Maximizing Detection Probability of Wetland-dependent Birds during Point-count Surveys in Northwestern Florida

—We conducted 262 call-broadcast point-count surveys (1–6 replicate surveys on each of 62 points) using standardized North American Marsh Bird Monitoring Protocols between 31 May and 7 July 2006 on St. Vincent National Wildlife Refuge, an island off the northwest coast of Florida. We conducted doubleblind multiple-observer surveys, paired morning and evening surveys, and paired morning and night surveys to examine the influence of call-broadcast and time of day on detection probability. Observer detection probability for all species pooled was 75% and was similar between passive (69%) and call-broadcast (65%) periods. Detection probability was higher on morning than evening (t 3.0, P 0.030) or night (t 3.4, P 0.042) surveys when we pooled all species. Detection probability was higher (but not significant for all species) on morning compared to evening or night surveys for all five focal species detected on surveys: Least Bittern (Ixobrychus exilis), Clapper Rail (Rallus longirostris), Purple Gallinule (Porphyrula martinica), Common Moorhen (Gallinula chloropus), and American Coot (Fulica americana). We detected more Least Bitterns (t 2.4, P 0.064) and Common Moorhens (t 2.8, P 0.026) on morning than evening surveys, and more Clapper Rails (t 5.1, P 0.014) on morning than night surveys. Received 2 March 2007. Accepted 6 October 2007. Maximizing detection probability of rare or inconspicuous birds during point-count surveys is essential so that sufficient individuals are detected to reliably estimate population trends (Lynch 1995). Wetland-dependent birds (i.e., rails and bitterns) are among the most inconspicuous groups of birds in North America. They vocalize infrequently and often occur in isolated wetlands making them difficult to monitor (Bystrak 1981, Gibbs and Melvin 1993). A marsh bird monitoring protocol was developed in 1999 for conducting standardized surveys for wetland-dependent birds across North America (Conway 2007). The protocol instructs surveyors to use call-broadcast after an initial 5-min passive period to increase vocalization probability of birds present during the survey period. However, callbroadcast might interfere with the observer’s ability to hear vocalizing birds during the survey (Conway and Nadeau 2006). A decrease in observer detection probability could potentially negate the benefits of increased vocali1 USGS, Arizona Cooperative Fish and Wildlife Research Unit, 325 Biological Sciences East, University of Arizona, Tucson, AZ 85721, USA. 2 St. Vincent National Wildlife Refuge, U.S. Fish and Wildlife Service, 479 Market Street, P. O. Box 447, Apalachicola, FL 32329, USA. 3 Corresponding author; e-mail: [email protected] zation probability. Few studies have examined the effects of call-broadcast on observer detection probability and participants using this protocol are not obligated to use methods (i.e., multiple-observer surveys) to account for variation among observers. Vocalization probability of wetland-dependent birds varies with time of day and diurnal patterns may vary regionally and among species (Conway and Gibbs 2001). Thus, surveyors should identify the optimal time of day to conduct surveys in their region to maximize detection probability. The Standardized North American Marsh Bird Monitoring Protocol (Conway 2007) instructs participants to conduct surveys in the morning or evening when birds are most vocal. Participants using the protocol are not obligated to define which of the two daily time periods is optimal in their region and few participants have attempted to do so. Many species of wetland-dependent birds are known to vocalize at night (e.g., Black Rail [Laterallus jamaicensis], Clapper Rail [Rallus longirostris], Virginia Rail [R. limicola], Yellow Rail [Coturnicops noveboracensis], and American Bittern [Botaurus lentiginosus]; Reynard 1974, Meanley 1985, Johnson and Dinsmore 1986, Gibbs et al. 1992, Bookhout 1995), yet few studies have attempted to examine the efficacy of night surveys. We compared morning to evening sur514 THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 120, No. 3, September 2008 veys and morning to night surveys to ascertain the optimal time of day to conduct surveys for wetland-dependent birds in the southeastern United States. Our specific objectives were to: (1) compare observer detection probability between passive and call-broadcast surveys in the context of the North American Marsh Bird Monitoring Protocol, and (2) ascertain the optimal time of day to conduct surveys for wetland-dependent birds in the southeastern United States.