Strange results for parametric trophic position

[Katharina Peters]

Hi,

I have an issue with the parametric calculation of the trophic position. The problem is that I get a mean value of 2.46, even though when I run the Bayesian model with these data it shows that the lower 95%CL is 3.43 and median is 3.66.

 

Have you had this issue before?

 

Many thanks,

Katharina

 

 

Katharina J. Peters, PhD

Research Associate

Cetacean Ecology Research Group | School of Natural and Computational Sciences | Massey University

Private Bag 102 904, North Shore, Auckland 0745, New Zealand 

* k.pe...@massey.ac.nz

[Claudio Quezada Romegialli]

Hi Katharina

That would be odd. Could you share the data you have so I can check it and see what's going on?

Cheers

Claudio

--
To download tRophicPosition package visit https://github.com/clquezada/tRophicPosition
---
You received this message because you are subscribed to the Google Groups "tRophicPosition" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trophicposition-s...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/trophicposition-support/1098D20B-40C5-47E3-AEFD-3F9337154FFD%40contoso.com.

[Katharina Peters]

Hi Claudio,

 

Did you have any luck finding out what gave those strange results?

 

Cheers,

Katharina

 

Katharina J. Peters, PhD

Research Associate

Cetacean Ecology Research Group | School of Natural and Computational Sciences | Massey University

Private Bag 102 904, North Shore, Auckland 0745, New Zealand 

* k.pe...@massey.ac.nz

 

From: Katharina Peters <K.Pe...@massey.ac.nz>
Date: Thursday, 25 November 2021 at 9:28 am
To: Claudio Quezada Romegialli <claudio...@upla.cl>
Subject: Re: [tRophicPosition-support] Strange results for parametric trophic position

 

Hi Claudio,

 

here are the data and the script I was using. I’m comparing two groups (cluster 1 and 2) of consumers and use two baselines (krill and SPOM), and I do this once using the Post TDF and once using TDF values from Borrell. The issue I described happens for both TDFs but only with cluster 2.

 

Any help is greatly appreciated 😊

 

Cheers,

Katharina

 

Katharina J. Peters, PhD

Research Associate

Cetacean Ecology Research Group | School of Natural and Computational Sciences | Massey University

Private Bag 102 904, North Shore, Auckland 0745, New Zealand 

* k.pe...@massey.ac.nz

 

[Claudio Quezada Romegialli]

Hi Katharina

I haven't had time until now, but I just looked at it and perhaps the two baselines model is not appropriate. As you have SPOM (baseline 1, lambda = 1) but also krill (baseline 2, lambda = 2) you are getting two baselines but they represent the same path of energy (pelagic) thus it is somewhat strange/odd using them to get alpha. Here are the results I get with parametricTP and lambda = 1:

It's ok having these rare results as one baseline is below the other, they do not represent different paths of energy. It says that 2 baselines and full models are not appropriate.

You could try calculating trophic position using each baseline (and different lambda) separately, but I'm not sure if that's what you want. Here are the results using Post's TDF and SPOM as baseline 1 (with lambda = 1):

They do look quite different

Let me know what you think. I am working on an individual model for trophic position. I will try to update both the GitHub and CRAN version as soon as possible.

Cheers

Claudio

[Katharina Peters]

Hi Claudio,

Thank you very much for taking the time to look at this. I think I will go ahead using only one baseline (krill) as that makes sense for my data. I wasn’t quite sure if it would work with two baselines, so you’ve helped me a lot!

All the best,

Katharina

 

Katharina J. Peters, PhD

Research Associate

Cetacean Ecology Research Group | School of Natural and Computational Sciences | Massey University

Private Bag 102 904, North Shore, Auckland 0745, New Zealand 

* k.pe...@massey.ac.nz

 

From: Claudio Quezada Romegialli <claudio...@upla.cl>
Date: Saturday, 11 December 2021 at 12:40 pm
To: Katharina Peters <K.Pe...@massey.ac.nz>
Cc: "trophicposi...@googlegroups.com" <trophicposi...@googlegroups.com>
Subject: Re: [tRophicPosition-support] Strange results for parametric trophic position

 

Hi Katharina

 

I haven't had time until now, but I just looked at it and perhaps the two baselines model is not appropriate. As you have SPOM (baseline 1, lambda = 1) but also krill (baseline 2, lambda = 2) you are getting two baselines but they represent the same path of energy (pelagic) thus it is somewhat strange/odd using them to get alpha. Here are the results I get with parametricTP and lambda = 1:

It's ok having these rare results as one baseline is below the other, they do not represent different paths of energy. It says that 2 baselines and full models are not appropriate.

 

You could try calculating trophic position using each baseline (and different lambda) separately, but I'm not sure if that's what you want. Here are the results using Post's TDF and SPOM as baseline 1 (with lambda = 1):

 

They do look quite different