Refugia -- Protection from Grazing

For a general introduction to consumers, see Consumers. For how surface properties create the physical basis for refugia, see Surfaces.

Why Refugia Are the Difference Between Stability and Collapse

Without refugia, a sufficiently hungry grazer would eat every last algal cell, drive its food to extinction, and then starve to death. The ecosystem would collapse in a boom-bust cycle with no recovery. Refugia — the physical hiding places where prey survives no matter how intense the grazing — are what prevent this. They guarantee a seed population that can regrow once grazing pressure eases, turning a one-way crash into a sustainable oscillation.

This makes refugia one of the most powerful levers in the model. Adding a rough, porous surface to a scenario can shift the outcome from extinction to coexistence, because the crevices and interstices that grazers cannot reach protect enough prey to keep the food web intact. The type, roughness, and amount of surface area in a scenario directly controls how much of each prey population is protected — and that, in turn, controls whether the ecosystem survives. This matches observations from real aquatic habitats: structural complexity promotes species coexistence.

The Core Concept

In a real ecosystem, a grazer cannot eat every last scrap of food. Some algae cells are tucked into tiny crevices in a rock. Some bacteria are buried deep inside a protective biofilm. Some ciliates are hiding in the narrow spaces between gravel particles that a copepod is simply too large to enter. These protected zones are called refugia, and they are critical for ecosystem stability.

In the model, refugia are represented through access fractions. Every grazer-food-surface combination has an access value between 0 and 1. If Copepods have an access of 0.25 to periphyton on gravel, that means Copepods can reach only 25% of the periphyton biomass growing on gravel. The other 75% is in refugia -- hidden in deep interstices between the gravel particles where a 1-2mm copepod simply cannot fit.

The complement of access IS the refugia fraction. Access of 0.25 means 75% refugia.

Maturity-Dependent Access

The access values in the tables below represent mature-biofilm conditions — surfaces with a well-developed EPS scaffold (high biofilm maturity M). On a freshly submerged surface with no biofilm (M = 0), prey are more exposed because there is no EPS structure to hide behind. The model interpolates between a "bare surface" access (higher, roughness-dependent) and the mature values listed here based on each surface's current M. Smooth surfaces like glass show the largest increase when bare, while rough surfaces like ceramic already provide geometric refugia regardless of biofilm state. As the biofilm matures over weeks to months, access gradually decreases toward the static values shown. See Biofilm Maturity for details.

Mouthparts Decide Who Can Reach the Mature Film

The mature-biofilm shelter is not the same for every grazer — it depends on the grazer's mouthparts. A snail drags a hard, mineralized rasping tongue (the radula) across the surface and can grind right through a cemented, tightly-attached film to the substrate beneath; this is why a nerite snail can clear hard green-spot algae that nothing else will touch. A cherry shrimp (Neocaridina), by contrast, grazes with fine, soft, brush-tipped legs. It readily picks off loose, young, or filamentous growth — a shrimp colony keeps soft biofilm and hair algae well cropped — but it cannot rasp through a mature, EPS-cemented mat. Aquarists know this as "shrimp pick at green-spot algae but can't control it."

The model captures this with a per-grazer biofilm penetration trait. Hard scrapers (snails and other radula/mandible grazers) are the reference: they face the full mature-biofilm shelter listed above. Soft-appendage grazers like shrimp face an elevated shelter — the more mature (cemented) the film, the more of it escapes them — so a maturing surface film increasingly slips out of a shrimp's reach even while a snail keeps grazing it. Crucially, this only affects the mature portion of the shelter: on a young, freshly-colonized surface (low M) a shrimp grazes just as effectively as a snail. The practical consequence is that a well-lit surface film in a shrimp-only tank tends to mature into a thin, persistent standing crop rather than being cropped to nothing, whereas the same tank with snails stays scraped clean.

The Basal Layer No Grazer Can Remove

Everything above is about how much of a surface a grazer can reach — an access fraction that scales the grazing rate. But a rate multiplier alone has a subtle failure mode: if grazing always removes some fraction of whatever biomass is left, a hungry enough community can grind a film all the way down to nothing, just more slowly. Real periphyton does not behave that way. A mature film is vertically layered: a loose overstory that grazers strip easily sits on a tightly-cemented basal layer — cells glued flat against the substrate and embedded in the EPS matrix (adnate diatoms like Cocconeis and Achnanthes, prostrate cyanobacteria, the holdfast bases of filamentous green mats). No metazoan grazer — not a hard snail radula, not a shrimp's soft brush — can rasp this basal crust off the substrate. Grazing shifts the community toward these prostrate forms rather than eliminating the film; heavy grazing drives periphyton down to a thin, non-zero grazed residual and holds it there, and a scraped surface re-greens within days by regrowth from the surviving base, not by slow re-seeding from the water column (Steinman 1996; Feminella & Hawkins 1995).

The model represents this with a basal refuge: a maturity-scaled floor on surface-attached algae (benthic green algae and pennate/centric diatoms) below which grazing simply stops. It is the mirror image of the access fraction — the two describe different layers of the same film:

• Access fraction / penetration — a rate control on the overstory: how fast the reachable film is grazed, and how much of the mature film a soft-mouthed shrimp can reach at all.
• Basal refuge — an absolute biomass floor on the understory: as the film is grazed down toward the floor, the reachable fraction smoothly falls to zero, so the crust is never removed no matter how hungry the grazers get.

The floor is a prey property, identical for every grazer — a snail's radula can no more scrape off the deepest EPS-embedded layer than a shrimp's brush can. (This is exactly what aquarists see: a thin green glass film persists under heavy snail and shrimp grazing.) It scales with biofilm maturity M: a freshly-colonized surface (M ≈ 0) has no cemented crust and can be grazed back to bare substrate, while a mature film holds the largest residual. How thin the persistent film is depends on the prey: filamentous green mats hold the most (they build the thickest EPS scaffold), tightly-adnate pennate diatoms a little less, and loosely-attached, mostly-planktonic centric diatoms the least.

The practical consequence is that surface films no longer go extinct under sustained grazing. Instead a grazed green or diatom film settles into a thin, persistent, maturity-scaled standing crop — the grazed residual the user sees on their glass — that regrows quickly whenever grazing eases. This is what turns a grazer–periphyton pair from a one-way crash into a bounded, self-restoring oscillation.

Surface-Specific Access

Each grazer has different access to food on different surfaces. The access value depends on the physical characteristics of the surface and the feeding mode of the grazer.

Copepod Access to Periphyton

Copepods are raptorial feeders that scrape and grab prey from surfaces. Their access depends on how exposed the surface is:

Copepod Access to Bacteria

Bacteria are very small (0.2-2 micrometers) and most live attached to surfaces in biofilms, making them hard for a 1-2mm copepod to reach:

Copepod Access to Ciliates (Microzooplankton)

Ciliates are 10-300 micrometers and many live on or near surfaces rather than free-swimming. Copepods hunt them raptorially:

Daphnia Access to Planktonic Algae

Daphnia are filter feeders working in the water column. Their access to planktonic algae is very high (95% default) because suspended algae are exactly what their filter mesh is designed to capture.

Daphnia Access to Periphyton

Daphnia cannot scrape surfaces. They can only capture periphyton cells that naturally detach and float into the water column. Smooth surfaces release cells more readily than rough ones:

Daphnia Access to Bacteria

Daphnia filter suspended bacteria from the water column. Like periphyton, they can only reach bacteria that are free-floating or loosely attached:

Density-Dependent Refugia

Static access fractions are only part of the story. In reality, when prey populations are small, the remaining individuals tend to occupy the very best hiding spots. As the population grows, those prime spots fill up and newcomers are forced into more exposed positions. The model captures this with density-dependent refugia.

The way it works: the static access value is multiplied by a scaling factor that depends on prey density. At zero prey density, the factor drops to a minimum (nearly all survivors are hiding). As prey density increases, the factor rises toward 1.0 (most prey are exposed). The transition follows a saturation curve controlled by two parameters:

• refugia_K_mol_L (half-saturation density) -- The prey density at which the scaling factor is halfway between its minimum and maximum. Below this density, refugia protection increases steeply. Above it, most additional prey are exposed.

• refugia_min_frac (minimum factor at zero density) -- How much the access shrinks at extremely low prey density. A value of 0.02 means that at near-zero density, effective access drops to 2% of the static access value.

Example: Copepod Hunting Periphyton

Copepods have static access of 0.80 to periphyton (default surface), with density-dependent refugia parameters: K = 5e-4 mol/L, minimum fraction = 0.02.

• At high periphyton density (well above K): effective access is close to 0.80 (nearly all prey exposed)
• At periphyton density equal to K: the scaling factor is about 0.51, so effective access is about 0.41
• At very low periphyton density: effective access drops to 0.80 times 0.02 = 0.016 (almost all survivors are hiding)

This creates powerful stabilization. As Copepods graze periphyton down, the remaining periphyton become harder and harder to find. The grazing rate drops faster than the prey population, preventing complete wipeout.

Which Interactions Have Density-Dependent Refugia

Not all interactions use density-dependent refugia. It is enabled for prey types where the "last survivors hide in the best spots" logic makes biological sense:

• Copepods on periphyton: K = 5e-4 mol/L, min = 0.02
• Copepods on microzooplankton (ciliates): K = 2.5e-5 mol/L, min = 0.02
• Daphnia on planktonic algae: K = 5e-5 mol/L, min = 0.105
• Daphnia on periphyton: K = 5e-4 mol/L, min = 0.02
• Every grazer on bacteria: K = see below, min = 0.08

Interactions without density-dependent refugia (like detritus) use their static access values at all prey densities.

Bacteria: a floor that depends on how much surface the tank has

Bacteria get the same kind of floor, with one difference that matters. For every other prey type the half-saturation density is a single number per litre of water. For bacteria that is only half the answer, because most of the bacteria in an aquarium are not in the water.

Free-swimming bacteria in productive fresh water number a few million per millilitre. The film growing on glass, gravel, and above all on filter media carries ten million to a billion cells per square centimetre. In a bare vessel the water-column population is the whole population. In an ordinary tank the film outweighs it many times over, and adding a filter adds far more bacterial habitat than it adds water.

So the bacterial half-saturation is built from two pieces:

• a water-column term, set from the observed abundance of free-swimming bacteria, which is the whole of the answer for a tank with no colonisable surface;
• a surface term, set from the observed density of biofilm per square centimetre, multiplied by how much colonisable area the tank has for each litre it holds.

The surface term counts texture, not bare area. Smooth glass gives a cell almost nowhere to hide and contributes very little; porous foam and sintered ceramic contribute nearly their full area, because the shelter is in the crevices. A 35 L tank with walls and gravel alone lands about five times the bare-vessel floor. Add a sponge filter and it lands about twenty-four times it.

This is the same reason a rinsed filter sets a tank back further than a water change does, and the reason a tank with generous media rides out a heavy grazing population that would strip a bare one. Note that for bulk prey pools like bacteria and microzooplankton, the static per-surface access values are combined into a surface-area-weighted average, so the effective access depends on the scenario's surface composition even without density-dependent scaling.

How Refugia and Surface Type Interact

The surface-specific access value and the density-dependent refugia factor work together multiplicatively. Consider Copepods eating periphyton on gravel:

1. Static access on gravel: 0.25 (75% in physical refugia from gravel interstices)
2. Density-dependent factor (at moderate periphyton density): maybe 0.60
3. Effective access: 0.25 times 0.60 = 0.15

So only 15% of periphyton on gravel are actually reachable by copepods at moderate densities. On smooth glass at the same density:

1. Static access on glass: 0.95
2. Density-dependent factor: still 0.60 (same prey density)
3. Effective access: 0.95 times 0.60 = 0.57

This means glass provides almost four times the food access of gravel, which is why surface choice has such a large impact on grazer-prey dynamics in the model.

Practical Implications

• Adding rough surfaces (gravel, porous ceramics) to a simulated jar creates refugia that protect algae and prevent grazer-driven crashes.
• Biofilm surfaces are designed as intentional refuges with very low access values across all grazers and food types.
• Density-dependent refugia prevent total extinction of prey species even on smooth surfaces -- as prey gets rare, the survivors effectively "disappear" from the grazer's perspective.
• Different grazers see different refugia. Daphnia (filter feeder) has no access to surface-attached food directly, only to cells that detach. Copepods (scraper/hunter) have high access to exposed surfaces but cannot enter narrow spaces. This means the same gravel surface provides different refugia for different grazer-periphyton pairs.

Further Reading

• Feeding Mechanics -- the Holling Type II functional response that refugia modify
• Surfaces -- the physical surface properties that create refugia
• The Food Web -- per-species diet tables and access fractions
• Stability and Failure -- how refugia prevent grazer-driven ecosystem collapse