Temperature falling to dewpoint

[Dan Robinson]

I did some searching on this simple question but haven't been able to find anything concrete.

I've always assumed that the temperature of air cannot fall below its dewpoint, due to latent heat warming when saturation occurs. In other words, the dewpoint is always less than or equal to the temperature. So, the temperature can only continue falling if the dewpoint falls (due to mixing). Is this true?

Applying this in a practical sense - assuming a low T/Td spread, little to no advection or mixing and good radiational cooling (clear skies), can a forecast low overnight temperature be estimated using the observed dewpoints?

For example, if the surface T/Td is 40/31F at 8PM in October with light winds and clear skies, you can normally count on frost. So, if observations show 40/34F with light winds and clear skies, can you safely say that frost is unlikely - IE, can you assume the temp will 'bottom out' at 34F and go no lower, even if 34F is reached at midnight?

 

[Jeff Snyder]

You are correct in saying that, at the surface, T >= Td (or RH <= 100%). For what it's worth, RH can get to 102-103% (supersaturation) in clouds, particularly in those with strong vertical velocities (or extremely clean air lol). With conditions favorable for strong radiational cooling -- clear skies and light winds -- the temperature will usually fall to the dewpoint. At that time, RH is 100%, and some of the water vapor will begin to condense out of the air. As this happens, the amount of water vapor "in" the air decreases slightly, lowering the Td. With a slightly lower Td, the RH dips <100%, thereby letting the temperature fall a bit more. In addition, the Td tends to decrease a bit at night anyway, owing largely to the ceasation of evapotranspiration (and regular evaporation should you be near a lake, river, or wet grounds), though this isn't a bit contributor during the late fall or winter anyway. So, if you have 41/30 T/Td, with clear skies and calm winds, the temp will likely fall to 30, while the Td drops a degree of two (except during winter, when vegetation isn't active), so you'd end up with, say, 29/29. The temp and Td will slowly fall in unison until near sunrise, probably ending near 27. This is a semi-reliable forecast tool, as long as you are forecasting light winds and clear skies. Of course, heat island and other effects can mess with this a bit.

 

[rdale]

"can you safely say that frost is unlikely - IE, can you assume the temp will 'bottom out' at 34F and go no lower, even if 34F is reached at midnight?"

Nope... Remember we're measuring the surface temp at 2 meters - 6 feet. So a "surface" temp of 34 is more than capable of being colder at the actual surface, so frost is a pretty good bet.

- Rob

 

[Doug Raflik]

Plus remember that there are hills and valleys throughout the land (well, maybe not in a good chunk of the plains) and colder air will settle in these valleys.

 

[Andrew Khan]

Plus remember that there are hills and valleys throughout the land (well, maybe not in a good chunk of the plains) and colder air will settle in these valleys.

This process actually takes place right here in the region that I live in...we live in sort of 'a valley'..

 

[Brian Boothe]

Here in Tallahassee, in the winter and spring we typically have the same occurance. The official station is out at the airport which is about 5 miles from downtown (as a crow flies) and in a little valley. Sometimes its really tough forecasting low temperatures due to the dewpoint dropping so much during the night. We've watched the temperature drop to saturation, and then watch the dewpoint and temperature drop together, sometimes 5-10 deg. Its rather frustrating sometimes.

 

[Morgan Palmer]

Of course, most airports are actually in more rural or suburban areas. Even for small towns, there could be enough heat generated (a minor heat island) to keep temps just a couple of degrees warmer. That's often all that separates frost from no-frost.

Obviously, if you say some people will get frost and some don't, they'll think you're wrong. That's a lot better than not forecasting frost and some does happen.

P.S. Austin Bergstrom Int'l Airport in Texas has unfortunately changed the climate records/etc quite a bit because it's often a good 10-15 degs colder there at night (well outside of town) than at the old airport right in the heart of the city. Fortunately, they've added a regular reporting station (Austin City) at Camp Mabry on the west side of town, but still a lot more representative of true Austin conditions than ABIA.

MP

 

[rdale]

Deleted...

 

[Morgan Palmer]

heh. perhaps the comment was misunderstood. I am in favor of forecasting frost even if the chances are small. I agree with you, Rob. Maybe I misstated my point, but I can't tell that I did!

 

[rdale]

Nevermind - I see what you meant now...

 

[Dan Robinson]

Thanks for the info guys! That makes sense that a T/Td would slowly fall together in unison through the night even after saturation was reached.