Ghosts in the data: false detections in VEMCO pulse position modulation acoustic telemetry monitoring equipment. Issue 1 (December 2015)

• Record Type: Journal Article
• Authors: Simpfendorfer, Colin; Huveneers, Charlie; Steckenreuter, Andre; Tattersall, Katherine; Hoenner, Xavier; Harcourt, Rob; Heupel, Michelle
• Is Part Of: Animal biotelemetry. Volume 3:Issue 1(2015)
• Page Start: 1
• Page End: 10
• Publication Date: 2015-12
• Subjects: False-positive data -- IMOS-AATAMS -- VEMCO VR2W -- Acoustic receiver performance; Animal radio tracking -- Periodicals; Biotelemetry -- Periodicals; Zoology -- Periodicals; Periodicals; Electronic journals; 590.721
• DOI: 10.1186/s40317-015-0094-z
• Languages: English
• ISSNs: 2050-3385

Abstract

Abstract Background False-positive data (better known as "false detections") in VEMCO VR2 acoustic telemetry monitoring studies that use pulse position modulation coding can cause biased or erroneous outcomes in data analysis. To understand the occurrence of false detections in acoustic monitoring data sets, the results of a range test experiment using eight acoustic receivers and 12 transmitters were examined. Results One hundred and fifty one tag ID codes were recorded, 137 of which were identified as likely from false detections, 12 were from test tags, and two were from tagged sharks. False detections accounted for <0.05 % of detections (918) in the experiment. False detection tag ID codes were not randomly distributed amongst the available codes, being more likely to occur at IDs close to tags used in the experiment. Receivers located near the bottom recorded the most false detections and tag ID codes from false detections. Receivers at the same depth did not differ significantly in the mean number of daily false detections. The daily number of false detections recorded by a receiver did not conform to a random pattern, and was not strongly correlated with daily receiver performance. Conclusions In an era of increasing data sharing and public storage of scientific data, the occurrence of false detections is of significant concern and the results of this study demonstrate that while rare they do occur and can be identified and accounted for in analyses.