Environmental Engineering

The big picture

Sometime in the last twenty-four hours you drank water that had been through a treatment plant, breathed air that was measured against a federal standard, and threw something away that went to an engineered facility with a liner under it. You almost certainly did not think about any of the three. That invisibility is the whole point. Environmental engineering is the discipline that stands between people and the consequences of their own waste, and it succeeds by being unnoticed.

Here is the field in one sentence: environmental engineering applies chemistry, biology, and fluid mechanics to keep contaminants out of the people, water bodies, air sheds, and soils that they would otherwise reach. That covers drinking water treatment, wastewater treatment, air pollution control, solid and hazardous waste management, and the cleanup of ground that somebody already ruined. It is a design discipline, not an observational one. An environmental scientist measures the nitrate in a stream and explains where it came from. An environmental engineer sizes the basin, picks the process, specifies the pump, and signs the drawing that will remove it.

Here is the plan for this lesson. First we look at where the discipline came from, because its origin as sanitary engineering explains its character. Then we lay out what the work actually is, area by area. Then we install the single habit of mind that runs under every lesson in this course, which is the mass balance. Then we draw the boundaries against the neighboring disciplines. Last, we meet the city whose numbers will follow you for the next fifteen lessons.

Where the discipline came from

Environmental engineering has an unusually clean origin story, and it starts with cholera. In 1854 a London physician named John Snow mapped cholera deaths around Broad Street in Soho and traced them to a single public water pump, which he had removed. Snow was not an engineer, and he could not prove the mechanism, since the germ theory of disease was not yet established. But he had shown that a water supply could kill a neighborhood, and that was an engineering problem.

What followed was the largest public health project of the nineteenth century. Edwin Chadwick's 1842 sanitary report argued that filth caused disease and that cities should be plumbed. After the Great Stink of 1858, when the Thames became unbearable enough to disrupt Parliament, Joseph Bazalgette built London an intercepting sewer system of roughly 132 kilometers of main sewers, sized on a per capita estimate and then, famously, doubled on the grounds that this was only going to be done once. The doubling is a good introduction to the profession's temperament.

Then came treatment. Slow sand filtration spread through European and American cities in the late 1800s. In 1908 Jersey City, New Jersey, became the first American city to continuously chlorinate its public water supply, following a court case over the quality of its Boonton Reservoir water. Within roughly two decades, typhoid fever, which had killed on the order of 30 deaths per 100,000 people per year in American cities, effectively vanished as a cause of death. Studies of twentieth century mortality repeatedly find that clean water and sewage treatment account for a large share of the gains in urban life expectancy, larger than any single medical intervention of the same period.

The field called itself sanitary engineering for most of a century. The rename to environmental engineering happened around 1970, and it was not cosmetic. The old field was about pathogens in water. The new field added air pollution, industrial chemicals, radiation, noise, hazardous waste, and eventually greenhouse gases, because the National Environmental Policy Act of 1970, the Clean Air Act Amendments of 1970, the Clean Water Act of 1972, and the founding of the U.S. Environmental Protection Agency in December 1970 created a legal demand for engineers who could design to all of it.

Key idea: The field began as sanitary engineering aimed at waterborne pathogens, and it earned its reputation by ending typhoid and cholera in industrial cities; the 1970 rename to environmental engineering marks the moment the law expanded the job to air, waste, and industrial chemicals.

What the work actually is

Practice divides by medium, which is the word the field uses for water, air, and land. Almost every job title sits in one of these areas.

Area	Central question	Typical deliverable
Drinking water	Is it safe to drink at every tap, all the time?	Process train design, basin and filter sizing, disinfection CT calculations, distribution models
Wastewater	Can this be discharged without harming the receiving water?	Aeration basin and clarifier sizing, permit compliance calculations, biosolids plans
Stormwater	Where does runoff go and what is it carrying?	Detention and green infrastructure design, permit reports, outfall monitoring
Air quality	What does this stack emit and where does it land?	Emission inventories, dispersion modeling, control equipment selection, permit applications
Solid waste	Where does the material go for the next thirty years?	Landfill cell and liner design, leachate and gas systems, diversion program analysis
Hazardous waste and remediation	What is in this ground and how do we stop it moving?	Site investigation, risk assessment, remedy selection, treatment system design
Industrial and compliance	How does this factory operate legally and cheaply?	Pretreatment design, pollution prevention audits, permits, reporting

Two features of this list are worth naming immediately. First, most of the work is process design: you are choosing and sizing a sequence of unit operations that each remove some fraction of some contaminant. That is an engineering activity with numbers in it, and this course will make you do those numbers. Second, nearly every deliverable in the right column exists because a statute or a permit demands it. Environmental engineering is the most regulation-driven engineering discipline there is, which is why the next lesson is about law before any lesson is about chemistry.

There is also a shape to the career that is worth knowing. Roughly half of environmental engineers work in consulting firms, which sell design and investigation services to municipalities and industry. A large share work for government, at EPA, at state environmental agencies, at the Army Corps of Engineers, and inside utilities. The rest work inside industry, managing compliance for a plant or a corporation. The U.S. Bureau of Labor Statistics counts on the order of 44,000 environmental engineers in the United States with a median annual wage around 100,000 dollars in the early 2020s, and the field is small enough that the professional community in any given state is a few hundred people who mostly know each other.

The mass balance habit

Here is the single mental tool that makes environmental engineering coherent. Draw a box around anything: a reactor, a lake, a city, a landfill cell, a human body. Then write the statement that all matter must obey.

Accumulation equals input minus output plus generation minus consumption.

That is it. Everything in this course is an application of that sentence. If nothing is being created or destroyed, and the system is at steady state so nothing accumulates, it collapses to the most useful statement in the field: what goes in must come out somewhere. Not sometimes. Always.

Worked example. A factory discharges 0.20 cubic meters per second of effluent containing 45 milligrams per liter of a conservative pollutant, meaning one that does not degrade, into a river flowing at 6.0 cubic meters per second and already carrying 3.0 milligrams per liter of it. What is the concentration downstream once mixing is complete? Take the box around the mixing zone. Mass in per second equals 0.20 times 45 plus 6.0 times 3.0, working in grams per second because a cubic meter per second times a milligram per liter gives grams per second: 9.0 plus 18.0 equals 27.0 grams per second. Volume out is 6.20 cubic meters per second. So the mixed concentration is 27.0 divided by 6.20, or 4.35 milligrams per liter.

Notice what that calculation just told you. The river's concentration rose by 1.35 milligrams per liter, and the factory's own 45 milligrams per liter effectively disappeared into a number that looks harmless. This is the logic behind the old maxim that the solution to pollution is dilution, and it is worth being honest that dilution genuinely works for some problems. A modest organic load into a large river really is degraded and diluted to nothing.

Dilution fails, however, in three specific ways, and each failure mode created a body of law you will meet in the next lesson. It fails when the contaminant persists, so the mass never leaves the system and simply spreads. It fails when the contaminant bioaccumulates, so organisms concentrate it back up from a dilute background, which is what mercury and PCBs and the per- and polyfluoroalkyl substances do. And it fails when there are many dischargers, because the assimilative capacity of the river is a fixed budget and the tenth factory finds it spent. Environmental regulation is largely the history of learning those three lessons in the field.

Key idea: Every process in this course is a mass balance, accumulation equals input minus output plus generation minus consumption, and at steady state that means whatever enters a treatment plant leaves it in the water, in the air, or in the solids, never nowhere.

A corollary you will keep meeting

The mass balance has an uncomfortable consequence that beginners routinely miss. Treatment does not destroy contaminants. It moves them and concentrates them.

A clarifier takes suspended solids out of water and produces sludge. A baghouse takes particles out of a gas stream and produces a hopper full of dust that may be a hazardous waste. A flue gas desulfurization scrubber takes sulfur dioxide out of the stack and produces gypsum slurry. Activated carbon takes per- and polyfluoroalkyl substances out of drinking water and produces spent carbon that is now a concentrated PFAS waste needing disposal. In every case the engineer has converted a dispersed problem into a concentrated one, which is genuinely progress, because concentrated problems can be contained and dispersed ones cannot. But the mass did not vanish, and the design is not finished until you have said where it went.

The exceptions, where contaminants really are destroyed, are the processes that break chemical bonds: biological oxidation in an aeration basin turns organic carbon into carbon dioxide and cell mass, thermal oxidation in an incinerator does the same faster, advanced oxidation processes generate hydroxyl radicals that shred organic molecules, and reductive dechlorination in an aquifer strips chlorine off solvents. Notice that all of these destroy organic compounds. Nothing destroys a metal. Lead, mercury, arsenic, and chromium can only be moved, immobilized, or changed in oxidation state.

Drawing the boundaries

Students confuse environmental engineering with three neighbors, so let us separate them cleanly. Environmental science, which at this school is Climate & Environmental Science (ENVS 210), studies how the earth system works: the carbon cycle, ecosystems, climate dynamics, the fate and transport of chemicals in nature. It is largely diagnostic. Environmental engineering is prescriptive and ends in a specification. You need the science, and this course borrows from it constantly, but the deliverable is different.

Civil engineering, which at this school is Civil Engineering & Infrastructure (ENGR 270), owns the built environment: structures, geotechnics, transportation, and water resources. Environmental engineering grew out of it and most environmental engineers still sit in civil engineering departments and take the civil licensure path. Civil engineering asks how big the pipe must be to carry the flow; environmental engineering asks what is in the flow and what has to come out. Where that course surveys water and wastewater infrastructure in a single lesson, this course spends six.

Chemical engineering, which at this school is Chemical Engineering Principles (ENGR 280), owns reactor design, separations, thermodynamics, and process control, generally in service of making a product. Environmental engineering uses the same reactor theory, and you will recognize plug flow and completely mixed reactors when they show up in Module 4. The differences are that our feedstock is uncontrolled, since you cannot specify what people flush, our product has no market value, and our constraint is a permit rather than a profit margin.

Key idea: Environmental science diagnoses, civil engineering builds the infrastructure, chemical engineering designs reactors for products, and environmental engineering designs processes whose product is a contaminant concentration low enough to satisfy a legal limit.

The city we will follow

Abstractions are hard to hold, so this course carries one worked city through every module. Call it Marley City, population 120,000, on a river in the American Midwest, with a modest industrial base and an aging downtown. Here are its founding numbers, and you will see all of them again.

Quantity	Value	Where it returns
Population	120,000	Everywhere
Drinking water demand	500 liters per person per day, so 60,000 cubic meters per day	Module 3, all three lessons
Wastewater flow	375 liters per person per day, so 45,000 cubic meters per day	Module 4, all three lessons
River flow at the outfall	8.0 cubic meters per second	Modules 2 and 4
Municipal solid waste	2.2 kilograms per person per day, so 264 tonnes per day	Module 6

Sanity-check the water numbers now, because checking your own figures is the habit that separates an engineer from a calculator. Marley City uses 60,000 cubic meters of water per day and produces 45,000 cubic meters of wastewater per day. The ratio is 0.75, and that is right: some water is consumed by irrigation, evaporation, and industrial processes and never returns to the sewer. If you ever compute a wastewater flow larger than the water supply, you have either made an arithmetic error or you have found infiltration of groundwater into leaky sewers, which is a real and expensive phenomenon we will size in Module 4.

Common misconceptions

• Environmental engineers are environmental advocates. Some are, but the job is technical and it serves clients on all sides. Environmental engineers design the emission controls for a refinery, write its permit application, and also write the state's inspection protocol. Competence, not position, is the credential.
• Treatment destroys pollution. Treatment overwhelmingly concentrates and relocates pollution. The residuals stream is part of the design, not an afterthought, and for metals there is no destruction available at all.
• The field is mostly about climate change. Greenhouse gases are now genuinely part of the work, and Module 5 covers them, but the daily practice of most environmental engineers is still water treatment, permits, compliance, and contaminated sites.
• Dilution never works. Dilution works well for degradable substances discharged in modest amounts to large water bodies, which is why receiving water assimilative capacity is a real regulated quantity. It fails for persistent, bioaccumulative, and multiply-sourced contaminants, and the law is built around those failures.

Recap

• Environmental engineering designs processes that keep contaminants away from people and ecosystems across water, air, and land, and it is prescriptive where environmental science is diagnostic.
• The field began as sanitary engineering, and chlorination and filtration essentially ended typhoid in American cities within two decades of 1908.
• The renaming around 1970 followed the statutes and the creation of EPA, which is why this is the most regulation-driven engineering discipline.
• Every calculation in the course is a mass balance: accumulation equals input minus output plus generation minus consumption.
• A worked mixing calculation put 0.20 cubic meters per second of 45 milligram per liter effluent into a 6.0 cubic meter per second river and got 4.35 milligrams per liter downstream.
• Treatment moves and concentrates contaminants rather than destroying them, except where bonds are actually broken, and metals are never destroyed.
• Marley City, population 120,000, uses 60,000 cubic meters of water per day, generates 45,000 cubic meters of wastewater per day, and produces 264 tonnes of solid waste per day.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Our mission and what we do. epa.gov
2. Centers for Disease Control and Prevention. (n.d.). Drinking water: public water systems. U.S. Department of Health and Human Services. cdc.gov
3. U.S. Bureau of Labor Statistics. (n.d.). Occupational Outlook Handbook: environmental engineers. U.S. Department of Labor. bls.gov
4. Encyclopaedia Britannica. (n.d.). Environmental engineering. britannica.com
5. Wikipedia. (n.d.). John Snow. Wikimedia Foundation. en.wikipedia.org

The big picture

In most engineering disciplines you choose the performance target yourself, or a client does. In environmental engineering the target arrives as a number in a permit, written by a state agency under authority delegated by a federal statute, and your job is to design something that meets it for the next twenty-five years while an inspector samples your effluent. That is a strange position to work from, and it means you cannot practice this discipline without knowing the law reasonably well.

You do not need to be a lawyer. You need to know five statutes, what each one regulates, who administers it, and how its requirements turn into numbers you can design against. This lesson gives you that, plus the single most important conceptual distinction in environmental regulation, which is the difference between a limit set by what technology can achieve and a limit set by what the receiving environment can tolerate. We finish by working a real effluent limit for Marley City, and the number we get will follow you into Module 4.

One orienting fact first. Nearly every one of these programs is administered by states, not by the federal government directly. EPA writes the rules and sets a floor, then approves a state to run the program, which the field calls primacy or delegation. So the permit on the wall of a Marley City treatment plant was issued by a state environmental agency, and the inspector who shows up works for that state. States may be stricter than the federal floor and some are considerably stricter. They may not be looser.

Five statutes in one table

Statute	Year	Regulates	Core mechanism
Clean Air Act	1970, amended 1977 and 1990	Emissions to outdoor air	National ambient standards, state implementation plans, source-specific emission standards, operating permits
Clean Water Act	1972	Discharges to surface waters	Discharge permits with numeric limits, water quality standards, total maximum daily loads
Safe Drinking Water Act	1974, amended 1986 and 1996	Water delivered by public water systems	Maximum contaminant levels, treatment techniques, monitoring and reporting
Resource Conservation and Recovery Act	1976, amended 1984	Solid and hazardous waste being generated now	Cradle-to-grave tracking of hazardous waste, disposal facility standards, land disposal restrictions
Comprehensive Environmental Response, Compensation, and Liability Act	1980, amended 1986	Contamination already released	Liability for cleanup, the National Priorities List, a federal cleanup process

The clean way to hold the last two apart is time. RCRA governs waste you are handling today and is forward-looking and prescriptive. CERCLA, which everyone calls Superfund, governs messes made in the past and is backward-looking and about liability. RCRA asks how you will manage this drum. CERCLA asks who is going to pay for the drum somebody buried in 1963.

Key idea: Five federal statutes structure the field by medium and by timing, they are almost always administered by states under delegated authority, and the practical output of all of them is a number in a permit that an engineer must design to meet.

The Clean Water Act and how a permit limit is born

The 1972 Clean Water Act made it unlawful to discharge a pollutant from a point source into navigable waters without a permit. A point source is a discrete conveyance: a pipe, a ditch, an outfall. Diffuse runoff from farms and streets is nonpoint source pollution and is largely outside the permitting system, which is why it is now the dominant remaining cause of water quality impairment in the United States. The permit itself is issued through the National Pollutant Discharge Elimination System, and every engineer in the water field will read hundreds of them.

A permit limit comes from one of two places, and this is the distinction to learn cold.

A technology-based limit asks what a well-run facility of this type can achieve, without reference to the receiving water. For municipal wastewater plants the answer is fixed in federal regulation as the secondary treatment standard: a 30-day average of 30 milligrams per liter for both five-day biochemical oxygen demand and total suspended solids, a 7-day average of 45 milligrams per liter, at least 85 percent removal of both, and pH between 6.0 and 9.0. Every municipal plant in the country must meet that whether it discharges into the Mississippi or into a creek.

A water-quality-based limit asks what the receiving water can tolerate. States adopt water quality standards, which consist of a designated use for each water body, such as aquatic life or drinking water supply, and numeric criteria that protect that use. If meeting the technology floor still leaves the stream violating its criteria, the permit writer computes a stricter limit by working backward from the criterion. That backward calculation is a mass balance, and you already know how to do it.

Worked example. Marley City discharges 45,000 cubic meters per day of treated effluent, which is 45,000 divided by 86,400 seconds, or 0.521 cubic meters per second, into a river. The state criterion for total ammonia nitrogen, protecting aquatic life chronically, is 1.9 milligrams per liter as nitrogen at pH 7 and 20 degrees Celsius. Background ammonia upstream is 0.05 milligrams per liter.

The tempting move is to use the average river flow of 8.0 cubic meters per second. Do that and the mass balance gives 8.0 times 0.05 plus 0.521 times the effluent concentration, all divided by 8.521, set equal to 1.9. That solves to 0.4 plus 0.521 C equals 16.19, so C equals 30.3 milligrams per liter, which is looser than what an ordinary plant already produces. The permit would not bind at all.

But permits are not written for average conditions. They are written for the design low flow, conventionally the 7Q10: the lowest average flow over seven consecutive days expected once in ten years. For this river the 7Q10 is 1.2 cubic meters per second. Redo the arithmetic. Mass in per second is 1.2 times 0.05 equals 0.06 grams, plus 0.521 times C. Total flow is 1.721 cubic meters per second. Setting the mixture to 1.9 gives 0.06 plus 0.521 C equals 1.9 times 1.721, which is 3.27. So 0.521 C equals 3.21, and C equals 6.2 milligrams per liter as nitrogen.

That is the permit limit, and it changes the plant. An ordinary secondary treatment plant that is not designed to nitrify discharges ammonia in the range of 15 to 30 milligrams per liter as nitrogen. To hit 6.2 the plant must convert ammonia to nitrate biologically, which means a longer sludge age, a bigger aeration basin, and roughly double the oxygen demand. In Module 4 we will size exactly that basin and compute exactly that oxygen bill. Notice what happened: one hydrologic statistic, the 7Q10, quietly decided the capital cost of a treatment plant.

Two more Clean Water Act mechanisms you should recognize. Section 303(d) requires states to list waters that fail their standards despite permits, and to write a total maximum daily load for each: the total pollutant load the water body can receive, divided among permitted dischargers as wasteload allocations, nonpoint sources as load allocations, plus a margin of safety. Section 404 requires a permit from the Army Corps of Engineers to place dredged or fill material into waters and wetlands, which is the provision that determines whether a project can build on a wetland. The scope of that jurisdiction narrowed substantially in 2023 when the Supreme Court held in Sackett v. EPA that the Act reaches only wetlands with a continuous surface connection to relatively permanent waters.

Key idea: Whichever is stricter governs, the technology-based floor or the water-quality-based limit, and the water-quality-based calculation is a mass balance run at a design low flow such as the 7Q10 rather than at average conditions.

The Safe Drinking Water Act

The Clean Water Act governs what leaves a pipe into a river. The Safe Drinking Water Act governs what leaves a tap into a person, and it applies to public water systems, defined as systems serving at least 25 people or 15 connections for at least 60 days a year. Private wells are not covered, which surprises people; roughly 13 to 15 percent of Americans drink from unregulated private wells and are their own regulators.

The Act's central device is a pair of numbers for each contaminant. The maximum contaminant level goal, or MCLG, is the concentration below which no known or anticipated adverse health effect occurs, with a margin of safety. It is a health goal, it is not enforceable, and for substances treated as having no threshold, principally genotoxic carcinogens, it is set at zero. The maximum contaminant level, or MCL, is the enforceable standard, and it must be set as close to the MCLG as is feasible using the best available technology, taking cost into consideration.

That gap between the two numbers is where the honest tension of the statute lives. Trichloroethylene has an MCLG of zero and an MCL of 5 micrograms per liter. Arsenic has an MCLG of zero and an MCL of 10 micrograms per liter, lowered from 50 in 2001 after a long fight over cost. The gap is not a scandal; it is the statute openly admitting that zero is unattainable and that the standard must be achievable by thousands of small utilities. About 90 contaminants are regulated this way. Where measuring the contaminant at the tap is impractical, EPA instead specifies a treatment technique, which is a required process rather than a number, as in the Surface Water Treatment Rule and the Lead and Copper Rule.

The Clean Air Act

The Clean Air Act is structurally different because air does not come in pipes. Its core is the National Ambient Air Quality Standards, which are concentrations in the outdoor air that must be achieved everywhere: primary standards protecting public health with an adequate margin of safety, including sensitive populations, and secondary standards protecting welfare such as crops, visibility, and buildings. The Supreme Court held in 2001 that EPA may not consider cost when setting the primary standards, which makes them unusual in American regulation.

Ambient standards do not by themselves control anybody. The mechanism that does is the state implementation plan: each state must submit a plan showing how it will attain and maintain each standard within its borders, and that plan contains the actual rules on actual sources. Areas that fail a standard are designated nonattainment and face progressively harsher requirements, including offsets requiring a new source to buy emission reductions from an existing one. Separately, EPA sets national performance standards for new sources by category and technology-based standards for 187 listed hazardous air pollutants, and large sources hold a single consolidated operating permit under Title V.

The most instructive piece for an engineer is the 1990 acid rain program, which capped total sulfur dioxide emissions from power plants and let plants trade allowances. It is the cleanest natural experiment in environmental regulation. Utility sulfur dioxide emissions fell from roughly 15.7 million tons in 1990 to under 1 million tons by the late 2010s, faster and at far lower cost than either industry or EPA had projected, because the trading system let reductions happen wherever they were cheapest rather than wherever a rule pointed.

RCRA and CERCLA

RCRA divides waste into two worlds. Subtitle C covers hazardous waste and imposes cradle-to-grave control: a generator must determine whether its waste is hazardous, obtain an identification number, use a manifest that tracks each shipment, and ship only to a permitted treatment, storage, and disposal facility. The 1984 amendments added land disposal restrictions, which prohibit landfilling hazardous waste until it has been treated to specified standards, a rule that transformed the industry by making treatment cheaper than disposal. Subtitle D covers ordinary solid waste, which states manage under federal minimum criteria that govern municipal landfill siting, liners, leachate collection, groundwater monitoring, and thirty years of post-closure care. Module 6 designs to those criteria.

CERCLA responds to sites that are already contaminated, and it was passed in December 1980 in direct response to Love Canal. Its power comes from a liability scheme that is strict, meaning no negligence need be shown, joint and several, meaning any one responsible party can be made to pay the whole cost, and retroactive, meaning conduct that was legal at the time still creates liability. Potentially responsible parties include current owners, past owners at the time of disposal, generators of the waste, and transporters who selected the site. The worst sites go on the National Priorities List, currently around 1,340 sites, and move through a defined sequence: remedial investigation, feasibility study, a record of decision selecting the remedy, design, construction, and five-year reviews forever after.

Key idea: RCRA is forward-looking and prescriptive about waste being handled now, while CERCLA is backward-looking and about who pays for contamination already in the ground, using strict, joint and several, retroactive liability.

Common misconceptions

• EPA runs these programs. EPA writes rules and sets floors, but states hold primacy for most permitting and enforcement, and the person who inspects a plant almost always works for a state agency.
• A drinking water MCL represents a safe level. The MCL is the feasible level. The health goal is the MCLG, which for carcinogens is zero, and the difference between them is an explicit judgment about cost and available technology.
• Permits are set by what the river can handle. Sometimes. The technology-based floor applies regardless, and only if the floor is insufficient to protect the water body does the stricter water-quality-based calculation control.
• Superfund is paid for by a tax on polluters. The dedicated excise taxes lapsed in 1995 and were only partly restored on chemicals in 2021, and in practice most cleanups are performed or paid for by identified responsible parties under enforcement agreements.
• An environmental impact statement can stop a bad project. The National Environmental Policy Act is procedural. It requires a federal agency to study and disclose environmental effects, not to choose the least damaging alternative.

Recap

• Five statutes structure the field: Clean Air Act for outdoor air, Clean Water Act for surface discharges, Safe Drinking Water Act for tap water, RCRA for waste now, CERCLA for contamination already released.
• States hold primacy for most programs, may be stricter than the federal floor, and may not be looser.
• Technology-based limits ask what a well-run facility can achieve; water-quality-based limits ask what the receiving water can tolerate; the stricter one governs.
• The secondary treatment standard is 30 milligrams per liter of BOD and TSS as a 30-day average with 85 percent removal.
• Working Marley City's ammonia limit at a 7Q10 of 1.2 cubic meters per second gave 6.2 milligrams per liter as nitrogen, which forces the plant to nitrify; at average flow the same criterion would have given a meaningless 30 milligrams per liter.
• Drinking water uses an unenforceable MCLG health goal and an enforceable MCL set as close to it as feasible, with treatment techniques where measurement is impractical.
• The Clean Air Act works through ambient standards implemented by state plans, and the 1990 sulfur dioxide trading program cut utility emissions from about 15.7 million tons to under 1 million.
• CERCLA liability is strict, joint and several, and retroactive, which is what makes it effective and what makes it litigated.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Laws and regulations. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). National Pollutant Discharge Elimination System (NPDES). epa.gov
3. U.S. Environmental Protection Agency. (n.d.). National primary drinking water regulations. epa.gov
4. U.S. Environmental Protection Agency. (n.d.). NAAQS table. epa.gov
5. U.S. Environmental Protection Agency. (n.d.). Superfund: CERCLA overview. epa.gov

The big picture

Every environmental design eventually runs into the same question, and it is not a technical question until you make it one: how clean is clean enough? You can always remove more. Each additional increment costs more than the last, often far more, and at some point the money would save more lives spent on something else. Somebody has to draw the line, and the framework that lets an engineer draw it defensibly is quantitative risk assessment.

This lesson teaches that framework and then works its arithmetic on a real contaminant, trichloroethylene in a drinking water well, until you can produce a number. Then it turns to the part of the subject that the arithmetic cannot capture. Risk assessment computes a probability for a hypothetical average person. It does not tell you who that person is, where they live, or what else they are already breathing. That question, the distribution of environmental burden across real communities, has its own history and its own effect on how an engineer designs, and the second half of the lesson takes it seriously.

The four steps

The structure comes from a 1983 National Research Council report so influential that the field simply calls it the Red Book. It separates risk assessment, which is technical, from risk management, which is a policy decision about what to do with the number. Keeping those apart is the whole point, because mixing them lets the desired answer drive the analysis.

1. Hazard identification. Can this substance cause harm at all, and what kind? Evidence comes from human epidemiology when it exists, animal bioassays, and mechanistic studies. The output is qualitative: this chemical is a probable human carcinogen, or it damages the developing kidney.
2. Dose-response assessment. How much harm per unit dose? The output is a number: a reference dose for non-cancer effects, or a cancer slope factor.
3. Exposure assessment. How much actually reaches the person, by what route, for how long? The output is an intake in milligrams per kilogram of body weight per day.
4. Risk characterization. Combine steps two and three, state the result, and be explicit about the uncertainty. The output is a probability or a hazard ratio, plus an honest paragraph.

Key idea: Risk assessment has four steps, hazard identification, dose-response, exposure, and characterization, and it is deliberately separated from risk management so that the policy preference cannot quietly select the technical answer.

Dose-response: the two shapes

Paracelsus said in the sixteenth century that the dose makes the poison, and the modern field is an elaboration of that sentence. Everything is toxic at some dose, including water and oxygen. The question is always the shape of the curve near the low doses people actually experience, and toxicology treats two shapes very differently.

For most non-cancer effects the curve is assumed to have a threshold. The body repairs damage up to some rate; below that dose nothing measurable happens. The analysis finds the highest tested dose at which no adverse effect appeared, the no-observed-adverse-effect level or NOAEL, and then divides by uncertainty factors to get a reference dose, a daily intake likely to be without appreciable risk over a lifetime.

Worked example. A chronic rat study finds a NOAEL of 5 milligrams per kilogram per day. Apply a factor of 10 for extrapolating from rats to humans, another 10 for variation among humans including sensitive individuals, and another 10 because the study was subchronic rather than lifetime. The product is 1,000. The reference dose is 5 divided by 1,000, or 0.005 milligrams per kilogram per day. Notice that these factors are conventions, not measurements. They are conservative by design and everyone in the field knows it.

For genotoxic carcinogens the assumption is different and stricter: no threshold. If a single molecule can in principle damage DNA in a way that leads to a tumor, then any dose carries some probability, and the dose-response line is extrapolated linearly from the observable high-dose data down through the origin. The slope of that line is the cancer slope factor, in units of inverse milligrams per kilogram per day. Multiply an intake by a slope factor and you get an excess lifetime cancer probability. This linear no-threshold assumption is a policy choice made in the face of unresolvable uncertainty, it is almost certainly conservative for some chemicals, and it is why the drinking water health goal for a carcinogen is zero.

Exposure and the arithmetic

Here is the standard intake equation for drinking water, and it is just bookkeeping. Chronic daily intake equals concentration times ingestion rate times exposure frequency times exposure duration, divided by body weight times averaging time.

Worked example. A private well serving a subdivision outside Marley City tests at 20 micrograms per liter of trichloroethylene, a degreasing solvent from an old machine shop. That is 0.020 milligrams per liter. Use the classic default residential exposure assumptions: ingestion 2 liters per day, exposure frequency 350 days per year, exposure duration 30 years, body weight 70 kilograms.

Numerator first, and keep the units visible. 0.020 milligrams per liter times 2 liters per day is 0.040 milligrams per day. Times 350 days per year is 14 milligrams per year. Times 30 years is 420 milligrams over the whole exposure.

Now the averaging time, and this is where students go wrong. For cancer, risk is treated as depending on the lifetime average dose, so the averaging time is a full 70-year lifetime: 70 times 365 equals 25,550 days. The denominator is 70 kilograms times 25,550 days, or 1,788,500. The chronic daily intake is 420 divided by 1,788,500, which is 2.35 times ten to the minus four milligrams per kilogram per day.

The oral cancer slope factor for trichloroethylene is 0.046 per milligram per kilogram per day. Excess lifetime cancer risk is intake times slope factor: 2.35 times ten to the minus four, times 0.046, equals 1.08 times ten to the minus five. That is roughly one additional cancer case per 93,000 people exposed at that level for thirty years.

For non-cancer effects the averaging time is only the exposure duration itself, because a threshold effect does not average out over years you were not exposed. So the denominator is 70 kilograms times 30 times 365, or 766,500. That intake is 420 divided by 766,500, or 5.48 times ten to the minus four milligrams per kilogram per day. The reference dose for trichloroethylene is 0.0005 milligrams per kilogram per day, so the hazard quotient is 5.48 times ten to the minus four divided by 5 times ten to the minus four, which is 1.10.

Read both results together. The cancer risk of about one in one hundred thousand sits inside the range EPA generally treats as acceptable for site cleanups, one in a million to one in ten thousand. The hazard quotient of 1.10 is above 1.0, which means the non-cancer margin of safety has been consumed. When a hazard quotient exceeds one you do not conclude that people will be harmed; you conclude that you have run out of the cushion the reference dose was built to provide, and that is enough to act on. In this case the non-cancer endpoint, not the cancer endpoint, is what drives the decision, which is the opposite of what most people expect.

Now run the calculation backward, which is what a remediation engineer actually does. What concentration corresponds to a one-in-a-million cancer risk? The target intake is ten to the minus six divided by 0.046, or 2.17 times ten to the minus five milligrams per kilogram per day. Multiply by the denominator 1,788,500 and divide by 2 times 350 times 30, which is 21,000. That gives 38.9 divided by 21,000, or 0.00185 milligrams per liter, which is 1.85 micrograms per liter.

The federal drinking water standard for trichloroethylene is 5 micrograms per liter. So the enforceable limit sits at a risk of about 2.7 in a million, not one in a million. That is not a scandal and it is not a secret: the maximum contaminant level goal for trichloroethylene is zero, and the enforceable standard is set where feasible treatment and analytical detection make it achievable for every water system in the country. When you understand that gap you understand most arguments about environmental standards.

Key idea: Cancer intake averages over a 70-year lifetime while non-cancer intake averages only over the exposure duration, which is why the same 20 micrograms per liter of trichloroethylene gave a cancer risk of about 1 in 93,000 and a hazard quotient of 1.10, with the non-cancer endpoint driving the decision.

Where the uncertainty actually lives

The arithmetic above is precise to three significant figures and the answer is not. Be honest about which inputs are soft. The slope factor comes from extrapolating animal tumor data across species and down three or four orders of magnitude in dose, and reasonable analyses of the same data set can differ by a factor of ten. The exposure assumptions are conventions: EPA's own Exposure Factors Handbook now supports an adult body weight closer to 80 kilograms and a higher drinking water intake for some populations, and swapping those in moves the answer. Exposure duration of 30 years is a residential default; people move. And the calculation covers one chemical by one route, while a real well contains a mixture and a real person also showers in it, which for a volatile compound like trichloroethylene can contribute inhalation exposure comparable to ingestion.

The professional response to all of that is not to abandon the method. It is to present the result as a range, to state the assumptions on the same page as the number, and never to write a risk estimate without the sentence explaining what would change it.

Environmental justice as a design constraint

Everything above computes a risk for a generic person. It says nothing about who that person is. The field of environmental justice exists because the answer to who turned out not to be random.

The usual starting point is Warren County, North Carolina, in 1982. The state chose a site in a rural, majority-Black county for a landfill to hold soil contaminated with polychlorinated biphenyls that had been illegally sprayed along highway shoulders. Residents lay down in the road in front of the trucks and more than 500 people were arrested. The protest did not stop the landfill, but it made the pattern visible and gave the movement its name.

Evidence followed. A 1983 U.S. General Accounting Office study of hazardous waste landfills in the Southeast found three of four sited in majority-Black communities. The United Church of Christ Commission for Racial Justice published Toxic Wastes and Race in the United States in 1987, the first national analysis, and reported that race was the strongest single predictor of the location of commercial hazardous waste facilities, stronger than income. In 1994 Executive Order 12898 directed federal agencies to identify and address disproportionately high and adverse human health or environmental effects on minority and low-income populations, and EPA now publishes EJScreen, a public mapping tool that overlays demographic data with environmental indicators.

An honest treatment has to include the causal argument, because engineers will hear it. Some researchers have asked which came first: were facilities sited in already-disadvantaged communities, or did the demographics of neighborhoods change after siting because property values fell and those with means left? Longitudinal work has found evidence for both mechanisms in different places. The important thing for an engineer is that the disagreement is about mechanism, not about the present distribution. The distribution itself is not seriously contested, and it is the distribution that determines who breathes what tomorrow.

What follows for design? Four things, all of them concrete.

First, siting is an engineering alternative. When you evaluate locations for a transfer station, a landfill, a compressor station, or a truck route, the demographic and existing-burden profile of each candidate belongs in the alternatives matrix alongside cost and haul distance, not in a separate document written later.

Second, cumulative burden is not captured by single-facility permitting. Each permit asks whether this source alone meets its standard. A neighborhood bounded by a highway, a rail yard, a refinery, and a port can host four facilities that each comply and still carry an aggregate exposure no single permit ever evaluated. Several states have begun writing cumulative impact requirements into their permitting statutes, and an engineer will increasingly be asked to quantify aggregate burden rather than facility-by-facility compliance.

Third, engagement is a design input. A public meeting held at 2 p.m. on a Tuesday, in English only, in a building reachable only by car, produces a record of no objection that means nothing. Residents routinely hold information the monitoring network does not: which days the smell is worst, which well went bad, which children are sick. Flint is the clearest example in living memory of what it costs to dismiss that information.

Fourth, monitoring commitments are part of the design. Fenceline monitoring with public data reporting converts a promise into a verifiable fact and is now standard practice at some refinery permits.

Key idea: Risk assessment produces an average, and environmental justice asks about the distribution around it, which reaches engineering practice through siting alternatives, cumulative burden analysis, real community engagement, and verifiable monitoring.

Common misconceptions

• A hazard quotient above 1 means people will get sick. It means the intake has exceeded the reference dose, which was already built with uncertainty factors of 100 or 1,000. It is a signal to act, not a prediction of illness.
• Risk assessment tells you what to do. It produces a number. The decision about acceptable risk, cost, and who bears the residual is risk management, and it is a policy judgment made by people with names.
• A one-in-a-million risk is the legal requirement. EPA generally manages site cleanups to a range from one in a million to one in ten thousand, and drinking water standards frequently sit above the one-in-a-million point because of feasibility, as trichloroethylene at 5 micrograms per liter does.
• Environmental justice is a public relations matter. It changes siting alternatives, permit conditions, cumulative impact analysis, and monitoring obligations, all of which are engineering scope with costs attached.
• Cancer and non-cancer intakes use the same arithmetic. They differ in averaging time, lifetime for cancer and exposure duration for non-cancer, which changed the same exposure by a factor of 2.3 in the worked example.

Recap

• Risk assessment runs in four steps and stays separate from risk management by design.
• Non-cancer effects are treated as having a threshold, giving a reference dose from a NOAEL divided by uncertainty factors; genotoxic carcinogens are treated as having none, giving a linear slope factor.
• A rat NOAEL of 5 milligrams per kilogram per day with uncertainty factors of 10, 10, and 10 gave a reference dose of 0.005.
• Trichloroethylene at 20 micrograms per liter gave a chronic daily intake of 2.35 times ten to the minus four and a cancer risk of 1.08 times ten to the minus five, about one in 93,000.
• The same exposure gave a non-cancer intake of 5.48 times ten to the minus four and a hazard quotient of 1.10, so the non-cancer endpoint governed.
• Back-calculating for a one-in-a-million risk gave 1.85 micrograms per liter, below the enforceable standard of 5, which reflects feasibility rather than pure risk.
• Warren County in 1982, the 1987 United Church of Christ study, and Executive Order 12898 in 1994 established that environmental burden is distributed unevenly, a finding whose mechanism is debated but whose distribution is not.
• Environmental justice enters practice through siting alternatives, cumulative burden, meaningful engagement, and fenceline monitoring.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Human health risk assessment. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Integrated Risk Information System (IRIS). epa.gov
3. Agency for Toxic Substances and Disease Registry. (n.d.). Toxicological profile for trichloroethylene. Centers for Disease Control and Prevention. atsdr.cdc.gov
4. U.S. Environmental Protection Agency. (n.d.). EJScreen: environmental justice screening and mapping tool. epa.gov
5. Wikipedia. (n.d.). Warren County PCB Landfill. Wikimedia Foundation. en.wikipedia.org

The big picture

This is the lesson that makes every later lesson possible, and it is the one students are most tempted to skim. Do not. Environmental engineering runs on concentrations and loadings, and the field expresses the same physical quantity in at least six different ways depending on who is talking. A laboratory reports milligrams per liter. A regulation says parts per billion. A chemist wants molarity. A water softening calculation demands milligram-equivalents. A drinking water standard for nitrate is written on a nitrogen basis while the laboratory reports the nitrate ion. An air quality standard is in parts per million by volume while the monitor reports micrograms per cubic meter.

Getting one of those conversions wrong does not produce a slightly wrong answer. It produces an answer off by a factor of four, or forty-four, and it has caused real overdoses of treatment chemicals and real reporting violations. By the end of this lesson you should be able to move among all of them without thinking, and you should have two identities memorized so thoroughly that they feel like arithmetic rather than chemistry.

The two identities that make everything easy

Here they are. Learn them now.

One milligram per liter equals one gram per cubic meter. Check it: a cubic meter is 1,000 liters, so 1 milligram per liter times 1,000 liters is 1,000 milligrams, which is 1 gram. This identity means you can switch between the units a lab uses and the units a flow calculation uses without touching a calculator.

One cubic meter per second times one milligram per liter equals one gram per second. This falls straight out of the first identity, and it is the reason every mixing and loading calculation in this course takes one line.

From those, mass loading in kilograms per day is concentration in milligrams per liter times flow in cubic meters per day, divided by 1,000.

Worked example. Marley City treats 60,000 cubic meters of water per day and applies a chlorine dose of 2.5 milligrams per liter. The chemical feed is 2.5 times 60,000 divided by 1,000, or 150 kilograms of chlorine per day. That is the number the operator orders against, and at a delivered price of roughly 1 dollar per kilogram for bulk sodium hypochlorite on a chlorine-equivalent basis, it is also the annual chemical budget line: about 55 tonnes and 55,000 dollars a year for one chemical at one plant.

Second worked example. The river past Marley City runs at 8.0 cubic meters per second carrying 15 milligrams per liter of suspended sediment. Using the second identity, that is 8.0 times 15 equals 120 grams per second. Over a day, 120 times 86,400 seconds equals 10,368,000 grams, or about 10.4 tonnes of sediment per day moving past the city. Sediment loads are why reservoirs fill in.

Key idea: One milligram per liter is one gram per cubic meter, and one cubic meter per second at one milligram per liter is one gram per second, so mass loading in kilograms per day equals milligrams per liter times cubic meters per day divided by 1,000.

Milligrams per liter, parts per million, and percent

In dilute water solutions, one milligram per liter equals one part per million by mass, and one microgram per liter equals one part per billion. The reason is that a liter of water has a mass of 1,000 grams, or one million milligrams, so one milligram in it is one part per million. That equivalence is an approximation resting on the density of water being 1.000 kilograms per liter, and it holds to well within analytical precision for fresh water and for ordinary wastewater.

It breaks down where the density departs from one. Seawater has a density near 1.025 kilograms per liter. Brines from reverse osmosis and from oil field production can reach 1.2. Thickened sludge is denser still. For those you must carry the density explicitly, and the correct relation is parts per million by mass equals milligrams per liter divided by density in kilograms per liter.

Percent is parts per hundred, so one percent by mass is 10,000 milligrams per liter in dilute solution. That conversion appears constantly in sludge work: a primary sludge at 4 percent solids contains 40,000 milligrams per liter of dry solids, or equivalently 40 kilograms of dry solids in every cubic meter. Keep that in your pocket for Module 4.

Air is different and this catches everyone. In water, parts per million is a mass ratio. In air, parts per million means parts per million by volume, which for an ideal gas is also a mole ratio. Those are not the same thing and there is no getting around the distinction. We come back to it below.

Molarity and equivalents

A chemist needs moles, because reactions happen mole for mole. Molarity is moles of solute per liter of solution, and the conversion is simply division by molar mass, with a factor of 1,000 to move from milligrams to grams.

Worked example. A groundwater contains 100 milligrams per liter of calcium ion. Calcium has a molar mass of 40.08 grams per mole. So the molarity is 0.100 grams per liter divided by 40.08 grams per mole, or 2.50 times ten to the minus three moles per liter, which the field writes as 2.50 millimolar.

Now the third layer. Many water chemistry calculations, especially softening, ion exchange, and alkalinity, care about charge rather than moles, because ions exchange charge for charge. The unit is the equivalent, and the equivalent weight is the molar mass divided by the charge magnitude. Calcium ion carries a charge of two, so its equivalent weight is 40.08 divided by 2, or 20.04 grams per equivalent. Our 100 milligrams per liter is therefore 100 divided by 20.04, or 4.99 milligram-equivalents per liter.

The word "as" and why it matters

Environmental reports constantly express a concentration as some other substance. This is not sloppiness. It is a device for making different ions comparable, and you must be able to read it.

As calcium carbonate. Hardness and alkalinity are both expressed as milligrams per liter as calcium carbonate, which has a molar mass of 100.09 and a charge of two, giving an equivalent weight of 50.04 grams per equivalent. To convert, take the concentration in milligram-equivalents per liter and multiply by 50.04. Our calcium at 4.99 milligram-equivalents per liter is therefore 4.99 times 50.04, or 250 milligrams per liter as calcium carbonate. Water above about 180 milligrams per liter as calcium carbonate is classed as very hard, so this groundwater will scale water heaters and eat soap.

As nitrogen. This one causes real violations. The drinking water standard for nitrate is 10 milligrams per liter as nitrogen. The nitrate ion has a molar mass of 62.00 and contains one nitrogen atom of mass 14.01, so the ratio is 62.00 divided by 14.01, or 4.43. A laboratory reporting 45 milligrams per liter as nitrate is reporting 45 divided by 4.43, or 10.2 milligrams per liter as nitrogen, which is a violation. The same laboratory reporting 45 milligrams per liter as nitrogen would be describing water four and a half times worse. Always read the basis. The same discipline applies to ammonia, which may be reported as ammonia or as nitrogen with a ratio of 17.03 to 14.01, and to phosphate, reported as phosphate or as phosphorus with a ratio of 94.97 to 30.97.

Species	Reported as	Multiply by to get the elemental basis
Nitrate, NO3	NO3	Divide by 4.43 to get milligrams per liter as N
Nitrite, NO2	NO2	Divide by 3.28 to get milligrams per liter as N
Ammonium, NH4	NH4	Divide by 1.29 to get milligrams per liter as N
Orthophosphate, PO4	PO4	Divide by 3.07 to get milligrams per liter as P
Sulfate, SO4	SO4	Divide by 3.00 to get milligrams per liter as S

Key idea: A concentration expressed as calcium carbonate or as nitrogen is a deliberate change of basis, and reading it wrong changes the number by factors of three to five, which is enough to turn a compliant sample into a violation.

Air concentrations

Now the air side. Gaseous pollutants are usually reported either as a volume fraction, in parts per million or parts per billion by volume, or as a mass concentration in micrograms per cubic meter. Converting requires the molar volume of an ideal gas, which at 25 degrees Celsius and 1 atmosphere is 24.45 liters per mole. At 0 degrees Celsius it is 22.4 liters per mole, and standards specify which reference condition applies, so check.

The working relation is that micrograms per cubic meter equals parts per million by volume times molar mass times 1,000, divided by 24.45.

Worked example. The national ambient standard for ozone is 0.070 parts per million as an eight-hour average. Ozone has a molar mass of 48.00. So 0.070 times 48.00 times 1,000, divided by 24.45, gives 3,360 divided by 24.45, or about 137 micrograms per cubic meter. That is worth knowing, because the World Health Organization guideline for the same averaging period is 100 micrograms per cubic meter, so you can now see directly that the international guideline is meaningfully stricter than the American standard.

Second worked example. The one-hour sulfur dioxide standard is 75 parts per billion, which is 0.075 parts per million. Sulfur dioxide has a molar mass of 64.07. So 0.075 times 64.07 times 1,000 divided by 24.45 gives 4,805 divided by 24.45, or about 197 micrograms per cubic meter.

Particulate matter has no molar mass and is therefore always reported as a mass concentration, in micrograms per cubic meter, never in parts per million. If you ever see particulate matter quoted in parts per million, someone has made an error.

Removal efficiency, and reporting honestly

Percent removal across a process is simply influent minus effluent, divided by influent, times 100. A plant taking 250 milligrams per liter of biochemical oxygen demand down to 15 achieves 250 minus 15, over 250, or 94.0 percent removal. Two traps live here. First, removal must be computed on the same basis, and if flows differ between the two points, as when a side stream returns, you must compare masses rather than concentrations. Second, high percentage removal is easy at high influent concentration and brutally hard at low; going from 250 to 15 is routine, while going from 15 to 5 may require an entire additional process train.

Finally, be careful about numbers near the bottom of the analytical range. A method detection limit is the lowest concentration that can be distinguished from zero with confidence; a reporting limit is the somewhat higher level at which the laboratory will state a numerical value. Results below these come back as non-detect, which is not the same as zero. The common convention when averaging is to substitute half the detection limit for a non-detect, and any report that does this must say so, because with many non-detects that convention can drive the whole answer.

Key idea: Percent removal must be computed on matched flows and matched bases, non-detect is not zero, and the last increment of removal always costs far more than the first.

Common misconceptions

• Parts per million always means the same thing. In water it is a mass ratio equal to milligrams per liter; in air it is a volume ratio requiring the molar volume to convert. They are different quantities that share a name.
• Milligrams per liter equals parts per million exactly. Only when the solution density is 1.000 kilograms per liter. In seawater, brines, and sludges you must divide by the density.
• Nitrate at 45 milligrams per liter is safely under the 10 milligram standard. Only if both numbers are on the same basis. Forty-five as nitrate is 10.2 as nitrogen, which is a violation.
• Hardness as calcium carbonate means the water contains calcium carbonate. It contains calcium and magnesium ions; expressing them as calcium carbonate is a charge-equivalent bookkeeping convention that lets different ions be added together.
• A non-detect means the contaminant is absent. It means the concentration is below what the method can distinguish from zero, and how non-detects are treated can dominate a computed average.

Recap

• One milligram per liter equals one gram per cubic meter, and one cubic meter per second at one milligram per liter equals one gram per second.
• Loading in kilograms per day equals milligrams per liter times cubic meters per day divided by 1,000; Marley City at 60,000 cubic meters per day and a 2.5 milligram per liter chlorine dose needs 150 kilograms per day.
• In dilute water, milligrams per liter equals parts per million; one percent equals 10,000 milligrams per liter, so 4 percent sludge is 40 kilograms of dry solids per cubic meter.
• Calcium at 100 milligrams per liter is 2.50 millimolar, 4.99 milligram-equivalents per liter, and 250 milligrams per liter as calcium carbonate, which is very hard water.
• Nitrate reported as the ion must be divided by 4.43 to compare against the 10 milligram per liter as nitrogen standard.
• Micrograms per cubic meter equals parts per million by volume times molar mass times 1,000 divided by 24.45, giving 137 for the 0.070 parts per million ozone standard.
• Percent removal must use matched flows and bases, and non-detect is not zero.

Sources

1. U.S. Geological Survey. (n.d.). Water Science School. U.S. Department of the Interior. usgs.gov
2. U.S. Environmental Protection Agency. (n.d.). Basic information about nitrate in drinking water. epa.gov
3. U.S. Environmental Protection Agency. (n.d.). NAAQS table. epa.gov
4. World Health Organization. (n.d.). Air quality, energy and health: guidelines. who.int
5. Wikipedia. (n.d.). Hard water. Wikimedia Foundation. en.wikipedia.org

The big picture

Two properties of water decide most of what an environmental engineer can do to it. The first is pH, along with the buffering capacity that resists changes in pH, because pH governs which chemical species exist, whether metals dissolve or precipitate, how well disinfectants work, and whether pipes corrode. The second is dissolved oxygen, because oxygen is what aquatic life breathes and what microorganisms consume, and the whole logic of wastewater treatment is the management of a single oxygen budget.

This lesson builds both, then introduces the measurement that ties them together, biochemical oxygen demand. You will finish with a worked BOD bottle calculation, an alkalinity number for a real water, and the ability to predict what happens to a river when Marley City's plant discharges into it.

The carbonate system

Natural water is not pure water. It contains dissolved carbon dioxide from the atmosphere and from decaying organic matter, and where it has touched limestone it contains dissolved carbonate minerals. Those species interconvert with pH, and this carbonate system is the master variable of water chemistry.

Carbon dioxide dissolves and reacts with water to form carbonic acid. Carbonic acid loses a proton to become bicarbonate, and bicarbonate loses another to become carbonate. The two dissociation constants have negative logarithms of about 6.35 and 10.33 at 25 degrees Celsius, and those two numbers tell you everything about which species dominates.

pH range	Dominant species	Where you find it
Below 6.35	Dissolved carbon dioxide and carbonic acid	Rainwater, acidic soft-water lakes, mine drainage
6.35 to 10.33	Bicarbonate	Nearly all natural surface and ground water, drinking water, wastewater
Above 10.33	Carbonate	Lime-softened water, soda lakes, some industrial streams

Note the middle row. Essentially every water an environmental engineer works with sits in the bicarbonate window, roughly pH 6.5 to 8.5. That is not a coincidence, it is the carbonate system holding it there.

One consequence worth knowing: pure water in equilibrium with atmospheric carbon dioxide has a pH near 5.6, not 7.0. That is the natural pH of rain. Precipitation is called acid rain when sulfuric and nitric acids from combustion push it well below 5.6, and in the northeastern United States before the acid rain program, rainfall pH near 4.2 was routine.

Alkalinity

Alkalinity is the capacity of a water to neutralize acid. Chemically it is the sum of bicarbonate, twice carbonate, and hydroxide, minus hydrogen ion, and in almost all natural waters the bicarbonate term dominates everything else. In the laboratory it is measured by titrating with standard acid to about pH 4.5, and it is reported in milligrams per liter as calcium carbonate.

Alkalinity matters to an engineer for three practical reasons. It buffers the water, so a given acid addition changes pH less. It is consumed by acidic treatment chemicals, principally coagulants, and running out of it causes the pH to crash. And it participates in corrosion control, because bicarbonate is part of what forms protective scale on pipe interiors.

Worked example. A source water contains 150 milligrams per liter of bicarbonate. The equivalent weight of bicarbonate is its molar mass of 61.02 divided by its charge of one, so 61.02 grams per equivalent. That gives 150 divided by 61.02, or 2.458 milligram-equivalents per liter. Multiply by the calcium carbonate equivalent weight of 50.04 and the alkalinity is 123 milligrams per liter as calcium carbonate. Waters below about 50 are considered poorly buffered and are the ones that give operators trouble.

Worked example, continued. Marley City's treatment plant doses aluminum sulfate, called alum, as its coagulant. Alum is acidic in water: it hydrolyzes to aluminum hydroxide floc and releases protons that consume alkalinity. Working the stoichiometry, one mole of hydrated alum with a molar mass of 594 consumes three moles of calcium bicarbonate, equivalent to 300 grams as calcium carbonate. The ratio 300 divided by 594 is 0.505, so each milligram per liter of alum consumes about 0.50 milligrams per liter of alkalinity as calcium carbonate.

At a normal dose of 30 milligrams per liter of alum, the plant consumes 15 milligrams per liter of alkalinity, dropping from 123 to 108. Nothing happens; the water is still well buffered. But suppose a spring runoff event triples the turbidity and the operator raises the dose to 200 milligrams per liter. Now 100 milligrams per liter of alkalinity is consumed, leaving 23. The pH falls sharply, coagulation itself becomes less effective because aluminum hydroxide precipitation has an optimum pH window around 6 to 7, and the finished water leaves the plant aggressive toward pipe. The fix is to feed lime or soda ash alongside the alum to restore alkalinity, and this is a routine operational calculation at every conventional plant.

Key idea: Alkalinity is acid-neutralizing capacity dominated by bicarbonate and reported as calcium carbonate, and each milligram per liter of alum consumes about 0.50 milligrams per liter of it, so a high coagulant dose on a poorly buffered water crashes the pH unless lime is added.

Dissolved oxygen

Oxygen is only slightly soluble in water, and that single fact organizes an entire branch of the profession. Air is 21 percent oxygen, so a room contains roughly 280 grams of oxygen per cubic meter. Water in equilibrium with that same air at 20 degrees Celsius holds about 9.1 grams per cubic meter, which is 9.1 milligrams per liter. Water carries about one thirtieth as much oxygen as air, and a fish must extract it from a far denser medium.

Saturation depends strongly on temperature, and moderately on pressure and salinity.

Temperature, degrees Celsius	Saturation dissolved oxygen, milligrams per liter, fresh water at sea level
0	14.6
10	11.3
20	9.1
25	8.3
30	7.6

Read that table as a warning. A warm summer river holds a quarter less oxygen than the same river in spring, and it holds it precisely when biological activity, and therefore oxygen consumption, is at its highest. Thermal discharges from power plants matter for exactly this reason. Altitude matters too: saturation scales roughly with barometric pressure, so at 1,500 meters where pressure is about 0.84 atmospheres, the 20-degree saturation falls from 9.1 to about 7.6 milligrams per liter. Salinity reduces it further, which is why seawater holds less oxygen than fresh.

The biological thresholds are simple and worth memorizing. Most fish are comfortable above 5 milligrams per liter, stressed between 2 and 5, and dying below 2. Water below about 2 milligrams per liter is called hypoxic, and at zero it is anoxic, where the chemistry turns reducing, sulfide forms, iron and manganese dissolve, and the water begins to smell.

Biochemical oxygen demand

Now the measurement that connects organic pollution to oxygen. Biochemical oxygen demand is the mass of oxygen that microorganisms consume while degrading the organic matter in a sample, conventionally measured over five days in the dark at 20 degrees Celsius. It is not a chemical species. It is a bioassay, and it is a strange, slow, imprecise test that the field has kept for over a century because it measures exactly the thing that matters: how much oxygen this water will steal from a river.

The procedure is simple. Fill a 300 milliliter glass bottle with a diluted sample plus aerated dilution water seeded with microorganisms. Measure the dissolved oxygen. Stopper it with no air bubble, incubate five days at 20 degrees, and measure the dissolved oxygen again. The difference, scaled by the dilution, is the BOD.

Worked example. An operator pipettes 5.00 milliliters of raw wastewater into a 300 milliliter bottle and fills the rest with dilution water. Initial dissolved oxygen is 8.8 milligrams per liter. After five days it is 3.2 milligrams per liter. The dilution fraction P is 5.00 divided by 300, or 0.01667. So the BOD is the depletion divided by P: 8.8 minus 3.2 is 5.6, divided by 0.01667, which is 336 milligrams per liter.

Two validity checks are built into the method and you should apply them every time. The depletion must be at least 2.0 milligrams per liter, or the measurement is lost in the noise; here it is 5.6, so we are fine. And the remaining oxygen must be at least 1.0 milligram per liter, or the microorganisms ran out of oxygen partway through and the result understates the true demand; here 3.2 remains, so we are fine. When both checks fail on the same sample, the operator must run a series of dilutions, which is why a BOD test in practice means five bottles, not one.

Why five days? Because the original English researchers chose it so the test would finish before the Thames reached the sea. It is a historical accident that became a global standard, and it captures only part of the total demand. Degradation follows approximately first-order kinetics: the BOD exerted by time t equals the ultimate BOD times the quantity one minus e to the minus k t.

Worked example, continued. For municipal wastewater the rate constant k is often near 0.23 per day, base e. At five days, one minus e to the minus 1.15 equals one minus 0.317, or 0.683. So a five-day BOD of 336 milligrams per liter implies an ultimate BOD of 336 divided by 0.683, or 492 milligrams per liter. The five-day test sees only about two thirds of the total oxygen the sample will eventually demand.

There is a second demand hiding in the same sample. Ammonia in the wastewater is oxidized to nitrate by nitrifying bacteria, and that reaction consumes 4.57 grams of oxygen for every gram of nitrogen oxidized. This nitrogenous demand normally starts after five to eight days because nitrifiers grow slowly, so it does not usually corrupt a five-day test, but in a nitrifying plant's effluent it can. The standard fix is a nitrification inhibitor, and the result is reported as carbonaceous BOD, or CBOD, which is what most permits actually limit.

Because the BOD test takes five days, operators also use faster surrogates. Chemical oxygen demand uses a strong chemical oxidant and gives an answer in two hours, but it oxidizes things bacteria will not, so it always reads higher; for municipal wastewater the ratio of COD to BOD is typically about 2 to 1. Total organic carbon is faster still. Neither replaces BOD in a permit, but a plant that establishes its own local ratio can use COD for daily control and BOD for compliance.

Key idea: BOD is a five-day bioassay of the oxygen a water will consume; a 5.00 milliliter sample in a 300 milliliter bottle depleting from 8.8 to 3.2 milligrams per liter gives 336 milligrams per liter, which at a rate constant of 0.23 per day corresponds to an ultimate BOD of 492.

Putting it on the river

Now combine everything. Marley City's plant discharges 0.521 cubic meters per second of effluent with a dissolved oxygen of 2.0 milligrams per liter into a river carrying 8.0 cubic meters per second at 8.0 milligrams per liter. The mixed dissolved oxygen is a mass balance: 8.0 times 8.0 plus 0.521 times 2.0, all over 8.521, which is 64.0 plus 1.04 over 8.521, or 7.63 milligrams per liter.

At 20 degrees the saturation is 9.09, so the initial oxygen deficit is 9.09 minus 7.63, or 1.46 milligrams per liter. From there two processes compete. Bacteria consume oxygen as they degrade the mixed BOD, deepening the deficit. Reaeration across the water surface replenishes it at a rate proportional to the deficit itself. The classic Streeter-Phelps analysis, first published in 1925 for the Ohio River, combines these into the oxygen sag curve: dissolved oxygen falls to a minimum some distance downstream, at the critical point, then recovers.

Two design consequences follow immediately. First, the critical point is downstream, often many kilometers, so a sample taken at the outfall tells you almost nothing about the damage. Second, everything that raises the deficit or slows reaeration makes it worse: warm water, low flow, deep slow pools rather than riffles. That is the physical reason permits are written at the 7Q10 low flow and at summer temperatures, which is exactly what we did in Lesson 2.

Common misconceptions

• Alkalinity is the same as pH. pH is the current hydrogen ion activity; alkalinity is the reservoir of acid-neutralizing capacity. A water can have pH 8 with almost no alkalinity, and it will swing wildly when anything is added.
• Alkaline and basic mean the same thing here. In this field alkalinity is a measured capacity in milligrams per liter as calcium carbonate, not a statement that the water has pH above 7.
• Rainwater should have pH 7. Equilibrium with atmospheric carbon dioxide gives pH about 5.6. Acid rain is defined against that value, not against neutrality.
• BOD measures the amount of organic matter. It measures the oxygen that microorganisms will consume degrading it, which depends on biodegradability, seed viability, temperature, and time, not on organic mass alone.
• The five-day BOD is the total oxygen demand. At a typical rate constant it captures roughly two thirds of the carbonaceous demand and normally none of the nitrogenous demand.
• Cold water is worse for aquatic life. Cold water holds substantially more dissolved oxygen; the dangerous condition is warm, slow, low water in late summer.

Recap

• The carbonate system has dissociation constants near pH 6.35 and 10.33, so nearly every water an engineer meets is in the bicarbonate window between them.
• Alkalinity is acid-neutralizing capacity, measured by titration to pH 4.5 and reported as calcium carbonate; 150 milligrams per liter of bicarbonate is 123 as calcium carbonate.
• Each milligram per liter of alum consumes about 0.50 milligrams per liter of alkalinity, so a 200 milligram per liter dose on a 123 alkalinity water leaves only 23 and requires lime.
• Dissolved oxygen saturation falls from 14.6 milligrams per liter at 0 degrees to 7.6 at 30, scales with barometric pressure, and drops with salinity.
• Fish are comfortable above 5 milligrams per liter, stressed between 2 and 5, and water below 2 is hypoxic.
• A 5.00 milliliter sample in a 300 milliliter bottle going from 8.8 to 3.2 milligrams per liter gives BOD of 336; at k equal to 0.23 per day the ultimate BOD is 492.
• Nitrification adds 4.57 grams of oxygen demand per gram of nitrogen, which is why permits usually limit carbonaceous BOD measured with an inhibitor.
• Marley City's discharge dropped the river from 8.0 to 7.63 milligrams per liter, an initial deficit of 1.46, with the true minimum occurring kilometers downstream on the oxygen sag curve.

Sources

1. U.S. Geological Survey. (n.d.). Dissolved oxygen and water. U.S. Department of the Interior. usgs.gov
2. U.S. Environmental Protection Agency. (n.d.). Indicators: dissolved oxygen. epa.gov
3. Encyclopaedia Britannica. (n.d.). Biochemical oxygen demand. britannica.com
4. Wikipedia. (n.d.). Alkalinity. Wikimedia Foundation. en.wikipedia.org
5. Wikipedia. (n.d.). Oxygen sag curve. Wikimedia Foundation. en.wikipedia.org

The big picture

Two kinds of biological problem run through the rest of this course. The first is too much fertility: nitrogen and phosphorus entering a water body and feeding an explosion of algae whose eventual decay strips the oxygen out. The second is pathogens: the bacteria, viruses, protozoa, and worms that make water dangerous to drink or swim in. Both problems are biological, both are measured indirectly, and both drive expensive treatment decisions.

This lesson gives you the working understanding of each that a design engineer needs. We will run Marley City's nutrient loads, and one of the numbers we compute here, the oxygen required to nitrify its ammonia, will show up again in Module 4 as a line in an energy budget. We will also work out why the field measures harmless bacteria in order to detect dangerous ones.

Nitrogen: the shape-shifter

Nitrogen is difficult because it exists in many oxidation states and moves among them biologically. For an engineer, five forms matter.

Organic nitrogen is nitrogen bound in proteins and urea, the form in which most of it arrives at a sewage plant. Bacteria convert it to ammonia by ammonification. Ammonia in water exists in two forms in equilibrium, the ammonium ion and un-ionized ammonia, with a dissociation point near pH 9.25. That equilibrium matters enormously because the un-ionized form is the toxic one to fish. The same total ammonia concentration is far more dangerous at pH 8.5 than at pH 7.0, and warmer water shifts it further toward the toxic form, which is why the ammonia criterion you used to compute Marley City's permit limit is written as a function of pH and temperature.

Nitrifying bacteria oxidize ammonia to nitrite and then to nitrate, consuming oxygen and alkalinity. Under anoxic conditions, where nitrate is present but oxygen is not, other bacteria perform denitrification, reducing nitrate to nitrogen gas which bubbles away. That last step is the only route by which nitrogen actually leaves a water system, and every biological nutrient removal process in the world is built around arranging for it.

Two stoichiometric constants govern the engineering, and you should carry them.

• Nitrification consumes 4.57 grams of oxygen and 7.14 grams of alkalinity as calcium carbonate per gram of nitrogen oxidized.
• Denitrification returns about 2.86 grams of oxygen equivalent and 3.57 grams of alkalinity as calcium carbonate per gram of nitrogen reduced.

Worked example. Marley City's plant receives 45,000 cubic meters per day of wastewater containing 40 milligrams per liter of total Kjeldahl nitrogen, which is organic nitrogen plus ammonia. The load is 40 times 45,000 divided by 1,000, or 1,800 kilograms of nitrogen per day. Check that against people: 1,800 kilograms divided by 120,000 residents is 15 grams per person per day, which sits right in the expected range of 12 to 15.

Now the consequences. Fully nitrifying that load requires 4.57 times 1,800, or 8,226 kilograms of oxygen per day, on top of the oxygen needed for organic carbon. Nitrification roughly doubles a plant's aeration demand, and aeration is typically half of a treatment plant's entire electricity bill. That single number is why nitrification is treated as a major upgrade rather than an adjustment.

The alkalinity consequence is sharper still. Nitrifying 40 milligrams per liter of nitrogen consumes 7.14 times 40, or 286 milligrams per liter of alkalinity as calcium carbonate. Typical municipal wastewater carries only 200 to 250. So a plant that nitrifies completely can run out of alkalinity, the pH in the aeration basin falls below 6.5, and the nitrifiers, which are pH-sensitive, slow down and then stop. The plant fails at the very thing it was upgraded to do. The remedies are to add caustic or lime, or to place an anoxic denitrification zone ahead of the aerobic zone so that the returned 3.57 grams per gram recovers half the alkalinity for free. That second option is why nutrient removal plants look the way they do, and we will build one in Module 4.

Key idea: Nitrification consumes 4.57 grams of oxygen and 7.14 grams of alkalinity per gram of nitrogen, so Marley City's 1,800 kilograms of nitrogen per day costs 8,226 kilograms of oxygen and threatens to exhaust the wastewater's own buffering, which is precisely why denitrification zones exist.

Phosphorus: the one that does not fly away

Phosphorus is simpler and in some ways harder. It has no significant gaseous form, so unlike nitrogen it cannot be removed by converting it to a gas. It exists in water mainly as orthophosphate, plus polyphosphates from detergents and organic phosphorus in biomass. It sorbs strongly to soil particles and settles into sediment, where it can be released again when the overlying water goes anoxic, a mechanism called internal loading that keeps lakes eutrophic for decades after their external inputs are cut.

Municipal wastewater typically carries 5 to 8 milligrams per liter of total phosphorus. For Marley City at 6 milligrams per liter, the load is 6 times 45,000 divided by 1,000, or 270 kilograms of phosphorus per day, which is 2.25 grams per person per day. Removal is achieved either chemically, by precipitating with iron or aluminum salts, or biologically, by cycling bacteria through anaerobic and aerobic zones so that they take up more phosphorus than they need and are then wasted with the sludge.

Eutrophication

The mechanism is a four-step chain and it is worth stating carefully because people usually stop at step two.

1. Nutrients enter a water body from fertilizer runoff, wastewater effluent, animal operations, and atmospheric deposition.
2. Algae and cyanobacteria, previously nutrient-limited, multiply rapidly into a bloom.
3. The bloom dies, and bacteria decompose the enormous mass of dead algae.
4. That decomposition consumes dissolved oxygen, driving the deeper water hypoxic and killing fish and bottom life.

Step three is the one people skip, and it is the whole problem. The algae themselves produce oxygen while alive. It is their death and decay that creates the oxygen deficit, and it happens on the bottom, below the thermocline, where the water is already cut off from the atmosphere.

Which nutrient to control? The answer follows from the Redfield ratio, the roughly constant elemental composition of marine plankton, about 106 carbon to 16 nitrogen to 1 phosphorus by moles, which works out to about 7.2 parts nitrogen to 1 part phosphorus by mass. Compare a water body's supply ratio to that. If it receives nitrogen and phosphorus in a mass ratio well above 7.2, phosphorus runs out first and is limiting. As a broad generalization, fresh water is phosphorus-limited, because nitrogen-fixing cyanobacteria can pull nitrogen from the atmosphere while nothing can fix phosphorus, and marine and estuarine systems are nitrogen-limited. That generalization is why phosphate was removed from laundry detergent in the 1970s and why coastal nitrogen limits are the current regulatory frontier.

Two American cases anchor this. The Gulf of Mexico hypoxic zone forms every summer where the Mississippi River delivers nutrients, principally nitrogen from Midwestern agriculture, into stratified coastal water. Its measured area has ranged from a few thousand to over 20,000 square kilometers, with the record of about 22,700 square kilometers set in 2017, against an interagency management target of 5,000. It has not been met, because the nitrogen comes from nonpoint agricultural runoff across a drainage basin covering roughly 40 percent of the contiguous United States, and nonpoint sources are largely outside the permitting system you learned in Lesson 2.

The second case is Lake Erie in August 2014, when a cyanobacterial bloom over the shallow western basin produced enough microcystin, a liver toxin, that Toledo, Ohio, told roughly 400,000 people not to drink or boil their tap water for two days. Boiling makes it worse, because it concentrates the toxin. That event moved cyanotoxins from an ecological concern to a drinking water treatment problem, and utilities on affected lakes now monitor for them and maintain treatment, typically activated carbon or ozone, capable of removing them.

Key idea: Eutrophication harms water bodies not through the bloom itself but through the oxygen consumed decomposing it, and the limiting nutrient is usually phosphorus in fresh water and nitrogen in marine systems, which is what sets the control strategy.

Microorganisms and why we measure the wrong ones

Four classes of waterborne pathogen matter. Bacteria such as Salmonella, Shigella, Vibrio cholerae, and pathogenic strains of Escherichia coli are roughly 0.5 to 5 micrometers and are readily killed by chlorine. Viruses such as norovirus, hepatitis A, and rotavirus are 20 to 100 nanometers, small enough to pass through many filters, and generally chlorine-susceptible. Protozoa, principally Giardia and Cryptosporidium, form environmentally hardy cysts and oocysts 4 to 15 micrometers across; Cryptosporidium is highly resistant to chlorine, which is the single most important fact in modern drinking water design. Helminths, parasitic worms, matter mainly in wastewater reuse and biosolids land application, where their eggs are the design-limiting organism.

Now the measurement problem. You cannot test drinking water for pathogens directly at the frequency required. The organisms are present at low concentrations, they are individually difficult and slow to culture, there are dozens of them, and by the time a positive result came back the water would be consumed. So the field measures an indicator organism instead: a harmless organism that is abundant in feces, easy and cheap to detect, at least as persistent in water as the pathogens, and absent where fecal contamination is absent.

Indicator	What it indicates	Typical use
Total coliforms	General sanitary quality, not necessarily fecal	Drinking water distribution system monitoring and treatment integrity
Fecal coliforms	Coliforms growing at 44.5 degrees Celsius, more specific to warm-blooded animals	Older recreational and shellfish standards
Escherichia coli	Fecal contamination specifically	Freshwater recreational criteria and drinking water follow-up
Enterococci	Fecal contamination, persists better in salt water	Marine recreational criteria

Enumeration uses one of two methods. Membrane filtration passes a measured volume through a fine filter, incubates the filter on selective medium, and counts colonies, reported as colony forming units per 100 milliliters. The most probable number method distributes the sample among many small wells with a medium that changes color when the target organism grows, then infers a concentration statistically from how many wells turn positive. An MPN result is a statistical estimate with genuine confidence intervals, not a count, which is why two labs can honestly report different numbers on the same water.

The federal recreational water criteria set an Escherichia coli geometric mean of 126 per 100 milliliters for fresh water and an enterococci geometric mean of 35 per 100 milliliters for marine water, with a companion statistical threshold value not to be exceeded by more than ten percent of samples. Those are the numbers behind every beach closure sign you have ever seen.

Indicators have a real weakness that you should be able to state. They correlate poorly with viruses and with Cryptosporidium, precisely the pathogens that are hardest to remove. A water can be coliform-negative and still carry infectious oocysts. That is why drinking water regulation does not rely on indicators alone but requires treatment performance, expressed as log removal. Each log is a factor of ten: one log is 90 percent removal, two log is 99 percent, three log is 99.9 percent, four log is 99.99 percent. A conventional surface water plant must achieve at least 3-log removal or inactivation of Giardia, 4-log of viruses, and 2-log of Cryptosporidium, and Module 3 is about how the process train earns that credit stage by stage.

The reason for that architecture is Milwaukee in the spring of 1993. Cryptosporidium oocysts passed through a treatment plant whose filtration was performing poorly, into a distribution system carrying a normal chlorine residual, and an estimated 403,000 people became ill in what remains the largest documented waterborne disease outbreak in United States history. At least 69 people died, mostly people with compromised immune systems. Chlorine was working exactly as designed and was irrelevant, because Cryptosporidium tolerates it. The regulatory response was a series of rules tightening filtration performance, requiring turbidity to be monitored continuously as a proxy for particle removal, and giving credit for ultraviolet light, which inactivates Cryptosporidium readily.

Key idea: The field measures cheap harmless indicators because pathogens cannot be measured fast enough, but indicators miss viruses and Cryptosporidium, so drinking water regulation adds required log removal credits earned process by process.

Common misconceptions

• Algal blooms kill fish by poisoning them. Some cyanobacteria do produce toxins, but the ordinary mechanism is oxygen depletion during decomposition of the dead bloom.
• Controlling nitrogen is always the answer. In most fresh water phosphorus is limiting and nitrogen control alone accomplishes little, because cyanobacteria can fix atmospheric nitrogen. In marine systems the reverse usually holds.
• Boiling handles any bad tap water. Boiling kills pathogens but concentrates chemicals, including the cyanotoxins in the Toledo event and nitrate and metals generally.
• A coliform-negative sample means the water is safe. Indicators track bacterial fecal contamination reasonably and viruses and Cryptosporidium poorly, which is why treatment performance is regulated separately.
• Chlorine handles everything. Cryptosporidium is highly chlorine-resistant, and that fact, demonstrated in Milwaukee in 1993, reorganized modern drinking water treatment around filtration integrity and ultraviolet light.
• Cutting nutrient inputs fixes a eutrophic lake quickly. Phosphorus stored in sediment is released again when the bottom goes anoxic, so internal loading can sustain the problem for decades.

Recap

• Nitrogen moves through organic, ammonia, nitrite, and nitrate forms, and only denitrification to nitrogen gas removes it from water.
• Un-ionized ammonia is the toxic form, which is why ammonia criteria depend on pH and temperature.
• Nitrification costs 4.57 grams of oxygen and 7.14 grams of alkalinity per gram of nitrogen; Marley City's 1,800 kilograms per day of nitrogen means 8,226 kilograms per day of extra oxygen and 286 milligrams per liter of alkalinity demand.
• Phosphorus has no gas phase, sorbs to sediment, and returns through internal loading; Marley City discharges 270 kilograms of phosphorus per day.
• Eutrophication harms through decomposition of the dead bloom, not the bloom itself; phosphorus usually limits fresh water and nitrogen usually limits marine systems, following a Redfield mass ratio near 7.2 to 1.
• The Gulf of Mexico hypoxic zone reached about 22,700 square kilometers in 2017 against a 5,000 target, driven by nonpoint agricultural nitrogen.
• Toledo's 2014 microcystin event shut off drinking water for roughly 400,000 people and made cyanotoxins a treatment problem.
• Indicator organisms substitute for pathogens, with recreational criteria of 126 Escherichia coli per 100 milliliters in fresh water and 35 enterococci per 100 milliliters in marine water.
• Milwaukee 1993 sickened an estimated 403,000 people with chlorine-resistant Cryptosporidium and produced the modern log removal framework.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Nutrient pollution. epa.gov
2. National Oceanic and Atmospheric Administration. (n.d.). Gulf of Mexico hypoxia. noaa.gov
3. Centers for Disease Control and Prevention. (n.d.). Cryptosporidiosis (Crypto). U.S. Department of Health and Human Services. cdc.gov
4. U.S. Environmental Protection Agency. (n.d.). Recreational water quality criteria and methods. epa.gov
5. Wikipedia. (n.d.). 1993 Milwaukee cryptosporidiosis outbreak. Wikimedia Foundation. en.wikipedia.org

The big picture

A conventional surface water treatment plant is a sequence of five or six operations, each removing a different fraction of a different thing, arranged so that no single failure reaches the consumer. That arrangement has a name, the multiple barrier principle, and it is the organizing idea of drinking water engineering. Protect the source. Remove particles. Remove more particles. Kill what is left. Keep it clean on the way to the tap. Each barrier is imperfect. The system works because they are independent.

This lesson covers the front half of that train, from the river to the bottom of the sedimentation basin, and by the end of it you will have sized a real basin for Marley City. The two numbers you produce, a surface area and a detention time, are the numbers a designer actually hands to a structural engineer.

Where the water comes from

The single largest determinant of what a plant must do is the source. Groundwater has passed through tens of meters of soil, which is an extraordinarily effective filter, so it is usually clear and low in organic matter and pathogens; often it needs only disinfection, though it may be hard, or high in iron, manganese, arsenic, or nitrate. Surface water from a river, lake, or reservoir is exposed to everything in the watershed and essentially always needs the full train. Reservoirs add their own complications, including seasonal stratification that pulls iron and manganese out of anoxic bottom sediment and algal blooms that create taste and odor problems.

Before spending money on treatment, spend it on the watershed. The classic demonstration is New York City, which draws most of its roughly 4 million cubic meters per day from the Catskill and Delaware watersheds and delivers it unfiltered, one of the few large systems in the country permitted to do so. It holds that filtration avoidance determination by running an aggressive watershed protection program: land acquisition, septic system upgrades, agricultural best management practices, and stream restoration across a basin of roughly 4,000 square kilometers. The program has cost well over a billion dollars over three decades, which sounds enormous until you compare it with the filtration plant it replaced, estimated in the billions and with an operating cost forever. The city did build a filtration plant for its smaller Croton supply, which came online in 2015 at a cost of roughly 3.5 billion dollars. The comparison is the single best argument in the field for treating source protection as an engineering alternative rather than as environmental sentiment.

The master operational parameter for surface water is turbidity, the scattering of light by suspended particles, measured in nephelometric turbidity units. It is not itself a health parameter. It matters because particles shield microorganisms from disinfectants, because turbidity correlates with the particles that carry pathogens, and above all because it can be measured continuously and instantly, which nothing biological can. Regulation therefore uses it as a surrogate for filtration performance: combined filter effluent turbidity must be at or below 0.3 units in at least 95 percent of measurements each month and must never exceed 1 unit. Good plants run continuously below 0.1.

Key idea: The multiple barrier principle stacks independent imperfect barriers, source protection is the cheapest of them, and turbidity is the surrogate that lets particle removal be verified continuously when pathogens cannot be.

Why particles will not settle

Now the central problem. The particles that make surface water cloudy are mostly colloids: clay, silt, humic organic matter, algae, and bacteria, typically 0.01 to 10 micrometers across. They do not settle, and the reason is not only that they are small. They carry a net negative surface charge, they attract a shell of counter-ions that moves with them, and any two of them approaching each other feel electrostatic repulsion long before they touch. Brownian motion keeps them stirred. A colloidal suspension can sit for months without clearing.

So calculate what settling alone can do. Stokes law gives the terminal velocity of a small sphere as gravity times the density difference times the diameter squared, all divided by eighteen times the viscosity. Take a 20 micrometer floc particle with a density of 1,050 kilograms per cubic meter in water at 998 kilograms per cubic meter and a viscosity of 1.002 times ten to the minus three pascal seconds. The numerator is 9.81 times 52 times the square of 20 times ten to the minus six, which is 9.81 times 52 times 4 times ten to the minus ten, or 2.04 times ten to the minus seven. The denominator is 18 times 1.002 times ten to the minus three, or 0.018. The velocity is 1.13 times ten to the minus five meters per second, which over a day is 0.98 meters. Just under one meter per day.

Hold that number. We are about to design a basin that removes particles settling at 30 meters per day. A 20 micrometer particle is thirty times too slow. The entire purpose of coagulation and flocculation is to fix that, and now you know exactly by how much.

Coagulation

Coagulation is the chemical destabilization of the suspension. Add a salt of a trivalent metal, most often aluminum sulfate, which everyone calls alum, or ferric chloride, or a prehydrolyzed polyaluminum chloride. Two things happen at once. The trivalent cations compress the electrical double layer and neutralize surface charge, so the particles stop repelling each other. And the metal hydrolyzes to a voluminous hydroxide precipitate that physically enmeshes particles as it forms, a mechanism the field calls sweep floc, which is what actually dominates at ordinary doses.

Coagulant must be dispersed almost instantly, because the hydrolysis reactions occur in fractions of a second. That is the job of the rapid mix: violent agitation for 30 to 60 seconds. Mixing intensity is quantified by the velocity gradient G, in inverse seconds, defined as the square root of power input divided by viscosity times volume. Rapid mix runs at G of roughly 700 to 1,000 per second.

Worked example. Marley City treats 60,000 cubic meters per day, which is 41.7 cubic meters per minute. A 60-second rapid mix needs a chamber of about 41.7 cubic meters. At G equal to 800 per second, the required power is G squared times viscosity times volume: 800 squared is 640,000, times 1.002 times ten to the minus three is 641, times 41.7 cubic meters gives about 26,700 watts, or a 27 kilowatt mixer. That is a real motor on a real drawing.

The dose itself is not calculated from theory. It is determined by a jar test: six beakers of the actual raw water, six different coagulant doses, identical rapid mix and slow mix and settling in a gang stirrer, then measure the settled turbidity in each. The optimum is read off the resulting curve, and it changes with season, temperature, and storm events. Every conventional plant runs jar tests routinely, and an operator who trusts a fixed dose year-round is an operator waiting for a bad week. For Marley City the current jar test gives 30 milligrams per liter of alum, which from Lesson 5 consumes 15 milligrams per liter of alkalinity, leaving 108 of the original 123.

One modern wrinkle. Coagulation also removes dissolved natural organic matter, and that matter is the precursor from which chlorine forms disinfection byproducts. Regulation therefore requires enhanced coagulation, meaning a dose and pH chosen to hit a specified percentage removal of total organic carbon, at plants where byproduct formation is a concern. It usually means more coagulant and a lower pH than turbidity removal alone would require, and it is a good example of a rule reaching back up the train to change a chemical dose.

Flocculation

Coagulation destabilizes; it does not aggregate. Flocculation is the gentle, prolonged mixing that lets destabilized particles collide often enough to grow. Two mechanisms drive the collisions: Brownian motion for the smallest particles, and induced velocity gradients for everything above about a micrometer. The design variables are the same G, now at 20 to 70 per second, and a detention time of 20 to 45 minutes, with the dimensionless product G times t typically between 30,000 and 150,000.

Good practice tapers the mixing. The first flocculation stage runs at higher G to promote collisions when particles are small and strong; later stages run lower so the large fragile floc is not torn apart. Getting this wrong in either direction is a classic operational failure: too little energy and nothing grows, too much and you shear the floc back into fragments that pass straight through the plant.

Worked example. Marley City uses three tapered stages at G of 60, 40, and 20 per second, 10 minutes each, for a total of 30 minutes. Total volume is 60,000 cubic meters per day times 30 minutes divided by 1,440 minutes per day, or 1,250 cubic meters. Using an average G of 40 per second and t of 1,800 seconds gives G times t equal to 72,000, comfortably inside the design range.

What size must the floc reach? Go back to Stokes and solve for the diameter that settles at the design overflow rate of 30 meters per day, which is 3.47 times ten to the minus four meters per second. Rearranged, diameter squared equals eighteen times viscosity times velocity, divided by gravity times the density difference: 0.018 times 3.47 times ten to the minus four gives 6.26 times ten to the minus six, divided by 510 gives 1.23 times ten to the minus eight, so the diameter is 1.11 times ten to the minus four meters, or about 110 micrometers. Flocculation must grow a 20 micrometer colloid into a 110 micrometer floc. That is a factor of five in diameter and about 170 in volume, and it explains why the flocculation basin is the largest quiet structure in the front half of the plant.

Key idea: Coagulation neutralizes charge in seconds at high mixing intensity while flocculation grows floc over half an hour at low intensity, and the target is a particle around 110 micrometers, because that is what settles at the 30 meters per day the basin is designed for.

Sedimentation and overflow rate

Now the basin, and one governing idea that surprises people. In ideal settling theory, whether a particle is removed depends only on whether its settling velocity exceeds the basin's overflow rate, defined as flow divided by surface area, in meters per day. Depth does not appear. The reason is that a deeper basin gives a particle more time to fall but also a longer distance to fall, and the two effects cancel exactly.

That result, which comes from the classical Hazen analysis, has a real design consequence: for settling, area is what you buy. Depth exists for other reasons, namely sludge storage, protecting the settled blanket from wind and density currents, and providing hydraulic stability. Typical basin depths are 3 to 5 meters and typical overflow rates for coagulated surface water are 20 to 40 meters per day.

Worked example, the one to remember. Marley City, 60,000 cubic meters per day, design overflow rate 30 meters per day.

• Required surface area: 60,000 divided by 30 equals 2,000 square meters.
• Use four rectangular basins, so 500 square meters each. At a length-to-width ratio of about 3 to 1, take 12.5 meters wide by 40 meters long.
• At a side water depth of 4.0 meters, each basin holds 2,000 cubic meters, so 8,000 cubic meters total.
• Detention time: 8,000 divided by 60,000 equals 0.133 days, or 3.2 hours, which is squarely in the normal 2 to 4 hour range.
• Check the horizontal velocity: 60,000 cubic meters per day divided by four basins is 15,000 per basin, divided by the cross-section of 12.5 times 4.0 equals 50 square meters, giving 300 meters per day, or 0.21 meters per minute. Slow enough not to scour the settled sludge.

Why four basins rather than one of 2,000 square meters? Redundancy. A plant must be able to take a basin out of service for cleaning or repair while still meeting demand. With one basin out, the remaining three carry 60,000 over 1,500 square meters, an overflow rate of 40 meters per day, which is at the top of the acceptable range but workable. That kind of check, what happens when one unit is down, separates a design from a calculation.

Where land is expensive, plate or tube settlers are installed in the basin: closely spaced inclined surfaces that multiply the effective settling area several times within the same footprint, because each plate is a short settling zone in its own right. Where floc is light, as with algae-laden or highly colored water, dissolved air flotation reverses the process, saturating water with air under pressure and releasing it as microbubbles that attach to floc and carry it upward to be skimmed.

The sludge nobody mentions

Apply the mass balance from Lesson 1. Everything removed in the basin is still on site. Alum sludge production has two components: the aluminum hydroxide formed from the coagulant, and the turbidity solids captured. From the stoichiometry, 594 grams of hydrated alum yields 156 grams of aluminum hydroxide, so a 30 milligram per liter dose produces about 7.9 milligrams per liter of hydroxide. If the raw water's turbidity corresponds to roughly 20 milligrams per liter of suspended solids, total residuals are about 28 milligrams per liter, which at 60,000 cubic meters per day is 1,680 kilograms of dry solids per day.

That gelatinous alum sludge is difficult to dewater, and its handling, by lagoons, mechanical dewatering, or discharge to the sanitary sewer, is a real cost and a real permit condition. Water treatment plants are point sources too.

Common misconceptions

• Deeper sedimentation basins settle better. In ideal settling theory removal depends only on overflow rate, which is flow over surface area. Depth serves sludge storage and hydraulic stability.
• Turbidity is a health standard. It is a surrogate for particle removal and filtration integrity that can be measured continuously, which is why regulation leans on it so hard.
• Coagulation makes particles settle. Coagulation destabilizes them chemically. Flocculation is what grows them to a settleable size, and the two require opposite mixing intensities.
• More mixing is better in flocculation. Excess velocity gradient shears grown floc back into fragments, and the design deliberately tapers the energy downward through the basin.
• A coagulant dose can be computed from raw water chemistry. It is determined empirically by jar testing on the actual water and changes with season, temperature, and storms.
• Water treatment plants do not generate waste. Marley City's front end produces about 1,680 kilograms of dry alum sludge per day, which is a permitted residual stream.

Recap

• Multiple independent barriers, starting with source protection, define drinking water practice, and New York City's unfiltered Catskill and Delaware supply shows watershed protection competing directly with capital treatment.
• Filtered water turbidity must be at or below 0.3 units in 95 percent of samples and never above 1 unit.
• A 20 micrometer floc particle settles at only about 0.98 meters per day by Stokes law, far too slow for a basin designed at 30.
• Coagulation neutralizes colloid charge and forms sweep floc, needs 30 to 60 seconds at a velocity gradient near 800 per second, and a 60,000 cubic meter per day plant needs about a 27 kilowatt rapid mixer.
• Doses come from jar tests, not from theory, and enhanced coagulation targets organic carbon removal to limit disinfection byproducts.
• Flocculation runs 20 to 45 minutes at tapered velocity gradients with G times t between 30,000 and 150,000; Marley City's three stages give 1,250 cubic meters and G times t of 72,000.
• The floc must reach about 110 micrometers to settle at 30 meters per day.
• Marley City's basins need 2,000 square meters of surface, built as four units 12.5 by 40 meters at 4 meters deep, giving a 3.2 hour detention time and still workable at 40 meters per day with one basin out of service.
• The front end produces roughly 1,680 kilograms per day of alum sludge that must itself be managed.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Drinking water treatment plant residuals and water treatment processes. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Surface Water Treatment Rules. epa.gov
3. New York City Department of Environmental Protection. (n.d.). Watershed protection. nyc.gov
4. U.S. Geological Survey. (n.d.). Turbidity and water. U.S. Department of the Interior. usgs.gov
5. Wikipedia. (n.d.). Flocculation. Wikimedia Foundation. en.wikipedia.org

The big picture

The back half of a drinking water plant does two things the front half cannot. Filtration removes the particles that escaped sedimentation, taking turbidity from a few units down to a tenth of one. Disinfection then inactivates whatever microorganisms remain, and unlike everything before it, disinfection is not a removal process at all. Nothing leaves the water. Organisms are chemically damaged so they cannot infect.

This lesson works both, and its centerpiece is a calculation every water engineer must be able to do from memory: the CT check. It asks whether a plant is actually achieving the disinfection credit it is claiming, and the honest answer for Marley City is going to depend on the season.

Filtration

Water leaving sedimentation still carries 1 to 5 turbidity units, and it must reach 0.3 or below. Filtration does that, and it is worth understanding that it is not a strainer. In a granular medium bed, the pores between sand grains are far larger than the particles being captured. Removal happens by depth filtration: particles are carried by the flow into contact with grain surfaces by interception, by settling onto the top of grains, and for the smallest particles by Brownian diffusion, and they then attach because coagulation earlier in the train destabilized them. That last clause is the important one. A filter fed uncoagulated water fails, no matter how fine its sand, because the particles bounce off instead of sticking.

The common configurations are worth distinguishing. Rapid sand filters run at 5 to 15 meters per hour through roughly 0.6 to 0.8 meters of graded silica sand. Dual media filters put a coarser, lighter anthracite layer above the sand, which lets larger particles penetrate deeper before being caught and roughly doubles the run length between backwashes; they now dominate new construction. Slow sand filters run a hundred times slower, at 0.1 to 0.4 meters per hour, and work partly biologically through a living surface layer; they need enormous land area but almost no chemicals and remain excellent for small, clean, cold-water supplies.

Worked example. Marley City, 60,000 cubic meters per day, dual media, design loading rate 10 meters per hour, which is 240 meters per day.

• Required filter area: 60,000 divided by 240 equals 250 square meters.
• Provide eight filters, so 31.25 square meters each, say 5 by 6.25 meters.
• Now the check that matters. One filter is always in backwash, and codes require the plant to meet demand with the largest unit out of service. With seven filters online, area is 218.75 square meters, so the loading rate rises to 60,000 divided by 218.75, or 274 meters per day, which is 11.4 meters per hour. Still within the 5 to 15 range, so the design holds.

As a filter loads with captured solids, head loss climbs. When it reaches the available head, or when effluent turbidity begins to break through, the filter is backwashed: flow is reversed at 35 to 50 meters per hour, sometimes with air scour, expanding the bed and carrying accumulated solids to waste. Backwashing costs water.

Worked example, continued. Backwash one 31.25 square meter filter at 37 meters per hour for 10 minutes. That is 37 divided by 60 times 10, or 6.17 meters of water depth, times 31.25 square meters, or about 193 cubic meters per backwash. Backwashing all eight filters once a day uses 1,542 cubic meters, which is 2.6 percent of the plant's 60,000 cubic meter production. Two to four percent is normal, and it is recovered at many plants by settling the backwash water and returning the supernatant to the head of the plant, which creates its own risk of recycling concentrated Cryptosporidium and is regulated for that reason.

One more operational fact with a health consequence. Immediately after backwash a filter passes a turbidity spike for several minutes while the bed re-establishes, called ripening. Since Milwaukee, the standard response is filter-to-waste: divert the first several minutes of production to drain rather than to the clearwell. A plant without filter-to-waste sends its dirtiest water of the day straight to the city several times daily.

Beyond granular media, membranes are now routine. Microfiltration and ultrafiltration, with pores of roughly 0.1 and 0.01 micrometers, are absolute barriers to protozoa and to most bacteria and earn high log removal credits by direct integrity testing rather than by turbidity surrogate. Nanofiltration and reverse osmosis go further, removing dissolved organics, hardness, and salts, at a substantial energy cost. And granular activated carbon, used as a filter medium or as a separate contactor, adsorbs the taste and odor compounds, pesticides, and organic precursors that nothing else removes.

Key idea: Filtration is depth capture on grain surfaces, not straining, and it only works on coagulated water; Marley City needs 250 square meters at 10 meters per hour, still runs at an acceptable 11.4 with one filter down, and spends about 2.6 percent of production on backwash.

Chlorine chemistry

Chlorine gas dissolved in water hydrolyzes to hypochlorous acid and hydrochloric acid, and hypochlorous acid then dissociates into hydrogen ion and hypochlorite ion. The dissociation point is pH 7.54 at 25 degrees Celsius, and here is why that number matters: hypochlorous acid is roughly 80 times more effective a disinfectant than hypochlorite ion, because the neutral molecule crosses cell membranes and the charged ion does not.

The consequence is direct and often surprising to beginners. At pH 6.5 about 90 percent of free chlorine is the potent hypochlorous form. At pH 8.5 about 90 percent is the weak hypochlorite form. The same measured residual, on the same water, is far less effective in alkaline water. Every CT table in the regulations is indexed by pH for exactly this reason, and a plant that raises pH for corrosion control has just made its disinfection harder.

Now breakpoint chlorination, which explains the strange shape of a chlorine demand curve. Add chlorine to water containing ammonia, and the first additions do not produce free chlorine at all; they produce chloramines, and the measured residual rises. Keep adding and the chloramines are oxidized and destroyed, and the residual falls even as you add more chlorine. Past that minimum, the breakpoint, further additions finally appear as free chlorine residual. The practical rule is that roughly 7.6 milligrams of chlorine is needed per milligram of ammonia nitrogen to reach breakpoint, and a plant that stops halfway is running on chloramines without meaning to.

The CT calculation

Disinfection performance depends on both how much disinfectant is present and how long the organism is exposed to it. The regulatory measure is CT, the residual concentration in milligrams per liter multiplied by the contact time in minutes, with units of milligram-minutes per liter. EPA publishes tables of required CT for each organism class, disinfectant, temperature, and pH, and a plant demonstrates compliance by computing its achieved CT and dividing by the required value.

The subtlety is the T. You cannot use the theoretical detention time, because real basins short-circuit: some water races through in a fraction of the nominal time. The regulations require T10, the time by which 10 percent of the water has passed through, determined by tracer study or estimated using a baffling factor applied to the theoretical detention time.

Baffling condition	Factor	Description
Unbaffled	0.1	No internal baffles, low length-to-width ratio, high inlet velocity
Poor	0.3	Single or no baffles, unbaffled inlet and outlet
Average	0.5	Baffled inlet or outlet, some intra-basin baffles
Superior	0.7	Perforated inlet baffle, serpentine interior, perforated outlet
Perfect plug flow	1.0	Pipeline flow, achieved only in a long transmission main

Worked example, the one to remember. Marley City has a 4,000 cubic meter clearwell with average baffling, treating 60,000 cubic meters per day. It maintains a free chlorine residual of 1.0 milligram per liter at pH 7 and the water is 10 degrees Celsius.

• Flow per minute: 60,000 divided by 1,440 equals 41.7 cubic meters per minute.
• Theoretical detention time: 4,000 divided by 41.7 equals 96 minutes.
• T10 with a baffling factor of 0.5: 96 times 0.5 equals 48 minutes.
• Achieved CT: 1.0 times 48 equals 48 milligram-minutes per liter.

Now, is 48 enough? That depends on what credit the plant needs from disinfection, and this is where the train logic pays off. The Surface Water Treatment Rule requires 3-log removal or inactivation of Giardia overall. A properly operated conventional treatment train, meaning coagulation, sedimentation, and filtration meeting turbidity limits, is credited with 2.5 logs. Disinfection therefore has to supply only the remaining 0.5 log.

The table value for 3-log Giardia inactivation by free chlorine at 10 degrees Celsius, pH 7, at a 1.0 milligram per liter residual is 87 milligram-minutes per liter. Scale it: 0.5 log out of 3 requires 87 times 0.5 divided by 3, or 14.5. The plant achieves 48. The ratio is 3.3, so it passes comfortably. Viruses are easier still; 4-log virus inactivation with free chlorine at 10 degrees needs a CT of only about 6.

Now change one thing. In February the source water falls to 0.5 degrees Celsius. Disinfection kinetics slow sharply with temperature, and the table value for 3-log Giardia rises from 87 to 165. The 0.5-log requirement becomes 165 times 0.5 divided by 3, or 27.5. The plant still achieves 48, so it still passes, but the safety ratio has fallen from 3.3 to 1.75. Winter is when CT compliance problems appear, and it is why plants raise their residual seasonally.

Now change a second thing, and watch the whole argument collapse. Suppose a filter breakthrough puts combined effluent turbidity above 1 unit and the plant loses its filtration credit for that period. Now disinfection alone must supply the full 3 logs: 165 milligram-minutes per liter at winter temperature, against 48 achieved. The plant is short by a factor of 3.4. Could the operator chlorinate out of it? Tripling the residual to 3.0 milligrams per liter gives a CT of 144, and the table value itself rises somewhat at higher residuals, so the answer is no, not reliably, and the taste and byproduct consequences of a 3 milligram per liter residual are severe. That is the lesson: you cannot chlorinate your way out of a filtration failure. The barriers are not interchangeable.

Key idea: CT is residual times T10, not theoretical detention time, and Marley City's 4,000 cubic meter clearwell at average baffling gives only 48 minutes of credited contact out of 96 nominal, which passes a 0.5-log Giardia requirement comfortably in summer, narrowly in winter, and not at all if filtration credit is lost.

The alternatives and their tradeoffs

Chloramine, formed deliberately by adding ammonia along with chlorine, is a weaker disinfectant but a far more persistent one, and it forms much less trihalomethane. Many large systems use free chlorine at the plant for primary inactivation and then convert to chloramine for the distribution system. The costs are real: chloramine supports nitrifying bacteria inside the distribution system, which consume the residual and depress pH, and, as Lesson 9 will show in detail, a switch to chloramine changed the chemistry of lead pipe scale in Washington, DC, with serious consequences.

Ultraviolet light inactivates organisms by damaging their DNA, and its dose is measured in millijoules per square centimeter. It is astonishingly good against protozoa: about 12 millijoules per square centimeter achieves 3-log Cryptosporidium inactivation, a target chlorine essentially cannot reach at any practical dose. It is poor against some viruses, with 4-log adenovirus inactivation requiring on the order of 186. It adds no taste, forms no byproducts, and leaves no residual whatsoever, so it never replaces a chemical disinfectant, it supplements one. The typical validated design dose is 40 millijoules per square centimeter.

Ozone is the most powerful of the common oxidants, excellent against protozoa, and it also destroys taste and odor compounds. It must be generated on site, it is energy-intensive, it leaves no residual, and where the source water contains bromide it forms bromate, a regulated carcinogen with a maximum contaminant level of 10 micrograms per liter.

Disinfection byproducts

Chlorine reacts with natural organic matter to form hundreds of compounds, of which two families are regulated: trihalomethanes, with a total maximum contaminant level of 0.080 milligrams per liter, and five haloacetic acids, at 0.060. Both are associated in epidemiological studies with modestly elevated bladder cancer risk over decades of exposure, and both are computed as a locational running annual average, which means a single bad sample at one point in the distribution system can put a utility out of compliance for a year.

Here is the tradeoff stated plainly, because it is the honest center of the subject. Too little disinfection produces acute microbial illness that kills people in days. Too much produces byproducts that raise a chronic cancer risk over decades. The two risks are not remotely comparable in magnitude: the microbial risk is far larger and far faster. The regulatory answer has therefore never been to disinfect less. It has been to remove the organic precursors first through enhanced coagulation or activated carbon, to move the point of chlorination later in the train, and to reduce water age in the distribution system so that byproducts have less time to form.

Common misconceptions

• Filters strain particles out. Pores are far larger than the particles captured; removal is attachment onto grain surfaces within the bed depth, and it fails entirely if the water was not coagulated.
• CT uses the basin's detention time. It uses T10, the time at which 10 percent of the water has passed, which for average baffling is half the theoretical value.
• A higher chlorine residual can compensate for any failure. Not for a filtration failure. In the worked winter case the plant would need more than three times its CT, which no practical residual delivers.
• Ultraviolet light disinfects the distribution system. It leaves no residual at all, so it is always paired with a chemical disinfectant for downstream protection.
• Chloramine is simply a safer chlorine. It reduces trihalomethanes and lasts longer, but it supports distribution system nitrification and can destabilize lead pipe scale.
• Byproduct concerns mean we should chlorinate less. The microbial risk is larger and faster. The engineering answer is to remove precursors and shorten water age, not to weaken the barrier.

Recap

• Filtration is depth capture requiring prior coagulation; rapid sand and dual media run at 5 to 15 meters per hour, slow sand at 0.1 to 0.4.
• Marley City needs 250 square meters of filter at 10 meters per hour, built as eight units, running at an acceptable 11.4 meters per hour with one out of service, and spending about 2.6 percent of production on backwash.
• Filter-to-waste handles the post-backwash ripening turbidity spike.
• Hypochlorous acid is about 80 times more effective than hypochlorite ion, with the dissociation point at pH 7.54, which is why CT tables are indexed by pH.
• Breakpoint chlorination requires roughly 7.6 milligrams of chlorine per milligram of ammonia nitrogen before a free residual appears.
• CT equals residual times T10; Marley City's 4,000 cubic meter clearwell at a 0.5 baffling factor gives 48 minutes and a CT of 48.
• Conventional treatment earns 2.5 of the required 3 logs of Giardia credit, so the required CT is 14.5 at 10 degrees and 27.5 at 0.5 degrees; the same plant fails outright if filtration credit is lost.
• Ultraviolet light achieves 3-log Cryptosporidium at about 12 millijoules per square centimeter but leaves no residual; ozone is strongest but forms bromate where bromide is present.
• Trihalomethanes are limited to 0.080 milligrams per liter and haloacetic acids to 0.060, both as locational running annual averages, and the answer is precursor removal rather than weaker disinfection.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Surface Water Treatment Rules. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Stage 1 and Stage 2 Disinfectants and Disinfection Byproducts Rules. epa.gov
3. Centers for Disease Control and Prevention. (n.d.). Water disinfection with chlorine and chloramine. U.S. Department of Health and Human Services. cdc.gov
4. World Health Organization. (n.d.). Guidelines for drinking-water quality. who.int
5. Wikipedia. (n.d.). Water chlorination. Wikimedia Foundation. en.wikipedia.org

The big picture

Everything in the last two lessons happened inside a fence. The water left that fence clean, and then it entered a buried network of pipe that in an American city of Marley City's size runs to roughly 850 kilometers, was installed over a period of 130 years, is made of at least six different materials, and cannot be inspected. Nationally there are more than 3.4 million kilometers of drinking water main, and the American Water Works Association has estimated something on the order of 250,000 to 300,000 main breaks a year, which is roughly one every two minutes.

The distribution system is where the barrier concept gets its hardest test, and it is where the most serious American drinking water failures of the last twenty-five years have occurred. Not one of them was a treatment plant failure. All of them were chemistry, oversight, and institutional failures in the pipes. This lesson is about that.

What a distribution system has to do

Four requirements, and they pull against each other.

Pressure. Normal service pressure runs roughly 275 to 550 kilopascals, which is 40 to 80 pounds per square inch. The binding constraint is the regulatory minimum of 138 kilopascals, or 20 pounds per square inch, which must be maintained at every point even during a fire flow. Below that, the system can draw contaminated groundwater in through joints and cracks, and a loss of pressure triggers a boil water notice for that reason.

Fire flow. This is the requirement that shapes everything. Domestic demand for a single-family neighborhood might be a few liters per second, while a fire flow requirement is commonly 60 to 250 liters per second for two to four hours. Mains are therefore sized for fire, not for drinking, and that oversizing is the direct cause of the water age problem below.

Storage. Elevated tanks and standpipes provide three volumes stacked together: equalization storage to cover the daily peak so the plant can run at a steady average rate, fire storage, and emergency storage for a plant outage. Elevated storage also sets system pressure by gravity, which is why a tank's overflow elevation, not a pump, defines the pressure in the district below it.

Water quality. The residual must survive to the last tap, and the water must not become a problem on the way.

Worked example. A dead-end branch main is 400 meters of 400 millimeter pipe, sized that way for a future hydrant, serving 20 homes at 2.6 people each and 500 liters per person per day. Demand is 20 times 2.6 times 0.5 cubic meters, or 26 cubic meters per day. Pipe volume is pi over four times 0.40 squared times 400, which is 0.1257 square meters times 400, or 50.3 cubic meters. Water age in that branch is 50.3 divided by 26, or 1.93 days, and that is before the water reaches the branch. Had the same street been served by a 200 millimeter main, the volume would be 12.6 cubic meters and the age 0.48 days, a quarter as long.

Why does age matter? Four things happen to old water. The disinfectant residual decays, and a system that leaves the plant at 1.0 milligrams per liter may arrive at 0.1 or at zero. Disinfection byproducts keep forming as long as chlorine and organic matter are in contact, so the trihalomethane maximum in a system is almost always at its far end, which is exactly why the Stage 2 rule switched to a locational running annual average. In chloraminated systems, nitrifying bacteria grow in the low-residual zones, consume ammonia, depress pH, and accelerate residual loss in a self-reinforcing cycle. And taste and odor complaints concentrate in the same places. The design responses are to loop mains rather than dead-end them, to install automatic flushing devices at unavoidable dead ends, to turn storage tanks over properly rather than letting a tank float on the system for weeks, and sometimes to install booster chlorination out in the network.

Two more issues deserve naming. Cross connections are physical links between the potable system and anything else, and a pressure reversal can pull that content backward into the main; backflow preventers and their annual testing are a genuine public health program. And non-revenue water, the gap between what a utility produces and what it bills, runs 10 to 30 percent. For Marley City at 15 percent that is 9,000 cubic meters per day, about 1.15 million dollars a year at a production cost near 0.35 dollars per cubic meter, leaking into the ground.

Key idea: Distribution mains are sized for fire flow rather than for drinking demand, and the resulting oversizing produces water age, which decays residual, grows byproducts, and feeds nitrification, so looping, flushing, and tank turnover are water quality measures rather than housekeeping.

Corrosion and the protective scale

Corrosion is an electrochemical process: metal at an anodic site gives up electrons and goes into solution, electrons flow through the metal to a cathodic site, and something in the water accepts them. It costs utilities enormous sums in pipe replacement, and it puts metal into the water people drink.

The thing that stops it is not the absence of corrosive water. It is a protective scale: a thin, adherent layer of mineral deposit on the pipe interior, built up over years, that physically separates the metal from the water. On a lead service line that scale may be lead carbonate, or, where free chlorine has been used for a long time, a lead dioxide layer in which lead sits in an oxidized and very insoluble form. Either way, the pipe is not really in contact with the water. The scale is.

Understand that and the central lesson of this lesson follows immediately: the danger is not corrosive water, it is changing water. A stable scale in equilibrium with a stable water chemistry releases almost nothing. Change the chemistry and the scale can dissolve or spall, and lead that has sat safely on a pipe wall for fifty years enters the water in weeks.

The levers an operator actually has are few.

• pH and alkalinity. Raising pH into roughly the 7.5 to 8.5 range and maintaining adequate alkalinity favors formation of insoluble carbonate scale. This is the cheapest control and often the only one a small system uses.
• Orthophosphate. Dosing 1 to 3 milligrams per liter as phosphate forms a lead phosphate scale that is extremely insoluble across a wide pH range. It is the workhorse of American corrosion control, it costs very little, and it is the single intervention whose absence defines the Flint failure.
• Indices. The Langelier Saturation Index compares actual pH to the pH at which calcium carbonate would be saturated, indicating whether water tends to deposit or dissolve carbonate scale. It is a useful screening indicator and a poor predictor of lead release specifically, and treating it as a lead control criterion is a known error.

Where lead comes from

Almost never from the source water, and almost never from the treatment plant. Lead is picked up between the water main and the glass, from three places. Lead service lines, the pipe connecting the main to the building, installed widely until the mid-twentieth century; EPA's most recent national inventory estimates roughly 9 million of them still in service. Lead solder used on copper plumbing, permitted until the 1986 Safe Drinking Water Act amendments banned it. And brass fixtures and meters, which could legally contain up to 8 percent lead until 2014, when the definition of lead-free was tightened to a weighted average of 0.25 percent on wetted surfaces.

Worked example. A 15 meter lead service line of 19 millimeter internal diameter holds pi over four times 0.019 squared times 15, which is 2.84 times ten to the minus four square meters times 15, or 4.25 liters. That is the water that sat in direct contact with lead pipe overnight, and it is the first 4.25 liters out of the tap in the morning.

Now look at the regulatory sample. The Lead and Copper Rule of 1991 requires a first-draw one liter sample after at least six hours of stagnation. In a house whose interior plumbing is copper, that first liter is interior plumbing water; the lead service line water is in liters two through five. The rule's own sample can therefore miss the service line entirely, which is one reason the 2021 revisions added a fifth-liter sample at sites known to have lead service lines. The regulatory number and the exposure are not the same quantity, and an engineer should never confuse them.

The rule works as a treatment technique rather than a maximum contaminant level, because lead concentration depends on the building, not on the utility. Utilities sample a set of high-risk homes, and if the 90th percentile of results exceeds the action level, the utility must optimize corrosion control, conduct public education, and begin replacing lead service lines. The action level was 15 micrograms per liter from 1991 onward. The Lead and Copper Rule Improvements, finalized in October 2024, lower it to 10 micrograms per liter and require water systems to replace lead service lines, generally within ten years. Copper has an action level of 1.3 milligrams per liter.

Key idea: Lead enters water from the service line, solder, and brass rather than from the source, protective scale rather than non-corrosive water is what holds it in place, and orthophosphate plus pH control is how that scale is maintained.

Washington, DC, 2001 to 2004

The District of Columbia's water supplier switched its residual disinfectant from free chlorine to chloramine in November 2000, primarily to reduce disinfection byproducts. Chloramine is a weaker oxidant, and the change lowered the oxidation-reduction potential of the water in the pipes. The lead dioxide scale that had built up under decades of free chlorine was stable only under those oxidizing conditions. Under the new chemistry it began converting to more soluble lead in the plus-two oxidation state, and lead levels in tap water rose sharply, with many homes far above the action level and some at extraordinary concentrations.

The episode took years to be recognized and reported publicly. The eventual fix was to add orthophosphate, which built a new and stable lead phosphate scale and brought levels down. The engineering lesson is precise and transferable: a change made to solve one water quality problem altered the chemistry governing another, and nobody had modeled the interaction. Byproduct control and corrosion control are coupled, and a disinfectant change in a system with lead pipe is a corrosion control decision whether or not anyone treats it as one.

Flint, Michigan, 2014 to 2016

Here are the facts in order. In April 2014, while under state-appointed emergency management and seeking to reduce costs, Flint switched its drinking water source from treated Lake Huron water purchased through Detroit to the Flint River, treated at the city's own long-idle plant. The Flint River water was more corrosive, in particular carrying higher chloride, which raises the chloride-to-sulfate mass ratio, a condition associated with galvanic lead release. No orthophosphate corrosion control was applied. That omission is the technical center of the entire event.

The protective scale in Flint's lead service lines and lead-soldered plumbing destabilized. Residents complained almost immediately about color, taste, odor, and rashes, and those complaints were repeatedly discounted. In parallel, the system experienced boil water advisories for coliform detections, and elevated trihalomethanes, and an outbreak of Legionnaires' disease across 2014 and 2015 in which 12 people died.

Two outside investigations broke the case open in 2015. A Virginia Tech research team led by Marc Edwards conducted independent citizen-collected sampling in August 2015 and found a citywide 90th percentile above the action level, with individual homes in the hundreds and one at several thousand micrograms per liter. In September 2015, Dr. Mona Hanna-Attisha, a pediatrician at Hurley Medical Center, published an analysis of children's blood lead data showing that the percentage with elevated levels had roughly doubled after the switch, and had risen more in the neighborhoods with the worst water. Flint returned to Detroit water in October 2015, a federal emergency was declared in January 2016, and service line replacement proceeded over the following years.

There is a further technical failure worth stating, because it is an engineer's responsibility. Sampling protocols in use at the time, in Flint and elsewhere, could bias results low: instructing residents to remove faucet aerators before sampling, to flush the line the night before, and to fill the bottle slowly all tend to reduce measured lead. Those practices were subsequently prohibited in federal guidance and in the rule revisions. A monitoring program that systematically underestimates the thing it monitors is not a monitoring program.

Put the whole event in one sentence, because this is what a course should leave you with: no unit process failed, no pipe collapsed, and nothing exploded; a chemical that costs on the order of a hundred thousand dollars a year for a city that size was not added, the people who could see the result were not believed, and a city was poisoned. That is what environmental engineering failure usually looks like. It is quiet and it is chemical and it is institutional.

Key idea: Washington and Flint were both scale destabilization events caused by a change in water chemistry without corrosion control analysis, and in both cases the technical fix was orthophosphate while the deeper failure was oversight and the dismissal of residents reporting what they could see.

What to do about lead now

Three layers. In the short term, corrosion control optimization, meaning orthophosphate at an adequate dose plus pH and alkalinity control, verified by monitoring rather than assumed. In the medium term, full service line replacement, and here a warning: a partial replacement, in which the utility replaces its half at the main and the homeowner's half stays lead, can raise lead release for months afterward through physical disturbance and galvanic coupling between the new copper and the remaining lead. Partial replacements are now discouraged for that reason. In the interim, point-of-use filters certified to the NSF and ANSI Standard 53 for lead reduction are effective, cheap, and the correct advice to give a household while the pipe is still in the ground.

Common misconceptions

• Lead comes from the water source or the treatment plant. It is picked up from lead service lines, lead solder, and brass fixtures between the main and the tap.
• Corrosive water is the hazard. Changing water is the hazard. A stable scale in equilibrium with stable chemistry releases very little; the danger arrives with a source switch or a disinfectant switch.
• Flint's treatment plant failed. The plant produced water that met most standards. Corrosion control was not applied, which is an omitted chemical rather than a broken process.
• Replacing part of a lead service line helps somewhat. Partial replacement frequently raises lead temporarily through disturbance and galvanic coupling, which is why full replacement is the standard.
• Bigger mains are always better. Mains sized generously for fire flow create water age, which decays residual, forms byproducts, and supports nitrification.
• A compliant 90th percentile means every home is safe. The rule is a treatment technique on a sampled population, and a first-draw liter can miss lead service line water entirely.

Recap

• Distribution must hold at least 138 kilopascals everywhere at all times, is sized by fire flow rather than domestic demand, and uses elevated storage for equalization, fire, and emergency volume.
• A 400 meter dead-end of 400 millimeter main serving 26 cubic meters per day gives a water age of 1.93 days, against 0.48 days for a 200 millimeter main.
• Water age decays residual, forms byproducts at the far ends of the system, and supports nitrification in chloraminated systems.
• Non-revenue water of 15 percent costs Marley City about 9,000 cubic meters per day and 1.15 million dollars a year.
• Protective scale, not non-corrosive water, is what keeps lead out of the water, and orthophosphate at 1 to 3 milligrams per liter plus pH control is how it is maintained.
• A 15 meter lead service line holds 4.25 liters, so the regulatory first-draw one liter sample can miss it entirely, which the 2021 revisions addressed with a fifth-liter sample.
• The action level was 15 micrograms per liter at the 90th percentile from 1991, lowered to 10 by the 2024 Lead and Copper Rule Improvements, which also require lead service line replacement generally within ten years.
• Washington's 2000 switch from free chlorine to chloramine destabilized lead dioxide scale; Flint's 2014 source switch without orthophosphate destabilized scale citywide, with a Legionnaires outbreak that killed 12 alongside it.
• Full service line replacement is the durable answer; partial replacement can make things worse, and certified point-of-use filters are the correct interim advice.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Lead and Copper Rule. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Basic information about lead in drinking water. epa.gov
3. Centers for Disease Control and Prevention. (n.d.). Lead in drinking water. U.S. Department of Health and Human Services. cdc.gov
4. American Water Works Association. (n.d.). Water distribution and lead service line resources. awwa.org
5. Wikipedia. (n.d.). Flint water crisis. Wikimedia Foundation. en.wikipedia.org

The big picture

A wastewater treatment plant does not choose its influent. Whatever 120,000 people put down a drain arrives, along with whatever leaks into the pipe on the way, at a flow rate that swings by a factor of five over a day and by a factor of three more in a storm. The plant has to handle all of it without a bypass. That is a fundamentally different design problem from a water treatment plant, which pumps a metered flow of a source it selected.

This lesson covers the collection system that delivers the problem and the first two stages of the plant that receive it. By the end you will have sized Marley City's primary clarifiers, and the numbers coming out of them, the BOD and solids that remain, will be the input to the activated sludge design in the next lesson.

The collection system

Sanitary sewers run by gravity wherever the topography allows, because gravity is free and does not fail during a power outage. They are laid on a slope chosen to satisfy two competing constraints. The pipe must be steep enough to maintain a velocity of at least 0.6 to 0.9 meters per second when flowing full or half full, so that grit and solids are scoured along rather than deposited; and it must be shallow enough in slope that excavation depth does not run away, since a sewer that keeps falling eventually needs a lift station to bring it back up.

That first constraint produces a result students find counterintuitive: an oversized sewer has less usable capacity, not more. Doubling a pipe's diameter for a given flow drops the velocity below the scouring threshold, solids settle in the invert, the effective cross-section shrinks, and the pipe eventually blocks. Sewers are sized close to their design flow deliberately.

Where gravity runs out, a lift station pumps sewage up through a pressurized force main to a higher gravity reach. Lift stations are the weak points of a collection system: they have mechanical equipment, they need power, and they are the usual cause of sanitary sewer overflows during storms and outages, which is why they carry standby generators and alarms.

Infiltration and inflow

Sewers leak in both directions. Infiltration is groundwater entering through cracked pipe, bad joints, and deteriorated manholes; it is a slow, seasonal, continuous flow that rises with the water table. Inflow is stormwater entering directly through improper roof leader and foundation drain connections, holes in manhole covers, and cross-connected catch basins; it is fast and it arrives with the rain. Together they are abbreviated as I and I, and in older systems they can double or triple the flow the plant has to treat, which means the community is paying to treat, disinfect, and pump clean rainwater.

Worked example. Marley City's plant averages 45,000 cubic meters per day in dry weather. Following a 25 millimeter rainfall, the plant records 96,000 cubic meters over the next 24 hours. The excess is 51,000 cubic meters. The sewershed covers 4,000 hectares, which is 40 million square meters, so the rain that fell on it totals 0.025 meters times 40 million, or 1,000,000 cubic meters. The fraction of rainfall that reached the sewer is 51,000 divided by 1,000,000, or 5.1 percent. That figure, which the field calls the R-value for rainfall-derived inflow and infiltration, typically runs 2 to 10 percent in American systems, so Marley City is unremarkable and still spends real money on it: 51,000 cubic meters of rainwater treated for one storm.

Reducing I and I means finding it, which is done by smoke testing, dyed water testing, and closed-circuit television inspection of the pipe, and then fixing it by lining, grouting, or replacing pipe and by disconnecting improper connections. It competes for money against plant capacity, and the honest comparison an engineer must make is whether a dollar spent on rehabilitation removes more peak flow than a dollar spent on a bigger plant.

Flow variation

Domestic sewage flow follows a daily rhythm: a minimum in the small hours, a morning peak, a lesser evening peak. Smaller systems swing more, because there is less averaging. Designers use peaking factors, with a peak hour of roughly 2.5 times average for a city of this size and a minimum of about 0.4 times average.

For Marley City: average 45,000 cubic meters per day, peak hourly 112,500 cubic meters per day which is 1.30 cubic meters per second, minimum about 18,000. Every hydraulic element, channels, screens, splitter boxes, clarifier weirs, must pass the peak without flooding, while every biological element must keep working at the minimum. That spread is the central difficulty of wastewater plant hydraulics.

Key idea: Sewers are sized close to their design flow so velocity stays above the scouring threshold, and infiltration and inflow add flow the community pays to treat; a 25 millimeter storm delivering 51,000 cubic meters to Marley City's plant represents an R-value of 5.1 percent.

Hydrogen sulfide and the crown of the pipe

One phenomenon in collection systems deserves its own section, because it destroys infrastructure and it kills people.

Sewage flowing slowly in a warm, full pipe goes anaerobic. In the slime layer on the submerged pipe wall, sulfate-reducing bacteria use sulfate as an electron acceptor and produce hydrogen sulfide, which partitions into the air space above the flow. On the moist, non-submerged upper wall, a different group of bacteria, principally Thiobacillus species, oxidize that hydrogen sulfide to sulfuric acid. The acid attacks the cement paste in concrete pipe from the inside, and the crown of the pipe corrodes away while the invert, which is submerged and protected, looks fine. Sewers have collapsed from the top down while an inspection from a manhole showed nothing wrong. The controls are to maintain velocity and avoid long detention, to ventilate force main discharge points, to dose nitrate or iron salts that suppress sulfide formation, and to specify corrosion-resistant liners or plastic pipe where conditions are aggressive.

The safety side is blunter. Hydrogen sulfide is detectable by smell at about 0.01 parts per million, but at roughly 100 parts per million it paralyzes the sense of smell, so the warning disappears exactly when the danger begins, and at several hundred parts per million it causes rapid collapse and death. It is also heavier than air, so it pools in the bottom of manholes and wet wells. Confined space entry without gas monitoring and ventilation remains a leading cause of death in this industry, and it frequently kills two people, the worker who went down and the colleague who went after them.

Preliminary treatment

The first structures in a plant protect the rest of it. Bar screens intercept rags, plastics, wipes, and debris. Coarse screens use 25 to 50 millimeter openings; fine screens run 6 to 10 millimeters and remove considerably more, at the cost of more material to handle. Modern plants fight a losing battle with so-called flushable wipes, which do not disperse and which rope together into masses that jam pumps.

Screenings quantities run roughly 0.005 to 0.05 cubic meters per 1,000 cubic meters treated. At a mid-range 0.015 for Marley City, that is 0.015 times 45, or about 0.68 cubic meters per day of wet screenings going to a landfill.

Grit chambers follow, and their purpose is to remove sand, eggshells, coffee grounds, and other dense inorganic particles that would otherwise abrade pumps and accumulate in digesters. The trick is selectivity: grit must settle out while organic solids stay suspended for the primary clarifier to handle. A horizontal-flow grit chamber is therefore designed to hold a velocity near 0.3 meters per second, fast enough to keep organics moving and slow enough to drop sand. Vortex and aerated designs achieve the same separation differently. Grit quantities run about 0.004 to 0.20 cubic meters per 1,000 cubic meters; at 0.03, Marley City produces roughly 1.35 cubic meters per day.

Primary clarifiers

Primary sedimentation is plain gravity settling, and it is the best bargain in the plant: no chemicals, no aeration, no biology, just detention time. It typically removes 50 to 65 percent of suspended solids and 30 to 35 percent of BOD.

Worked example, the one that feeds the next lesson. Marley City, 45,000 cubic meters per day average, 112,500 peak hourly, design overflow rate 35 meters per day at average flow.

• Required area: 45,000 divided by 35 equals 1,286 square meters.
• Use four circular clarifiers 20 meters in diameter. Each has an area of pi over four times 400, or 314 square meters, so 1,257 square meters total. Actual average overflow rate is 45,000 divided by 1,257, or 35.8 meters per day.
• Peak check: at 112,500 cubic meters per day the overflow rate becomes 89.5 meters per day, inside the usual peak criterion of 80 to 120, so solids will not be washed over the weirs during the morning peak.
• At a 3.5 meter side water depth, total volume is 4,400 cubic meters and the detention time at average flow is 4,400 divided by 45,000, or 0.098 days, which is 2.35 hours, in the normal 1.5 to 2.5 hour band.
• Weir loading: peripheral weir length is pi times 20, or 62.8 meters per clarifier, so 251 meters total. At average flow that is 45,000 divided by 251, or 179 cubic meters per day per meter of weir, within the usual 125 to 250 limit. Exceed it and the upward velocity near the weir entrains settled solids.

Now the mass balance, and carry these numbers forward. Raw wastewater at 250 milligrams per liter of BOD and 250 of suspended solids at 45,000 cubic meters per day delivers 11,250 kilograms per day of each. Primary treatment removing 60 percent of solids and 33 percent of BOD gives:

Stream	BOD	Suspended solids
Raw influent	250 milligrams per liter, 11,250 kilograms per day	250 milligrams per liter, 11,250 kilograms per day
Removed in primary	3,712 kilograms per day	6,750 kilograms per day
Primary effluent to secondary	167.5 milligrams per liter, 7,538 kilograms per day	100 milligrams per liter, 4,500 kilograms per day

Look at what gravity just bought. It removed 3,712 kilograms of BOD per day that will now never have to be aerated. At roughly one kilogram of oxygen per kilogram of BOD, and at the 1 to 2 kilowatt-hours per kilogram of oxygen that a real aeration system delivers, primary settling is saving something like 4,000 to 7,000 kilowatt-hours a day of electricity. That is why primary clarifiers survive in an era of processes that could technically do without them.

The removed solids leave as primary sludge: 6,750 kilograms per day of dry solids, thick and highly putrescible, typically drawn at 4 to 6 percent solids. At 4 percent and a density near 1,010 kilograms per cubic meter, the volume is 6,750 divided by 0.04 times 1,010, or about 167 cubic meters per day of sludge to be handled. Lesson 12 handles it.

Key idea: Four 20 meter primary clarifiers give Marley City 1,257 square meters, an average overflow rate of 35.8 meters per day, 2.35 hours of detention, and a peak rate of 89.5, removing 6,750 kilograms per day of solids and 3,712 of BOD by gravity alone and leaving 167.5 milligrams per liter of BOD for the biological stage.

What else arrives

Two influent realities shape modern practice. First, industrial contributions. A single plating shop or dairy can deliver a load that overwhelms a municipal plant or poisons its biology with metals. The Clean Water Act's pretreatment program addresses this by requiring industrial users to treat their waste to categorical standards before discharging to a sewer, and larger cities administer their own permits under it. When you hear that a plant upset was caused by a slug discharge, this is the program that failed.

Second, fats, oils, and grease. Restaurant grease congeals in cold sewers, captures wipes and debris, and builds the blockages that cause most dry-weather sanitary sewer overflows. Grease interceptor requirements on food service establishments are the standard control, and enforcing them is unglamorous, cheap, and genuinely effective.

Common misconceptions

• A bigger sewer has more capacity. Below the scouring velocity of roughly 0.6 meters per second, solids deposit in the invert and the effective capacity falls.
• Infiltration and inflow are the same thing. Infiltration is groundwater entering slowly through defects; inflow is stormwater entering directly and immediately through improper connections. They are found and fixed differently.
• Sewers corrode from the flowing sewage. The invert is protected. Biologically generated sulfuric acid destroys the crown, above the waterline, which is why a manhole inspection can miss an imminent collapse.
• Hydrogen sulfide is safe as long as you can smell it. It paralyzes the sense of smell near 100 parts per million, so the warning vanishes as the hazard becomes lethal.
• Primary clarifiers are obsolete. They remove roughly a third of the BOD with no energy input, saving thousands of kilowatt-hours a day of aeration at a plant this size.
• Grit chambers are just small clarifiers. They are tuned to a velocity near 0.3 meters per second precisely so that they capture sand while letting organic solids pass to the primary clarifier.

Recap

• Gravity sewers are sloped to keep velocity above 0.6 to 0.9 meters per second, so oversizing reduces rather than increases effective capacity.
• Infiltration is slow groundwater entry; inflow is direct storm entry; Marley City's 25 millimeter storm added 51,000 cubic meters, an R-value of 5.1 percent.
• Peak hourly flow is about 2.5 times average, so Marley City's hydraulics must pass 112,500 cubic meters per day while its biology must work at 18,000.
• Sulfate reduction in the flow plus sulfide oxidation on the crown produces sulfuric acid that destroys concrete pipe from the top down, and hydrogen sulfide kills workers in confined spaces after deadening their sense of smell.
• Screens produce about 0.68 cubic meters per day of screenings and grit chambers about 1.35 cubic meters per day at this plant.
• Four 20 meter primary clarifiers give 1,257 square meters, 35.8 meters per day at average flow, 89.5 at peak, 2.35 hours detention, and 179 cubic meters per day per meter of weir.
• Primary treatment removes 6,750 kilograms per day of solids and 3,712 of BOD, leaving 167.5 milligrams per liter of BOD and 100 of solids for secondary treatment.
• Primary sludge is 6,750 kilograms per day of dry solids, roughly 167 cubic meters per day at 4 percent.
• Industrial pretreatment and grease interceptor programs control the two influent problems that most often disable a plant or a sewer.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Primer for municipal wastewater treatment systems. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Sanitary sewer overflows and the national pretreatment program. epa.gov
3. Centers for Disease Control and Prevention, National Institute for Occupational Safety and Health. (n.d.). Hydrogen sulfide. cdc.gov
4. Water Environment Federation. (n.d.). Resources on collection systems and wastewater treatment. wef.org
5. Wikipedia. (n.d.). Sanitary sewer. Wikimedia Foundation. en.wikipedia.org

The big picture

Everything so far has been physics and chemistry. Now the plant becomes a farm. The activated sludge process, invented in Manchester, England in 1914 by Edward Ardern and William Lockett, cultivates a dense mixed population of bacteria and protozoa in an aerated tank, feeds them the organic matter in the wastewater, and then settles them out and sends most of them back to eat again. That recycling of the biomass is what makes the process work and what gives it its name: the sludge is activated because it is alive and returned rather than discarded.

The operator of a wastewater plant is not running a machine. They are managing a livestock herd of roughly ten thousand kilograms of microorganisms whose appetite, health, and settling behavior determine whether the permit is met. This lesson teaches the three numbers that control that herd, sizes Marley City's basin, and then discovers that the ammonia limit we computed back in Lesson 2 forces the basin to be substantially larger than the BOD alone would require.

The vocabulary

The contents of an aeration basin are called mixed liquor. Its solids concentration is mixed liquor suspended solids, or MLSS, typically 2,000 to 4,000 milligrams per liter in a conventional plant. Roughly 70 to 85 percent of that is organic, and the organic fraction, mixed liquor volatile suspended solids or MLVSS, is used as the proxy for actual living biomass because you cannot count bacteria in a plant every day.

Mixed liquor flows to a secondary clarifier. The settled biomass is split: most is return activated sludge, pumped straight back to the head of the aeration basin, and a small deliberate fraction is waste activated sludge, removed from the system entirely. That wasting rate is the single most important control an operator has, because it sets how long the average organism stays in the system.

Three parameters govern the design.

• Hydraulic retention time is basin volume divided by influent flow: how long the water stays. Typically 4 to 8 hours.
• Solids retention time, also called sludge age or SRT, is the mass of solids in the system divided by the mass wasted per day: how long the organisms stay. Typically 4 to 15 days. Because solids are recycled and water is not, sludge age is far longer than hydraulic retention time, and that decoupling is the central trick of the process.
• Food-to-microorganism ratio, or F over M, is the BOD applied per day divided by the mass of volatile solids in the basin, in kilograms of BOD per kilogram of MLVSS per day. Conventional plants run 0.2 to 0.5.

Sludge age is what selects the population. At a short sludge age the fast-growing carbon-eating heterotrophs dominate and slow growers wash out. At a long sludge age slower organisms can establish, and the crucial slow growers are the nitrifiers, which is why every nitrifying plant is a long sludge age plant. Longer sludge age also means the biomass spends more time in endogenous decay, so less waste sludge is produced per unit of BOD treated, which is a genuine operating saving.

Key idea: Activated sludge decouples how long the water stays from how long the organisms stay, and sludge age, controlled by the wasting rate, is what determines which organisms live in the tank and therefore what the plant can do.

Sizing Marley City's aeration basin

From the last lesson, primary effluent carries 167.5 milligrams per liter of BOD at 45,000 cubic meters per day, a load of 7,538 kilograms per day.

Worked example, step by step. Design at an F over M of 0.30 kilograms of BOD per kilogram of MLVSS per day, with an MLSS of 3,000 milligrams per liter and a volatile fraction of 0.80, giving an MLVSS of 2,400 milligrams per liter, which is 2.4 kilograms per cubic meter.

• Required biomass: 7,538 divided by 0.30 equals 25,127 kilograms of MLVSS that must be present in the basin.
• Required volume: 25,127 divided by 2.4 kilograms per cubic meter equals 10,470 cubic meters.
• Hydraulic retention time: 10,470 divided by 45,000 equals 0.233 days, or 5.6 hours, squarely in the 4 to 8 hour range.
• Build it as four parallel basins of about 2,620 cubic meters each, say 40 meters long by 13 meters wide by 5 meters of side water depth, which allows one to be dewatered for maintenance.

Now the sludge age. Sludge production is estimated from an observed yield, the mass of biomass produced per mass of BOD removed, typically 0.4 to 0.6 kilograms of volatile solids per kilogram of BOD removed for a conventional plant. BOD removed is 167.5 minus about 10 milligrams per liter remaining, or 157.5, which is 7,088 kilograms per day. At an observed yield of 0.40, waste activated sludge production is 0.40 times 7,088, or 2,835 kilograms of volatile solids per day.

Sludge age is then the mass in the system divided by the mass leaving it: 25,127 divided by 2,835 equals 8.9 days. At 20 degrees Celsius that is comfortably enough for nitrification.

The winter problem

Here is where the course's numbers collide, and it is the most instructive moment in the module. Back in Lesson 2 we computed Marley City's water-quality-based ammonia limit at the 7Q10 low flow and got 6.2 milligrams per liter as nitrogen. That limit requires the plant to nitrify, and it applies in every season.

Nitrifiers grow slowly, and their growth rate falls steeply with temperature. In a Midwestern city the mixed liquor may reach 12 degrees Celsius in February. The minimum sludge age at which nitrifiers can maintain themselves at that temperature is on the order of 5 to 7 days, and design practice applies a safety factor of roughly 2 to 2.5 to cover peak loads and diurnal swings, which puts the required design sludge age near 15 days. Our basin gives 8.9. It will nitrify in July and lose nitrification in February, which is precisely when the receiving water is least able to assimilate ammonia and when the un-ionized fraction is lowest but the load is highest.

So redesign for winter. Holding the same wasting rate, a 15 day sludge age requires 15 times 2,835, or 42,525 kilograms of MLVSS in the system. At the same 2.4 kilograms per cubic meter, that is 17,719 cubic meters, a hydraulic retention time of 9.4 hours. The basin has grown by 69 percent purely to satisfy a nitrogen limit, and that ratio is a fair statement of what nutrient limits cost in capital.

Two honest footnotes. Observed yield actually falls as sludge age rises, because more of the biomass is consumed in endogenous respiration, so the true required volume is somewhat less than 17,719; a full design would iterate. And a designer would compare this against raising the MLSS instead, which shrinks the basin but loads the clarifier harder, as the next section shows. Real design is the negotiation between those two tanks.

The oxygen bill

Now compute what it costs to run. Carbonaceous oxygen demand is approximately the ultimate BOD removed minus the oxygen equivalent of the biomass hauled away, since carbon that leaves as sludge was never oxidized. Converting five-day BOD to ultimate with the factor of 0.68 from Lesson 5, the removed load of 7,088 kilograms per day corresponds to 10,424 kilograms of ultimate demand. Subtract 1.42 times the 2,835 kilograms of sludge wasted, or 4,026, and carbonaceous demand is about 6,398 kilograms of oxygen per day.

Nitrification adds its own. The influent carries 1,800 kilograms of nitrogen per day, of which roughly 12 percent of the volatile solids produced is assimilated into new cells, about 340 kilograms, leaving 1,460 kilograms actually oxidized. At 4.57 kilograms of oxygen per kilogram of nitrogen, that is 6,672 kilograms of oxygen per day.

Total oxygen requirement is about 13,070 kilograms per day, and note what that says: nitrification roughly doubles the aeration demand, exactly as Lesson 6 promised. A good fine-bubble diffused aeration system in the field delivers on the order of 1.2 kilograms of oxygen per kilowatt-hour, so aeration alone needs about 10,900 kilowatt-hours per day. A plant of this size uses roughly 0.5 kilowatt-hours per cubic meter treated, or 22,500 kilowatt-hours per day in total, so aeration is close to half the plant's entire electricity consumption. This is why blower control, dissolved oxygen setpoints, and fine-bubble diffuser maintenance are not trivia. Running the basin at 4 milligrams per liter of dissolved oxygen instead of the 2 that biology needs can waste a substantial fraction of that energy for no benefit.

Key idea: Marley City's basin needs 10,470 cubic meters for BOD alone at a 8.9 day sludge age, but 17,719 cubic meters to hold a 15 day sludge age for winter nitrification, and the resulting oxygen demand of about 13,070 kilograms per day makes aeration roughly half the plant's electricity bill.

Secondary clarification, and the criterion that actually governs

The biology only counts if the biomass can be separated from the water. A secondary clarifier does two jobs at once: it clarifies, producing a low-solids effluent, and it thickens, producing a concentrated return stream. Those two jobs give it two independent design criteria, and you must check both.

Worked example. Marley City, 45,000 cubic meters per day, MLSS 3,000 milligrams per liter, return sludge ratio 0.5.

• Overflow rate criterion. Use 24 meters per day at average flow. Required area is 45,000 divided by 24, or 1,875 square meters. Note that this uses influent flow only, because the return sludge leaves through the bottom rather than over the weir.
• Solids loading criterion. Use 4.0 kilograms per square meter per hour. The solids arriving are carried by the influent plus the return: 45,000 times 1.5, or 67,500 cubic meters per day, at 3.0 kilograms per cubic meter, giving 202,500 kilograms per day, or 8,438 kilograms per hour. Required area is 8,438 divided by 4.0, or 2,110 square meters.
• Solids loading governs, as it usually does at ordinary MLSS levels. Provide four clarifiers of 26 meters diameter, area 531 square meters each, 2,124 square meters total. The actual overflow rate is then 45,000 divided by 2,124, or 21.2 meters per day, comfortably under the criterion.

Notice the design tension this exposes. Raising MLSS shrinks the aeration basin, because the same biomass fits in less volume, but it raises the solids loading on the clarifier proportionally and therefore grows the clarifier. Every activated sludge design is a trade between those two structures, and the reason membrane bioreactors are attractive is that they abolish the clarifier entirely by filtering the mixed liquor through membranes, which permits MLSS of 8,000 to 12,000 milligrams per liter and a very small basin, at a substantial energy and membrane replacement cost.

Settleability and what goes wrong

Whether the clarifier works depends on how the sludge settles, which is measured by the sludge volume index. Settle a liter of mixed liquor for 30 minutes, read the settled volume in milliliters, multiply by 1,000, and divide by the MLSS in milligrams per liter; the result is the volume in milliliters occupied by one gram of sludge.

Worked example. A liter of Marley City mixed liquor settles to 360 milliliters in 30 minutes at an MLSS of 3,000 milligrams per liter. The index is 360 times 1,000 divided by 3,000, or 120 milliliters per gram. Below 100 is excellent, 100 to 150 is good, and above 150 indicates bulking, in which the sludge blanket rises, solids wash over the weir, and the plant violates its suspended solids permit while the biology upstream is working perfectly.

Bulking is almost always caused by filamentous organisms, long thread-like bacteria that bridge between floc particles and hold the sludge open. The usual triggers are a persistently low food-to-microorganism ratio, low dissolved oxygen, a nitrogen or phosphorus deficiency in an industrial-heavy influent, and septic influent high in sulfide. The structural remedy is a selector, a small high-concentration contact zone at the head of the basin where floc-forming organisms out-compete filaments for readily available substrate. The emergency remedy is chlorinating the return sludge, which works and which also damages the culture, and is therefore the thing an operator does at 2 a.m. rather than the thing a designer plans for. A related nuisance, thick brown foam driven by Nocardia-type filaments, is managed by keeping sludge age down and by physically removing the foam rather than letting it recirculate.

The alternatives

Trickling filters spray wastewater over a bed of rock or plastic media coated in biological film, using far less energy because air moves by natural draft, at the cost of less control and generally poorer effluent. Rotating biological contactors mount discs on a slowly turning shaft, half submerged. Moving bed and integrated fixed-film systems add suspended plastic carriers to an activated sludge basin, growing attached biomass that increases capacity without enlarging the tank, which is one of the most common ways to retrofit a plant for nitrification without building a new basin. Membrane bioreactors, as noted, replace the clarifier with membranes and produce an effluent clean enough to feed reverse osmosis directly, which is the foundation of the potable reuse systems in the next lesson.

Whatever the configuration, the result at Marley City is an effluent near 10 milligrams per liter of BOD and 10 of suspended solids, against a permit of 30 and 30. Overall removal from the raw 250 is 96 percent, and the plant runs with margin because permits are written on averages and a plant with no margin fails on its worst day.

Common misconceptions

• The clarifier is where treatment happens. Treatment happens in the aeration basin. The clarifier separates the biomass, and it fails for physical reasons while the biology is fine.
• Return sludge and waste sludge are the same stream. Return sludge keeps the culture concentrated and never leaves the system; waste sludge is the deliberate removal that sets sludge age.
• More dissolved oxygen is always better. Above roughly 2 milligrams per liter the biology gains nothing, and aeration is about half the plant's electricity, so excess dissolved oxygen is pure waste.
• Secondary clarifiers are sized by overflow rate. At ordinary MLSS the solids loading criterion governs, as it did here at 2,110 square meters against 1,875.
• Bulking means the plant is overloaded. Filamentous bulking is more often caused by a persistently low food-to-microorganism ratio, low dissolved oxygen, or nutrient deficiency than by excess load.
• Nitrification is a chemical addition. It is a slow-growing population that can only be held by running a long enough sludge age at the coldest temperature of the year.

Recap

• Activated sludge recycles settled biomass so that sludge age is decoupled from hydraulic retention time, and the wasting rate is the operator's primary control.
• At an F over M of 0.30 and an MLVSS of 2,400 milligrams per liter, Marley City's 7,538 kilogram per day BOD load needs 25,127 kilograms of biomass in 10,470 cubic meters, a 5.6 hour retention time.
• Waste sludge production of 2,835 kilograms per day gives a sludge age of 8.9 days, enough to nitrify in summer and not in winter.
• Holding a 15 day sludge age for 12 degree winter nitrification requires 17,719 cubic meters, a 69 percent larger basin, which is what the Lesson 2 ammonia limit of 6.2 milligrams per liter actually costs.
• Oxygen demand is about 6,398 kilograms per day carbonaceous plus 6,672 nitrogenous, roughly 13,070 total, near half the plant's electricity at about 1.2 kilograms of oxygen per kilowatt-hour.
• Secondary clarifiers must satisfy both overflow rate and solids loading; solids loading governed here, giving 2,110 square meters provided as four 26 meter units.
• Raising MLSS shrinks the basin and grows the clarifier, which is the central trade in every activated sludge design.
• A sludge volume index of 120 milliliters per gram is good; above 150 indicates filamentous bulking, addressed structurally by a selector and in emergencies by chlorinating the return sludge.
• The finished effluent is near 10 milligrams per liter of BOD and solids against a 30 and 30 permit, about 96 percent removal from the raw wastewater.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Wastewater technology fact sheets. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Energy efficiency for water and wastewater utilities. epa.gov
3. Water Environment Federation. (n.d.). Activated sludge and secondary treatment resources. wef.org
4. Encyclopaedia Britannica. (n.d.). Wastewater treatment. britannica.com
5. Wikipedia. (n.d.). Activated sludge. Wikimedia Foundation. en.wikipedia.org

The big picture

Marley City's plant now meets its BOD and solids permit. Three problems remain. Its effluent still carries nitrogen and phosphorus, which the permit limits and the river cannot absorb. It produces about ten tonnes of dry solids a day that have to go somewhere. And the water it discharges is clean enough that a growing number of communities have concluded it should not simply be thrown away.

This lesson closes all three. It is also where the numbers from Lesson 2, Lesson 6, and Lesson 11 finally resolve: the ammonia limit that forced nitrification, the alkalinity that nitrification threatened to exhaust, and the aeration bill that nitrification doubled all get settled by one process rearrangement.

Biological nitrogen removal

Nitrification converts ammonia to nitrate. It does not remove nitrogen; the nitrate is still there, and total nitrogen limits are increasingly common in nutrient-sensitive watersheds. Removal requires denitrification: anoxic conditions, meaning nitrate present but dissolved oxygen absent, plus a carbon source for the bacteria to oxidize.

Where does the carbon come from? The obvious answer is the influent, which is full of BOD. But the influent enters the front of the plant and the nitrate is produced at the back. The standard solution reverses the geometry. In the Modified Ludzack-Ettinger configuration, an anoxic zone is placed first, before the aerobic zone, and nitrate-rich mixed liquor is pumped backward from the end of the aerobic zone into that anoxic zone through an internal recycle. Raw carbon meets returned nitrate at the head of the plant, and the denitrifiers get a free carbon source.

How much nitrogen can that remove? Only the nitrate you actually bring back can be denitrified, so removal is set by the recycle ratios. With an internal recycle ratio of 3 and a return activated sludge ratio of 0.5, the fraction of nitrified nitrogen delivered to the anoxic zone is 3.5 divided by 4.5, or 78 percent, and the remaining 22 percent leaves in the effluent as nitrate.

Worked example. From Lesson 11, Marley City oxidizes 1,460 kilograms of nitrogen per day. At 78 percent capture, 1,136 kilograms per day are denitrified and 324 escape as nitrate, which at 45,000 cubic meters per day is 7.2 milligrams per liter. Add about 1 milligram per liter of residual ammonia and 1.5 of unbiodegradable organic nitrogen and the effluent total nitrogen is roughly 9.7 milligrams per liter, down from an influent 40. That is a normal biological nutrient removal result, and pushing below about 8 requires a second anoxic zone fed with purchased methanol, because by then the influent carbon is gone. The cost curve steepens sharply at that point, which is worth remembering whenever a permit limit is being negotiated.

Two dividends fall out of the same rearrangement, and they are large.

Alkalinity. Lesson 6 warned that nitrifying 40 milligrams per liter of nitrogen would consume 286 milligrams per liter of alkalinity while typical wastewater carries only 200 to 250. Recompute with real numbers. Nitrogen actually oxidized is 1,460 kilograms per day, or 32.4 milligrams per liter, consuming 7.14 times 32.4, or 232 milligrams per liter. Nitrogen denitrified is 1,136 kilograms per day, or 25.2 milligrams per liter, returning 3.57 times 25.2, or 90. Net consumption is 142 milligrams per liter. Against an influent alkalinity of 220, that leaves 78 milligrams per liter, enough to hold the pH where nitrifiers work. The anoxic zone paid for the buffering, and a plant without one would be buying caustic soda every day.

Oxygen. Denitrification uses nitrate rather than oxygen as the electron acceptor, which recovers 2.86 kilograms of oxygen equivalent per kilogram of nitrogen reduced: 2.86 times 1,136, or 3,249 kilograms of oxygen per day. Against the 13,070 kilogram per day demand computed in Lesson 11, that is a 25 percent cut in the aeration bill, roughly 2,700 kilowatt-hours a day. Nutrient removal, done this way, partly pays for itself.

Key idea: Placing the anoxic zone first and recycling nitrate backward denitrifies 78 percent of the oxidized nitrogen using influent carbon, which drops effluent total nitrogen to about 9.7 milligrams per liter, returns 90 milligrams per liter of alkalinity, and cuts aeration demand by roughly a quarter.

Phosphorus removal

Phosphorus has no gas phase, so it can only leave in the sludge. There are two routes.

Chemical precipitation adds iron or aluminum salts that form insoluble metal phosphate. It is reliable, instantly controllable, and it makes more sludge.

Worked example. To remove 5 milligrams per liter of phosphorus, note that 5 divided by the atomic mass 31 gives 0.161 millimoles per liter. Real plants need a molar ratio of iron to phosphorus near 1.8 rather than 1.0, because the metal also hydrolyzes, so 0.29 millimoles per liter of iron are required. As ferric chloride at 162.2 grams per mole, that is 47 milligrams per liter, which at 45,000 cubic meters per day is about 2,115 kilograms of ferric chloride per day. That is a chemical delivery every day or two, forever, plus the extra sludge it generates.

Enhanced biological phosphorus removal instead exploits organisms that behave strangely when cycled through an anaerobic zone and then an aerobic one: anaerobically they take up simple organic acids and release phosphate, aerobically they take phosphorus back up well beyond their metabolic need, a behavior called luxury uptake. Waste that biomass and the phosphorus leaves with it. A well-run biological system reaches 0.5 to 1.0 milligrams per liter with no chemicals, though it is more fragile than a chemical feed and often backed up by one. Typical permits sit at 1.0 milligram per liter of total phosphorus, while plants discharging to the Chesapeake Bay or the Great Lakes have been pushed to 0.1 or below, where the cost per kilogram removed is many times what it is at the start of the curve.

Effluent disinfection

Permits generally limit fecal indicator bacteria, commonly around 200 fecal coliforms per 100 milliliters as a monthly geometric mean. Chlorine works, but a chlorine residual discharged to a river is itself acutely toxic to fish, so a chlorinating plant must then dechlorinate with sulfur dioxide or sodium bisulfite. That two-chemical arrangement is why ultraviolet disinfection has become the default for new plants: no chemicals, no residual to remove, no byproducts, and the absence of a residual is harmless here because nothing downstream needs protecting. The catch is that ultraviolet performance depends on the water transmitting light, so a plant with poor solids removal loses disinfection immediately, which couples clarifier performance directly to the bacteria limit.

The solids stream

Now follow the mass. Marley City produces 6,750 kilograms per day of primary solids and 2,835 kilograms per day of waste activated volatile solids, which at a volatile fraction of 0.8 is about 3,540 kilograms per day of total solids. Together, roughly 10,290 kilograms per day of dry solids. Handling them typically consumes a third to a half of a plant's operating budget, and it is the part of the job that new engineers are least prepared for.

The sequence is thicken, stabilize, dewater, dispose. Thickening comes first because every downstream tank is sized by volume. Primary sludge thickens by gravity to 4 to 6 percent; waste activated sludge, light and flocculent, will not, and is thickened by dissolved air flotation, gravity belt, or rotary drum instead. Taking it from 0.8 percent to 5 percent cuts its volume by more than a factor of six, and every cubic meter removed here is a cubic meter of digester that need not be built.

Anaerobic digestion is the standard stabilization process at this scale and one of the most elegant unit operations in the field. In a heated, sealed, mixed tank at about 35 degrees Celsius with a solids retention time of 15 to 20 days, a consortium of bacteria hydrolyzes, ferments, and converts organic matter to methane and carbon dioxide. It destroys 45 to 55 percent of the volatile solids, kills most pathogens, removes most of the odor, improves dewaterability, and produces fuel.

Worked example. Feed 10,290 kilograms per day of dry solids at a volatile fraction of 0.75, so 7,718 kilograms of volatile solids per day. Destroy 50 percent, or 3,859 kilograms per day. At 0.95 cubic meters of biogas per kilogram of volatile solids destroyed, gas production is 3,666 cubic meters per day. At 65 percent methane that is 2,383 cubic meters of methane, and at 35.8 megajoules per cubic meter the energy content is about 85,300 megajoules per day, or 23,700 kilowatt-hours per day of thermal energy.

Run that through a combined heat and power engine at 35 percent electrical efficiency and you get roughly 8,300 kilowatt-hours per day of electricity, plus recoverable jacket and exhaust heat that easily covers digester heating. The plant's total demand is about 22,500 kilowatt-hours per day, so the digesters supply on the order of 37 percent of the plant's electricity from what used to be a disposal problem. A number of plants have reached full energy neutrality by adding high-strength organic waste such as restaurant grease and food waste to the digesters, a practice called co-digestion, which raises gas production substantially at almost no capital cost.

Size the digesters. At 4 percent feed solids and a density near 1,010 kilograms per cubic meter, feed volume is 10,290 divided by 40.4, or 255 cubic meters per day. At a 20 day solids retention time the required volume is 5,100 cubic meters, built as two 2,550 cubic meter tanks roughly 20 meters in diameter, so one can be taken down without stopping the plant.

Smaller plants without digesters use aerobic digestion, simpler and producing no gas, or lime stabilization above pH 12. Dewatering follows, by belt filter press, centrifuge, or screw press with a polymer conditioner, producing a cake at 18 to 25 percent solids that a front loader can handle.

Key idea: Anaerobic digestion destroys about half the volatile solids and turns them into 3,666 cubic meters per day of biogas, worth roughly 8,300 kilowatt-hours of electricity daily, which is about 37 percent of the entire plant's power.

Biosolids

What leaves the dewatering equipment is biosolids, the term for treated sewage solids that meet the federal standards for beneficial use. After digestion, Marley City has 10,290 minus 3,859, or about 6,431 kilograms per day of dry solids remaining; dewatered to 20 percent cake that is 32 tonnes per day of wet material, roughly a truckload and a half every day, every day, forever.

The governing rule is the federal biosolids regulation, 40 CFR Part 503, and it sets three kinds of requirement: ceiling and pollutant concentration limits for nine metals, pathogen reduction, and vector attraction reduction, meaning the material must not attract flies and rodents. Pathogen reduction comes in two grades. Class B biosolids have substantially reduced pathogens and may be land applied with restrictions on crop harvesting, grazing, and public access. Class A biosolids have pathogens reduced below detection, generally by composting, heat drying, or thermophilic digestion, and can be sold in bags to the public; the products sold by several large cities are Class A biosolids.

Worked example. Land application is limited by the agronomic rate, the amount that supplies the nitrogen a crop needs and no more, so that nitrate does not leach to groundwater. Take biosolids at 4 percent nitrogen with 30 percent of it available in the first year, and a corn crop needing 170 kilograms of nitrogen per hectare. Each dry tonne supplies 1,000 times 0.04 times 0.30, or 12 kilograms of available nitrogen. The application rate is 170 divided by 12, or 14.2 dry tonnes per hectare. Marley City produces 6,431 kilograms per day, which is 2,347 dry tonnes a year, so it needs about 165 hectares of cropland a year, and it needs it reliably, within haul distance, with cooperative landowners.

Be honest about the controversy, because it is live. Land application recycles nutrients and organic matter and avoids landfilling, and decades of monitoring have found the metals limits protective. It also faces two genuine problems. Odor and truck traffic generate real local opposition, and the burden falls disproportionately on rural and low-income communities, which is the Lesson 3 analysis applied to a specific facility decision. And per- and polyfluoroalkyl substances concentrate in biosolids because they partition to solids and survive digestion; several states have restricted land application on those grounds and Maine banned it outright in 2022. The alternatives cost more and carry their own emissions. There is no clean answer today, and an engineer should say so.

Water reuse

The last idea in this module is that treated effluent is a resource. Reuse comes in a hierarchy of increasing barrier requirements.

Type	Typical treatment	Example use
Non-potable reuse	Secondary plus filtration and disinfection	Landscape and agricultural irrigation, cooling towers, toilet flushing, dust control
Indirect potable reuse	Advanced treatment, then an environmental buffer such as an aquifer or reservoir	Groundwater recharge, reservoir augmentation
Direct potable reuse	Advanced treatment with engineered storage, straight to the water plant or distribution	Drought response and permanent supply in water-scarce cities

The reference project is the Groundwater Replenishment System of the Orange County Water District in California, which takes secondary effluent through microfiltration, reverse osmosis, and ultraviolet light with hydrogen peroxide as an advanced oxidation step, then injects and percolates the product into the groundwater basin that supplies drinking water wells. Operating since 2008 and expanded twice, it now produces on the order of 490,000 cubic meters per day, enough for roughly a million people, at a cost per cubic meter that compares favorably with imported water or seawater desalination in that region.

Direct potable reuse, with no environmental buffer, has a longer history than most people expect: Windhoek, Namibia, has practiced it since 1968 out of necessity. In the United States, Big Spring, Texas, started in 2013, Wichita Falls ran a temporary system during the 2014 drought, and California adopted comprehensive direct potable reuse regulations in December 2023.

The public objection, usually phrased as toilet to tap, deserves a straight engineering answer rather than a dismissal. It is this: every water supply downstream of another city is already indirect potable reuse, and the Mississippi, Ohio, and Thames have functioned that way for a century. The difference in an engineered reuse system is that the number of barriers is larger, each one is instrumented and continuously monitored, and the residence time between treatment and consumption is known rather than assumed. The engineering question is never whether water has been used before. It is how many independent barriers stand between the last use and this glass, and how you would know if one of them failed.

Common misconceptions

• Nitrification removes nitrogen. It converts ammonia to nitrate, which is still nitrogen in the effluent. Only denitrification to nitrogen gas removes it.
• Nutrient removal always costs more energy. Denitrification returned 3,249 kilograms of oxygen equivalent per day at Marley City, cutting aeration demand by roughly a quarter, and it recovered 90 milligrams per liter of alkalinity.
• Sludge is a nuisance byproduct. Digested at this plant it supplies about 37 percent of the site's electricity, and co-digestion of food waste and grease can carry a plant to energy neutrality.
• Class A and Class B biosolids differ in metals content. They differ in pathogen reduction. Metals limits apply to both.
• Land application rates are set by how much sludge is available. They are set by the agronomic nitrogen requirement of the crop, which for Marley City means about 165 hectares a year.
• Potable reuse is a novel and unproven idea. Windhoek has done it since 1968 and Orange County recharges roughly 490,000 cubic meters per day; the real question is barrier count and monitoring, not novelty.

Recap

• An anoxic zone placed first with internal recycle denitrifies 78 percent of oxidized nitrogen at recycle ratios of 3 and 0.5, giving about 9.7 milligrams per liter of effluent total nitrogen.
• That arrangement returns 90 milligrams per liter of alkalinity, leaving 78 of the influent's 220, and recovers 3,249 kilograms per day of oxygen equivalent.
• Chemical phosphorus removal needs an iron to phosphorus molar ratio near 1.8, about 2,115 kilograms per day of ferric chloride here; biological removal reaches 0.5 to 1.0 milligrams per liter without chemicals.
• Ultraviolet light has displaced chlorine for effluent disinfection because a chlorine residual is toxic to fish and would have to be removed again.
• Marley City generates about 10,290 kilograms per day of dry solids, thickened, then digested at 35 degrees Celsius for 20 days in 5,100 cubic meters of tank.
• Digestion destroys 3,859 kilograms per day of volatile solids and yields 3,666 cubic meters per day of biogas, about 8,300 kilowatt-hours per day of electricity, roughly 37 percent of plant demand.
• Roughly 6,431 kilograms per day of dry biosolids remain, 32 tonnes per day of 20 percent cake, requiring about 165 hectares a year at an agronomic rate of 14.2 dry tonnes per hectare.
• Part 503 sets metals limits, pathogen classes A and B, and vector attraction reduction; PFAS in biosolids is an unresolved and consequential problem.
• Reuse runs from non-potable through indirect to direct potable, with Orange County at about 490,000 cubic meters per day and Windhoek practicing direct potable reuse since 1968.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Biosolids laws and regulations. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Water reuse and recycling. epa.gov
3. U.S. Environmental Protection Agency. (n.d.). Nutrient removal from wastewater. epa.gov
4. Orange County Water District. (n.d.). Groundwater Replenishment System. ocwd.com
5. Wikipedia. (n.d.). Anaerobic digestion. Wikimedia Foundation. en.wikipedia.org

The big picture

Water problems have edges. A river has banks, a plume has a boundary, a pipe has a discharge point that somebody owns. Air has none of that. Emissions leave a stack and become everyone's, mixing across county and national borders within hours, and the concentration a person breathes depends less on how much was emitted than on what the atmosphere happened to be doing that morning.

That single fact organizes air quality engineering. You cannot dilute your way out of a problem you share with everyone downwind, you cannot sample the receiving body meaningfully at one point, and the same emission rate produces harmless concentrations on a windy afternoon and lethal ones under a stagnant inversion. This lesson builds the pollutants, the standards, the formation chemistry, and then the meteorology, and it ends by computing exactly how much difference the weather makes.

The six criteria pollutants

The Clean Air Act designates a small set of common pollutants for which EPA publishes health criteria and sets national ambient standards. There are six, and their standards are worth knowing approximately.

Pollutant	Primary standard	Main sources
Fine particulate matter, PM2.5	9.0 micrograms per cubic meter annual, 35 over 24 hours	Combustion of all kinds, plus secondary formation from sulfur dioxide, nitrogen oxides, ammonia, and organics
Coarse particulate matter, PM10	150 micrograms per cubic meter over 24 hours	Road dust, construction, agriculture, crushing operations
Ozone	0.070 parts per million over 8 hours	Entirely secondary, formed from nitrogen oxides and volatile organic compounds in sunlight
Nitrogen dioxide	100 parts per billion over 1 hour, 53 annual	Combustion at high temperature: vehicles, power plants, boilers
Sulfur dioxide	75 parts per billion over 1 hour	Combustion of sulfur-bearing fuel, chiefly coal and oil, and smelting
Carbon monoxide	9 parts per million over 8 hours, 35 over 1 hour	Incomplete combustion, historically dominated by motor vehicles
Lead	0.15 micrograms per cubic meter, rolling 3-month average	Smelters, piston aircraft fuel, battery recycling; formerly gasoline

Two structural points about that table. The fine particle annual standard was tightened from 12.0 to 9.0 micrograms per cubic meter in February 2024, which moved a large number of counties into nonattainment overnight and is the current center of gravity in American air regulation. And note that only lead has a national standard expressed as a mass concentration alongside particles; everything gaseous is in parts per million or billion, so the conversion you learned in Lesson 4 is used constantly. Ozone at 0.070 parts per million is about 137 micrograms per cubic meter.

Distinguish primary pollutants, emitted directly, from secondary pollutants, formed in the atmosphere from precursors. Sulfur dioxide, carbon monoxide, lead, and much of the coarse particle load are primary. Ozone is entirely secondary; nothing emits it. Much of fine particulate matter is secondary as well. That distinction is the reason air regulation is hard: you cannot control a secondary pollutant by putting a device on a stack, because no stack emits it.

Key idea: Six criteria pollutants carry national ambient standards, the fine particle annual standard tightened to 9.0 micrograms per cubic meter in 2024, and the two most damaging pollutants, ozone and fine particulate matter, are largely secondary and therefore cannot be captured at a stack.

How ozone forms, and why it misbehaves

Ground-level ozone is formed by sunlight acting on nitrogen oxides and volatile organic compounds. The skeleton of the chemistry is short. Nitrogen dioxide absorbs ultraviolet light and splits, releasing an oxygen atom that combines with molecular oxygen to make ozone. Ozone then reacts with nitric oxide to regenerate nitrogen dioxide and consume itself. Left alone that cycle reaches a steady state with very little ozone in it.

Volatile organic compounds break the cycle open. Their oxidation produces peroxy radicals that convert nitric oxide back to nitrogen dioxide without consuming ozone. Now the loop runs forward with nothing to stop it, and ozone accumulates through the day, peaking in mid to late afternoon downwind of the city that emitted the precursors.

The consequence is a genuinely counterintuitive control problem. The response of ozone to its two precursors is not additive, and it depends on which is in relative shortage.

• In a volatile organic compound limited regime, typical of dense urban cores with heavy traffic, cutting nitrogen oxides can raise local ozone, because nitric oxide had been destroying ozone directly. Cutting organic compounds lowers it.
• In a nitrogen oxide limited regime, typical of suburban and rural downwind areas and of most regions on multi-day scales, cutting nitrogen oxides lowers ozone efficiently.

The field's favorite evidence is the weekend effect: in several major urban areas, ozone was historically higher on weekends than weekdays even though weekend truck traffic and therefore nitrogen oxide emissions were much lower. Less nitric oxide meant less ozone destruction downtown. That result confused regulators for years and it is the reason ozone control is designed on regional photochemical models rather than on intuition.

Fine particles

Particulate matter is classified by aerodynamic diameter because that is what determines where it deposits. Particles above about 10 micrometers are caught in the nose and throat. PM10 reaches the lower airways. PM2.5, at 2.5 micrometers and below, reaches the alveoli, where gas exchange happens and where the body has no mechanical clearance mechanism. Ultrafine particles below 0.1 micrometers can cross into the bloodstream.

Composition matters as much as size. A large fraction of urban PM2.5 is secondary: sulfate formed by oxidation of sulfur dioxide, nitrate formed from nitrogen oxides, ammonium pairing with both from agricultural ammonia, and secondary organic aerosol formed from oxidized hydrocarbons. This is why the acid rain program cut fine particle concentrations across the eastern United States as a side effect, and why agricultural ammonia is now an active regulatory question in Europe and California.

The health evidence for fine particles is among the strongest in environmental epidemiology. The Harvard Six Cities study published in 1993 followed adults in six American cities and found mortality tracking fine particle concentration after controlling for smoking and other factors, and large American Cancer Society cohort analyses replicated the association. Meta-analyses associate long-term exposure with cardiovascular and respiratory mortality and with lung cancer. The World Health Organization attributes on the order of 4.2 million premature deaths a year globally to ambient air pollution, and roughly 6.7 million when household air pollution from cooking fuels is included, which makes air pollution one of the largest environmental health burdens in the world by a wide margin.

Key idea: Ozone responds non-additively to its precursors, so cutting nitrogen oxides can raise urban ozone while lowering it regionally, and fine particles do the most damage because they reach the alveoli and are substantially secondary, formed from sulfur, nitrogen, ammonia, and organic precursors.

Meteorology: the variable nobody controls

Three atmospheric properties set the concentration produced by a given emission rate.

Wind speed dilutes linearly. Double the wind and halve the concentration.

Mixing height is the depth of atmosphere available for vertical dilution. On a sunny afternoon, surface heating drives convection and the mixing height may reach 1,500 to 2,000 meters. On a clear calm night it can collapse below 100.

Atmospheric stability determines whether a rising parcel of air keeps rising. Air cools with height at about 9.8 degrees Celsius per kilometer when it rises without exchanging heat. If the actual atmosphere cools faster than that, a displaced parcel keeps accelerating and the atmosphere is unstable and disperses well. If it cools more slowly, or warms with height, the parcel sinks back and the atmosphere is stable.

When temperature actually increases with height, the condition is a temperature inversion, and it puts a lid on the atmosphere. Two kinds matter. A radiation inversion forms on clear calm nights when the ground radiates heat away and chills the air touching it; it is shallow and usually burns off by mid morning. A subsidence inversion forms when a high pressure system sinks and compressively warms a layer aloft; it can persist for days over a whole region, and it is the killer.

The box model

The simplest quantitative dispersion tool treats a city as a box. Air enters upwind clean, picks up emissions uniformly across the city, mixes through the mixing height, and leaves downwind. At steady state, concentration equals the area emission rate times the along-wind city length, divided by wind speed times mixing height.

Worked example. A city 15 kilometers across, so an area of 225 square kilometers, emits 50 tonnes per day of carbon monoxide. Wind is 2 meters per second and mixing height is 400 meters.

• Area emission rate: 50,000 kilograms per day divided by 2.25 times ten to the eighth square meters equals 2.22 times ten to the minus four kilograms per square meter per day, which divided by 86,400 seconds is 2.57 times ten to the minus nine kilograms per square meter per second.
• Concentration: 2.57 times ten to the minus nine, times 15,000 meters, divided by 2 times 400, equals 3.86 times ten to the minus five divided by 800, or 4.82 times ten to the minus eight kilograms per cubic meter, which is 48.2 micrograms per cubic meter.
• In the units of the standard: 48.2 times 24.45 divided by 1,000 times 28.01 gives 0.042 parts per million, against an eight-hour standard of 9. Comfortable by a factor of over 200, which is a fair picture of modern American urban carbon monoxide.

Now impose a subsidence inversion. Mixing height collapses from 400 meters to 50, and wind falls from 2 meters per second to 0.5. Both terms are in the denominator, so the concentration rises by 400 over 50 times 2 over 0.5, a factor of eight times four, or 32. Carbon monoxide reaches 1.35 parts per million, still below the standard.

Run the same 32-fold multiplier on fine particles and the answer changes character completely. Suppose the same city emits 3 tonnes per day of primary PM2.5. The area emission rate is 1.54 times ten to the minus ten kilograms per square meter per second, and the ordinary-day concentration from local primary sources alone is 2.89 micrograms per cubic meter, unremarkable. Under the inversion it becomes 92 micrograms per cubic meter, nearly three times the 35 microgram 24-hour standard, and that is before any secondary formation or any regional transport.

That calculation is the entire explanation of air pollution episodes. Emissions did not change. The atmosphere did.

History confirms it in the bluntest possible way. In Donora, Pennsylvania, in October 1948, a subsidence inversion trapped emissions from a zinc smelter and steel works in a narrow river valley for five days; roughly 20 people died and something like 6,000 of the town's 14,000 residents became ill. In London in December 1952, a cold, windless, five-day inversion over a city burning soft coal in millions of open grates produced the Great Smog, with contemporaneous estimates of about 4,000 excess deaths and later analyses putting the toll considerably higher. Neither event involved an unusual emission. Both produced landmark legislation, the British Clean Air Act of 1956 and, in the United States, the sequence of federal air statutes that culminated in 1970.

Key idea: A box model shows concentration varying inversely with wind speed and mixing height, so a subsidence inversion multiplying both terms produced a 32-fold increase here, turning a routine 2.89 micrograms per cubic meter of fine particles into 92 with no change in emissions at all.

Stacks and plumes

For a single elevated source the box model is too crude, and the standard tool is the Gaussian plume model, which treats the plume as spreading with a normal distribution in the crosswind and vertical directions, with spread parameters that grow with downwind distance and depend on stability class. You do not need to run one here, but you should know its three lessons.

First, ground-level concentration from an elevated source is zero at the stack, rises to a maximum some distance downwind, and then declines. The worst exposure is never at the fence line.

Second, what matters is the effective stack height, the physical height plus the plume rise from the exhaust's buoyancy and momentum. Hot, fast exhaust climbs well above the stack, and concentration at the ground falls roughly with the square of effective height, which is why raising a stack is such an effective local remedy.

Third, and this is the cautionary part, it is only a local remedy. Through the 1960s and 1970s American and British utilities built very tall stacks to solve ground-level sulfur dioxide problems near power plants. It worked locally and it worked exactly as designed. It also injected sulfur into the free atmosphere where it traveled hundreds of kilometers, oxidized to sulfuric acid, and fell as acid deposition on forests and lakes in New England, eastern Canada, and Scandinavia. Dispersion is not treatment. The tall stack is the clearest case in the field of a solution that relocated a problem across a border, and it is why the 1990 amendments attacked total sulfur mass rather than local concentrations.

Common misconceptions

• Ozone is emitted by cars. Nothing emits ozone. Cars emit nitrogen oxides and volatile organic compounds, which form ozone in sunlight hours later and kilometers downwind.
• Cutting nitrogen oxides always lowers ozone. In volatile organic compound limited urban cores it can raise local ozone, which is why the weekend effect exists and why regional photochemical modeling is required.
• Particulate matter is mostly soot from stacks. Much urban PM2.5 is secondary, formed in the atmosphere from sulfur dioxide, nitrogen oxides, ammonia, and organic vapors.
• Air pollution episodes are caused by spikes in emissions. They are caused by inversions. In the worked case a fixed emission rate produced a 32-fold concentration increase from meteorology alone.
• The highest ground concentration is right at the stack. For an elevated source it is some distance downwind, and it falls roughly with the square of effective stack height.
• A taller stack solves the problem. It solves the local problem and exports the mass, which is precisely how acid deposition was created.

Recap

• Six criteria pollutants carry ambient standards; fine particles tightened to 9.0 micrograms per cubic meter annually in 2024 and ozone stands at 0.070 parts per million over eight hours.
• Ozone is entirely secondary and fine particles are substantially secondary, so neither can be captured at a stack.
• Ozone chemistry runs on nitrogen dioxide photolysis with volatile organic compounds breaking the null cycle, producing volatile organic limited and nitrogen oxide limited regimes and the weekend effect.
• PM2.5 reaches the alveoli, is strongly associated with cardiovascular and respiratory mortality since the Harvard Six Cities study, and ambient air pollution is linked to roughly 4.2 million premature deaths a year worldwide.
• Concentration falls with wind speed and mixing height, and inversions, especially multi-day subsidence inversions, remove both.
• The box model gave 48.2 micrograms per cubic meter of carbon monoxide, or 0.042 parts per million, on an ordinary day for a 15 kilometer city emitting 50 tonnes per day.
• Collapsing mixing height to 50 meters and wind to 0.5 meters per second multiplied concentrations by 32, turning 2.89 micrograms per cubic meter of fine particles into 92, well above the 24-hour standard.
• Donora in 1948 and London in 1952 were inversion events, not emission events, and both produced landmark legislation.
• Ground concentration from a stack peaks downwind and falls with the square of effective stack height, but tall stacks export mass, which is how acid deposition was manufactured.

Sources

1. U.S. Environmental Protection Agency. (n.d.). NAAQS table. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Ground-level ozone basics. epa.gov
3. U.S. Environmental Protection Agency. (n.d.). Particulate matter (PM) basics. epa.gov
4. World Health Organization. (n.d.). Ambient (outdoor) air pollution. who.int
5. Wikipedia. (n.d.). 1948 Donora smog. Wikimedia Foundation. en.wikipedia.org

The big picture

Now the equipment. A gas stream leaving an industrial process carries particles, acid gases, organic vapors, or all three, and the engineer's job is to select and size a device that removes enough of them to satisfy a permit at a cost the plant can carry. The choices are few and each has a characteristic strength, a characteristic weakness, and a characteristic residue.

That last word matters. Every control device in this lesson obeys the mass balance from Lesson 1: it converts an air problem into a solid or a liquid problem. Then we turn indoors, where most people actually breathe, and close with greenhouse gases, where the arithmetic of global warming potential produces a result about wastewater plants that surprises nearly everyone.

Particulate control

Four devices dominate, separating cleanly by particle size and by cost.

Cyclones spin the gas so inertia throws particles to the wall. They have no moving parts, are cheap, tolerate high temperature, and are useless below about 10 micrometers, exactly the size range that matters for health, so their real role is pre-cleaning ahead of a more expensive device.

Electrostatic precipitators charge particles in a corona discharge and collect them on grounded plates, periodically rapped so the dust falls into hoppers. They handle enormous gas volumes at low pressure drop, which made them standard on coal-fired power plants. Their performance follows the Deutsch-Anderson equation: efficiency equals one minus e to the minus the product of drift velocity and collection area divided by gas flow.

Worked example. A precipitator treats 50 cubic meters per second of flue gas with a drift velocity of 0.10 meters per second and 3,000 square meters of plate. The exponent is 0.10 times 3,000 divided by 50, or 6.0. Efficiency is one minus e to the minus 6, or 0.99752, so 99.75 percent.

Now push it. Reaching 99.9 percent needs an exponent of 6.908, or 3,454 square meters, about 15 percent more plate. Reaching 99.99 percent needs 9.21, or 4,605 square meters, 53 percent more than the original. That exponential states a rule running through this whole course: each additional nine of removal costs more than the one before it. It also explains why precipitator performance degrades when ash resistivity changes, which is what happened across the American power fleet after plants switched to low-sulfur western coal.

Fabric filters, universally called baghouses, pass the gas through woven or felted bags, building a dust cake that does most of the actual filtration. They routinely achieve 99.9 percent and better, and unlike precipitators their efficiency holds for submicron particles, which is why they displaced precipitators for fine particle compliance and for mercury control by activated carbon injection. They are sized by air-to-cloth ratio, gas flow divided by cloth area, in meters per minute.

Worked example. The same 50 cubic meters per second is 3,000 cubic meters per minute. At a conservative air-to-cloth ratio of 1.0 meter per minute for a reverse-air design, the required cloth area is 3,000 square meters. At roughly 5 square meters per bag, that is about 600 bags, which is a building. Pulse-jet designs run 2 to 4 meters per minute and are correspondingly smaller. The penalty is pressure drop, typically 1 to 2 kilopascals, which the induced draft fan must overcome every second the plant runs.

Wet scrubbers, particularly venturi designs, accelerate the gas through a throat where water is injected, and the enormous relative velocity drives particles into droplets that are then separated. Efficiency rises with pressure drop, so a high-energy venturi reaches fine particles at a large and permanent fan power cost. They handle sticky, explosive, or hot streams and remove gases at the same time, and they produce a contaminated wastewater stream, converting an air permit problem into a water permit problem.

Device	Effective range	Typical efficiency	Residue produced
Cyclone	Above 10 micrometers	70 to 90 percent	Dry dust
Electrostatic precipitator	Wide, degrades with resistivity changes	99 to 99.9 percent	Dry ash, possibly hazardous
Fabric filter	Down to submicron	99.9 percent and above	Dry dust, plus spent bags
Venturi scrubber	Fine particles at high pressure drop	90 to 99 percent	Contaminated wastewater

Key idea: Cyclones pre-clean, precipitators move huge volumes cheaply, baghouses win on fine particles, and scrubbers handle difficult streams while creating a wastewater; the Deutsch-Anderson exponential means every additional nine of removal costs more plate area than the last.

Gaseous pollutant control

Absorption transfers a gas into a liquid, and its great application is flue gas desulfurization. In a wet limestone scrubber, sulfur dioxide is absorbed into a limestone slurry and oxidized to calcium sulfate dihydrate, which is gypsum.

Worked example. A 500 megawatt coal unit burns 200 tonnes per hour of coal at 2 percent sulfur, so 4 tonnes per hour of sulfur. Converting to sulfur dioxide multiplies by 64.07 over 32.06, giving 8.0 tonnes per hour of sulfur dioxide. At 95 percent removal, 7.6 tonnes per hour are captured, which is 7,600 divided by 64.07, or 118.6 kilomoles per hour. At a stoichiometric ratio of 1.05, limestone consumption is 124.5 kilomoles per hour times 100.09 kilograms per kilomole, or about 12.5 tonnes per hour of limestone. Gypsum production is 118.6 times 172.17, or about 20.4 tonnes per hour.

Run that for 7,000 operating hours a year and the plant consumes roughly 87,000 tonnes of limestone and produces roughly 143,000 tonnes of gypsum. The mass balance is unforgiving and completely visible: a device that keeps sulfur out of the sky creates 143,000 tonnes a year of solids to be sold or landfilled. Much of it goes to wallboard manufacturers, a genuinely good outcome that also made that industry dependent on coal plants now closing.

Adsorption binds molecules onto a solid surface, and activated carbon is the workhorse: a gram of it has hundreds of square meters of internal surface. It is the standard control for volatile organic compounds at low concentration, for mercury injected into flue gas ahead of a baghouse, and, as Lesson 16 shows, for per- and polyfluoroalkyl substances in water. The carbon saturates and must be regenerated or disposed of, which is again a concentrated residue.

Thermal and catalytic oxidation destroy organic vapors rather than capturing them. Thermal oxidizers run at 750 to 1,000 degrees Celsius; catalytic units do the same at 300 to 500 over a precious metal catalyst, saving fuel but vulnerable to poisoning. These are among the few genuine destruction technologies in the field.

Nitrogen oxide control either prevents formation or reduces chemically. Low-nitrogen-oxide burners and staged combustion lower peak flame temperature, since thermal formation is exquisitely temperature-sensitive. Selective catalytic reduction injects ammonia or urea ahead of a catalyst, reducing nitrogen oxides to nitrogen and water at 80 to 90 percent removal; the non-catalytic version works in a narrow high-temperature window, cheaper and less effective. Both risk ammonia slip, unreacted ammonia leaving the stack, which is itself a fine particle precursor.

The three-way catalytic converter deserves special mention as one of the most consequential environmental technologies ever deployed. It simultaneously oxidizes carbon monoxide and hydrocarbons and reduces nitrogen oxides, which is only possible in a narrow window around the stoichiometric air-fuel ratio of about 14.7 to 1. Holding that window requires the oxygen sensor and closed-loop fuel injection every modern car has, which is why emission control and engine management became one system. It also required removing lead from gasoline, since lead poisons the catalyst permanently, and that phaseout produced one of the clearest public health improvements ever measured: average American blood lead levels fell by roughly three quarters between the late 1970s and the early 1990s.

Indoor air quality

Now the part of air quality that regulation barely touches and that dominates actual exposure. Americans spend on the order of 90 percent of their time indoors, and indoor concentrations of many pollutants exceed outdoor concentrations, sometimes by large factors, because sources are close, volumes are small, and ventilation is limited.

Radon is the most serious. It is a radioactive noble gas produced by the decay of uranium in soil and rock, it seeps into buildings through foundation cracks and sump openings, and its decay products lodge in lung tissue. It is the second leading cause of lung cancer in the United States after smoking, associated with roughly 21,000 deaths a year. EPA's action level is 4 picocuries per liter, testing costs very little, and mitigation by sub-slab depressurization, essentially a fan and a pipe drawing soil gas from beneath the slab and venting it above the roof, typically costs a few thousand dollars. Few interventions in environmental engineering have a better ratio of lives saved to money spent, and most homes have never been tested.

Other indoor sources are a short list. Combustion appliances, especially unvented gas stoves and space heaters, produce carbon monoxide, nitrogen dioxide, and fine particles, with measured kitchen nitrogen dioxide during cooking exceeding outdoor ambient standards. Volatile organic compounds and formaldehyde come from building materials, furnishings, adhesives, and cleaning products. Mold following moisture intrusion drives a substantial share of complaints.

The universal control is ventilation, and it can be calculated. At steady state, the indoor concentration equals the outdoor concentration plus the generation rate divided by the ventilation rate.

Worked example. A classroom holds 25 people, and each adult at rest exhales about 0.0052 liters per second of carbon dioxide, so generation is 0.13 liters per second. Ventilate at 5 liters per second per person, which is 125 liters per second. The steady-state rise above outdoor air is 0.13 divided by 125, or 1.04 times ten to the minus three by volume, which is 1,040 parts per million. Add an outdoor background near 420 and the room sits at about 1,460 parts per million.

That number is the reason carbon dioxide monitors became common in schools. Carbon dioxide at those levels is not itself harmful; it is a proxy for how much of the air in the room has already been through somebody's lungs, and therefore for the concentration of everything else people emit, including respiratory aerosols. Double the ventilation to 10 liters per second per person and the rise halves to 520, putting the room near 940, under the usual 1,000 part per million rule of thumb. Where ventilation cannot be increased, filtration helps: a MERV 13 filter in the air handler, or portable units with high-efficiency filters sized by clean air delivery rate, remove particles without bringing in outside air.

Key idea: Most exposure happens indoors, radon alone is associated with about 21,000 American lung cancer deaths a year and is cheaply fixable, and a steady-state ventilation calculation shows a classroom at 5 liters per second per person sitting near 1,460 parts per million of carbon dioxide, which is a proxy for everything else in the air.

Greenhouse gases and the arithmetic that surprises people

The physics of the greenhouse effect belongs to Climate & Environmental Science (ENVS 210). What belongs here is the accounting an environmental engineer actually performs, and it rests on one concept: global warming potential, the warming caused by a mass of a gas over a chosen horizon, conventionally 100 years, relative to the same mass of carbon dioxide. Multiply a mass by its global warming potential and you get carbon dioxide equivalent.

Gas	100-year global warming potential	Typical environmental engineering source
Carbon dioxide	1	Energy use, combustion, incineration
Methane	About 28	Landfills, anaerobic digesters, collection systems, leaks
Nitrous oxide	About 273	Wastewater nitrification and denitrification, combustion, agriculture
Sulfur hexafluoride	About 24,300	Electrical switchgear

Now apply it to the plant we built in Module 4, because the result changes how you would operate it.

Worked example one, methane leakage. Marley City's digesters produce 3,666 cubic meters per day of biogas at 65 percent methane, or 2,383 cubic meters of methane. Suppose 5 percent leaks from seals, pressure relief, and flare inefficiency: 119 cubic meters per day, which at a methane density near 0.67 kilograms per cubic meter is 80 kilograms per day. At a global warming potential of 28, that is 2,236 kilograms of carbon dioxide equivalent per day. Meanwhile the biogas that did get burned generated 8,300 kilowatt-hours per day of electricity, avoiding grid emissions of roughly 3,320 kilograms of carbon dioxide at 0.4 kilograms per kilowatt-hour. A 5 percent leak therefore erases about two thirds of the climate benefit of the entire digestion system.

Worked example two, nitrous oxide. Biological nitrogen removal releases a small fraction of the treated nitrogen as nitrous oxide, particularly under transient or oxygen-limited conditions. The default emission factor in international greenhouse gas inventory guidance is on the order of 1.6 percent of the nitrogen load. Marley City treats 1,800 kilograms of nitrogen per day. At 1.6 percent, 28.8 kilograms per day leaves as nitrous oxide nitrogen, which as nitrous oxide is 28.8 times 44 over 28, or 45.3 kilograms per day. At a global warming potential of 273, that is 12,400 kilograms of carbon dioxide equivalent per day, which dwarfs both the methane leak and the electricity savings and is comparable to the plant's entire energy footprint.

Take the lesson seriously, because it generalizes. When a gas has a global warming potential in the hundreds, a fraction of a percent of a mass flow can dominate a facility's carbon inventory. Anyone who computes a treatment plant's footprint from its electricity bill alone will get the wrong answer, and process control aimed at minimizing nitrous oxide formation, chiefly by avoiding oxygen-limited nitrification and ammonia spikes, is a legitimate climate intervention.

One more area where this profession has direct leverage: refrigerants. The Montreal Protocol of 1987 phased out ozone-depleting chlorofluorocarbons and is the most successful international environmental agreement ever concluded, but the hydrofluorocarbons that replaced them are potent greenhouse gases. The Kigali Amendment of 2016 and, in the United States, the American Innovation and Manufacturing Act of 2020 commit to phasing hydrofluorocarbon production down by 85 percent, and leak detection and recovery on large refrigeration systems is now ordinary environmental compliance work.

Common misconceptions

• Electrostatic precipitators are the best particulate device. They move enormous volumes cheaply but degrade with ash resistivity changes; baghouses beat them on fine particles and now dominate new fine particle compliance.
• Scrubbers are a clean solution. They convert an air problem into a contaminated wastewater stream, and a limestone scrubber produces 143,000 tonnes of gypsum a year at the worked plant.
• Catalytic converters simply burn pollutants. They oxidize carbon monoxide and hydrocarbons while simultaneously reducing nitrogen oxides, which only works near a stoichiometric air-fuel ratio and required leaded gasoline to be eliminated.
• Indoor air is cleaner than outdoor air. Indoor concentrations frequently exceed outdoor ones, and people spend about 90 percent of their time inside.
• Carbon dioxide in a classroom is the health hazard. It is a proxy for ventilation adequacy and for everything else occupants emit, not the hazard itself at those levels.
• A treatment plant's carbon footprint is its electricity use. At the default emission factor, nitrous oxide from nitrogen removal alone can exceed the entire energy footprint.

Recap

• Cyclones pre-clean above 10 micrometers, precipitators handle large volumes, baghouses reach submicron particles, and scrubbers handle difficult streams while producing wastewater.
• A precipitator with a drift velocity of 0.10 meters per second, 3,000 square meters, and 50 cubic meters per second reaches 99.75 percent; 99.99 percent would need 4,605 square meters, 53 percent more.
• A baghouse on the same stream at an air-to-cloth ratio of 1.0 meter per minute needs 3,000 square meters of cloth, roughly 600 bags.
• A 500 megawatt unit on 2 percent sulfur coal makes 8.0 tonnes per hour of sulfur dioxide; 95 percent scrubbing consumes 12.5 tonnes per hour of limestone and produces 20.4 tonnes per hour of gypsum.
• Selective catalytic reduction removes 80 to 90 percent of nitrogen oxides with ammonia, at the risk of ammonia slip; three-way catalysts required unleaded gasoline and blood lead fell about three quarters after the phaseout.
• Radon is associated with roughly 21,000 American lung cancer deaths a year, has an action level of 4 picocuries per liter, and is mitigated cheaply by sub-slab depressurization.
• A classroom of 25 at 5 liters per second per person reaches about 1,460 parts per million of carbon dioxide; doubling ventilation brings it to about 940.
• Methane has a global warming potential near 28 and nitrous oxide near 273, so a 5 percent digester gas leak erased two thirds of the digestion benefit and nitrous oxide at a 1.6 percent emission factor produced 12,400 kilograms of carbon dioxide equivalent per day.
• The Montreal Protocol succeeded against ozone depletion, and the Kigali Amendment and the 2020 American Innovation and Manufacturing Act now phase down the hydrofluorocarbons that replaced the original refrigerants.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Air pollution control technology fact sheets. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Health risk of radon. epa.gov
3. U.S. Environmental Protection Agency. (n.d.). Understanding global warming potentials. epa.gov
4. Centers for Disease Control and Prevention. (n.d.). Radon in the home. U.S. Department of Health and Human Services. cdc.gov
5. Wikipedia. (n.d.). Flue-gas desulfurization. Wikimedia Foundation. en.wikipedia.org

The big picture

Everything so far has been about keeping contaminants out of water and air. This module is about the material left over when nothing else applies: the mass that has to be put somewhere, permanently, on land. It is the least glamorous part of environmental engineering and it is the part that touches every household every week.

The numbers set the scale. EPA's most recent complete national accounting, for 2018, put United States municipal solid waste generation at about 292 million tons, roughly 4.9 pounds per person per day, which is about 2.2 kilograms. Of that, about 32 percent was recycled or composted, about 12 percent was combusted with energy recovery, and about 50 percent went to landfill. Marley City, at 120,000 people and 2.2 kilograms per person per day, generates 264 tonnes of municipal solid waste per day. This lesson follows all of it.

Component	Approximate share of generation	Engineering significance
Paper and paperboard	23 percent	High heating value, strong market for cardboard, weak for mixed paper
Food waste	22 percent	Generates landfill methane and leachate strength; divertible to composting or digestion
Plastics	12 percent	High heating value, very low actual recycling rate, persistent in landfill
Yard trimmings	12 percent	Readily composted and banned from landfill in many states
Metals	9 percent	Highest commodity value, especially aluminum
Rubber, leather, textiles	9 percent	Growing fast, with almost no recovery
Wood	6 percent	Combustible, though treated wood raises a hazardous waste question
Glass	4 percent	Heavy, abrasive, low value, and contaminates paper when broken

The solid waste management hierarchy orders the options: source reduction first, then reuse, then recycling and composting, then energy recovery, then landfill. It is a useful ordering and it is also routinely misapplied, because it ranks options by general preference rather than by the life-cycle outcome of a specific material in a specific place. Lesson 16 returns to that.

The modern landfill

An engineered landfill has almost nothing in common with a dump. Federal criteria under RCRA Subtitle D, in force since the early 1990s, closed tens of thousands of open dumps and small unlined sites and replaced them with a smaller number of large regional facilities that are, in engineering terms, containment structures with instrumentation.

Siting comes first. A landfill may not be located near an airport runway where birds would endanger aircraft, in a floodplain, in wetlands, near a Holocene fault, in a seismic impact zone, or in unstable ground, absent specific demonstrations. Those restrictions alone eliminate most of any given county.

The base is a composite liner, and the word composite is the whole idea: two dissimilar barriers in intimate contact, so that a defect in one is not a defect in the other. The federal minimum is a flexible membrane liner at least 1.5 millimeters thick, in practice high-density polyethylene, laid directly on at least 0.6 meters of compacted clay with a hydraulic conductivity no greater than ten to the minus seven centimeters per second. Liquid penetrating a pinhole in the geomembrane then meets clay that transmits it at a few centimeters per year.

Above the liner sits the leachate collection system: a granular drainage layer, perforated pipes on a graded subgrade, and a sump with pumps. Its performance requirement is a head limit: no more than 30 centimeters of leachate may stand on the liner. That matters because flow through a defect scales with the head driving it, so keeping the liner drained is as important as the liner itself.

Waste is placed in cells, compacted by heavy steel-wheeled machines, and covered daily with about 15 centimeters of soil to control odor, litter, vectors, and fire. When an area reaches final grade it receives a final cover, a low-permeability barrier with drainage and vegetated soil above, designed to shed water so infiltration and therefore leachate generation collapse. Groundwater monitoring wells ring the site upgradient and downgradient. After closure the owner must maintain cover, leachate system, gas system, and monitoring for 30 years, with posted financial assurance to guarantee it.

Worked example: capacity and life. Marley City generates 264 tonnes per day and diverts 32 percent, so 180 tonnes per day reach the landfill, or 65,700 tonnes per year.

• At a compacted in-place density of 700 kilograms per cubic meter, that occupies 65,700,000 divided by 700, or 93,857 cubic meters per year of waste volume.
• Daily and intermediate cover typically consume about 20 percent additional airspace, bringing the total to about 112,600 cubic meters per year.
• A cell 400 meters by 400 meters, excavated and mounded to an average 20 meters of waste thickness, provides 3,200,000 cubic meters.
• Life: 3,200,000 divided by 112,600 equals 28.4 years.

Notice how much leverage the density and diversion figures carry. Raising compaction from 700 to 850 kilograms per cubic meter, achievable with better equipment and more passes, extends that life by about four years at no capital cost, which is why landfill operators care intensely about compaction. And every tonne diverted upstream is airspace not consumed: at 45 percent diversion instead of 32, the site would last about 35 years.

Worked example: leachate. Leachate generation is a water balance on the surface. Take annual precipitation of 900 millimeters, an open active area of 5 hectares, which is 50,000 square meters, and an infiltration fraction of 30 percent, the rest running off or evaporating. Infiltration is 0.9 times 0.30, or 0.27 meters per year, over 50,000 square meters, giving 13,500 cubic meters per year, or about 37 cubic meters per day. Once an area is capped the infiltration fraction falls to a few percent, which is the main reason final cover earns its cost.

Leachate is difficult water. Young leachate in a landfill's acid phase can carry BOD of 2,000 to 20,000 milligrams per liter and ammonia in the hundreds to low thousands. As the site matures the organic strength falls but the ammonia does not, and it is ammonia that dominates leachate management for decades. Options are hauling or piping it to the municipal wastewater plant, which needs a pretreatment permit under the program from Lesson 10 and can upset that plant if the ammonia load is large relative to it, treating it on site biologically, or recirculating it through the waste in a bioreactor landfill.

Key idea: A modern landfill is a containment structure with a composite liner, a leachate collection system limited to 30 centimeters of head, daily and final cover, gas control, groundwater monitoring, and 30 years of funded post-closure care; Marley City's 180 tonnes per day fills a 3.2 million cubic meter cell in 28 years and generates about 37 cubic meters per day of leachate.

Landfill gas

Buried organic matter decomposes anaerobically, exactly as it does in the digesters from Lesson 12 but far more slowly and with no temperature control. The product is landfill gas, roughly half methane and half carbon dioxide with trace organics, and it presents three problems at once: it is explosive between about 5 and 15 percent methane in air and migrates laterally through soil into basements, it smells, and it is a potent greenhouse gas.

Worked example. A standard first-order model treats each tonne of waste as yielding a total methane potential, commonly taken near 100 cubic meters per tonne, released at a rate constant near 0.04 per year. For a site filled steadily at 65,700 tonnes per year, generation after 28 years approaches 100 times 65,700 times one minus e to the minus 1.12, which is 100 times 65,700 times 0.674, or about 4.43 million cubic meters of methane per year, roughly 12,100 cubic meters per day.

Burned, that is 12,100 times 35.8 megajoules, or about 433,000 megajoules per day, which is 5.0 megawatts of thermal power. Through a reciprocating engine at about 33 percent efficiency it becomes roughly 1.65 megawatts of electricity, continuously, for decades. Landfill gas to energy is a mature business for exactly this reason.

Now the climate arithmetic from Lesson 14, which is the part that matters most. Gas collection systems are good but not perfect; collection efficiencies of 60 to 85 percent are typical over a site's life. At 75 percent collection, 25 percent of 4.43 million cubic meters, or 1.11 million cubic meters of methane, escapes each year. At 0.67 kilograms per cubic meter that is 743 tonnes of methane, and at a global warming potential of 28 it is about 20,800 tonnes of carbon dioxide equivalent per year from one mid-size landfill. Landfills are among the largest anthropogenic methane sources in the United States, on the order of one seventh of national methane emissions, and improving collection efficiency and diverting food waste are two of the cheapest available climate interventions in this field.

Waste to energy

Combustion reduces waste volume by about 90 percent and mass by about 75 percent, and it recovers energy. Roughly 60 to 75 facilities operate in the United States, handling about 12 percent of municipal solid waste, a much smaller share than in Japan or in several northern European countries where land is scarce.

Worked example. Municipal solid waste has a heating value near 10,500 kilojoules per kilogram, about a third of coal. Marley City's 180 tonnes per day therefore carries 180 times 10,500 megajoules per tonne, or 1,890,000 megajoules per day, which is 21.9 megawatts of thermal power. Net electrical efficiency at a mass burn plant is only about 20 percent, giving roughly 4.4 megawatts of electricity, or 105,000 kilowatt-hours a day, enough for a few thousand homes.

Why only 20 percent, when a modern power plant reaches 40 or more? Because the flue gas contains chlorine from plastics, and hot chloride attacks superheater tubes, so steam temperature must be held well below what thermodynamics would otherwise allow. That is a materials constraint setting a thermodynamic ceiling, and it is a clear case of a contaminant in the feed limiting an entire process.

The control train on a modern combustor is the full sequence from Lesson 14: selective non-catalytic reduction for nitrogen oxides, a dry or semi-dry scrubber for acid gases, activated carbon injection for mercury and dioxins, and a baghouse, all under continuous emissions monitoring. It works. Dioxin emissions from American municipal waste combustors fell by well over 99 percent after the standards of the 1990s, moving the sector from dominant national dioxin source to a minor one, which is among the clearest technology-forcing successes in the Clean Air Act's record.

Two residues remain. Bottom ash, roughly 10 percent of input mass, is usually non-hazardous and used as aggregate in several countries. Fly ash, around 3 percent, concentrates metals and requires stabilization; in the United States the two are commonly combined and managed as non-hazardous after passing the leaching test described in the next lesson, a practice both defensible and genuinely contested.

The siting objection is real and belongs in the open. Waste to energy facilities and transfer stations are disproportionately located in low-income communities and communities of color, and Chester, Pennsylvania, a small majority-Black city hosting one of the country's largest incinerators and receiving waste from several states, is the case most often cited. Apply the framework from Lesson 3: this is a cumulative impact and siting alternatives question, and it is an engineering question.

Recycling, honestly

Recycling is where public belief and engineering reality diverge most sharply, so here it is plainly.

A material recovery facility is a factory. It sorts commingled recyclables with screens, magnets, eddy current separators for aluminum, optical sorters, and human pickers, and it costs real money per tonne to run, commonly 70 to 100 dollars. Whether recycling makes or loses money depends entirely on what the sorted commodities sell for, and those markets are violently volatile.

Material	Rough commodity value	Reality
Aluminum cans	Very high, often above 1,000 dollars per tonne	Recycling uses about 5 percent of the energy of primary production; permanently worth doing
Cardboard	Moderate and volatile	Strong markets and high recovery; the workhorse of the industry
Steel cans	Moderate	Separated magnetically and reliably recycled
PET and natural HDPE bottles	Moderate	The only plastics with dependable end markets
Mixed paper	Near zero, sometimes negative	Collapsed after 2018 and has not fully recovered
Glass	Usually negative	Heavy to haul, abrasive on equipment, and when broken it contaminates paper
Mixed plastics 3 through 7	Negative	Collected in many programs, landfilled or burned in most of them

Two events shaped the current situation. Single-stream collection, everything in one cart, spread through the 2000s because it roughly doubles participation. It also roughly doubles contamination, which now runs 17 to 25 percent, and broken glass in a single-stream cart is much of what makes mixed paper hard to sell. Then in 2018 China implemented its National Sword policy, restricting imports of recovered paper and plastics under a contamination limit of 0.5 percent that essentially no American single-stream bale could meet. The market that had absorbed a large fraction of United States recovered material closed, municipalities that had been paid for recyclables began paying to have them taken, and some suspended programs entirely.

Be precise about plastics, because this is where wishful thinking is most expensive. The United States plastic recycling rate is on the order of 5 to 9 percent. The resin identification code stamped inside the triangle is an identification symbol, not a claim that the item is recyclable in your program. Putting doubtful items in the cart hoping they will be handled, which the industry calls wishcycling, raises contamination and makes the whole bale less marketable, so it is worse than throwing them away.

None of that means recycling fails. It means the honest picture is material-specific: aluminum, steel, cardboard, and clear PET work economically and environmentally, glass works where collected separately and used locally, and mixed plastics do not work. The interventions with demonstrated results are deposit return systems, which recover 60 to 90 percent of containers against roughly 20 to 30 percent in curbside-only jurisdictions and produce clean uncontaminated material, and extended producer responsibility, which shifts end-of-life cost onto the producers who choose the packaging. The latter has operated across much of Europe for decades and has been enacted in several American states since 2021, and it changes the incentive at the one point in the chain where packaging design is actually decided.

Finally, food waste. It is 22 percent of the stream, the single largest driver of landfill methane and leachate strength, and straightforwardly divertible to composting or to anaerobic digestion, including co-digestion at the wastewater plant built in Lesson 12. Several states have banned commercial food waste from landfill, and of everything here food diversion has the best ratio of environmental benefit to technical difficulty.

Common misconceptions

• Landfills are holes where garbage rots away. A lined landfill is a dry containment structure designed to limit water entry, and buried waste degrades very slowly; newspapers have been excavated legible after decades.
• The liner is the barrier. The liner is half of a composite system, and keeping leachate head below 30 centimeters matters as much as the membrane, since flow through any defect scales with driving head.
• Landfill gas is mostly carbon dioxide, so it is harmless. It is about half methane, explosive between 5 and 15 percent in air, and at 75 percent collection this one site still leaks about 20,800 tonnes of carbon dioxide equivalent a year.
• Incinerators are dirty by nature. Dioxin emissions from American municipal waste combustors fell by more than 99 percent after the standards of the 1990s; the live objections now are siting equity, ash, and carbon.
• If it carries a recycling symbol it gets recycled. The resin code identifies the polymer. Only PET and natural HDPE have dependable markets, and the national plastic recycling rate is roughly 5 to 9 percent.
• Putting doubtful items in the cart beats trashing them. Wishcycling raises contamination and can make an entire bale unsalable.

Recap

• United States generation is about 292 million tons a year, roughly 2.2 kilograms per person per day, about 32 percent recycled or composted, 12 percent combusted, and 50 percent landfilled.
• Marley City generates 264 tonnes per day and landfills 180 after diversion.
• Subtitle D landfills require siting restrictions, a composite liner of 1.5 millimeter geomembrane over 0.6 meters of clay at ten to the minus seven centimeters per second, leachate collection limited to 30 centimeters of head, cover, gas control, groundwater monitoring, and 30 years of funded post-closure care.
• At 700 kilograms per cubic meter and 20 percent cover airspace, Marley City consumes 112,600 cubic meters a year and a 3.2 million cubic meter cell lasts 28.4 years.
• A water balance with 900 millimeters of rain, 5 hectares open, and 30 percent infiltration gives 37 cubic meters per day of leachate, whose long-term problem is ammonia rather than organics.
• Gas generation reaches about 12,100 cubic meters of methane per day, worth about 1.65 megawatts of electricity, while a 25 percent collection shortfall still emits about 20,800 tonnes of carbon dioxide equivalent a year.
• Waste to energy at 10,500 kilojoules per kilogram gives 21.9 megawatts thermal and about 4.4 electrical, limited to 20 percent efficiency by chloride corrosion of superheater tubes.
• Dioxin emissions from American combustors fell over 99 percent after the 1990s standards, while siting equity, ash, and carbon remain genuine objections.
• Recycling economics are material-specific: aluminum, steel, cardboard, and clear PET work while mixed paper, glass, and mixed plastics generally do not, and contamination of 17 to 25 percent plus China's 2018 National Sword restrictions reshaped the industry.
• Deposit return systems recover 60 to 90 percent of containers, extended producer responsibility moves cost to the packaging decision, and food waste diversion has the best benefit-to-difficulty ratio available.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Facts and figures about materials, waste and recycling. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Municipal solid waste landfills. epa.gov
3. U.S. Environmental Protection Agency. (n.d.). Landfill Methane Outreach Program. epa.gov
4. U.S. Environmental Protection Agency. (n.d.). Energy recovery from the combustion of municipal solid waste. epa.gov
5. Wikipedia. (n.d.). Landfill gas. Wikimedia Foundation. en.wikipedia.org

The big picture

This last lesson covers the material nobody wants: waste dangerous enough to need its own legal regime, ground that somebody already ruined, and a family of chemicals that will outlast everyone reading this. It closes with the profession itself, because after fifteen lessons of processes and numbers you should know what the job actually is and how you become licensed to sign for it.

Keep one thread in view. Every technique here is an application of the mass balance from Lesson 1 and the risk framework from Lesson 3. Nothing is destroyed unless a bond is broken. Nothing is a hazard unless a pathway connects it to a receptor. Those two sentences will carry you through the whole subject.

Is it hazardous? The RCRA determination

Under RCRA Subtitle C, a generator must determine whether its waste is hazardous, and the determination follows a fixed sequence. Is it a solid waste at all, which in the statute's odd usage includes liquids and containerized gases? Is it specifically excluded, as household waste, domestic sewage, and certain mining and fossil fuel wastes are? Then it is hazardous if it is either listed or characteristic.

Listed wastes appear on one of four lists: F for non-specific sources such as spent solvents, K for specific industrial processes, and P and U for discarded commercial chemical products, P being the acutely hazardous ones. A listed waste is hazardous regardless of concentration and drags its listing along even when mixed with something else, which is why the field is so insistent about not combining waste streams.

Characteristic wastes behave dangerously, and there are four characteristics: ignitability, broadly a flash point below 60 degrees Celsius; corrosivity, pH at or below 2 or at or above 12.5; reactivity, meaning explosive or generating toxic gas; and toxicity, determined by a leaching test. That test is the source of the most common misunderstanding in this subject. The toxicity characteristic leaching procedure extracts a milled sample with a mildly acidic solution meant to simulate landfill conditions, then measures 40 regulated constituents in the extract against fixed limits, for example 5.0 milligrams per liter for lead and arsenic, 1.0 for cadmium, 0.2 for mercury, and 0.5 for benzene.

Worked example. A demolition contractor has soil containing 800 milligrams per kilogram of total lead. Is it a hazardous waste? You cannot tell. Total concentration is not the criterion. Run the leaching procedure, and suppose the extract measures 3.2 milligrams per liter of lead, below the 5.0 limit. The soil is not a characteristic hazardous waste, even at 800 milligrams per kilogram, because what the rule regulates is the tendency to leach into groundwater, not the amount present. Change the soil chemistry, or run the same test on a soil at 300 milligrams per kilogram of highly soluble lead, and the answer can reverse. Engineers routinely stabilize a waste with cement or phosphate specifically to fail the leaching test in the good direction, which is a legitimate and widely used treatment strategy called solidification and stabilization.

Once a waste is hazardous, cradle-to-grave applies: an identification number, accumulation time limits that depend on the generator's monthly quantity category, a manifest tracking each shipment, shipment only to a permitted treatment, storage, and disposal facility, and biennial reporting. The land disposal restrictions added in 1984 forbid landfilling hazardous waste until it has been treated to specified standards, a rule that transformed the industry by making treatment cheaper than disposal, and the corrective action provisions require permitted facilities to clean up their own past releases.

Key idea: A waste is hazardous if it is listed or if it exhibits ignitability, corrosivity, reactivity, or toxicity, and toxicity is decided by a leaching test rather than by total concentration, which is why 800 milligrams per kilogram of lead can be non-hazardous and why stabilization is a real treatment.

Love Canal and the machinery it created

Between 1942 and 1953 the Hooker Chemical Company disposed of roughly 21,000 tons of chemical waste in an abandoned, partly excavated canal near Niagara Falls, New York, then capped it with clay. In 1953 it sold the land to the local school board for one dollar, with a deed provision disclaiming liability and noting the waste. A school and a residential neighborhood were built on and around it.

Through the 1970s, after unusually wet years raised the water table, chemicals surfaced in basements, in the schoolyard, and in surface pools. Residents documented health complaints and investigations confirmed contamination. In August 1978 the New York State Health Commissioner declared a health emergency and recommended evacuating pregnant women and children under two, and President Carter declared a federal emergency, the first ever for a non-natural disaster. A second declaration in 1980 brought the total relocated to roughly 950 families. In December 1980, Congress passed CERCLA. The connection is that direct.

Notice what the case actually teaches an engineer: the disposal was, by the standards of its era, not obviously improper, the clay cap was a real engineering measure, the deed disclosed the waste, and it still ended in a neighborhood being abandoned. Long-lived hazards outlast the institutions that create them, which is why CERCLA liability is retroactive and why post-closure care is measured in decades.

The machinery CERCLA built runs in a fixed order: preliminary assessment and site inspection, scoring under the Hazard Ranking System, listing on the National Priorities List, a remedial investigation to characterize the site and a feasibility study to compare alternatives, a proposed plan with public comment, a Record of Decision selecting the remedy, design and construction, long-term operation, and five-year reviews forever whenever waste is left in place. Roughly 1,340 sites are currently listed and a few hundred have been deleted since 1983. The parallel brownfields program handles the far more numerous lightly contaminated properties under liability protections created in 2002, and it is where most contaminated-site engineers actually spend their careers.

Investigating a site

Investigation is staged so money is spent in proportion to what is known. A Phase I environmental site assessment involves no sampling at all: historical maps, aerial photographs, regulatory databases, title records, interviews, and a walk of the property. It identifies recognized environmental conditions and, in a transaction, establishes the all appropriate inquiries supporting a purchaser's liability defense. A Phase II assessment collects data: soil borings, monitoring wells, soil gas surveys, and laboratory analysis.

Everything then gets organized into a conceptual site model, a chain running source, release mechanism, transport medium, exposure point, exposure route, receptor. Its power is that a risk requires every link. Break any one and the pathway is incomplete. That is why a paved cap, a fence, a deed restriction, or extending a municipal water main can be a legitimate remedy even though the contamination remains: they sever the pathway rather than remove the mass. It is also why an engineer must be able to say honestly which links a proposed remedy actually breaks.

Two transport facts drive most decisions. Dissolved contaminants move with groundwater but slower, because sorption onto aquifer solids retards them by a factor that can be two, ten, or a hundred depending on the compound and the aquifer's organic carbon. More consequential is the distinction between light and dense non-aqueous phase liquids. Gasoline and fuel oil are lighter than water and float on the water table, where they are relatively accessible. Chlorinated solvents such as trichloroethylene and tetrachloroethylene are denser: released in quantity they sink through the aquifer, pool on low-permeability layers, and dissolve slowly for decades. A dense phase pool is a source that keeps regenerating the plume, and failing to recognize one is the most common reason a cleanup runs thirty years without finishing.

Remediation, and the honest arithmetic of pump and treat

The technology menu is short and each entry has a proper use.

• Pump and treat extracts contaminated groundwater, treats it above ground, and discharges or reinjects it.
• Soil vapor extraction and air sparging pull volatile compounds from the unsaturated zone and strip them from groundwater; excellent for volatile organics, useless for metals.
• In situ chemical oxidation injects permanganate, persulfate, or peroxide to destroy organics in place.
• Bioremediation stimulates microorganisms: aerobic degradation for petroleum, and for chlorinated solvents reductive dechlorination, in which specific bacteria strip chlorine atoms one at a time, requiring an injected electron donor such as emulsified vegetable oil or lactate.
• Permeable reactive barriers place a treatment wall, classically granular zero-valent iron, across the plume path so groundwater treats itself as it flows through.
• Thermal methods heat the subsurface with steam or electrical resistance, expensive and effective on dense phase source zones.
• Containment uses slurry walls, sheet piling, and caps to stop movement; excavation digs the material out and hauls it to a permitted facility.
• Monitored natural attenuation relies on documented natural degradation, with sampling to prove it is actually happening.

Worked example, the one that changed the field's expectations. A trichloroethylene plume 150 meters wide sits in a sand aquifer 12 meters thick with a hydraulic conductivity of 30 meters per day and a hydraulic gradient of 0.004.

• Darcy flux: 30 times 0.004 equals 0.12 meters per day.
• Groundwater flow through the plume cross-section: 0.12 times 12 meters times 150 meters equals 216 cubic meters per day, which is 2.5 liters per second. That is the minimum extraction rate to capture the plume hydraulically, and hydraulic containment is a genuine achievement: nothing moves downgradient any more.
• Now the mass. At 500 micrograms per liter, which is 0.5 grams per cubic meter, extraction removes 216 times 0.5, or 108 grams per day, which is 39 kilograms per year.
• Suppose 2,000 kilograms of trichloroethylene is present in the aquifer, most of it sorbed to solids or held as dense phase liquid. At 39 kilograms a year, cleanup takes 51 years, and that is the optimistic case.

The real behavior is worse and is worth naming precisely. Concentrations fall quickly at first as readily mobile mass is removed, then flatten into a long tail as the remaining mass desorbs slowly from solids and dissolves slowly from the dense phase. If pumping stops, concentrations rebound as equilibrium re-establishes. Reviews of American pump-and-treat systems have repeatedly found that few reached their cleanup goals within their projected timeframes. The lesson is not that pump and treat is useless. It is that it is a containment technology being asked to do mass removal, and the modern approach is to attack the source zone aggressively with thermal, chemical oxidation, or excavation, then use pumping or a reactive barrier to manage what remains.

Key idea: Capturing a 150 meter plume in this aquifer takes 216 cubic meters per day, which contains it completely, but at 500 micrograms per liter it removes only 39 kilograms a year against 2,000 kilograms present, so pump and treat contains well and cleans slowly, and source zones must be attacked directly.

PFAS: the current frontier

Per- and polyfluoroalkyl substances are a family of thousands of synthetic compounds built around chains of carbon atoms saturated with fluorine. The carbon-fluorine bond, at roughly 485 kilojoules per mole, is the strongest single bond in organic chemistry, and it is why these compounds resist heat, acid, base, oxidation, and biological degradation almost completely. That property is exactly what made them useful in non-stick coatings, stain and water repellents, food packaging, chrome plating, and above all in aqueous film-forming firefighting foam at airports, refineries, and military bases, and it is exactly what makes them a permanent environmental problem. The popular name, forever chemicals, is chemically accurate.

They are essentially everywhere. National biomonitoring has detected PFAS in the blood of the great majority of Americans, and nationwide sampling has found them in drinking water supplies serving a large share of the population. The evidence associates several of them with kidney and testicular cancer, thyroid disease, elevated cholesterol, developmental effects, and suppressed immune response including reduced antibody response to vaccination.

In April 2024 EPA finalized the first national drinking water standards for the group. Perfluorooctanoic acid and perfluorooctane sulfonic acid each carry a maximum contaminant level of 4.0 nanograms per liter with health goals of zero, three more compounds are limited at 10 nanograms per liter, and mixtures are handled with a hazard index. Register that number in the units from Lesson 4. Four nanograms per liter is 4 parts per trillion, or 0.004 micrograms per liter, against a lead action level of 10 micrograms per liter. The standard is roughly two and a half thousand times lower, and it sits at the edge of what commercial laboratories can reliably measure. In May 2024 EPA separately designated those two compounds as CERCLA hazardous substances, bringing the full Superfund liability machinery of this lesson to bear on them.

Treatment is where the mass balance bites hardest. Three technologies work at drinking water scale: granular activated carbon, anion exchange resin, and reverse osmosis or nanofiltration. All three are separation processes. None destroys anything. Each produces a concentrated residual, spent carbon, spent resin, or reject brine, each now a concentrated PFAS waste needing disposal, and each less effective against the short-chain compounds replacing the long-chain ones. Destruction technologies exist but are immature: high-temperature incineration is used and its completeness is contested, while supercritical water oxidation and plasma methods are promising and not yet routine. Compliance costs run to billions nationally, and utilities that did nothing wrong are recovering part of it through large settlements with manufacturers. PFAS is the clearest current illustration of a principle worth carrying out of this course: the property that makes a chemical useful is often the same property that makes it a permanent problem.

Life-cycle thinking

This course has treated pollution as something to be captured at a point. The alternative frame asks whether it needed to be created. Life cycle assessment, standardized in ISO 14040 and 14044, evaluates burdens across raw material extraction, manufacturing, distribution, use, and end of life, in four steps: define goal and scope including a functional unit, compile an inventory, translate it into impact categories, and interpret.

The functional unit does most of the work and is where analyses go wrong. Comparing a ceramic mug with a paper cup is meaningless until you state the function, something like delivering 500 servings of coffee. Do that honestly and the answer becomes conditional: the mug carries a large manufacturing burden and wins only after enough uses, with the crossover depending heavily on how the washing is done. That conditional answer is the correct one, and an engineer's job is to present it rather than pick a headline.

The related discipline is pollution prevention, given a statutory hierarchy in the 1990 Pollution Prevention Act: prevent at the source, then reduce, then recycle, then treat, and only then dispose. Notice how much of this course has lived at the treatment step, and that the steps above it are cheaper wherever they are available. Notice too that treatment has its own footprint: water and wastewater systems consume roughly 2 percent of United States electricity, so a plant that treats more thoroughly emits more carbon somewhere else. That tension is genuine and does not resolve neatly.

The profession

Environmental engineers number roughly 44,000 in the United States, with a median annual wage around 100,000 dollars in the early 2020s and modest projected growth. About half work in consulting, selling investigation and design services; the rest are split among government at every level, utilities, and industry.

The credential that matters is the professional engineer license, and the path is standard. Earn an ABET-accredited engineering degree. Pass the Fundamentals of Engineering examination, usually in the final year, which makes you an engineer in training. Work roughly four years under a licensed engineer. Pass the Principles and Practice of Engineering examination in environmental engineering. Apply to a state board, which licenses you in that state, and maintain continuing education after that. What the license buys is legal authority and legal responsibility: only a licensed engineer may seal drawings for public works, and the seal means a named individual with a license to lose has judged the design adequate. Beyond licensure, the American Academy of Environmental Engineers and Scientists offers board certification as a specialty credential.

The ethical obligation is stated first in every engineering code: hold paramount the safety, health, and welfare of the public. In this field that clause is not decorative. Reread Lesson 9. In Flint the technical facts were knowable, and the people who eventually made them public were an academic engineer and a pediatrician working outside the institutions responsible. Somebody with the relevant knowledge is usually inside the room first. Being that person, and being willing to write it down and say it out loud, is the actual content of professional responsibility, and it is the last thing this course has to teach you.

Common misconceptions

• High total concentration makes a waste hazardous. For the toxicity characteristic it is the leaching test that decides, which is why 800 milligrams per kilogram of lead can be non-hazardous and why stabilization is a treatment.
• Superfund cleanups remove all the contamination. Many remedies break the exposure pathway rather than remove mass, which is why five-year reviews continue indefinitely wherever waste is left in place.
• Pump and treat cleans an aquifer. It contains a plume reliably and removes mass slowly, 39 kilograms a year against 2,000 present in the worked case, with tailing and rebound making the real performance worse.
• Chlorinated solvents float like gasoline. They are denser than water, sink through the aquifer, and pool on low-permeability layers as a source that keeps regenerating the plume.
• PFAS treatment destroys the compounds. Carbon, ion exchange, and membranes separate and concentrate them into a residual that still needs disposal; destruction technologies remain immature.
• A life cycle assessment gives a clean verdict. The answer depends on the functional unit and the use assumptions, and presenting the conditional result honestly is the engineer's job.

Recap

• A waste is hazardous under RCRA if it is listed on the F, K, P, or U lists or exhibits ignitability, corrosivity, reactivity, or toxicity, and toxicity is judged by a leaching test rather than total concentration.
• Cradle-to-grave tracking, land disposal restrictions, and corrective action are the operative obligations once a waste is hazardous.
• Love Canal, where about 21,000 tons of waste was buried between 1942 and 1953 and a neighborhood was later evacuated, produced CERCLA in December 1980.
• The Superfund process runs from assessment and Hazard Ranking through the National Priorities List, remedial investigation and feasibility study, Record of Decision, construction, and five-year reviews; roughly 1,340 sites are currently listed.
• A conceptual site model is a chain from source to receptor, and breaking any single link eliminates the pathway, which is why caps, fences, and water line extensions are legitimate remedies.
• Dense non-aqueous phase liquids such as chlorinated solvents sink and pool, acting as long-term sources, while petroleum floats on the water table.
• Capturing the worked plume takes 216 cubic meters per day and removes 39 kilograms a year against 2,000 kilograms present, so source zones must be attacked with thermal, chemical, or excavation methods.
• PFAS resist degradation because of the carbon-fluorine bond at about 485 kilojoules per mole, and the 2024 drinking water standards set 4 nanograms per liter for two compounds, roughly 4 parts per trillion.
• Carbon, ion exchange, and membranes separate PFAS into concentrated residuals rather than destroying them.
• Life cycle assessment depends on the functional unit, the pollution prevention hierarchy puts source reduction ahead of treatment, and water systems use about 2 percent of United States electricity.
• Licensure runs from an ABET degree through the Fundamentals of Engineering exam, about four years of supervised practice, and the environmental Principles and Practice exam, and the seal carries personal legal responsibility.

Sources

1. U.S. Environmental Protection Agency. (n.d.). Hazardous waste identification and RCRA overview. epa.gov
2. U.S. Environmental Protection Agency. (n.d.). Superfund cleanup process. epa.gov
3. U.S. Environmental Protection Agency. (n.d.). Per- and polyfluoroalkyl substances (PFAS). epa.gov
4. Agency for Toxic Substances and Disease Registry. (n.d.). PFAS and your health. Centers for Disease Control and Prevention. atsdr.cdc.gov
5. U.S. Bureau of Labor Statistics. (n.d.). Occupational Outlook Handbook: environmental engineers. U.S. Department of Labor. bls.gov
6. Wikipedia. (n.d.). Love Canal. Wikimedia Foundation. en.wikipedia.org