Segmentation of transects

[Mathilde Le Moullec]

Hei the distance sampling group,

It is a very basic question. I have not found how the segmentation is actually done for Distance and DSM R package.

How am I supposed to characterize a segment by its coordinates, one coordinate at each extremity of the segment or the midpoint coordinate? Using the midpoint coordinate, the projection is not correct and my transects are shortened by half a segment length at each extremity. If I define the start and end coordinates of the segments, the projection is correct. In the latter case, I get two coordinates = two rows in my segdata table for one segment with its respective single covariate (here the proportion of vegetated surface in the covered area of the segment).  How can I have one covariate per row (= per segment) in my table segdata?

Here is an illustration for 3 segments (UTM coordinates): I thus have 4 coordinates but only 3 covariates. 

> segdataSK

   Area Transect.Label Sample.Label        x       y Effort      veg

1     1              1          1-1 426271.0 8742023 1525.3 58.48137

2     1              1          1-2 427796.3 8742023 1525.4 83.00654

3     1              1          1-3 429321.7 8742023 1525.3 73.77207

4     1              1          1-4 430847.0 8742023 1525.3       NA

In the mexdolphins dataset, segdata have a different depth covariate for each row, I must have missed an elementary point of the data organization…

Thanks in advance for the precious help.

Mathilde

[David Lawrence Miller]

Hi Mathilde & other listfolk!

Each segment in a DSM analysis is given the location of its midpoint
(centroid). The "Effort" column accounts for this issue in the
modelling, as it defines the length of the segment -- so this is not a
problem during model fitting.

This will however look a little goofy if you plot the segments, as you
say the segments will be short. For plotting it's worth having an extra
set of columns in your data (e.g. "x_start", "x_end", "y_start",
"y_end") that you can use when you want the correct segment locations to
be plotted.


I hope this helps!

best,
--dave

> --
> You received this message because you are subscribed to the Google
> Groups "distance-sampling" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to distance-sampl...@googlegroups.com
> <mailto:distance-sampl...@googlegroups.com>.
> To post to this group, send email to distance...@googlegroups.com
> <mailto:distance...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/distance-sampling/f8963b25-8dd7-4e14-8d6a-55ae024e7c44%40googlegroups.com
> <https://groups.google.com/d/msgid/distance-sampling/f8963b25-8dd7-4e14-8d6a-55ae024e7c44%40googlegroups.com?utm_medium=email&utm_source=footer>.
> For more options, visit https://groups.google.com/d/optout.

[Mathilde Le Moullec]

Of course it helps! Thanks for this precision David

> <mailto:distance-sampling+unsub...@googlegroups.com>.