Evaluating the performance of the Bayesian mixing tool MixSIAR with fatty acid data for quantitative estimation of diet

Abstract

We test the performance of the Bayesian mixing model, MixSIAR, to quantitatively predict diets of consumers based on their fatty acids (FAs). The known diets of six species, undergoing controlled-feeding experiments, were compared with dietary predictions modelled from their FAs. Test subjects included fish, birds and mammals, and represent consumers with disparate FA compositions. We show that MixSIAR with FA data accurately identifies a consumer’s diet, the contribution of major prey items, when they change their diet (diet switching) and can detect an absent prey. Results were impacted if the consumer had a low-fat diet due to physiological constraints. Incorporating prior information on the potential prey species into the model improves model performance. Dietary predictions were reasonable even when using trophic modification values (calibration coefficients, CCs) derived from different prey. Models performed well when using CCs derived from consumers fed a varied diet or when using CC values averaged across diets. We demonstrate that MixSIAR with FAs is a powerful approach to correctly estimate diet, in particular if used to complement other methods.

Complementary use of stable isotopes and fatty acids for quantitative diet estimation of sympatric predators, the Antarctic pack-ice seals

“…The equivalent to discrimination values for FA data are calibration coefficients, which are used to account for selective FA metabolism and hence changes in the predator’s FA proportions relative to diet consumed (Iverson et al 2004 ; Guerrero and Rogers 2020 ). We used the calibration coefficients derived from captive harbour seals, Phoca vitulina , fed herring for over a year (Rosen and Tollit 2012 ), and applied them to our three study species.…”

“…We used the calibration coefficients derived from captive harbour seals, Phoca vitulina , fed herring for over a year (Rosen and Tollit 2012 ), and applied them to our three study species. Prior to the Bayesian mixing model analysis, the FA signatures of prey items were multiplied by the calibration coefficients of harbour seals; thus, taking prey signatures to the predator space (Bromaghin 2017 ; Guerrero and Rogers 2020 ). This was done prior to analysis since MixSIAR treats discrimination as additive values, whereas FA calibration coefficients are multiplicative values.…”

“…For diet estimation analyses, we used dietary and extended dietary FAs, according to Iverson et al ( 2004 ), but discarded those FAs with high calibration coefficients (> 3), since they indicate greater influence of consumer metabolism or preferential accrual rather than a reflection of diet (Guerrero and Rogers 2020 ). Thus, the number of FAs used for analyses, for each consumer species, was 12.…”

“…It is usual that several prey species share similar SI values and therefore they need to be combined into a single prey group (e.g. Hückstädt et al 2012 ; Goetz et al 2016 ; Botta et al 2018 ); making it difficult to obtain a good taxonomic resolution of the predator’s diet (Guerrero and Rogers 2020 ). Considering these limitations, the combined use of these two biotracers has been proposed as a more accurate technique to estimate diet composition (Dalsgaard et al 2003 ).…”

“…Guerrero and Rogers ( 2020 ) tested the performance of MixSIAR with FA data for diet estimation of captive animals undergoing feeding experiments, including different consumer species such as fish, birds and marine mammals, with different diet compositions. MixSIAR accurately estimated consumers’ main dietary items, and identified shifts in diet and absent prey (Guerrero and Rogers 2020 ). To date, however, MixSIAR has not been used to study the diet of wild animals based on their FAs.…”