AB 8:299-304 (2010)  -  DOI: https://doi.org/10.3354/ab00239

Data-processing artefacts in three-dimensional dive path reconstruction from geomagnetic and acceleration data

Kozue Shiomi1,2,*, Tomoko Narazaki3, Katsufumi Sato2, Kenichiro Shimatani4, Nobuaki Arai5, Paul J. Ponganis6, Nobuyuki Miyazaki7

1Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113-8657, Japan
2International Coastal Research Center, Ocean Research Institute, The University of Tokyo, 2-106-1 Akahama, Otsuchi, Iwate 012-1102, Japan
3Graduate School of Frontier Sciences, The University of Tokyo, 1-15-1 Minamidai, Nakano, Tokyo 164-8639, Japan
4The Institute of Statistical Mathematics, 10-3 Midori-cho, Tachikawa, Tokyo 190-8562, Japan
5Graduate School of Informatics, Kyoto University, 36-1 Yoshida-honmachi, Sakyo, Kyoto 606-8501, Japan
6Center for Marine Biotechnology and Biomedicine, Scripps Institution of Oceanography, University of California San Diego, La Jolla, California 92093-0204, USA
7Ocean Research Institute, The University of Tokyo, 1-15-1 Minamidai, Nakano, Tokyo 164-8639, Japan

ABSTRACT: Tri-axis magnetism and acceleration data loggers have recently been used to obtain time-series headings and, consequently, the 3-dimensional dive paths of aquatic animals. However, problems may arise in the resulting calculation process with multiple parameters. In this study, the dive paths of loggerhead turtles and emperor penguins were reconstructed. For both species, apparently unrealistic movements were found. Time-series heading data of turtles showed small regular fluctuations synchronous with stroking. In the dive paths of penguins, infrequent abrupt changes in heading were observed during stroke cycles. These were unlikely to represent true behaviours according to observations of underwater behaviour and tri-axis magnetism and acceleration data. Based on the relationship between sampling frequency and frequency of body posture change, we suggest that (1) the changes in the animals’ posture concurrent with strokes and (2) the mismatched treatment (i.e. filtering and non-filtering) of the acceleration and magnetism data caused the artefacts. These inferences are supported by the results of simulations. For data sets obtained at a given sampling frequency, the error pattern in calculated dive paths is likely to differ depending on the frequency and amplitude of body posture changes and in swim speed. In order to avoid misinterpretation, it is necessary to understand the assumptions and inherent problems of the calculation methods as well as the behavioural characteristics of the study animals.


KEY WORDS: 3D dive path · Low-pass filter · Sampling frequency · Stroke activity · Data logger · Dead-reckoning


Full text in pdf format  
Cite this article as: Shiomi K, Narazaki T, Sato K, Shimatani K, Arai N, Ponganis PJ, Miyazaki N (2010) Data-processing artefacts in three-dimensional dive path reconstruction from geomagnetic and acceleration data. Aquat Biol 8:299-304. https://doi.org/10.3354/ab00239

Export citation
Mail this link - Contents Mailing Lists - RSS
- -