Water in the atmosphere is present as vapour, liquid and ice. It is just a small amount of the Earth's total water supply. I'm wondering how long it will take to rain down all the water from the atmosphere and "replace" it with new water by evapotranspiration.

This will give an indication for the maximum accuracy of weather forecasting while atmospheric water plays an essential role in climatological conditions.


The term you are looking for is called residence time or residency time. This UCAR Science article suggests it is about 9 days on average. This should be verifiable by dividing the average global precipitable water by the average global daily precipitation.

I don't see any reason why the residency time of water would offer any hint of weather predictability limits. If the equations to predict evapotranspiration used in forecast models have skill, shouldn't the 9 days not be the limit (and how limiting a factor it is would depend upon how skillful those equations are)? The residence times of vehicles on roadways is probably near/less than an hour, yet we can forecast traffic quite reasonably well in advanced in at least many circumstances (and that's prone to the much more volatile nature of human choice/error)!

Indeed I think the "residence time" of energy in the lower troposphere, if such an idea even has sense, may well be shorter than a day for the Earth given the continual heavy IR absorption and reemission rates in the atmosphere, and the dominant input being solar energy. But we can forecast much longer periods than that well. We have great understanding of insolation into the Earth-air system, and then quite solid skill at predicting its transfer, even with the complexity of clouds and other feedbacks. So we can outdo such a timescale even if energy passes through the system indeed quite quickly. Likewise, there seems to be no reason for the 9 days to be a particularly meaningful limit on the atmosphere.

  • 1
    $\begingroup$ Thanks for your answer. I now understand that the meteorological equations in weather/climate models make reliable (water) balances over a longer time than the residence time in the atmosphere. I'm trying to identify all uncertainties in flood protection systems and the weather forecast is an important factor. $\endgroup$ – 3TW3 Aug 23 '17 at 7:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.