The Mold CFU Threshold Response Protocol for Cannabis (2026)

When a yeast-and-mold count comes back high, the protocol is a fast, documented sequence, not a panic. Confirm the result and isolate the affected room, identify whether the load is on surfaces or in the air, then run the matched response: disinfect hard surfaces with ProKure V, reset an empty room fast with ProKure G, and hold ongoing air quality with ProKure D while plants are in the room. Re-test to prove the room is back in range, and log every step. Because state thresholds vary and change, confirm your current limit with your regulator rather than a remembered number.

The six-figure scenario this protocol exists to prevent

Picture the moment every head grower is trying to avoid. A room is two or three weeks from harvest, the crop already carries six figures on the books, and a yeast-and-mold screen comes back above the state's colony-forming-unit (CFU) threshold. The batch is held. Now you are choosing between remediation, a retest, or destruction, and every day of delay is money and risk. Nobody saw it coming because nothing in the file showed the load climbing, so there was no early signal and no documented response ready to run.

That scenario is the whole reason a CFU threshold response protocol exists. The protocol is not a product you spray when the number is bad. It is a written, rehearsed sequence that does three things: it catches a rising load before the gating test, it responds in a matched way when a count is high, and it proves on paper that the room came back into range. A grower with that protocol treats a high count as a known procedure with assigned steps. A grower without it treats the same count as a crisis, and crises are where expensive mistakes happen, like spraying the wrong product on a canopy or destroying a lot that a documented remediation and retest could have saved.

The distinction that matters most, and the one most protocols blur, is between the compliance test and your own monitoring. The compliance test is the finished-flower yeast-and-mold screen that gates whether a batch can be sold. Your own monitoring is the air sampling and environmental data you collect through the cycle to see the load before it reaches that gate. The protocol lives mostly in the second world, because by the time the compliance test fails, your options are expensive. The point of the response protocol is to make the gating test a formality you already expect to pass, because you have been watching and responding upstream the whole time.

Run the cost math once and the protocol stops looking like overhead. A single failed pre-sale mold test on a mature crop can erase six figures in one batch, plus the remediation labor and the licensing exposure that comes with a compliance miss. Against that, the ongoing ClO2 air program, ordered at roughly a $2,466 average order by established operations reordering around 13.6 times a year, is a rounding error on the crop value you already have on the books. The expensive path is almost always the one where a grower skips the air-quality layer or the monitoring to save a few hundred dollars a cycle, then loses a harvest to a spike that a documented protocol would have caught and closed. Prevention is the cheaper number in nearly every scenario a cultivation director will model, and the response protocol is the insurance that makes the prevention investment pay off.

There is a psychological trap in the crisis moment worth naming, because it drives the most damaging mistakes. A grower staring at a high count wants to do something dramatic, and the dramatic options are usually the wrong ones: dousing the canopy with a surface disinfectant that is not labeled for foliage, running an empty-room gas product with plants still in the room, or destroying a lot that a documented remediation and retest could have saved. The protocol exists precisely to replace the dramatic reflex with a calm sequence. When the steps are written down and rehearsed, the high count becomes a procedure, and procedures do not panic.

What a CFU threshold actually is, and why you confirm it with your regulator

A CFU count is a measure of viable mold and yeast, reported as colony-forming units, usually per gram of flower or per volume of sampled air. A threshold is the regulatory line above which a result fails. The trap growers fall into is treating a number they heard at a conference, or read on a blog, as their standard. It is not. The exact threshold, the method, and whether a state screens for specific Aspergillus species on a presence-absence basis rather than a straight count all vary by state and change over time.

So the first move in any response protocol is not a product. It is confirming your current threshold and action levels against your own state regulator, and keeping a copy of the rule you relied on in your file. This is not bureaucratic caution; it is the thing that makes every later step defensible. If your protocol says "respond when air sampling trends toward the action level," the action level has to be a real, sourced number tied to the current rule, not a figure someone remembered. We will not print a specific CFU threshold here, because doing so would hand you a number that may already be stale for your state. The protocol is built to work against whatever your current, confirmed threshold is.

What your file needs to show on the threshold question is a chain, not a memorized figure: the accredited lab result, tied to a room and a date, with the method stated, measured against the current regulatory limit you have on file, and, where a result approached or crossed that limit, the corrective action and the re-test that followed. That chain is what turns a raw count into evidence of a managed program, and it is what an auditor or an owner actually wants to see.

It helps to understand why thresholds move, because it changes how you treat any number you are given. Cannabis sits in a regulatory gap at the federal level, so each state wrote its own testing rules, its own CFU limits, and in some cases its own species-level screening, and those rules get revised as labs, methods, and public-health thinking evolve. A number that was correct for your state two years ago may have tightened, loosened, or shifted from a count-based limit to a presence-absence screen for specific organisms. The practical consequence is that your SOP should cite the current rule by reference and show your monitoring addresses the organisms that rule names, rather than freezing a figure in the document. If the standard moves between when you wrote the SOP and when the auditor reads it, a file that points to the live source protects you; a file that quotes a stale number does not.

Build an internal action level that sits below the regulatory limit, and write it into the protocol as the trigger for response. Waiting until a count crosses the legal threshold to act is waiting until the batch is already in jeopardy. A sensible program responds when its own monitoring trends toward the action level, giving the room time to come back down before a gating test ever runs. Set that action level against your confirmed current limit, document the logic, and you have converted the threshold from a pass-fail cliff into an early-warning system.

Reading the result: surfaces, air, or incoming genetics

A high count tells you there is a problem; it does not tell you where the problem lives. Before you reach for any product, work out which of the three paths the load is traveling, because the path determines the response. Mold reaches flower by three routes: airborne spores that settle and germinate, contaminated surfaces that reseed a clean room, and incoming genetics (clones, mothers, tissue) that carry it in. A response that treats the wrong path leaves the real source untouched and the next test fails again.

Start with the pattern. A single hot spot in one room, right after a turnover, points at surfaces the last crop left dirty. A load that climbs steadily across a cycle, or shows up room-wide, points at the air. A problem that appears right after you introduced new clones points at incoming genetics. These are not mutually exclusive, and a bad situation is often all three at once, which is exactly why the response is a system rather than a single product. Your air sampling and environmental records are what let you read the pattern; without them you are guessing, and guessing is how a surface problem gets an air-only fix.

Each path has its own tell, and learning to read them quickly is most of the skill. The surface path shows up as a sharp, localized spike right after a turnover, because a dirty bench or tool reseeded an otherwise clean room, and it tends to concentrate where the contaminated surface was. The air path shows up as a gradual climb across the cycle or a room-wide elevation with no single hot spot, because spores are circulating and settling everywhere the air reaches, which is also why the dense back half of flower, with its poor internal airflow, is where air-driven loads do the most damage. The genetics path shows up as a problem that appears shortly after new clones or mothers entered, often tracking back to the specific bench or row the new stock occupied. A grower who can place a high count into one of those three patterns has already narrowed the response from a shotgun to a matched tool.

This reading step is also where the biosecurity framing earns its place, and it is where GroClarity's cultivation-IPM position differs from a generic cleaning instruction. Prevention and response are a program with stages, owners, and proof at each step, not a chore you do when the number is bad. A program reads the path, matches the response, documents it, and re-tests. A chore grabs the nearest sanitizer and hopes. The whole protocol below is built so that each step has a trigger, a matched product, and a record.

One more reading discipline matters: separate the organism from the symptom. A yeast-and-mold count is a bulk measure, but the organism driving it changes the stakes. Aspergillus is a human-safety organism that states screen for specifically, so an Aspergillus-driven count is a harder stop than a general mold load, and your response and your documentation should reflect that. Botrytis and powdery mildew behave differently again, favoring humidity swings and dense canopy. You do not need a mycology lab to run the protocol, but noting what your testing and your eyes tell you about the likely organism sharpens both the response and the record you keep.

The response protocol, week by week

Run the response as an ordered sequence. These steps are written so a facility manager can execute them without guessing, but treat them as a draft SOP and confirm every product claim against the current label before you post it. Each step names a trigger, an action, and the record to keep.

Step one, confirm and isolate (day zero to day one). A high count is not acted on until it is confirmed, because a single reading can be a sampling artifact. Confirm with a repeat sample from the affected room, and while you wait, isolate the room: limit traffic, tighten airflow so you are not pushing spores into clean rooms, and flag the batch in your tracking so nothing moves until the picture is clear. Record the original result, the confirmation sample, the isolation actions, and the time. Isolation is not optional even when you suspect a false positive, because the cost of containing a room that turned out clean is a day of caution, while the cost of not containing a room that turned out hot is cross-contamination into your clean rooms and a bigger problem next cycle. Treat containment as the cheap insurance it is, and record it so the file shows you acted immediately rather than waiting for certainty.

Step two, read the path (day one to day two). Using the confirmation result plus your environmental and air-sampling history, decide whether the load is on surfaces, in the air, in incoming genetics, or across more than one. Write down the determination and the evidence for it. This is the decision that routes every later step, so it belongs in the file, not just in the grower's head.

Step three, surface response with ProKure V (day two, or at the next empty-room window). Where surfaces are implicated, disinfect the hard non-porous surfaces, benches, trays, reservoir walls, tools, drop-down hoses, with ProKure V at its labeled use. V is suitable for use with plants in the room because it treats surfaces, not foliage, so routine surface disinfection does not require evacuating the crop. Record the date, the surfaces treated, the dilution, and the contact time.

The execution detail matters for efficacy, so write it into the step. Pre-clean any visibly soiled surface with a non-disinfectant cleaner first, because organic soil consumes the oxidizer before it can disinfect, and a count that stays high after a surface pass often traces to skipping the pre-clean. Activate V per its label, drop the pouch into an opaque container with the correct water volume for 100 ppm, agitate for one minute at 15-minute intervals until full activation at one hour, then apply by sprayer, mop, or sponge to visible wetness and hold the full 10-minute contact time before the surface is worked again. The crew wears gloves if hands will contact the solution; no respirator is required at 100 ppm. Do not extend V to the canopy because a count is scary; it is a surface product, and spraying foliage with a disinfectant is both off-label and the kind of panic move the protocol exists to prevent.

Step four, empty-room reset with ProKure G (at turnover, room empty). When the situation calls for a full-space reset and the room can be cleared, run ProKure G. G is the fast-release gas for an empty room, labeled not for use with plants, crops, or feed, so people and plants are out during treatment and you ventilate before re-entry. Record the treatment and the safe re-entry time on the caution placard and in the turnover log. If genetics were implicated, this is also the window to quarantine and inspect incoming stock before the next crop enters.

Dose G by room volume, not by guess. A 10-gram pouch covers roughly 1,000 cubic feet and a 25-gram pouch roughly 2,250 cubic feet for standard odors, with multiple pouches placed strategically for even coverage in a larger or heavily loaded room. Activate the white pouch in a plastic cup with about 1.5 inches of water, place it on a stable, level surface in a central location, turn the lights off, and allow at least 4 to 6 hours or overnight for the system to work before ventilating and turning the lights on for one hour ahead of re-entry. Because the gas is heavier than air, keep the unit high and central and use a fan to distribute it, away from direct drafts that would channel it out before it spreads. The reason G sits in the turnover window rather than the active-grow window is the plant restriction, and the reason it is worth the empty-room downtime is that a fast full-space reset breaks the surface-to-air carryover that a surface wipe alone leaves behind.

Step five, hold the air with ProKure D (plants back in, ongoing). Once the crop is back in the room, run ProKure D for ongoing air quality across the grow. D is the slow-release gas that improves air quality and controls odors, holding for up to 30 days per deployment, and it runs with plants present. Log the placement date so the up-to-30-day window is on paper, and keep your air sampling running so you can see the load staying down rather than assuming it.

Size D to the room and hang it correctly. A 10-gram pouch treats approximately 1,000 cubic feet, a 50-gram approximately 4,000, and a 95-gram approximately 8,000; for a larger room, place pouches equally apart rather than oversizing one in a corner. Hang high and central using the trifold punch holes, or place the unit on a stable, level surface, so the slow-released gas distributes through the canopy rather than pooling. Across a typical flower cycle the up-to-30-day window gives you a natural replacement rhythm, a second deployment around day 30 so you do not leave an air-quality gap in the dense back half of flower when airflow is worst and a rebound load is most likely. D is the layer that keeps the room from drifting back toward the action level after you have done the surface and empty-room work, which is why it is the ongoing step rather than a one-time blast.

Step six, re-test and close the loop (after the label interval). The response is not finished until a re-test shows the room back in range. Pull a follow-up sample after the appropriate interval, file the result next to the original and the corrective action, and only then clear the batch or the room in your tracking. A high count with a documented response and a clean re-test is proof the program works; a high count with no recorded follow-up is the finding an auditor writes up.

Sample the re-test the same way you sampled the original, same points, same method, same accredited lab, so the two numbers are actually comparable and a drop is a real drop rather than an artifact of sampling somewhere cleaner. If the re-test is still high, resist the urge to repeat the same step harder; go back to the path reading, because a persistent count usually means you treated one path while the real source sat on another. A surface response that does not move the number points at an air or genetics source you missed, and that re-reading is more valuable than a second identical pass. Close the loop in the tracking system only when the re-test confirms the room is back under your action level, and keep both results, the action between them, and the disposition in one place so the whole event reads as a single closed story.

Step seven, review and adjust the cadence. After the loop closes, look at what the pattern told you and tighten the preventive cadence so the same path does not reopen: more frequent surface passes, a shorter D replacement interval, tighter humidity control, or a stricter quarantine on incoming genetics. The response protocol feeds the prevention protocol; each event should make the next one less likely.

A note from Dr. Bernie Lorenz on reading the number, not fearing it

Dr. Bernie Lorenz, GroClarity's Chief Science Officer, frames the CFU question the way a scientist should: the number is information, not a verdict to panic over. A high count tells you the microbial pressure in a room crossed a line; it does not tell you the crop is lost, and it does not tell you which lever to pull. The discipline is to let the data route the response. Read whether the load is on surfaces or in the air, match the product to the path and to its own label, document what you did, and re-test to confirm. That is biosecurity as a science, built on results rather than on a reflex to spray something.

The other thing Dr. B would say is that the cheapest place to deal with a CFU problem is upstream, before the gating test, which means the real work is the monitoring that catches a rising load early and the air-quality layer that keeps the pressure down between tests. A grower who treats mold prevention as a continuous program, with data on the page, is not gambling on a single reading at the finish line. He is managing the environment around the plants rather than reacting to the plants themselves, which is the whole ProKure approach: treat the room, not the crop.

He would also push back on the instinct to chase a single number. One reading is a data point, not a trend, and a protocol built on single readings overreacts to noise and underreacts to slow climbs. The scientific move is to watch the trend: a count that ticks up two samples in a row toward your action level is a clearer signal than a single high reading that may be a sampling artifact, and it gives you time to respond while the response is still cheap. That is why confirmation is step one and why the air sampling runs continuously rather than only at test time. The number is the start of the inquiry, not the end of it.

Finally, Dr. B would insist that honesty about what the products do is part of the science, not a legal afterthought. ProKure V kills on labeled surfaces; ProKure G resets an empty room; ProKure D holds air quality with no kill claim and no EPA registration. Overstating any of those, calling D a fungicide, implying G can run around plants, claiming V treats the canopy, is not just a compliance risk, it is bad science that leads a grower to trust a product to do something it does not do. The protocol works because each claim is true and matched to the right product, which is exactly the posture that keeps a crop and a license safe at the same time.

The V, G, and D system: matching the product to the path

The response protocol works because each product does a distinct job and is documented to its own label. Mixing them up is where growers break a label rule or leave a gap, so keep the three straight.

ProKure V is the surface half of the system. It is a water-activated liquid, a hospital-type disinfectant for cultivation facilities, EPA Reg. No. 87508-3-89334, used at 100 ppm for a 10-minute contact time on hard non-porous surfaces, and it leaves no residue. Its label kill list on those surfaces includes Aspergillus niger (ATCC 9642), Fusarium solani (ATCC 36031), Botrytis sp. and Penicillium digitatum (ATCC 201167), plus bacteria such as MRSA (ATCC 33591), and it kills SARS-CoV-2 on hard non-porous surfaces in 30 seconds. The reason V belongs in a CFU response is the combination bleach and quats cannot match: a hospital-type surface kill list, no residue, and suitability for use with plants in the room, applied to surfaces rather than foliage. V is the smallest share of what cultivators order, but it is the disinfectant half of the system and it carries the only kill claims.

ProKure G is the empty-room air reset. It is the fast-release gas, EPA Reg. No. 87508-2-89334, a deodorizing delivery system that works in about 4 to 6 hours, and it controls odor-causing bacteria, mold, and mildew in the air of a space. Its label restriction is absolute: not for use with plants, crops, or feed, so G runs only when the room is empty of people and plants, and you ventilate for one hour before re-entry. Keep G's restriction matched to G alone and never attach it, or V's kill claims, to the wrong product in your SOP.

ProKure D is the ongoing air-quality layer that runs with the crop. It is the slow-release gas that improves air quality and controls odors and provides ClO2 protection for up to 30 days per pouch, sized by volume (a 10-gram pouch treats about 1,000 cubic feet, a 50-gram about 4,000, a 95-gram about 8,000). The compliance line is verbatim and non-negotiable: ProKure D is not an EPA-registered product and not labeled for sale as a pesticide under 40 CFR. It makes air-quality and odor statements only, carries no kill claims and no registration number, and must never be documented as a registered fungicide. D and G together are the air half of the system that most of the volume goes toward, because the air is where a CFU test is won or lost, while V handles the surfaces. Pull the SDS, labels, and EPA registration references for V and G from prokuresolutions.com for your file, and buy the products that match your room on shop.groclarity.com.

The reason the three work as a system rather than as alternatives is that a CFU problem almost never lives on one path alone. A dirty bench reseeds the air; a loaded air column settles back onto clean surfaces; a contaminated clone feeds both. Treat only the surfaces and the air reloads them; treat only the air and the surfaces refill it. The matched sequence, V on the surfaces, G to reset the empty room, D to hold the air through the grow, closes every path a spore can travel, which is why a single-product response so often produces a clean reading followed by a rebound two weeks later. Keep each product in its lane and the lanes cover the whole road.

A short note on where V sits in the order of priority, because the purchase data can mislead. V is the smallest share of what cultivators order, roughly a tenth of volume against the air products, which tempts a grower to treat it as optional. It is not. V carries the only kill claims in the line and it is the surface half of a CFU response; the fact that growers buy more air product than surface product reflects how often each job comes up, not the importance of the surface step when a count is high. In a response protocol, V is the step that actually reduces the viable load on the surfaces a spore is reseeding from, so it earns its place regardless of its share of the invoice.

Turning the response into a downloadable SOP your crew runs every time

A protocol that lives in one grower's head fails the week that grower is on vacation. The response sequence above becomes durable only when it is a written SOP your crew can follow and your auditor can read. Build it as a checklist that mirrors the seven steps: confirm and isolate, read the path, surface response, empty-room reset, hold the air, re-test and close the loop, review and adjust. Next to each step, name the trigger, the responsible person, the product and its labeled use, and the record to capture.

Keep the SOP honest about documentation, because the record is half the point. For each event you want the original result, the confirmation, the path determination and its evidence, the product applications with dates and operators, the re-test, and the final disposition of the batch or room. Store it where it survives a staff change, a shared drive or a compliance system, not a notebook on a desk. Over several events you build a history that shows the program responding and closing loops, which is exactly what reassures ownership and satisfies a state auditor.

For a multi-room or multi-state operation, run the identical SOP at every site. The value of standardization is that one response protocol, documented the same way everywhere, means an auditor who learns your file at one location already knows where everything lives at the next, and a cultivation director is not defending a patchwork of improvised responses room to room. The protocol scales because it is written, matched to product labels, and recorded the same way every time.

Make the checklist usable under pressure, not just complete on paper. A grower reaching for it the morning a count comes back high wants a one-page flow: the trigger at the top, the seven steps as checkable lines, the product and labeled use beside each, and a short records column to initial and date as the work gets done. Keep the deeper reference, the full product label language, the dilution tables, the threshold source, as attachments behind the one-pager rather than in the flow itself, so the person executing is not reading three pages to find the next action. The test of a good SOP is whether a competent crew member who has never run this exact event can execute it correctly from the document alone. If it needs the head grower standing over it to make sense, it is notes, not an SOP.

Review the SOP after every event and after any threshold change. Each closed loop teaches you something, which path reopened, which step was slow, where a record was missing, and the fix belongs in the next version of the document. When your state revises its CFU limit or its screening method, update the sourced threshold and your internal action level in the same pass. A living SOP that absorbs each event and each rule change stays defensible; a static document written once and never revisited slowly drifts out of alignment with both your operation and the regulation, which is its own audit exposure.

Next step

A high CFU count is survivable when you have a protocol ready and expensive when you do not. Confirm your current threshold with your state regulator so your standard is sourced. Write the seven-step response into an SOP your crew can run without you, match each step to the right ProKure product and its label, and keep the record that proves each loop closed. Keep your own air sampling running between compliance tests so you see a rising load early, because the cheapest response is the one you make before the gating test. When you are ready to build the program, scope air testing and a standardized protocol through the groclarity.com contact form, pull the SDS and labels for V and G from prokuresolutions.com, and order the products that match your rooms on shop.groclarity.com.

Related resources