Title

Evaluating Avian Communities of the Blanco River Valley Using Occupancy Modeling and Landowner Conducted Surveys

Document Type

Thesis

Comments


Presented to the Graduate Council of Texas State University–San Marcos in Partial Fulfillment of the Requirements for the Degree Master of Science, May 2008.

Committee Members Approved:
Dr. Thomas R. Simpson
Dr. John T. Baccus
Dr. Floyd W. Weckerly

Approved:
J. Michael Willoughby, Dean of the Graduate College

Abstract


Abundance and distribution of species tend to be linked, so when outside forces cause changes in population size there is a change in the number of sites occupied. Presence-nonpresence surveys are a simple method for monitoring these changes and obtaining valuable information on avian assemblages facing development and are arguably more accurate than point counts. The need for reliable and cost-effective surveys is a constant challenge for biologists. Non-governmental agencies, private organizations or “citizen scientists” may be an answer to this problem. Landowners in Texas may perform bird censuses as one of the requirements to maintain agricultural tax status by managing for wildlife. These data are public information and can be used to track broad-scale changes across a landscape through time. The study of animal distributions at a large spatial scale has benefited immensely from collaborative work with amateur ornithologists. Using the Coefficient of Jaccard, I found landowner surveys had a 51% similarity to my presence-nonpresence surveys conducted in all seasons and 56% similarity to spring and summer. These results do not show a strong similarity between the surveys. Common, year-round species such as Northern Cardinal [spring ψ = 0.9958 (SE = 0.0369), summer ψ = 0.9110 (SE = 0.0453), and winter ψ = 0.8897 (SE = 0.0546)] and Carolina Chickadee [ψ = 1.00 SE = 0.000), p = 0.5580 (SE = 0.0304)] had high occupancy and high probabilities of detection and were detected on landowner surveys. Secretive species, such as the Yellow-billed Cuckoo [ψ = 1.00 (SE = 0.000, p = 0.1017 (SE = 0.0227)], had low detection probabilities but were also detected by landowner surveys. Eleven species were equally detectable every season and considered year-round residents. Species whose occupancy varied seasonally (n = 10) declined from spring to winter, except for the Eastern Phoebe. Four species varied in occupancy and detectability seasonally, but only the Eastern Phoebe and Canyon Wren had detection probabilities increase from spring to winter. All but four of these species were detected on landowner surveys. Occupancy results suggest landowner and contractor surveys may have limited value for use by biologists but important changes such as larger sample size and additional seasons might increase value for wildlife biologists.