This is what your inputs predict over time. The KPI tiles at the top give you the headline numbers; the chart and table give you the full time-resolved view.

The numbers shown are bulk-water values — they describe how chlorine and DBPs evolve in a closed sample of water. In a real network, the FAC at a tap will be lower than the bulk prediction here because of pipe-wall demand. The MAV compliance flags assume the predicted bulk DBP levels are reached at the sampling point, which is conservative (real DBP at a tap is usually slightly higher than bulk, due to additional contact time).

Time-to-threshold is the hour at which bulk FAC drops to the network minimum set in Settings. Use the Wall Effect tab to estimate how much sooner this happens in your distribution system.

This tab breaks the total THM and HAA values down into the individual species at a chosen time point and compares them to the MAVs.

For THMs, bromide drives the shift from chloroform toward the brominated species (BDCM, DBCM, bromoform). Even small bromide concentrations (>0.05 mg/L) can produce significant BDCM, which has the most stringent MAV (60 µg/L). For HAAs, the di- and tri-chlorinated species dominate in low-bromide waters; brominated HAAs become significant at higher bromide.

Pick a contact time below — the bars and table refresh automatically.

The core model predicts bulk decay only. In a real distribution system, chlorine is also consumed at the pipe wall — by biofilm, corrosion products, tuberculation, and reactions with pipe materials. Total in-pipe decay is usually well-described by combining the bulk reaction with a first-order wall reaction:

C_pipe(t) = C_bulk(t) · exp(−k_wall · t)

The wall coefficient k_wall depends on pipe material, age, condition, and flow velocity. The reference table below gives literature ranges. You can either:

1. Estimate from a field observation — give the tool an observed FAC at a known residence time in your network and it will back-calculate the implied k_wall. This is the most useful method, because it matches the model to your real system.
2. Enter a wall coefficient directly from the reference table or your own data.

The "in-pipe" curve below shows what the bulk model would predict if the wall effect were applied to your scenario.

The defaults that ship with the tool will be roughly right for most NZ surface waters, but every catchment is different. This tab lets you:

• Learn how to run a chlorine decay test and a DBP formation potential (DBPFP) series for your water.
• Fit your data — paste time/FAC measurements and the tool will fit the capacity-limited decay parameters (fast and slow demand capacities and their rate constants) automatically.
• Adjust coefficients manually in the table below, with help-text explaining what each one does.

The expandable sections below walk through the test protocols and the math, so it's a fairly self-contained workflow. Once you've fitted your data, hit "Apply to current preset" and the rest of the tool will use your calibrated coefficients.

A chlorine decay test characterises how chlorine reacts with the constituents in your water. It can be done in any well-equipped operations lab.

What you need:

• A clean sample of your water, at the temperature and pH it experiences in service
• A stock sodium hypochlorite solution of known strength
• A DPD colorimeter or amperometric titrator for FAC measurement
• Headspace-free glass bottles in a temperature-controlled bath or incubator
• A clock and a notepad

Protocol:

1. Collect your sample. Filter it through 0.45 µm if you want to characterise dissolved-only reactions. Use it within a few hours.
2. Adjust pH to the value you want to model (target pH ±0.1). Bring the sample to the test temperature.
3. Dose with a chlorine concentration similar to what you use operationally. For DBP formation potential, dose at 5 × DOC so chlorine isn't limiting.
4. Mix briefly, then dispense into 10–12 amber bottles, filling each completely (no headspace). Cap and incubate at the test temperature.
5. Sample one bottle at each timepoint: 0.25, 0.5, 1, 2, 4, 8, 24, 48, 96, 168 h. Measure FAC immediately.
6. If you want DBPs as well, fix the chlorine residual at each timepoint with a quench reagent (e.g. sodium thiosulphate or sodium sulphite) and send to a lab for THM/HAA analysis.

Tips:

• The fast phase happens in the first 1–4 hours — don't skip the early timepoints.
• If chlorine drops below 0.2 mg/L before the end of the test, your dose was too low — repeat with a higher dose.
• Run a blank (sample with no chlorine) to check for background interferences.

A DBPFP test tells you how much THM and HAA your water will form per unit of chlorine consumed. It is the basis for the k_THM and k_HAA coefficients in the model.

Standard conditions (ICR / Standard Methods 5710):

• Temperature: 20 °C (or your site temperature if you want a site-specific fit)
• pH: buffered to 7.0 ± 0.2 (phosphate buffer)
• Chlorine dose: enough to give a residual of 3–5 mg/L at the test endpoint (typically 4–6 × DOC)
• Contact time: 7 days (168 h) for "ultimate" formation
• Headspace-free, dark, sealed

Outputs:

• FAC measured at start and endpoint (and ideally a kinetic curve in between)
• Individual THM and HAA species measured at the endpoint

Calculating yields:

• k_THM = total THMs (µg/L) ÷ chlorine consumed (mg/L)
• k_HAA = total HAAs (µg/L) ÷ chlorine consumed (mg/L)

For most NZ humic / fulvic waters these come out around 22–30 µg THM and 25–40 µg HAA per mg Cl₂ consumed. If you fit values well outside that range, double-check the test conditions.

This is the only place where data leaves the tool. Three options:

• PDF report — formatted document with the tabs you select, ready to attach to an email or store with a project record. Uses your browser's print-to-PDF.
• CSV download — the time-resolved data and key results, ready to open in a spreadsheet for further analysis.
• JSON save — a single file containing every input, every coefficient and all results. Re-load it later to pick up exactly where you left off.

The PDF and CSV always reflect the current model state; if you change an input after generating one, just regenerate.