Tuesday, May 12, 2015

The high desert gets soaked, but not as bad as Oklahoma

After days and days of high-precipitation thunderstorms with embedded tornadoes and large hail, Oklahoma is finally getting a bit of a break.  Just how wet was it there?

The Oklahoma Mesonet recently published rainfall totals for the past two weeks.  Here's what it looked like:

Huge rainfall totals across the state.  Over 12 inches in some locations in only 2 weeks.  The annual percentage of normal rainfall has spiked through the roof.  So far through the roof, in fact, that it's maxing out the colorbar on the Oklahoma Climate Survey plot, with most places greater than 180% of their total rainfall.
Needless to say, the latest monthly outlook from the Climate Prediction Center has drought conditions in Oklahoma and Texas "improving" over the next month...
But drought persists elsewhere---California is seeking no help and the persistent troughing has kept most of the heavy rain-makers to the south of the upper midwest, causing drought conditions to develop in Minnesota and the Dakotas.

Eastern Washington and Oregon have also been running a bit low on their precipitation and are in "drought" conditions as well.  A warmer-than-average winter has allowed the western side of the Cascades to see about normal precipitation, but without much snowfall accumulating in the mountains.  This is bad news for the eastern sides of the states, where they don't see as much precipitation, but rely on meltout from the snowpack to supply them with water for agriculture.

Here's a plot showing the snow water equivalent normals and what we've had this year at Stampede Pass, which is in the Cascades southeast of Seattle.
The light blue line is the normal Snow Water Equivalent, or the mass of water in the snowpack.  The dark blue is this year.  You can see that by this time of year we should be just dropping down from our peak snow in mid April, but this year we're already at zero snow.  Looking at a map of percentage of average snowpack as of last week, we can see pretty much everyone is basically melted out (or at least below 25% of where they should be).
But there is some hope for the eastern side of the mountains!  An upper-level low is churning its way through Oregon at this time.  It's moving slowly and a bit further south than usual, and this is allowing a fair amount of moisture to get to the eastern sides of the Cascades.  Here's the latest water vapor satellite image.
We can see the swirl of the low with its attendant moisture streaming up through Nevada and into Idaho.  That low is forecast to slowly lift north-northeastward over the next couple of days.   With cold air underneath that trough, we're expecting some destabilization and a few rounds of thunderstorms to develop.  All of this in an area that does not get much rain to begin with.  Eastern Washington and Oregon are basically deserts; here's a map (from the PRISM group) of annual average precipitation across Oregon.  Much of the high desert east of the Cascades sees 10-15 inches or less of rainfall per year.
Here's a similar map for Washington, with the same story away from the higher terrain:
Our middle- to long-range forecasts are calling for several inches of rain by the end of the week east of the Cascades.  This is an unusual setup.  Here's the total rainfall accumulation by Friday evening from this morning's local WRF model run.
Large swaths of area getting 1-3 inches of rainfall.  Great for farmers!  It may seem like these numbers are small in comparison to Oklahoma, but keep in mind that these are places that are lucky to get a tenth of an inch of rain with a good convective shower.  There are usually a few thunderstorms that occur throughout the year that will give a more thorough soaking; this will be one of those periods.

Monday, May 4, 2015

Forecasting a Subtropical Storm

Model forecasts have now been consistently hinting that a small, but powerful storm could form off the eastern Florida coast by the end of this week.  This storm isn't quite "tropical" but it's not quite "extratropical", so we use the term "subtropical" to describe it, as it would have characteristics of both tropical and extratropical cyclones.  Let's see what this looks like in the model forecast fields.

First an overview of the large-scale setup.  Here's the GFS forecast (from HOOT) for this Wednesday afternoon at 300mb, right before this cyclone is forecast to really get going:
Winds are shaded and heights are contoured.  We see two main "belts" of air here.  First, our friend the polar jet stream is on its way north, as is typical during the spring time.  There's a deep trough forecast over the Pacific Northwest (ending our wonderfully warm and clear weekend we had been enjoying...).  There's also a negatively-tilted ridge over the northern plains and into the upper midwest.

Further south, we have a strengthening subtropical jet stream, most visible over the eastern Pacific into Baja California as a strengthening jet streak there.  However, extending east a little further we can see that the height gradient associated with the subtropical jet, while weaker, is showing a trough over the southeastern US and Florida.  A trough there means that there must be a pocket of slightly colder air over the southeast, with some kind of gradient in temperature to the south.  We know that extratropical cyclones get their energy from temperature gradients, and we usually see those develop in association with a trough aloft.  So, seeing this setup, one might expect an extratropical cyclone to develop off the southeast coast as that trough moves east-northeastward.

Well, a cyclone is indeed forecast to develop there.  Let's zoom in now and take a look at what the deterministic ECMWF forecast (courtesy of Wunderground) from last night is saying about this developing cyclone.  Let's start with the sea-level pressure (blue contours), precipitation (colored shading), and 1000-500mb thickness (yellow contours) for 21Z Thursday:
There is a clear surface low forming off the southeast coast.  Note that the forecast precipitation is somewhat "comma" shaped: it's NOT a symmetrical circle.  Asymmetrical (comma-shaped) precipitation patterns are closely associated with extratropical cyclones.  You can even see what looks like a hint of cold and warm fronts in the precipitation: very extratropical. Tropical cyclones tend to be much more circular and symmetric.  So, again this looks more like an extratropical cyclone.

But let's look at the thickness contours (the yellow lines).  In the core of that storm, there's actually a small solid contour of thickness, at 5700m.  This is a positive thickness anomaly in the center of the storm.  What does that mean? Higher thicknesses mean that the column of air being measured is warmer on average.  So, a positive thickness anomaly would point more towards this storm being a warm-core cyclone.  Tropical cyclones are warm-core cyclones; whereas extratropical cyclones are cold-core cyclones (they tend to be colder in the center).  So, this warm-core characteristic seems more indicative of a tropical cyclone.

Let's compare this to the GFS forecast for a slightly earlier time (09Z Thursday; the GFS model develops this storm more quickly):

Here the precipitation might be slightly less clearly comma-shaped, though there is a large dry slot to the south of the low.  The warm-core idea is again shown with higher thickness values in the center of the storm.  So different models have varying degrees of warm-core-ness.

The thickness plots above are based on temperatures in the 1000mb to 500mb layer.  Let's explore temperature further by looking at the 850mb temperatures from the ECMWF forecast, which would be right in the middle of that layer:
You can see that there are definitely temperature gradients in the vicinity of the cyclone, with cooler air being pulled down from the north on the western side of the storm and warmer air being pulled north on the east side of the storm.  So the warm "core" we see in the thickness isn't as pure as we might expect from a fully tropical cyclone.  Let's check the GFS forecast, here at 21Z:
The GFS definitely has a pool of warmer air at the center of the storm, making a stronger case for a warm-core cyclone.

One tool to help decide if a cyclone is tropical or extratropical (or something else) is looking at something called tropical cyclone phase diagrams (from work by Robert Hart at Florida State).  Here's the forecast phase diagram for the cyclone in the GFS forecast:

The type of cyclone basically depends upon the thickness (the y-axis on the diagram) and the thermal wind (which is related to temperature gradients; the x-axis on the diagram).  By measuring these two things and plotting them as the cyclone evolves over time, we can see what "region" of the plot the storm falls into.  Here, the storm's trajectory (the dots along a line going from A to Z) are all in the lower right part of the chart---that "red" sector is labeled "symmetric warm-core", aka tropical cyclone.  An extratropical cyclone would be "asymmetric cold-core".  So, interestingly enough, according to this one method of looking at cyclones, they would forecast the storm forecast to develop off the southeast coast as a tropical cyclone.

As we've seen above, the latest forecasts show this particular cyclone to have aspects of both a tropical (warm-core) and extratropical (cold-core) cyclone.  So what do we do?  We call it a "subtropical" cyclone, sort of a catch-all term that describes storms that have both warm- and cold-core attributes.  More thorough definitions are at this page and this page.  If the maximum winds are less than 34 knots, it will be labeled a "subtropical depression".  Greater than that, it will become a "subtropical storm" and actually given a name from the 2015 hurricane name list by the NHC (which would make this Subtropical Storm Ana, should it get strong enough).

Regardless of which model you believe, both bring heavy rain and strong winds to the Florida, Georgia and Carolina coasts Thursday and Friday.  The ECMWF forecast has it just reaching subtropical storm strength at this time...
But there remains much uncertainty.  Not only are the GFS and ECMWF different, but the ECMWF ensemble forecasts also have a fair amount of spread.  Here's a plot comparing the ECMWF ensemble mean sea-level pressure forecast on the left with the single high-resolution deterministic forecast (what I've shown above) on the right:
While it's clear on the right that there is a strong low off the Carolinas, we don't really see that in the ensemble mean on the left panel.  What we do see in the left panel is a big blob of purple in that region, which means "greater than normal uncertainty".  In other words, different forecast runs of the ECMWF have different answers as to how strong and where exactly a storm may form off the coast, if one even forms at all.  So, we'll have to be watching all week to see what develops.