Are Hot Springs Safe? Burns, Amoeba, and Bacteria by the Numbers

Yes, with one large caveat and one small one. The large caveat is heat. The small one is a rare amoeba that gets far more headlines than cases. Three separate hazards get folded into the single question "are hot springs safe," and they differ sharply in frequency and severity. The numbers below come from the four sources listed at the bottom.

The short version: more than 20 people have died from burns after entering or falling into Yellowstone's hot springs. The brain-eating amoeba, Naegleria fowleri, was tied to between 0 and 6 US cases per year across 1978–2018, 85 in total, from all water activities combined, not just hot springs. Those two facts belong in different categories.

The short answer

Three hazards. Three different sizes.

1. Scalding and falling in. This is the hazard with a documented death count attached to a named place: Yellowstone reports more than 20 deaths from burns, and calls burns "a common cause of serious injury and death in the park." Treat thermal features, and their runoff, as burn hazards.
2. Naegleria fowleri, the "brain-eating amoeba." It causes primary amoebic meningoencephalitis, PAM. In a 2023 systematic review, 72.7% of the cases died. It does live in hot springs, including recreational soaking sites in Grand Teton National Park. And the US rate from all water activities combined was 0 to 6 cases a year across 1978–2018. Both halves are true at once.
3. Ordinary bacteria. At Hot Springs National Park, the National Park Service says the water's high temperature "kills most harmful bacteria" and that the water is "monitored to U.S. standards for safe drinking water." The same page says bacteria "may still be present" and could affect people with compromised immune systems if water is inhaled.

What this article will not give you: a safe soaking temperature, a safe soaking time, or your odds per dip. The sources cited here do not establish any of those. Anyone who hands you such a number without a source is guessing.

Hazard 1: Scalding and falling in

Start with the hazard that carries a documented death toll, because it is also the one that gets skipped once the conversation turns to microbes.

The National Park Service's Yellowstone safety page puts it plainly: "Burns from thermal features are a common cause of serious injury and death in the park. The ground surrounding these features may look solid, but it can be just a thin crust with super-heated water below." And the number: "More than 20 people have died from burns suffered after they entered or fell into Yellowstone's hot springs."

Two details on that page matter for anyone who has ever thought about "just testing" a wild pool.

First, it is not only the pool. "Even thermal runoff can be extremely hot and cause burns." The stream draining a hot spring can burn you before you reach the source.

Second, in Yellowstone the question of whether to soak has been answered by law. "For your safety and for the protection of thermal features in Yellowstone, it is illegal to swim or bathe in any water that is entirely of thermal origin." Pets are kept out of thermal areas and off the boardwalks "to protect our pets from being burned or killed in hot springs." The same page adds a hazard that does not fit the postcard image of a hot spring: "Toxic gases may accumulate to dangerous levels in some hydrothermal areas. If you begin to feel sick while exploring a geyser basin, leave the area immediately."

Why is this water so hot in the first place? A hot spring is groundwater that has been heated at depth, following the geothermal gradient, and returned to the surface. Our explainer on how hot springs work covers the plumbing. For a sense of scale: earlier work on Grand Teton's Huckleberry and Polecat hot springs reports outflow temperatures of 40–60 °C. The soaking sites sampled in the 2024 Grand Teton study averaged 41.8 °C at Huckleberry site H1 and 42.3 °C at Polecat site P1, the two highest means, and one Huckleberry site reached 46.1 °C in July. Those are different measurements from different studies, but together they say what a bather needs to hear: the water at a spring is not one temperature. Where it emerges and where people sit can be very different numbers, and the season moves them.

The pattern to hold onto: scalding does not depend on a rare organism showing up. It depends on where you put your foot, and the ground can lie.

Hazard 2: The brain-eating amoeba

Naegleria fowleri is the organism behind the headlines. A 2023 systematic review in the International Journal of Environmental Research and Public Health defines the disease it causes in one sentence: "Primary amoebic meningoencephalitis (PAM) is a rare but lethal infection of the brain caused by a eukaryote called Naegleria fowleri." Rare and lethal. Both words are doing work.

Lethal

The review screened 461 studies and kept 21 for qualitative analysis. Across those, "72.7% of the cases succumbed to mortality." The cases "were distributed globally." The youngest was an 11-day-old boy. The eldest was 75. No age range opts out.

The longer record is starker than that percentage suggests, and it comes with its own caveat. The Grand Teton paper cites the worldwide count: "Between 1937 and 2018, there were 381 confirmed PAM cases worldwide and only 7 confirmed survivors, although it is possible that some cases may go undiagnosed." Seven survivors out of 381 confirmed cases across eighty-one years. The 72.7% figure and the 381/7 figure are different denominators — one is the mortality across the 21 studies the 2023 review selected, the other is the confirmed worldwide record — and the undiagnosed-cases caveat belongs with the second one.

Rare

The same review anchors its numbers to the US record: "the incidence of PAM associated with water activities ranged between 0 and 6 cases per year (n = 85 cases in total)," using data from 1978 until 2018. Eighty-five cases. Four decades. That figure covers water-associated PAM generally and is not specific to hot springs; the cited sources do not break it down by type of water body. Of the 21 studies, 18 "reported a previous history of water contact activities, such as swimming at a lake, river, or waterpark."

Where it lives

This is the part that makes the hot-spring connection real rather than hypothetical. The review lists the environments where the amoeba "can be discovered": "natural hot springs, ponds, rivers, freshwater lakes, drinking water distribution systems, untreated swimming pools, fountains, hospitals, thermal waters, untreated drinking water, and waterparks." The Grand Teton paper's version of the list: "naturally or artificially heated spas, hot springs, water distribution systems, freshwater lakes and ponds, rivers, sediments, and effluents from power plants."

So: a real organism, in real hot springs, causing a disease that killed 72.7% of the reviewed cases, at a national rate of 0 to 6 cases a year from every kind of water activity put together. The organism is dangerous and the infections are extremely rare; neither fact cancels the other.

What the Grand Teton study actually found

The 2024 paper in ACS ES&T Water by Stauffer and colleagues is the study that put "brain-eating amoeba" and "national park hot springs" in the same sentence. Its methods and its limits both matter.

The setup

Five sites, all popular for recreational soaking, across three thermal areas of Grand Teton National Park and the John D. Rockefeller Jr. Memorial Parkway: three at Huckleberry Hot Springs (H1, H2, H3), one at Kelly Warm Springs (K1), and one at Polecat Hot Springs (P1). Each site was sampled three times: July 2016 (midsummer), November 2016 (late autumn), and March 2017 (early spring). That is 15 site-visits, and every percentage below is out of 15. The samples were collected in 2016 and 2017; the paper was published in 2024.

The results

• Site H1 at Huckleberry: "N. fowleri was detected with real-time PCR from low-volume noncultured and large-volume cultured water samples at every time point." H1 had a mean water temperature of 41.8 °C.
• Kelly Warm Spring (K1) "had no detections from water samples and had the lowest average water temperatures (mean of 27.5 °C)."
• Polecat (P1) had the highest mean temperature, 42.3 °C.
• Huckleberry sites H2 and H3 swung with the season: H2 ranged 18.6–46.1 °C and H3 ranged 15.8–44.8 °C, with the maximum in July and the minimum in March.
• pH ran from 6.97 (H2, March 2017) to 8.01 (H3, July 2016), which the authors call "neutral to slightly alkaline."
• Detection rates across all sites and time points: water 60% (noncultured) and 67% (cultured); biofilm 20% and 20%; sediment 7% and 33%.
• E. coli turned up too: "present in most samples (>1 CFU/100 mL in 80% of samples collected) at low densities."

The authors' own summary: "These results provide the first detections of N. fowleri in Grand Teton National Park and provide new insights into the distribution of pathogenic N. fowleri and other nonpathogenic Naegleria spp. in natural thermal water systems in northern latitudes." Earlier work at Huckleberry and Polecat had already found that "Naegleria spp. isolates accounted for almost half of the amoebae observed," so the genus was known to be there. The 2024 paper is the first to report the pathogenic species, N. fowleri, in the park.

The sentence that matters more than the rest

This paper detected the organism in water, biofilm, and sediment. It reports no infection, no illness, and no case from those springs. Detection is not a case. A headline that says "brain-eating amoeba found in national park hot springs" is accurate. A reader who comes away thinking "people got sick at Grand Teton" has heard something the paper does not say. The sources cited here do not establish any PAM case contracted at Grand Teton, Yellowstone, or Hot Springs National Park.

The scope is narrow too: the findings cover five sites in one park, sampled at three time points, and describe what was in those pools on those trips. The sources cited here do not establish that any other specific hot spring, developed or wild, is safe or unsafe.

Why the 46-degree figure is not a safety rule

One number from the Grand Teton paper gets pulled out of context and turned into advice, so it needs its own section.

The paper states: "The laboratory-determined upper thermal limit for growth is 46 °C, and the amoeba can persist for short durations (24–96 h) at 48 °C." It adds: "The amoeba also can persist at lower temperatures with viable cells isolated from water distribution biofilm samples collected at 10 °C."

What 46 °C is: the temperature above which the amoeba stops multiplying, measured in a laboratory.

What 46 °C is not: a line above which water is amoeba-free, or a line below which water is dangerous. Three reasons, all from the same passage of the same paper.

1. Stopping growth is not the same as dying. The amoeba persists for one to four days at 48 °C, two degrees above its growth limit.
2. It tolerates cold. Viable cells have been isolated from biofilm at 10 °C, which is colder than the lowest temperature recorded at any of the five Grand Teton soaking sites (15.8 °C, at H3 in March).
3. The cited sources publish no temperature at which the amoeba is killed outright. Chasing a hotter pool to get above its range means chasing the temperatures behind Hazard 1 — the burns that Yellowstone attributes more than 20 deaths to.

The Kelly Warm Spring result cuts the other way and is just as easy to over-read. K1 had no detections in water, and it was the coolest site, with a mean of 27.5 °C. That is one site sampled three times. The sources cited here do not establish a temperature below which the amoeba is absent, and the 10 °C biofilm finding rules out treating temperatures down to 10 °C as reliably amoeba-free.

In short: the amoeba's temperature range overlaps the temperature range people like to soak in, at both ends. Temperature does not sort hot springs into safe and unsafe.

How exposure happens and what the illness looks like

The route matters more than anything else in this article, because it is the one thing a reader can act on. Both papers state it plainly. The Grand Teton paper: "PAM is primarily contracted through recreational activity with untreated water where N. fowleri enters the nose then migrates along the olfactory nerves and through the cribriform plate where it invades the central nervous system." The systematic review describes the same path in more detail: the amoeba "enters the nasal cavity, attaches to the nasal mucosa, burrows into it, and then travels through the cribriform plate and onwards along the olfactory nerves to the olfactory bulb."

The cited sources identify water entering the nose as the primary established route. That is why the review's two exposure histories are what they are: "Significant exposure to freshwater either from recreational activities or from a habit of irrigating the nostrils preceded onset." The recorded exposure histories were recreational freshwater contact — the swimming-at-a-lake-river-or-waterpark history that 18 of the 21 studies recorded — and nasal irrigation, meaning water deliberately put into the nose.

Speed is the other reason diagnosis is hard. The review: once the amoeba "enters the body, the onset of disease can be rapid, ranging from 2 to 8 days but sometimes quicker." The Grand Teton paper says PAM "usually results in death within 3–7 days after the onset of symptoms."

Symptoms, per the review: "The symptoms at early presentation included fever, headache, and vomiting, while late sequalae showed neurological manifestation." The diagnostic problem: "An accurate diagnosis remains a challenge, as the symptoms mimic bacterial meningitis."

The cited sources publish no per-exposure probability and no ranked list of protective behaviours. This article stops where they stop. What it can say is that both exposure histories in the review involve freshwater, and one of them involves putting water into the nose on purpose.

Hazard 3: Ordinary bacteria

The third worry is the one with the least evidence in this set of sources, so this section is short.

The National Park Service's safety page for Hot Springs National Park addresses it directly, in the context of visitors collecting the thermal water to drink: "The thermal water at Hot Springs National Park is nearly 4,400-year-old spring water in its natural state. The water's high temperature kills most harmful bacteria, and it is monitored to U.S. standards for safe drinking water. Bacteria may still be present and could affect those with compromised immune systems or other medical conditions if water is inhaled."

That statement applies only to Hot Springs National Park's monitored water. NPS says the heat kills "most harmful bacteria" and that the water is monitored to drinking-water standards, while cautioning that bacteria may still be present — with a specific caution for immunocompromised people and for inhalation. The same page adds a practical note for anyone filling a bottle: "the water's high temperatures could cause harmful chemicals to be released from some containers. Food grade storage containers that can handle hot water are recommended."

Hot Springs National Park monitors its water and says so. A wild pool at the end of a trail carries no such statement. The Grand Teton study gives one glimpse of what an unmonitored soaking site can look like: E. coli above 1 CFU/100 mL in 80% of samples, "at low densities." The authors report it as a measurement, not as an illness, and the sources cited here do not establish any bacterial illness from soaking at those sites.

The cited sources publish no count of bacterial infections from hot spring soaking and no comparison between developed and undeveloped springs. Whether a given pool has been tested by anyone is a separate question from what thermal water is used for elsewhere — our guides to what geothermal energy is used for and to geothermal energy in Iceland cover that side of it.

Putting the three hazards in order

Here is the ranking the evidence supports, with the number that justifies each position.

1. Scalding. More than 20 deaths at one park. Yellowstone's answer was to make bathing in thermal water illegal. The ground can be a thin crust. Runoff burns. Gases accumulate. This is the hazard with the measurable toll.
2. Naegleria fowleri. 0 to 6 US cases a year across 1978–2018, 85 total, from all water activities. 72.7% of reviewed cases fatal, and worldwide "between 1937 and 2018, there were 381 confirmed PAM cases worldwide and only 7 confirmed survivors," a record the authors note "may go undiagnosed" in some cases. Detected in water samples from recreational soaking sites at Grand Teton, in 60% of site-visits. The cited sources establish no PAM case contracted at those springs.
3. Bacteria. Present at low densities in 80% of Grand Teton samples. Monitored to drinking-water standards at Hot Springs National Park, where NPS still flags a risk for immunocompromised people if water is inhaled. The cited sources publish no illness count.

What the sources do let you act on:

• In Yellowstone, do not enter thermal water. It is illegal, and the death count is the reason given.
• Anywhere, treat the ground around thermal features, and the runoff from them, as burn hazards. "The ground surrounding these features may look solid."
• If you feel sick in a geyser basin, leave.
• Know what the PAM cases had in common: freshwater recreation, or nasal irrigation. Know the early symptoms: fever, headache, vomiting. Know that they mimic bacterial meningitis.
• If you are immunocompromised, the one park that publishes a bacterial caution ties it to inhaling the water.
• Collecting hot water to drink? Use food-grade containers rated for hot water.

What they do not let you do: pick a "safe" temperature, set a timer, or compute your odds. The sources cited here do not establish a safe temperature, a safe duration, or a per-soak probability, and this article will not pretend otherwise.

What We Could Not Source

The Centers for Disease Control and Prevention publishes pages on Naegleria fowleri. Every cdc.gov page tried for this article returned HTTP 403 from our environment, under two different browser identities. No CDC figure, quote, or guidance is used here, and CDC does not appear in the source list below. A 403 is a constraint, not a licence to paraphrase from memory.

Also not established by the sources cited here, and therefore not stated anywhere above:

• Any per-soak or per-visit probability of infection.
• Whether any specific commercial or developed hot spring is safe or unsafe. The Grand Teton work covers five sites in one park at three time points.
• A safe soaking temperature or a safe soaking duration.
• A temperature that makes water amoeba-free. The 46 °C figure is a laboratory growth limit, and the organism persists 24–96 hours at 48 °C.
• Any PAM case contracted at Grand Teton, Yellowstone, or Hot Springs National Park.
• Any count of bacterial illness from hot spring soaking.
• Anything about whether soaking is beneficial. That is a separate question with a separate evidence base, covered in our guide to whether hot springs are good for you.