The Bayesian methods (e.g. MixSIAR) use the mean and standard deviation of each resource in the estimation of resource contribution in diets.
In my case, I have two main resources: littoral and pelagic.
For the pelagic one, I have only the d13C and d15N mean, resulting in loss of observed variation (SD) for the pelagic resource.
Other resource (multi-sample), littoral, have been processed in order to have the mean and the SD.
Do you know a diet estimation method in line with this case?
Thanks for helping
Best regards
Pierre
Stable Isotope Mixing Models for Estimating Source Proportions
https://www.epa.gov/eco-research/stable-isotope-mixing-models-estimating-source-proportions
IsoConc: Concentration-dependent stable isotope mixing model
ISOCONC1_01.xls is a Microsoft Excel 2000™ spreadsheet which performs calculations for the concentration-weighted stable isotope mixing model. This dual-isotope model takes into account isotopic element concentration differences among the sources in determining the proportional contributions of sources to a mixture. The user supplies the isotopic signatures for each source and the mixture, as well as the isotopic element concentrations for each source. Separate estimates are made of the contributions of each source for total mass and each isotopic element.
Download Isoconc1_01.xls [Excel 2000 spreadsheet](185 K)
Check https://esajournals.onlinelibrary.wiley.com/doi/pdf/10.1002/ecs2.2083
Also check https://bioone.org/journals/journal-of-mammalogy/volume-93/issue-2/11-MAMM-S-158.1/Converting-isotope-values-to-diet-composition--the-use-of/10.1644/11-MAMM-S-158.1.pdf
Is there some literature value you may use instead/in addition of the values you measured? Anyway, whatever its use, the representativity of one value alone is questionable, so having other values may increase the robustness of your mixing model and of your values.
I remember some manuscripts that estimate it using only d13c of the consumer and of the 2 sources of interest....
I will look for it and come back here.
Hi again Pierre,
If i am not wrong, this is the paper: https://esajournals.onlinelibrary.wiley.com/doi/full/10.1890/0012-9658%282002%29083%5B0703%3AUSITET%5D2.0.CO%3B2
And the formula was something like:
Relative contribution of source 1 to the consumer = (d13C of the consumer - d13C of source 2) / (d13C of source 1 - d13C of soource 2)
You also could look for manuscripts presenting SD values for the same or similar sources that you have... And use these values to perform more powerful analyses as for example, MixSIAR in R.
Hi Pierre,
I strongly agree with Pierre Cresson. Searching in literature for stable isotope information of the same resource(s) and adding these values to yours would make your model more valid.
Hi Pierre,
Another thing that you can do in addition to searching values in the literature is to run your model under different values of standard deviations and number of replicates for the pelagic endpoint. Such sensitivity analysis can be used to test how the dispersion of your pelagic endpoint affects your dietary estimations and the credible intervals, which is a good thing to motivate the robustness of the analysis.
Hope this helps :)
To reply your other question... Nope, i don't know :)
In your situation I would use Mix SIAR and run model by model recording the results by assigning a R object to each model. The good thing is that you would need to vary just one (or two) parameters in the model, so with a few models it should be enough.
Cheers!
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics