for TMCF exceeds that of tropical rain forest (Hamilton et al., 1995). The way in which the fragmentation of tropical forest affects their species richness varies greatly depending on the scale at which one studies the phenomenon (Haila, 2002; Laurance et al., 2002; Arellano and Halffter, in press). Both the remaining fragments of * Corresponding author at present address: 6 Suffolk Walk, University of East Anglia, Norwich NR4 7TU, UK. E-mail address: [email protected] (E. Pineda). 0006-3207/$ - see front matter # 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocon.2003.08.009

forest and the group of plant communities generated by human activities which surround them participate in the species dynamics of the landscape (Gustafson and Gardner, 1996; Gascon et al., 1999). To understand the heterogeneity of the modified landscape it is important to realize that, from the perspective of the organisms that inhabit it, the landscape is not environmentally uniform (Haila, 2002), particularly in tropical montane ecosystems. In Mexico, TMCFs are discontinuously distributed over less than 1% of the territory at mid-elevations of mountain ranges (800–2000 m asl). Inhabited by 10% of Mexico’s flora (Rzedowski, 1998) and 12% of its terrestrial vertebrates (Flores-Villela and Gerez, 1994). TMCFs are recognized as the most species-diverse ecosystems per unit area. This high species diversity is related to the remarkable environmental heterogeneity and complex biogeographical history, where tropical, temperate and numerous endemic elements are present

500

E. Pineda, G. Halffter / Biological Conservation 117 (2004) 499–508

(Rzedowski, 1998; Challenger, 1998; Campbell, 1999a). However, in Mexico TMCF has decreased mainly as a result of its transformation into cattle pastures, coffee and fruit plantations, and human settlements (Challenger, 1998; Williams-Linera et al., 2002). The habitat fragmentation caused by human activity has acted on a heterogeneous and species-rich ecosystem, the natural distribution of which is both restricted and discontinuous. The effects of landscape fragmentation on species diversity, i.e. the alpha (within community), beta (among communities) and gamma (landscape) diversities, can be assessed through the use of an indicator group (Halffter, 1998). The species richness of the entire fragmented landscape, gamma diversity, is the result of both alpha and beta diversities, which reflect the heterogeneity of the landscape (pre- or post-fragmentation). As such, gamma diversity is a function of the sensitivity of species to differences or changes in the landscape, so in modified landscapes where the original landscape (pre-fragmentation) was homogeneous, we expect the total number of species present to be similar to the number of species in the richest patch (whether it is a forest remnant or some other mosaic habitat). In this scenario the species composition in the remaining patches would be subsets of the richest one and consequently, gamma diversity would be more greatly influenced by alpha diversity. On the other hand, species richness in heterogeneous landscapes will be notably higher than that of the richest patch, and the number of species for the landscape would be a result of the distinctness or dissimilarity in the composition of the species assemblages of the different patches of vegetation that make up the landscape. Thus, high values of beta diversity are expected. This way of partitioning the elements of diversity is central to understanding the contribution of local and regional processes to species diversity (Schluter and Ricklefs, 1993; Lande, 1996). Furthermore, when both the forest fragments and the habitats that have been transformed by human activities are considered, appropriate conservation and management guidelines can be established. Amphibians have been proposed as good bioindicators of habitat quality (Wyman, 1990; Wake, 1991; Blaustein et al., 1994) because (1) most have a biphasic life cycle, and require different habitats and food as larvae and adults; (2) they exhibit local diversity in their reproductive modes; (3) their highly permeable skin makes them vulnerable to chemical and physical changes in both terrestrial and aquatic habitats; and (4) they exhibit low vagility and strong philopatry (Crump, 1974; Wyman, 1990; Wake, 1991; Blaustein et al., 1994; Lips, 1998). It has been suggested that habitat transformation and fragmentation are the most obvious causes of the recorded declines in the global amphibian population (Alford and Richards, 1999; Heyer, 2000; Blaustein and

Kiesecker, 2002), however there are few studies of the relationship between landscape fragmentation and amphibian diversity at tropical latitudes to substantiate this (Zimmerman and Bierregard, 1986; Tocher et al., 1997; Gascon et al., 1999; Vallan, 2000). We partition diversity into its components (alpha, beta and gamma) to evaluate the influence of the fragmentation of cloud forest on frog diversity and to identify the determinants of species diversity on the local and mesoscale of a tropical montane landscape. We also evaluate the spatial distribution patterns of frog assemblages, as well as the roles of the forest remnants and the mosaic of anthropogenic habitats in the maintenance of frog diversity. Finally, we identify the frog guilds most affected by habitat transformation. This study aims to contribute to a better understanding of the response of frog assemblages to habitat transformation in highly diverse and spatially heterogeneous systems, thereby helping to determine the appropriate strategies for the conservation of both the remnants of the original forest and the habitats created by humans and the species that inhabit them.

2. Methods 2.1. Area and study sites The study was carried out in the surroundings of Xalapa, Veracruz, Mexico (19 340 N lat., 96 560 W long.), between 1120 and 1590 m elevation in the upper region of the La Antigua River Basin. The mean annual temperatures in this region range from 17  C at higher elevations to 20  C at lower elevations, and total annual precipitation ranges from 1500 to 1900 mm (Soto and Go´mez, 1990). The main edaphic component is of volcanic origin and landforms are diverse and complex. Elongated hills along with their valleys and some plateaus are common (Geissert and Campos, 1993). One hundred and fifty years ago, prior to the boom in coffee production began during the 19th century, the dominant land cover type was TMCF, also known as bosque meso´filo de montan˜a, but its transformation to agricultural and urban systems has had an increasingly important impact since the end of the 19th century to the middle of the 20th century (Challenger, 1998; Williams-Linera, 2002). At present, remnants of undisturbed TMCF cover only 10% of the original area (Williams-Linera et al., 2002) and these are immersed in a mosaic of cattle pastures, shaded coffee plantations (the two most extensive systems in the area), sugar cane and citrus plantations, secondary vegetation, and human settlements. The study sites were selected based on an analysis of digital aerial images (INEGI, 1995) and we hiked through the area for confirmation of site suitability. The

E. Pineda, G. Halffter / Biological Conservation 117 (2004) 499–508

criteria for selection were that: (1) each site should represent just one type of land use, (2) all sites must have at least one stream to ensure that the absence of water was not a limiting factor for the anuran fauna that require water to complete their life cycle, and (3) the distance between sites should be greater than 500 m to ensure the independence of samples in the measurement of local diversity. The ten sites selected (Fig. 1, Table 1) include five fragments of TMCF and five patches of habitats of human origin (three shaded coffee plantations and two cattle pastures). In TMCF fragments there is a high diversity of tree species, mainly Quercus, Liquidambar, Ulmus, Clethra, Carpinus, Oreopanax and Platanus, in addition to an abundance of epiphytes and lianas. Coffee plantations are the traditional shaded, polyspecific type (Moguel

501

and Toledo, 1999), where cultivated species of the genera Inga, Citrus, Musa, Psidium, Persea and such native elements as Ulmus, Heliocarpus, Oreopanax and Platanus provide shade for the coffee plants. Epiphytes are also abundant in this habitat type. There are few trees in the pastures; Liquidambar, Quercus and Acacia are used to provide the cattle with shade. 2.2. Frog sampling Anurans were collected from all possible microhabitats of the ten sites, using a time-constrained technique (Scott, 1994) from May to November during 1998, 1999 and 2000. Two people searched each site at dusk, night and dawn. The first individuals captured for each species were preserved in 70% alcohol as voucher specimens. The rest of the individuals caught were returned to the place of capture after identification. To ensure that a minimum of 80% of the estimated total number of frog species were captured at each site and for the entire landscape (thus validating comparisons) we used an asymptotic species accumulation function (Sobero´n and Llorente, 1993) and three non-parametric estimators: ICE, Chao2 and Bootstrap (Colwell, 1997). Sampling effort was measured in person h (one search=1.5 h2 people=3 person h). Search time at each site ranged from 33 to 39 person h and was 345 person h for the entire landscape. 2.3. Data analysis

Fig. 1. The locations of the ten study sites in a mountainous landscape in central Veracruz, Mexico. The sites are described in Table 1. Table 1 Characteristics of the ten sites studied in a fragmented landscape in central Veracruz, Mexico Site

Habitat type

Size (ha)

Shape

Canopy cover S.D (%)

Altitude (m asl)

1 2 3 4 5 6 7 8 9 10

TMCF TMCF TMCF TMCF TMCF Shaded coffee Shaded coffee Shaded coffee Pasture Pasture

72 62 55 17 16 122 104 41 40 11

2.10 3.20 2.21 3.38 3.79 1.46 1.39 1.34 1.73 2.97

922.4 892.8 843.7 882.6 833.2 607.2 527.1 4811.1 73.8 52.6

1510 1550 1570 1560 1180 1180 1250 1120 1420 1530

TMCF: Tropical montane cloud forest.

For our purposes gamma diversity is defined as the total number of species found throughout the landscape. Alpha, or local, diversity describes the species richness or the number of species that occupy a clearly delimited site where, for our purposes, the latter is the minimum unit in terms of space and time that contains an assemblage or community. Beta diversity can be defined in two different ways. (1) To determine the diversity among sites we use the mean number of species not found in each of the sites studied (Lande, 1996), and calculated as  ¼    where  is the number of species in the entire landscape and  is the mean species richness of the sites studied. This provides a measurement based on the number of species which is therefore comparable with alpha diversity. This way of measuring beta diversity is also used in the additive partitioning (Veech et al., 2002) of species diversity. (2) To determine the diversity between sites, i.e. the distinctness of species composition between sites, also known as complementarity (e.g. Colwell and Coddington, 1995) which can be expressed as



Sj þ Sk  2Vjk  100 Sj þ Sk  Vjk

502

E. Pineda, G. Halffter / Biological Conservation 117 (2004) 499–508

where Sj and Sk are the number of species recorded for sites j and k respectively, and Vjk is the number of species common to both sites. Complementarity (C) varies from 0 (where the species lists for two sites are identical) to 100% (where the lists are totally different). To compare species richness among sites, we used fitted species accumulation curves generated with the program EstimateS 5.01 (Colwell, 1997; randomised 500 times). We chose four variables to describe site attributes: area, shape, canopy cover and altitude; and five traits to describe landscape heterogeneity: distance between sites, differences in area, shape, canopy cover and altitude between sites, in order to detect any association with alpha and beta diversity, respectively. The computer program ArcView 3.2 was used to analyse digital aerial photographs (INEGI, 1995) and to obtain area, shape and distance between sites. p Shape ffiffiffiffiffiffiffi (Sh) (Patton, 1975) was calculated as Sh ¼ P=2 A where P is the perimeter of the site and A is the area. The minimum value for shape is 1 when the site is a perfect circle, and this value increases as the perimeter increases in relation to the area. Canopy cover was measured using a Crown spherical densiometer at 20 points, each separated by 20 m, at each site. A Thommen altimeter was used to obtain site altitude. Table 1 shows the attributes of the sites. A multiple regression model (backward stepwise) was used to detect those site variables related to alpha () diversity, and to generate a predictive model of species richness for sites within the landscape. Data were analysed using the statistical package STATISTICA 5.1 (StatSoft, 1997). To determine whether the traits of landscape heterogeneity were related to beta diversity between sites, we used the Mantel test (Miller, 1997) which estimates the association between two independent dissimilarity matrices that describe the same entities set, and tests whether the association is stronger than one would expect by chance (Sokal and Rohlf, 1995). The association between beta diversity and the five dissimilarity matrices (distance between sites and differences in area, shape, canopy cover and altitude) was tested for all possible pairs of the ten sites. There were 500 permutations in each analysis. To examine whether the composition of anuran assemblages at the sites of the fragmented landscape fit a non-random pattern, we used an analysis of nestedness patterns. Nestedness refers to a non-random distribution of species where species assemblages of sites with lower richness are subsets of the biota at the richer sites (Wright and Reeves, 1992). If extinction or colonization of species is similar or predictable in each site of a landscape, then assemblages in the smaller or more transformed sites are expected to be subsets of the assemblages found in bigger or better-conserved sites. However, if local extinction or colonization is totally random and independent of both site and species identity, then a nested distribution pattern will not occur.

Of the several measurements that exist to evaluate nestedness (Wright et al., 1998) we used the T metric (Atmar and Patterson, 1993) calculated with the Nestedness Temperature Calculator (Atmar and Patterson, 1995) which measures the deviation of an observed presence–absence matrix from a perfectly nested one of equal size. Values for T range from 0 for assemblages that are perfectly nested to 100 for assemblages that are completely randomly ordered. T was calculated for all study sites, as well as separately for TMCF remnants and for the surrounding mosaic of habitats. The statistical significance for the T value of the observed matrix was assessed by Monte Carlo simulations (500). In order to detect how the frog guilds are affected by forest transformation, all species were characterised with respect to three attributes: size, habit and reproductive mode, based on field observations and bibliographic information (Duellman, 1970; Lee, 1996; Mendelson, 1997; Vogt et al., 1997; Campbell, 1999a, b; Campbell and Savage, 2000). Size was classified according to mean SVL (snout-vent length): (1) small (< 31 mm), (2) medium (31–50 mm) and (3) large ( > 50 mm). The classes of habit attribute were: (1) arboreal, (2) terrestrial and (3) fossorial. Reproductive mode was classified by both oviposition site and the site where the larva develops, the two criteria suggested by Crump (1974), resulting in three modes: (1) eggs and larvae in water, (2) eggs out of water and larvae in water, and (3) eggs and larvae out of water. The analysis of size, habit and reproductive mode was based on observations of their frequency of occurrence, and then tested using Spearman’s rank correlation test (Zar, 1984) using those site attributes for which there was a significant relationship with alpha diversity.

3. Results 3.1. Alpha and gamma diversity The gamma diversity of the landscape was 21 anuran species belonging to six families (Table 2). Seventy-six percent (16 species) belong to the Hylidae and Leptodactylidae families. All species were recorded for TMCF and 62% of the total were found in the surrounding mosaic of habitats (13 species in shaded coffee plantations and four species in cattle pastures). Eight species–Bufo cristatus, Hyla arborescandens, H. mixomaculata, H. picta, H. taeniopus, Eleutherodactylus berkenbuschi, E. spatulatus and Leptodactylus labialis–were found exclusively in TMCF, while none was exclusive to the surrounding vegetation. Three species—Eleutherodactylus pymaeus, Hyla miotympanum and Rana berlandieri—were found in all habitat types and the last two were present in almost all ten sites.

503

E. Pineda, G. Halffter / Biological Conservation 117 (2004) 499–508

Table 2 Matrix representing observation frequencies of frog species registered and their natural history traits as recorded at ten sites of a fragmented landscape in central Veracruz, Mexico Species

Site 1

Bufonidae Bufo cristatus Wiegmann, 1833 Bufo valliceps Wiegmann, 1833 Centrolenidae Hyalinobatrachium fleischmanni (Boettger, 1893) Hylidae Hyla arborescandens Taylor, 1939 Hyla mixomaculata Taylor, 1950 Hyla miotympanum Cope, 1863 Hyla picta (Gu¨nther, 1901) Hyla taeniopus Gu¨nther, 1901 Smilisca baudini (Dume´ril y Bibron, 1841) Scinax staufferi (Cope, 1865) Leptodactylidae Eleutherodactylus alfredi (Boulenger, 1898) Eleutherodactylus berkenbuschi Peters, 1870 Eleutherodactylus cystignathoides (Cope, 1877) Eleutherodactylus decoratus Taylor, 1942 Eleutherodactylus mexicanus (Brocchi, 1877) Eleutherodactylus pygmaeus Taylor, 1937 Eleutherodactylus rhodopis (Cope, 1867) Eleutherodactylus spatulatus Smith, 1939 Leptodactylus labialis (Cope, 1877) Microhylidae Gastrophryne usta (Cope, 1866) Ranidae Rana berlandieri Baird, 1854 Species number

3

Size 2

5

6

2

3 9 2 4

3

2 8

4

3 3 6

4

6

5

8 2

2

3

8

6 3 4 8 1 3

6 13

3 4 5 4 2 2

7 11

8

10

3

4

L L

T T

1 1

5

S

A

2

L M M S L L S

A A A A A A A

2 1 1 2 2 1 1

M L S S M S M S M

T T T T T T T T T

3 3 3 3 3 3 3 3 2

S

F

1

L

T

1

3

4

6

6

4 2 2

5

2 5 3

3 5 1

1 1 4 5

2

2 10

4

5

4

3 2 5 7

4 6

5 3

3

2

2 4 1

1

7 10

6 10

6 8

RM

7

2 2

1

4

9

Habit

4 8

4 7

3

4 3

The sites are listed in Table 1. Size (snout-vent length): S, small species ( <31 mm); M, medium species (31–50 mm); L, large species ( >50 mm). Habit: T, terrestrial species; A, arboreal species; F, fossorial species. RM=Reproductive mode: 1, eggs and larvae in water; 2, eggs outside water and larvae in water; 3, eggs and larvae outside water.

The Clench Model and non-parametric methods for estimating species richness indicated that the completeness of inventory for each site was above 80%, while for the entire landscape, it was above 94% (Table 3). With the sampling effort applied at each site, the species accumulation curves reached an asymptotic phase (Fig. 2a). For the entire landscape, the species accumulation curve reached a plateau with a sampling effort of 210 person h, and at around 300 person h the unique and duplicate records practically vanished (Fig. 2b). Frog alpha diversity averaged 8.3 species and differed notably among sites, ranging from 13 species in a fragment of TMCF (62% of the species found over the entire landscape) to only three species (14% of the total for the landscape) in the two poorest sites, both of which were pastures. Two of the anthropogenic habitat sites (6 and 7, both shaded coffee plantations) were richer than or equal to several TMCF fragments (sites 3, 4 and 5) in accumulated species (Fig. 2). Multiple regression showed that alpha diversity increases logarithmically with increasing canopy cover (as the most important variable) and with the area of

the site (F2, 7=89.38, P < 0.001, R2=0.95). Hence, the resulting model is Log10 =0.41252+0.00604 C +0.00178 A, where  is species number, C is canopy cover and A is area of site. Finally, the gamma diversity of the entire landscape can be expressed in an additive way as: 21 species ()=8.3 ()+12.7 (). 3.2. Complementarity and nested subsets Mean complementarity between sites was 67% (range: 30–92%; Table 4), i.e. on average, two thirds of the anuran species in a given pair of sites are present exclusively in one site or another. The Mantel test showed that the difference in the species composition of assemblages between sites increases with (1) increasing difference in altitude (ZMantel=7172, P=0.002, r=0.60) and (2) increasing difference in canopy cover (ZMantel=1259, P=0.003, r=0.56). The analysis of nested subset patterns indicated that anuran assemblages in the fragmented landscape exhibited

504

E. Pineda, G. Halffter / Biological Conservation 117 (2004) 499–508

Table 3 Observed and estimated species richness for ten study sites and the entire landscape Site

Number Number of species estimated Completeness of species observed ICE Chao2 Bootstrap Clench

1 2 3 4 5 6 7 8 9 10 Entire Landscape

13 11 10 8 10 10 8 7 3 3 21

13.4 11.4 10.1 8.0 10.0 10.9 8.4 7.0 3.0 3.0 21.0

13.5 11.3 10.0 8.0 10.0 11.0 8.5 7.0 3.0 3.0 21.0

13.6 11.7 10.0 8.3 10.5 10.9 8.5 7.1 3.1 3.0 21.2

15.5 13.8 12.4 9.6 12.5 12.5 10.0 8.6 3.7 3.6 22.4

84 97 80 98 81 100 83 100 80 100 80 92 80 95 81 100 81 100 83 100 94 100

Completeness=percent of estimated richness (minimum–maximum). The sites are listed in Table 1.

a nested distribution throughout all the study sites (T=36.12 , P=0.014), and among the sites of the anthropogenic habitats (T=16.75 , p=0.021), but not among the TMCF fragments (T=45.80 , P=0.69). 3.3. Frog guilds Canopy cover and site area, the two site variables that approached statistical significance when checked for possible correlation with alpha diversity, were related in a different way to the variation in the frog guild at each site. With the decreasing canopy cover the proportion of large species also decreased (Spearman: rs=0.778, P=0.008, n=10), ranging from 47% at site 4, to 0% at site 9. The percentages of those species with eggs that develop out of water and larvae that develop in water also diminished (Spearman: rs=0.634, P=0.049, n=10) from 24% at site 4, to 0% at sites 7, 9 and 10. The variation in the relative proportions of the remaining frog guilds was independent of canopy cover. A reduction in fragment or patch area is associated with a decrease in the proportion of terrestrial species (Spearman: rs=0.782, P=0.007, n=10) from 86 to 45% (sites 7 and 9, respectively); and an increase in the proportion of arboreal species (Spearman: rs=0.782, P=0.007, n=10) from 14% to 50% at the same sites. Because there was only one fossorial species (Gastrophryne usta) we decided to exclude it from the analysis. Finally, there was a negative correlation between the area of the site and the percentage of species with aquatic eggs and larvae (Spearman: rs=0.784, P=0.007, n=10), ranging from 32 to 77% (sites 1 and 10, respectively). The variation in the relative proportions of species of the remaining frog guilds was independent of the area of the sites.

Fig. 2. Species accumulation curves for frogs at the ten study sites and for the entire mountainous landscape in central Veracruz, Mexico. (A) Comparisons for all sites were made at the asymptotic phase. Black marks indicate TMCF fragments: =site 1, &=site 2, ~=site 3, - -*- -=site 4, ^=site 5, and white marks indicate anthropogenic habitats: &=site 6, ~=site 7, ^=site 8, &=site 9, - -*- -=site 10. (B) Observed species= , non-parametric richness estimators (ICE=^, Chao 2=* and Bootstrap=~), unique=& and duplicate=& species (rare species) for the entire landscape. Note the similarity in the number of observed species and species richness estimates starting at 210 person h sampling effort.

Table 4 Matrix representing the complementarity values (%) between sites Site

2

3

4

5

6

7

8

9

10

1 2 3 4 5 6 7 8 9 10

40

47 50

38 42 62

79 76 82 80

79 76 57 88 46

69 64 71 86 62 50

75 71 79 85 58 30 33

77 83 70 78 92 82 78 75

77 83 70 63 82 82 78 75 50

Average complementarity is 67%. The sites are described in Table 1.

E. Pineda, G. Halffter / Biological Conservation 117 (2004) 499–508

4. Discussion Our results show that beta diversity has a greater influence on gamma diversity than does alpha diversity. Frog species seem to respond differently to habitat fragmentation, hence the notable variation in the species number and composition of the assemblages in the different components of the landscape. Differences in the degree of habitat transformation and in the altitudinal zone where the TMCF was originally distributed seem to be relevant determinants of frog diversity in the landscape studied. 4.1. Alpha, beta and gamma diversity With estimated inventory completeness over 80%, and an asymptotic species accumulation curve for each site, the comparison of species richness among inventories is valid (Colwell and Coddington, 1995). The disappearance of unique and duplicate frog records throughout this landscape reduced the risk of including tourist and transitory species in the analysis, which is desirable as these do not regularly inhabit in the zone and are not part of the species pool (Gaston, 1996). Differences in species richness at the sites implies that frogs respond strongly and differentially to the degree of habitat transformation, the latter represented by decreasing canopy cover and patch area. As such, the size of species assemblages varies across the components of this fragmented landscape. Interestingly, two of the shaded coffee plantations were equally rich or richer in frog species than three of the TMCF fragments. These two plantations had 23–36% less canopy cover than the TMCF fragments, but were also notably larger (2–7 times). It seems that large, anthropogenic habitats with structural and environmental characteristics similar to those of coffee plantations with polyspecific shade can support the same number of frog species as a smaller forest fragment, or even more. This may be an indication that while diminishing of canopy cover is associated with a reduction in structural complexity, larger areas offer more microhabitats. Alpha diversity does not appear to be an important determinant of gamma diversity because even the species number recorded in the richest forest fragment (just 62%) does not resemble the richness of the entire landscape. A similar trend was observed for dung beetles (59% in the richest forest remnant) in this landscape (Arellano and Halffter, in press). However, this value is lower than those reported for frogs in other fragmented tropical forests. In Manaus, Brazil, the richest fragment was occupied by 86% of the total number of species found in all the fragments together (Tocher et al., 1997). While in a rainforest in Madagascar, Vallan (2000) reported that 81% of the frog species recorded for the entire landscape were

505

found in the richest fragment. In these studies, in contrast to our findings, gamma diversity is clearly influenced to a greater extent by alpha diversity. Complementarity between sites was high, averaging 67%. The wide variation in species composition among sites is associated with the altitudinal zone over which the cloud forest occurs, and with the different degrees of habitat transformation, especially with respect to the decrease in canopy cover. Altitude and the differences in canopy cover among sites, appear to be the two most important environmental factors in determining beta diversity at this landscape, and consequently in determining gamma diversity. Substantial changes in plethondontid salamander assemblages have also been observed along the same altitudinal zone; in fact, for these organisms there is a total species turnover between the highest (1800 m asl) and lowest (1000 m asl) elevations in this landscape (Wake et al., 1992). Differences in canopy cover between sites imply differences in their capacity to support several frog guilds (large species, small species and species with reproductive mode 2, as discussed below). The composition of the frog species assemblages tends to diverge as canopy cover becomes more dissimilar between sites. Thus, canopy cover affects not just the species number at each site, but also appears to influence the species composition of assemblages. Complementarity values for other biological groups have been reported for the same landscape. The average complementarity for trees and shrubs between forest fragments was 84% and 80%, respectively (Williams-Linera, 2002). For dung beetles the average complementarity between forest, shaded coffee plantations and pastures was 41% (Arellano and Halffter, in press). Frog species assemblages display a nested distribution among all study sites, as well as among the anthropogenic habitat sites, but not among TMCF fragments. This suggests that prior to the fragmentation of the landscape the species pool had a heterogeneous distribution, possibly owing to the high environmental diversity and the complex biogeographical history of this type of vegetation in Mesoamerica (Rzedowski, 1998; Challenger, 1998; Campbell, 1999a), in conjunction with the patchy distribution commonly displayed by amphibians (Zimmerman and Bierregaard, 1986). Subsequently, habitat fragmentation has influenced local extinction and colonization processes which have in turn resulted in the nested arrangement of diversity among sites. Therefore, a change from random to nonrandom distribution of frog species in the landscape studied may have been influenced by habitat fragmentation. Nestedness is common in nature (Wright et al., 1998), however the level of nestedness we report among the forest fragments (T=45.8 , P=0.69) is notably lower than that reported in other studies of frog diversity, and was not significant. In Amazonia, Brazil, the nestedness value of T among rainforest fragments was

506

E. Pineda, G. Halffter / Biological Conservation 117 (2004) 499–508

2.7 (Zimmerman and Bierregaard, 1986). In the midwestern United States, Kolozsvary and Swihart (1999) reported a T of 7.29 . In Madagascar, Vallan (2000) reported a T of 22.07 ; all statistically significant, in contrast to the value we report. We believe one reason for this is the distribution of the assemblages throughout this landscape prior to its fragmentation by human activities. The complex topography of the area suggests that the distribution of species assemblages was most likely heterogeneous. It is worth noting that the landscape we are studying is located within a transition interface between a tropical and a temperate zone, adding another dimension of variation to the microclimate and environmental heterogeneity associated with the altitudinal range.

disturbed habitats. At the same landscape a similar result is reported for dung beetles by Arellano and Halffter (in press). Thus, although the anthropogenic habitats do not provide species complementarity to the diversity of the TMCF studied, they (especially the shaded coffee plantations) clearly have an important influence on the connectivity between fragments and on landscape dynamics. Similar findings have been reported for frogs in a fragmented tropical forest in Amazonia, Brazil (Gascon et al., 1999) and for other vertebrate groups (Laurance et al., 2002), although the mosaic of habitats that they studied complemented the diversity of the whole landscape matrix by 8–25%, depending on the taxa. These results confirm that the mosaic of vegetation surrounding forest fragments is anything but a biological void.

4.2. Forest fragments and anthropogenic habitats 4.3. Frog guilds and habitat transformation TMCF fragments are of notable importance for eight frog species. Five of these—Bufo cristatus, Hyla arborescandens, H. mixomaculata, H. taeniopus and Eleutherodactylus berkenbuschi—have only been recorded in well-preserved TMCF in different locations (Duellman, 1970; Mendelson, 1997; Campbell and Savage, 2000), suggesting that these species display a high degree of sensitivity to the transformation of cloud forest. The other three species, Hyla picta, Eleutherodactylus spatulatus and Leptodactylus labialis, have also been recorded in different types of vegetation and in disturbed habitats, but their altitudinal limits coincide with TMCF’s altitudinal zone (Lee, 1996; Vogt et al., 1997; Campbell, 1999a). Thus, the eight species mentioned above would appear to be vulnerable to local extinction if TMCF continues to be reduced. These species need, therefore, to be carefully monitored in conservation efforts. Collectively, the TMCF fragments are suitable for supporting the 21 frog species recorded at the landscape, and 62% of these species seem to be supported in the anthropogenic habitats as well. However, the capacity of anthropogenic habitats to support frog fauna varied notably between the shaded coffee plantations and the cattle pastures. The relatively high number of frog species registered in the coffee habitat (13 species collectively, the 62% mentioned above) suggests that this environment is indeed attractive to some species and, in addition to meeting their habitat requirements, allows them to move freely across the landscape, owing perhaps to its structural and environmental similarities to the original forest (Perfecto et al., 1996; Moguel and Toledo, 1999). On the other hand, the very low number of frog species in the cattle pastures and the total absence of species exclusively associated with highly sunny environments, suggests that in the original landscape (prior to fragmentation by humans) there was no frog fauna that was exclusive to gaps or to intensively

With diminishing canopy cover the proportions of large species and species with reproductive mode 2 (eggs outside water, larvae in water) decreased, and with the reduction of patch area, the proportion of terrestrial species also decreased. This suggests that the loss of canopy cover and reduction in area result in the loss of the conditions required to support these species guilds. The reasons for such tendencies could be physiological and ecological. Higher temperatures, lower soil and atmospheric humidity, as well as increasing wind velocity are some of the consequences of canopy removal (Saunders et al., 1991; Murcia, 1995). Amphibians need to keep their skin wet to allow gaseous exchange and depend on external heat to regulate their internal temperature (Duellman and Trueb, 1994). Larger species have a greater surface area for exchange than the smaller ones and would tend to lose greater quantities of water as relative humidity decreases as a result of canopy destruction. The eggs of species with reproductive mode 2 are exposed to the atmosphere and with reduced humidity would be vulnerable to desiccation. Furthermore, the simplification of the vegetation structure could reduce the availability of oviposition sites (in the vegetation above streams) for most of the species registered that exhibit reproductive mode 2. A reduction in patch area is related to decreasing environmental heterogeneity at ground level, resulting in the loss of microhabitats, breeding sites and territory for several species (Zimmerman and Bierregaard, 1986). As such, the species living at ground level would be the most affected. In contrast, with the reduction of patch area the proportions of arboreal species and species with reproductive mode 1 increase. For arboreal species, one reason may be that in spite of diminishing patch area, where the canopy is high there is a high degree of vertical heterogeneity. This may buffer adverse microclimatic changes and limit the disappearance of microhabitats

E. Pineda, G. Halffter / Biological Conservation 117 (2004) 499–508

used by tree-dwelling species. For species with reproductive mode 1 (eggs and larvae develop in water), smaller patch area is generally related to a decrease in the availability of the microhabitats (ponds and streams) required for oviposition. Therefore, a low proportion of species with reproductive mode 1 is expected to occupy smaller patches. However, in this study two species, Hyla miotympanum and Rana berlandieri, were observed breeding in streams or in ponds at every study site, independent of canopy cover or patch size. This suggests that for these species the presence of a body water had a stronger positive effect than the negative effects of canopy removal or reduction in patch size. Crowding may even occur as a result of the scarcity of bodies of water in small patches. Both species seem to survive and reproduce in several habitats independent of disturbance levels because the key factor is the presence of any body of water.

507

the manuscript. Finally, we gratefully acknowledge Clara Sampieri, Julia´n Bueno and Claudia Moreno who generously provided assistence with fieldwork. Financial support for this research was received from the Mexican Consejo Nacional de Ciencia y Tecnologı´a (CONACyT, Project 37514-V), the Comisio´n Nacional para el Uso y Conocimiento de la Biodiversidad (CONABIO Project 093-01), ORCYT-UNESCO (Project 883.612.2) and IDEA WILD. Eduardo Pineda gratefully acknowledges the award of a graduate scholarship from CONACyT. This paper has been written in partial fulfilment of the requirements for Eduardo Pineda as a candidate for a doctoral degree of the Graduate Studies Division at the Instituto de Ecologı´a, A.C., Xalapa, Veracruz, Mexico.

References 5. Conclusions Human activity does not generate only two types of extreme scenarios: well-conserved and highly-impoverished habitats. Rather, the anthropogenic habitats located throughout the landscape represent a gradient of transformation, and these habitats have different effects on the species dynamics and biodiversity. Habitat fragmentation does not act directly on the species assemblages. Changes in environmental conditions act on each species (specifically on each population) in an independent manner, so the change in the number of species of each assemblage is the cumulative result of a series of specific events. From a conservation perspective, the largest and richest cloud forest fragment is not sufficient to preserve the anuran fauna of this landscape. The maintenance of a set of forest fragments extensively distributed (altitudinally and horizontally) in proximity to shaded coffee plantations, appears to be an appropriate strategy for maintaining frog diversity (and the diversity of other flora and fauna) of the landscape.

Acknowledgements We are grateful to Juan Francisco Ornelas, Guadalupe Williams-Linera, Gustavo Casas-Andreu, Federico Escobar, and Claudia Moreno for their comments and suggestions on earlier versions of this manuscript. We are also thankful to Diana Pe´rez-Staples for partially translating an earlier version of this manuscript. We thank Bianca Delfosse who carefully reviewed and corrected the English text of the manuscript and for offering valuable suggestions, and Jonathan Campbell and one anonymous reviewer for comments that improved

Alford, R.A., Richards, S.J., 1999. Global amphibian declines: a problem in applied ecology. Annual Review of Ecology and Systematics 30, 133–165. Arellano, L., Halffter, G., 2003. Gamma diversity: derived from and a determinant of alpha diversity and beta diversity. An analysis of three tropical landscapes. Acta Zoolo´gica Mexicana 90, 27–76. Atmar, W., Patterson, B.D., 1993. The measure of order and disorder in the distribution of species in fragmented habitat. Oecologia 96, 373–382. Atmar, W., Patterson, B.D., 1995. The Nestedness Temperature Calculator: A Visual Basic Program Including 294 Presence Absence Matrices. AICS Research Inc. University Park, NM and The Field Museum, Chicago, IL. Blaustein, A.R., Wake, D.B., Sousa, W.P., 1994. Amphibians declines: Judging stability, persistence, and susceptibility of populations to local and global extinctions. Conservation Biology 8, 60–71. Blaustein, A.R., Kiesecker, J.M., 2002. Complexity in conservation: Lessons from global decline of amphibian populations. Ecology Letters 5, 597–608. Campbell, J.A., 1999a. Distribution patterns of amphibians in Middle America. In: Duellman, W.E. (Ed.), Patterns of Distribution of Amphibians: A Global Perspective. The Johns Hopkins University Press, Baltimore, pp. 111–210. Campbell, J.A., 1999b. Amphibians and reptiles of Northern Guatemala, the Yucatan and Belize. University of Oklahoma Press, Norman. Campbell, J.A., Savage, J.M., 2000. Taxonomic reconsideration of Middle American frogs of the Eleutherodactylus group (Anura: Leptodactylidae): a reconnaissance of subtle nuances among frogs. Herpetological Monographs 14, 186–292. Challenger, A., 1998. Utilizacio´n y conservacio´n de los ecosistemas terrestres de Me´xico. Pasado, presente y futuro. Comisio´n Nacional para el Conocimiento y Uso de la Biodiversidad, Universidad Nacional Auto´noma de Me´xico, Agrupacio´n Sierra Madre, SC, Me´xico, DF. Colwell, R.K., 1997. EstimateS: Statistical Estimation of Species Richness and Shared Species from Samples. Version 5.01. Available: http://viceroy.eeb.uconn.edu/EstimateS. Colwell, R.K., Coddington, J.A., 1995. Estimating terrestrial biodiversity through extrapolation. In: Hawksworth, D.L. (Ed.), Biodiversity Measurement and Estimation. Chapman & Hall, New York, pp. 101–118.

508

E. Pineda, G. Halffter / Biological Conservation 117 (2004) 499–508

Crump, M.L., 1974. Reproductive strategies in a tropical anuran community. Miscellaneous Publications of the Museum of Natural History of the University of Kansas 61, 1–68. Duellman, W.E., 1970. The Hylid Frogs of Middle America. Monographs of the Museum of Natural History of the University of Kansas, Kansas City. Duellman, W.E., Trueb, L., 1994. Biology of Amphibians. The Johns Hopkins University Press, Baltimore. Flores-Villela, O., Gerez, P., 1994. Biodiversidad y conservacio´n en Me´xico: Vertebrados, vegetacio´n y uso de suelo. Comisio´n Nacional para el Uso y Conservacio´n de la Biodiversidad, Universidad Nacional Auto´noma de Me´xico, Me´xico, DF. Gascon, C., Lovejoy, T.E., Bierregaard, R.O., Malcolm, J.R., Stouffer, P.C., Vasconcelos, H.L., Laurence, W.F., Zimmerman, B., Tocher, M., Borges, S., 1999. Matrix habitat and species richness in tropical forest remnants. Biological Conservation 91, 223–229. Gaston, K.J., 1996. Species richness: measure and measurement. In: Gaston, K.J. (Ed.), Biodiversity: A Biology of Numbers and Difference. Blackwell Science, London, pp. 77–113. Geissert, D., Campos, A., 1993. Los paisajes morfoedafolo´gicos del a´rea de influencia de la ciudad de Xalapa. In: Lo´pez-Moreno, I.R. (Ed.), Ecologı´a urbana aplicada a la ciudad de Xalapa. Instituto de Ecologı´a, AC, Xalapa, pp. 65–79. Gustafson, R.J., Gardner, R.H., 1996. The effect of landscape heterogeneity on the probability of patch colonization. Ecology 77, 94– 107. Haila, Y., 2002. A conceptual genealogy of fragmentation research: From Island Biogeography to Landscape Ecology. Ecological Applications 12, 321–334. Halffter, G., 1998. A strategy for measuring landscape biodiversity. Biology International 36, 3–17. Hamilton, L.S., Juvik, J.O., Scatena, F.N., 1995. Tropical Montane Cloud Forests. Springer-Verlag, New York. Heyer, W.R., 2000. Amphibian declines and extinctions: the situation in Latin America. In: Pe´faur, J.E. (Ed.), Ecologı´a Latinoamericana. Actas III Congreso Latinoamericano de Ecologı´a. Universidad de Los Andes, Me´rida, pp. 47–55. INEGI, 1995. Ortofotos Digitales (Digital Orthophotographs) E14b26f, E14b27d, E14b36c and E14b37a. Veracruz, Me´xico. Kolozsvary, M.B., Swihart, R.K., 1999. Habitat fragmentation and the distribution of amphibians: patch and landscape correlates in farmland. Canadian Journal of Zoology 77, 1288–1299. Lande, R., 1996. Statistics and partitioning of species diversity, and similarity among multiple communities. Oikos 76, 5–13. Laurance, W.F., Lovejoy, T.E., Vasconcelos, H.L., Bruna, E.M., Didham, R.K., Stouffer, P.C., Gascon, C., Bierregaard, R.O., Laurance, S.G., Sampaio, E., 2002. Ecosystem decay of Amazonian forest fragments: a 22-year investigation. Conservation Biology 16, 605–618. Lee, J.C., 1996. The Amphibians and Reptiles of the Yucata´n Penı´nsula. Cornell University Press, Ithaca. Lips, K.R., 1998. Decline of a tropical montane amphibian fauna. Conservation Biology 12, 106–117. Mendelson III, J.R., 1997. A new species of toad (Anura: Bufonidae) from Oaxaca, Me´xico with comments on the status of Bufo cavifrons and Bufo cristatus. Herpetologica 53, 268–286. Miller, M.P. 1997. Tools for population Genetic Analyses (TFPGA) 1.3: A Windows Program for the Analysis of Allozyme and Molecular Population Genetic Data. Available: http://www. public.asu.edu/-mmille8/tfpga.htm. Moguel, P., Toledo, V.M., 1999. Biodiversity conservation in traditional coffee systems of Mexico. Conservation Biology 13, 11–21. Murcia, C., 1995. Edge effects in fragmented forests: implications for conservation. Trends in Ecology and Evolution 10, 58–62. Patton, D.R., 1975. A diversity index for quantifying habitat edge. Wildlife Society Bulletin 394, 171-173. In: Forman, R.T.T. (Ed.),

Land Mosaics: The Ecology of Landscapes and Regions. Cambridge University Press, Cambridge, p. 142. Perfecto, I., Rice, R.A., Greenberg, R., Vander Voort, M.E., 1996. Shade coffee: a disappearing refuge for biodiversity. Bioscience 46, 598–608. Rzedowski, J., 1998. Diversidad y orı´genes de la flora faneroga´mica de Me´xico. In: Ramammoorthy, T.P., Bye, R., Lot, A., Fa, J. (Eds.), Diversidad Biolo´gica de Me´xico: Orı´genes y distribucio´n. Universidad Nacional Auto´noma de Me´xico, Me´xico, DF, pp. 129–145. Saunders, D.A., Hobbs, R.J., Margules, C.R., 1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5, 18–32. Schluter, D., Ricklefs, R.E., 1993. Species diversity: an introduction to the problem. In: Ricklefs, R.A., Schluter, D. (Eds.), Species Diversity in Ecological Communities: Historical and Geographical Perspectives. The University of Chicago Press, Chicago, pp. 1–10. Scott Jr., N.J., 1994. Complete species inventories. In: Heyer, W.R., Donnelly, M.A., McDiarmid, R.W., Hayek, L.A.C., Foster, M.S. (Eds.), Measuring and Monitoring Biological Diversity. Standard Methods for Amphibians. Smithsonian Institute Press, Washington, DC, pp. 84–92. Sobero´n, J.M., Llorente, J.B., 1993. The use of species accumulation functions for the prediction of species richness. Conservation Biology 7, 480–488. Sokal, R.R., Rohlf, F.J., 1995. Biometry, 3rd ed. W.H. Freeman and Company, New York. Soto, E.M., Go´mez, M., 1990. Atlas clima´tico del municipio de Xalapa. Instituto de Ecologı´a, AC, Xalapa. StatSoft, Inc., 1997. STATISTICA for Windows. Version 5.1. Tulsa. Available: http://www.statsoft.com. Tocher, M.D., Gascon, C., Zimmerman, B.L., 1997. Fragmentation effects on a central Amazonian frog community: a ten-year study. In: Laurance, W.F., Bierregaard, R.O. (Eds.), Tropical Forest Remnants: Ecology, Management, and Conservation of Fragmented Communities. The University of Chicago Press, Chicago, pp. 127–137. Vallan, D., 2000. Influence of forest fragmentation on amphibians diversity in the nature reserve of Ambohitantely, highland Madagascar. Biological Conservation 96, 31–43. Veech, J.A., Summerville, K.S., Crist, T.O., Gering, J.C., 2002. The additive partitioning of species diversity: recent revival of an old idea. Oikos 99, 3–9. Vogt, R.C., Villarreal, J.L., Pe´rez-Higareda, G., 1997. Lista notada de anfibios y reptiles. In: Gonza´lez-Soriano, E., Dirzo, R., Vogt, R.C.Historia Natural de Los Tuxtlas. Universidad Nacional Auto´noma de Me´xico, Me´xico, DF 1, pp. 507–522. Wake, D.B., 1991. Declining amphibian populations. Science 253, 860. Wake, D.B., Papenfuss, T.J., Lynch, J.F., 1992. Distribution of salamanders along elevational transects in Mexico and Guatemala. Tulane Studies in Zoology and Botany Supplies Publications 1, 303–309. Williams-Linera, G., Manson, R.H., Isunza, E.V., 2002. La fragmentacio´n del bosque meso´filo de montan˜a y patrones de uso del suelo en la regio´n oeste de Xalapa, Veracruz, Me´xico. Madera y Bosques 8, 73–89. Williams-Linera, G., 2002. Tree species richness complementarity, disturbance and fragmentation in a Mexican tropical montane cloud forest. Biodiversity and Conservation 11, 1825–1843. Wright, D.H., Reeves, J., 1992. On the meaning and measurement of nestedness of species assemblages. Oecologia 92, 416–428. Wright, D.H., Patterson, B.D., Mikelson, G.M., Cutler, A., Atmar, W., 1998. A comparative analysis of nested subset patterns of species composition. Oecologia 113, 1–20. Wyman, R.L., 1990. What’s happening to the amphibians? Conservation Biology 4, 350–352. Zar, J., 1984. Biostatistical Analysis, 2nd ed.. Prentice Hall, New Jersey. Zimmerman, B.L., Bierregaard, R.O., 1986. Relevance of the equilibrium theory of island biogeography and species-area relations to conservation with a case from Amazonia. Journal of Biogeography 13, 133–143.