Modeling Threatened California Red Legged Frog (Rana draytonii) Habitat with MaxEnt Using Environmental DNA Presence via Quantitative PCR

As a sensitive and federally threatened species, the California Red Legged Frog (Rana draytonii) is known as a bioindicator because it most often occurs in complex healthy ecosystems capable of sustaining a diverse web of species. As humans affect natural habitats and climate change alters the environment, locating and monitoring sensitive species and their habitats is vital to supporting biodiversity and ecosystem health. Once ranging from Northern Mexico to the northern reaches of the California coast and Sierra Mountains, habitat of R. draytonii has been reduced to primarily the central coast of California. Environmental DNA ( eDNA) can be invaluable in locating rare, largely nocturnal, and protected species such as R. draytonii. This study will identify the habitat of R. draytonii using eDNA from water samples from several central coastal California watersheds and use environmental data - including invasive predatory American bullfrog (Lithobates catesbeianus) presence- to create a MaxEnt habitat model for R. draytonii. Through this model, I ask: Which environmental variables best predict R. draytonii presence and is the presence of the invasive L. catesbeianus a significant predictor? The best model F-score was 0.14 and contained 11 variables. Based on permutation importance, top variables were precipitation, percent conifer coverage, percent grass coverage, and mean temperature. Per the MaxEnt model, L. catesbeianus presence positively correlated with R. draytonii presence but was not significant when considered alone through Welch's Student t-test results. A parallel model of R. draytonii presence given climate change based on relative concentration pathways (RCP) 4.5 and 8.5 for the year 2041-2045 did not validate sufficiently and was not presented here so as to not be misleading. I recommend continuing to refine R. draytonii habitat models by including variables such as Chytrid fungus presence, food source density, catchment ephemeral waterbody percentage, and distance between waterbodies. Currently, I recommend land managers use the predictive map and consider this model's top-scoring variables to inform land conservation, corridor formation, and habitat improvement to support this species and, by definition, myriad species now and into the future.