Table 4.

DIC model selection results for the δ¹³C, δ¹⁵N, δ³⁴S stable isotope (CNS) Bayesian mixing models (MixSIAR) used to estimate diets of leopard seals based on stable isotope values from plasma and red blood cells (RBC) sampled from seals at Cape Shirreff, Livingston Island in 2013, 2014, and 2017 (n = 23)

Model	DIC	∆DIC	Median variance (σ)
CNSplasma ~ (1|Year)	32.77379	–	σyear = 0.772
CNSplasma ~ Sex + (1|Year)	33.85499	1.0812	σyear = 0.911 σSex = 1.684
CNSplasma ~ (1|Year/Seal ID)	40.84313	8.0693	σyear = 0.753 σSeal ID = 0.103
CNSplasma ~ Mass +  (1|Year)	45.78776	13.014	σyear = 0.764 σMass = 0.091
CNSRBC ~ (1|Year)	29.58741	–	σyear = 0.193
CNSRBC ~ Sex + (1|Year)	32.38981	2.8024	σyear = 0.210 σSex = 0.281
CNSRBC ~ Mass + (1|Year)	39.74659	10.1592	σyear = 0.195 σMass = 0.063
CNSRBC ~ (1|Year/Seal ID)	107.9329	78.3455	σyear = 0.178 σSeal ID = 0.134

Median variance (σ) reflects the 50% quantiles of variance for each model effect. Year and seal ID were random effects, sex was a fixed effect, and mass was a continuous effect

CNS_plasma: plasma tissue, CNS_RBC: red blood cell tissue