Estimates of tidal‐marsh bird densities using Bayesian networks. Issue 1 (19th September 2018)

• Record Type: Journal Article
• Authors: Wiest, Whitney A.; Correll, Maureen D.; Marcot, Bruce G.; Olsen, Brian J.; Elphick, Chris S.; Hodgman, Thomas P.; Guntenspergen, Glenn R.; Shriver, W. Gregory
• Is Part Of: Journal of wildlife management. Volume 83:Issue 1(2019)
• Page Start: 109
• Page End: 120
• Publication Date: 2018-09-19
• Subjects: Bayesian network -- density -- model‐based -- monitoring -- Northeast USA -- predictive model -- tidal‐marsh birds; Wildlife management -- Periodicals; Zoology -- Periodicals; 333.954
• DOI: 10.1002/jwmg.21567
• Languages: English
• ISSNs: 0022-541X

Abstract

ABSTRACT: Conserving tidal‐marsh bird communities requires strategies to address continuing pressures from human development to the effects of increasing rates of sea‐level rise. Knowing tidal‐marsh bird distributions and population sizes are important for developing these strategies. In the Northeast United States, where estimates of sea‐ level rise are 3 times higher than the global average, 5 bird species are tidal‐marsh specialists: clapper rail ( Rallus crepitans ), willet ( Tringa semipalmata ), Nelson's sparrow ( Ammospiza nelsoni ), saltmarsh sparrow ( A. caudacuta ), and seaside sparrow ( A. maritima ). We used a regional marsh bird survey to develop Bayesian network models to identify factors that influence patch‐scale species density and to estimate regional population sizes. We modeled species density as a function of habitat covariates at the patch, local, landscape, and regional spatial scales. Densities were most sensitive to patch location and dimension, patch geomorphic setting, indices of human development, and changes in mean sea level. We estimated 110, 000 clapper rails (95% CI = 61, 000–159, 000), 111, 000 willets (95% CI = 70, 000–152, 000), 7, 000 Nelson's sparrows (95% CI = 4, 000–10, 000), 60, 000 saltmarsh sparrows (95% CI = 40, 000–80, 000), and 234, 000 seaside sparrows (95% CI = 112, 000–356, 000) from the United States–Canada border to, and including, the mouth of the Chesapeake Bay, Virginia, USA. Our abundance estimates can be used to identify priority conservation areas at multiple geographic scales and our models help identify key habitat and landscape components for tidal‐marsh restoration and management to benefit tidal‐marsh birds and can be modified for other species. © 2018 The Wildlife Society. Abstract : Our Bayesian models help identify marsh patches that support breeding tidal‐marsh specialist birds in the Northeast United States and provide information that helps focus conservation actions to maintain tidal‐marsh bird communities. Our results support conserving patches with lower rates of sea‐level rise and prioritizing tidal‐marsh restoration and management techniques that maintain suitable habitat for nesting tidal‐marsh birds.