Wastewater treatment in WESS plants: compartments, chemistry, and diagnosis

A WESS (Water-Efficient Sanitation System) is a packaged, decentralised treatment plant: it takes the full sewage stream of a building or small settlement, treats it on site through a train of compartments, and returns the water for non-potable reuse — typically toilet flushing. Because the water recirculates rather than discharging to a sewer, a WESS behaves differently from a conventional works in three ways that matter for everything in this guide:

First, it is a closed loop. Anything the biology and physical polishing steps cannot remove — dissolved salts above all — accumulates cycle after cycle. A municipal works never sees this; a WESS lives with it, and "reboot" (partial or full replacement of the loop volume with fresh water) is a legitimate, scheduled operation rather than a failure.

Second, it is small. Hydraulic and organic shock loads that a large works would absorb — a weekend of heavy use, a drum of harsh cleaning chemicals, a power outage — can wipe out the microbial community of a plant with a few cubic metres of biological volume. Recovery then needs reseeding, not patience.

Third, it is operated lightly. There is no resident process engineer. The plant must therefore be diagnosable: a structured set of measurements, taken on a site visit with field instruments plus a small laboratory suite, must be enough to say what is going wrong, where in the train, and what to do about it. That structure — compartment taxonomy, measurement bands, diagnosis catalogue, mitigation library — is the WESS protocol, and it is what the rest of this guide teaches.

Every WESS variant is described against the same ten-compartment archetype. A given plant will not have all ten — a small septic-plus-wetland system may only have C1, C3 and C9 — but every real compartment maps onto one archetype, and every measurement band, flag and diagnosis in the protocol is anchored to that mapping. The train runs from raw influent to reuse product:

First principles: oxidation, reduction, and where the energy comes from

A redox reaction is an electron transfer. The species that loses electrons is oxidised; the species that gains them is reduced (the schoolroom mnemonic OIL RIG — Oxidation Is Loss, Reduction Is Gain — is all you need). There is nothing biological in the definition: iron rusting is iron oxidised by oxygen. What biology adds is harnessing. A microbe is an engine that couples one oxidation to one reduction — it strips electrons from a fuel (the electron donor) and delivers them to an electron acceptor, capturing the energy released in between as ATP. Everything the biological compartments do is a version of this one transaction.

In a WESS the dominant fuel is the sewage itself:

The electrons must land somewhere, and the acceptors are not equivalent: each pays a different energy price. Ranked from richest to poorest, they form the acceptor ladder that organises the whole train:

The ordering is thermodynamic, not managerial: no valve enforces it. While oxygen is present every other acceptor waits, because the organisms using oxygen out-compete the rest for the same electrons; when oxygen runs out the community shifts to nitrate, then sulphate, then CO₂. (In the gap below −100 mV, fermenters also work without any external acceptor at all, splitting organics into the volatile fatty acids that feed methanogenesis — the ABR chain of §3.) This self-sorting is why compartment zoning works: the designer engineers the acceptor supply — blow air into C4, exclude it from C5, keep C3 sealed — and the right community assembles itself in each tank.

The redox ladder: one number that places you in the train

The oxidation–reduction potential (ORP) is the field measurement of all of the above. A platinum electrode dipped in the compartment adopts a voltage set by the dominant redox couples in the liquid, read against an Ag/AgCl reference: strongly positive means the water is electron-poor and oxidising (acceptors abundant), strongly negative means electron-rich and reducing (donors queuing for scarce acceptors). It is a summary voltage, not a concentration — but because each regime on the acceptor ladder holds a characteristic window, the one number places the compartment on that ladder. The protocol's compartment windows are simply the ladder mapped onto the train — which is why a single probe dip can tell you a compartment is in the wrong regime before any laboratory result comes back. An anoxic zone reading +150 mV is not "slightly high"; it has become an aerobic zone, and its denitrifiers have stopped working.

Compartments in full: hardware, chemistry, readings and mitigations

Each WESS compartment is treated here as a complete unit: the hardware as built, the chemistry it runs, the values the field kit and the laboratory should return when it is healthy, and the failure modes with their mitigations. The bands repeat the protocol’s operational triggers, so a reading taken at any of these units can be placed immediately. Thresholds are decision aids under calibration rather than validated regulatory limits.

C1 · Conservancy / septic tank

A sealed reinforced-concrete box receives the raw stream and holds it in near-total oxygen absence (ORP −300 to −100 mV). Two processes begin at once. Coarse and settleable solids drop into a sludge blanket, taking roughly half the suspended load out of the flow within a day of quiet residence. In the liquid above, hydrolysis breaks particulate organics into soluble molecules and ammonification converts organic nitrogen to ammonium, which is why NH₄-N rises through this compartment while COD begins to fall.

Urine carries most of the nitrogen, and urea hydrolyses within hours:

Failure here is almost always arithmetic rather than biological: the blanket grows until the tank stops settling, and raw load passes through to compartments never sized for it. The worked case of §8 shows the signature (septic COD of 5 691 mg/L against the 3 000 action level) and the catalogue names it SEPTIC_OVERDUE. The mitigation is M01, a desludge with a standing sludge-judge cadence so the next one is scheduled rather than discovered. Composite sampling (M11) belongs here whenever a reading will carry weight, since a grab sample near the inlet of a stratified tank can say almost anything. Persistent grit and rag loads point back to what the community puts down the drain, which is M13 territory.

C3 · Anaerobic reactor (ABR / UASB)

This compartment does the heavy COD removal at zero aeration cost. Influent passes up (UASB) or over-and-under (ABR baffles) through a retained anaerobic biomass, and the three-stage chain of §3 runs to completion: hydrolysis, fermentation to volatile fatty acids, then methanogenesis at ORP −400 to −200 mV. Most of the methane leaves by the acetate route, the rest from hydrogen:

Two failure modes dominate. ABR_WASHOUT (TSS above 500 leaving the reactor) follows hydraulic surging or a collapsed blanket; the response is M06, an integrity check with chemically-enhanced backwash where membranes follow, together with M01 to restore the solids inventory. Souring is the chemical failure: an organic shock makes acid faster than the methanogens clear it, pH slides below 6.5, and the methanogens quit, which makes more acid. That loop does not self-correct. Breaking it needs load relief, alkalinity, and in the worst case a reboot with live-culture reseed (site-walk step 10). The causes worth ruling out on the same visit are toxic loads and chlorinated water routed back through the biology, both M13 and plumbing questions rather than dosing ones.

C4 · Aerobic reactor

Air changes everything. At DO of 2–4 mg/L and ORP of +50 to +200 mV, heterotrophs burn the residual soluble COD and the nitrifiers run the first leg of the nitrogen relay. Both reactions consume something besides oxygen, and both costs appear in the readings:

AERATION_FAIL is the acute failure: a blower or diffuser outage stops nitrification and aerobic COD removal within hours and the whole plant slides down the redox ladder. The mitigation is M02, an aeration audit and rebalance of setpoint and blower duty. The chronic failures are quieter. Complete ammonia removal with acidic effluent is unbuffered nitrification; the acid of the first reaction has exhausted the water’s alkalinity, and below pH 6.5 the nitrifiers throttle themselves, trading an acid problem for an ammonia one. Over-aeration wastes power, shears floc, and poisons C5 through the recycle. Filamentous bulking grows out of low DO, low F:M or surfactant loading and is read at the settler but caused here; the settling root-cause tree of site-walk step 4 assigns it, and M10, M13 and M14 carry the responses that are not aeration.

C5 · Anoxic reactor

Between −100 and 0 mV, with oxygen excluded, facultative heterotrophs switch to nitrate respiration and run the relay’s second leg. The unit needs three things in the same water at the same time: nitrate (delivered by the recirculation), soluble carbon (delivered by the influent), and the absence of dissolved oxygen (protected by the mixer doing what a diffuser must not).

DENIT_FAIL announces itself as high effluent nitrate under low effluent ammonia: the first leg worked and the second never ran. Three causes cover the field cases. Under-recirculation starves the zone of nitrate, and M03 raises the recycle to 3–4 times forward flow. Carbon starvation follows over-aerated or over-settled influent, and M04 doses an external carbon source. Oxygen poisoning rides the recycle from an over-aerated C4, which makes the cure M02 upstream rather than anything in this tank. The worked case is the standing reminder that plumbing outranks chemistry here: EP had carbon in abundance (septic COD 5 691) and still put 70.9 mg/L of nitrate in its effluent, because the reactants never met.

C6 · Secondary settling tank

No reaction is intended here; the unit is physics applied to a biological product. Bacteria that grew at the right sludge age flocculate into dense aggregates bridged by extracellular polymer, and an hour of quiet lets them fall while clarified water leaves over the weir. The settled biomass returns to C4 to work again, which makes this tank half of the activated-sludge loop rather than a mere polisher.

One unintended reaction matters. Sludge that sits too long in the hopper goes anoxic, denitrifies whatever nitrate it carried in, and the nitrogen bubbles float whole rafts of solids to the surface. Rising sludge in an otherwise healthy plant is that mechanism, and the fix is faster sludge return rather than anything chemical.

The diagnostic pair SVI/DSVI separates the two failure families. SVI of 210 with DSVI of 95 settles fine once diluted, so the tank simply holds too much sludge: an inventory finding, M01 territory, fixed by wasting cadence. SVI and DSVI both high is bulking proper, and the step-4 root-cause tree assigns it to low DO, low F:M or surfactants, which routes the mitigation to M02, M13 or a reboot-and-reseed after community failure. Solids that escape this tank do double damage, first as effluent TSS, then as the particulate shielding that lets E. coli survive a correct chlorine residual at C9.

C7 · Membrane filtration (UF / MBR)

The membrane is the train’s one absolute barrier. An ultrafiltration wall carries pores of roughly 0.01–0.1 µm, and rejection at that scale is geometric: water and dissolved species pass, solids and bacteria cannot, the smallest viruses are marginal. Nothing biological or chemical is asked of the unit. In a WESS it either polishes the settler’s effluent or, in the MBR form drawn above, replaces the settler entirely, in which case the SVI questions of C6 stop mattering and membrane care takes their place.

The physics that runs the unit fits in one law, a resistance in series:

The protocol leaves C7 unbanded because there is no window to hold; there is only integrity to prove and fouling to pace. The readings that matter are trends:

Failures split into the gradual and the abrupt. Fouling is the gradual one, and it comes in families: cake that air scour and backwash remove daily, and pore blocking and biofilm that they cannot, which is what the chemically-enhanced backwash of M06 exists for — hypochlorite or acid drawn backwards through the fibres on a maintenance cadence. The abrupt failure is a broken fibre, and it is why permeate turbidity earns continuous attention: an intact membrane makes particulate shielding at C9 impossible, so elevated E. coli with an adequate chlorine residual downstream of an MBR points at a fibre rather than at dosing. Two cautions close the unit. CEB chemicals leave by the drain, never through the biology — the same routing discipline as chlorinated reuse water. And a membrane run hard to postpone cleaning saves nothing; R_c compacts under pressure, and cake that backwash would have lifted becomes fouling only chemistry can reach.

C9 · Disinfection / contact tank

Chlorine dosed into water speciates within seconds, and the active pair it forms does the killing:

The protocol reads this unit jointly or not at all: a count and a residual, interpreted together (site-walk steps 13–15). Elevated count with adequate residual is shielding, and the cure lives at C6 or C7 rather than in the dosing room. Elevated count with no residual is DISINFECT_FAIL proper, met by M08: repair the dosing schedule and put the operator log behind it. Over-dosing has its own red line at 5 mg/L for by-products and user acceptance. One design trap is specific to recirculating systems and worth checking by tracing pipes rather than by sampling: chlorinated water routed back through the biological compartments sterilises the plant’s own workforce.

C10 · Treated effluent storage / reuse tank

Chemistry here is subtraction: nothing upstream removes dissolved salts, so everything the users add — urine, detergents, water hardness — accumulates around the loop while evaporation concentrates it. TDS therefore rises in every healthy WESS, and the operative measurements are the level and the slope. A second, slower process runs in storage: the chlorine residual decays with time and temperature, so water that passed at the dosing point can fail at the furthest tap, and any regrowth shows up there first.

TDS_LOOP is a calendar problem wearing a chemistry costume. The remedy is M12, a reboot SOP that schedules partial (30–50%) or full drain-and-refill against the trend instead of waiting for the biology to pay the osmotic tax above 1 500 mg/L, with fresh water secured before the drain starts. M09 puts an inline EC sensor on the loop so the trend is continuous rather than per-visit, and M11 keeps the samples that matter defensible. A plant flying green on every other band while TDS climbs is healthy; the correct response is a booking, not a mitigation hunt.

Carbon: from COD to methane

Organic load is tracked as COD (chemical oxygen demand — everything oxidisable) and BOD (the biologically available fraction). Raw domestic sewage arrives with COD in the hundreds to thousands of mg/L; a healthy train removes well over 95% of it. The anaerobic stages (C1, C3) do the heavy lifting cheaply: hydrolysis breaks particulate organics to soluble molecules, fermentation converts these to volatile fatty acids, and methanogens finish the job:

What anaerobic biology leaves behind — residual soluble COD — is finished aerobically in C4, and any last hard-to-degrade colour and organics are adsorbed in C8. This division explains two common flags: high TSS leaving the ABR (C3) means the sludge blanket is being washed out and the plant is exporting its own biomass downstream, and high effluent COD (C9) with everything else healthy usually means the polishing adsorbent (GAC/zeolite) is saturated and needs regeneration, not that the biology has failed.

Nitrogen: the two-step relay that fails in two directions

Urine makes domestic wastewater nitrogen-rich. In the septic and anaerobic stages, organic nitrogen is ammonified to ammonium — which is why NH₄-N rises through the early train even as COD falls. The aerobic stage then nitrifies:

The relay therefore fails in two distinguishable directions, and the effluent tells you which. High effluent ammonia means the first leg failed — nitrification incomplete, usually an aeration or pH problem in C4. High effluent nitrate with low ammonia means the first leg worked and the second failed — denitrification lagging, usually carbon or recirculation. The two diagnoses have entirely different mitigations, which is why the protocol keeps NH₄-N and NO₃-N as separate banded parameters at C9. Nitrite (NO₂-N) sitting between the two steps is normally near zero; measurable nitrite means the relay is stalling mid-handoff — a sensitive early-warning signal, and more toxic than either neighbour.

Phosphorus: the element biology barely touches

Ordinary heterotrophic growth takes up only ~1–2% P by mass, so a conventional biological train removes little phosphorus. Persistent high effluent TP is thus a design finding, not an operational lapse: the mitigation is chemical dosing (alum or ferric, jar-tested first) at the sedimentation stage, precipitating phosphate as a metal salt into the sludge.

Salinity: the closed-loop tax

Every flush adds salts — urine, detergents, water hardness — and evaporation concentrates them, while nothing in the train removes them. TDS therefore ratchets upward in any reuse loop; the operative questions are only how fast (the protocol trends %/week between visits) and when to intervene. Above ≈ 1500 mg/L, osmotic stress starts to suppress the biology itself and users notice the water. The remedy is the reboot: a partial (30–50%) or full drain-and-refill, scheduled deliberately rather than suffered as a mystery decline.

Disinfection: chemistry with a shield problem

Chlorine dosing (typically tablet-fed) must hold a free residual of about 0.2–0.5 mg/L at the point of reuse: enough to keep killing, little enough to avoid taste, odour and disinfection by-products. Two subtleties dominate WESS practice. First, chlorine is consumed by ammonia and organics before it becomes free residual — so an upstream biological failure silently eats the disinfectant dose. Second, particulate shielding: E. coli embedded in suspended solids survives contact even when the residual reads adequate. That is why the protocol interprets a lab E. coli count jointly with the field chlorine reading: elevated count + adequate Cl₂ → fix upstream solids; elevated count + no residual → fix the dosing. One more design trap is specific to recirculating systems: if chlorinated water is routed back through the biological zones, the disinfectant sterilises the plant's own workforce.

Settling: the physics that biology can break

Separation in C6 depends on bacteria flocculating into dense, settleable aggregates. The sludge volume index ($\mathrm{SVI} = \mathrm{SV}_{30}/\mathrm{MLSS}$, the 30-minute settled volume per unit mixed-liquor solids) measures this: 80–150 mL/g settles well; above 150 the sludge is bulking — almost always filamentous overgrowth, encouraged by low DO, low F:M ratio, or surfactant loading; far below 80, pin-point floc that leaves a turbid supernatant. Persistent surface foam is read with it: light tan and thin is healthy; dense, dark and persistent points to filaments or heavy detergent use in the community.

The protocol's sixteen parameters split naturally into what a field kit reads in minutes and what needs a laboratory (with a SANAS chain of custody for anything compliance-facing). Each parameter earns its place by answering a specific diagnostic question:

A reading only means something at a compartment: DO of 3 mg/L is healthy in C4 and a failure in C5. The protocol therefore bands every (parameter, compartment) pair: GREEN within the healthy window, RED beyond the action threshold, AMBER between them. Each flag carries a plain-language label naming the mechanism, so the flag set reads as a symptom list.

Flags are symptoms; a diagnosis is a named failure mode defined as a logical combination of them (flag keys read PARAM_side[_band]@compartment). The catalogue below is the protocol's §4.5 — nine mechanisms that between them cover the great majority of WESS field failures. Note how each one leans on the chemistry of §3: the logic is only credible because the mechanism is.

Mitigations are catalogued with a type, a default owner (who acts: the site operator, the technology provider, the WESP programme, or the user community), and default impact / effort scores (1–5) used to rank recommendations — high-impact, low-effort actions surface first in every generated report.

Notice the shape of the library: the two highest-leverage actions (M01, M08) are cheap operator disciplines, not capital works. The pattern repeats across the field programme — most red flags trace to maintenance cadence and dosing habits before they trace to design. Training and awareness items score low on immediate impact but are the only mitigations that prevent recurrence.

The protocol's reference dataset samples two sister plants (EP and WP, 7de Laan, March 2026 laboratory campaign) on the same day with the same methods — a natural controlled comparison. Three points per plant: the septic compartment (C1), the pumping chamber after the anaerobic stage (C3), and final effluent (C9).

Read EP's numbers as a chain of §3's chemistry. Ammonium is near zero in the septic tank but 72.8 mg/L in the pumping chamber — ammonification proceeding through the anaerobic train exactly as expected. Effluent ammonia is 0.29: nitrification is working perfectly. But that nitrogen had to go somewhere, and there it is — 70.9 mg/L of effluent nitrate. The relay's first leg ran; the second (DENIT_FAIL) never started: with septic COD at 5 691 the plant has carbon in abundance, so the fault is plumbing — recirculation or anoxic-zone integrity — not chemistry supply. Around that central finding cluster the rest: septic desludge overdue (SEPTIC_OVERDUE), phosphorus never addressed by design (TP_HIGH), the reuse loop past its salinity threshold (TDS_LOOP), and residual COD at the polish (COD_RESIDUAL). Meanwhile disinfection is flawless — a reminder that a plant can hand you sparkling, pathogen-free water that is quietly failing four other ways.

WP, fed a lighter load, is green nearly everywhere on the same day with the same lab: the comparison isolates operation and load, not technology, as EP's problem. Its prioritised mitigation list writes itself from §7: M01 (desludge — impact 5, effort 1) first, then M03/M04 for the nitrate, M12 to book the reboot, M05 for phosphorus.

The de-risking site visit is itself a decision tree — seven phases, seventeen core steps, with remedial branches that fire on the readings just taken. Each gate below is a transcribed decision table from the protocol; amber-style ranges ("monitor if…") sit between the green pass and the red branch.