Predicting and mapping Plethodontid salamander abundance using LiDAR-derived terrain and vegetation characteristics

Abstract

<jats:p>Aim of the study: Use LiDAR-derived vegetation and terrain characteristics to develop abundance and occupancy predictions for two terrestrial salamander species, Plethodon glutinosus and P. kentucki, and map abundance to identify vegetation and terrain characteristics affecting their distribution.Area of study: The 1,550-ha Clemons Fork watershed, part of the University of Kentucky’s Robinson Forest in southeastern Kentucky, USA.Materials and methods: We quantified the abundance of salamanders using 45 field transects, which were visited three times, placed across varying soil moisture and canopy cover conditions. We created several LiDAR-derived vegetation and terrain layers and used these layers as covariates in zero-inflated Poisson models to predict salamander abundance. Model output was used to map abundance for each species across the study area.Main results: From the184 salamanders observed, 63 and 99 were identifdied as P. glutinosus and P. kentucki, respectively. LiDAR-derived vegetation height variation and flow accumulation were best predictors of P. glutinosus abundance while canopy cover predicted better the abundance of P. kentucki. Plethodon glutinosus was predicted to be more abundant in sites under dense, closed-canopy cover near streams (2.9 individuals per m2) while P. kentucki was predicted to be found across the study sites except in areas with no vegetation (0.58 individuals per m2).Research highlight: Although models estimates are within the range of values reported by other studies, we envision their application to map abundance across the landscape to help understand vegetation and terrain characteristics influencing salamander distribution and aid future sampling and management efforts.Keywords: Zero-inflated Poisson model; Kentucky; Cumberland plateau; Plethodon glutinosus; Plethodon kentucki.</jats:p>

Light Detection and Ranging (LiDAR)-Assisted Detection of Rock Outcrops in Appalachian Hardwood Forests

“…Such datasets have previously been used to identify fine-scale features such as forested wetlands (Lang and McCarty 2009), below-canopy cave openings (Weishampel et al 2011), and even the presence of invasive plant taxa (Asner et al 2008). Data derived from LiDAR and representing vegetation height and terrain complexity have also recently been used to improve surveys and abundance predictions for several terrestrial salamander species within central Appalachian hardwood forests (Contreras et al 2020). Our results extend this applicability to the detection of rock outcrops embedded within Appalachian hardwood forests and associated surveys of rock outcrop specialists.…”

“…Ground points (e.g., points reflecting the ground surface) from airborne LiDAR datasets can then be rendered into high-resolution elevation models that visualize terrain features beneath thick forest cover and at a resolution sufficient enough to identify microhabitat features associated with the forest floor, including ground cover characteristics, fine-scale hydrology and terrain, and variability in forest soils (Liu 2008;Akay et al 2009;Murphy et al 2011). In fact, high-resolution terrain and vegetation height information derived from LiDAR data have recently been used to improve occupancy and abundance modeling for some lungless salamander taxa within the central Appalachian region (Contreras et al 2020). Such LiDAR datasets, which are becoming more widely available via state-level geospatial programs, may therefore provide researchers with an ability to remotely identify rock outcrops within forest ecosystems more effectively than more traditional methods, such as visual inspections of aerial orthoimagery.…”