Date of Graduation

5-2024

Document Type

Thesis

Degree Name

Master of Science in Biology (MS)

Degree Level

Graduate

Department

Biological Sciences

Advisor/Mentor

Roberts, Caleb P.

Committee Member

DeGregorio, Brett A.

Second Committee Member

Mortensen, Jennifer L.

Keywords

Avian Communities; Boundary Detection; Community Ecology; Landscape Ecology; Spatial Covariance

Abstract

Globally, grasslands are declining in range, which is leading to a decline in grassland-dependent taxa. These declines can be attributed to a variety of issues, a major one being the encroachment of woody species into historic grasslands. There is a new framework for strategizing grassland management, referred to as “Defend the Core,” which calls for proactive management of intact grasslands, meaning conserving remaining grasslands by preventing woody encroachment in the first place. To successfully defend the core, it is imperative that we develop rapid methods to identify intact grassland cores for management priority. One method that can be used for identifying grassland cores is identifying and tracking spatial regimes, which can be done using the relatively new metric of spatial covariance. Spatial covariance can measure the degree of coexistence between two vegetative functional groups, essentially quantifying boundaries between two cover types. To identify grassland regimes, or intact grassland cores, we can calculate the spatial covariance between tree cover and grass cover. However, this metric was recently developed, and knowledge gaps remain. This metric has never been tested in grasslands of the southeastern United States and has not been used as a covariate for occupancy probability modeling. Additionally, spatial covariance has never been related to established on-the-ground metrics of vegetation structure. For this thesis, we used the Rangeland Analysis Platform cover data to calculate spatial covariance between tree and grass cover in historic grassland sites across the Mississippi Alluvial Valley and the Arkansas River Valley in central and east Arkansas. These sites have undergone varying management interventions and are highly fragmented in comparison to rangelands of the western US, where spatial covariance has previously been tested. In Chapter I, we related spatial covariance to a variety of established metrics for quantifying vegetation structure. We collected vegetation data at our study sites and used generalized linear models to understand the relationship between each vegetation measurement and spatial covariance. We found that there were no strong relationships between on-the-ground vegetation metrics and spatial covariance, only further emphasizing that spatial covariance is identifying and measuring something that is not captured by many traditional field-collected vegetation metrics. In Chapter II, we used spatial covariance in an occupancy framework to understand avian responses to spatial covariance and boundaries of grassland cores. We used a multi-species occupancy model to predict occupancy probability of the grassland bird community and individual grassland bird species and mapped these results across a known grassland core. We found that all grassland birds included in our models avoided areas with negative spatial covariance, or boundaries between grass and tree cover, and were more likely to occupy intact grassland cores that are free of boundaries. In this thesis, we delve deeper into the recently developed spatial covariance metric. We find that this metric is promising for identifying grassland cores from an avian perspective, which only helps better strategize management and protection of grasslands.

Download

Share

COinS