Thursday, August 9, 2012

What Causes Hurricanes (and Some Mars Dust Storms)?

It's not uncommon that I hear incorrect explanations for the origin or behavior of various atmospheric phenomena.  This post was inspired by a recent event where an individual insisted that "extreme pressure variations" are needed to produce a hurricane.  Below, I describe the actual mechanisms that produce a hurricane.  It turns out the same physics may also be responsible for producing some Mars dust storms.

Consider first a motionless atmosphere with no horizontal pressure variations at all.  That is to say, there is no wind and if you were to look at a weather map you would see no high or low pressures.  It is from this state that hurricanes (also known as tropical cyclones) develop.  Immediately, you can see that if this is indeed the initial scenario, then the assertion that "extreme pressure variations" are necessary is exactly wrong.

In reality, there is no such idealized state in the atmosphere.  There is always some amount of pressure variation and therefore some amount of wind.  The process I will describe below works just as well as long as these winds and pressure variations are small.

From this rather boring initial state, heat a chunk of air near the surface from the sun.  It becomes buoyant like a hot air balloon, and rises.  In the tropics, it is not unusual that such rising motion will develop into individual thunderstorms.  Inside a thunderstorm, water vapor condenses, which releases heat.  This heating causes the air to expand.  (Slightly technical here: this is an adiabatic expansion, meaning that no net energy is being added; even though there is heating due to condensation, it is internal heating and represents no external energy input).  This expansion has the effect of causing a weak low pressure to develop near the surface and weak high pressure to develop aloft.  Once there is a low pressure at the surface, horizontal winds develop and begin to flow toward the storm.

If the thunderstorm is over water, the wind now flowing toward the storm interacts with the underlying ocean.  The interaction is actually quite complex, but I'll distill the essential and relevant parts here.  If the ocean is warmer than the air, the air will be heated through turbulent eddies transporting the heat of the ocean upward.  The strength of the turbulent eddies and therefore the efficiency of the exchange is related to the wind speed and the difference in temperature between the atmosphere and the ocean surface.  If the wind is strong (all other things being equal) the exchange will be strong.  If the temperature difference is strong (all other things being equal) the exchange will be strong.

In the case we have so far, the winds are rather weak, so the exchange is rather weak.  Nonetheless, energy is being added to the atmosphere and carried toward our thunderstorm. (Actually, it is entropy that is being transported, but it's not important at this level of discussion, and I digress...)  As the air flows toward the weak low pressure, it expands.  Normally, this expansion would cause cooling.  However, the warm ocean underneath keeps the temperature of the air nearly constant as it expands.  In other words, the heat the air receives from the ocean nearly balances the cooling due to expansion.  This is called an isothermal (same temperature) expansion process.

Once at the storm, the air rises, condenses, and heats the atmosphere.  This strengthens the low pressure at the surface.  In turn, this increases the wind speed flowing toward the storm.  And this in turn increases the heat exchange between the atmosphere and the ocean (recall that the magnitude of the exchange is related to the wind speed).

It should be clear that what we have is a feedback process.   It's a process that reinforces itself.  As more energy is picked up by the air flowing towards the storm, the more heating is produced in the storm, the lower the pressure drops, and the stronger the winds get.  If you look at the entire system of air flowing toward the storm, air rising in the storm, air flowing out of the storm aloft, and then sinking away from the storm, you have a complete cycle.  Approximately 30 years ago the thermodynamics of this system was recognized in two landmark papers (you can get them here and here).  It turns out to be a sort of Carnot engine.  The process was called "Wind-Induced Sensible Heat Exchange", because it coupled the wind to the exchange of heat energy with the ocean.  While the fine details of this mechanism are still being debated in the literature, the overall process is now well established.   In essence, the engine extracts heat from a source, in this case the ocean and turns it into mechanical work (in this case increasing wind speed).

For simplicity sake, I've left out a lot of details, but these are not relevant to the big picture here.  There is at least one detail that is relevant:  The Earth rotates.  This means that as the air moves, it will experience an apparent deflection.  If the movement is over a sufficiently long time, this will result in the air beginning to rotate around the low pressure rather than flowing toward it.  If nothing else were to happen, there would be no air flowing to the storm, just going around it.  That would be the end of the storm.  But, there is friction.  Friction keeps some component of the motion always flowing toward the low pressure.  Thus, the feedback process may continue.  Furthermore, because friction is also a function of wind speed, the strong the winds, the stronger the friction, which dictates how much energy can flow to the storm.

The above process is the basic concept of how hurricanes develop.   No "extreme pressure variations" are needed.  The storm can develop from a completely motionless  atmosphere with no pressure variation at all.  It actually depends on the lack of an initial pressure variation.

Why don't all storms develop into hurricanes?  There's lots of things that can mess up the feedback process.  One is that the system moves over land.  Then there is no heat exchange with the ocean.  Another is that the initial storm can be too close to the equator.  The deflection of winds at the equator is zero, so no rotation can develop.  If the storm moves to higher latitudes, the ocean can become too cold and this limits the energy that can be provided.  Another major hindrance to storm formation is the presence of strong wind shear.  Wind shear is a change of wind speed or direction with height.

Here's the thing about wind shear:  Wind shear results when there are strong pressure variations.  For example, the jet stream is a result of strong large scale pressure differences.  Hurricanes and jet streams just don't get along.  The strong winds that result from the strong pressure differences literally rips the hurricane apart by shearing the top of it off.

Not only do hurricanes not form because of extreme pressure variations, they are prohibited by extreme pressure variations, because such variations produce wind shear.

Obviously, once a hurricane forms, it has strong pressure variations related to the storm itself.  This is not what causes the storm however.  The strong pressure variations are a result of the storm, not the other way around.  The hurricane is produced by the Wind-induced Sensible Heat Exchange--a process that transfers heat from the ocean to the atmosphere.  The process depends upon having a relatively flat pressure field for it to start and proceed.  Again, if extreme pressure variations are present the associated wind shear will disrupt the entire process.

I mentioned early on that a similar process may  occur in some Mars dust storms.  I published a paper about this a couple of years ago (here).  Dust essentially provides the heating on Mars.  As air flows toward the storm, dust is lifted, and that dust is radiatively heated by the sun.  Winds increase in the disturbance and this causes more dust to be lifted.  Thus, I named the process Wind-Enhanced Interaction of Radiation and Dust (WEIRD)  in recognition of the WISHE analog to hurricanes.


Rafael Romo Mulas said...
This comment has been removed by the author.
Rafael Romo Mulas said...

An extremely interesting and even entertaining post, with great teaching power!!

I am an Earth Science bachelor student from Italy, trying to make his final thesis on martian dunes. I would like to compare some simulated dune changes to observed ones, and I think I will need some wind simulations too. Because of this, even if this might not be the best place to ask, I think you might help me understand if the MRAMS is in any way available to researchers/students and how.
Since I can't find info about this matter anywhere, I seriously doubt it is available, but I thought it might have been better to ask anyway.
Thanks in advance for any reply!


Scot Rafkin said...

Hi Rafael,

I would be interested in a potential collaboration and I am willing to provide you with the MRAMS code. Please send me an email: rafkin-@-boulder.-swri.-edu (remove dahses).