Track reconstruction

A common interest is to be able to make inferences based on where tagged animals have been

This will provide relevant information regarding ecology and behaviour

It allows detailed information about movement, opening the door to various aspects related to movement ecology

We will refer to 2D location \( (x,y) \) or 3D location \( (x,y,z) \)

pseudotrack - obtained from dead-reckoning (initial point + directions + speeds)

georeferenced track - “anchored”“ with known locations (besides the first one)

For that purpose, we can distinguish two types of tag data

• data with direct (even if noisy) information on location \( (x,y) \) or \( (x,y,z) \)) - like DTag data
• data without direct information about location (except a starting point) - like a GPS enabled tag

The world is not black and white, so there are a few hybrid instances of location information

• obtained from the tag (e.g. a GPS reading when animal comes to the surface)
• from an independent source (e.g. georeferenced visual sightings when animal surfaces)

Dead reckoning, or ded (from deduced) reckonig: obtaining a position from an initial position, a direction and a speed

A car was traveling in a straight line from A to B, separated by 40 km. Given it was half way through 10 minutes after it started, where was the car 15 minutes after it started?

Each new position is based on the previous estimated position

Errors do not cancel each other, so a track obtained by dead reckoning will drift from the original track

Track accuraccy will decrease with time - see practical exercise 1

Given a set of locations, what is the most likely track?

Given a set of locations, what is the most likely track?

straight lines between points?

Given a set of locations, what is the most likely track?

allow some inercia

Given a set of locations, what is the most likely track?

allow measurement error

Given a set of locations, what is the most likely track?

choice is driven by assumptions on the movement and on the measurements: method used will have an impact on inferences made!

R code can be included in a slide like so. It will be executed when the presentation is compiled. By default, output generated by the code is also included in the slide.

summary(cars)

     speed           dist       
 Min.   : 4.0   Min.   :  2.00  
 1st Qu.:12.0   1st Qu.: 26.00  
 Median :15.0   Median : 36.00  
 Mean   :15.4   Mean   : 42.98  
 3rd Qu.:19.0   3rd Qu.: 56.00  
 Max.   :25.0   Max.   :120.00  

To include code but not its output, use the R code chunk option “results='hide'”.

summary(cars)

You can also display code without evaluating it by including the option eval=FALSE in the R code chunk. For example, the matlab code below wouldn't work in R, but we can display it anyway.

Learn more about R code chunk options online: https://yihui.name/knitr/options.

my_matrix = eye(7);

We can have bold and italicized text (and any other formatting you can specify in RMarkdown: see https://www.rstudio.com/wp-content/uploads/2016/03/rmarkdown-cheatsheet-2.0.pdf).

You can also include equations inline: \( E = MC^2 \) (be careful not to leave whitespace adjacent to the $)

Or in display mode:

\[ e^{i\phi} = \text{cos}(\phi) + i\text{sin}(\phi) \]

For more help and examples, check out https://support.rstudio.com/hc/en-us/articles/200486468-Authoring-R-Presentations.