sim07 — Transport Coupling

Open the interactive visualization — renders the grid heatmap (play/pause, step slider) and all metric charts from results.json.

One-line: Tests whether a structure-sourced transport field T with a mass threshold M_c produces the trace→actor crossing (H7) as a phase transition — extending sim06's minimal Grassé stigmergy with the Mahadevan mechanism in lumped form.

Result: NULL. No phase transition in M_c. Sweeping M_c from inert to fully active decreases stability monotonically (0.876 → 0.739) and fragments the structure (57 → 128 pillars). The crossing detector never fires for any M_c or transport_coupling. Structure- sourced scalar transport alone is insufficient; venting disperses the very cue that recruits deposits.

What sim07 adds to sim06

The only addition is a transport field T on the grid:

• T is sourced by structure above M_c: T += transport_gain · max(0, M − M_c). Below M_c, no sourcing — the structure is inert (sim06 regime).
• T diffuses (toroidal) and decays.
• T → P coupling vents pheromone from high-T (saturated) to low-T (gaps): P += transport_coupling · (T_neighbor_avg − T_local).

Agents are unchanged from sim06 (Grassé deposit rule, pheromone following). No self- maintenance emission — the transport field IS the new mechanism.

Sign correction. The DESIGN.md sketch wrote the coupling as (T_local − T_neighbor_avg), which would increase pheromone at structure (positive feedback — wrong direction). The prose ("saturated pillars shed their pheromone to their flanks") describes venting, so the implemented sign is (T_neighbor_avg − T_local). This correction is documented in the 2026-07-27 daily report.

Results (default run, seed 42)

Transport modestly consolidates (fewer pillars, more mass) but does NOT cross — stability drops slightly (criterion 1 needs ≥ 0.90).

Perturbation / self-repair

Both recover mass after a 25%-area damage at 60% of steps — but recovery is driven by the deposit rule (termites wander back), not by transport. The circularity safeguard is not satisfied: repair does not track T. This is consistent with the null — T is not the causal layer.

The M_c sweep (the phase-transition test)

No jump. A smooth degradation, not a phase transition. A transport_coupling sweep (0.0 → 0.80) confirms: no value crosses, stability stuck ~0.85.

Why the null is informative

The negative feedback is real but its effect has the wrong sign for consolidation: venting pheromone away from saturated pillars disperses the cue that attracts deposits, fragmenting rather than consolidating. A lumped linear advection of a scalar cue does not reproduce the Mahadevan mechanism, where directed flow carries the cue along channels to where building should continue. The mechanism needs geometry (directed flow), not just a venting scalar.

H7 refinement (from this null)

The crossing is not produced by "structure sources a scalar transport field" alone. Two candidates remain:

1. Directed transport — channel geometry that carries cue to building fronts, not away from them (the Mahadevan mechanism's directionality, lost in the lumped scalar).
2. External multi-rate driver — the diurnal oscillation (H4) the structure rectifies into directed flow; sim07's lumped T has no external clock. Candidate sim08.

Run

cd simulations
uv run python3 sim07_transport_coupling/sim07.py selftest     # sanity checks
uv run python3 sim07_transport_coupling/sim07.py run         # baseline+transport+perturb -> results.json
uv run python3 sim07_transport_coupling/sim07.py sweep_plot  # M_c sweep -> output/*.png

Visualization: visualize.html (renders results.json + the M_c sweep).

Cross-references

• [[../concepts/environmental-physics-coupling]] — the concept this sim tests
• [[../concepts/stigmergic-consolidation]] — the negative-feedback gap
• [[../hypotheses/hypotheses.md]] — H7 refined
• sim06 — the null result this sim builds on
• King, Ocko & Mahadevan (PNAS 2015); Ocko, Heyde & Mahadevan (PNAS 2019)