Habitat characteristic profiles: an intuitive approach to evaluate species-habitat relationships as demonstrated on several Texas bird species
Date
2021-12
Authors
Plappert, Joseph L.
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Identifying the habitat characteristics that matter most to a species is crucial to understanding its basic ecology and conservation needs. Although species-habitat relationships are often considered complex and best understood with large, multi-faceted models, a simpler approach may prove fast, cost-effective, and powerful. I used logistic regression models to generate habitat-characteristic profiles (HCPs), a graphical interpretation method wherein a single habitat variable is plotted on the x-axis and the
probability of species occurrence is on the y-axis. For a group of eight bird species, I evaluated four habitat variables, all related to vegetation structure: canopy cover, contagion index (a measure of spatial heterogeneity), broadleaf:juniper ratio, and mean canopy height.
All four variables were measured remotely with canopy cover, contagion index, and broadleaf:juniper ratio being generated from remote-sensing satellite imagery of the National Agricultural Imagery Program and mean canopy height coming from LiDAR data of the Global Ecosystem Dynamics Investigation. All eight bird species are “species of conservation concern”, as identified by the Oaks and Prairies Joint Venture (OPJV). Model building was completed using seven years (2012-2019) of OPJV point count species presence-absence data. These data originate from 19 survey routes, in six Texas counties, for a total of 478 points. I used ArcGIS to circumscribe a 250 m radius circular buffer around each point and subsequently derive each of the four habitat variables within each buffer.
Logistic regression models were developed to examine the effect of each habitat variable separately and in combination with the other variables. The models were compared using AIC. Competitive models (ΔAIC < 3) were used to generate HCPs for each variable and species. HCPs proved to be an effective method for understanding and displaying species-habitat relationships and comparing among species. For most species, vegetation structure appeared to strongly influence species occurrence (habitat use) within the relatively small local area of the 250 m radius buffers. Furthermore, percent canopy cover alone was sufficient in explaining patterns of habitat use for the majority of focal species. This suggests
that species-habitat relationships may be relatively simple, contrary to many habitat characterizations that sometimes include dozens of habitat variables. As a further assessment of the HCPs, I conducted model validation using two sets of independently collected species presence-absence data. I found that the models performed well at predicting probability of occurrence for all species (AUC > 0.5). In a time where many species are in steep decline, a quick method of evaluating species-habitat relationships could prove very beneficial. By utilizing HCPs, the habitat associations of any species could be quickly evaluated with minimal survey effort potentially resulting in better conservation outcomes.
Description
Keywords
Birds, Habitat, Habitat-characteristic profile, Canopy cover, Contagion index, Canopy height, Ecology, Grassland, Shrubland, Texas, Edwards Plateau, Oaks and prairies, Freeman Center, Kerr Wildlife Management Area
Citation
Plappert, J. L. (2021). <i>Habitat characteristic profiles: an intuitive approach to evaluate species-habitat relationships as demonstrated on several Texas bird species</i> (Unpublished thesis). Texas State University, San Marcos, Texas.