This is the second part of the Sources of Neutrinos series. In Part 1 we covered solar neutrinos, the largest natural neutrino flux on Earth. We now turn to a much higher-energy source that bombards the planet constantly: neutrinos produced by cosmic rays striking the atmosphere.
How they’re made
Cosmic rays are mostly high-energy protons (about 90%) plus heavier nuclei (helium, carbon, iron). They arrive at the top of Earth’s atmosphere with energies ranging from a few hundred MeV up to 10²⁰ eV (the highest-energy known particles). Their flux: roughly 1,000 per square meter per second at the top of the atmosphere across all energies.
A cosmic-ray proton entering the atmosphere typically interacts with a nitrogen or oxygen nucleus at about 15 km altitude. The collision is hadronic — it produces pions ($\pi^\pm$), kaons ($K^\pm$, $K^0$), and other secondaries. These charged mesons then decay: $$\pi^+ \to \mu^+ + \nu_\mu, \quad \pi^- \to \mu^- + \bar\nu_\mu$$
The muons themselves then decay (with a roughly 2.2 microsecond lifetime in their rest frame): $$\mu^+ \to e^+ + \nu_e + \bar\nu_\mu, \quad \mu^- \to e^- + \bar\nu_e + \nu_\mu$$
For every charged pion that decays in flight, you get one muon plus one muon neutrino. If the muon then also decays in flight, you get an electron, an electron neutrino, and another muon neutrino. Total per pion decay chain: two muon neutrinos and one electron neutrino. The expected ratio is therefore $\nu_\mu / \nu_e = 2$ — at energies where muons mostly decay before hitting the ground.
The energy spectrum
Atmospheric neutrinos span roughly 100 MeV to 100 TeV — six orders of magnitude. The flux peaks near 1 GeV, with the spectrum falling steeply at higher energies because the parent cosmic rays do too.
At low energies (sub-GeV), the muons have time to decay before reaching the ground, so the $\nu_\mu / \nu_e$ ratio is close to 2.
At higher energies (above ~30 GeV), muons reach the ground before decaying — they’re relativistic and time-dilated. The neutrinos that do arrive are mostly direct $\nu_\mu$ from pion and kaon decay. The ratio rises substantially above 2.
Above ~100 TeV, “prompt” neutrinos from heavy-flavor (charm, bottom) hadron decays become important. These produce a roughly equal mix of $\nu_e$ and $\nu_\mu$ and are theoretically predicted but not yet conclusively observed.
Why path length matters
Cosmic rays strike Earth’s atmosphere uniformly from all directions. Atmospheric neutrinos arrive at a detector from all directions too — including up from below, after having traveled through the Earth itself.
Detector geometry gives you a natural distance ruler:
- Directly down-going neutrinos travel ~15 km from their production point in the upper atmosphere to the detector.
- Horizontally moving neutrinos travel a few hundred kilometers.
- Up-going neutrinos travel through the Earth — up to about 12,800 km for a neutrino entering on the opposite side of the planet.
Combined with the wide energy spectrum, this gives an extraordinary range of L/E (path length over energy) — exactly the variable that drives oscillation.
The 1998 discovery
Super-Kamiokande’s 1998 measurement of atmospheric neutrinos was a turning point in particle physics. The collaboration compared the rates of up-going vs down-going muon neutrinos. If neutrinos didn’t oscillate, the rates should be equal — cosmic rays come from all directions equally, and Earth is essentially transparent to ~GeV neutrinos.
The data showed a clear asymmetry: muon neutrinos arriving from below were significantly suppressed compared to those arriving from above. The suppression depended on energy and direction in exactly the way predicted by neutrino oscillation. The announcement at Neutrino ‘98 in Takayama, Japan, established neutrino oscillation as a real phenomenon and won Takaaki Kajita a share of the 2015 Nobel Prize.
The interpretation: muon neutrinos from below have traveled thousands of kilometers — long enough at GeV energies to oscillate into tau neutrinos, which Super-K mostly couldn’t see at the time. Muon neutrinos from above traveled only ~15 km and hadn’t yet oscillated.
Atmospheric neutrinos today
Super-K continues to measure atmospheric neutrinos and refine the oscillation parameters. Its combined data set is decades long.
IceCube’s DeepCore — a denser sub-array of strings near the center of IceCube — is optimized for the ~10-100 GeV atmospheric range. IceCube has produced atmospheric-neutrino oscillation measurements competitive with Super-K, with very different systematics (the South Polar ice vs ultra-pure water).
The atmospheric-neutrino flux is also the dominant background for accelerator long-baseline experiments. T2K and NOvA both have to model and subtract atmospheric backgrounds carefully to extract their oscillation signal from the beam. The same is true for DUNE and Hyper-Kamiokande.
What’s next
The next generation will exploit atmospheric neutrinos for independent measurements of the neutrino mass ordering. The MSW effect alters the oscillation pattern for atmospheric neutrinos passing through Earth’s core. The sign of the matter effect depends on the mass ordering: enhanced for normal ordering with neutrinos and inverted ordering with antineutrinos, suppressed in the opposite cases.
IceCube-Upgrade (planned for 2027 deployment) will instrument the existing DeepCore region more densely for atmospheric-neutrino mass-ordering measurements. KM3NeT/ORCA in the Mediterranean is similarly optimized. Both should produce competitive mass-ordering measurements by the early 2030s, complementary to JUNO’s reactor-baseline approach.
The next part of this series turns to a man-made neutrino source: nuclear reactors.
Frequently asked
What are atmospheric neutrinos?
Neutrinos produced when cosmic rays — mostly protons — strike nuclei in Earth's upper atmosphere. The collisions produce pions and kaons that decay into muons and neutrinos. The muons in turn decay into electrons and more neutrinos. The result is a steady drizzle of muon and electron neutrinos with energies ranging from sub-GeV to TeV, arriving from all directions.
What energies do atmospheric neutrinos have?
The atmospheric neutrino spectrum spans about six orders of magnitude — from 100 MeV up to 100 TeV. The flux peaks around 1 GeV. The spectrum is well-modeled by cosmic-ray flux measurements combined with hadronic cascade simulations, with about 10–20% uncertainty across most of the range.
Why are atmospheric neutrinos useful for oscillation studies?
They naturally provide two key variables: a wide range of energies (sub-GeV to TeV) and a wide range of path lengths (10 km from directly overhead to 12,800 km from straight down through the Earth). The combination L/E spans almost five decades, exactly the range needed to map the oscillation pattern. Atmospheric data established neutrino oscillation in 1998 at Super-Kamiokande.
Are atmospheric neutrinos still important today?
Yes. Super-Kamiokande and IceCube continue to measure atmospheric neutrinos for high-precision oscillation parameter fits. They also provide the main background that long-baseline accelerator experiments like T2K and NOvA must subtract. The next generation — DUNE, Hyper-K, IceCube-Upgrade — will use atmospheric neutrinos for independent measurements of the mass ordering.
Cite this article 5 formats
APA
Neutrino Times Editorial Team. (2026, February 22). Sources of Neutrinos — Part 2: Atmospheric neutrinos and cosmic rays raining down. Neutrino Times. https://neutrino-times.com/articles/sources-of-neutrinos-part-2-atmospheric/
Chicago
Neutrino Times Editorial Team. "Sources of Neutrinos — Part 2: Atmospheric neutrinos and cosmic rays raining down." Neutrino Times, February 22, 2026. https://neutrino-times.com/articles/sources-of-neutrinos-part-2-atmospheric/.
MLA
Neutrino Times Editorial Team. "Sources of Neutrinos — Part 2: Atmospheric neutrinos and cosmic rays raining down." Neutrino Times, 22 Feb. 2026, https://neutrino-times.com/articles/sources-of-neutrinos-part-2-atmospheric/.
BibTeX
@misc{neutrino-times-sources-of-neutrinos-part-2-atmospheric,
author = {Neutrino Times Editorial Team},
title = {Sources of Neutrinos — Part 2: Atmospheric neutrinos and cosmic rays raining down},
howpublished = {Neutrino Times},
year = {2026},
month = {feb},
url = {https://neutrino-times.com/articles/sources-of-neutrinos-part-2-atmospheric/},
note = {Accessed: 2026-02-22}
} RIS
TY - GEN TI - Sources of Neutrinos — Part 2: Atmospheric neutrinos and cosmic rays raining down AU - Neutrino Times Editorial Team PY - 2026 DA - 2026-02-22 PB - Neutrino Times UR - https://neutrino-times.com/articles/sources-of-neutrinos-part-2-atmospheric/ ER -