Stumbling about in the snow

Let's talk a bit about the forecasting of this blizzard event in the northeast yesterday and today.  A lot of people are writing about what went into this forecast and why New York and Philadelphia did not get nearly as much snow as was forecast (thought Boston and points east did!).  The common theme?  Forecasters didn't adequately communicate the uncertainty regarding this forecast.

Just how uncertain was the forecast?  Here's an example of the predicted snowfall accumulation at JFK airport in New York City from several of our model systems---the NAM (the reds and greys), the GFS (the blues) and the Rapid Refresh (green) starting at 12Z yesterday (yesterday morning).

The numbers drop over time after reaching their peak because this forecast factors in melting and compaction of the snow which will decrease its depth over time.  You'll notice that all of the higher forecasts (20-34") are all red, orange or grey---these estimates all were based on forecasts from the NAM model.  On the other hand, the lower forecasts (6-12") are all based on the GFS (or RAP) forecasts.  In total, this is a lot of uncertainty---as few as 6" or as much as 34".  6" is a normal, everyday winter snow.  36" will keep a city shut down for days.  And this is the morning before the event, when leaders and other interested parties have to make decisions about committing resources, closing things down, etc.  What is a forecaster to do?

It turns out that the GFS had the low pressure center moving further east.  Here's the 30 hour GFS surface pressure forecast (the black contours) from the 12Z run, so this this a forecast for where the low would be around the middle of the day today:

Note the center of low pressure is forecast to be southeast of Martha's Vineyard, with a longitude east of Cape Cod.  This GFS forecast kept the heavier snow bands east over Long Island, but not over New York City, hence the lower snow totals that we saw above.  Here's the exact same forecast, but from the ECMWF model (thanks to Cliff Mass for grabbing these):

Here the low pressure center is further to the west---centered straight south of Cape Cod at this time.  The ECMWF had snow bands further west as well---over New York City.  The NAM happened to agree more with the ECMWF as to the position of the low.

From reading the National Weather Service forecast discussions, this seemed to be what led to the decision to go with the higher snow totals.  This is from the 630 AM EST forecast discussion yesterday at the New York Weather Service Office:

"THE 00Z ECMWF REMAINED VERY CONSISTENT WITH ITS
PREVIOUS RUNS. IT DID LOWER QPF SLIGHTLY...BUT THE OVERALL EVOLUTION
OF THE SYSTEM REMAINED SIMILAR TO ITS 12Z RUN. DID NOT WANT TO MAKE
DRASTIC CHANGES TO THE EXPECTED SNOW AMOUNTS AND HEADLINES WITH
JUST ONE MODEL CYCLE. IN FACT...THE LATEST 06Z NAM HAS COME INTO
CLOSER AGREEMENT WITH THE 00Z ECMWF."

Earlier in the discussion, though, they suggest that the spread in the models, particularly at such short forecast lead times, was more than they would like.  And again in the 1:34 PM EST discussion after the 12Z models shown above had come in:

"THE STORM APPEARS ON TRACK. THE 12Z GFS IS AN OUTLIER FROM THE 12Z
NAM AND 12Z ECMWF. HAVE UPPED SNOWFALL TOTALS ACCORDINGLY WITH
MOST AREAS IN THE 20-30 RANGE FOR THIS EVENT."

So that was the logic.  And, in the past, this logic has served us well.  The European model (the ECMWF) has long been considered the "gold standard" of our weather prediction, usually surpassing the US GFS model in terms of forecast skill.  Indeed, it was the ECMWF's superior performance to the GFS in predicting the path of Hurricane Sandy that recently led to a large investment from Congress to improve the GFS.

What's more, even by 00Z yesterday evening, it was hard to distinguish which model is doing better.  The 12-hour forecasts from the GFS, ECMWF and NAM had the low in virtually the same place at that time, so it was hard to determine which way it would go.  It's also challenging when the low is off the coast, as it was here, to get an accurate fix on what's going on...there are limited surface observations over the ocean, so we have to rely on satellite data which can be tricky to interpret.

So what ended up happening?  The low actually moved to the east, as the GFS had been predicting.  Here's a map put together by @anthonywx showing the analyzed position of the low today at 15Z (the black L) and where the ECMWF forecast it would be 3 hours later (the red L, same as the position shown in the maps above):

So the GFS solution ended up playing out---the low was further east as was a lot of the heavy snow.  This didn't stop Boston and Massachusetts from getting heavy snow, but it definitely pulled things away from New York and Philadelphia.

Note the difference in the position of the low---120 miles.  At a quick glance, as far as a 30-hour model forecast goes, that really isn't too bad.  It just so happened that there was a strong gradient in snowfall on the western edge of this storm, at that gradient happened to pass through the largest city in the country.

There was a lot of uncertainty, but forecasters used knowledge that has served them well in the past to refine their forecast in support of decision makers.  Personally, I think that this is a good case of "better safe than sorry".  Would giving the public better estimates of the actual forecast uncertainty (which was large in this case) have helped people make better decisions?  Should we have put up graphics that said the snowfall could be anywhere from 6"-36" in New York City?  What was the actual forecast uncertainty? Again, I think that it was right of the forecasters to use their experience to try and refine that number, even if it may not have been the best refinement here.  Could an automated system have done better?  Much will be debated in the days to come.

And, for what it's worth, the GFS brings another Nor'easter through the area next Monday.  Who (or what) will we trust then?

High risk of severe storms today for northern Illinois

******UPDATE AT 2 PM CDT******
The Storm Prediction Center has issued a "Particularly Dangerous Situation" Tornado Watch for eastern Iowa, southwestern Wisconsin and northwestern Illinois this afternoon.

They expect "explosive" storm growth over the next few hours with a moderate potential for strong tornadoes and severe winds.  This watch is probably somewhat in response to the 18Z Davenport sounding:

That backing of the low-level wind field to more southerly flow has indeed happened, which is ramping up the low-level wind shear.  We also see that the cap has rapidly eroded with steepening lapse rates in the low-levels.  Sounding analogs (which look at previous soundings from severe weather events for similarities) show an 80% match with previous tornadic events.  Definitely be on the lookout this afternoon

----------------------------------------------------------------------------------

It looks like it could be a rough day for severe weather in northern Illinois.  The Storm Prediction Center has issued a relatively rare "high risk" for northern Illinois and Indiana, including the Rockford and Chicago areas.

There has actually been a fair bit of uncertainty regarding the development and evolution of thunderstorms in this area over the past 24 hours.  Yesterday several models were pointing to a powerful squall line (or quasi-linear convective system if you prefer that term) moving across northern Illinois early this morning.  Take the NCAR EnKF-initialized 3km WRF forecast from yesterday  morning--here's their forecast reflectivity over the upper midwest for 15Z this morning:

 This suggested an intense derecho-type storm (note the bow-echo appearance in this simulated radar composite) that was to move across southern Wisconsin and northern Illinois this morning bringing strong winds and heavy rain.  However, here's what the radar actually looked like at 15Z this morning:

No big thunderstorm complex---only some areas of rain.  There were some thunderstorms earlier, but they were not nearly as strong as many of the models had suggested.  We could tell even last night that this big derecho-type storm was not going to materialize as the precursor storms did not get their act together further west.

However, that brings us to today's forecast. There's a very big consequence of the lack of storms this morning--the entire purpose of thunderstorms is to re-stabalize the atmosphere by convectively removing pre-existing instability.  Had we had thunderstorms move through this morning, they probably would have done just that--cooled and dried out the lower levels of the atmosphere a bit and leave us less unstable today.  But that didn't happen.  Now we get another day of heating an already conditionally unstable airmass and this will destabilize the air even more...setting up today's potential for severe weather.

Let's look at the setup this morning.  A low pressure center has been getting its act together in the central Plains all night.  Here's the surface mesoanalysis from 12Z this morning:

You can see the cluster of storms that was there earlier, though, as I said, not nearly as strong as some models were indicating.  If we look at the Davenport sounding from this morning we see reasons for concern:

You can click on that image to make it larger.  One of the striking features of this sounding is actually the wind profile--almost uniform wind direction all the way through the troposphere.  This enhances the possibility of these derecho/straight-line wind events as storms will quickly tend to organize in a line perpendicular to a uniform wind profile. The surface dewpoint this morning is in the low 60s in Davenport and predicted to rise.  We also see relatively high dewpoints all the way up to around 850mb and near saturation above 500mb (that may have something to do with outflow aloft from the storms to the north.  All in all this sounding shows a lot of moisture in the air--meaning heavy rain is definitely a possibility.

We also have a fair bit of instability.  The sounding profile between about 800mb and 500mb shows a pattern indicative of what is called an "elevated mixed layer" (EML).  This layer of the atmosphere is marked by extremely steep lapse rates (temperatures falling off rapidly with height--basically dry adiabatic) and well-mixed moisture content.  This actually leads to this layer being relatively dry--you can see that the dewpoint is well below the temperature for much of this layer. The presence of an EML enhances the potential for very strong downdrafts--all that moisture being lofted in the thunderstorm can evaporate into that dry layer, causing a lot of cooling within that layer.  This cooled air becomes negatively buoyant and sinks down to the ground, creating powerful downdrafts and organizing a "cold pool" beneath the storms.  This kind of setup again favors organized, bow-echo sorts of storms.

Another concern is the potential for tornadoes.  The wind profile in that sounding doesn't show a lot of low-level directional shear, but this will change as the surface low moves closer.  Here's the surface analysis from 17Z this morning:

As the low has moved closer, notice that the surface winds have become more southerly across eastern Iowa and northern Illinois.  This is creating more directional wind shear in the low-level wind profile.  The low pressure center is also forecast to deepen today as we see a shortwave trough approaching from the west in the 500mb analysis, helping to give the low an added boost.

The convection-allowing models we have are a bit messy with the convection they are developing today.  Here's the Composite Reflectivity forecast from the HRRR model for 23 Z this evening:

 Lots of storms popping up all over the place.  It's difficult to tell if there's any forecast organization.  I ran WRF simulation last night with a few different parameters and had this forecast for 23Z tonight:

Similar picture with a lot of discrete storms across northern Illinois. It looks like the models are making any capping inversion relatively easy to break which may explain why we seem to have storms popping up everywhere.  My WRF run has the storms starting to organize into bowing line segments east of Chicago around 3Z tonight:

All in all it looks like we're in for a crazy day weather-wise across northern Illinois and Indiana.  We'll have to be watching to see how this develops.

Ensembles in hurricane track forecasts

Let's look at the tropics today.  As most people are aware by now, Hurricane Katia has formed in the eastern tropical Atlantic and is chugging along westward, forecast to become a major hurricane.  Katia is a Cape Verde type hurricane, which means that it formed just off the coast of Africa (near the Cape Verde Islands).  Because of this, it will go through a long, slow trek across the tropical Atlantic before approaching the Caribbean or the US.

Where is Katia headed?  We can use the NCEP's experimental GFS ensemble to get an idea of possible tracks.  Remember than a model ensemble involves running the same model dozens of times with slightly different initial conditions each time.  We then get several possible outcomes.  If all the outcomes are similar, we have pretty good confidence in the forecast.  However, if they are vastly different, there's low confidence in the forecast.  Here's several tracks from the 20-member GFS ensemble:

10-day track forecasts for Katia from GFS 20-member ensemble.  http://ruc.noaa.gov/tracks

The current position of the storm is about where the dot that says "OBS" is located--it's a little to the left of the name "Katia".  From that point, every white dot further westward represents the mean location of the storm over all the ensemble members every 48 hours.  So, each subsequent dot to the west is 2-days later.  We can see that the tracks all agree fairly well out to about 6 days, so we have high confidence in our track forecast out to that time.  We start to see some spread after that, but only one of the ensemble members actually brings Katia into the US coast in Maine.  It seems far more likely that Katia will stay out to sea, perhaps impacting Bermuda, but staying away from the US coast.That's still several days away, however.

Another way of graphically looking at ensemble track forecasts is to look at the spread of the ensemble members about the mean.  For instance, say that at 48 hours out you calculated the mean position of the center of the storm from all of your ensemble members.  You could then go to each individual ensemble member model and calculate how far the storm in that particular model was from the mean position.  You can then find the mean error (a rough standard deviation) of the ensemble members from the mean location.

What does this tell us?  If the ensemble members are all doing different things, then the average difference between any one ensemble member and the mean position will be rather large.  But, if all the ensemble members have the storm in about the same place, then the average difference between any one ensemble member and the mean position will be rather small.

We can plot this on maps.  First, we can plot the mean position of the storm out of all the ensemble members at various forecast times.  Then, around each of those points, we can draw a circle whose radius represents the average distance between the position of the storm in each ensemble member and the mean position of the storm.  As uncertainty in the position grows, the circles will get larger.  As circles start overlapping each other, you get the charactaristic "cone" shape we so often see in hurricane forecasts.  Here's an example for the first 168 hours of the Katia track:

GFS Ensemble 168 hour track forecasts and uncertainties.  http://www.esrl.noaa.gov/psd/forecasts/gfsenkf/.

In the figure above, each white circle with a number in it represents the mean position of the storm in the GFS ensemble. The number tells you what hour of the forecast it is, starting at 6Z this morning.  As time goes on, the colors transition up the rainbow from purple to blue to green to yellow to red.  The colored areas represent the circles drawn around each mean position to show the uncertainty.

You can see that for the first 24 hours of the track forecast, the colored swath formed by those overlapping circles is very small--it's hardly visible.  Remember that small circles means that there is low uncertainty--the ensemble has a good idea that this is going to be the track of the storm.

However, after that, the circles start widening considerably.  By the time we get to the end at 168 hours, the uncertainty has grown quite a bit.  In theory, the ensemble suggests that by 168 hours (that's seven days from now), the center of the storm could be anywhere in the big red circular area at the end of the track.  Even with that large area of uncertainty, you can still clearly see a track emerging in the colored swath.  The ensemble is pretty confident that the storm will continue moving northwestward over the next week or so.

Let's look at another example.  There's currently a disturbance (identified as "93L" by the folks at the Hurricane Center) in the central Gulf of Mexico.

GOES visible image of invest 93L at 1715Z, Sept. 1, 2011.

It may not look like much now, but most of our models are developing that junky-looking area of clouds and storms into a tropical depression and tropical storm in the next few days.  Since this storm is forming in the Gulf, it probably poses a much greater threat to the US than it's looking like Katia will.  What do ensemble track forecasts say about this one?

GFS Ensemble 168 hour track forecasts and uncertainties.  http://www.esrl.noaa.gov/psd/forecasts/gfsenkf/.

This plot looks like a big mess.  Note that there doesn't seem to be a clear "track" developing like we saw in the Katia case.  The average position of the storm over the next week jumps all over the place in the north-central Gulf.  By seven days out, the average position really doesn't seem to have changed much, and the red "circle of uncertainty" covers pretty much the entire Gulf Coast (except for west Florida).

So what can we take from this graphic?  It gives us a couple of possibilities to watch out for:

1. There's clearly a lot of uncertainty in the forecast track, if the models suggest that the storm could be anywhere in the northern Gulf over the next several days.  This makes forecasting the track very difficult, and as such it's possible that a lot of people on the Gulf Coast might be put under tropical storm or hurricane watches and yet never really see significant impacts of the storm.  This could be a very difficult storm to predict, and everyone will have to be cautious.
2. Since, after seven days, the center of the storm really hasn't moved much (and the storm may never even have made landfall in that time, despite being very close to land), this storm may be a very slow mover and stick around for a while.  What does that mean?  Prolonged periods of heavy rain and windy conditions on the Gulf Coast.  The storm winds don't have to be too powerful--it doesn't even have to be a hurricane--for the storm to cause lots of flooding damage if it sticks around for a while and keeps bringing wave after wave of rain to the coast.

As of early this afternoon, the NHC still has this as an invest area in the Gulf--it's not a depression or a tropical storm yet.  However, a hurricane hunter aircraft currently investigating the area has found winds both measured and estimated to be at tropical storm strength.  So, should a closed circulation center be found, this could become tropical storm Lee in the next 24 hours.  We'll have to wait and see.

Digging into a "simple" weather model

I often use complex, 4-dimensional weather models and show their output here as an indication of what our weather forecasts look like.  But today I wanted to show how a much simpler weather model works--and how sensitive these models can be to their parameters.

I wrote up what would be considered a small surface parameterization scheme to simply forecast the air temperature at the surface at a point over the next 24 hours.  This is based on equations from the book Parameterization Schemes: Keys to Understanding Numerical Weather Prediction Models by David Stensrud.  It's not a book for the novice meteorologist--you definitely have to have gone through at least some undergraduate meteorology to follow what's going on there.  But the first chapter basically discusses what I did.

So what is a "parameterization scheme"?  It's a way of trying to predict something that we can't explicitly measure or figure out.  When it comes down to it, a lot of features of weather models are parameterized.  We can't track every single raindrop or every single little breeze of air being tossed and turned through trees and buildings.  Yet, these processes collectively are extremely important and feed back into the larger-scale weather processes.  So, we try to estimate their net effect using parameterization schemes.

I wrote up a very basic one that looks at several radiative processes and tries to predict the surface temperature over the next 24 hours.  I started it at around 1:00 PM local time here in Seattle using my "optimal" initial conditions.  Here are the parameters:

• Latitude: 47.6928 N, Longitude 122.3038 W
• Day of year: 229
• Initial surface air temperature: 72 Fahrenheit
• Precipitable water: 1.27 cm
• Ground reservoir temperature: 12 Celsius (54 Fahrenheit)
• Cloud fraction: 0.0
• Bowen ratio: 0.9

 I'll get into what some of these mean in a minute. Here's the output:

I initialized the model at 72 degrees Fahrenheit.  You can see that the model expects us to warm up a little bit this afternoon, then cool off as the day progresses, bottoming out to around 46 degrees as a low overnight.  Then, the sun comes up at about 17 hours in and you can see the temperature starts rising again.

Is that low temperature tonight reasonable?  Here's a 24-hour meteogram showing the past 24 hours at this location (my weather station at my house, actually...):

Temperature is shown by the red curve in the top panel.  You can see that last night the low temperature reached was 48.2 Fahrenheit--not too far from our 46 Fahrenheit forecast.  So at least the model seems to be reasonable.

Now lets look at these different parameters that we can adjust.  I'm not going to change the latitude, longitude, day of year or initial surface air temperature, because those are all things we directly know and observe.  So, let's look at precipitable water.  This is something that is measurable (if you have a weather balloon, which sadly I do not...), and I recently wrote a blog post describing it.  It describes the total amount of water vapor in the air overhead.  So what happens if we adjust this a bit--say, increasing it to 2.0 cm or decreasing it to 0.5 cm?  Here's what happens:

In the graph above, the "optimal" run from before is shown in black.  The temperature results when the precipitable water is increased to 2.0 cm is in red and the temperature results when the precipitable water is decreased to 0.5 cm is in blue.  You can see that there is a pretty strong effect on the temperature path.  The temperature stays warmer when there is more water in the atmosphere, mostly because water has a higher heat capacity than air.  Therefore, moist air retains more heat, keeping it warmer longer.  Notice that with the lower precipitable water values, the temperature immediately began decreasing after the start of the model--the air was so dry that it immediately started cooling off.  However, with more precipitable water, the greater heat capacity of the water vapor keeps the air from losing heat as quickly.  This still allows the temperature to build a little before it starts cooling.

So that's the effect of precipitable water.  What about the effect of the ground reservoir temperature?  This temperature is, in simple terms, just the average temperature of the soil underfoot.  You might guess that a warmer soil temperature would help keep the temperature warmer, particularly overnight as the air cools down.  Does this actually happen?  Let's try increasing and decreasing the ground reservoir temperature by 2 degrees Celsius.  Here's the results:

Just like we might have guessed--if the ground is warmer (the run with the red curve), the air temperature also stays warmer.  Notice that same "reservoir" effect like we saw with precipitable water--the air temperature still continues to rise a bit once we start the model with a warmer ground temperature, whereas with a cooler ground temperature the model air temperature cool immediately.  Both moisture and soil temperature are important for determining how warm or cold we get and when we reach those high and low points.

The next parameter is called "cloud fraction" and it's simply an estimate of how much cloud cover there is in the sky--1.0 is completely overcast and 0.0 is completely clear.  Since it's a clear day today in Seattle, my optimal value is 0.0.  But, we can always try increasing it.  What if we had spotty cumulus clouds covering half the sky (cloud fraction = 0.5)?  Or what if it were overcast?  We might expect it to stay cooler during the day because the sun is being blocked out.  But what about at night?  Here are some results:

Well these results look a bit different.  The red line shows the temperature forecast with a cloud fraction of 0.5, and the blue line shows the temperature trace with a cloud fraction of 1.0.  We can now begin to see some of the hazards of using parameterizations and incompleteness in the model. With more clouds, we kind of expected the clouds to block out the sun and keep it cooler during the day.  That doesn't seem to have happened.  However, at night things stay much warmer with more clouds--this is definitely realistic.  Without clouds, as the earth cools at night, most of that energy would just be lost to space.  Clouds, however, can absorb much of that radiation and re-emit it back down toward the earth's surface.  Thus cloudy nights are warmer than clear nights, and the model definitely seems to capture this.

But, our intuition still tells us that more clouds should keep us cooler during the day.  So what's wrong here?  My guess is that while the model seems to be handling the impact of clouds on longwave radiation (the radiation emitted from the earth) well, it doesn't handle the potential blocking effects of clouds on shortwave radiation (radiation from the sun) well, if at all.  This gives us a clue that we may want to fix something in the model.

Let's try experimenting with the final parameter, the Bowen ratio.  This is a somewhat complex parameter that technically describes the ratio between sensible and latent heat emitted from the surface.  What this is talking about (in simpler terms) is, out of all the heat being emitted from the surface, how much of it is being lost through evaporating water (latent heat) and how much is being lost through simply radiating away the heat (sensible heat).  The AMS glossary gives typical values for this ratio:

• 5 over semi-arid regions
• 0.5 over grasslands and forests
• 0.2 over irrigated orchards or grass
• 0.1 over the ocean

These values make quite a bit of sense--over semi-arid regions (like deserts), there's not a lot of water vapor--so a Bowen ratio of 5 means that 5 times as much heat is released from the desert surface as sensible heat as opposed to heat being lost to evaporating water vapor.  Over the ocean, on the other hand, there's lots of water present--so the amount of heat released from the ocean as sensible heat is only 1/10 of the amount released through the evaporation of water from the surface.

Let's experiment with this value.  Here in Seattle there are a lot of trees and water, so maybe the Bowen ratio is closer to, say, 0.5.  But, on the other hand, this is a city with lots of pavement and buildings.  Those surfaces seem like they'd be hotter (and wouldn't have a lot of moisture).  So, maybe the Bowen ration is more like 2.0.  Let's see what happens:

Wow.  Some significant differences here.  Increasing the Bowen ratio really warms us up a lot, though we cool off to about the same temperature overnight.  Decreasing the Bowen ratio (the blue line) really cools us off--it makes us lose that reservoir of heat and we start cooling off immediately.  You can see that during the next day we start warming up really slowly.  Since the lower Bowen ratio means that more energy is going to evaporating water, there's less energy being used to increase the warmth of the air itself (the warmth we sense--hence the term: "sensible" heat).  Neither of these extreme values seem realistic--we haven't gotten up to 80 in Seattle, nor is the temperature plummeting this afternoon.  So, sticking to a middle value--0.9 to 1.0 is probably good.

This was a very simple model, a "parameterization scheme", if you will.  You can see that just by varying these parameters by a little bit, the results can change quite a lot.  We also can use our intuition and knowledge about the way temperatures typically go to both identify flaws in the model (like how it's probably not handling the effects of clouds on shortwave radiation well) and try to estimate unknown parameters (like trying to pick the right Bowen ratio). 

But I emphasize that this is a very, very simplified model over just one location.  I mean, none of these parameters change throughout the run of the model and there is no advection--this model has no idea if a cold front moves through or if the day starts clear and ends cloudy or anything like that.  And even in this simple model it's very sensitive to changes.  Imagine trying to forecast temperature AND moisture AND winds AND pressure for thousands of points across the country.  It's amazing that our models do as well as they do.

Hope you found this useful and not too boring.  If you'd like a copy of the Python script I used to run this simple model, let me know...

More moderate risks...and differences

We're definitely entering the heart of the spring severe weather season now.  In the wake of last weekend's moderate risks in northern Iowa and Wisconsin where several large and damaging tornadoes were reported, yet another potent shortwave is digging into the central part of the country.

The SPC has issued a moderate risk for eastern Oklahoma and western Arkansas today:

Fig 1 -- SPC day 1 convective outlook for April 14, 2011.  From the SPC.

And then as this trough and surface low move eastward, another moderate risk has been issued for an area in central Mississippi and Alabama with a slight risk area throughout much of the southeast.

Fig 2 -- SPC day 2 convective outlook for April 15, 2011. From the SPC.

Based on this morning's 12Z analyses, there is a sharp shortwave across the intermountain west at 300mb:

Fig 3 -- 300mb analysis of winds (colors) and heights (contours) valid 12Z, April 14, 2011.  From the HOOT website.

Now, compare this to the 12Z RUC model (the Rapid Update Cycle model) initialization:

Fig 4 -- 12Z RUC initialization for April 14, 2011.  From the HOOT website.

Overall the two are very similar--and they should be, as the RUC model should be initialized off of the 12Z observations.  But I do notice some small, but important differences.  The most critical is in the placement of the axis of that trough in the west.  In the objective (computer-drawn) analysis of the observations in figure 3, the axis of the trough is placed through far western Idaho and into central Utah.  However, in the RUC initialization, the trough axis seems to be located in eastern Idaho and down into central Colorado--further east than in the objective analysis.  Which do we believe?

One thing we can check is the water vapor imagery from that time.  Water vapor images retrieve radiation that is typically emitted from the upper-levels of the troposphere--around the 300mb level we're looking at in the maps above.  The strong jet streaks up there perturb and contort the water vapor field so that we can rather easly trace the location of most major jet axes aloft on the water vapor images.  Here's the water vapor image from 12Z this morning:

Fig 5 -- 12Z GOES-W water vapor image for April 14, 2011.  From the HOOT website.

See the sharp contrast between the drier air (the blacker shades) and the moister air (the white and blue shades) that stretches from eastern Nevada through southwestern Utah and into northern Arizona and New Mexico?  That roughly marks the center of the jet streak aloft at that time.  I tried to highlight it using my mouse that has a very poor response time on the image below:

Fig 6 -- Same as figure 5, but annotated with approximate location of the center of the jet streak.

If we assume that the winds are more or less parallel to the height contours (which isn't a bad assumption--we should be near geostrophic balance up there...) then we see that the actual trough axis agrees more with the RUC model initialization--it probably stretches into eastern Colorado at 12Z instead of being further west in Utah.  So...the model seems more believable after all.  This once again highlights the need to be cautious when looking at objective analyses--analyses that were done by a computer trying to fit contours to sparse data.  Using higher-resolution observations like the satellite imagery can help to correct any errors.

As the exit region of this jet streak aloft emerges from over the Rockies this afternoon, we've seen a surface low develop which is now in northwestern Oklahoma as of early this afternoon:

Fig 7 -- RUC surface analysis of temperature (colors), mean sea-level pressure (contours) and winds (barbs) for 18Z, April 14, 2011.  From the HOOT website.

 Note the wind field and temperature fields are a little different from what we usually see.  The warm sector to the east of the low is very expansive--the warm front boundary between colder air to the north and warmer air to the south extends northward out of the analyzed low pressure center before turning sharply eastward in northern Kansas and stretching across the midwest into western Pennsylvania.  This seems like an odd structure for a warm front, as usually we don't see them heading northward out of a low-pressure center.  Also the winds to the northeast of the low are more easterly than southerly--this doesn't do much for advecting in moister air needed to fuel thunderstorms.  Therefore, we might guess that the greater threat for severe storms would be further south to the southeast of the low where winds are southerly all the way down to the Gulf of Mexico.  This area probably has more moisture advecting in.  A check of the dewpoint analysis from the RUC shows that this is indeed where the higher dewpoints are located:

Fig 8 -- RUC 15Z dewpoint temperature analysis with wind barbs for April 14, 2011. From the HOOT website.

And we begin to see why the SPC has placed their moderate risk in that particular region.  Why not further south where the moisture is greater? The capping inversion seems to be stronger there and our jet streak aloft seems to be staying further north.  Without that upper air support that comes from the divergence induced by the jet streak exit region, it becomes much more difficult for storms to form.

Upper-level support is fine, but what about forcing near the surface?  We need something to provide lift.  Usually this is accomplished by a frontal boundary, but we saw that the warm front was way too far north and the cold front really didn't seem to have fully developed yet.  However, the dewpoint map shows the primary mechanism being looked to for lift--convergence of the winds along the dryline.  There is a notable dryline present in the dewpoint map from early this afternoon--the sharp boundary between moister air to the east and drier air to the west in western Oklahoma and north Texas is accompanied by winds shifting from southerly to the east and westerly to the west.  This means winds are converging near the dryline, and convergence at the surface typically forces air to rise.  As the dryline mixes eastward across the state of Oklahoma this afternoon, that convergence zone will move with it and storms will form as surface air is lifted.

Why haven't storms fired already?  Our old friend the capping inversion is still in place--but just barely.  The Norman forecast office launched a special sounding at 18Z this afternoon (around 1 PM CDT).  Here it is:

Fig 9 -- 18Z sounding from Norman, OK (KOUN) on April 14, 2011. From the SPC.

A small capping inversion remains right at around 800mb.  However, the lapse rate (the rate temperature is changing with height) below that inversion is very steep--nearly dry adiabatic.  This is a highly unstable layer--and that probably means that a lot of mixing of air is going on in that near-surface layer.  With such a small cap remaining, this mixing will almost assuredly erode the cap away in the near future.  The dewpoint is registered at 56 degrees Fahrenheit at the surface here.  Any small increases in moisture due to advection from the south will be magnified as increases in the convective potential energy.  Already there is a surface CAPE value of some 2064 J/kg.  Very unstable.

And I won't say much about wind shear, but it's rather strong--winds shift from 10 knots out of the south at the surface to 40 knots out of the southwest at 700mb in the sounding above.  Not the largest shear I've seen, but definitely there.  Furthermore, as evening approaches and the low-level jet kicks up, that shear will undoubtedly increase.  However, by then the threat will have shifted to eastern Oklahoma as the dryline will have already mixed through Norman.

One final interesting note about the movement of this low-pressure center.  The RUC model has the low pressure center moving northward along that north-south oriented warm front boundary before settling in at the point where the warm front suddenly turned east.  Here's the forecast surface map from the RUC model for 10 PM CDT tonight:

Fig 10 -- RUC 12 hour forecast of surface temperature (colors), winds (barbs) and mean sea-level pressure (contours) for 03Z, April 15, 2011.  From the HOOT website.

So the low is forecast to move almost due north throughout the day, according to the RUC model.  Also we can see that as the low moves north the cold front is finally beginning to show up, and it is beginning to overtake the dryline from central Kansas down into northern Oklahoma.  At least according to this model.

However, if we look at a map of surface pressure changes in the last three hours, we see something slightly different:

Fig 11 -- 3-hour pressure changes and wind vectors as of 19Z, April 14, 2011.  From the College of DuPage.

The area of strongest pressure falls over the last few hours has been in far eastern Kansas near the Kansas City area.  To me, this implies that the low pressure center is actually moving more easterly than the RUC model would suggest.  This would speed up the timeline of events a little bit as then the dryline would push through Oklahoma somewhat faster. If these strong pressure falls are a sign of the low pressure center's movement, then I wouldn't be surprised to see the risk area extended slightly eastward as the day wears on.  We can also watch the water vapor image to try and trace the location of the jet streak's exit region over time this afternoon.  If it seems that the jet streak is also moving further east, the surface low will probably try to stay with it as that divergence aloft is what's supporting the low.  We'll just have to see.

Regardless, I expect some impressive-looking storms in eastern Oklahoma today.  The tornadic potential will be high, so be alert and listen for warnings if you live in an at-risk area.

Moderate risk again...and model uncertainty

Another day, another moderate risk of severe weather in the southern US forecast for tomorrow.  Severe weather season seems to be here...

Fig 1 -- SPC day 2 convective outlook from 1730Z, Saturday, Feb. 26, 2011.

The evolution of this system seems to be straightforward--at first.  We start with the current upper-level conditions.  Like we saw with the last severe weather system, we start out with a deep trough off the coast of California this morning:

Fig 2 -- 300mb height (contoured) and winds (shaded) at 12Z, Feb. 26, 2011. 

This trough is forecast to move eastward and become more neutrally tilted (the axis will be pointed more north-south) by Sunday morning.  Both the GFS (shifting to 500mb here so that I can do a comparison...):

Fig 3 -- GFS 36 hour forecast of 500 mb heights (contoured) and winds (shaded) at 00Z, Monday, Feb. 28, 2011. 

And the ECMWF:

Fig 4 -- ECMWF 48 hour forecast of 500mb heights (contoured) and winds (shaded) at 00Z, Monday, Feb. 28, 2011.

Show similar upper-air patterns on Sunday evening.  The exit region of the jet around the base of the trough (particularly the left exit region since this jet seems to be straight and not very curved) would be the favored location for divergence aloft.  Divergence aloft means low pressure and lift at the surface.  So we'd generally expect widespread lift in the eastern Oklahoma-Arkansas-southern Missouri region on Sunday night.  This corresponds well with the western part of the SPC's convective risk above.

However, from here on out things start diverging in the models (no pun intended...).  Compare the GFS forecast for 500mb on Monday morning:

Fig 5 -- GFS 48 hour forecast of 500mb heights (contoured) and winds (shaded) at 12Z, Monday, Fe.b 28, 2011.

To the ECMWF forecast at the same time:

Fig 6 -- ECMWF 48 hour forecast of 500mb heights (contoured) and winds (shaded) at 12Z, Monday, Feb. 28, 2011.

There are subtle, but significant differences in the wind field between the two models.  Notice the location of the maximum wind in the jet.  The ECMWF has the maximum over southeastern Missouri whereas the GFS has the maximum back over southeastern Oklahoma.  The ECMWF also places the critical coupling area between the anti-cyclonically curved jet over eastern Canada and the left exit region of the jet over the central US somewhat further west and north than in the GFS model.  These differences make it difficult to predict how the surface low is expected to move from Sunday night into Monday morning, since the areas of divergence aloft associated with the jet patterns are different in the two models.

But I'm not the only one having difficulty placing the surface low.  We'll look at what the GFS forecasts to happen at the surface overnight.  On Sunday night at 00Z, the center of the low is decently well-defined over western Oklahoma.

Fig 7 -- GFS 36 hour forecast of surface temperature (shaded), sea-level pressure (contoured) and winds (barbs) at 00Z, Monday, Feb. 28, 2011.

Note the strong baroclinic zone stretching east from the low center.  A baroclinic zone is just a fancy name for an area with a strong temperature gradient (that technically should be moving around as well, but...that's difficult to tell).  We would normally associate the temperature gradient stretching east from the low with the warm front.  But, compare the winds to the south and the north of that boundary--those are pretty strong winds to the north of it, blowing out of the north.  If the cold air to the north is advancing, this looks to be a cold front instead of a warm front.  But that wouldn't be consistent with our normal depiction of a surface low-pressure system--a cold front typically doesn't stretch to the northeast of the low center.  We'd expect to see a cold front in that location if the low center was actually further northeast --say, over Missouri or Illinois.  But here is what the GFS shows for the surface pattern just six hours later:

Fig 8 -- GFS 42 hour forecast of surface temperature (shaded), sea-level pressure (contoured) and winds (barbs) at 06Z, Monday, Feb. 28, 2011.  

The center of low pressure has been shifted to the northwest--now over Illinois and Indiana!  This can't be the same low pressure center--even driving on the interstate at 70 mph it takes 14 hours or so to reach Indiana from Oklahoma.  That would imply that the surface low would have had to have moved along at somewhere around 140 mph to move that far in that amount of time.  Highly unlikely.  So we have a reformation of the low pressure center further to the northeast along the baroclinic zone.  But this GFS model only gives us one perspective on where and when this transition will happen.

Here is the SPC's SREF forecast for low pressure centers from all of its different ensemble member models.  At 00Z Sunday night, we see that there's generally good agreement in the placement of the low center in northwestern Oklahoma/southwestern Kansas.

Fig 9 -- SREF 33 hour forecast of member surface low pressure centers for 00Z, Monday, Feb. 28, 2011.

Moving just six hours later, we can see that there is considerable spread among all the members, where they have the position of the surface low (or lows--some model members have two low centers at the same time) at various places along the baroclinic zone:

Fig 10 -- SREF 39 hour forecast of member surface low pressure centers for 06Z, Monday, Feb. 28, 2011.

So there is some ambiguity where the actual center of the low pressure is going to be.  As many storm chasers know, the further one gets from the low pressure center, supercell probabilities tend to weaken.  The low pressure center serves as a focal point for low-level wind shear and the further you get from that, usually the wind shear will tend to weaken somewhat.  Not always the case, though.

So why this ambiguity in tracking where the low center is going to be?  It all goes back into the feedback between the the temperature gradients below, which drive the winds aloft, which generate divergence aloft, which forms low pressure at the surface, which advects the temperature gradients around...and the feedback continues.  Those subtle differences in the placement of the jet maxima and the dynamics of the winds aloft that we saw in the GFS and ECMWF models earlier have drastic implications for just where the surface low will be at any given time.

So what caused the divergence of the GFS and ECMWF upper air pattern forecasts?  After all--they seemed to agree very well through early Sunday evening.  Then they started differing.  But by Sunday evening, we know that convection is going to start forming.  Everyone agrees on that.  The place where convection is likely to form would be along the fronts (the baroclinic zones) where the low-level convergence and lift are maximized.  So we're likely going to have lots of storms firing along the fronts--which is also right underneath the jets.  What do storms do?  They tend to lift lots of warm, moist air into the upper atmosphere and can draw down cool air from aloft.  This radically alters the thermal structure of the atmosphere in areas of convection.  Since the position and strength of the jets themselves is tied to the thermal structure of the atmosphere, this in turn alters the jet pattern.

So, differences in how models handle convection can feed back into altering how the models forecast large-scale, upper-level patterns.  If some models fire more convection or lift more warm air aloft in the convection, they can change the wind pattern aloft to keep things in balance.  Models that aren't as vigorous in their convective prediction may not alter their wind patterns aloft as much.  I'm guessing that differences in how the GFS and ECMWF forecast convection are leading to the differences in their upper air patterns.  I would also suggest that the same thing is happening with the different SREF model members--and, since it all feeds back, that's why 's so hard to place the low pressure center...

Regardless, there's good agreement that moisture at low levels will be there:

Fig 11 - - 33 hour forecast of 2m dewpoint temperature from the mean of the SREF model members at 00Z, Monday, Feb. 28, 2011.

The mean of the SREF models suggests 60+ degree Fahrenheit dewpoints as far north as southern Missouri and southern Illinois on Sunday night (note the strong gradient of moisture over central Oklahoma and Texas showing a classic dryline or cold front pattern).  There is some concern about an elevated mixed layer capping convection over Oklahoma--temperatures are somewhat warm at 850mb aloft.  But if that can be overcome, there is more than enough moisture to get some good storms.  Winds aloft are conducive for lift and shear also looks respectable.  So the potential for severe weather is definitely there.