It's the temperature inside the watch, and that temperature is needed for the baro-altimeter. The watch is warmed by your body, and insulated if you wear long sleeves.
It always amazes me that people expect a watch worn on the body will somehow reflect the external temperature. A sad indictment on however many years of schooling, and people remain unaware of the concept of body heat.
As jim mentions, the point of the internal temperature sensor is to calibrate the barometer (temperature affects air pressure).
The internal HR sensor is equally useless...a random number generator really...but it basically is what it is. I was furious with mine for awhile because it would refuse to bind to a Garmin HR strap (actually multiple HR straps) but I have since figured a way to do it with general success so I no longer hate the watch. Kind of like it actually.
It always amazes me that people expect a watch worn on the body will somehow reflect the external temperature. A sad indictment on however many years of schooling, and people remain unaware of the concept of body heat.
As jim mentions, the point of the internal temperature sensor is to calibrate the barometer (temperature affects air pressure).
Also a few degree can be accepted, but now in every training it shows totally wrong temperature curve in the report. Just now, running 17km, temperature was always 5degree, and the curve shows value between 21 and 14°C.
2. If to use for calibrate the barometer, then it is even worse use the wrong temperature. Totally illogical answer.
1.The curve always seems right to me. A downward shaping curve as the face of the watch gradually goes from room temperature to outside temperature when cold, and an upward sloping curve when the ambient external temperature is higher than room temperature. The back of the watch presumably remains close to 37C (body temperature), while the internal temperature of the watch remains somewhere in between that of the face and the back.
2. The internal barometer sensor and the air inside the barometer port being sampled presumably are at the temperature of the watch itself. And that is what it should be calibrated to. As an M. Eng, I am sure you could tell me what the logical result would be if you used the external temperature to calibrate a barometric sensor that is located INSIDE the watch. We could have discussion about the speed of gas exchange through the small port opening, and what this might mean in terms of air temperature WITHIN the watch, but there is not enough published data in the specs to reach a conclusion.
If you want to know the external temperature, use a Tempe sensor. When cycling rather than running, my F5 is mounted on the handlebars and records temperatures much closer to actual. It does record slightly higher temperatures than official metereological observations when in direct sunlight, but this isn't surprising.
The way these barometers work is by measuring the resistance of pressure sensitive micro resistors inside a sealed box with a membrane that flexes with the external air pressure, so they're air (and water) tight but pressure sensitive as a whole. The resistance is sensitive to temperature as well, so the unit contains a temperature sensor to compensate for that, and that stage is precise. I think it's likely that that's factored into the elevation calculation as well, which will be compromised, but it isn't the only reason for the sensor.
I think it's likely that that's factored into the elevation calculation as well.
I have seen that claim often in these forums, but I have not been able to reproduce it on my Fenix 3.
At unchanged altitude reading, the difference between ambient pressure and barometric pressure is the same at any temperature I have tried. But I have not tried more than 100 m above sea level, so the watch may use a different calculation at higher altitudes.
Also, I do not think it would be sensible to do such a correction in the elevation calculation since the temperature of the watch may be a worse approximation to the actual outside temperature, compared to the watch just relying on the standardized temperatures in the ISA standard atmosphere.
Thanks, that's interesting. People have said that the error increases as you get further from the calibration elevation, which would make sense if it isn't trying to use temperature, which it probably should if you're high enough for the temperature to fall. Error prone, though, as I said before.
