Biofilter Recovery After a Pump Failure in a Tilapia RAS
A 60 m³ freshwater recirculating system holds 8 t of Nile tilapia at 27 °C and is fed 160 kg/day of 32% crude protein feed. The moving-bed biofilter contains 15 m³ of media at 500 m²/m³, with a design nitrification rate of 0.55 g TAN/m²/day at 27 °C. Five days ago a return pump failed for six hours and the filter media sat without flow. Today: TAN 3.2 mg/L and rising, nitrite-N 6.0 mg/L, chloride 25 mg/L, pH 6.9 (down from 7.4), alkalinity 60 mg/L as CaCO3, DO 5.5 mg/L, and mortality has reached 40 fish/day. Make-up water is 5% of system volume per day at 10 °C and 30 mg/L alkalinity. The operator wants to add molasses to 'grow the bacteria back faster' and to stop feeding until ammonia reads zero.
1. Build the nitrogen and alkalinity budget to show why this system cannot recover on its own.
Read the worked answer
TAN load. Using the standard figure of about 30 g TAN per kg of feed: 160 kg feed/day × 30 g = 4.8 kg TAN/day entering the system.
Filter capacity. 15 m³ × 500 m²/m³ = 7,500 m² of media. At the design rate of 0.55 g/m²/day, full capacity is 7,500 × 0.55 = 4.13 kg TAN/day - already below the load, so the filter was marginal at best before the failure. After six hours without flow, much of the biofilm is anoxic and a fraction has sloughed; a realistic recovery estimate is 30–50% of design, i.e. 1.2–2.1 kg TAN/day. The shortfall of roughly 2.7–3.6 kg/day across 60 m³ is 45–60 mg/L/day of accumulating TAN. Observed TAN is rising more slowly than that, which tells you heterotrophic assimilation and some surviving nitrifiers are still working - but not enough.
Alkalinity. Nitrification consumes about 7.14 g of alkalinity as CaCO₃ per gram of TAN oxidised. Oxidising even 4.0 kg TAN/day demands ~29 kg CaCO₃/day. Make-up water supplies only 3 m³ × 30 mg/L = 0.09 kg/day. Total system alkalinity is 60 mg/L × 60 m³ = 3.6 kg - about two hours of nitrification demand.
The feedback loop. Alkalinity is consumed → pH falls (already 7.4 → 6.9) → nitrification rate falls sharply below pH 7.0 → TAN accumulates further → and nitrite accumulates because Nitrobacter/Nitrospira recover more slowly than Nitrosomonas. The 6.0 mg/L nitrite-N with only 25 mg/L chloride is a Cl:NO₂-N ratio of about 4:1, far below the 10:1 minimum, so methaemoglobin toxicity is compounding the ammonia toxicity. The system is in a self-reinforcing decline.
2. Give the 48-hour intervention in priority order, with quantities, and state what you would not do.
Read the worked answer
Do first: protect the fish.
- Add sodium chloride to raise chloride to at least 100 mg/L against 6 mg/L nitrite-N. Need +75 mg/L Cl⁻ ≈ 125 mg/L NaCl ≈ 7.5 kg in 60 m³. This is the single fastest way to cut acute mortality.
- Dose sodium bicarbonate. Target 100–150 mg/L alkalinity and hold it there. To lift from 60 to 120 mg/L is 60 mg/L × 60 m³ = 3.6 kg immediately, then ~25–30 kg/day to match nitrification demand (about 0.18 kg bicarbonate per kg feed). Dose by metering pump, not by hand.
- Cut feed by 60–70% (to 50–60 kg/day), not to zero. Zero feed starves the biofilm of the very ammonia it needs to rebuild, and it does not remove ammonia already in the water. Feeding to appetite is what must stop; the fish can go to 50 kg/day for a week without loss of growth.
- Increase exchange to 15–20% per day, but temper the 10 °C make-up water - a 10 °C inflow at 20%/day will shock a 27 °C system. Run make-up through the heater or introduce it slowly.
- Reseed the biofilter with a commercial nitrifier culture or sludge from a healthy system, and keep DO in the filter above 4 mg/L.
What I would not do:
- Not add molasses now. A carbon source drives heterotrophic growth that competes for oxygen and surface area, and at pH 6.9 with a failing filter the dominant risk is oxygen and ammonia, not nitrate. Reconsider only after TAN and nitrite are stable and DO is above 6 mg/L.
- Not chase TAN to zero. That is a two-to-three-week goal, not a shutdown criterion.
- Not use zeolite as the main fix - its capacity is small relative to a 4.8 kg/day load in freshwater and it is exhausted within hours.
- Not raise pH with lime while TAN is high; use bicarbonate, which buffers without a sharp pH spike.
3. Set a two-week recovery plan with the specific numbers that trigger the next step.
Read the worked answer
Days 1–3 - stabilise. Hold chloride ≥100 mg/L, alkalinity 100–150 mg/L, pH 7.2–7.6, DO ≥6 mg/L in the tank. Feed 50 kg/day. Test TAN, nitrite, pH, alkalinity and DO twice daily. Accept a slow rise in TAN as long as it stays under 4 mg/L; the trigger to increase exchange to 30% is TAN >4 mg/L or nitrite-N >10 mg/L.
Days 4–7 - confirm the filter is turning. The decisive signal is a falling TAN with a still-high or rising nitrite: that means ammonia oxidisers have recovered and nitrite oxidisers are the new bottleneck. When TAN falls for three consecutive readings, raise feed by 10 kg/day and hold for two days before the next increment. Keep dosing bicarbonate at 20–30 kg/day, adjusted fortnightly from the alkalinity trend rather than a fixed rate.
Days 8–14 - rebuild to full load. Step feed 70 → 100 → 130 → 160 kg/day only if TAN stays <1 mg/L and nitrite-N <2 mg/L after each step. Target alkalinity 100–150 mg/L, chloride ≥100 mg/L, DO ≥6 mg/L, pH >7.2.
Hard triggers to back off: TAN >4 mg/L, nitrite-N >10 mg/L, pH <6.8, DO <5 mg/L, or mortality above 40 fish/day for two days - in any of these, drop feed by 30% and increase exchange.
Engineering follow-up within 30 days: install a standby return pump with automatic changeover and an alarm on filter flow, because the whole event traces to a single six-hour stoppage. Also add continuous alkalinity or pH-controlled bicarbonate dosing, since this system was already running a nitrogen load above its design capacity before the failure.