Water central | Water Storage-book | Water testing-download | Wild water wisdom-article | Water Conservation Central | Fecal coliform counts | Stormwater | Rainwater harvesting | Slow sand filters

Fecal Coliform Bacteria Counts: What They Really Mean About Water Quality

Translating general and fecal coliform bacteria counts to more understandable measures can help you understand and relate them to the real world. A deeper understanding water quality can help you better design and use water systems and find safe water for drinking, swimming, boating and recreation.

Few people understand the commonly used measurements for microbiological water quality...

What the heck does it mean that there are "52 MPN fecal coliforms/100ml of water?

Is it good to drink? To wash dishes in? To bath in? Irrigate with?

The average person—or even engineers and scientists without a public health or microbiology background—wouldn't have a clue.

In fact, these units are so obscure that even people who work with them every day for years and make important decisions based on test results often have little sense of how to relate contamination either to cause or effect in a quantitative way.

Most public health professionals would say that there would be a hazard. But they would be hard pressed to come up with an assessment of the size of the hazard that was accurate within a factor of a hundred. Many would be off by a factor of ten thousand. (Read on and then see how these questions are worked out below).

^ Top ^Learn to do your own testing | Few understand units | Policy w/o understanding | Alternative measurements | Examples

Concentration blindness

Further obscuring the picture, almost all standards are expressed in terms of concentration, not total quantity of organisms.

A spokesman for the Santa Barbara sewage treatment plant once calmly explained that discharge of untreated sewage to the ocean during intense rainfall was not an issue, because "the dilution factor is so great." This is an exceptionally clear case of "concentration blindness."

50,000 kilograms a day of untreated human feces flushed to the ocean in a billion gallons of storm water in a storm that overwhelms the facility is much more harmful to swimmers than the same amount of treated human feces in a million gallons of sewer water. The amount of water that carries it to the ocean is irrelevant considering the relative size of the ocean—what matters is how many pathogens end up in the ocean. If anything, feces delivered in a giant slug of fresh water are worse, as the large flow of less dense fresh water might tend to float to the surface where human exposure is more likely.

However, according to standards for effluent, which are based on concentration, the untreated scenario appears 1,000 times less dangerous as it is, because the pathogens are delivered in 1,000x more water.

^ Top ^Learn to do your own testing | Few understand units | Policy w/o understanding | Alternative measurements | Examples

Indicator connection varies

General coliforms, E. Coli, and Enterococcus bacteria are the "indicator" organisms generally measured to assess microbiological quality of water. However, these aren't generally what get people sick. Other bacteria, viruses, and parasites are what we are actually worried about.

Because it is so much more expensive and tedious to do so, actual pathogens are virtually never tested for. Over the course of a professional lifetime pouring over indicator tests, in a context where all standards are based on indicators, water workers tend to forget that the indicators not the thing we actually care about.

What are these indicators?

The more closely related the animal, the more likely pathogens in their feces can infect us.

Human feces are the biggest concern, because anything which infects one human is likely to infect another. There isn't currently a quantitative method for measuring specifically human fecal bacteria (though expensive genetic studies can give a presence/absence result).

Ingesting human strangers' feces via contaminated water supply is a classic means for infections to spread rapidly. The more pathogens an individual carries, the more hazardous their feces. Ingesting feces from someone who is not carrying any pathogens may gross you out, but it can't infect you. Infection rates are around 5% in the US, and approach 100% in areas with poor hygiene and contaminated water supplies.

Keep in the back of your mind that the ratio of indicators to actual pathogens is not fixed. It will always be different, sometimes very different. Whenever you are trying to form a mental map of reality based on water tests, you should include in the application of your water intuition an adjustment factor for your best guess of the ratio between indicators and actual pathogens.

"Best guess?!" I can imagine precision obsessed regulators cringing. Well, it can hardly be better to ignore the fact that the number and virulence of pathogens present in samples with the same number of fecal coliform indicators can be enormously different, simply because checking for the pathogens themselves is too cumbersome.

These are the factors to include in your mental indicator to pathogen adjustment factor:

^ Top ^Learn to do your own testing | Few understand units | Policy w/o understanding | Alternative measurements | Examples

Policy is being made and infrastructure built with incomplete understanding of hazards

Important public health and engineering decisions are often made with a fuzzy idea of the hazards.

There is a tendency to tighten policy and overbuild facilities until the number of coliforms per 100ml at some point in the process is compliant. If the actual hazard is not in focus, seeking the simple assurance of a compliant reading is understandable.

However, this is a poor design guide compared to real understanding. Consider:

^ Top ^Learn to do your own testing | Few understand units | Policy w/o understanding | Alternative measurements | Examples

Alternative measurements that broaden understanding

In order to incorporate water quality considerations into my designs in a quantitative way, I first had to convert the measurements to units I could understand. Most other water quality measurements are ratios: parts per million, or parts per billion.

The beauty of this kind of measurement is that by multiplying the concentration by the volume of water it is possible to figure out how much actual stuff you're talking about, as in the kitchen sink and beach front septics examples above.

It turns out that one fecal coliform bacteria per 100 milliliters closely equates to one part per billion of feces, or one milligram per cubic meter (you can see how I did the conversion below).

One part per billion of fecal matter is an infinitesimal amount of contamination; about a grain of sand in five 55 gallon drums, or about how much water someone drinks in three years. However, this fails the minimum standard for drinking water quality in most of the developed world, which is zero general coliforms per 100mL (zero fecal coliforms would be a more reasonable standard; zero general coliforms requires disinfection, which in practice can mask actual pathogens by killing off indicators. In the non-industrialized world, water sources with <50 fecal coliforms/ 100mL were anecdotally regarded as "good for drinking" by locals, in my experience).

Converting to concentration and absolute quantities enables you to estimate what could account for a given level of contamination, or what level of contamination would result from a given action. For example, a buttwipe (ahem) diluted in a swimming pool of water yields a feces concentration of about 1 part per billion.

Measuring organisms per 100 mL, you can't easily relate a case of contamination either to cause or effect in a quantitative way.

Without further ado, here is a table which shows conventional units and standards, and their conversion to parts per million, parts per billion, and the novice-friendly units of of buttwipes or turds per swimming pool...

Typical Water Quality Standards, Unit Conversions, and Examples of Fecal Coliform Levels in Water

Screenshot of spreadsheet: Typical Water Quality Standards, Unit Conversions, and Examples of Fecal Coliform Levels in Water

To find the conversion factor from any unit to any other, find the bold number 1's, then read across to the other column. For example:

Things that jumped out for me after making this conversion table:

^ Top ^Learn to do your own testing | Few understand units | Policy w/o understanding | Alternative measurements | Examples

Examples of using these units to understand reality better

Back to the questions we opened with:

If a kid sneaks a swim in a 50,000 gallon drinking water tank, what is the likely level of contamination?

Layperson version: 50,000 gallons is twice as big as a swimming pool. One buttwipe per swimming pool is the drinking standard, so half a buttwipe per swimming pool does not exceed the drinking standardization most likely the kid wiped his butt last time they pooped, so probably well below the actual health issue threshold.

Scientist version: 50,000 is about 200 m3. A buttwipe is about 100 mg. So there is about 0.5 mg per m3 un-wiped, perhaps 0.05 mg/ m3 if they wiped, which does not exceed the drinking standard.

Additional considerations: If the kid is known to have an infectious condition, the cause for alarm is greater. Also, the amount of fecal matter may not be average, it probably won't all come off in the water, and the part of it that does will not be evenly distributed. Most likely, the particles will sink to the bottom or top. If the person who put in the tank read our Water Storage book and the inlets and outlets are designed optimally for water quality, almost none of it will make it into the water distribution system.

If there is a bear poop bleeding into the edge of a swiftly flowing river (25,000 gallons per minute), how likely are you to get sick from drinking the water?

Layperson version: That's a swimming pool every minute. If The bear poop takes 1000 minutes, (16 hours) to dissolve, that's a thousandth of a turd per swimming pool, or ten times the drinking standard.

Scientist version: The bear bowel movement is 1,000,000 milligrams. The flow is about 100 m3 per minute, times 1000 minutes = 100,000 m3 of water. 1,000,000 divided by 100,000 is 10 mg feces/ m3 - ten times the drinking standard.

Additional considerations: A bear poop is probably bigger than a human poop. However, bear pathogens are less likely to infect humans. The swift flow will probably distribute the fecal matter pretty evenly through the water column before long, without much settling. This illustrates a feature of rivers: while on average they are likely to be clean, infectious level pulses of pathogens are likely to come through.

(Note: you can estimate flow by multiplying the width times the depth of the channel times half the speed of the surface. Ten meters wide, two meters deep on average and a meter per second is about ten cubic meters per second.)

If you irrigate your fruit trees with kitchen sink water, how likely is a kid to get sick from eating the dirt under the trees?

A risk assessment analysis of this scenario is viewable in the Arizona Water CASA greywater study. Note that they assume from the high level of indicators that there is a level of pathogens in the water corresponding to nearly a gram a day of fecal matter entering the kitchen sink. This could be accounted for by ten people wiping their butts with their hands only and washing them off in the kitchen sink—an unlikely scenario, I dare say. Also, note that they assume that 100% of the dirt the child eats will come from the greywater-irrigated area, 365 days in a row.

Considering that even with these wild assumptions, the risk was on the order of 1 in 10,000 of the kid getting sick, the risk is probably not significant.

The fruit itself does not pose a risk, as pathogens don't pass from the soil into the plant's vascular system.

^ Top ^Learn to do your own testing | Few understand units | Policy w/o understanding | Alternative measurements | Examples

Do your own tests

Much of what I learned about microbiological water quality came from doing hundreds of tests myself (Want to learn how to test your own water? See our Water Quality Testing download). I use coliscan plates and an chicken egg incubator. What this system lacks in precision is made up for by the fact that you can afford to test many more samples, so you can begin to understand what is going on. (The total setup to do 100 coliscan "easy gel" tests costs less than $200, enough for ten lab tests if you find a very inexpensive lab).

When I test water, the thing I'm least interested in is the results of that particular test. That just shows the indicator bacteria at that spot at that moment.
I'm more interested in knowing how the quality is different at different points in the flow, or different times.
What I'm most interested in: using the test results to further refine my water intuition and understanding.
The better grounded you water intuition, the more accurate your mental model of how the quality is at every point in the system, at any time, with fewer tests.

Instead of grabbing one twenty dollar test from a water tank, I'll test the water the spring as far in as I can reach, at the outlet of the spring box, the bottom of the spring box, the inlet to the tank, the outlet to the tank, the kitchen tap, and in a drinking glass.

If you picked a random rural home system and did this series of tests, you'd could easily find that the bacteria levels varied by a factor of ten. One costly, highly precise test from a certified lab—the standard approach—lends a false sense of accuracy to the results. If you did two such tests a few hours or a few feet apart they could easily be out of agreement by a wide margin.

Numerous tests provide a rich data set from which you can learn a lot about what makes bacteria levels rise and fall.

Once you know that, of course you can make a better system, better judge the safety of water for drinking or swimming, etc.

For potable water, the accuracy of results is more critical, and you need a more sensitive test. I use coliscan membrane filtration tests and Hach presence absence with MUG tests for design of potable systems. Then, if I need certified results, I send one last sample to a certified lab.

^ Top ^Learn to do your own testing | Few understand units | Policy w/o understanding | Alternative measurements | Examples


Unit derivation notes for the scientifically inclined

The numbers in the spreadsheet above were jiggled so the alternative units came out to be even orders of magnitude from the conventional units and each other.

In some cases the number used was close to the middle of the range, in others it is off the average by 30% or more. Overall, the alternative measures which are represented as equivalent on the table are within a factor of two or three of actual equivalency.

This degree of precision is sufficient for this area of study, where data from samples that should be identical routinely varies by a factor of two or three.

The conventional measurements use indicator organisms. There is a few orders of magnitude difference in coliforms per gram of feces for different mammals, so the precedent for allowing imprecision of a large order is well established.

Here's the assumptions and math:

If you want to get into these numbers more deeply, see our Water testing-download, a download packet which includes the editable spreadsheet the calculations were done in.

^ Top ^Learn to do your own testing | Few understand units | Policy w/o understanding | Alternative measurements | Examples