Summary A majority of environmental studies describe microbiomes at coarse scales of taxonomic resolution (bacterial community, phylum), ignoring key ecological knowledge gained from finer‐scales and microbial indicator taxa. Here, we characterized the distribution of 940 bacterial taxa from 41 streams along an urbanization gradient (0%–83% developed watershed area) in the Raleigh‐Durham area of North Carolina (USA). Using statistical approaches derived from macro‐organismal ecology, we found that more bacterial taxa were classified as intolerant than as tolerant to increasing watershed urbanization (143 vs 48 OTUs), and we identified a threshold of 12.1% developed watershed area beyond which the majority of intolerant taxa were lost from streams. Two bacterial families strongly decreased with urbanization: Acidobacteriaceae (Acidobacteria) and Xanthobacteraceae (Alphaproteobacteria). Tolerant taxa were broadly distributed throughout the bacterial phylogeny, with members of the Comamonadaceae family (Betaproteobacteria) presenting the highest number of tolerant taxa. Shifts in microbial community structure were strongly correlated with a stream biotic index, based on macroinvertebrate composition, suggesting that microbial assemblages could be used to establish biotic criteria for monitoring aquatic ecosystems. In addition, our study shows that classic methods in community ecology can be applied to microbiome datasets to identify reliable microbial indicator taxa and determine the environmental constraints on individual taxa distributions along environmental gradients.

When a forest is fragmented, this increases the amount of forest edge relative to the interior. Edge effects can lead to loss of animal and plant species and decreased plant biomass near forest edges. We examined the influence of an anthropogenic forest edge comprising cattle pasture, coconut plantations, and human settlement on the mantled howler (Alouatta palliata), white-faced capuchin (Cebus capucinus), Central American spider monkey (Ateles geoffroyi), and plant populations at La Suerte Biological Research Station (LSBRS), Costa Rica. We predicted that there would be lower monkey encounter rate, mean tree species richness, and diameter at breast height (DBH) in forest edge versus interior, and that monkeys would show species-specific responses to edge based on diet, body size, and canopy height preferences. Specifically, we predicted that howler monkeys would show positive or neutral edge effects due to their flexible folivorous diet, large body size, and preference for high canopy, capuchins would show positive edge effects due to their diverse diet, small body size, and preference for low to middle canopy, and spider monkeys would show negative edge effects due their reliance on ripe fruit, large body size, and preference for high upper canopy. We conducted population and vegetation surveys along edge and interior transects at LSBRS. Contrary to predictions, total monkey encounter rate did not vary between the forest edge and forest interior. Furthermore, all three species showed neutral edge effects with no significant differences in encounter rate between forest edge and interior. Interior transects had significantly higher mean tree species richness than edge transects, and interior trees had greater DBH than edge trees, although this difference was not significant. These results suggest that forest edges negatively impact plant populations at La Suerte but that the monkeys are able to withstand these differences in vegetation.

Microbial community composition and diversity change along chemical gradients, leading to the expectation that microbial community information might provide new gradient characterizations. Here we examine stream bacteria composition and diversity along a strong chemical gradient in Central Appalachian streams. Coal mining in the region generates alkaline mine drainage (AlkMD), causing dramatic increases in conductivity, alkalinity, sulfate and metals sufficient to degrade stream macrobiota communities throughout the ecoregion. In this study, we examined the relationship between water and biofilm chemistry and biofilm bacteria taxonomic composition in streams where active and reclaimed surface coal mines occupied 0-96% of watershed surface area. We incubated wood veneers in each stream site for 4 months to develop biofilms on similar substrates. We sampled water chemistry at the time of deployment and collection, and after 1 month. Following incubation, we collected biofilms for microbial and chemical characterization. Microbial composition was determined by pyrosequencing 16S rRNA amplicons. Biofilm subsamples were analyzed by inductively coupled plasma mass spectrometry to determine metal concentrations. Our results show that microbial community composition differed significantly between AlkMD-exposed and AlkMD-unexposed sites, and that compositional dissimilarity increased with AlkMD loading. Diversity was not correlated with pH or extent of upstream mining, but instead correlated with biofilm concentrations of Cd, Mn, Zn and Ni. Within mined sites, the extent of upstream mining was negatively correlated with taxonomic richness. Despite major compositional shifts, functional capacity predicted with PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) correlated with mining in only 3 of 43 level-2 KEGG (Kyoto Encyclopedia of Genes and Genomes) Orthology groups.

Land cover change often alters the chemical regime and reduces the diversity of sensitive taxa in downstream aquatic ecosystems. The consistently elevated ionic strength associated with surface coal mines has been implicated in extirpating sensitive taxa throughout many Appalachian streams. We quantified secondary production at three sites spanning a gradient of mining impacts in the Mud River (West Virginia, U.S.A.) by sampling macroinvertebrates monthly from 2012 to 2013. Not only do we observe significant changes in aquatic insect community structure driven by the loss of sensitive taxa in mined watersheds, but we show that these losses translate directly to depressed biomass throughout the year becoming most apparent when pollutant concentrations rise during summer baseflow. These distinct seasonal patterns result in threefold decreases in total insect production and EPT production ∼1‐km downstream of an unmined reach. Farther downstream, where pollutant concentrations are much higher, total annual productivity is similar to the unmined reach, but suffers from a 31% loss of taxa comprising production and altered timing of that production. Mayflies were the insects most notably affected by the chemical alteration. Whereas mayflies accounted for ∼14% of total production in our upstream, unmined site, they only accounted for 0.2% of production at the most impacted site. We conclude that elevated ionic strength depresses insect production by preventing sensitive taxa from completing their life cycles in mining‐impacted streams. Because surface coal mining dominates the Central Appalachian landscape, such altered production patterns are likely a common landscape feature impacting regional food webs.

Rivers represent natural edges in forests, serving as transition zones between landscapes. Natural edge effects are important to study to understand how intrinsic habitat variations affect wildlife as well as the impact of human-induced forest fragmentation. We examined the influence of riparian and anthropogenic edge on mantled howler, white-faced capuchin, Central American spider monkeys, and vegetation structure at La Suerte Biological Research Station (abbreviated as LSBRS), Costa Rica. We predicted lower monkey encounter rate, tree species richness, and median dbh at both edge types compared to interior and that monkeys would show species-specific responses to edge based on size and diet. We expected large, folivorous-frugivorous howler monkeys and small, generalist capuchins would be found at increased density in forest edge, while large, frugivorous spider monkeys would be found at decreased density in forest edge. We conducted population and vegetation surveys along interior, riparian, and anthropogenic edge transects at LSBRS and used GLMM to compare vegetation and monkey encounter rate. Tree species richness and median dbh were higher in forest interior than anthropogenic edge zones. Although spider monkey encounter rate did not vary between forest edges and interior, howler monkeys were encountered at highest density in riparian edge, while capuchins were encountered at highest density in anthropogenic edge. Our results indicate that diverse forest edges have varying effects on biota. Vegetation was negatively affected by forest edges, while monkey species showed species-specific edge preferences.Our findings suggest that riparian zones should be prioritized for conservation in Neotropical forests.Abstract in Spanish is available with online material. K E Y W O R D SAlouatta, Ateles, Cebus, edge effects, forest fragmentation, natural forest edge, river edge, tropical rain forest