Modelled mid‐trophic pelagic prey fields improve understanding of marine predator foraging behaviour

Biophysical interactions are influential in determining the scale of key
ecological processes within marine ecosystems. For oceanic predators, this
means foraging behaviour is influenced by processes shaping the
distribution of prey. However, oceanic prey is difficult to observe and
its abundance and distribution is regionally generalised. We use a
spatiotemporally resolved simulation model to describe mid-trophic prey
distribution within the Southern Ocean and demonstrate insights that this
modelled prey field provides into the foraging behaviour of a widely
distributed marine predator, the southern elephant seal. From a five-year
simulation of prey biomass, we computed climatologies of mean prey biomass
(average prey conditions) and prey biomass variability (meso-scale
variability). We also compiled spatially gridded metrics of seal density
and diving behaviour from 13 years of tracking data. We statistically
modelled these metrics as non-linear functions of prey biomass (both mean
and variability) and used these to predict seal distribution and
behaviour. Our predictions were consistent with observations (R2adj =
0.23), indicating that seals aggregate in regions of high mesoscale
activity where eddies concentrate prey. Here, seals dived deeper (R2marg =
0.12, R2cond = 0.51) and spent less time hunting (R2marg = 0.05, R2cond =
0.56), likely targeting deep but profitable prey patches. Seals generally
avoided areas of low eddy activity where prey was likely dispersed. Most
seals foraged south of the Subantarctic Front, despite north of the front
exhibiting consistently high simulated prey biomasses. This likely
reflects seal prey or habitat preferences, but also emphasises the
importance of mesoscale prey biomass variability relative to regionally
high mean biomass. This work demonstrates the value of coupling
mechanistic representations of prey biomass with predator observations to
provide insight into how biophysical processes combine to shape species
distributions. This will be increasingly important for the robust
prediction of species’ responses to rapid system change.