differences between estimates using dsm and another method (kriging)

[Kevin Chan]

I've been doing a comparison of the dsm distance sampling method and the finite population block kriging method using the sptotal package on moose in my region and I have some questions about differences I’ve noticed in the results.

 We have done 2 separate surveys over the past 5 years which isn’t great temporally but historical data suggest there shouldn’t be huge variation in the population density. The FPBK was run trying to count all moose within 4km x 4km grid cells and estimating over the study area while the distance sampling survey was flown with 10km spacing across a much larger area and projecting onto the same grid but expanded out to fill the entirety of the distance sampling study area.

 When I predict the dsm over the same area as the FPBK grid I find that the estimate for the kriging analysis is about 2 to 2.5 times higher than the distance sampling survey even though we don’t have a correction factor for moose that were not detected. My questions are what could be causing the difference I’m seeing here and are there issues in the distance sampling survey design (e.g. line spacing) that could cause underestimation (if that is what the issue is). Other examples comparing these 2 methods have shown comparable results though usually lower estimates from the distance sampling methods, though this is the largest difference I’ve seen. 

[Stephen Buckland]

How are complete counts carried out in a 4x4 square?  How do you avoid double-counting, due to animal movement during the count?  And might more animals move into the square during the count?

For the aerial line transects, can animals be missed on the line because they are hidden by canopy?

Steve Buckland

---

From: distance...@googlegroups.com <distance...@googlegroups.com> on behalf of Kevin Chan <wkkc...@gmail.com>
Sent: 01 May 2026 16:10
To: distance-sampling <distance...@googlegroups.com>
Subject: {Suspected Spam} [distance-sampling] differences between estimates using dsm and another method (kriging)

 

You don't often get email from wkkc...@gmail.com. Learn why this is important

I've been doing a comparison of the dsm distance sampling method and the finite population block kriging method using the sptotal package on moose in my region and I have some questions about differences I’ve noticed in the results.

 We have done 2 separate surveys over the past 5 years which isn’t great temporally but historical data suggest there shouldn’t be huge variation in the population density. The FPBK was run trying to count all moose within 4km x 4km grid cells and estimating over the study area while the distance sampling survey was flown with 10km spacing across a much larger area and projecting onto the same grid but expanded out to fill the entirety of the distance sampling study area.

 When I predict the dsm over the same area as the FPBK grid I find that the estimate for the kriging analysis is about 2 to 2.5 times higher than the distance sampling survey even though we don’t have a correction factor for moose that were not detected. My questions are what could be causing the difference I’m seeing here and are there issues in the distance sampling survey design (e.g. line spacing) that could cause underestimation (if that is what the issue is). Other examples comparing these 2 methods have shown comparable results though usually lower estimates from the distance sampling methods, though this is the largest difference I’ve seen. 

--
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.
To view this discussion visit https://groups.google.com/d/msgid/distance-sampling/28b5bb5b-8f7f-42f6-80c6-1a4a1d85706dn%40googlegroups.com.

[Kevin Chan]

We try to be as complete as possible but we have about 700m spacing for our line transects. we usually are able to see the same moose on the next line or 2 over and we account for that with our gps. we generally cover a cell within 15-20 minutes and they generally do not move that much due to deep snow at the time of our survey. I'm sure some animals were missed on our survey and we didn't account for imperfect detection as some of the newer variants of the block kriging try to do due to time constraints. typically though our landscape is fairly open and our trees are smaller as i am sampling close to the arctic circle. 

[Stephen Buckland]

Ok. So what truncation distance do you use in your ds analysis? It sounds like detection would probably stay at or close to certain to well beyond the next line. And in your complete count approach, can you easily identify which animals have already been counted from a previous line? There’s no obvious reason that I can see why the 2 methods differ so much.

Stephen T. Buckland

CREEM, The Observatory, Buchanan Gdns, St Andrews KY16 9LZ, Scotland

e-mail st...@st-andrews.ac.uk

The University of St Andrews is a charity registered in Scotland:No SC013532

---

From: distance...@googlegroups.com <distance...@googlegroups.com> on behalf of Kevin Chan <wkkc...@gmail.com>

Sent: Saturday, May 2, 2026 4:14:01 AM
To: distance-sampling <distance...@googlegroups.com>
Subject: Re: {Suspected Spam} [distance-sampling] differences between estimates using dsm and another method (kriging)

 

To view this discussion visit https://groups.google.com/d/msgid/distance-sampling/1b2bfaa8-8a23-4543-a764-3a7532c4715fn%40googlegroups.com.