The relationship between microhabitat and abundance of amphibians in the highlands of the Andes in Ecuador AbstractIn ecology, different scales have been implemented to study the relationship that exists between organisms and their environment. Microhabitat is the smallest one of these scales and can influence the distribution of organisms, like amphibians. In this study, I try to test what relationship exists between the abundance of amphibians and the microhabitats present in the high Andean areas of southern Ecuador. I use a dataset of the biodiversity of amphibians and classification of seven microhabitats (bunchgrasses, marshes, shrubs, water, rocks, intervention and forest), in six river basins. The relationship between abundance and microhabitat was calculated using a quasi-Poisson model. In total, 187 individuals and six species of amphibians were recorded. The microhabitat of marshes and bunchgrasses showed to be significant in the abundance of amphibians. The presence of these two microhabitats increases the number of amphibian individuals. This result can help in the elaboration of conservation plans focused on the protection of microhabitats that are being used by amphibians and thus guarantee their survival.IntroductionGlobally, ecosystems have been studied mostly on a large scale, as a single landscape. However, biodiversity and environmental variables in an ecosystem are heterogeneous. Several studies have shown that groups of fauna will not occupy the whole ecosystem as their habitat throughout its life, but only use a small scale of the ecosystem, depending on the environmental conditions that it has [1] [2] [3] [4] [5]. This is how Morris in 1987 defined the term microhabitat as “A discrete fraction of the habitat, with specific niche requirements, directly or indirectly influencing the distribution of species”. Thus, began to implement scales in ecology to study an ecosystem and its biodiversity.It is considered that there is a hierarchy in the selection of a microhabitat, where the organism first chooses a place to live and then decides about the use of other sectors [6]. As a result, individuals choose places where they maximize their survival and reproductive capacity [7]. In the case of amphibians, the decision to occupy a microhabitat would be linked to its physiology. The permeability of the skin and eggs of amphibians makes them able to absorb elements of the environment [8]; which also makes them organisms very sensitive to changes in temperature and humidity [9]. This sensibility in conjunction with the low dispersion capacity of amphibians limits the colonization of new habitats and causes several species of this group to have high specificity and fidelity to a microhabitat.In the tropics of South America, studies of the relationship between amphibian diversity and the microhabitat that they occupy have generally focused on areas of low montane forest [10] [11] [12]. The high Andean regions (> 3000 m a.s.l.) present a lack of studies in this aspect of ecology, probably due to the low richness of amphibians that occur at this altitude. In this work, I present the relationship between seven microhabitats and the abundance of amphibians in high Andean areas of Ecuador.MethodsData CollectionThe data was provided by the Biology Laboratory of the University of Azuay and its project “Characterization of the biodiversity of plants, birds and amphibians of the high Andes of the province of Azuay, Ecuador”.In this project, 6 river basins were used as sampling units. Within each river basin, information was obtained on the richness (number of species) and abundance (number of individuals per species) of amphibians. Characterization of microhabitats present in each river basin was also carried out. This classification was made based on the vegetation present in the study area, modifying a previous classification made by the staff of the Missouri Botanical Garden [13]. A total of seven microhabitats were identified: bunchgrasses, shrubs, marshes, forests of paper tree (Polylepis sp.), water (rivers and ponds), rocks and intervention (bare soil). The data were obtained during three surveys covering the dry and rainy season of 2018.Statistical AnalysisTo test if there is a relationship between the type of microhabitat and the abundance of amphibians I run a quasi-Poisson GLM to account for overdispersion, so the dispersion parameter is not fixed at 1. In the structure of the model, I use abundance of each amphibians species by river basin as the response variable and the percentage of each of the seven types of microhabitats present in each river basin as explanatory variables. I report results as statistically significant if p equals or is smaller than 0.05 I use a Chi-squared test to estimate the goodness-of-fit of the model. Because the quasi-Poisson model returns an NA value for log-likelihood, I followed the recommendation of Bolker (2017) where I extract the log-likelihood of a Poisson model of the data and then continue my analysis with the select model using a quasi-Poisson GLM. I use R version 3.1.5 [14] for all analysis and plotting.ResultsThe final dataset included a total abundance of 187 individuals of amphibians grouped into six species: Gastrotheca pseustes, Pristimantis riveti, Pristimantis cryophilius, Pristimantis sp. 1, Pristimantis sp. 2 and Pristimantis sp.3. Due to taxonomic doubts, three of the six species could only be identified up to genus level.The log-likelihood test reveals that a simpler model of Shurbs+Forest+Marshes+Bunchgrasses (AIC: 180.67, df: 5) is better than a full model using the seven microhabitats (AIC:182.67, df:6).  The result of the quasi-Poisson model shows that two microhabitats, marshes and bunchgrasses, influence amphibian abundance (Table 1). The effect size indicates that for every 1% of bunchgrasses that increases, the number of amphibian individuals is increased by 1.19. In the case of marshes, for every 1% that this microhabitat increases, the number of individuals is increased by 1.14.Table 1. Results of the quasi-Poisson model between abundance and four microhabitats in high Andes of EcuadorVariableEstimatePr (> | t |)(Intercept)-8.9940.087Forest-0.1880.311Shrubs0.1060.087Marshes0.1080.010Bunchgrasses0.1760.050 DiscussionsThe abundance of amphibians in the Andean highlands of Ecuador seems to be related to the microhabitats present within the ecosystem, where the areas of bunchgrasses and marshes can positively influence the increase of amphibian individuals.The high Andean zones are characterized by abiotic variables that define the development of the life forms that inhabit these ecosystems. Low atmospheric pressure, desiccant winds, clarity of the atmosphere, high levels of radiation and the perpendicularity of the sun’s rays are the main characteristics found in the Andes above 3000 m a.s.l. [15]. Both plant and animal species of the high Andes have developed unique adaptations that allow them to withstand these abiotic variables.Bunchgrasses is a plant form typical and the most highly distributed of the high Andes. This could explain that it has a high significance in the abundance of amphibians. Being a widely distributed microhabitat, bunchgrasses provide a greater opportunity for refuge for the amphibians that inhabit the Andes. In addition, bunchgrasses are usually plant formations that can reach up to 1.5 m. This height, allows the bunchgrasses to develop adaptation strategies to withstand extremely cold temperatures, creating microclimates around them that prevent their freezing at night [16]. The amphibians could take advantage of this microclimate to ensure their survival also on nights with temperatures below zero.One aspect that organisms use to choose a microhabitat to survive is their reproductive mode [17]. Five of the six registered species belong to the genus Pristimantis. This genus of amphibians, during the reproductive cycle, deposit the eggs in the soil, under grasses or leaf litter, and its development is direct, that is, there is no tadpole stage [18] [8]. Due to this characteristic, these species are not linked to water bodies. However, although this genus of amphibians does not depend on rivers or ponds to reproduce, if they need high levels of humidity to ensure the survival of their eggs [19]. Marshes provide a microhabitat with high levels of humidity and low vegetation, which would facilitate the birth of the individuals and their survival until the adult stage.Forests and shrubs were not significant during our analysis. In the Andes, tree formations tend to appear scattered and very isolated [20]. In Ecuador, these arboreal formations called patches are usually made up of trees of the genus Polylepis. Although our analysis does not show a relationship with amphibian abundance, Polylepis patches can have a barrier effect against abiotic variables, as was shown in studies in the high Andes of Colombia [21].  The lack of use of this microhabitat by amphibians may not only be linked to the low presence of forest in the Andes but also to the morphology of the amphibian species present in this place. All amphibian species recorded in this work, except for G. pseustes, belong to a group called Terrarana. This group of amphibians has begun to exploit more terrestrial habits, including the ability to “run” instead of jumping at any threat [22]. Therefore, they are more adapted to live at ground level and cannot obtain many advantages when climbing trees.Bibliography[1]A. Blanco, Reparticion de microhabitats y recursos troficos entre especies de Bufonidae y Leiuperidae (Amphibia:Anura) en áreas de Bosque Seco tropical de la región Caribe-Colombia, Bogotá: Universidad Nacional de Colombia, 2009.[2]G. Amat and O. Vargas, “Caracterización de microhábitats de la artropofauna en páramos del Parque Nacional Natural Chingaza Cundinamarca, Colombia,” Caldasia, vol. 79, no. 16, pp. 539-550, 1991.[3]T. Martin, “Are microhabitat preferences of coexisting species under selection and adaptive?,” Ecology, vol. 2, no. 79, pp. 656-670, 1998.[4]N. Clemann, J. Merville, N. Ananjeva, M. Scroggie, K. Milto and E. Kreuzberg, “Microhabitat occupation and functional morphology of four species of sympatric agamid lizards in the Kyzylkum Desert, central Uzbekistan,” Animal Biodiversity and Conservation, vol. 2, no. 31, pp. 51-62, 2008.[5]E. Jorgensen, “Small mammal use of microhabitat reviewed,” Journal of Mammalogy, vol. 3, no. 85, pp. 531-539, 2004.[6]A. Díaz de Pascual, “Caracterización del hábitat de algunas especies de pequeños mamíferos de la selva nublada de Monte Zerpa, Mérida,” Ecotropicos, vol. 1, no. 6, pp. 1-9, 1993.[7]J. Brown, “Habitat selection as an evolutionary game,” Evolution, vol. 3, no. 44, pp. 732-746, 1990.[8]W. Duellman and L. Trueb, Biology of Amphibians, Nwe York: McGraw Hill, 1986.[9]J. Méndez-Narváez, “Diversidad de anfibios y reptiles en hábitats altoandinos y paramunos de la cuenca del río Fúquene, Cundinamarca, Colombia,” Biota Colombiana, vol. 15, no. 1, pp. 94-103, 2014.[10]K. Cruz, “Diversidad y preferencia de microhábitats de la herpetofauna del bosque protector “Pedro Franco Dávila” (Jauneche) y del área provincial natural de recreación “Cerro de Hayas” (Naranjal),” Universidad de Guayaquil, Guayaquil, 2017.[11]F. Vargas and F. Castro, “Distribución y preferecia de microhábitat en anuros (Amphibia) en bosque maduro y áreas perturbadas en Anchicayá, Pacífico colombiano,” Caldasia, vol. 21, no. 1, pp. 95-109, 1999.[12]C. García, F. Castro and H. Cardenas, “Relación entre la distribución de anuros y variables del hábitat en el sector La Romelia del Parque Nacional Natural Munchique (Cauca, Colombia),” Caldasia, vol. 2, no. 27, pp. 299-310, 2005.[13]J. Luteyn, “WeTropicos,” 2018. [Online]. Available: http://www.mobot.org/MOBOT/research/paramo_ecosystem/introduction.shtml. [Accessed 15 February 2019].[14]R Core Team, “R: A language and environment for statistical computing,” R Foundation for Statistical Computing, Vienna, Austria, 2018.[15]I. Hedberg and O. Hedberg, “Tropical-alpine lifeforms of vascular plants,” Oikos, pp. 297-307, 1979.[16]D. Serrano and R. Galárraga, “El páramo andino: características territoriales y estado ambiental. Aportes interdisciplinarios para su conocimiento,” Estudios Geográficos, vol. LXXVI, no. 278, pp. 369-393, 2015.[17]M. Crump, “Reproductive strategies in a tropical anuran community,” Misc.Publ.Mus.Nat.Hist.Univ.Kansas, vol. 6, no. 1, pp. 1-68, 1974.[18]W. Duellman, “The biology of an equatorial herpetofauna in Amazonian Ecuador,” Misc.Publ.Mus.Nat.Hist.Kansas, vol. 65, pp. 1-352, 1978.[19]E. Pineda and G. Halffter, “Species diversity and habitat fragmentation: frogs in a tropical montane landscape in Mexico,” Biological Conservation, vol. 117, no. 5, pp. 499-508, 2004.[20]C. Josse, F. Cuesta, G. Navarro, V. Barrena, E. Cabrera and E. Chacón-Morena, Ecosistemas de los Andes del norte y centro. Bolivia, Colombia, Ecuador, Perú y Venezuela, 1st ed., Lima: Secretaría General de la Comunidad Andina, 2009.[21]S. Buitrago, K. Pulido and L. Vanegas, “Environmental variability and physiological responses from Polylepis cuadrija (Rosaceae) in fragmented environment in the paramo de la Rusia (Colombia),” Revista de biología tropical, vol. 61, no. 1, pp. 351-361, 2013.[22]S. Hedges, W. Duellman and M. Heinicke, “New World direct-developing frogs (Anuran:Terrarana): Molecular phylogeny, classification, biogeography, and conservation,” Zootaxa, vol. 1737, no. 1, pp. 1-182, 2008.[23]B. Bolker, “The Comprehensive R Archive Network,” 29 October 2017. [Online]. Available: https://cran.r-project.org/web/packages/bbmle/vignettes/quasi.pdf. [Accessed 16 February 2019].[24]D. Morris, “Ecological scale and habitat use,” Ecology, vol. 68, no. 2, pp. 362-369, 1987.