Backcountry Drinking Water: What Filters Can and Cannot Do

What's in there

The things that make backcountry water a problem are mostly biological, and mostly they arrive by way of something that lived upstream.

Protozoa are the classic backcountry offenders. Giardia, which travels as a cyst about 8 to 12 microns across — large, as these things go, and correspondingly easy to strain out. Cryptosporidium, smaller at 4 to 6 microns, and considerably more stubborn for reasons that matter later. Both arrive from mammal faeces: beaver, cattle, deer, other people. Both are tough-walled, dormant, and patient.

Bacteria — Campylobacter, E. coli, Salmonella — are smaller again, roughly 0.2 to 5 microns, and travel the same route.

Viruses are the ones that change the equation. Norovirus, hepatitis A, rotavirus: 0.02 to 0.1 microns, which is to say ten to a hundred times smaller than a bacterium. In remote North American backcountry they're an uncommon risk because they're largely human-specific and need human sewage. In heavily used areas, in developing regions, and anywhere downstream of settlement, they're a live consideration.

Then there's everything that isn't alive. Mine drainage carrying dissolved metals. Agricultural runoff. Harmful algal blooms, which produce toxins in the water rather than merely floating in it. Fuel spills. None of these are organisms and none of them can be strained.

A filter is a size barrier

Backcountry filters work by forcing water through pores small enough to block protozoa and bacteria. An absolute rating describes the largest opening, so compare absolute ratings rather than a vague “nominal” claim.

Viruses are smaller than the 0.1- and 0.2-micron pores common in these filters. When viruses are a credible risk, use a purifier, boil, or follow filtration with a disinfectant and its full contact time.

Chemicals, and the patience they ask for

Chemical disinfection works differently. It doesn't remove anything; it kills. And killing takes time, which is where people go wrong.

Chlorine dioxide handles bacteria and viruses in around thirty minutes. Against Cryptosporidium it needs up to four hours, because that thick oocyst wall is genuinely, impressively resistant to chemical attack. This is not a suggestion buried in the small print. It is the difference between treated water and water with tablets in it.

Iodine and plain chlorine are worse against Cryptosporidium — unreliable enough that they shouldn't be your answer where crypto is plausible, which is anywhere with cattle or heavy recreational use.

Three things lengthen the required contact time: cold water, cloudy water, and — with iodine — certain organisms regardless. Cloudiness matters because particles physically shield microbes from the chemical. If the water is silty, let it settle or pre-filter it through a bandana first, then treat, and give it the full time.

Don't blend methods by instinct. Half a chemical protocol plus half a filter protocol is not a complete protocol.

UV devices can inactivate germs in clear water when the lamp, batteries, vessel, and exposure sequence work correctly. They remove no sediment, leave no disinfectant residual, and are poor choices for visibly cloudy water unless it is clarified first.

Boiling

Boiling is the oldest method and it remains, in pure biological terms, the most thorough: bring water to a rolling boil for one minute, or three minutes above about 6,500 feet, and every pathogen on the list is dead. The water reaching boiling point does most of the work; the hold time is margin.

What boiling cannot do is remove anything dissolved. Heat doesn't touch heavy metals, and it concentrates rather than removes them. It also costs fuel, which is why nobody boils all their drinking water on a long trip.

As a backup when a filter fails, it's excellent. As a primary method, it's a car-camping technique.

The dirty side and the clean side

A common treatment failure is not equipment failure. It is somebody putting the dirty hose down on top of the clean bottle cap.

Establish a physical and verbal separation and be slightly pedantic about it. The dirty bag is the dirty bag. It keeps its own color, its own corner of the pack, its own name — say it out loud, so nobody in the group absent-mindedly fills the wrong one. Keep untreated hands and untreated water away from clean bottle threads and outlets. Water that has passed correctly through a filter and then been poured into a bottle you rinsed in the creek is untreated water.

Collect from the clearest place you can reach without standing in the current or breaking down the bank. Facing upstream, holding the bag just under the surface rather than dredging the bottom, gets you cleaner water and a filter that lasts longer.

Follow the exact backflush or cleaning interval for the filter rather than borrowing another model's schedule. If flow collapses and the required cleaning parts are missing, use the named backup or turn around.

Cold, and what it does to a filter

Hollow-fibre filters are bundles of tiny tubes with water inside them. Freeze one and the expanding ice splits the fibres — and the damage is completely invisible from outside. The filter looks fine. It flows fine. It is now a tube that passes everything.

So: a filter that's been used sleeps in the tent with you, often in the foot of the sleeping bag, on any night that might drop below freezing. If you suspect it's frozen, retire it. There is no field test, and it is not worth guessing.

Follow the backflushing schedule. Watch for a flow rate that's dropping and won't recover, for cracks, for a lost O-ring. Carry a backup method that's independent of the first — tablets weigh almost nothing and don't care how cold it got.

What the blue line does—and does not—tell you

Sit down with current land-manager information before the trip and mark your sources in three categories: reliable, seasonal, and unknown. Then find the longest dry stretch and check that your containers hold enough to cross it with margin — drinking, cooking, and reserve, plus whatever a dog needs.

A watercourse printed on a map means somebody surveyed water there once. Drought, season, upstream diversion, a beaver dam, a fire closure, or a contamination event can all make it a dry bed with a name.

Do not treat water with visible scum, mats, paint-like color, or a bad odor as an ordinary filter problem. Harmful algal toxins may remain after common backcountry treatment. Keep dogs out too. If untreated water was consumed and significant diarrhea, vomiting, fever, dehydration, or neurologic symptoms follow, seek medical advice and report the exposure details.

The creek that wasn't there

A route description promises a seasonal creek four miles in. The current ranger report can't confirm it's running.

The honest version of this calculation is simple: if you walk in on empty bottles and the creek is dry, you have no margin to get back out. So you carry water for the whole out-and-back plus cooking plus reserve, you shorten the route to fit that weight, and you save the longer line for a season when the sources are confirmed. The filter comes along. It just isn't counted as water.

And at the dry creek bed, you don't dig in the mud, and you don't drink from the stagnant pool where the deer have been standing, and you don't push on to the next unverified mark on the map because you're thirsty and it's probably fine. You turn around at the time you wrote down before you left.

Which is the whole thing, really. A filter is a very good sieve. It works on what it can reach, at the scale it was built for, and it has no opinion about whether there's any water in front of you. That part is still yours to get right — and it's the part that rewards sitting down with a map a week ahead, which is a quieter and more pleasant task than most of the alternatives.