Density estimates at the sample level

[Dylan Brown]

Hi everyone,

I have 2 questions:

#1

I am currently running an analysis on line transect data collected for
pronghorn and mule deer. I have successfully attained density estimates for
the global and stratum levels. However, I recently realized that it would
benefit me to calculate density estimates at the sample level as my design
is one of repeated measures. If I can make DISTANCE produce density
estimates at the sample level then I will have a density estimate for each
repeated measure and be able to run a separate analysis with this data. I
become confused when I have to tell DISTANCE how to calculate cluster size
in the model definition window. I assumed it should be calculated at the
sample level. Is this correct?

#2

When I try to run the analysis I receive a series of warning messages
stating that encounter rate cannot be estimated empirically when estimating
encounter rate by sample. DISTANCE then sets encounter rate variance to
zero. Is this a problem? If not is there a way to make DISTANCE show all
the warning messages rather than truncating them at 30 so I know if
something else went wrong during the analysis?

Thanks in advance.

Dylan Brown

--
distance-sampling list home page:
http://www.jiscmail.ac.uk/lists/distance-sampling.html
To unsubscribe, go to the home page and click "Join or leave the list",
or email jisc...@jiscmail.ac.uk with the following message body:
leave distance-sampling
Before posting to the list, read the list guidelines at
http://www.jiscmail.ac.uk/files/distance-sampling/
distance-sampling.welcome
Any questions, email the list owner at
distance-sam...@jiscmail.ac.uk
--

[Distance Development]

Hi Dylan, (list)

#1

To get cluster size at the sample level, you need to tick the
appropriate box in the Model Definition | Estimate | Quantities to
estimate and levels of resolution section. However, it usually doesn't
make sense to do size bias regression [ie regression of log(cluster
size) against estimated g(x)] at this level, as there are too few
samples for the regression to do anything sensible. Also, it seems
unlikely that mean cluster size truly varies at this level in most
circumstances. So, it often makes more sense to estimate population
mean cluster size at a higher level (such as stratum). However, you can
do it if you want to. (Note that if you do estimate cluster size at a
higher level, then you may run into issues estimating variance as the
transect-level estimates are not independent. This is also true for
estimating the detection function at higher levels, and has been
discussed in previous posts I expect. If this may be a problem for you,
and you can't find something sensible in previous posts, let me know and
I'll write more.)

#2

It's not a problem for you, since you aren't interested in the estimates
of encounter rate variance by line. I guess you're going to subtract
the density or encounter rate estimates between time periods on the same
line (or some other more sophisticated method of repeated measures
analysis) and then use the between-line variation in encounter rate or
density difference (or something like that for the more sophisticated
methods) as your measure of variance.

There's no way to stop it only colour coding the first 30 errors and
warnings, but all of them will be shown in the log window, and the title
bar that separates the log window from the list of log messages tells
you the total number (e.g., it says "Log messages - 123 warnings, 0
errors").

Hope this helps,

- Len

--
Len Thomas l...@mcs.st-and.ac.uk http://www.creem.st-and.ac.uk/len/
Centre for Research into Ecological and Environmental Modelling
The Observatory, University of St Andrews, Scotland KY16 9LZ
Tel: (0)1334-461801 Fax: (0)1334-461800 Admin tel: (0)1334-461842
The University of St Andrews is a charity registered in Scotland

[Dylan Brown]

[Distance Development]

--
Len Thomas l...@mcs.st-and.ac.uk http://www.creem.st-and.ac.uk/len/
Centre for Research into Ecological and Environmental Modelling
The Observatory, University of St Andrews, Scotland KY16 9LZ
Tel: (0)1334-461801 Fax: (0)1334-461800 Admin tel: (0)1334-461842
The University of St Andrews is a charity registered in Scotland