Habitat suitability modelling for species at risk is sensitive to algorithm and scale: A case study of Blanding's turtle, Emydoidea blandingii, in Ontario, Canada

Don't worry, be happy: Habitat selection of Blanding's Turtles (Emydoidea blandingii) living in a reference condition in Georgian Bay

Few areas within the Great Lakes basin are currently free from impact of human activities, and it is important to study these reference conditions for comparison with degraded sites in those regions. Here, we use radio telemetry to investigate habitat use, movement, and habitat selection of a population of the endangered (Federally in Canada) Blanding's turtle (Emydoidea blandingii, BLTU) inhabiting a mostly undisturbed archipelago located at the northern shore of Mnidoo gamii (Georgian Bay), Ontario over two active seasons (May to September 2021 and 2022). We found a mean home range of 16.21 ha for females (n = 7) and 15.10 ha for males (n = 7). Of the five habitat classes (Marsh, Open Water, Rock, Peatland, and Forest), females used all except Peatland during the nesting season, and both sexes used all five habitat classes throughout both active seasons in 2021 and 2022. Disproportionate habitat use was detected at the landscape scale but not at the home range scale which was consistent with the hypothesis that adult Blanding's turtles residing in relatively undisturbed sites with abundant habitat types use all habitat types according to their availability. We also observed the use of open, deep water by Blanding's Turtles as travel corridors for nesting and mating. Effective future conservation strategies should prioritize the protection and connectivity of relatively undisturbed wetlands, forests, and rock barrens in this region and use this study as a reference condition to compare BLTU habitat use and movement across disturbance gradients within Georgian Bay.

Predicting the Habitat Suitability of Melaleuca cajuputi Based on the MaxEnt Species Distribution Model

Gelam tree or Melaleuca cajuputi (M. cajuputi) is an important species for the local economy as well as coastal ecosystem protection in Terengganu, Malaysia. This study aimed at producing a current habitat suitability map and predicting future potential habitat distribution for M. cajuputi in Terengganu based on Species distribution modeling (SDM) using the Maximum Entropy principle. Our modeling results show that for the current potential distribution of M. cajuputi species, only 10.82% (1346.5 km2) of Terengganu area is suitable habitat, which 0.96% of the areas are under high, 2.44% moderate and 7.42% poor habitat suitability. The model prediction for future projection shows that the habitat suitability for M. cajuputi would decrease significantly in the year 2050 under RCP 4.5 where the largest contraction from suitable to unsuitable habitat area is about 442.1 km2 and under RCP 2.6 is the highest expansion from unsuitable to suitable habitat area (267.5 km2). From the future habitat suitability projection, we found that the habitat suitability in Marang would degrade significantly under all climate scenarios, while in Setiu the habitat suitability for M. cajuputi remains stable throughout the climate change scenarios. The modeling prediction shows a significant influence on the soil properties, temperature, and precipitation during monsoon months. These results could benefit conservationist and policymakers for decision making. The present model could also give a perception into potential habitat suitability of M. cajuputi in the future and to improve our understanding of the species’ response under the changing climate.

Effects of landscape composition on wetland occupancy by Blanding’s Turtles (Emydoidea blandingii) as determined by environmental DNA and visual surveys

Habitat loss and degradation have led to the extinction of many species worldwide. The endangered Blanding’s Turtle (Emydoidea blandingii (Holbrook, 1838)), a semi-aquatic freshwater turtle, occupies a wide range of wetlands and landscapes primarily in southeastern Canada and the Great Lakes region of the United States. We explored whether the probability of wetland occupancy by Blanding’s Turtles is affected by the surrounding landscape. We used visual surveys, environmental DNA, and Atlas data to document the presence of Blanding’s Turtles in wetlands in Ottawa, Ontario, Canada. We tabulated landscape composition at multiple scales surrounding the wetlands to determine whether landscape composition can predict wetland occupancy. Generally, wetlands surrounded by forest and other undisturbed lands were most likely to harbour Blanding’s Turtles, whereas those surrounded by more human-disturbed lands were least likely to harbour Blanding’s Turtles. Larger wetlands and a high proportion of wetlands in the surrounding landscape also increased the probability of occupancy by Blanding’s Turtles. Finally, older wetlands were more likely to be occupied by Blanding’s Turtles. The ability to estimate a species’ probability of occupancy can aid in conservation efforts, such as critical habitat delineation.

Background matching and camouflage efficiency predict population density in four-eyed turtle (Sacalia quadriocellata)

Background matching is an important way to camouflage and is widespread among animals. In the field, however, few studies have addressed background matching, and there has been no reported camouflage efficiency in freshwater turtles. Background matching and camouflage efficiency of the four-eyed turtle, Sacalia quadriocellata, among three microhabitat sections of Hezonggou stream were investigated by measuring carapace components of CIE L*a*b* (International Commission on Illumination; lightness, red/green and yellow/blue) color space, and scoring camouflage efficiency through the use of humans as predators. The results showed that the color difference (ΔE), lightness difference (ΔL(*)), and chroma difference (Δa(*)b(*)) between carapace and the substrate background in midstream were significantly lower than that upstream and downstream, indicating that the four-eyed turtle carapace color most closely matched the substrate of midstream. In line with these findings, the camouflage efficiency was the best for the turtles that inhabit midstream. These results suggest that the four-eyed turtles may enhance camouflage efficiency by selecting microhabitat that best match their carapace color. This finding may explain the high population density of the four-eyed turtle in the midstream section of Hezonggou stream. To the best of our knowledge, this study is among the first to quantify camouflage of freshwater turtles in the wild, laying the groundwork to further study the function and mechanisms of turtle camouflage.

A Current Perspective on the Historical Geographic Distribution of the Endangered Muriquis (Brachyteles spp.): Implications for Conservation

The muriqui (Brachyteles spp.), endemic to the Atlantic Forest of Brazil, is the largest primate in South America and is endangered, mainly due to habitat loss. Its distribution limits are still uncertain and need to be resolved in order to determine their true conservation status. Species distribution modeling (SDM) has been used to estimate potential species distributions, even when information is incomplete. Here, we developed an environmental suitability model for the two endangered species of muriqui (Brachyteles hypoxanthus and B. arachnoides) using Maxent software. Due to historical absence of muriquis, areas with predicted high habitat suitability yet historically never occupied, were excluded from the predicted historical distribution. Combining that information with the model, it is evident that rivers are potential dispersal barriers for the muriquis. Moreover, although the two species are environmentally separated in a large part of its distribution, there is a potential contact zone where the species apparently do not overlap. This separation might be due to either a physical (i.e., Serra da Mantiqueira mountains) or a biotic barrier (the species exclude one another). Therefore, in addition to environmental characteristics, physical and biotic barriers potentially shaped the limits of the muriqui historical range. Based on these considerations, we proposed the adjustment of their historical distributional limits. Currently only 7.6% of the predicted historical distribution of B. hypoxanthus and 12.9% of B. arachnoides remains forested and able to sustain viable muriqui populations. In addition to measurement of habitat loss we also identified areas for conservation concern where new muriqui populations might be found.

Habitat-forming bryozoans in New Zealand: their known and predicted distribution in relation to broad-scale environmental variables and fishing effort

Frame-building bryozoans occasionally occur in sufficient densities in New Zealand waters to generate habitat for other macrofauna. The environmental conditions necessary for bryozoans to generate such habitat, and the distributions of these species, are poorly known. Bryozoan-generated habitats are vulnerable to bottom fishing, so knowledge of species’ distributions is essential for management purposes. To better understand these distributions, presence records were collated and mapped, and habitat suitability models were generated (Maxent, 1 km² grid) for the 11 most common habitat-forming bryozoan species: Arachnopusiaunicornis, Cellariaimmersa, Cellariatenuirostris, Celleporariaagglutinans, Celleporinagrandis, Cinctiporaelegans, Diaperoeciapurpurascens, Galeopsisporcellanicus, Hippomenellavellicata, Hornerafoliacea, and Smittoideamaunganuiensis. The models confirmed known areas of habitat, and indicated other areas as potentially suitable. Water depth, vertical water mixing, tidal currents, and water temperature were useful for describing the distribution of the bryozoan species at broad scales. Areas predicted as suitable for multiple species were identified, and these ‘hotspots’ were compared to fishing effort data. This showed a potential conflict between fishing and the conservation of bryozoan-generated habitat. Fishing impacts are known from some sites, but damage to large areas of habitat-forming bryozoans is likely to have occurred throughout the study area. In the present study, spatial error associated with the use of historic records and the coarse native resolution of the environmental variables limited both the resolution at which the models could be interpreted and our understanding of the ecological requirements of the study species. However, these models show species distribution modelling has potential to further our understanding of habitat-forming bryozoan ecology and distribution. Importantly, comparisons between hotspots of suitable habitat and the distribution of bottom fishing in the study area highlight the need for management measures designed to mitigate the impact of seafloor disturbance on bryozoan-generated habitat in New Zealand waters.