Roads as ecological traps for giant anteaters

Identifying wildlife road crossing mitigation sites using a multi-data approach - A case study from southwestern Costa Rica

Roads are one of the most widespread structures that drive habitat loss and fragmentation. But they also restrict animal movement and drive landscape-level impacts on biodiversity. The South Pacific of Costa Rica is known for its high levels of biodiversity, but little has been done to reduce road impacts upon wildlife communities. To understand these impacts and advise on possible mitigation action, we used three key data approaches: 1. Camera traps, to survey wildlife activity along two major road sections that dissect the region's protected areas and biological corridors. Seventy-eight camera traps were deployed in secondary forest patches at different distances (between 200 m and 1 km) from the roads for six months and covariates were collected to explain the patterns found. 2. Citizen science data extracted from iNaturalist to identify roadkill "hotspots" along the roads. And 3. Circuitscape analysis, to assess how landscape structure could influence animal movement. Camera traps recorded 30 terrestrial species. Ocelots and agoutis displayed a negative effect of distance from protected area, while the Apex predators displayed a positive effect toward higher forest cover and vegetation density. Circuitscape analysis showed high connectivity throughout most of the area. Only a few locations showed higher flow (bottle neck locations), which coincided with roadkill "hotspots" identified through citizen science direct observations (70 observations of 21 species). Amalgamating data from the different analyses allow us to identify four key wildlife crossing locations (one of less priority) along the Inter-American Highway. We strongly recommend the placement of under/overpasses in these locations, with the aim to ensure wildlife safe movement and connectivity of wildlife populations in the region. Culvert modifications in the area could also be considered to incorporate wildlife underpasses at a reduced cost.

Estimation and interpretation problems and solutions when using proportion covariates in linear regression models

Proportion variables, also known as compositional data, are very common in ecology. Unfortunately, few scientists are aware of how compositional data, when used as covariates, can adversely impact statistical analysis. We describe here how proportion covariates result in multicollinearity and parameter identifiability problems. Using simulated data on bird species richness as a function of land use, we show how these problems manifest when fitting a wide range of models in R, both in a frequentist and Bayesian framework. In particular, we show that similar models can often generate substantially different parameter estimates, leading to very different conclusions. Dropping a covariate or the intercept from the model can solve the multicollinearity and parameter identifiability problems. Unfortunately, these solutions do not fix the inherent challenges associated with interpreting parameter estimates. To this end, we propose focusing the interpretation on the difference of slope parameters to avoid the inherent unidentifiability of individual parameters. We also propose conditional plots with two x-axes and marginal plots as visualization techniques that can help users better interpret their modeling results. We illustrate these problems and proposed solutions using empirical data from the North American Breeding Bird Survey. The practical and straightforward approaches suggested in this article will help the fitting of linear models and interpretation of its results when some of the covariates are proportions.

Background and common lesions in the female reproductive organs of giant anteaters (Myrmecophaga tridactyla)

The giant anteater (Myrmecophaga tridactyla) is a vulnerable species in South America and is considered endangered or near extinction in Central America. Therefore, studies describing the reproductive characteristics of this species are pivotal for its conservation. Thus, this study aimed to provide a morphological description of the female reproductive tissues of this species. We collected tissue samples from six female giant anteaters and performed gross, morphological, and histochemical analyses. Five adult subjects and one juvenile were included in the study. In the ovary, classifications were made according to the follicle and oocyte sizes: primordial, primary, secondary, early antral, or antral. Typical follicles with a single oocyte surrounded by a simple or stratified layer of cubic epithelium, atretic follicles, corpora lutea, corpora albicans, and ovarian cysts were also observed. No ovarian lesions were observed. By contrast, endometritis, metritis, mucometra, and endometrial cysts were identified in the uterus. Uterine alterations in these subjects were frequent and could affect reproduction.

Do the roadkills of different mammal species respond the same way to habitat and matrix?

While road network expansion connects human settlements between themselves, it also leads to deforestation and land use changes, reducing the connectivity between natural habitat patches, and increasing roadkill risk. More than 30% of registered mammal roadkills in Brazil are concentrated in four species: Cerdocyon thous (crab-eating fox); Euphractus sexcinctus (six-banded armadillo); Tamandua tetradactyla (collared anteater) and Myrmecophaga tridactyla (giant anteater), the latter being categorized as vulnerable by IUCN redlist. Our aim was to understand how these animals’ roadkills could be related to the land use proportions on landscapes all over the Brazilian territory, and investigate if the roadkill patterns differ among species. We collected secondary data on mammal roadkills (N = 2698) from several studies in different regions of Brazil. Using MapBiomas’ data on land use and land cover, we extracted landscape composition around each roadkill sample. Through the proportion of land use and land cover in the area of influence where the roadkill occurred, we built binomial GLM models and selected the best ones by Akaike Information Criteria. For crab-eating fox and the six-banded armadillo, the best models include matrix coverage resulting in increased roadkill risk, while both anteaters’ species have a habitat and a matrix component in their best models, with an interaction between the variables. These four species seem to be roadkilled in different landscape arrangements, but in all scenarios, anthropic areas had an important influence over the models. For habitat-dependent and more sensible species, such as Tamandua tetradactyla and Myrmecophaga tridactyla, the amount of matrix influencing the roadkill risk depends on habitat availability in the landscape. It changes the strength and direction of the effect according to the proportion of natural areas in the region, while with generalist species such as Cerdocyon thous and Euphractus sexcinctus, the quantity of human-modified coverage increases the risk.