Habitat Quality, Not Patch Size, Modulates Lizard Responses to Habitat Loss and Fragmentation in the Southwestern Amazon
As deforestation frontiers expand in the tropics, understanding species responses is critical to inform efficient land-use management policies. Here we evaluated lizard responses to changes in landscape, habitat patch, and quality in a deforestation frontier in southwestern Brazilian Amazonia. We first considered overall lizard assemblages and then two species subgroups of contrasting thermoregulation strategies. We sampled lizards at 21 forest patches (42–7,035 ha) subject to moderate habitat disturbance (e.g., cattle intrusion and selective logging). Based on 6,000 pitfall trap-days and 60,000 m of visual surveys, we recorded 215 individuals representing 15 species. Contrary to expectations, species richness and abundance were not explained by any of the variables considered, including matrix quality, proportion of forest cover, forest patch area and shape, age since isolation, arthropod biomass, canopy openness, litter volume, tree density, and fire history. However, the composition of persisting species was determined by the structural complexity of the adjacent matrix. The abundance of species that avoid direct sun exposure to regulate their body temperature increased under more-complex structures of adjacent matrix areas and decreased within patches more intensively affected by past fire events. Disturbance within forest patches likely decreased the patch-matrix contrast, contributing to the absence of forest area effects. Nevertheless, species-specific microhabitat requirements dictated part of the lizard species that were able to persist in the fragmented landscape. Strategies aiming to maximize lizard diversity should embrace retention of the habitat quality within forest patches and forest corridors/high-stature vegetation in at least some sites connecting forest remnants. À medida que as fronteiras de desmatamento se expandem nos trópicos, entender a resposta das espécies torna-se crítico para tomar medidas eficientes do manejo do uso do solo. Neste estudo, avaliamos as respostas da comunidade de lagartos a alterações ao nível da paisagem, fragmento e qualidade do habitat, numa fronteira de desmatamento no sudoeste da Amazônia brasileira. Primeiro consideramos toda a comunidade de lagartos e depois dois sub-grupos de espécies com comportamentos de termorregulação contrastantes. Amostramos os lagartos em 21 fragmentos (42 – 7,035 ha) sujeitos a níveis moderados de perturbação (por exemplo, invasão pelo gado e corte seletivo de madeira), utilizando armadilhas tipo-queda e busca ativa. No total, amostramos 6.000 armadilhas/dia e 60.000 m percorridos, respectivamente. Com base neste esforço, capturamos 215 indivíduos representando 15 espécies. Ao contrário do que esperávamos, nem a riqueza nem a abundância de espécies foram explicadas pelas das variáveis consideradas, incluindo qualidade da matriz, proporção de cobertura florestal, área e forma do fragmento florestal, tempo desde o isolamento, grau de abertura do dossel, voluma da liteira, densidade de árvores e histórico de fogos. No entanto, a composição de espécies foi determinada pela complexidade da matriz ao redor dos fragmentos. A abundância das espécies que evitam a exposição direta à luz solar para regular a sua temperatura corporal aumentou em fragmentos cercados por matrizes estruturalmente mais complexas e diminuiu em fragmentos anteriormente sujeitos a maior intensidade ao fogo. A perturbação dos fragmentos, ao diminuir o contraste estrutural com a matriz, provavelmente contribui para o fato de não terem sido observados quaisquer efeitos da área dos fragmentos. Ainda assim, a persistência de lagartos foi determinada por requerimentos espécie-específicos ao nível do micro-habitat. Estratégias que visem maximizar a diversidade de lagartos devem passar pela retenção da qualidade do habitat dos fragmentos, bem como de corredores/maior complexidade da estrutura da vegetação conectando remanescentes florestais.ABSTRACT
RESUMO
Location of the 21 forest patches surveyed within the Jaurú and Cabaçal river basins, Mato Grosso, southern Brazilian Amazon. Patches surveyed are numbered according to Supplementary Table S1. Inset map shows the location of the study area (red rectangle) within the Brazilian Amazon biome. The sampling site inset illustrates the sampling design used at each forest patch in terms of pitfall trapping and intensive visual searches. Distances between traps are indicated in the figure. See main text for a detailed description of the sampling design.
Nonmetric Multidimensional Scaling (NMDS) ordination plot based on the Bray-Curtis similarity matrix of lizard species composition representing both 21 surveyed forest patches and 15 recorded lizard species. Site symbols are sized proportionally to forest patch size. Species names are colored according to species thermoregulation strategy: heliophile (in red) and heliophobe (in green; for complete species names, see Table 1).
Estimates of averaged models and their 95% confident intervals for predictors of (A) species richness, (B) species abundance (log10x), (C) species composition (denoted by the NMDS axis 1), and (D) abundance of heliophile and (E) heliophobe individuals (log10x). Predictors included matrix complexity (matrix), forest cover (cover), years since patch isolation (age), patch shape (shape), litter depth (litter), tree density (trees), arthropod biomass (arthropod), canopy openness (canopy), and presence of fire (fire). Statistically significant negative and positive coefficients are color-coded in red and blue, respectively.
Relationships between the species abundance of heliophile (red circles) and heliophobe individuals (green circles; log10x) and degree of vegetation complexity in the matrix surrounding surveyed forest patches. Line represents the adjusted model and shaded area represents the 95% confidence region. P-values are indicated for the strongest relationships (i.e., P < 0.05).
Contributor Notes
