Tuesday, November 22, 2011

The Pineapple Express

After a brief reprieve here in Seattle (and an inch or so of rain last night), another shortwave trough is moving up through the area and rain is on its way in again.  So the rain will continue.  But, it's going to be a rainy week or so ahead as well.  Why is this?  A particular weather event known as a Pineapple Express (not the movie...the real thing).  Let's look at what makes a Pineapple Express event.

In simplest terms, a Pineapple Express is a heavy rain event here on the west coast due to a plume of moisture moving northeastward out of the tropics.  Such moisture plumes are often referred to as atmospheric rivers.  In stylized form, you'll often see these events depicted as giant arrows of moisture running into the coast:
But we can do better than that.  Here's a 105-hour forecast of total column water vapor (all the moisture in a column above your head) for next Tuesday:
The reds, whites and blues indicate lots of water vapor in the atmosphere over those locations.  You can see that the water vapor is concentrated in a relatively narrow stream stretching from the tropical Pacific up north into southern British Columbia and northern Washington.  This is the hallmark of one of these Pineapple Express events--a "river" of moisture brought up from the tropics to the west coast.  Strong winds usually accompany this plume of moisture, and since it has tropical origins (often near Hawaii), we get the name--Pineapple Express.  Moisture on an express route from the land of pineapples.

Of course, since this air has a tropical origin, it tends to be much warmer than usual.  Here's a forecast of 850mb temperatures for that same time next Tuesday:
 You can see that a plume of warmer air accompanies this moisture.  This means that most of the precipitation will fall as rain--the freezing level will be at several thousand feet.  Only the higher mountains will see snow.  You can also see on this map the forecast from strong, southwesterly winds in the plume, helping advect that warmth and moisture northwards.

Pineapple Express events are often set up by a deep trough aloft over the central Pacific.  This helps orient the pressure field aloft to promote southwesterly flow out of the tropics and into the west coast of the United States.  There is even some evidence out there that suggests this sort of pattern may in turn be linked to a phenomenon in the tropical Pacific and Indian Ocean known as the Madden-Julien Oscillation, or MJO.  However, that's a topic for another post.  In the meantime, there's a lot of water to look forward to in Seattle.

Monday, November 21, 2011

Torrential rains for Seattle

The rainy season is here in Seattle.  The forecast for the next few days says it all...
A series of upper-level disturbances moving over the Seattle area combined with a favorable flow pattern aloft are helping to bring in copious amounts of moisture to the Pacific Northwest over the next few days.  Here's a capture of the satellite water vapor image from late this afternoon.  You can see a plume of moisture stretching from the Pacific Northwest far out into the Pacific.  All that moisture is headed ashore in the next few days.

How much rain are we talking about here?  A lot.  Here's a forecast map from our 4-km WRF model of precipitation totals over the next 72 hours over western Washington.  We're maxing out the scale of the color bar on this map here...
When you do the division correctly with that color bar, that works out to between 4-10 inches of liquid (!) precipitation over most of western Washington over a 72 hour period.  Will this verify?  Who knows.  It seems a bit excessive to me.  But I have done several independent WRF runs looking just at Seattle, and I consistently am getting total accumulations of around 5 inches over the next several days.  It's going to be very, very wet...

Tuesday, November 15, 2011

Flow with no low

A quick look at the infrared satellite image over the eastern US today shows an elongated area of high clouds and deep convection stretching from Texas up through the eastern Great Lakes.
Looking at the radar composite also confirms that there's a lot of convection and rain going on associated with these clouds.  The Storm Prediction Center even has a slight risk for severe weather out for the southern Mississippi valley.
However, this pattern doesn't look much like what we typically see in mid-latitude cyclones.  Usually we have a pretty well-defined low pressure center with fronts.  Above this low pressure center there's often a jet streak well-positioned to give divergence aloft and the lift needed to generate lots of clouds and precipitation.  But where is the surface low on the above images?  There's no characteristic swirl of the cloud pattern or precipitation (except for the little low over eastern North Dakota).  Looking at a surface map of temperature and pressure also doesn't reveal any strong low-pressure centers.

What we do see, however, is what meteorologists call a baroclinic zone--a zone where there are relatively strong horizontal temperature gradients.  Notice that along the same corridor where we're seeing clouds and rain, there's a pretty well-defined temperature gradient with 60s to the south (he green colors) and 40s to the north (the light blue colors).  This looks kind of like a front, and you might even analyze it as a stationary front.  Baroclinic zones like this are often associated with convergence in the wind field, and you can see that there are signs of some convergence in the surface winds, particularly over the southern plains.  It turns out that weak convergence in the lower atmosphere along this baroclinic zone is providing just enough lift to support all of this rain and cloudiness.

One other feature of baroclinic zones, however, is that they provide the "fuel" for surface low pressure centers to feed on and deepen.  There's a lot of potential energy that can be tapped in a region where the temperature is changing rapidly.  We also know that baroclinic zones can strengthen the winds aloft (through the thermal wind relation).  In fact, a check of the 300mb winds aloft shows that we do indeed have a strong jet streak roughly paralleling the baroclinic zone.

Since this baroclinic zone probably slopes northward as you go up in the atmosphere (for reasons I'll explain in another blog) it makes sense that we see the jet streak to the north of the strongest temperature gradients at the surface.

But let's look at this upper-air setup for a moment.  Notice that there does look to be a shortwave trough of sorts across the central plains--troughs often bring in vorticity that can aid in surface low development.  We also might try applying that four-quadrant jet streak model I've talked about in previous blog posts to look for regions of convergence and divergence associated with this jet streak.  From that model, we know that divergence aloft often occurs in the right entrance region and the left exit region of a straight jet streak.  If we imagine air entering the jet streak over the Great Lakes in the map above, it's moving from southwest to northeast.  This would place the right entrance region somewhere over northern Texas, Oklahoma, Arkansas or Missouri.  It would place the left exit region somewhere over Ontario and Quebec.

Now, divergence aloft leads to upward motion and a lowering of the pressure at the surface.  Here's a diagram from another blog post I did about this particular fact:

So, if we see an area of potential divergence aloft (particularly over a low-level baroclinic zone), we might look for a surface low to form there.  The surface plot I showed above was at 16Z (about 10 AM CST).  Here's a plot from 20Z--just four hours later:
It now looks like we have a surface low center developing in north Texas and another area of low pressure developing in Ontario.  Pretty amazing.  We could have probably forecasted something like this to happen because of the three ingredients I have been talking about:
  1. The presence of a rather well-defined baroclinic zone in the lower atmosphere.
  2. A shortwave trough axis moving in aloft.
  3. The divergent region of a jet streak aloft moving overhead, causing the pressure at the surface to fall.
So now we are finally getting our better-defined surface lows.  The question is which low will actually become dominant and start directing the flow.  Low pressure centers need to feed on the energy in baroclinic zones to grow and deepen.   Notice on the western side of  the Ontario low there is a very strong pressure gradient, implying very strong winds out of the northwest.  Also, the temperatures accompanying those winds are bitterly cold--those pinks and purple represent temperatures in the teens.   In contrast, there don't look to be very strong temperature gradients around the Texas low--in fact it seems kind of orphaned out there away from the baroclinic zone.  You can also see that the primary baroclinic zone (the region where temperature is changing rapidly) looks to have moved further north--it's up over northern Illinois through northern Kansas now. 

These factors point to the northern low becoming the dominant low pressure center here.  With very cold air racing down the back-side of the low, this surge of air will quickly overtake the existing, somewhat stationary baroclinic zone and turn it into a full-fledged cold front.  And, indeed, that's what the models are forecasting.  Here's the NAM model forecast for tomorrow at 15Z:

The low center in eastern Ontario still isn't very well defined, but flow around it has brought in much colder air, sharpening the baroclinic zone again and turning it into a cold front.  However, this baroclinic zone/cold front is still in the same geographical area as it was today, so I wouldn't expect any break from the rain any time soon...

Wednesday, November 9, 2011

Five crazy weather events you could be looking at right now

1) The extremely deep low in the Bering Sea
An unusually deep low pressure center moved through the Bering Sea last night and is now moving northward through the Chukchi Sea in western Alaska.  At its peak intensity, the low was down to 943 mb (or so we assume).  It was still at 946 mb, at least according to this morning's 12Z surface analysis from the NAM-WRF model over Alaska:
Hurricane-force winds are being reported with this storm, which is unusually far north to be of such a deep magnitude.  Here's the Alaska surface observations from 06Z early this morning.  Look at the wind observations in the Bering Sea.  Anywhere you see a triangle-shaped flag on the wind barb, that means winds over 50 knots were reported.  Amazing.
NOAA runs an ocean wave modeling system called WAVEWATCH III that forecasts wave heights over areas in time.  Here's their 12Z analysis of wave heights across the Alaska region.
Within that orange area in the center of the Bering Sea, they're showing waves as high as 9-10 meters--that's 30 feet or more.  In fact, some reports have come in from buoys and ships reporting 40-foot waves with a wave period of only 14 seconds.  That means that if you were on a boat, every 14 seconds you would fall 40 feet, then rise 40 feet on the waves.  Spectacular.  You can also see on this analysis map the grayed-out area to the north.  This represents areas that are already completely iced over for the winter.  That area is still pretty far north, so villages on the western Alaska Coast that would normally be protected from these waves by sea ice don't have that protection yet.  Major coastal flooding and storm surge is expected.

2) The anomalously deep cut-off trough forecast over the eastern Pacific
Anyone who was planning on heading down to southern California this weekend hoping to escape the winter weather may be in for a bit of a shock.  The same highly-amplified upper-air pattern that helped get that Alaska storm going is going to help spin off a deep, cut-off, upper-level low that looks like it will hover off the coast of California for the next few days.  Here's the 36-hour forecast for 500mb heights and temperature on Thursday evening from the UW-WRF 36km model:
You can see that this low is paired with a very strong ridge of higher heights over the central Pacific and that the main jet stream (where the iso-height lines are packed closely together) is much further north.  This means that the low center is going to hang around for a while.  Notice that with counterclockwise flow around that low height center it's directing onshore flow right into California in addition to bringing in colder air aloft.  This combination is going to help bring cloudiness and rain showers to that area over the next few days.

We can compare this trough and ridge to the climatological normal values for 500mb heights to see just how unusual of a pattern this is. Here's the CPC's 3-day forecast for 500mb height anomalies over this area:
Anywhere surrounded by dashed red lines is anomalously high and anywhere surrounded by dashed blue lines is anomalously low.  You can see that, particularly with respect to that ridge in the central Pacific, this pattern in somewhat unusual.  It will definitely make for some interesting weather.

3) First snowfall across parts of the plains and the upper midwest
With the jet stream shifted so far north due to this huge ridge over the northern Pacific, it's getting to be kind of hard to make other shortwaves move along and propagate away.   The same surface low that helped create severe thunderstorms in Oklahoma the other day has been slowly moving northeastward and getting better organized.  It's currently located over western Michigan:

Note the very tight isobar packing on the western side of the low.  This implies strong northwesterly winds in that area, helping to advect in some very cold air.  Behind this low, temperatures look cold enough to support snow.  Here's the RUC analysis of 1000-500mb thickness (an indication of how cold the lowest part of the atmosphere happens to be) for 18Z today.  Anything north of the solid blue line should be cold enough to support snow:
As the low moves eastward and the cold front continues to swing south, areas in eastern Illinois and Wisconsin and into Michigan should see temperatures start to fall over the next few hours, eventually to a point where any precipitation that falls could be snow.  Combined with the high winds behind the low pressure center, we can see why winter weather advisories have been posted for parts of Wisconsin.

4) Sub-tropical and now Tropical Storm Sean

Take a look again at the surface analysis that I showed two images ago.  Notice the rather deep low analyzed off the coast of the southeastern US?  That's tropical storm Sean (formerly sub-tropical storm Sean), a late-season tropical storm that should stay well-away from land.  On visible satellite, we can see that it has a good spiral shape, but it's not very symmetrical--there are bands within the storm with little cloud content and no well-defined eye.  These are all evidence that Sean isn't particularly strong--but still a tropical storm.

The National Hurricane Center official forecast has Sean staying a tropical storm for a while, feeding off the warm waters in the Gulf Stream.  There is some evidence in the dynamical models and satellite imagery to suggest that Sean may strengthen to a weak category one hurricane, though, so I wouldn't be surprised to see it upgraded sometime soon. However, they don't project it to hit mainland North America at all.  It may clip Bermuda, but otherwise it's staying well out to sea.

5) The "tropical storm" that formed in the Mediterranean Sea

If you really want an unusual cyclone, for the last several days a warm-core low developed and lingered over the western Mediterranean Sea.  The CIMSS satellite blog had several loops of satellite images as the cyclone reached its peak intensity.  Based on the warm-core structure and that the peak winds reached tropical storm strength, this was considered to be a tropical cyclone.  In the Mediterranean.  Strange...

As you can see, there's been a lot going on in the weather recently.  Plenty to keep any meteorologist, from amateur to professional, occupied for quite some time...

Monday, November 7, 2011

Late-season severe weather in the southern plains

It's been a busy two weeks for me, so I apologize for not doing more blog posts.  Hopefully this week I'll be able to get back into the regular swing of things.

Today we have an interesting weather event setting up for the southern plains.  In an upper-air scenario that looks more like what we see in springtime, we have an upper-level trough across the desert southwest this morning.

The jet streak and wind fields associated with this upper-level trough are providing enough upper-level divergence to support a growing area of low-pressure, currently analyzed in west Texas.

See the trough of lower surface pressures extending northeastward through western Oklahoma out from the center of the low?  That trough marks the location of a relatively stationary front.  You can also see a bit of convergence in the surface winds along that boundary.  Notice that throughout much of eastern Texas and Oklahoma the winds are out of the south or southeast--straight out of the Gulf of Mexico.  This is bringing in some unusually humid air to the region.  Here's the Oklahoma Mesonet's dewpoint map for this morning:


You can see dewpoints in the mid-60s across much of the eastern part of the state.  However, once you cross that stationary frontal boundary across northwestern Oklahoma dewpoints drop into the mid-40s.  A strong moisture gradient is another sign of a front.

One other thing we can check is the temperature structure aloft to see if the atmosphere is unstable enough to support convection. Here's this morning's 12Z sounding from Norman, OK:

You can see that we are fairly saturated at low levels (we should be with dewpoints in the 60s at this time of year...) so we look to see if the lapse rates (the rate temperature is changing with height) are greater than the saturated adiabatic lapse rate to find instability.  On average, the saturated (or moist) adiabatic lapse rate is around 6.4 degrees Celsius per kilometer.  In the lower-left corner of the skew-T diagram above, they list the lapse rates over various layers of the atmosphere.  You can see that (except for the layer right at the surface) all the lapse rates are greater than that 6.4 degrees Celsius per kilometer benchmark.  This tells us that the atmosphere is indeed unstable with respect to saturated air.  This is good for convection.

So it appears we have all the ingredients--upper-level support, low-level convergence near a surface boundary, a conditionally unstable atmosphere, and lots of moisture near the surface.  November just seems a little late in the year to be expecting severe weather.  However, take a look at this climatology of tornado events in the Oklahoma City area from the National Severe Storms Laboratory:
The x-axis is given as the day of the year (out of 365).  Today (Nov. 7) is day 311, so that's where we are on the graph.  You can see the clear maximum during the spring time, but there is a secondary, smaller maximum in the autumn.  The autumn is once again a period when the orientation of the upper-level jets and the cold polar air are fluctuating rather wildly, so it makes sense that we see another period of slightly more active weather now.  As such, seeing tornadoes in early November is by no means historically unprecedented.  It's just not what we usually think about during this time of year.

With respect to the development of severe weather today, the Storm Prediction Center has already put out a tornado watch for north Texas and southwestern Oklahoma, citing strong low-level wind shear for the tornadic potential.  Already storms are forming around that stationary frontal boundary, as we can see in this visible satellite image:

What's interesting is that there's a lot of cloud cover filling in the entire "warm sector" of this storm--the entire region where there is warm moist air advecting in from the Gulf of Mexico.  The extensive cloud cover over eastern Oklahoma and eastern Texas is going to prevent those areas from seeing a lot of sunlight today and, consequently, they won't warm up as much at the surface. This works against promoting strong instability in those regions.

But notice that in western Oklahoma, the cloud deck isn't completely overcast.  There are lots of small cumulus clouds but no big mass of clouds blocking out the sun.  This is allowing western Oklahoma to heat up far more than eastern Oklahoma, making it more unstable there.  We can see these temperature differences between overcast eastern Oklahoma and partly sunny western Oklahoma in the Mesonet temperature observations:

Western Oklahoma has warmed into the mid-70s while eastern Oklahoma is still mostly in the mid-60s.  Those clouds make a big difference.  You can also see that it's much colder behind that front in northwestern Oklahoma.  Anyhow, this is one reason why the Storm Prediction Center is focusing their severe weather threat on western Oklahoma--they're going to warm up more, making them more unstable, and they're closer to the convergence associated with the frontal boundary.  All in all, it will be an interesting day to watch things develop.  Particularly if you have access to the dual-pol radar data coming out of the KVNX radar in northern Oklahoma...