Table 1 Possible system errors that can affect the utility of animal dead-reckoning within the ‘tilt-compensated compass’ framework

Derived heading

Errors in deriving static acceleration (postural) estimates

During bouts of high centripetal (turning) acceleration

Free-falling behaviour

[65, 72,73,74,75]

Gyro-integrated data [cf. 76, 77]

Using Euler angles (angle of rotation about each axis of a given coordinate system)

The orientation of the device with respect to the earth’s frame of reference (cf. Additional file 1: Text S2) can only be defined reasonably at angles less than perpendicular or less than a 180° inversion (dependent on pitch and roll equations used—cf. “VPC dead-reckoning procedure in R” section) from their longitudinal and lateral axes of ‘normal’ posture (otherwise, unstable measures arise from the Gimbal lock singularity complex [cf. 78], whilst x, y and z values can become inversed and/or represent different ‘surge’, ‘sway’ and ‘heave’ planes)

[21, 79,80,81,82]

Quaternion-estimated heading [cf. 83, 84, 85]

Tag placement/dislodgment

In line with the above—range for accurate angular (pitch and roll) measures are restricted in one or more dimensions

Heading will be biased according to the degree of displacement about the z-axis

[19, 86]

Ensure tag orientation is noted during deployment and retrieval operations (and subsequently used in corrections)

Variations in the strength and declination of magnetic fields

Animals that undertake long journeys (regionally/globally)

Environmental and man-made magnetic noise (iron distortions)

[87]

Ensure at least one magnetic calibration procedure is carried out (see Additional file 1: Text S3 for details) and apply magnetic declination offset to heading values where required

Derived speed

Deviations of the DBA ~ speed relationship

Load bearing

Moving over a deformable substrate/changeable incline

Changing gait

Moving within fluid media

Gliding/thermalling behaviour

[12, 65, 68, 69, 88,89,90,91]

Iteratively modulate the gradient and/or intercept within the DBA–speed linear regression according to environmental circumstance and mode of movement [cf. 68]

By-pass DBA (e.g., use speed/acoustic sensors, step/tail-/wing beat frequency, vertical speed, etc. [e.g., 47])

Both

External forces (e.g., current vectors in air- and water flow)

Decreases the signal-to-noise ratio of motion sensor data

Affects the relationship between an animal’s (longitudinal axis) direction of travel from their true vector of travel

Some animals do not always move in the same direction as their anterior–posterior axis

[11, 15, 46, 92]

Smooth postural (and pre-derivative data)/DBA estimates [cf. 67]

Incorporate current flow vectors within the dead-reckoning procedure