Effects of Coffee Agroecosystems on Moth Diversity in San Luis, Costa Rica Nathaniel Talbot Department of Biology, University of Oregon ABSTRACT landscape. Fragmentation studies based on Island Biogeography Theory and Hubble Unified Theory typically view the degraded area surrounding forests as uninhabitable, however, most agroecosystmes actually serve as viable habitat to a subset of natural communities as w ell as a number of colonizing species. As a follow up to Rickets et al (2000) who studied moths in a tropical, agricultural landscape, this study focuses on how moth diversity is affected by coffee plantations in San Luis, Costa Rica. Moths were collected from three sites: a forest, a sun grown coffee farm and a coffee farm grown shaded banana trees. The samples were then divided into morphospecies and many were identified to family. Shannon Weiner diversity indices found species diversity to be highest in forest (H'=4.256), lowest in shade grown coffee (H'=3.413) and intermediate in sun grown coffee (H'=3.863). Multiple comparisons determined all three sites significantly different in diversity. Beta diversity was highest between forest and shade grown coff ee(0.309), intermediate between forest and sun grown coffee(0.358), and lowest between the two coffee farms(0.418).The forest also had a larger representation at the family level along with more rare species. These results give many implications as to how effectively agroecosystems harbor biodiversity and whether this is due to structural or geographical dynamics. RESUMEN La mayora de los bosques primarios de Costa Rica han sido convertidos en tierras agrcolas. Los estudios sobre fragmentacin ba sados en las teora de biogeografa de islas y en la teora unificada de Hubble presentan a la zona degradada alrededor de estos bosques como inhabitable. Sin embargo, la mayora de los agrosistemas en realidad sirven como hbitats viables para un subgrup o tanto de comunidades naturals as como para cierto nmero de especies colonizadoras. Este estudio determin cmo la diversidad de polillas en tierras agrcolas tropicales es afectada por los cafetales en San Luis, Costa Rica, como parte de una continuac in del estudio de Rickets et al (2000). Las polillas fueron recolectadas en tres lugares diferentes: Un bosque, un cafetal sin sombra y un cafetal con plantas de banano. Las muestras fueron clasificadas por especies morfolgicas y muchas fueron identif icadas al nivel de Familia. Los ndices de diversidad de Shannon Weiner demostraron que la diversidad de especies fue mayor en el bosque (H'=4.256), menor en el cafetal con bananos (H'=3.413) e intermedia en el cafetal sin sombra (H'=3.863). Las comparac iones mltiples mostraron que los tres sitios diferan significativamente en diversidad. La diversidad beta fue mayor entre el bosque y el cafetal con sombra (0.309), intermedia entre el bosque y el cafetal sin sombra (0.358) y menor entre los dos cafetal es(0.418). El bosque tambin tena una mayor representacin al nivel de Familia y de especies raras. Estos resultados tienen implicaciones acerca de cun efectivos son los agroecosistemas para preservar la biodiversidad y si esto es debido a sus dinmica s estructurales o geogrficas. INTRODUCTION The transformation of primary forest into human dominated landscape has profound impacts on species composition. Even the reduction of large tracts of forest into fragments is observed to indirectly reduce th e species richness of the remaining area via edge effects, territory insufficiencies and the invasion of ruderal species (Power 1996). Currently, forests are being
converted to human dominated landscapes at annual rates of 1 4% of their current area (Dobs on et al 1997). The products of this transformation are row crop agriculture and urban areas, now (Vitousek et al 1997). Costa Rica, although a leader in global environmental awareness, is no exception to a rising rate of deforestation, having reduced its forest cover to one quarter of its original extant since 1940 (Inbar 2001). Much of this primary forest has been replaced with pasture, tree farms, and banana, pineapple and coffee plantations. After deforestation takes place, remnant fragments become important refuges for what tropical species remain. Unfortunately, the reduction of viable habitat makes survival unlikely for species with large body sizes, high d egrees of specialization, or unique movement patterns (Bierregaard and Dale). Moreover, fragmentation often eradicates rare species altogether and reduces the effective population size of more common species, making them more prone to extinction (Schelhas and Greenburg). Studies that involve population dynamics, extinctions, and movement of species in and between fragments are based on Island Biogeography Theory and Hubble Unified Theory. These models view fragments as islands or depositories of biodiversit y surrounded oceans, largely uninhabitable by natural populations. Although interesting and useful, these models fail to incorporate agricultural landscapes as habitat to a subset of natural species as well as a number of colonizers. Since agriculture is s uch an integral part of Costa Rican economic activity, it is important to consider these agroecosystems as viable habitat for some species and to determine which types of agroecosystems confer the least impact on local biodiversity. For example, certain st ructural and compositional traits of agroecosystems better emulate natural habitats and therefore may be more conducive to conserving local species, however, this phenomenon is relatively understudied (Perfecto and Snelling, 1995). When compared to large r taxa such as birds and mammals, there have been very few comparisons of arthropod species richness between nearby natural and artificial habitats (Power 1996). Past studies comparing the presence of arthropods in various agricultural settings to that of forests have shown a general trend toward reduced richness in agroecosystems, while differences in diversity between agroecosystms are subtler. Different taxa seem to respond differently to these transformed habitats. A study by Estrada et. all (1998) exam ined Coleoptera richness among disturbed habitats. They found beetle richness to be highest in forest fragments followed then by mixed and cacao plantations, forest edge, coffee, citrus and all spice, and lastly, pasture. Another study assessed the diversi ty of the ant community in several coffee plantations as a function of vegetative complexity and shade provided by the canopy (Perfecto and Snelling, 1995). They found a positive correlation, suggesting that agroecosystems with shade providing plants have higher diversity because they more effectively mimic forests. A comparison of butterfly diversity between shade grown coffee farms and forest fragments in Costa Rica showed a significantly higher diversity in the forest (Inbar, 2001). In contrast, another Costa Rican report found that while forests host the most shade loving, forest dependent species, overall butterfly richness between forest and coffee did not differ (Baptiste, 2001). Finally, a study in Indonesia showed that coconut, rice, corn, and soybe an plantations housed a smaller number of Lepidopterans than nearby primary forest but showed an increase in certain herb feeding families such as Noctuidae and Sphingidae (Holloway et al, 1990). Moths are especially useful for biogeographical or conservat ion studies because many are host specific and hence can be used as reliable indicators for local forest health and plant diversity (Janzen1988). Also, moths play crucial community roles as pollinators, herbivores and prey items (Janzen 1987). Therefore, a change in moth diversity and abundance due to anthropogenic land transformation could have extensive ecological implications. Moreover, deforestation and fragmentation could reduce or nchecked and become detrimental agricultural pests. By understanding which agroecosystems are more
successful in maintaining moth diversity, more effective agricultural methods can be implemented in order to sustain biodiversity as a whole. Ricketts et. al appraised moth richness between four habitats(coffee, shade coffee, pasture and mixed farm) in southern Costa Rica and tested whether richness was affected by distance from remaining forest fragments (2000). No significant differences in richness between habitats were found, however, sites more than 3.5 km from the forest had much lower richness than those within one kilometer of the forest. Three possible reasons are given for the lack of difference between habitats: 1) Moths do not utilizing these habita ts and are merely passersby during between forest flight, 2) the areas used by moths are not directly in the farms per se, but pocketed in small areas of natural vegetation, riparian zone, or fallow pastures homogenously scattered throughout the local agri culture mosaic and 3) the vagility of moths conceals whether or not the moths are using the resources of particular agroecosystems. In continuation of Rickets et al, this study examined moth diversity, abundance, and composition between three transformed habitats in San Luis, Costa Rica: a sun grown coffee farm, a coffee farm shaded with banana trees and a forest fragment. It simplified the 2000 study by examining fewer, but similar agroecosystems, sampling each site more extensively and controlling for ni ghtly environmental variability. Finally, composition and distribution were analyzed and used to determine whether local moth populations are utilizing the two farms equally, differently, or at all, and whether differences are do to structural or geographi cal disparity. MATERIALS AND METHODS STUDY SITES This experiment was conducted in San Luis, Costa Rica during the months of October and November. The lifezone of San Luis is classified as pre montane moist forest and sits at an elevation of 1.300 m (Hayes and Laval 1989). Three sites of capture were chosen, each representing a distinct ecosystem: a 0.5 hectare plot of coffee shaded with banana and other larger trees, a 0.5 hectare plot of sun grown coffee bordered with wind breaks, and a forest pastu re edge located at the southwest tip of the Monteverde Cloud Forest. The former two were set in Finca la Bella, a community of strictly organic farmers, controlling for the various effects of chemical use. Also, in both farms the capture sites were oriente d where they could be seen from all localities of the plot. The forest site was located at the base of a forested hillside, approximately 25 m away from the edge. All three lay essentially in a straight line, with shade coffee and forest sites approximatel y 1 kilometer apart and sun coffee equidistant from the others. Capture sites consisted of 1.5 m x 1.5 m white sheets hanging between two trees, illuminated by a 6 watt black light. The backside of each sheet was obscured by either trees or dense brush in order to prevent attracting moths from other areas. METHODOLOGY Data were collected over 7 evenings from all three sites. Lights were turned on at sundown and returned to after two hours for collection. Moths were picked up one at a time from the sheets instantly killed by squeezing the thorax and placed into a net. No more than 25 minutes were spent at each site regardless of the number of moths on the sheet. The weather and moon phase was also recorded each night
The samples were later separated into morphospecies based on body shape, wing pattern, and antennae. Recent studies have shown that morphospecies richness can be used as an accurate indicator of actual species richness (e.g., Oliver and Beattie 1996). Therefore, for the remainder of this pa per morphospecies will be referred to as species. Very small (less than one centimeter wide), non descript individuals, that were too similar to be accurately distinguished, were clumped into two general morphospecies: grey/blacks and brown/tans. This was done to avoid error and inconsistency while dividing them. Finally, wing color pattern and venation was used to identify as many morphospecies to family as possible. RESULTS A total of 843 individuals were collected representing166 different species an d nine families. A total of 72, 83, and 122 species were found in shaded coffee, sun coffee, and forest, respectively (Fig 1). The Friedman Test showed that the three sites differed significantly in abundance and richness (richness: x 2 =11.214, df= 2, p=0.0 030, abundance: x 2 = 10.333, df= 2, p=0.0057). This test also controlled for abiotic factors such as weather and moon phase affecting the presence of moths between nights. Nights in which there were too many moths to make accurate comparisons were excluded from the Friedman Test. A multiple comparison analysis (Zar 1984) proved both richness and abundance to be significant between shade coffee and forest Shannon Weiner diversity indices upheld this trend, finding species diversity to be highest in forest (H'=4.256), lowest in shade grown coffee (H'=3.413) and intermediate in sun grown coffee (H'=3.863). Using a modified T test(Za r 1984), difference in diversity was found to be significant between all three habitat pairs (Sites A and C: t=3.315, df=346.503, p<0.05. Sites A and B: t= 3.398, df= 395.659, p<0.05. Sites B and C: t=4.175, df=525.714, p<0.05). Lastly, the Jaccard test fo und species turnover to be lowest between shade coffee and forest ( = 0.309), intermediate between sun coffee and forest ( = 0.358), and highest between the two coffee farm ( = 0.418). A cumulative catch curve denotes a trend toward the leveling off of spe cies richness in both coffee farms (Fig 2). The forest site, however, shows less of an asymptote, suggesting that it still had a considerable number of species yet to be collected. 73 of the 166 species were identified to family. All of the nine identified families were represented in the forest site while the other two sites each contained seven families (Fig 3). Dioptidae and Lasioptidae were absent from the sun grown coffee farm while Dioptidae and Sphingidae were absent from the shade grown coffee farm Moreover, only one sphingid individual was found in the sun grown farm. The representation of species in Arctiidae was near twice as great in the forest than in the two plantations. Geometridae and Noctuidae was underrepresented in the forest site compar ed to the two plantations. Sixty of the 166 species were represented by only one individual. 54,44, and 40 rare species were collected from the forest, shade coffee, and sun coffee, respectively (Figs 4,5 and 6). The two groups containing all small, non descript moths accounted for 20% of all collected individuals. They consisted of 36% of the shade coffee sample, 23%
of the sun coffee sample, and only 9% of the forest sample. The forest site contained considerably fewer of these drab individuals than th e other two sites, however, it contained several more species with relatively large numbers. Overall Shannon Weiner evenness was highest in forest (E=0.886) followed by coffee (E=0.874) and then coffee banana (E=0.802). This distribution of moths was more similar between the two coffee farms than between the forest and either coffee farm. DISCUSSION Overall trends suggest that more species of moths preferred intact forest to either of the two agroecosystems. The deficiency of species in the coffee farms m ay be due to several factors, none being mutually exclusive. First, the female moth is very particular about which substrate to lay her eggs on, probably sensing certain chemicals in the host plant that will be important to the developing caterpillars (Cov ell 1984). This high host specificity would make finding an agreeable substrate in a coffee farm unlikely. Next, highly specialized pollination syndromes have evolved between many species of moths and plants, for example, between Sphingidae and Meliacea, R ubiacea, and Orchidacea (Haber 1983). Therefore, the habitat of many moth species is probably restricted to that of their nectar giving mutualist. The moths that can establish and meet their living needs in an uniform habitat like a coffee farm will likely flourish and dominate, in turn reducing diversity even further (Root 1973). Furthermore, increases in wind and rain impede profoundly affect moth numbers. Finally, the o penness of farms may increase the exposure of moths to avian predators such as birds and bats. Interestingly, diversity, richness and abundance were all lower in shade grown coffee than sun grown coffee, contradicting the findings of Rickets et al (2000) t hat richness does not significantly differ between agroecosystems at similar distances from the forest. It also rebuts the theory that shade and vegetative complexity confers higher richness canopy (Perfecto and Snelling, 1995). In fact, H' of sun grown co ffee was truly intermediate, resembling shade grown coffee no more closely than forest. These findings might be explained by the geography of the area. While Rickets et al (2000) found richness declines at distances greater that 3.5 km from the forest; the coffee farms in this study may have been more sensitive to distance. While the sun grown coffee farm lay approximately 0.5 km from the forest, the shaded coffee farm was nearly twice that distance away. This extra 0.5 km may have been enough distance to p revent many species from arriving in the shaded coffee farm. Another possibility for sun grown coffee housing more species is that different localities in the agricultural mosaic around a forest act as sinks for species, but rely heavily on a continuous fl ow of individuals from the forest to avoid local extinction (Rickets et al). If this is the case, according to island biogeography theory, farms farther from the forest should exhibit smaller species richness at equilibrium (Wilson and McArthur 1967). The agroecosystems seemed to sustain certain moth families more effectively than others (Fig 3). The lack of sphingids and arctiidae in the two farms is striking. It seems that this deficiency is largely compensated for by an abundance of geometrids and noctu ids. No discerning morphological or ecological traits of these families seem to explain these trends. In fact, my results opposed the findings of past studies. Holloway et al (1990) found that Sphingidae is actually more abundant in soybean, coconut, rice, and
corn plantations while Rickets et al (2000) captured significantly higher numbers of Geometeridae in the forest. The latter study, however, did find less noctuids in the forest The two farms showed a greater proportion of small, nondescript moths than the forest, particularly in the shade grown coffee (Figs 5 and 6). It is possible that many of these moths are generalists that thrive in disturbed areas, having been liberated fr om oppressive competition and predation imposed by the forest (Root 1973). Dissimilarly, the forest edge attracted more rare species (species with only one individual in sample) (Fig 4). This was to be expected because tropical forests are generally charac terized by the existence of many specialized and uncommon species. Moreover, the cumulative catch curve suggests that there was still a considerable number of rare species in the forest that had not yet been collected (Fig 2). Species turnover was lowest b etween shade coffee and forest ( = 0.309), intermediate between sun coffee and forest ( = 0.358), and highest between the two coffee farms ( = 0.418). This overlap is strikingly low compared to that of Rickets et al (2000), and indicates that three sites a re providing very different resources to facilitate distinct communities. It also indicates that there are many species, possibly colonizers, found exclusively in these agroecosystems. The neighboring sites shared more of the same species than the two dist ant sites, implying that the species were loosely confined to a certain area or chose not to fly far from a their general locality. It is important to point out that the sun coffee, although roughly equidistant from the other two sites, had considerably mo re species overlap with the shade coffee than the forest. This suggests that factors besides distance are affecting species composition between sites and that the environments of the two farms support relatively alike communities. It must be considered, ho wever, that low overlap was possibly an artifact of incomplete sampling. In conclusion, this study upholds the findings of Rickets et al(2000) that forest harbors greater moth diversity than agroecosystems. However, compared to the former study, these resu lts showed much greater difference in diversity, richness, and abundance between agroecosytems. It seems that distance from a forest plays a fundamental role in maintaining local biodiversity, in this case more so than vegetative complexity. Still, much wa s left unknown about whether moths are actually utilizing agroecosystems, what the exact degree to which distance from forest affects diversity. A more extensive st udy could provide further knowledge on the biogeography and movement of moths, habitat requirements and host specificity, and the regulation of agricultural pests. ACKNOWLEDGEMENTS ing a piece thanks for the company during the cold, wet evenings of moth carnage. You know how to tie a lot of cool knots and for that I hold you in the highest esteem. Mr. Hy thanks for the guidance. Richard Whitten, you were incredibly helpful with the identification process and your moth guides were of great value. A very special thanks to my host mother Xinia Leiton for letting me keep dead moths in the freezer. Lastly, thanks to the families in Finca la Bella for opening your beautiful farms to me. I love you all.
LITERATURE CITED Baptiste, Anne. 2001. Agricultura l Land Use and Butterfly Communities. CIEE Fall 2001. Monteverde, Costa Rica. Estrada A. 1998. Dung and Carrion Beetles in Tropical Rainforest Fragments and Agricultural Habitats in Los Tuxtlas, Mexico. Journal of Tropical Ecology 14 :577 593 Hayes, M.,L aval R. 1989 The Mammals of Monteverde Tropical Science Center. San Jose, Costa Rica. Hayber W.A. 1983. Checklist of Sphingidae. In Janzen D.H. Natural Costa Rican History p.646. The University of Chicago Press. Chicago, Illinois. Aspects of the biogeography and ecology of the Seram moth The Natural History of Seram, edited b y I. Edwards and J. Proctor, pp. 37 62. Andover, MD: Intercept. Inbar M. 2001. Butterfly diversity and Abundance in Sun Grown Coffee Plantations ver ses Forest Fragment. CIEE Summer 2001. Monteverde, Costa Rica. Janzen, D.H. 1973. Sweep samples of tropical foliage insets: effects of seaons, vegetation types, elevation, time of day, and insularity. Ecology 54: 687 709. Janzen D.H. Natural Costa Rican History pp.620 622. The University of Chicago Press. Chicago, Illinois. Janzen, D.H. 1988. Ecological characterization of a Costa Rican dry forest caterpillar fauna. Biotropica 20:120 135. Oliver, I., and A.J. Beattie.1996. Invertebrate morpospecies as surrogates for species: a case study. Conservation Biology, 10 :99 109. Perfecto I., Snelling R. 1995. Biodiversity and Transformation of Tropical Agroecosystem: Ants in Coffee Plantations. Ecological Applications. 5 : 1084 1097 Power,A.G. 1996. Arthropod Diversity in Forest Patches and Agroecosystems of Tropical Landscapes. Forest Patches in Tropical Landscapes Eds. J. Schelhas and R. Greenburg. Pp. 91 102. Island Press. Washington, DC. Rickets T.H. et al. 2001. Countryside Diversity in Forest Patches a nd Agroecosystems of Tropical Landscapes. Conservation Biology 15 : 378 388
Root, R.B. 1973. Organization of a plant arthropod association in simple and diverse habitats: the fauna and collards( Brassica oleracea ) Ecol. Monogr. 43: 95 124. Zar, J.H. 1984. Biostatistical Analysis Prentice Hall, Englewood Cliffs, New Jersey. Pp.199 and 229. FIGURES Figure 1. Number of morphospecies collected at each site over all seven nights. Richness is consistently and significantly d ifferent between sites. Richness fluctuates similarly for all sites between nights.
Figure 2. Cumulative catch curve for all three sites. The richness in the forest (yellow) rises more quickly than the sun coffee (pink) or sh ade coffee (blue). The curves for both farms level of sooner than that of the forest. Figure 3. Family composition of total moth sample collected in San Luis, Costa Rica. Nine families were represented in the forest while only s even were found in each coffee farm.
Figure 4. Frequency distribution of number of species vs. number of individuals at the forest site in San Luis, Costa Rica.* Small, non descript individuals were clustered into these two grou ps (see methods). Figure 5. Frequency distribution of number of species vs. number of individuals at the sun grown coffee site in San Luis, Costa Rica.* Small, non descript individuals were clustered into these two groups (see methods).
Figure 6. Frequency distribution of number of species vs. number of individuals at the shade grown coffee site in San Luis, Costa Rica.* Small, non descript individuals were clustered into these two groups (see methods ). Astericks
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xlink http:www.w3.org1999xlink xmlns:xsi http:www.w3.org2001XMLSchema-instance
leader 00000nas 2200000Ka 4500
controlfield tag 008 000000c19749999pautr p s 0 0eng d
datafield ind1 8 ind2 024
subfield code a M39-00369
Efectos de los agroecosistemas de caf en la diversidad de mariposas en San Luis, Costa Rica
Effects of coffee agroecosystems on moth diversity in San Luis, Costa Rica
The majority of Costa Ricas primary forests have been fragmented and converted to agricultural landscape. Fragmentation studies based on Island Biogeography Theory and Hubble Unified Theory typically view the degraded area surrounding forests as uninhabitable, however, most agroecosystems
actually serve as viable habitat to a subset of natural communities as well as a number of colonizing species. As a follow up to Rickets et al (2000) who studied moths in a tropical, agricultural landscape, this study focuses on how moth diversity is affected by coffee plantations in San Luis, Costa Rica. Moths were collected from three sites: a forest, a sun-grown coffee farm and a coffee farm grown shaded banana trees. The samples were then divided into morphospecies and many were identified to family. Shannon-Weiner diversity indices found species diversity to be highest in forest (H'=4.256), lowest in shade-grown coffee (H'=3.413) and intermediate in sun-grown coffee (H'=3.863). Multiple comparisons determined all three
sites significantly different in diversity. Beta diversity was highest between forest and shade-grown coffee(0.309), intermediate between forest and sun-grown coffee(0.358), and lowest between the two coffee farms (0.418). The forest also had a larger representation at the family level along with rarer species. These results give many implications as to how effectively agroecosystems harbor biodiversity and
whether this is due to structural or geographical dynamics.
La mayora de los bosques primarios de Costa Rica han sido convertidos en tierras agrcolas. Los estudios sobre fragmentacin basados en la teora de biogeografa de islas y en la teora unificada de Hubble presentan a la zona degradada alrededor de estos bosques como inhabitable. Sin embargo, la mayora de los agroecosistemas en realidad sirven como hbitats viables para un subgrupo tanto de comunidades naturales as como para cierto nmero de especies colonizadoras. Este estudio determin cmo la diversidad de polillas en tierras agrcolas tropicales es afectada por los cafetales en San Luis, Costa Rica, como parte de una continuacin del estudio de Rickets et al (2000). Las polillas fueron recolectadas en tres lugares diferentes: Un bosque, un cafetal sin sombra y un cafetal con plantas de banano. Las muestras fueron clasificadas por especies morfolgicas y muchas fueron identificadas al nivel de Familia. Los ndices de diversidad de Shannon Weiner demostraron que la diversidad de especies fue mayor en el bosque (H'=4.256), menor en el cafetal con bananos (H'=3.413) e intermedia en el cafetal sin sombra (H'=3.863). Las comparaciones mltiples mostraron que los tres sitios diferan significativamente en diversidad. La diversidad beta fue mayor entre el bosque y el cafetal con sombra (0.309), intermedia entre el bosque y el cafetal sin sombra (0.358) y menor entre los dos cafetales (0.418). El bosque tambin tena una mayor representacin al nivel de Familia y de especies raras. Estos resultados tienen implicaciones acerca de cun efectivos son los agroecosistemas para preservar la biodiversidad y si esto es debido a sus dinmicas estructurales o geogrficas.
Text in English.
Costa Rica--Puntarenas--Monteverde Zone--San Luis
Plantaciones de caf
Fragmentacin del hbitat
Costa Rica--Puntarenas--Zona de Monteverde--San Luis
Tropical Ecology Fall 2004
Ecologa Tropical Otoo 2004
t Monteverde Institute : Tropical Ecology