I'll admit I'm just speculating because I don't know much about weather and weather data collection, but is it not possible there's a 3 degree difference between 2 different places of measurement in Tokyo (i.e. both are correct to some degree?)
> while high resolution current weather information from Japan Metrological Agency
They are complaining about location precision. It might be very coarse (equivalent to US zip code size). It’s a valid complaint. I use Dark Sky on iOS because it’s accurate to 10 meters or so.
