Key takeaways
Three things to remember
- 01
Skin temperature deviation measures your overnight skin temperature relative to your personal baseline, not your absolute body temperature; what matters is the direction and persistence of the deviation, not the number itself.
- 02
The signal is most valuable as an early warning indicator: illness onset, training overload, and immune activation typically produce elevation 12 to 24 hours before you feel symptoms.
- 03
Alcohol, hard training, and luteal phase onset all produce elevation; reading temperature alongside HRV and resting heart rate is the only way to distinguish their causes.
How it works
The Mechanism
During sleep, the body actively transfers heat away from the core to the periphery. Skin temperature at the wrist or finger rises as the body cools the core, which is part of the mechanism that initiates and maintains deep sleep. This thermoregulatory process is tightly coupled to sleep staging: core body temperature drops during slow-wave sleep, and wrist skin temperature rises correspondingly.
When the body is fighting inflammation, whether from illness, intense training, or immune activation, it elevates core temperature slightly to accelerate immune function. This shows up in wrist skin temperature as a sustained overnight elevation above your personal baseline. The signal is often visible in wearable data 12 to 24 hours before subjective symptoms appear, making temperature deviation one of the earliest illness onset signals available from consumer devices.
Hormonal cycles also shift baseline skin temperature. In women, skin temperature rises by roughly 0.3 to 0.5 degrees Celsius during the luteal phase following ovulation and drops sharply just before menstruation. This is a normal circadian signal reflecting progesterone's thermogenic effect. Wearables that track this over months can identify the luteal phase with reasonable accuracy, which is useful for cycle tracking. Understanding the source of a temperature elevation, whether training load, illness, or normal hormonal cycle, requires reading it alongside HRV, resting heart rate, and sleep quality.
In practice
Why It Matters
Temperature deviation catches what your subjective sense of wellness misses.
Skin temperature deviation is most valuable as an early warning signal and as a pattern tracker over time. A single elevated reading after a hard training day or a poor night of sleep is expected and not concerning. Elevation persisting for three or more consecutive nights, especially alongside suppressed HRV and elevated resting heart rate, is a reliable signal of physiological stress that warrants reduced training load. For women, monthly patterns in temperature data can serve as a practical cycle phase indicator integrated directly into recovery interpretation.
Common misconception
Temperature deviation is not a measure of body temperature in the clinical sense. A deviation of plus 1.5 degrees Celsius does not mean you have a fever; it means your overnight skin temperature was 1.5 degrees above your personal norm. The absolute value is largely meaningless: what matters is the direction and persistence of the deviation from your own baseline.
Reference ranges
What a Healthy Range Looks Like
Below baseline
-1.5–-0.5°C
Uncommon at rest; may follow intense cold exposure or reflect extreme fatigue
Normal
-0.5–0.5°C
Within personal baseline variation; expected for most recovery nights
Slightly elevated
0.5–1.5°C
Mild physiological stress; common after hard training, alcohol, or luteal phase onset
Elevated
>1.5°C
Significant physiological stress; illness onset, intense training block, or sustained immune activation
Compare to your own established baseline, not to any published absolute range. Normal variation is roughly plus or minus 0.5 degrees Celsius night to night. Sustained elevation above 1.0 degrees for three or more nights warrants attention regardless of the absolute value.
Signs it is disrupted
- Three or more consecutive nights above your personal baseline alongside suppressed HRV
- Temperature elevation appearing 12 to 24 hours before cold or flu symptoms develop
- Persistent elevation after a training block that does not resolve during a recovery week
- Regular monthly pattern of elevation lasting 10 to 14 days in the second half of the cycle in women
What moves it
How to Improve It
Cool sleep environment. Sleeping in a room at 65 to 68 degrees Fahrenheit supports the core body cooling process that makes skin temperature deviation readings more stable and reduces night-to-night variation.
Reduce training load. When skin temperature has been elevated for more than two consecutive nights alongside suppressed HRV, reducing training volume by 40 to 50% for three to five days allows immune and recovery processes to resolve.
Alcohol elimination. Even one to two drinks elevate overnight skin temperature by blocking the normal thermoregulatory cooling process, producing a reliable false-positive elevation in recovery scores.
Consistent sleep timing. Irregular sleep timing disrupts the body's temperature-based sleep initiation process, adding noise to baseline temperature readings and making deviation signals harder to interpret.
Where to see it
Which Devices Track It
- Oura Ring
- Oura measures skin temperature at the finger, reports nightly deviation from a 30-day rolling personal baseline, and uses it as a direct input to the Readiness Score; the finger is a high-signal location for thermoregulatory measurement.
- WHOOP
- WHOOP measures skin temperature at the wrist and tracks deviation from a recent rolling baseline; it is a contributing factor to WHOOP's Recovery Score but is not displayed as a standalone metric by default.
- Apple Watch
- Apple Watch Series 8 and later measures wrist temperature during sleep and reports nightly baseline deviation; values are displayed in the Health app and used for cycle tracking, but Apple does not surface them as a recovery signal in Vitals.
- Garmin
- Garmin watches with Body Battery measure skin temperature during sleep as one input to Body Battery energy estimation, though temperature deviation is not reported as a standalone metric on most models.
Go deeper