At this writing coastal Bangladesh is still picking up the pieces from the devastation wreaked by Cyclone Sidr, a tremendously powerful tropical cyclone that packed peak winds of at least 250 km/h (however public attention seems to have fallen off). I've noted earlier that such a storm gets its power from the release of latent heat on condensation of water vapor to liquid water. The East coast is getting much needed rain, while Southern California is seeing a return of hot dry winds (the Santa Ana). It turns out that water in the atmosphere is maddeningly difficult to predict. It is also crucial to climate and ecology. The first thing to do, then, is to describe it.
The amount of water vapor in the atmosphere is generically known as the humidity. Which has the higher humidity, South Pole or Hawaii?
Since nothing is easy when it comes to water, there are at least 4 common measures of humidity, all of them useful under some conditions, none of them ideal for all situations. They are:
- specific humidity
- relative humidity
- dewpoint temperature
- mixing ratio
In the lab, one also uses the vapor pressure. A couple of other things to keep in mind before going on. Air acts as an ideal gas, more or less, at least for our purposes, so is subject to the ideal gas law: PV = nRT.
Here P = pressure, where standard sea level pressure is 1013.25 mbar, or, if you are an SI purist, 101325 Pascals. T is temperature in Kelvins, n the number of moles of air (basically the amount of air you have. A mole is either a subterranean rodent or a number of things, like "couple" or "dozen" or "gross". In this case, a mole is not equal to 2 or 12 or 144 things but 6.023 x1023 things. 1 mole of air has 6.023 x1023 molecules of air and at 273K or 0o Celsius and 1013.25 mbar pressure, takes up 22.4 liters). R is the ideal gas constant, 8.314 J/(mole K) (it is equal to Nk, where k is Boltzmann's constant and N is 6.023 x1023, if you prefer).
Rewrite as V = nRT/P, so as P decreases or T increases, the volume increases. Pressure falls off with altitude faster than temperature does, so a rising air parcel expands and cools, a sinking air parcel compresses and warms. Remember the wildfires in S. California? That was fanned by the Santa Ana, air flowing downslope (sinking) and compressing and warming as it did so.
Specific Humidity
This is how much water vapor you have in a volume of air, units are typically grams (of water vapor)/cubic meter (of air). This is clearly a measure of how much water vapor you have in an air parcel, but the volume of this air parcel changes as it rises and sinks even as the amount of stuff in that parcel is constant. We say that specific humidity is not conserved with respect to vertical motion. This is also known as absolute humidity.
Mixing Ratio
This is the amount of water vapor you have in an amount of air in the same units, such as grams (water vapor)/kilograms (air), or volume (water vapor)/volume (air). In the case of volume mixing ratio, we usually use terms like percent, parts-per-million, parts-per-billion, etc. 1 percent mixing ratio means that for every 100 molecules of air, 1 is water (or whatever). 380 ppm CO2 means that for every million molecules of air, 380 are CO2. Etc. Mixing ratio is conserved with vertical motion, so it's quite convenient in that respect.
Those were the easy ones. Because water can (and somewhere always does) exist in all three phases in the atmosphere, we have to consider the concept of saturation, which is the amount of water vapor an air parcel can hold before that water vapor condenses (or deposits). That amount is strongly temperature dependent, and ~doubles every 10o C. So an air parcel at 0o C can hold 3.5 g water vapor/kg air; at 10o C 7 g/kg; at 20o C 14 g/kg; etc -- here's a nice graphic. This brings us to
Relative Humidity
This is the ratio of water vapor that is actually in the air parcel compared to how much the air parcel can hold at the temperature of the air parcel. An air parcel at 20o C with 7 g/kg absolute humidity has 50% RH. Cool that parcel down to 10o C and the relative humidity goes to 100% -- although the amount of water hasn't changed. Cool that down to 0 and you will have to lose 3.5 g/kg, in the form of condensation. Cold air at 95+% RH may be much drier than warm air at 70% RH. This is why winter air in cold regions is so dry -- the air can't hold much water vapor. RH tells you nothing about how much water vapor is in the air, it tells you how much the parcel has compared to how much it could take at that temperature.
Why are we interested in this? In part it's because we're sensitive to both absolute humidity and to relative. If it's warm out and RH is high, evaporative cooling is ineffective and it feels "sticky". Mostly, though, we're interested in saturation. When RH reaches 100% (in the presence of tiny "condensation nuclei"; in the absence of CN water supersaturates quite readily) the water vapor turns into liquid water. This means clouds, release of latent heat, and possible precipitation. It also affects vertical motion in the atmosphere: parcels with high RH that are forced to lift will condense readily, releasing latent heat, which makes them more buoyant still.
Dewpoint Temperature
This is the temperature at which saturation occurs in that air parcel (if it is below zero it's technically the frostpoint temperature). The closer the dewpoint T is to the actual temperature, the closer to saturation and the higher the relative humidity. High dewpoint temperatures also indicate high absolute humidity. Dewpoint is usually how humidity is described on weather maps (image from weather.unisys.com):
The image is of a large midlatitude cyclone centered over Wisconsin at the time of the report. Each station reports 3 numbers (a full report has many more numbers and symbols, BTW). The upper left is temperature, in F (the rest of the world uses C of course), the lower left is the dewpoint temperature. So Little Rock, AR was showing 60F with a dewpoint T of 50F; NOLA was reporting 77/74; NYC 57/57. Of the three Little Rock had the lowest RH and absolute humidity; NYC the highest RH (100%); and NOLA the highest absolute humidity.
Unlike other gases in high concentrations (N2, O2, CO2, eg), water vapor distributions are far from uniform -- water vapor is not well-mixed. This is because it has a short residence time -- about 1-2 weeks -- and because its concentration feeds back on its lifetime. Humid airmasses are less dense than dry airmasses (dry air has a molecular weight of 28.9 g/mol; water has a molecular weight of 18 g/mol, so a moist air mass will have a lighter component absent in dry air) which means moist airmasses are easily lifted. As such an airmass lifts, the temperature drops, and when the temperature reaches the dewpoint, we get condensation. This releases latent heat, which causes the parcel to be more buoyant, etc etc. If that condensation leads to precipitation, water leaves the atmosphere. Here's a typical "sounding", this one from Miami (taken from the same website above)
This is a thermodynamic or skew-T diagram. There's a lot of info here, which will have to await another post. The things to note are that temperature is plotted at a 45o angle (the diagonal blue lines) vs altitude, in both meters and pressure (pressure is a common proxy for altitude -- it drops off logarithmically with altitude) and that the two white lines are atmospheric temperature on the right and dewpoint on the left. The closer the two lines are the closer you are to saturation. We see that the surface layer is relatively moist, and RH actually reaches 100% at 800 mbar (2065 m), but just above that is a very dry layer throughout most of the middle atmosphere (at 700 mbar the temperature is 9C and the dewpoint -35C). From the wind flags on the right it is clear that the surface layer is coming from the east, whereas the dry layer is coming from the west. I said this is "typical", but not in the sense that this is what you usually see over Miami, rather that water vapor is poorly mixed. Here's a range of soundings from across the US.
Back to the picture at the top: which has higher humidity? Hawaii, of course, by far. But the relative humidity was higher at South Pole.
Change
With all that in mind, let's think about what happens if global temperatures increase. The first thing to note is that evaporation will increase so absolute humidity (expressed either as mixing ratio or specific humidity) will increase. The second thing to note is that it makes no statement about relative humidity, or how close we are to saturation, and therefore condensation and precipitation. In fact, RH remains pretty much conserved at 70-80% (actually it appears to drop a little at the tropopause and in the lower stratosphere), but even that doesn't tell us whether a warming world is a drier one or a wetter: RH can be maintained by an active water cycle with high rates of evaporation and precipitation, or by one with low evaporation and precipitation. But it makes a difference to the temperature feedback term: water vapor is the primary GHG (its lifetime is so short that it acts as a feedback rather than a forcing, however) and greater absolute humidity means greater surface heating ... to a point. At saturation, condensation may occur and generate clouds, which have high albedo and will contribute a net cooling term. Precipitation will also remove water from the atmosphere. So the sign (not just the magnitude) of the water vapor feedback is not easy to predict. Clouds remain among the most difficult things to get right in any climate model.
One thing we can say is that with greater absolute humidity, very cold regions will see more precip, so you would expect to see the East Antarctic ice sheet and the middle of Greenland accumulate ice, even as the edges lose ice, which is exactly what is happening. Here's an image of Antarctic temperature trends. The gains, alas, do not appear to be enough to compensate for the losses, either in Antarctica or in Greenland
It turns out that warming introduces changes which act in opposite directions. The first is that the greater absolute humidity will increase greenhouse warming. Because 80% of the atmosphere by mass is in the troposphere, to maintain energy balance there will be a cooling of the lower stratosphere. Like this (from here):
The line marked "lower troposphere" does not correspond to the surface record, which warms even more strongly. Part of the cooling of the lower stratosphere (LS), something over half, is not due to energy balance (to first order anyway), but to increased humidification of the LS. This destroys ozone, which would otherwise absorb some of the UV radiation and convert that to heat. The increased water vapor in the LS is partly because of increased methane, which produces water when oxidized, and partly a direct increase in water vapor (because of greater humidity in the lower atmosphere, in which increased biomass burning also appears to play a role).
Anyway, the upshot is that the atmosphere becomes warmer at the surface, and colder at the top. This steepens the lapse rate and decreases stability: warm air rises so if the temperature falls off more rapidly with altitude the air column will increasingly want to turn over (whether it actually will or not is not so easy to say and will have to await yet another post).
The second factor is a result of the ice-albedo feedback. The Arctic high latitudes will warm up relatively more than temperate regions as the ice melts (the Arctic has far less ice than Antarctica and is sea ice so it's sensitive to water temperature). Here's a projection from the IPCC's AR4:
The various figures are for low, medium, and high emission scenarios in the 2020-2029 and 2090-2099 decade ranges (we're currently approaching if not actually higher than the high scenario). Note the strong warming trends at the poles, especially the North pole. This is the ice albedo feedback at work -- because there is so much ice in the Antarctic, the South polar region responds more slowly than the North.
Now the Earth receives more energy from the sun at the equator than it radiates, and radiates more than it receives at the poles (good figure here, the parent article is also well worth reading). This leads to poleward heat transport across the midlatitudes, which is what generates midlatitude weather systems, especially the midlatitude cyclone. Lifting of air masses in the cyclone lead to saturation, condensation, and precipitation. If warming leads to a decrease in this thermal gradient (as it certainly will) the loss of thermal contrast will tend to suppress midlatitude cyclone formation, even as the atmosphere humidifies. Right now we think (PDF alert) that will lead to fewer but stronger weather systems in midlatitudes (the contiguous US is dominated by midlat systems). The net result will be lower precipitation (IOW drying out) throughout most of the US, especially in the West. Or so we think: this is not one of the most robust of projections.
OK, that got a little long. I hope you at least appreciate that the behavior of water vapor is complex and that there are a number of ways -- more often than not confused in the general media -- to describe humidity. I touched on its importance to climate, but it is also key to vegetation zones, soil types, ecological niches, etc.
Next up: the Redfield ratio (unless it's something else)