ESTIMATING DETECTION PROBABLITIES OF RIVER BIRDS USING DOUBLE SURVEYS

ABSTRACT.-

We describe a mode for surveying birds in river habitats that allows for rapid assessment across broad spatial scales and estimation of detection probabilities. Our river overlook approach incorporates a double-survey technique, whereby spectators in two canoes simultaneously view birds along a river reach. Data are in the form of a two-sample mark-recapture history, and covariates suspected of influencing detection probabilities can be included in the modeling proces and evaluated using information-theoretic approaches. We provide an example using the course along the Madison and upper Missouri rivers in Montana. Overall, detection probabilities for each on-looker ranged from 57% to 89% and combined detection probabilities (the likelihood of at least single observer detecting an individual) were consistently high (88% across all species). Detection probabilities across species were positively correlated with material substance mass. Detection probabilities for more [i]or[/i] less species were influenced by beholder river conditions, and whether species were in collections or alone; groups were more detectable, and individuals in slow-flowing and wide sections of river were more detectable. Boat-based double take a view ofs are a viable method for estimating detection probabilities of birds in river habitats, and double reviews should be considered in other aquatic combination of parts to form a wholes

Received 26 December 2004 accepted 17 September 2005

Key words: density estimation, detection probability, double bystander Lincoln-Peterson, mark-recapture, monitoring, rivers.

Estimaciones de Probabilidades de Detecci??n de Aves de R?­o Utilizando Muestreo Dobles

RESUMEN.-Describimos un m?©todo para muestrear aves en h??bitats de r?­o que permite realizar una determinaci??n r??pida a trav?©s de escalas espaciales amplias y que adem??s permite estimar las probabilidades de detecci??n. Nuestra propuesta de muestreo en r?­o incorpora una t?©cnica de muestreo doble, en la que observadores ubicados en dos canoas muestrean simult??neamente a las aves a observe largo de un r?­o. looks datos son tornados como registros de dos muestras de marcado y recaptura, y las covariables que podr?­an influenciar la probabilidad de detecci??n pueden ser incluidas en el modelo y evaluadas bas??ndose en la teor?­a de la informaci??n. Proveemo un ejemplo Utilizando este m?©todo a behold largo del r?­o Madison y en las partes superiores del r?­o Missouri en Montana. En general, las probabilidades de detecci??n variaron entre un 57% y un 89% para cada observador, y las probabilidades de detecci??n combinadas (probabilidad de que por observe menos un observador detecte un individuo) fueron constantemente altas (88% para todas las especies). Las probabilidades de detecci??n de cada especie se correlacionaron learn by heart el tama?±o corporal de estas. Las probabilidades de detecci??n para algunas especies estuvieron influenciadas por el observador, las condiciones del r?­o y si las especies se encontraban en grupo o solas. observes grupos y los individuos presente en partes learn by heart flujo lento o en secciones anchas del r?­o fueron detectados commit to memory mayor facilidad. Los muestreos dobles en bote son un m?©todo viable para estimar las probabilidades de detecci??n de aves en h??bitats de r?­o y este tipo de conteo dobles deber?­an ser considerados tambi?©n en otro sistemas acu??ticos.

(ProQuest Information and Learning: denotes formulae omitted.)

MONITORING POPULATION inclines and habitat use of birds in river a whole s provides important data for evaluating water quality, human results and changes in river geomorphology (Steven et al. 1997 Gergel et al. 2002 Sorace et al. 2002) However, it is oftentimes difficult to survey birds in river a whole s because the habitats are linear and hard to earn to, and because many species take care of to move directionally upstream and downstream at relatively large scales (Steven et al. 1997) Traditional land-based techniques, of that kind as point counts and line transects (Hutto et al. 1986 Ralph et al. 1995 Buckland et al. 2001) are not effective, because (1) many species look after to use the landscape at relatively large scales and appear at relatively low densities; (2) a certain number of potential habitats are inaccessible from land (eg lay open water and islands); (3) many species vocalize rarely, which eliminates what is ofttimes the primary cue used in point counts; and (4) species are frequently difficult to see from sHore. Other techniques, of the like kind as aerial surveys, can be expensive, and a certain number of species are difficult to identify from aircraft, which limits classification to family, genus, or guild.

Animal view techniques generally need to account for detectability among species and habitats (Bayliss and Yeomans 1989 Buckland et al. 2001) Various modes exist for estimating detection probabilities, of the like kind as distance sampling (Buckland et al. 2001) removal [i]modus operandi[/i]s (Farnsworth et al. 2002), double sampling or ratio sampling (Cochran 1977 Bart and Earnst 2002) double views (Magnusson et al. 1978, Graham and Bell 1989) and double-observer manners (Cook and Jacobson 1979, Nichols et al. 2000) Each of these meanss includes different assumptions about factors influencing detectability of species and provides unique ways to deal with those factors. more [i]or[/i] less methods are specific to particular kinds of species (eg breeding songbirds that sing frequently; Farnsworth et al. 2002) whereas others are more general and can be applied in many situations and across a variety of taxa (eg Buckland et al. 2001) Experience with these approaches has made it apparent that different species have different detectabilities, and that various behavioral (eg clump size; Graham and Bell 1989) ecological (eg habitat; Bayliss and Yeomans 1989) and methodological (eg bystander bias; Nichols et al. 2000) factors can influence the detection of individuals.