This is a genuinely different model from the 3D per-polyp version, not a flattened copy of it. Instead of 520 independent point-polyps, the reef flat is a continuous 2D field viewed from above, and its two state variables — reactive-oxygen index ROS(x,y) and symbiont density N(x,y) — diffuse between neighbouring patches of tissue and are solved on a grid with finite differences every frame:
∂ROS/∂t = k₁·I(x,y)·max(0, T_local(x,y) − T_opt(x,y))² − k₂·ROS + D_ROS·∇²ROS
∂N/∂t = −k₃·max(0, ROS − ROS_thresh)·N + k₄·N·(1 − N/N₀) + D_N·∇²N
I(x,y) = I₀ · exp(−k_att · depth(x,y)) (Beer–Lambert light attenuation)
T_local(x,y) = T_sea − β · depth(x,y) (shallow-lagoon vs deep-channel cooling)
The reaction terms (k₁..k₄, ROS_thresh) are the same tested constants as the 3D per-polyp engine, so the two are directly comparable. What's new here is the coupling to a real reef-flat bathymetry map: light intensity falls off exponentially with local water depth (the standard Beer–Lambert law used in reef optics, k_att ≈ 0.12 m⁻¹ for clear lagoon water), and deeper channels run measurably cooler than shallow flats because they mix more slowly with the sun-warmed surface layer — the "depth refugia" effect documented on real reefs. Together they mean a single sea-temperature reading produces very different local stress at different points on the same reef, without needing any per-cell randomisation.
The diffusion terms D_ROS·∇²ROS and D_N·∇²N are the other new mechanism: reactive-oxygen and free-living symbiont cells both spread a little through the shared boundary layer and connected tissue between adjacent coral heads, so a bleaching patch grows outward as a genuine propagating front rather than flickering independently cell by cell — visible as the tissue-health map smoothing and spreading rather than looking like static noise.
- Sea temperature sets T_sea; deep channels lag behind it via the −β·depth cooling term, so they bleach later and recover a genetic/thermal refugium.
- Light irradiance sets I₀; shallow patches see nearly all of it (little water to attenuate through) while deep channels are shielded — this is why reef flats bleach worse than nearby deep water even at the same sea temperature.
- Bathymetry view shows the depth map driving both effects; switch back to see how closely the resulting bleaching pattern tracks the seafloor topography.
- Cool to 26 °C tests recovery: N regrows via the logistic term wherever enough symbionts survived, spreading through the diffusion term into fully-bleached neighbours too.