THIRSTY PLANET
The Paraná river floodplain seen from orbit

Plain Water: Turbidity

The first thing anyone judges water by is whether they can see through it, and the water industry turned that glance into a number a century ago. What turbidity is and how a nephelometer measures it by the light the water scatters, why a cloudy water is a dangerous water even when the cloud itself is harmless, how a turbidity reading was the warning that a city of 400,000 sick people ignored in 1993, why the filters at a water works are judged on a tenth of a unit, and what a river in flood carries past the intake.

· By , founder of Aguato · 8 min read

From orbit a great river in flood is the colour of milky coffee, a brown ribbon braided across a green plain, and where it meets the sea the brown spreads for tens of kilometres before it fades. The colour is mud: clay and silt lifted from the banks and the fields upstream, held in the water as particles too small to sink while the water moves. A bucket of it, left to stand, will clear from the top down over a day and leave a layer of fine sediment at the bottom. The amount of light that the water blocks on the way down, before it has cleared, is the first thing a person judges water by and the first thing the water industry measures, and the industry's name for it is turbidity.

This article is about what turbidity is and how it is measured, why a cloudy water is a dangerous water even when the cloud is only clay, how a turbidity reading was the warning a city ignored in 1993, and why a water works is judged on a tenth of a unit.

A cloud of particles

Turbidity is cloudiness. It is the loss of clarity caused by particles suspended in the water that are too small to settle quickly and too large to dissolve: clay and silt, fine organic debris, algae and other plankton, the flocs of iron and aluminium that a treatment plant makes on purpose, and bacteria. It is a property of the water as it looks rather than a measure of any one thing in it, and two waters with the same turbidity can carry very different loads: a litre of water made cloudy by fine white clay may hold a tenth of the mass of one made equally cloudy by coarse dark silt, because turbidity depends on how the particles scatter light, which depends on their size, shape and colour as much as their number. The measure of mass is a different test, total suspended solids, done by filtering a litre and weighing what is caught. Turbidity is the quick one, done in seconds by a sensor, and the one a plant runs continuously.

Turbidity, in NTUWhat it looks like
Below 0.3Clear in any glass; the limit for a filtered drinking water in the United States
1Clear in a glass; faintly cloudy in a white bucket; the WHO's target before disinfection
5Visibly hazy to most people; the WHO's limit for acceptability
20 to 50Cloudy; a lake after a storm; a slow river
100 to 1,000Opaque; a river in flood; mud

The nephelometer

The measurement is of scattered light. A beam is shone into a sample and a detector set at ninety degrees to the beam counts the light that the particles turn sideways; clear water scatters almost nothing and the reading is near zero, and the more particles, the more light reaches the detector. The instrument is called a nephelometer, from the Greek for cloud, and its unit is the nephelometric turbidity unit, calibrated against a standard suspension of a polymer called formazin that is made to a recipe so that every instrument in the world reads the same cloud the same way. Older methods held a candle under a tube of water and poured until the flame could no longer be seen, which gave the Jackson candle unit, and the oceanographers of the nineteenth century lowered a white disk on a rope until it vanished, which is the Secchi disk and is still used on every lake survey on Earth. The nephelometer replaced them because it works at the low end, where a drinking water lives: a Secchi disk cannot tell 0.3 from 1, and a candle cannot either, and the difference between those two readings is the difference between a filter that is working and one that is not.

A nephelometer. A beam into the water, a detector at ninety degrees, and the scattered light counted.
A nephelometer. A beam into the water, a detector at ninety degrees, and the scattered light counted.

Why the cloud matters

The particles that make a water cloudy are mostly harmless in themselves; a mouthful of silt is not a disease. The reason turbidity is the number the whole industry watches is what the particles do for the things that are. Bacteria, viruses and parasites in water attach to particles and hide inside the flocs, and a particle shields them from the chlorine, the ultraviolet lamp or the ozone that the disinfection article on this site describes: a dose that kills the bacteria in a clear water fails in a cloudy one because the chlorine cannot reach what is inside the cloud and the lamp's light is scattered before it arrives. Turbidity also feeds the chlorine, because organic particles react with it and use it up, so that a cloudy water needs a larger dose to leave the same residual and makes more of the byproducts the article on that subject explains. And for one organism the cloud is the whole story. Cryptosporidium is a parasite of the gut, passed in the faeces of cattle and people as a tough spore a few thousandths of a millimetre across, and chlorine at the doses used in drinking water does not kill it; the only barrier a conventional plant has against it is the filter, and a filter removes it the way it removes any other particle of that size. If the particles are getting through, so is the parasite. The turbidity meter on the filter outlet is the parasite alarm.

Milwaukee

In March and April 1993 the water works on the south side of Milwaukee, Wisconsin, drew from Lake Michigan at a time when the spring melt and heavy rain had carried the runoff of the farmland and the city's rivers into the lake near the intake, and the turbidity of the raw water rose. The plant's coagulation, which the article on that subject describes, did not keep up, and the turbidity of the finished water, which had run at a tenth of a unit, rose to above one and stayed there for days. The readings were within the rules of the time, which allowed up to 1 NTU on average and 5 at the limit, and nobody acted. In the two weeks that followed, about 400,000 people in the city, a quarter of its population, fell ill with watery diarrhoea, the pharmacies ran out of antidiarrhoeal medicine, and the cause, when it was found in the stools of the sick and then in the ice that had been made from the water, was Cryptosporidium. About a hundred people died, most of them with weakened immune systems. It remains the largest waterborne outbreak in the recorded history of the United States, and it changed the rule: the turbidity of filtered water in the United States must now be at or below 0.3 NTU in at least 95 percent of measurements each month and never above 1, and every filter is monitored separately, every fifteen minutes, so that one filter breaking through cannot hide in the average of the others.

The Milwaukee outbreak, 1993
CauseCryptosporidium through a conventional plant's filters
Warning signFinished water turbidity rose from about 0.1 to above 1 NTU for days
The rule then1 NTU average, 5 maximum: the readings were legal
IllAbout 400,000, a quarter of the city
DeathsAbout 100, mostly the immunocompromised
The rule since0.3 NTU in 95 percent of samples, each filter monitored separately

At the works

A water works is organised around getting the number down. Raw water from a river or a reservoir arrives at anything from a few units to hundreds in flood; coagulation and flocculation bind the particles into flocs, settling removes most of them and brings the water to a unit or two, and the filters, beds of sand and anthracite a metre deep, take out the rest. A filter's turbidity meter tells its whole story: a clean filter after backwashing reads a few tenths and then, as the bed ripens, falls to a tenth or less; it holds there for a day or two as the bed fills with floc; and when the bed is full the number begins to climb, slowly and then fast, which is breakthrough, and the operator backwashes before it reaches the limit. In a flood, when the raw water is at hundreds of units, the plant slows down, doses more coagulant, backwashes more often, and in the worst cases shuts the intake and runs on the reservoir until the river clears, which is why the boil water notices that follow a storm in a small town are usually about turbidity rather than about any germ that was found.

Secchi disks. The white disk is lowered until it vanishes, which has been the lake surveyor's turbidity for a century and a half.
Secchi disks. The white disk is lowered until it vanishes, which has been the lake surveyor's turbidity for a century and a half.

The number at the tap, for a well run plant on a river source, is a tenth of a unit, which is clearer than most bottled water, and it is held there because of what the clarity proves rather than for the look of the glass: that the filter is intact and that nothing the size of a parasite is getting through.

Beyond the works

Turbidity is the river's number too. The sediment a river carries is its load, the eroded soil of its basin on the way to the sea, and the engineers measure it at the intakes and the dams: a dam fills with the silt its river carries, which is why reservoirs have a design life, and the Yellow River, the most turbid large river on Earth, has raised its own bed above the plain with the silt it drops. The clearing of a forest or the ploughing of a slope upstream shows at the intake as turbidity, and the plants downstream of a logged catchment or a construction site pay for it in coagulant. In the sea it is the brown plume off a river mouth that smothers a reef, and in a lake it is the algae of the phosphorus article, which make a different kind of cloud, green rather than brown, that the Secchi disk has tracked for a century as the lakes of the world have been fed.

What it teaches

Turbidity is the plain water article about the measurement that stands in for all the others. It is the simplest number in the industry, light scattered by a cloud, and it is read continuously at every filter in every works because it is the only number that can say, in real time, whether the barrier against the parasites is holding. Milwaukee's filters were legal and 400,000 people were ill; the rule that followed moved the line from a unit to a third of one, which is the difference between a cloud a person can see and one only an instrument can. A river in flood is hundreds. The water that leaves the works is a tenth, and the tenth is the point.

A tenth of a unit at the tap, a third of a unit at the limit, and 400,000 people in a city whose readings were within the rules.

Sources

  1. World Health Organization (2017). Guidelines for Drinking Water Quality, 4th edition with addendum; and the 2017 technical brief Water Quality and Health, Turbidity. Median turbidity below 1 NTU, ideally below 0.5 NTU, before disinfection.
  2. US Environmental Protection Agency, Interim Enhanced Surface Water Treatment Rule (1998) and Long Term 1 rule: combined filter effluent at or below 0.3 NTU in at least 95 percent of monthly measurements, never above 1 NTU.
  3. Mac Kenzie, W.R. et al. (1994). A massive outbreak in Milwaukee of Cryptosporidium infection transmitted through the public water supply. New England Journal of Medicine 331. About 403,000 cases.
  4. APHA, AWWA and WEF, Standard Methods for the Examination of Water and Wastewater, method 2130 B, nephelometric turbidity; ISO 7027.
  5. Secchi, A. (1865). Relazione delle esperienze fatte a bordo della pontificia pirocorvetta l'Immacolata Concezione. The first Secchi disk measurements.
  6. Photographs: opener: Paraná River Floodplain, Northern Argentina by ISS Expedition 27 crew (Public domain) via Wikimedia Commons; inline: This instrument, called a nephelometer, measures light scattering in real time by Mary O'Neill (Public domain) via Wikimedia Commons; inline: Kinwamakwad Secchi Disks by Jason Kurtzweil (CC BY-SA) via Wikimedia Commons.