Something is not quite right: effects of two land uses on anuran diversity in subtropical grasslands

Although habitat modification is considered one of the main
causes of biodiversity loss, the relative contribution of different rural
land uses to biodiversity conservation are far less known. Additionally,
the realization of the multidimensionality of biodiversity demands studies
integrating variation of functional traits and phylogenetic information as
complements to address the effects of land use on the structure of animal
communities. Herein, we investigated the effects of land use (i.e.,
intensive agricultural and extensive livestock rearing) on functional and
phylogenetic diversity of anuran communities in farmland ponds from the
Uruguayan savanna ecoregion, while considering the effects of local factors
(i.e., water depth) on species composition. We surveyed adults and tadpoles
in 22 ponds and quantified five traits related to tadpole feeding, habitat
use, and predator avoidance. Tadpole identification was corroborated by DNA
barcoding based on a fragment of the mitochondrial 16S rRNA gene. We
observed a decline in phylogenetic mean nearest taxon distance associated
with increase of surrounding agricultural land use. While land use
intensification did not affect richness (functional or phylogenetic), ponds
in livestock ranches hosted about four times more tadpoles than
agricultural ponds. Functional evenness decreased with water depth,
although such relationship disappeared when considering phylogenetic
non-independence. Our results indicated that specific anuran clades were
more sensitive to intensification in land use, reinforcing a recent view of
phylogenetic homogenization following habitat conversion. Additionally, our
study suggests that extensive cattle grazing over wide native pastures may
provide an alternative more compatible with conservation than short-term
crops in subtropical grasslands.
Methods
In the spring of 2015 (10–16 October), we sampled tadpoles in 22 ponds located 5.5–316 km apart in southern Brazil. In order to ensure that the landscape composition surrounding each pond did not change over different scales, we defined circular areas (1,000 m radius) and measured land-cover data , using Qgis 2.18.16 . We based analysis on Google Earth imagery, using a land-cover classification for 2015 produced within the scope of the MapBiomas initiative. At each pond, we performed eight dip-net sweeps (30 cm diameter, 2mm mesh), each covering approximately 1 m2 and distributed at different pond depths (i.e., four sweeps near the edge of the pond and four sweeps in deep water near the centre of the pond). Sweeps were pooled into one sample per pond (3.5 L plastic bucket), where the collected tadpoles were euthanized with a benzocaine solution and sorted into series based on their morphology (i.e. coloration, body shape, eyes position, tail fin proportion, and mouth position).  We also registered the presence of all observed adults, visually or acoustically, during tadpole sampling. In laboratory, each tadpole without tail clip was rinsed in distilled water and then measured for a set of eight morphological traits on an automated stereomicroscope – Leica M205A.