The North American river otter is a semiaquatic carnivore that occupies freshwater habitats across most of North America, and is a conservation symbol across most of its range. It is used as an indicator species, a keystone species, an umbrella species, and an example of conservation success. The broad distribution of river otters suggests a broad range of habitat tolerances, but they appear to be highly sensitive to anthropogenic habitat alterations. This apparently conflicting robustness across habitat types and sensitivity to habitat change has baffled researchers for decades. Using morphometric, modeling, and synthetic approaches I explored how otters relate to their environment on a broad scale to gain a better understanding of the conditions to which they are robust, the conditions to which they are sensitive, and the mechanisms by which they adapt to varying environments. Using Maximum Entropy species distribution models I determined that river otter distributions are not strongly affected by climate or macro-environmental variables. Using geometric morphometric methods to examine how cranial shape varies across geographic and ecological space, I determined that morphological variation which may affect feeding and locomotion occurs more intensely at local than broad scales, indicating local morphological adaptation is not strong in this species. Finally, a review of the phylogeography and ecology of river otters and related species indicated that this species evolved from a low-speciation lineage that tends to produce species that can occupy a wide variety of environments without undergoing evolutionary change. Overall I determined that river otters do indeed have a broad ecological niche, and do not respond strongly to climatic or environmental differences or changes across their habitats through altering their distributions or locally adapting. Evidence suggests that river otters may respond strongly to anthropogenic alterations of their habitats because anthropogenically-induced habitat alterations tend to have strong consequences for aquatic food chains, and otters may be more reliant on robust food webs than they are on other aspects of their habitats. These findings have implications for how we think of otter conservation and the conservation of species and ecosystems that are strongly affected by otter presence, as well as what otters indicate about their habitat quality. Additionally, these results may shed light on the ecologies of other mustelid carnivores.
In the first chapter of my dissertation I conducted a review of the phylogenetics and biogeography of North and South American river otters to investigate how geography and environmental change have driven river otter evolution in the Americas, and how this informs the ecology of the modern species. I focused first on reviewing the biogeography and evolutionary history of Lontra and Pteronura, and then on the modern ecology, threats, and conservations status of each of the four modern species of Lontra and the single extant species of Pteronura. I chose to review these species because they overlap in geographic and environmental space, and because their shared history provides means for an evolutionarily-grounded examination of relative rarity, specialization, and level of conservation concern. I found that speciation between American otters primarily occurs allopatrically, and there is little functional differentiation in response to allopatric speciation, though they do appear to have the ability to adapt in response to extreme conditions when necessary. Additionally, allopatric speciation primarily occurs in response to changes in waterway connectivity, which is also responsible for changes in population connectivity within the modern species. Most otters have less specific habitat requirements than previously thought, as all species of Lontra persist across a wide variety of climates and semi-aquatic environments, and most of them do so without exhibiting a strong evolutionary response. All American river otters exhibit a strong sensitivity to anthropogenic habitat destruction, though several of them have also shown some ability to coexist with humans. Evidence from this study provides strong indication that this has to do with how human activities near waterways affect food webs, and more generally that on closer examination many habitat components previously thought to be requirements for otters may be better interpreted as indicators of food availability.
In the second chapter of my dissertation I use species distribution modeling to examine river otter associations with climate and environment across their range to explore their apparently conflicting robustness to habitat change and sensitivity to human presence. Specifically, I examined the climate and landcover variables that constrain the geographic distribution of otters. I obtained otter occurrences from GBIF and BISON biodiversity databases, climate variables from BIOCLIM, and environmental layers from NASA’s SEDAC program. I built Maximum Entropy species distribution models at 80 and 150 km spatial thinning and varying numbers of background points. The combined model at 80 km spatial thinning and default number of background points produced the highest quality models. Six climatic and landcover variables explained over 10% of otter distributions each: open shrubland, net primary productivity, urban/built, water, annual mean temperature, and precipitation of coldest quarter. Of these variables I determined that water, annual mean temperature, and precipitation of coldest quarter likely have biological significance. However, the predicted range map generated by these models do not match river otter distributions generated by the IUCN and NatureServe. This is likely due to incomplete occurrence data because of low reporting in parts of the species range. I conclude based on these data that river otters have broad climatic and habitat tolerances (as there are six weakly predictive variables as opposed to 1-2 strongly controlling variables) and that local habitat factors, such as intact-ness of riverine food webs, may have greater impact on otter distributions than broad regional variables. Additionally, I call for improved monitoring and reporting of this and other broadly-distributed species to ensure we can adequately track their habitat requirements and conservation status.
In the third chapter of my dissertation I use geometric morphometrics to explore the role of cranial morphological variation in otter persistence across the array of otter habitats. I address two research questions: 1. Is there morphological variation and structure in river otters across subspecies? 2. Is morphological variation in river otters best explained by a pattern of isolation by distance or isolation by ecology? I obtained 100 river otter crania from across the species range from museums. I 3-D scanned them using a Geomagic 3-D scanner and landmarked them using the IDAV Landmark program. To answer Q1 I conducted discriminant function analyses Procrustes ANOVAs and a Principal Components Analysis. None of these tests revealed strong morphological patterning, indicating there is not morphological differentiation across subspecies in cranial shape. T address Q2 I conducted Mantel tests and a Multiple Matrix Regression with Randomization (MMRR) on the relationship between morphological distance and geographic, climatic, and environmental distance. Both the Mantel test and the MMRR results indicated no significant relationship between morphological distance and climatic or environmental distance in otter crania. Both indicated a weak but negative relationship between morphological distance and geographic distance, indicating morphological variation is greater at short geographic distances and lower at broad geographic distances. I conclude that there is significant cranial variation between individuals, but little definable structure in this variation. The weak but significant (and potentially complex) relationship between geographic and morphological distance indicates the possibility that diversifying selection across smaller spatial scales may be more important than differentiation across broader populations, possibly indicating individual specialization within a generalist population.
