Your Landfill Leachate Treatment Plant Was Designed for Water That No Longer Exists

Most landfill leachate treatment plants are designed to treat water that stops existing about three years after commissioning. I don't mean that as a figure of speech. Leachate chemistry moves as a landfill ages, it keeps moving for decades after the site closes, and standard practice is to size a treatment train around one characterization study and then behave as if that study is a permanent fact. It isn't. That single assumption is the most expensive one in leachate management, and almost everyone makes it.

So the argument here is narrow. The technology mostly works. Reverse osmosis works, biological treatment works, recirculation does what it says on the tin. What doesn't work is treating any of them as a fixed answer to a moving input. Leachate is a data problem before it's a treatment problem, and the sites that treat it that way are the ones whose discharge permits survive contact with year seven.

Leachate won't hold still

Start with what's actually in the pipe. Young leachate, from a cell that's still filling, is acidic and loaded with readily biodegradable organics; the BOD-to-COD ratio sits above 0.5, which is a polite way of saying the bugs will eat it. That figure and the aging picture come from Kjeldsen's much-cited 2002 review of MSW leachate composition, which is still the reference most engineers reach for.

Now give the same cell a decade. The easy carbon is gone, the BOD-to-COD ratio collapses, pH climbs past neutral, and ammonia, which the landfill does not biodegrade, simply accumulates. By the methanogenic phase ammonia is the dominant pollutant, often well past a thousand milligrams per liter, and it isn't leaving on its own. A site can hit zero waste to landfill on this year's tonnage and still be making ammonia-heavy leachate from what it buried twenty years ago. Diversion changes the future mix. It does nothing to the legacy water.

Then there's weather. (Rainfall doesn't ask your design basis for permission.) A wet season can multiply flow severalfold and dilute strength; a dry spell concentrates it. What got landfilled matters too, because a cell that took a lot of construction debris makes different leachate than one full of household organics, and few sites bury a consistent mix year over year.

So the influent drifts on at least three axes at once: age, weather, and waste mix. A plant tuned to the average of those axes is tuned to a condition that rarely actually occurs.

Reverse osmosis moves the problem downstream

Reverse osmosis is the default answer for landfill leachate treatment, and as a membrane process it's a good one. It's also, physically, a splitter and not a destroyer, and that distinction is where the money hides.

Push leachate through an RO train and you get two streams. The clean one, the permeate, is roughly 70 to 85 percent of the volume and can often meet discharge limits (per the ultra-high-pressure RO work Saltworks has published). The other stream is the concentrate: 15 to 30 percent of what you started with, carrying nearly everything the membrane rejected. You didn't remove those contaminants. You put them in a smaller, meaner bucket that now has to go somewhere. It's the same trap as waste-to-energy ash, where the residual nobody wanted to think about turns into the offtake problem that quietly decides the economics.

And RO doesn't even catch the contaminant that mature leachate is worst for. Ammonia slips through the membrane; typical removal lands around 50 to 60 percent (Waste Today's technical writeup on leachate ammonia lays this out), which is exactly backwards from what you want, because ammonia is the thing that climbs as the site ages. So the plant that looked spotless at commissioning, when the leachate was young and RO-friendly, is the same plant that tripped its ammonia limit five years later once the chemistry turned. Reverse osmosis leachate treatment isn't wrong. It's just answering a question the influent keeps re-asking.

Biological treatment carries the same disease, by the way. A nitrification stage that thrives on the carbon-to-nitrogen balance of a young cell can stall on an old one, where the easy carbon is gone and the ammonia has spiked, and the organisms that strip nitrogen need conditions the raw water no longer hands them. So the polishing step people reach for to fix RO's ammonia gap is itself age-coupled. There's no leachate treatment technology exempt from the influent moving underneath it. There's only how fast you notice it move.

Now the concentrate has to go somewhere, and that's the point at which landfill leachate disposal cost stops being a line item and becomes the whole conversation. Truck it off-site and you're paying mostly for transport, not treatment: one 2023 survey of U.S. landfill practice in the Journal of the Air and Waste Management Association pegged trucked disposal anywhere from a few cents a gallon to about twenty, which at the far end of a long haul works out near $55/m3 (my conversion from their per-gallon range). Evaporate it and you're buying energy. Deep-well it and you're buying a permit and a geology. There's no free exit.

And the concentrate gets nastier the harder you squeeze it. Drive an evaporator or an extra high-pressure stage far enough and the dissolved-solids load climbs toward 140,000 milligrams per liter (again per Saltworks), at which point scaling and corrosion start writing your maintenance schedule for you. Shrinking the volume is not the same move as shrinking the problem. Past a point you're just swapping a disposal cost for a maintenance one, and the maintenance one shows up every week.

And the off-site route, the one a lot of sites quietly depended on, is closing. When the EPA sampled 200 landfills back in 2021 it found PFAS in the leachate at 95 percent of them, across 63 compounds, and it's now writing effluent limits for landfill discharge under the Subtitle D category (Waste Dive covered both the study and the rulemaking). The fallout came fast: the cheap disposal arrangement many operators leaned on for years broke around 2022, when publicly owned treatment works that used to take leachate as a revenue line started refusing loads over PFAS liability. That's why a growing number of operators are eyeing on-site thermal destruction for the worst concentrate rather than trusting an outside taker who can change their mind.

Recirculation buys time, not an exit

Leachate recirculation gets sold as a disposal strategy, and it isn't one. It's a stabilization strategy that happens to reduce how much you truck.

The bioreactor idea is sound. Pump leachate back through the waste mass, keep the moisture up, and the biology runs faster, so the site stabilizes in years instead of decades. Fine. But recirculating leachate doesn't destroy the contaminant mass; it moves it in a loop, and the fractions that won't degrade, ammonia first among them, build up as the water cycles. So you can end up with less leachate that's harder to treat, not more leachate that's easier. Recirculation is a genuine tool for leachate management. It isn't the same thing as making the problem leave.

There's a newer wrinkle worth sitting with. If leachate is the main pathway PFAS uses to escape a modern landfill, then pumping leachate back into the waste mass is pumping PFAS back in, and every cycle concentrates the very compounds neither the biology nor the membranes will touch. Recirculation was designed for a world where the hard-to-treat fraction eventually broke down inside the cell. Some of it doesn't break down, and we've known that for a while now.

The standard pitch on a treatment skid reads about like this: "Our system meets discharge limits and cuts leachate volume by eighty percent, so your disposal problem is solved."

That volume-reduction number is true and the promise around it is misleading. It describes the good day: young leachate, a fresh membrane, a cooperative season. It says nothing about the concentrate you now own, the ammonia the membrane let past, or what the influent looks like in year eight once the cell has gone methanogenic. A spec sheet is a photograph. Leachate is a film.

Treat it as a data problem

If the influent won't sit still, the plant has to watch it. That's the whole recommendation. Why does the industry keep sizing these plants to a single grab sample and then act surprised when the water turns? Watching is less exciting than a new membrane, which is probably why the watching part gets undersold.

Concretely: instrument the raw leachate and the concentrate for the things that actually move. Conductivity, a decent ammonia probe, pH, ORP, flow, and a UV-254 reading as a quick proxy for organic load. Log it, trend it, and feed it forward so the treatment train adjusts to the water it's about to see rather than the water it saw at design. The RO skid's PLC (usually a Siemens or an Allen-Bradley) can act on a set point that tracks the influent instead of a constant. This is the unglamorous core of what vendors package as waste-intelligence software, and the unglamorous core is the part that earns its keep.

Here's the analogy I keep coming back to, from a NIR moisture model on an RDF dryer I worked on in 2024. The sensor drifted about three percent a month, so we stopped treating calibration as a commissioning task and moved it inside the control loop, recalibrating against grab samples on a schedule the drift set, not the manual. Leachate is the same shape of problem. The influent drifts, therefore characterization can't live in a binder as a one-time report. It has to be a live input to the loop.

And no, this doesn't mean bolting an AI dashboard onto the SCADA and calling the plant autonomous. Most of what gets pitched as smart leachate control is a rule engine with a model bolted on the side, and a model that has never seen a methanogenic-phase influent can't manage one. Tune the inputs and the loop first. The clever part can come once you can trust the boring part underneath it.

The design principle that falls out of all this is deeply unfashionable, and the leachate treatment systems that actually hold up over a permit's life already run on it: size for a range, not a point. Give the ammonia step real headroom for the methanogenic years it will eventually see. Spec the membrane for a dirtier feed than commissioning shows you. Treat that first characterization study as the opening data point in a long series, not as the specification carved above the door. It costs more up front. It costs a great deal less than a permit exceedance in year six and the emergency retrofit you scramble to build chasing it.

Where this argument stops

Scale is where this breaks down. A small or closed landfill making a modest, stable trickle doesn't need a live sensor stack, and instrumenting one doesn't make economic sense; trucking everything to a taker who'll still have you is cheaper and simpler. Arid sites with low infiltration drift more slowly and can get away with less watching. And the instruments themselves aren't free or honest: an ammonia ISE fouls in raw leachate within days and grows its own drift, so the monitoring layer needs its own maintenance discipline or it lies to you with a straight face. I've chased a clean-looking dataset that turned out to be a fouled probe more than once. The sensor stack is a commitment, not a gadget you install and forget.

None of that rescues the snapshot approach, though. It just marks the floor below which the cheaper, dumber option wins outright, which is a smaller landfill than most operators think.

So, the position again, sharper. A landfill stops accepting waste and keeps making leachate for another thirty years; RCRA Subtitle D writes that post-closure care clock into 40 CFR Part 258, and the water doesn't stop when the clock runs out. If your landfill leachate treatment plant only understands the leachate you had at ribbon-cutting, you haven't built a treatment plant. You've built a monitoring gap with pumps, and the gap widens every year the chemistry moves and nobody checks. The problem outlives the landfill. Design the plant like it knows that.

Sources & Notes

• The leachate aging picture, the BOD-to-COD ratio above 0.5 in young leachate, and the point that ammonia rises rather than falls with age all trace to Kjeldsen and colleagues, Present and Long-Term Composition of MSW Landfill Leachate: A Review (Critical Reviews in Environmental Science and Technology, 2002).
• The PFAS-at-95-percent figure comes from EPA's September 2021 study of 200 landfills and the effluent-guideline rulemaking that followed, as reported by Waste Dive; EPA's own program page for the Landfills effluent guidelines tracks the Subtitle D revision.
• RO permeate and concentrate split, and the ammonia-passthrough problem, draw on Saltworks' write-up on high-pressure RO for leachate and Waste Today's piece on treating ammonia in landfill leachate.
• Trucked off-site disposal costs are from a 2023 leachate-practice survey in the Journal of the Air and Waste Management Association; the per-cubic-meter figure is my own conversion of their per-gallon range and should be read as order-of-magnitude.
• The NIR-moisture-drift example is from my own commissioning work on an RDF dryer in 2024; the drift rate and the move-calibration-into-the-loop fix are what we measured and did on that line.