Effect of Inter-Habitat Matrix on Tropical Evergreen Forest Remnants : An Emprical Test of Matris-Tolerrance Hypothesis on Butterflies

Abstract

Recent studies on habitat fragmentation has highlighted the importance of inter-habitat matrix and the need for a shift from traditional binary perspective to a 'matrix composition' ·perspective for understanding species diversity patterns in human modified landscapes . Matrix-tolerance model provides a simple framework for understanding the sensitivity of species to fragmentation by simultaneous study of native forest patches and surrounding matrix. According to this model, abundance of a species in the matrix is inversely proportional to its vulnerability to fragmentation. This study tested the matrix-tolerance hypothesis on butterflies of tropical evergreen forest remnants, surrounded by a matrix of coffee plantation and paddy field in Kodagu region of Western Ghats. Western Ghats is rich in butterfly fauna represented by 333 species of which 33 are endemic to the biogeographic region. The region is also characteristic in having human modified habitats with relatively large proportion of native forest patches which calls for a landscape level approach for conservation. In order to test the matrix tolerance hypothesis sampling was carried out in 16 sites which included 7 in fragments (sacred groves), 7 in matrix (5 in coffee plantations, 2 in paddy field) and 2 in large contiguous forest (Reserve Forest). For abundance estimation of butterflies, time constrained surveys were carried out 3 times in each site between January 10 and April 8, 2013. An index related to fragmentation vulnerability was generated for each species using abundance data obtained from time-constrained surveys. The Fragmentation Vulnerability Index, FVI was defined as the ratio between average abundance of the species in the fragments and average abundance in the continuous patch. Spearma'n's rank correlation was then used to test the association between rank FVI and rank matrix abundance. Various microclimatic and habitat variables were measured in matrix sites and Generalized Linear Models (GLMs) were constructed to test their association with estimated species richness. Analyses were carried out using SPSS 16 and R (version 3.0.1, R core team 2013) software. The present study shows that the relation between abundance of species in the inter-habitat matrix and its vulnerability to fragmentation is positive and significant (rs= 0.514, p = 0.001, N = 37). Species which are capable of tolerating the matrix are found to be less vulnerable to fragmentation. Some species deviate from the expected model in their response to habitat fragmentation which could possibly be explained by examining species specific traits. Fragment size didn't have a significant effect on relation between matrix abundance and fragmentation vulnerability index (F=0.758, p=0.387, df=l), unlike expected from previous studies. This could either be due to smaller range of fragment sizes in the present study or due to confounding factors like patch isolation and patch quality which can mask the effect of patch size. Among various habitat 'and microclimatic variables, canopy cover was found to be 'important predictor of ciliated species richness in matrix sites (weight = 0.8, ~ = 0.8, P = 0.008). Coffee plantations generally have a moderate level of canopy cover which is favorable for most butterfly species. But compared to paddy field~ the canopy cover is higher in coffee plantations and this could be the reason for increase in species richness with increasing canopy cover in matrix sites.