Monday, October 31, 2011

Learn about dual-pol radar...if you're in Seattle

Just an announcement today for those of you interested in learning about the new dual-pol radar systems that have no been installed in all Washington State radars and are continuing to be installed in radars across the country.  I will be giving a talk for the Puget Sound chapter of the American Meteorological Society about these new radar systems, how they work, what new products they'll include, and how we might interpret these products in the context of Pacific Northwest weather.

The talk will be at 6:30 PM this Wednesday, Nov. 2, at the Northeast Branch of the Seattle Public Library (6801 35th Ave. N.E.).  It's a free talk and open to the general public.  I invite any and all of you who may be interested to attend.

Tuesday, October 25, 2011

A Later-Season Hurricane

The "official" hurricane season lasts from the beginning of June through the end of November, so we're beginning to near the end of it.  In fact, we're getting to a time of the year when there are fewer hurricanes forming in the Atlantic.  Here's a chart from the National Hurricane Center showing the annual distribution of Atlantic hurricane frequency:

You can see that the number of tropical cyclones typically peaks in mid-September and then trails off.  However, interestingly enough, notice that there does appear to be a secondary maximum (though not nearly as prominent) in mid- to late-October.  Is there some physical reason for this?  I'm not sure.  But, the fact remains that here we are in late October and we have another hurricane to track:  Hurricane Rina.  (Oddly enough, the name "Rina" was selected to replace the previously-retired name "Rita" from this cycle of hurricane names.  I think that sounds a bit too similar, but...it's the way it is...).  Here's a visible satellite image of Rina from early this afternoon:
You can see that Rina is a fairly well-organized storm with a clear eye and that it isn't experiencing much wind shear.  How can you tell that there's not much wind shear going on from this image?  Look at the high-level cirrus clouds coming off of the storm--the wispy, fibrous like clouds on the outer edges of the circulation.  Notice that you can see those clouds pretty much all the way around the storm (maybe not so much on the western side, but they're still mostly surrounding the storm).  If we had strong wind speed shear (that is, winds strongly increasing with height in a direction that's NOT circling the storm), we'd see all of that cirrus "outflow" on one particular side of the storm, since the upper level winds would blow off all the clouds in one direction.  Because there's cirrus pretty much everywhere, we can guess that there's not a lot of vertical wind shear.  That's good for hurricanes, as it allows them to stay organized.

Strength-wise, Rina is a category 2 hurricane, forecast to become a stronger category 3 storm in the next day or so.  Warm sea-surface temperatures in the western Caribbean are helping fuel this storm.  Also, that lack of strong wind shear is allowing the storm to stay together and not get blown apart.

In contrast to the usual paradigm of hurricane forecasting, while we're pretty sure that the hurricane is going to strengthen, we aren't so sure about where the hurricane is going to go.  In a previous blog post, I talked about using model ensembles to forecast tropical cyclone paths. Remember that ensembles involve running multiple models or multiple perturbed versions of the same model to get many different possible forecasts of the future.  We can see how well these forecasts agree to make decisions about where the storm may go.

Well, here's an example of an ensemble forecast of Rina's path from this morning's ECMWF model ensemble.  Each yellow line represents a different ensemble member's forecast of Rina's path:
The black line shows the previous path of the center of the storm and where it is now.  But look at how many different paths the ensemble members are forecasting for this storm!  Some are sending it south into Nicaragua.  Some are bringing it northwest, then abruptly southwest into Belize.  Many are bringing it into the Yucatan Peninsula.  And a few even have it crossing the Yucatan and then curving back northward to hit Florida.  That's a lot of uncertainty in the path of this storm.  The National Hurricane Center's forecast favors the Yucatan path (and you can see that a fair number of the ensemble members above do too).  Here's the NHC's forecast:

The cone of white surrounding the storm indicates their uncertainty in where the hurricane center will be.  The forecasters at the Hurricane CEnter use their expertise and experience with past hurricanes to try to rule out some of the more unlikely forecasts.  As such, you can see here that they really don't believe the models that are taking the hurricane into Nicaragua or Belize.  But there's still a lot of uncertainty in their forecast track.  Furthermore, they haven't ruled out a landfall in Florida yet.  But Rina is a slow-moving storm, and we will have until the end of the week to see what Rina does and where it goes before determining if it will hit the US.

Thursday, October 20, 2011

Wild and wet day Chicago

Chicago (and much of the upper midwest) had quite the rainy and windy weather starting yesterday and into today.  So what's causing all this?

First, let's check the upper-level pattern.  Below is last night's 300mb analysis.  You can see that there is a deep, cut-off trough over the Mississippi River valley.  Strong jet streaks are notable on either side of the trough axis.

Most notably, the exit region of the eastern jet streak--where air is leaving the jet--is almost right over the location of the surface low.  Since the exit regions of jet streaks (particularly cyclonically-curved jet streaks) tend to be associated with areas of divergence aloft, this helps lower the pressure at the surface.  Notice in the surface map below that the surface low pressure center was right over Ohio--right under that exit region.
That's a rather spectacular surface low, there--the lowest pressures are down under 988mb.  Notice also that the isobars--the pressure contours--are very closely packed together, particularly on the northwestern side of the low.  Closely-packed isobars imply a very strong pressure gradient, and strong pressure gradients lead to high winds.  Very high winds.

As of 11AM this morning, here's the highest wind gusts as reported by WFO Chicago:

SITE                      PEAK GUST
(COASTAL STATIONS)          (MPH)

WAUKEGAN HARBOR..............59
CHICAGO CRIB.................62
CALUMET HARBOR...............52
BURNS HARBOR.................55
MICHIGAN CITY................62


So over 60mph wind gusts in some locations.  There was also a 73 mph wind gust reported up north in Racine, Wisconsin.  It looks like sustained winds have been measured in the 20-30 mph range for much of that area.  Wave heights on Lake Michigan are also reaching upwards of 15 feet.  This is helped by the fact that the winds are generally blowing out of the north, meaning that they travel a long ways down the lake before reaching the shore.  This gives waves pushed by the wind a long distance over which they can build up height.  You can see a good video of the wind and waves from the WGN news site here.

Another story has been the rainfall with this storm.  Since the upper-level low has been cut-off from the mean flow, it has been slow to move out of the region.  As such, this storm and its associated rainfall have lingered over the area for over 48 hours.  Here's a map of the precipitation totals from yesterday and today:

Some areas in northern Indiana and Ohio have seen upwards of 3 inches over the past 24 hours.  Also notice in northern Michigan how there's an area of higher precipitation totals on the western shore of Lake Huron. With the low pressure center over Ohio and counter-clockwise winds around the low, you can imagine that in northern Michigan the winds have been out of the east.  Therefore it looks like we're seeing some lake effect enhancement of the rainfall as those easterly winds blow from the lake into northern Michigan.  I'd also suspect that the band of higher precipitation seen in northwestern Indiana is also due to lake effect enhancement as those winds come out of the north down Lake Michigan.

Winds today are still pretty strong, but they've backed to become northwesterly across the Chicago area.  Here's a surface map from the College of DuPage site valid at 18Z today.
You can see that winds speeds are still at 15-20 knots, even inland, but the winds have become more northwesterly.  Also note that the temperatures are really starting to cool off--it's only in the mid 40s everywhere.

This storm is forecast to finally move off to the east in the next 12 hours or so, leaving behind much cooler temperatures.  There are frost and freeze advisories in effect for much of the upper midwest as cold air moves in behind this powerful storm.

Tuesday, October 18, 2011

Mountain Mixing Brings Warmth

Today is forecast to be a very warm day in Seattle--probably our last time to get into the upper 60s this year.  So what's causing this climatologically unusual day of warmth?

It starts with the upper-air pattern.  We're under a ridge at the moment, and with that ridge is a pool of very warm temperatures aloft along the coast.  Here's this morning's 700mb analysis:
Warm temperatures and high heights are very prevalent across western Washington.  All this warm air has a curious and somewhat counter-intuitive effect, though.  With warm air comes a fair amount of rising motion and a decrease in the density of the air.  This works to lower the pressure at the surface underneath this pool of warm air aloft.  The result is what we call a thermal trough--an elongated area of lower pressure at the surface in response to warming.  Here's the sea-level pressure and temperature analysis from this morning:

That map is a bit cluttered, but you can see the area of lower pressure long the coast.  Contrasting with that, there is an area of higher pressure in the interior of Washington on the Columbia Plateau.  See the large number of isobars (the black contours) that are running north-south along the axis of the Cascade Mountains?  That implies a very strong pressure gradient between the high Columbia Plateau and the Puget Sound lowlands.  The pressure is much higher to the east of the Cascades than it is to the west.  In fact, this morning the pressure in Yakima (east of the Cascades) has gotten to be almost 6 mb greater than the pressure in Seattle (west of the Cascades).  That's quite a difference.

The result is that air wants to push from the east to the west across the mountains.  Of course, the mountains make it difficult for air to move through.  As air crosses the Cascades, it's at a higher elevation than down here near sea-level in Seattle.  As a result, the air pushing across the Cascades this morning has stayed above the cold air that pooled in Seattle last night.  We can see this on the wind profiles from this morning in Seattle:
The above image is a series of soundings taken from the Seattle wind/temperature profiler at Sand Point.  Time increases to the left, so the most recent sounding (as of 11AM this morning local time) is on the left side of the image.  The heights are given in meters on the y-axis.  You can se that starting around 9Z this morning (the time marked by 18/09 on the bottom) that we had easterly winds above 500 meters or so with northerly winds below that.  The easterly winds above are from air coming out the higher elevations over the mountains.  The northerlies near the surface are due to the colder air near the surface moving toward that area of low pressure just off the coast.  So we have two air masses moving different directions.  You can even see that there  is a pronounced temperature difference between the two.  The red contours show temperature in degrees Celsius.  Where the winds are out of the east, the temperatures were in the 14-15 degrees Celsius range.  Below that, the temperatures were down around 10 degrees Celsius.  Two different air masses with two different temperatures.

But what's interesting to watch is how the height of the transition zone has been decreasing as this day has begun.  If you watch over time, the level where the winds switch from northerly to easterly has been dropping rather steadily.  At 11 AM it was only measured at around 200 meters above ground.  The winds aloft have really picked up, too--there are easterly winds measured at around 40 knots just 500 meters above the ground while at the surface they are barely registering out of the north.

As the sun warms the land below, it will basically destroy the cold pool of air that has been near the ground all night in the lowlands.  As the air warms, we'll get more movement and mixing.  The constant high winds just above the ground are also helping to promote this mixing.  Eventually, that warmer air mass with its easterly wins should mix all the way down to the surface.  Not only will this bring the slightly warmer air to the surface, but that air will warm even further as it decreases in elevation.  Since the pressure is higher near the surface, descending air compresses and, consequently warms.  So, as soon as this air aloft reaches the surface, we expect a big spike in temperatures.

Some higher elevations have already experienced this warmer air.  Here's the 11 AM surface map for the Puget Sound area:
Note the temperatures (the red numbers) down in the Puget Sound area--mostly low to mid 50s.  But look just inland to temperature in the foothills of the Cascades.  There are some mid-60s being reported there--with easterly winds as well.  Locations like North Bend, Index and Puyallup are high enough in elevation that they're already in the warmer air and easterly winds coming off of the mountains.  Hopefully that air will get down to us soon and cause our temperatures in Seattle to soar a bit too.  I'm looking forward to a warm day...

Wednesday, October 12, 2011

A warming week for the west?

Showers and thunderstorms are currently moving across the central US with another area of rain in the northeast.  These areas of rain are associated with two different surface lows and we're expecting clearing skies in their wake.

However, I want to look at the longer-range forecasts from the global weather models for the rest of this week and into next week.  Let's start by setting the stage.  Here's a map of 500mb heights this morning across the entire northern hemisphere.

We can see two types of waves in this pattern.  One is "longwaves".  These waves have a wavelength of tens of thousands of kilometers.  See the big troughs over the central US, the north-central Pacific and the north-central Atlantic?  Those big waves are the longwaves.  Embedded in these larger scale patterns are smaller wiggles in the 500mb height contours.  These represent "shortwaves" which only have wavelengths on the order of 100s of kilometers.  For example, that little bulge in the 500mb height lines over the central Appalachians is a little shortwave.  That little shortwave is what's helping to support the rain currently over the northeastern United States.

I bring this map to your attention because I want to point out where the lowest 500mb heights are located.  These are areas shaded in white and blue.  You can see that those regions are generally on the opposite side of the planet from the United States--they're generally over northern Asia, Greenland and north of Europe. Since lower 500mb heights are generally associated with colder air, this means that we shouldn't have to worry about any strong cold air masses moving southward into the US, at least for a few days.

Here's the GFS forecast for the 500mb pattern on Friday:
The lowest 500mb heights are still off over Asia and Greenland.  We see that the GFS is developing an unusually deep cut-off low in the middle of the Pacific.  Since this is well offshore, I wouldn't be concerned about this yet.  Another trough is present over the northeastern US.  However, most of the central and western part of the country seems to be under a ridge between the two troughs.  Ridges are generally associated with higher pressure (and often warmer temperatures).  So, this gives some hope for warming weather going into the end of the week.

Now if we get into the really long range forecasts for next week, the pattern looks like it's going to get more amplified.  Here's the forecast 500mb heights for next Tuesday:
Notice that a pocket of lower 500mb heights (and consequently, cooler air) has broken off and is moving southward through northern Canada.  This is leading toward more height falls and general troughing over the northeast.   However, that cut-off low in the Pacific is still forecast to be hanging around (though weakening).  The result--with a cut-off low and longwave trough in the pacific and a deepening trough over the eastern US--implies a rather strong ridge in between.  You can see here that there is forecast to be a very big ridge over the western US.

And it's not only the GFS model--here's the ECMWF forecast of 500mb heights for the same day:

A very dynamic pattern, with a big trough over the eastern US and a soaring ridge over the western US.

So how will this affect the temperature?  I still and somewhat leery about looking at surface temperatures from a 130-160 hour forecast.  But, just as a hint, here's the ECMWF 850mb temperatures for next Tuesday:

The temperatures are represented by the colors in degrees Celsius. Notice how the temperature pattern closely matches the height pattern in the maps above--a trough of cooler temperatures over the eastern US and a ridge of warmer temperatures over the western US.  You can see that even as far north as Seattle, the ECMWF is forecast 850mb temperatures around 20 degrees Celsius--that's 68 degrees Fahrenheit.  Since surface temperatures are usually somewhat warmer than the 850mb temperatures, we may see a warm beginning to next week out here.  That would be a nice break from the cool rainy weather we've had all week...

Saturday, October 8, 2011

Heavy rain where it's needed the most

If you remember, during the late summer there was a lot of talk about the wildfires in central and west Texas.  That area has been stuck in extreme or exceptional drought conditions for much of this year. Here's the official drought monitor image from NOAA as of October 4th:

You can see that pretty much all of Texas is in some pretty dire straits when it comes to needing rainfall.  But it appears that there is some relief in sight.  Take a look at the radar composite tonight over central Texas:

A large area of slowly moving, rather heavy rain has planted itself over much of central Texas.  There have been some strong wind gusts reported with some of the more intense cores, but no real severe weather.  Just lots of heavy rain.  I'll update this on Sunday morning with the 24-hour rainfall totals from across central Texas.

So what's helping to cause this rain?  Another one of these cut-off lows--like the one that was parked over the midwest and mid-Atlantic states a week ago--has set up over the Rockies, as seen on the 300mb analysis from this evening:
What's interesting about this particular setup is that there doesn't seem to be a strong or organized surface low associated with this upper-level low.  A check of the surface map shows that there's an "inverted trough" of low pressure, but really only seems like its there because you have to have a relatively "low" pressure point between the sprawling high over the eastern US (which has brought delightful temperatures to much of the upper midwest) and the high over the northern Rockies today.
I find it somewhat unusual to see the weather pattern more dominated by sprawling high pressure centers as opposed to powerful lows.  But, here it is.  Notice how as we approach winter the pressure pattern up north is really starting to get intense.  A powerful surface low is analyzed over the southeastern tip of Alaska and there's a very tight pressure gradient across northern Ontario and Quebec.  Tight pressure gradients imply strong winds, and you can see the cold temperatures up there as well.  Soon that sort of stuff will be headed our way...

However, with respect to today's setup, steady southerly and southeasterly winds on the western periphery of the east-coast high are helping to bring lots of relatively deep moisture out of the Gulf of Mexico and into the southern Plains.  This 850mb analysis of relative humidity shows a nice plume of moisture from the Gulf up into central Texas.
This moisture is obviously contributing to their heavy rainfall there.

Of course, while this rain is a welcome relief to drought-weary Texans, heavy prolonged rainfall after periods of drought can create flash floods.  If the ground is very, very dry it often can't absorb more water at the same rate that the rainfall is dropping water to the ground.  The result can be a lot of standing water and runoff into streams.  So--even though it's nice to get the rain, Texans should be vigilant for unpredictable flash floods.

Tuesday, October 4, 2011

Pressure rises, pressure falls

One of the simplest variables that we often look at to predict the weather is the atmospheric pressure.  In and of itself, pressure can tell us a lot about what kind of weather to expect.  For instance, I'm sure most of you have seen barometers like this one (taken from the barometer Wikipedia page...):

You've seen that the pressure scale is often accompanied by weather descriptions (and in this case, little pictures).  Rain is associated with lower pressure, "change" or sometimes "partly cloudy" is often observed with average pressure (around 1000mb or 29.5 inches of mercury) and "fair" or "sunny" is associated with higher pressures.  In general this holds true, though each location is different.  However, let me show you a curious phenomenon that occurred the the other day at the weather station at my house.  The following image is a meteogram of weather variables over a 24-hour period.
It says "60-hour" meteogram at the top, but it's really only 24 hours.  This is from Sunday evening into Monday morning. We had rain moving into the Seattle area that evening, and you can see in the bottom panel showing rainfall that we got quite a bit of rain from around 5:00 PM to 7:00 PM local time--almost 0.2 inches.  Now, all day the pressure had been falling pretty steadily--if we depended on the "barometer forecast", we would have expected this rain coming as the pressure fell.

But look more closely at the pressure pattern (the middle, brown-colored panel).  While the pressure did fall throughout the day, something curious happened right when it started raining--the pressure actually increased a bit and stayed slightly higher until the rain had stopped.  Higher pressures associated with rain?

But this makes perfect sense when we think of the air movements that are associated with rain.  To get such large precipitation rates that we were seeing there, there probably were some sort of convective motions going on to help produce extra-heavy rain.  Most convective processes in the atmosphere boil down to two parts--updraft motion helping to feed moisture into the storm and also downward motion accompanying the precipitation.

 Rain falls with cold air exiting the thunderstorm, and this colder air is more dense than the surrounding air.  Furthermore, the raindrops themselves exert a drag on the air as they fall, pulling air down with them.  The net result is a slight increase in downward pressure at the surface underneath areas of heavy rain.  Colder, denser, sinking air combined with downward drag on the air as the rain falls combine to increase the pressure at the surface.  So, this little "blip" of higher pressure actually makes sense.

You can see this effect a little bit in the temperature curve (the red line in the top panel) as well.  As the rain accompanies cooler air down to the surface, you can see that the temperature suddenly begins to decrease rather sharply once the rain begins.  It continues to cool at this steady rate until the rain stops.  At the same time, the dewpoint temperature (the green line in the top panel) also shows a significant rise while it was raining.  Some of that liquid water falling to the surface evaporates into the air (notice that the air temperature was well above the dewpoint, so the air was not saturated).  This extra evaporation due to the presence of so much liquid water dramatically increased the dewpoint temperature over a short period of time.

This was a bit of micro-meteorology here, but it just illustrates why you can't just say, "lowering pressure means rain".  The falling rain itself can cause the pressure to increase, too...