Images

Laura S. Deming, Ph.D., is a 2025 graduate of the Ph.D. Program in Environmental Studies and Sustainability at Antioch University, New England

Dissertation Committee:

  • Lisabeth Willey, PhD, Chairperson
  • James W. Jordan, PhD, Committee Member
  • Michael T. Jones, PhD, Committee Member

Dr. Deming, at work in the field, holding a wood turtle

Dr. Deming in the field.

Keywords

wood turtle, floodplain, geomorphology, LiDAR, habitat use, geomorphic complexity

Document Type

Dissertation

Publication Date

2025

Abstract

Wood turtles occupy cold, low-gradient riverine habitat throughout the northeastern United State and southeastern Canada. Stream alterations, such as dams, channel straightening, and floodplain development have diminished the natural hydrologic and sedimentation regimes that result in complex channel and floodplain geomorphology. Studies throughout the species’ range have quantified many aspects of their habitat, but none have quantified components of geomorphology of wood turtle sites. The goal of this study was to explore the role of stream channel and floodplain geomorphology in wood turtle habitat use, with specific objectives to (1) quantify components of stream channel and floodplain geomorphic complexity, (2) explore the role of floodplain geomorphic complexity in wood turtle habitat use, and (3) explore the role of floodplain vegetation structure in wood turtle habitat use. I calculated metrics of geomorphic complexity for five floodplains along a stream in the White Mountain Ecoregion. Stream channel metrics included gradient, sinuosity, and the number and area of point bars along each floodplain stream segment. Floodplain metrics were calculated using 1-meter resolution bare earth lidar, following a methodology developed by Scown et al. Three metrics of elevational variability (standard deviation, skewness, and standard deviation of curvature) were calculated at 6 scales from 3x3 pixels (4.41 m2) to 13x13 pixels (82.81 m2) to produce 18 elevation variables. Spatial variability was quantified by calculating a Moran’s I value of spatial autocorrelation for each of the 18 variables at 9 scales ranging from 5-meter to 100-meter radius around each point to yield 162 variables of spatial variability. I used these variables to compare geomorphic condition among the five floodplains. In one floodplain with previously collected wood turtle data, I used generalized linear models to test the association between each variable and wood turtle habitat use, represented by a kernel use density (kud) estimate calculated from the wood turtle observation points. I also calculated three components of vegetation structure from classified lidar for this floodplain, and used generalized linear models to test the association between each vegetation variable and wood turtle habitat use, represented by the kud. Elevation variables were similar for the five floodplains, and highlighted geomorphic features within the floodplain, such as scroll ponds, abandoned channels, and steep slopes. These variables were the most useful for predicting wood turtle habitat use within the floodplain. Moran’s I variables were not useful for predicting wood turtle habitat use or highlighting floodplain features but provide a means of comparing patchiness of variability among sites. Vegetation metrics were more useful than the elevation variables for predicting wood turtle presence, suggesting that, as measured in this study, vegetative structure may play a stronger role than geomorphic complexity in wood turtle habitat use. However, combined vegetation and elevation models were the best models overall, indicating that metrics of geomorphic complexity may provide useful information for identifying high value features or areas within the floodplain. Results of this study show that certain metrics of terrain elevational complexity were positively associated with wood turtle habitat use within the floodplain, and these metrics highlighted features, such as scroll ponds and steep slopes, known to be important to wood turtles. Variables of vegetation structure showed positive association with wood turtle habitat use for values representing habitat favored by wood turtles and had stronger predictive value than elevation metrics. However, the best overall model for predicting wood turtle habitat use included metrics of both vegetation structure and geomorphic complexity, confirming the relevance of geomorphic complexity in evaluating wood turtle habitat. Given the current availability of high-resolution lidar, such metrics can be readily calculated for areas of interest, contributing valuable information for identifying and evaluating known and potential wood turtle sites. Metrics of geomorphic complexity also provide a means of evaluating floodplain geomorphic condition across multiple sites and broad geographic regions. These, and other measures of stream and floodplain geomorphic complexity provide a means for establishing a quantitative framework for evaluating geomorphic condition of riverine ecosystems at multiple scales and may be useful in stream restoration efforts. This dissertation is available in open access at AURA, https://aura.antioch.edu/, and OhioLINK ETD Center, https://etd.ohiolink.edu.

Comments

ORCID No.: 0009-0009-5788-4293

Bio:

Laura Deming is a wildlife biologist at Moosewood Ecological and adjunct instructor at Antioch University, in Keene, NH. A native of South Portland, Maine, she earned a B.S. from Dartmouth College and M.S. from the University of Vermont School of Natural Resources. She was a senior biologist at New Audubon from 1992 to 2018, where she coordinated numerous field surveys for birds, turtles, vernal pool species, and bats. Laura recently joined the Moosewood Ecological team, and has worked on several projects across the state, including surveys for vernal pools, birds, turtles, snakes, and fish, conducting water quality and invertebrate sampling, and evaluating municipal regulatory and planning documents to assist community conservation efforts. For her doctoral research, Laura used high-resolution airborne LiDAR to evaluate geomorphic complexity of floodplains and how metrics of complexity are associated with wood turtle habitat use. In her free time, she enjoys birding, gardening, painting, hiking, and teaching her two dogs, Kaya and Finn, to find things.

Download

Share

COinS