PS 49-99
A comparison of traditional avian survey methods and recorded acoustical surveys in three habitat types

Wednesday, August 12, 2015
Exhibit Hall, Baltimore Convention Center
Sophia Christel, School of Earth, Energy, and Environmental Sciences, Stanford University, Stanford, CA

With climate change threatening ecosystems worldwide, recognition and monitoring of indicator species is a useful tool for early identification of climate-induced ecological changes.  Birds are excellent indicator species, but the methods traditionally used for surveying bird communities (point counts and transects) can be impractical in certain terrains (e.g. dense forest), and require highly trained field scientists for accurate bird identification.  Acoustical recording is increasingly used for surveying birds, and shows great potential.  However few studies have addressed the potential of habitat type or human presence to affect the reliability of this new method.  I sought to assess the efficacy of recordings for surveying avian communities in different habitat types, and without human interference.

I conducted species richness surveys using point counts, transects, and recordings in three distinct habitats within the 1200-acre Jasper Ridge Biological Preserve in Woodside California. Surveys were conducted between February 22nd and June 5th 2014.  All counts, including recordings, began at dawn, and recorders functioned autonomously, eliminating human interference.  Recordings were listened to with a filter to reduce ambient noise, but no other manipulations or software were used to enhance analysis.


Across all three habitats, in-person and recorded point counts yielded similar counts of species (to 95% significance level), but in-person transects produced higher counts (p = 3.03x10-8) than recorded “transects” (audio duration similar to the duration of the in-person transect).  This pattern held for the riparian and woodland habitats (in-person transects higher with respective p-values of 7.83x10-4 and 4.59x10-15), but in the chaparral the in-person and recorded transects did not produce significantly different richness counts.  The differences between point-type (in-person and recorded) and transect-type methods were expected and can be largely attributed to survey duration, which was approximately seven times longer for transects than for points.  Some variation is also due to variable habitat heterogeneity; more heterogeneous habitats showed greater variation in richness counts (and species lists) between point and transect, and between in-person and recorded transect counts.  My results suggest that, at least at Jasper Ridge, the use of an acoustical recorder could effectively replace in-person point counts with no significant deterioration of data quality, and no need for software-enhanced analysis.  My results do not suggest the use of recorders for the replacement of in-person transects, except in the chaparral.