mizerReef model for a generic Caribbean reef (10 functional groups)
Source:R/data.R
caribbean_10_model.RdA calibrated mizerReef object for a generic Caribbean coral reef
with ten functional groups, parameterised from field data collected at
Karpata Reef, Bonaire (FORCE dataset, Dryden 2017). Functional groups span
six predator guilds, parrotfish, farming damselfish, other herbivores, and
invertebrates. The model includes competitive predation refuge (using
karpata_refuge) and benthic resources (algae and detritus).
Format
A mizerReef object (extends MizerParams) with
10 species, competitive refuge, and algae/detritus components.
Details
Built from caribbean_10_species,
caribbean_10_interaction, and karpata_refuge;
calibrated to match observed biomasses and built with mizerReef 2.0.0 /
mizer 3.1.0. Detritus resource dynamics are tuned via tuneUR() so that
they are genuinely at steady state (dB/dt = 0) for the model's
calibrated abundances. Algae biomass is likewise tuned by tuneUR() to
its own steady state, but for a fixed, literature-informed production
rate (algae_growth, left at newReefParams()'s default of
2000 grams per square meter per year – see
setAlgaeParams) that is not itself retuned to match
consumption, since real algal production is not driven by grazer demand;
an earlier untuned version of this object (algae biomass not yet solved
for its steady state) is archived at
inst/archive/caribbean_10_model_untuned.rda for reference.
Recalibrated in 2026 against a corrected senescence-mortality formula
(reefSenMort()), the same underlying issue that required
recalibrating caribbean_3_model. Biomass, growth, and
reproduction were all re-tuned (verified: repeated reefSteady()
calls leave biomass unchanged). Final biomass is within 79-121\
six FORCE-observed targets (pred_eng, pred_grab,
pred_inv, pred_plank, parrotfish, herbs);
growth (age at maturity) matches its target closely for five of those six,
and reasonably (within a factor of ~3) for pred_grab after
correcting its own age_mat target (see
caribbean_10_species's documentation). eels and
farm_damsel – which have no FORCE-observed biomass target – were
given soft, literature-derived targets and brought up from near-zero to a
modest, non-vanishing presence, deliberately not matched exactly (doing so
costs disproportionate accuracy on the six well-observed groups). A
version of this model that ignores eels/farm_damsel entirely
and matches the six FORCE-observed groups more tightly is archived (not
bundled as a package dataset, to avoid doubling the recalibration burden)
at inst/archive/caribbean_10_model_tight_biomass_fit.rds – load
with mizer::readParams(). The complexity-increases-biomass
relationship central to the associated manuscript/thesis chapter was
explicitly re-verified against this recalibrated model (competitive vs.
non-complex refuge), not merely assumed to still hold. See
inst/scripts/Caribbean_10_model-calibration.R's design note for the
full reasoning, including why growth-matching required tuning one species
at a time rather than mizerExperimental's usual batched workflow.