Hurricane impacts on plant-herbivore networks along a successional chronosequence in a tropical dry forest

Forest Ecology and Management xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco

Hurricane impacts on plant-herbivore networks along a successional chronosequence in a tropical dry forest☆ N. Luvianoa, E. Villa-Galavizb, K. Boegec, A. Zaldívar-Rieverónd, E. del-Vala,

⁎

a Instituto de Investigaciones en Ecosistemas y Sustentabilidad, Universidad Nacional Autónoma de México, Antigua Carretera a Pátzcuaro No. 8701, Col. Ex-Hacienda de San José de La Huerta, C.P. 58190 Morelia, Michoacán, Mexico b School of Biological Sciences, University of Bristol, Life Sciences Building, Bristol BS8 1TQ, UK c Instituto de Ecología, Universidad Nacional Autónoma de México, Circuito Exterior s/n, Ciudad Universitaria, C.P. 04510 Ciudad de México, Mexico d Colección Nacional de Insectos, Insituto de Biología, Universidad Nacional Autónoma de México, Circuito Exterior s/n, Ciudad Universitaria, C.P. 04510 Ciudad de México, Mexico

A R T I C L E I N F O

A B S T R A C T

Keywords: Chronosequence Ecological networks Extreme events Herbivores Hurricane Secondary succession Tropical dry forest

Hurricanes are one of the major natural forces affecting plant and animal dynamics in areas where tropical dry forests (TDF) are located. Most studies of hurricane impacts have evaluated the effects on plants, but few have addressed how interactions between plants and animals are affected. We investigated the effect of hurricane Jova on plant-lepidopteran interaction networks along a TDF successional gradient in the Chamela region on the Mexican Pacific coast. We surveyed plant-lepidopteran larvae trophic interactions during 2007–2010 and resurveyed this association in 2012 and 2013, following the landfall of a category 2 hurricane. The parameters of plant-lepidopteran interaction networks changed after the hurricane. Network size, number of compartments and specificity (H2) diminished, while connectance and robustness increased. Network differences throughout the chronosequence remained similar to the patterns observed in previous studies showing larger networks and higher levels of robustness but lower connectance in late successional stages than in recently abandoned pastures. It is interesting that even though the hurricane affected plant-lepidopteran interaction networks, the topology related to forest succession remained unaltered after the passage of the hurricane, demonstrating the long legacy of the human footprint on ecosystems.

1. Introduction

2008) and plant animal interactions (Ferguson, 1995; Pascarella, 1998; Rathcke, 2001; Angulo-Sandoval et al., 2004; Horvitz et al., 2005). However few studies have documented the impact in plant-animal networks (Sánchez-Galván et al., 2012). Following the passage of a hurricane, the most evident damage is the loss of canopy foliage, which translates into increased light intensities and temperatures within the forest and a decrease in relative humidity (Lugo, 2008), and therefore modifying the microhabitat for animal species. Lugo (2008) proposed that the main post-hurricane impacts on animals are mediated through changes in the ecological space available to organisms, habitat modification, as well as to increased landscape heterogeneity and variability in ecosystem processes. Depending on the life histories of the species involved, the hurricane impact can thus present wide variation. For example, while bird communities, particularly nectarivores and fruit/ seed eaters, showed a general decline following Hurricane Hugo in the US Virgin Islands (Askins and Ewert, 1991; Wauer and Wunderle, 1992), whereas bats showed decreased abundance and boas increased

Tropical dry forests (TDF) are located in tropical and subtropical regions of the world, where hurricanes are one of the major natural forces affecting plant and animal communities. Hurricanes are an example of large and infrequent disturbances that have profound effects on ecosystems (Lugo, 2008). Post-hurricane plant dynamics involve the growth of secondary succession species replacing large trees with slow growth rates (Denslow, 1980; Zimmerman et al., 1994; Denslow, 1995). However, while TDF species have evolved with these occasional large disturbances over a long time period, we are unsure how they will react to the increasing hurricane magnitude, frequency and intensity predicted by recent climatic models. Particularly, hurricanes are already becoming more severe and frequent on the Pacific coast of Mexico than they were previously recorded (Knutson et al., 2015). Most investigations studying hurricane impacts have addressed the effects on vegetation, including changes in animal communities (Lugo,

☆ ⁎

Part of the Special Issue “Resilience of tropical dry forests to extreme disturbance events: An interdisciplinary perspective from long-term studies”. Corresponding author. E-mail address: [email protected] (E. del-Val).

http://dx.doi.org/10.1016/j.foreco.2017.09.011 Received 30 May 2017; Received in revised form 30 August 2017; Accepted 2 September 2017 0378-1127/ © 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Luviano, N., Forest Ecology and Management (2017), http://dx.doi.org/10.1016/j.foreco.2017.09.011

Forest Ecology and Management xxx (xxxx) xxx–xxx

N. Luviano et al.

2.3. Sampling protocol

their home range in Puerto Rico (Wunderle et al., 2004). In the case of invertebrates, the impacts of a hurricane are strikingly noticeable. Population counts on invertebrates after Hurricane Hugo in Puerto Rico showed dramatic reductions for three species of snails and two walking-sticks (Willing and Camilo, 1991). For lepidopterans, Torres (1992) documented an outbreak of 15 generalist species probably due to a flush of new foliage that occurred after the hurricane and the ability of these species to include novel host plants in their diet. This outbreak was subsequently controlled by Dipteran and Hymenopteran parasitoids. However, changes in plant-animal interaction networks following a hurricane strike have never been fully investigated. Lepidopteran communities studied along a successional gradient following anthropogenic disturbances in a TDF of the Mexican Pacific coast have revealed marked changes in richness, diversity and species composition (López-Carretero, 2010). Moreover, lepidopteran–plant interaction network in the same locality has been shown to differ among successional stages (Villa-Galaviz et al., 2012). Lepidopterans are a key group of animals that are highly relevant for pollination, and therefore their conservation is crucial to maintain ecosystem functions and services (Bawa et al., 1985; Haber and Frankie, 1989). Because hurricanes on the Mexican Pacific coast have been predicted to increase in frequency and magnitude, it is important to investigate the effect that these extreme events have on biodiversity and ecosystem functioning, including plant-animal interactions, in order to determine and document resilience of this ecosystem. Here we investigated the effect of category 2 Hurricane Jova on lepidopteran diversity and on the plantlepidopteran interaction network across a TDF successional gradient. In particular, we addressed the following questions: (i) are mature forest plant-herbivore networks more resilient to hurricane disturbance than early successional forests? and (ii) is the footprint of the successional gradient on plant-herbivore networks maintained after the hurricane?

From 2007 to 2010, we sampled lepidopteran larvae every year in 11 plots corresponding to the experimental design of the CIECO-UNAM Tropical Forest Management project (MABOTRO). This project features a successional chronosequence of abandoned agricultural fields and cattle ranching pastures (Avila-Cabadilla et al., 2009). This chronosequence consists in plots of four stages of secondary succession: initial (fields abandoned for four years), early secondary forest (6–9 years of abandonment), late secondary forest (13–16 years of abandonment) and mature forest. There were three replicates for each successional stage, with the exception of the initial stage, which only had two due to a fire in the third replicate. Plot size was at least 1 ha (the mature forest plots are immersed in the biosphere reserve and are larger) and the distance between plots of the same successional stage was 3 km. Within each plot, in a defined area of 20 × 50 m, we established four parallel 2 × 20 m transects separated by 10 m. With the exception of lianas, all woody plants with stems ≥1 cm in diameter and ≥50 cm in height within these transects were labeled. During the rainy season of each year during the period 2007–2010, we surveyed caterpillars on all leaves and stems of all marked plants within transects up to 3 m in height. In adult trees, which represented only 10% of all marked individuals, a subsample of three branches – consisting of approximately 100 leaves - was taken for each tree. Caterpillars were recorded and reared in the laboratory for subsequent taxonomic identification of adults (see Villa-Galaviz et al., 2012 for details). We also generate DNA Barcoding sequences (∼650 bp of the cytochrome oxidase I mitochondrial DNA gen; Hebert et al., 2003) to molecularly characterize presumptive species using the Barcode Index Number (BIN) system (Rhatnasingham and Hebert, 2013). For this, one leg of each adult specimen, or the head capsule, if the caterpillar died was removed for DNA extraction using the EZ-10 Spin Kit minipreps DNA Genomic Column kit (BIO Basic, Toronto, Ontario, Canada). The amplification protocol and primers employed followed Ceccarelli and Zaldívar-Riverón (2013). Sequences were edited with the program Sequencher version 4.0.5 (Gene Codes) and manually aligned based on their translated amino acids. The Barcoding sequences, locality details and GenBank accession numbers can be retrieved from the project file ‘Lepidoptera from Chamela, Mexico’ (MXBLP), which is found in the projects section of the Barcode of Life Data Systems (www.boldsystems. org). In a previous study of this system, we reported plant-herbivore networks to be highly resilient during forest succession (Villa-Galaviz et al., 2012). For each successional stage replicate, we constructed a bipartite network and calculated the following network structure descriptors: network size (number of lepidopteran and plant species interacting), connectance (fraction of realized interactions from the possible total) (Dunne et al., 2002), robustness (based on the species fraction that survives after removing one species, the value is obtained after calculating the area below the extinction curve) (Memmott et al., 2004; Burgos et al., 2007), number of compartments (network subgroups not connected with other subgroup) (Tylianakis et al., 2007), and network specialization (H2, measures the deviance between the realized interactions for one species and the expected for each species with respect to the total possible interactions in the network) (Blüthgen, 2006). When comparing these network parameters among successional stages, we did not find differences between secondary and mature forest and only the recently abandoned fields showed significant differences (Villa-Galaviz et al. 2012). To evaluate whether this resilience in plant-herbivore networks across succession remained after the occurrence of an extreme event such as Hurricane Jova, we sampled the same plots during 2012 and 2013 and compared plant-herbivore network structure metrics before and after the hurricane. Linear mixed effect models (R nlme package) were used to

2. Materials and methods 2.1. Study site This study was carried out in the Chamela-Cuixmala region of Jalisco, Mexico (CCBR, 19 °22′–19 °39′ N, 104 °56′–105 °10′ W), comprising the Biosphere Reserve (13,142 ha of conserved land) and its surroundings. The average annual rainfall is 795.3 mm but varies greatly from year to year (from 366 to 1329 mm) and is mainly concentrated (87%) within the rainy season from June to October (Maass et al., this issue). The vegetation within the reserve primarily consists of TDF (1149 plant species with an average canopy height of 6 m) and semi-deciduous forest established along larger streams (average canopy height of 10 m) (Lott et al., 1987). The surrounding area consists of a mosaic of secondary succession forests, agricultural fields and cattle grazing pastures (Sánchez-Azofeifa et al., 2009). The TDF found at the Chamela-Cuixmala region is considered one of the most diverse of its kind, with 1200 plant species and a high levels of endemism (Lott et al., 1987; Trejo and Dirzo, 2000). The invertebrate inventory is still very limited; 1877 arthropod species have been described in the reserve, 583 of which are lepidopterans (Pescador-Rubio et al., 2002).

2.2. Hurricane Jova Hurricane Jova struck the coasts of Colima and Jalisco, Mexico on October 10th 2011 as category 2 in the Saffir Simpson scale. During the hurricane, maximum wind intensities reached 205 km/h with gusts of up to 250 km/h. By October 13th, it lost strength and was reduced to a tropical storm category. Jova lasted 168 h, advancing 2000 km at an average speed of 12 km/h (Brennan, 2012) and producing 187.9 mm of precipitation in two days at the Chamela biological station.

2

Forest Ecology and Management xxx (xxxx) xxx–xxx

N. Luviano et al.

from 134 in 2007 to 270 in 2008, while in 2012 and 2013 it significantly decreased to 122 and 89 species, respectively (F(1,56) = 21.8, P < 0.001). Lepidopteran richness and abundance showed no differences between successional plots (richness: F(3,8) = 2.3, P = 0.15 and abundance F(3,8) = 3.3, P = 0.08), independently of the hurricane strike (successional stage vs. hurricane interaction not significant; Table 1).

determine whether there were differences between lepidopteran richness, abundance and network parameters in each different successional stage and between pre-hurricane and post-hurricane years (Crawley, 2002). For these models, successional stage and pre or post hurricane conditions were used as fixed effects while the year of study and plot were considered as the random effects. For the analysis of connectance, number of compartments and robustness, network size was also used as a covariate because of the correlation between this variable and the other descriptors. For each network, we defined core or peripheral lepidopteran and plant species using the recent formula proposed by Dáttilo et al. (2013) that considers generalist species and provides stability to the network. We analyzed changes in these core species using linear mixed effects models with the same structure as described above. All analyses were performed in R (R Development Core Team, 2014).

3.2. Plant-lepidopteran networks Following the hurricane, plant-lepidopteran networks showed patterns similar to those previously found in the secondary succession plots. Early successional stages had smaller networks and were less robust than the older ones (F(3,8) = 8.8, P = 0.006. Table 1 and Fig. 1). In contrast, connectance, number of links and compartments did not present differences between successional stages (P > 0.05), while specificity (H2) was greater in early successional stages (P < 0.05). However, after the hurricane Jova stroke the region, network size, number of compartments and specificity (H2) decreased while connectance and robustness increased, although the number of links between species was not affected (Table 1 and Fig. 1). The hurricane did not have differential effects on network parameters, as a function of the successional stage (i.e., significant hurricane × succession interaction was not significant). We found 56 different lepidopteran species that represented core species in some networks during particular years. However, only 1 to 7 core species were found in any given network and there was a high level of species substitution. Common core species were Orgiya sp (Erebidae, present in 47% of networks), Psychidae sp. 1 (in 45% of networks), Psychidae sp. 2 (in 29% of networks) and Geometridae sp. 1 (in 9% of networks). The number of core lepidopteran generalist species in the networks after the hurricane was similar to those previously observed, depending on the particular successional stage. Generalist core herbivore species ranged from 1 in the early successional stage to 2.6 in the late successional stage (Fig. 1). There were no significant differences in core generalist species among successional stages (F(3,8) = 2.07, P = 0.18) or after the hurricane (F(1,56) = 0.24, P = 0.62) and the interaction between successional stage and hurricane was also non-significant (F(3,56) = 0.14, P = 0.93). In terms of core plant species, we found 45 generalist host species. Particularly important were Casearia nitida (present in 24% of networks), Apoplanesia paniculata (in 22% of networks), Caesalpinia caladenia (in 22% of networks) and Croton roxanae (in 20% of networks). The number of core plant species ranged from 0 (early successional stage in 2009) to five species (late successional stage in 2010), depending on the successional stage and year, but variation was considerably wide among years that there were no significant differences between successional stages (F(3,8) = 1.77, P = 0.23), hurricane effect (F(1,56) = 2.47, P = 0.12) or in the interaction between succession vs. hurricane effect (F(3,56) = 2.6, P = 0.06; Fig. 1 and Table 1).

3. Results 3.1. Diversity During the wet season of 2012 and 2013, we recorded 1774 and 1084 caterpillars respectively. Lepidopteran abundance in previous years ranged from 737 to 2715, therefore abundances after the hurricane were similar to previous years (F(1,56) = 0.40; P = 0.53, Table 1). Lepidopteran species richness in years prior to the hurricane ranged Table 1 Mixed linear-model analysis testing effects on different network and diversity parameters considering the effect of hurricane, successional stage and successional stage vs. hurricane interaction. Significant effects are marked in bold. Parameters

D.F.

F value

P value

Lepidopteran richness Hurricane Successional stage SS vs Hurricane

1.56 3.8 3.56

21.76 2.26 0.95

0.0001 0.16 0.42

Lepidopteran abundance Hurricane Successional stage SS vs Hurricane

1.56 3.8 3.56

0.4 3.33 0.1

0.52 0.07 0.96

Network size Hurricane Successional stage SS vs Hurricane

1.51 3.7 3.51

17.01 7.24 0.65

0.001 0.015 0.59

Connectance Hurricane Successional stage SS vs Hurricane

1.49 3.7 3.49

28.96 13.39 1.31

0.0001 0.003 0.28

H2 Hurricane Successional stage SS vs Hurricane

1.51 3.7 3.51

20.1 0.77 2.29

0.0001 0.73 0.09

Number of compartments Hurricane Successional stage SS vs Hurricane

1.50 3.7 3.50

23.14 1.41 2.44

0.0001 0.32 0.07

Robustness Hurricane Successional stage SS vs Hurricane

1.50 3.7 3.50

265.48 7.8 1.67

< 0.0001 0.01 0.18

No. Caterpillar core species Hurricane Successional stage SS vs Hurricane

1.56 3.8 3.56

0.24 2.07 0.14

0.62 0.18 0.93

No. plant core species Hurricane Successional stage SS vs Hurricane

1.56 3.8 3.56

2.47 1.77 2.06

0.12 0.23 0.06

4. Discussion Severe natural disturbances can alter ecosystems in different ways. In particular, hurricanes represent natural disturbances in coastal forests throughout the world. The effects of hurricanes on biodiversity depend on their frequency, magnitude and duration (Flynn et al., 2010, Kupfer et al., 2008, Ross et al., 2001). The impact of hurricane strikes is most evident in the forest canopy and this causes a sudden increase in light availability within the forest (Brokaw and Grear, 1991, Zimmerman et al. 1994, Vandermeer et al., 1996). The density and basal area of tropical dry forest trees was found to diminish by up to 42% after Hurricane George stroke Puerto Rico (Rojas-Sandoval et al., 2014), while on the Caribbean coast of Mexico, at Puerto Morelos, Quintana Roo, Sánchez-Sánchez and Islebe (1999), found a 33% reduction in stem density following Hurricane Gilbert. 3

Forest Ecology and Management xxx (xxxx) xxx–xxx

N. Luviano et al.

/ŶŝƟĂů

>ĂƚĞͲƐƵĐĐĞƐƐŝŽŶĂů

ĂƌůLJͲƐƵĐĐĞƐƐŝŽŶĂů

b 700

Ă 60

600 500

Abundance

Richness

50 40 30 20

400 300 200

10

100

0

Đ 80

2007

2008

2009

2010

2012

0

2013

2

Connectance 2008

2009

2010

2012

Compartments

Links

0 2007 2008 2009 2010 2012 2013

g 0.7

h

2012

2013

2007

2008

2009

2010

2012

2013

2007

2008

2009

2010

2012

2013

0.4 0.2

14 12 10 8 6 4 2 0 1 0.8

H2

0.5

Robustness

2010

Ĩ 16

0.5

0.3

0.6 0.4 0.2

0.1

0

2007 2008 2009 2010 2012 2013

j

6

Lepcore species

2009

0.6

2013

1

i

2008

0 2007

1.5

-0.1

2007

d 0.8

70 60 50 40 30 20 10 0

Ğ

DĂƚƵƌĞĨŽƌĞƐƚ

5

2007

2008

2009

2010

2012

2013

2007

2008

2009

2010

2012

2013

5 4

4 3

3

2

2

1

1 0

0 2007

2008

2009

2010

2012

2013

Fig. 1. Lepidopteran diversity (richness and abundance) and plant-herbivore network parameters in the study area, comparing years before (2007–2010) and after (2012, 2013) the impact of Hurricane Jova (denoted with the dotted line). Showing mean ± EE for each parameter per successional stage per year.

4

Forest Ecology and Management xxx (xxxx) xxx–xxx

N. Luviano et al.

and after the hurricane. This finding is particularly relevant, because through their interaction with most of the species within the network (Dáttilo et al., 2013), generalist species contribute to the stability and functioning of the system in a greater extent than the peripheral species. A previous investigation in our study system found differences in plant-lepidopteran networks between successional stages; in particular, recently abandoned pastures present smaller, more connected and less robust networks than later successional stages (Villa-Galaviz et al., 2012). We found that following the strike of hurricane Jova, differences between successional stages remained similar: old-growth forests and late successional stages present larger networks with greater robustness, while in recently abandoned pastures networks with greater connectance prevailed. Therefore human landscape modifications in the region appear to be long-lasting and surpass natural disturbances. One could argue that the resilience we found was due to the fact that our studied transects were not severely affected by the hurricane; however, we did record the loss of individual plants and many large broken branches. We are therefore confident than the observed patterns are representative of the damage inflicted by Hurricane Jova in the forest. In a recent study in the same forest, Saavedra and collaborators (2017) have shown that during forest succession, reorganization of herbivore communities has an important effect on community dynamics and composition, regardless of the resilience of plant-herbivore interaction networks. In particular, resident (early-arriving) herbivore species increase their likelihood of persistence across forest successional stages, and negatively affect persistence of colonizing (late-arriving) species (Saavedra et al., 2017). Hence, it would be interesting to know whether a major environmental disturbance such as a hurricane has also an effect on the reorganization of herbivore communities. Because hurricanes in the Pacific coast of Mexico are predicted to become more frequent and severe (Knutson et al., 2015), there is a need to further understand the implications of consequence of global changes for biodiversity conservation and maintenance of ecosystem functions. This study provides evidence that plant-herbivore interactions in the tropical dry forest appear to regenerate following anthropogenic disturbances (such as land-use changes) and also after the considerable impact of a relatively mild hurricane (category 2). Whether this would be the case with the occurrence of more severe and frequent hurricanes remains an open question. It is interesting that the signature of the network topology related to successional stages remained unaltered after the hurricane, illustrating the long-lasting legacy of the human footprint on the ecosystem. Nevertheless, it is important to consider that the studied TDF may be resilient to both anthropogenic and natural disturbances thanks to the nearby presence of a large conservation area nearby (i.e. the Chamela-Cuixmala Biosphere Reserve), providing a source of plant propagules and associated herbivore species. In this context, we suggest that conservation planning should always ensure the resilience capacity of the system by providing large areas of conserved forest in addition to any other management strategies adopted.

Increased light availability also facilitates the production of new leaves in many species and hence a sudden increase in resource availability for folivores. Some studies have reported that canopy openness is a determinant factor for lepidopteran richness and abundance (Barlow et al., 2007; Hawes et al., 2009) and the effects of a hurricane strike are therefore expected to be evident in lepidopterans, as has been observed in other insects. For example, Osborn (1935) reported a decrease in the abundance of Pectinophora gosypiella (Lepidoptera: Gelechiidae), following Hurricane San Felipe in Puerto Rico, and Wolcott (1941) reported a population decrease in Icerya purchasi (Hemiptera: Margarodidae) following Hurricane San Ciprián, while Coccus viridis (Hemiptera: Coccidae) presented increased abundance in open coffee fields. More recently, Grimmbacher and Stork (2009) investigated changes in beetle assemblages in a fragmented tropical forest in Queensland, Australia, following the passage of tropical cyclone Larry. These authors found an increased abundance in drought tolerant species since they are able to cope with greater light incidence experimented after the hurricane. However, species communities were similar to their pre-hurricane condition as only 5% of the species were affected. Fragmentationinduced differences in the beetle communities were also unaltered by the hurricane. The impacts of hurricanes on particular arthropod communities can also have regional effects. For example, hurricanes impacting the Gulf of Mexico coast in Brownsville (USA) and Tampico (Mexico) in 1933 promoted an unusual presence of tropical butterflies in Texas (Neck, 1934). This phenomenon was also observed 30 years later following Hurricane Beulah when tropical lepidopterans were again recorded in Texas (Doyle, 1970; Kendall, 1970; Heitzman and Heitzman, 1972). Lepidopteran outbreaks were also recorded in Puerto Rico following Hurricane Hugo in 1989; in this case, the species observed were open habitat specialists (Torres, 1992). After Hurricane Rita struck the Louisiana coast, sugar cane farmers observed an increased abundance of the caterpillar Diatraea saccharalis (Beuzelin et al., 2009). In this study, we found a decrease in lepidopteran richness but not in abundance. Since year-to-year fluctuations in lepidopteran larval abundance are quite wide in this system (i.e. species turnover is ca. 50% between years and total abundance can double from one year to the next) (López-Carretero, 2010), caterpillar richness recorded after Hurricane Jova fell into the previously observed range and did not differ significantly from that of previous years. Plant-animal interaction networks are increasingly being studied to further understand community structuring and resilience to different environmental disturbances (Forup and Memmott, 2005; Lewinsohn et al., 2006; Bascompte, 2010). López-Carretero et al. (2014), for example, studied variations in plant-herbivore interactions over different seasons in a tropical coastal forest of Mexico and found that network specialization declines during dry periods. These authors argued that more specialized species tend to be restricted to periods of less stressful conditions while generalists are present throughout the year. DiazCastelazo et al. (2009) also compared a plant-ant interaction network with a 10 year separation period and found similar network parameters, indicating high network resilience to changes in the number of interactions as well as to the arrival of new species across years. Nevertheless, few studies have investigated the effects of hurricanes on ecological networks. The only study to explore this was that of Sánchez-Galván et al. (2012), that evaluated a plant-mutualist ant network in la Mancha in Veracruz, Mexico, prior to and following the impact of Hurricane Karl in the Gulf of Mexico. These authors found strong resilience in the network proprieties, as the impact of the hurricane did not alter the network attributes. We found that after hurricane Jova, some parameters in our plant-lepidopteran network were negatively affected (network size, number of compartments and specificity), but connectance and robustness actually increased. These results suggest that our plant-lepidopteran network had some resilience to this disturbance, probably due to the core generalists that persisted before

5. Conclusions The structure of plant-lepidopteran interaction networks showed some changes following the passage of category 2 Hurricane Jova but robustness of the network increased. This result suggests that some TDF food web properties could be resilient to both anthropogenic and natural perturbations if large natural reserves are present, highlighting the importance of conservation and enlargement of existing natural protected areas. However, it remains uncertain whether this type of interaction can continue to show resilience in the face of the increased frequency and severity of hurricanes expected as a result of climate change.

5

Forest Ecology and Management xxx (xxxx) xxx–xxx

N. Luviano et al.

(Hesperiidae and Libytheidae). J. Res. Lepidopt. 10, 284–286. Horvitz, C.C., Tuljapurkar, S., Pascarella, J.B., 2005. Plant-animal interactions in random envrionments: habitat-stage elasticity, seed predators and hurricanes. Ecology 86, 3312–3322. Kendall, R.O., 1970. Three hairstreaks (Lycaenidae) new to Texas and the United State. J. Lepidopt. Soc. 24, 59–61. Knutson, T.R., Sirutis, J.J., Zhao, M., Tuleya, R.E., Bender, M., Vecchi, G.A., Villarini, G., Chavas, D., 2015. Global projections of intense tropical cyclone activity for the late twenty-first century from dynamical downscaling of CMIP5/RCP4.5 scenarios. J. Clim. 28, 7203–7224. Kupfer, J.A., Myers, A.T., McLane, S.E., Melton, G.N., 2008. Patterns of forest damage in a southern Mississippi landscape caused by hurricane Katrina. Ecosystems 11, 45–60. Lewinsohn, T.M., Jordano, P., Bascompte, J., 2006. Structure in plant-animal interaction assemblages. Oikos 113, 174–184. López-Carretero, A., 2010. Composición y diversidad de lepidópteros en la cronosecuencia sucesional del bosque tropical caducifolio, consecuencias sobre la herbivoría de Casearia nitida. Mexico DF, UNAM, pp. 86. López-Carretero, A., Boege, C., Díaz-Castelazo, K., Rico-Gray, V., 2014. Evaluating the spatio-temporal factors that structure network parameters of plant-herbivore interactions. PloS ONE 9, e110430. Lott, E.J., Bullock, S.H., Solís-Magallanes, J.A., 1987. Floristic diversity and structure of upland and arroyo forests in coastal Jalisco. Biotropica 19, 228–235. Lugo, A.E., 2008. Visible and invisible effects of hurricanes on forest ecosystems: an international review. Austral Ecol. 33, 368–398. Maass, J.M., Ahedo-Hernández, R., Araiza, S., Verduzco, A., MArtínez-Yrízar, A., Jaramillo, V.J., Parker, G., Pascual, F., García-Méndez, G., Sarukhán, J., this issue. Long-term (33 years) rainfall and runoff dynamics in a tropical dry forest ecosystem in western Mexico: management implications under extreme hydrometeorological events. For. Ecol. Manage. Memmott, J., Waser, N.M., Price, M.V., 2004. Tolerance of pollination networks to species extinctions. Proc. R. Soc. Lond. Ser. B-Biol. Sci. 271, 2605–2611. Neck, R.W., 1934. Effects of 1933 hurricanes on butterflies of central and southern Texas. J. Lepidopt. Soc. 31, 67–68. Osborn, H., 1935. Insects of Puerto Rico and The Virgin Islands: Homoptera (excepting the Sternorhynchi). N.Y. Acad. Sci. 14, 111–260. Pascarella, J.B., 1998. Hurricane disturbance, plant-animal interactions, and the reproductive success of a tropical shrub. Biotropica 30, 416–424. Pescador-Rubio, A., Rodríguez-Palafox, A., Noguera, F.A., 2002. Diversidad y estacionalidad de Arthropoda. In: Noguera-Alderte, A.N., Vega-Rivera, J.H., Quesada, M. (Eds.), Historia natural de Chamela. Insituto de Biología, UNAM, Mexico DF. Rathcke, B.J., 2001. Pollination and Predation Limit Fruit Set in a Shrub, Bourreria succulents (Boraginaceae), after Hurricanes on San Salvador Island, Bahamas. Biotropica 33, 330–338. RDevelopmentCoreTeam, 2014. R: A Language and Environment for Statistical Computing R Foundation for Statistical Computing. Viena, Austria. Rhatnasingham, S., Hebert, P.D.N., 2013. A DNA-based registry for all animal species: the Barcode index number (BIN) system. PloS ONE 8, e66213. Rojas-Sandoval, J., Meléndez-Ackerman, E.J., Fumero-Caban, J., García-Bermúdez, M.A., Sustache, J., Aragón, S., Morales, M., Fernández, D.S., 2014. Effects of hurricane disturbance and feral goat herbivory on the structure of a Caribbean dry forest. J. Veg. Sci. 25, 1069–1077. Ross, M.S., Carrington, M., Flynn, L.J., Ruiz, P.L., 2001. Forest succession in tropical hard wood hammock sof the Florida keys:effects of direct mortality from hurricane Andrew. Biotropica 33, 23–33. Sánchez-Azofeifa, G.A., Quesada, M., Cuevas-Reyes, P., Castillo, A., Sánchez-Montoya, G., 2009. Land cover and conservation in the area of influence of the Chamela-Cuixmala Biosphera Reserve, Mexico. For. Ecol. Manage. 258, 907–912. Sánchez-Galván, L.R., Díaz-Castelazo, C., Rico-Gray, V., 2012. Effect of Hurricane Karl on a plant–ant network occurring in coastal Veracruz, Mexico. J. Trop. Ecol. 28, 603–609. Sánchez-Sánchez, O., Islebe, G.A., 1999. Hurricane Gilbert and structural changes in a tropical forest in south-eastern México. Global Ecol. Biogeogr. 8, 29–38. Torres, J.A., 1992. Lepidoptera outbreaks in response to successional changes after the passage of Hurricane Hugo in Puerto Rico. J. Trop. Ecol. 8, 285–298. Trejo, I., Dirzo, R., 2000. Deforestation of seasonally dry tropical forest: a national and local analysis in Mexico. Biol. Conserv. 94, 133–142. Tylianakis, J.M., Tscharntke, T., Lewis, O.T., 2007. Habitat modification alters th estructure of tropical host-parasitoid food webs. Nature 445, 202–205. Villa-Galaviz, E., Boege, K., del-Val, E., 2012. Resilience in plant-herbivore networks during secondary succession. PloS ONE 7, e53009. Wauer, R.H., Wunderle, J.M., 1992. The effect of Hurricane Hugo on bird populations on St. Croix, U.S Virgin Islands. Wilson Bull. 104, 656–673. Willing, M.R., Camilo, G.R., 1991. The effect of hurricane hugo on six invertebrate species in the luquillo experimental forest of Puerto Rico. Biotropica 23, 455–461. Wolcott, G.N., 1941. The dispersion of the cottony cushion-scale in Puerto Rico in eight years. Caribbean Forest. 2, 132–135. Wunderle, J.M.J., Mercado, J.E., Parresol, B., Terranova, E., 2004. Spatial ecology of Puerto Rican boas (Epicrates inornatus) in a hurricane-impacted forest. Biotropica 36, 555–571. Zimmerman, J.K., Everham, E.M., Waide, R.B., Lodge, D.J., Taylor, C.M., Brokaw, N., 1994. Responses of tree species to hurricane winds in subtropical wet forest in Puerto Rico: implications for tropical tree life histories. J. Ecol. 82, 911–922.

Acknowledgements We thank Eloy Castro and members of the Laboratorio de Interacciones Bióticas en Habitats Alterados for field work assistance, and Ruben Pérez Ishiwara, Raúl Ahedo and Salvador Araiza for logistic support. We are grateful with Wesley Dáttilo for his guidance in the analysis. This project was sponsored by SEP CONACyT grant 2012179045 and Papiit-UNAM IN211916. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.foreco.2017.09.011. References Angulo-Sandoval, P., Fernández-Marín, H., Zimmerman, J.K., Aide, M., 2004. Changes in patterns of understory leaf phenology and herbivory following hurricane damage. Biotropica 36, 60–67. Askins, R.A., Ewert, D.N., 1991. Impact of Hurricane Hugo on bird populations on St. John, U.S Virgin Islands. Biotropica 23, 481–487. Avila-Cabadilla, L., Stoner, K.E., Henry, M., Alvarez-Añorve, M., 2009. Composition, structure and diversity of phyllostomid bat assemblages in different successional stages of a tropical dry forest. For. Ecol. Manage. 258, 986–996. Barlow, J., Overal, W.L., Arauyjo, I.S., Gardner, T.A., Peres, C.A., 2007. The value of primary, secondary and plantation forests for fruit-feeding butterflies in the Brazilian Amazon. J. Appl. Ecol. 44, 1001–1012. Bascompte, J., 2010. Structure and dynamics of ecological networks. Science 329, 765–766. Bawa, K.S., Bullock, S.H., Perry, D.R., Coville, R.E., Grayum, M.H., 1985. Reproductive biology of tropical lowland rain forest trees. II Pollination systems. Am. J. Bot. 72, 346–356. Beuzelin, J.M., Reagan, T.E., Akbar, W., Cormier, H.J., Flanagan, J.W., Blouin, D.C., 2009. Impact of hurricane Rita storm on sugarcane borer (Lepidoptera: Crambidae) management in Louisiana. J. Econ. Entomol. 102, 1054–1061. Blüthgen, N., Menzel, F., Blüthgen, N., 2006. Measuring specialization in species interaction networks. BMC Ecol. 6. Brennan, M.J., 2012. Tropical Cyclone Report. Hurricane Jova (EP102011) 6–12 October 2011. National Hurricane Center. Burgos, E., Ceva, H., Perazzo, R.P.J., Devoto, M., Medan, D., Zimmerman, M., Delbue, A.M., 2007. Why nestedness in mutualistic networks? J. Theor. Biol. 249, 307–313. Ceccarelli, F.S., Zaldívar-Riverón, A., 2013. Broad poly- phyly and historical biogeography of the Neotropical wasp genus Notiospathius (Braconidae: Doryctinae). Mol. Phylogenet. Evol. 69, 142–152. Crawley, M.J., 2002. Statistical Computing. An Introduction to Data Analysis using SPlus. John Wiley & Sons, Ltd., Chichester, UK. Dáttilo, W., Guimaraes Jr, P., Izzo, T.J., 2013. Spatial structure of ant-plant mutualistic networks. Oikos 122, 1643–1648. Denslow, J.S., 1980. Patterns in plant species diversity during succession under different disturbance regimes. Oecologia 46, 18–21. Denslow, J.S., 1995. Distrubance and diversity in tropical rain forests: the density effect. Ecol. Appl. 5, 962–968. Diaz-Castelazo, C., Guimaraes Jr., P., Jordano, P., Thompson, J.N., Marquis, R., RicoGray, V., 2009. Changes of a mutualistic network over time: reanalysis over a 10-year period. Ecology 91, 793–801. Doyle, J.E.I., 1970. Field notes on three skippers in Texas (Hesperiidae). J. Lepidopt. Soc. 24, 212. Dunne, J.A., Williams, R.J., Martinez, N.D., 2002. Food-web structure and network theory, the role of connectance and size. Proc. Natl. Acad. Sci. U. S. A. 99, 12917–12922. Ferguson, B.G., 1995. Recruitment and decay of a pulso of Cecropia in Nicaraguan rain forest damaged by Hurricane Joan: relation to mutualism wirh Azteca ants. Biotropica 27, 455–460. Flynn, F.B., Uriarte, M., Crk, T., Pascarella, J.B., Zimmermann, J.K., Aide, M.T., 2010. Hurricane disturbance alters secondary forest recovery in Puerto Rico. Biotropica 2 (42), 149–157. Forup, M.L., Memmott, J., 2005. The restoration of plant pollinator interactions in hay meadows. Restorat. Ecol. 13, 265–274. Grimmbacher, P.S., Stork, N.E., 2009. How do beetle assemblages respond to cyclonic disturbance of a fragmented tropical rainforest landscape ? Oecologia 161, 591–599. Haber, W.A., Frankie, G.W., 1989. A tropical hawkmoth community: costa rican dry forest sphingidae. Biotropica 21, 155–172. Hawes, J., Silva, M.C., Overal, W.L., Barlow, J., Gardner, T.A., Peres, C.A., 2009. Diversity and composistion of Amazonian moths in primary, secondary and plantation forests. J. Trop. Ecol. 25, 281–300. Hebert, P.D.N., Cywinska, A., Ball, S.L., de Waard, J.R., 2003. Biological identifications through DNA barcodes. Proc. R. Soc. Lond. Ser. B-Biol. Sci. 270, 313–321. Heitzman, J.R., Heitzman, R.L., 1972. New butterfly records for the United States

6