Testing novel methods to assess how habitat associations and invasive species influence northern leopard frog (Lithobates pipiens) persistence in northern Colorado

Abstract

Land cover change, invasive species, and disease are among the leading causes of amphibian decline worldwide but are often studied in isolation and over relatively short time scales. Understanding how species respond over time to interacting stressors is essential to inform conservation decisions. The northern leopard frog (Lithobates pipiens) is a classic example of an amphibian experiencing declines related to these stressors: populations in the western U.S. are threatened by interactions among land cover change, the invasive American bullfrog (Lithobates catesbeianus), and the fungal pathogen Bd (Batrachochytrium dendrobatidis). In northeastern Colorado, northern leopard frog distribution has declined for decades, while bullfrog populations have expanded. American bullfrogs, native to the eastern U.S., threaten northern leopard frog persistence as a predator, competitor, and asymptomatic carrier of the fungal pathogen Bd. Land cover and habitat have changed substantially in northeastern Colorado in recent decades, resulting in natural wetland loss, often through the replacement of ephemeral ponds and wetlands with artificial permanent water bodies that favor invasive species encroachment. Understanding how northern leopard frog distribution is influenced by interacting stressors over time relies on effective monitoring programs. However, effective monitoring is challenging because declining native and recently arriving invasive taxa are inherently difficult to detect. Unaccounted for variation in species' detection across space (e.g., locations) or time can bias our understanding of how species distributions (e.g., occurrence, persistence) are influenced by interacting stressors over time. Comparing species' detection probabilities across multiple detection methods enables the optimization of monitoring protocols that can then be used to address hypotheses related to species declines and associated conservation actions. The first chapter of this thesis compares the efficacy of three amphibian detection methods—eDNA, automated acoustic recorders, and visual encounter surveys— for native northern leopard frogs and invasive bullfrogs. We explored hypotheses related to how environmental (e.g., temperature, pH) and survey-specific (e.g., survey date, search time) predictors influence detection probabilities for each species and method. We collected detection-nondetection information for both amphibian species at 38 potential breeding sites in Colorado's South Platte River Basin and used an occupancy modeling approach to estimate resulting detection probabilities. We hypothesized that detection probabilities for all three methods would vary with environmental factors, some survey-specific considerations, and be associated with species' distinct life histories and habitat preferences. We found that northern leopard frog detection probability was higher for visual encounter surveys and eDNA than for automated acoustic recorders, where detection was lower but peaked in early spring. Northern leopard frog eDNA detection was highest in neutral pH conditions and visual encounter survey detection increased slightly with longer search times. Bullfrog eDNA and automated acoustic detection were higher than that of northern leopard frogs, with comparable visual encounter detection probabilities for both species. Bullfrog eDNA and visual detection increased in warmer water temperatures and acoustic detection increased later in the season. These results aligned with each species' breeding habitat preferences, behaviors, and seasonal emergence. Using our results, we provided examples of how to combine methods to create optimized monitoring programs for given objectives. The second chapter of this thesis explores how northern leopard frog persistence is influenced by interactions among habitat features and bullfrog presence. To predict species' responses to interacting threats at a landscape-level, researchers benefit from estimating persistence (or conversely, local extirpation), and colonization. Because the effects of invading species on native species persistence occurs over broad time scales (e.g., decades) at a landscape level, understanding northern leopard frog response to bullfrog invasion benefits from long-term monitoring data, which does not currently exist in our study area. Instead, we conditioned on previous occurrence (detection only) records of northern leopard frogs to estimate persistence by visually resurveying those previously occupied sites. We recorded the contemporary presence of bullfrogs and habitat features (e.g., emergent vegetation, canopy cover, surrounding land cover) to explore hypotheses related to how northern leopard frog persistence was influenced by interactions among bullfrog presence and habitat. We used a two-species static occupancy approach to investigate hypotheses related to bullfrog presence and habitat features. We predicted that bullfrog presence and anthropogenic land cover (urban, agricultural) would interact to decrease northern leopard frog persistence while bullfrog absence and grassland cover would increase persistence probability. We expected multiple stressors (e.g., bullfrogs and low emergent vegetation) would reduce persistence and beneficial habitat features (e.g., high emergent vegetation) would mitigate the negative effects of bullfrog presence. Contrary to expectations, bullfrog presence was associated with northern leopard frog persistence, suggesting bullfrogs may colonize remaining northern leopard frog habitats. Northern leopard frogs were more likely to persist at sites with bullfrogs when surrounding grassland cover was high; thus, grassland mediates the potential negative effects of invasive bullfrogs. Ultimately, long term detection-nondetection information for both the native and invasive species are needed to predict whether northern leopard frogs will persist in the future, but our results represent a time-specific description of where they have persisted. Our findings suggest that grassland conservation and restoration is an important tool to support the persistence of northern leopard frogs amidst bullfrog encroachment. Northern leopard frog declines have been pervasive in northeastern Colorado, prompting an urgent need for better informed conservation efforts and monitoring programs. Monitoring methods can be combined in novel ways to achieve desired species' detection probabilities, yielding improved estimates of their distributions/occurrences and better-informed conservation initiatives. Conservation efforts for northern leopard frogs should consider the important role of surrounding grassland for northern leopard frog persistence, especially in areas where bullfrogs co-occur. In this thesis, we provide suggestions on how to improve monitoring for northern leopard frogs and bullfrogs, evaluate how northern leopard frog persistence is influenced by interacting threats, and suggest future directions for management and conservation actions.