History Repeats Itself

The 2014 Atlantic Hurricane season was, for the most part, a bitter disappointment.  Only 8 named storms formed from June 1-November 30, the fewest since 1994.  All of those storms were restricted to July 1-October 28.  For those of us who enjoy the thrill of tracking hurricanes and maybe even making our own predictions, there was little to be excited about.  Of those 8, 4 had a direct impact on land, with two of those being weak tropical storms.

On the other hand, the 2014 Eastern Pacific Hurricane season was incredibly active, with the most named storms since 1992.  There were 20 named storms that formed in those waters.  However, this season had a wider active frame than 1992.  In 1992, the storms formed from June 1-October 30, whereas this season had storms form from May 22-Nov 5.  There were 9 major hurricanes in the Eastern Pacific, which is more than the Atlantic had in total named storms!

This is quite the interesting situation, so some numbers shall provided to make a point.  This table shows the number of named storms experienced each year in both the Atlantic and Eastern Pacific oceans.  From this table, not much alone can be determined

The above chart does not show much that can immediately be concluded.  This can be said, however: there are three major peaks on the Pacific graph in 1985, 1992, and 2014.  Those peaks, correspond to summers immediately following La Niña, El Niño, and El Niño events respectively.  I will leave the rest up to you, the reader, to conclude, because I do not have any concrete answers to this mystery.  1992 and 2014 are consistent with what is known about El Niño, but 1985 makes an answer into a conundrum

The Difficulties in a Florida Forecast

Each of the last two days, the National Weather Service offices in Peninsular Florida have been caught off guard by the predictable daily pattern.  On each day, the forecast was predicted to be normal: hot and humid with thunderstorms.  On both days, there was a tornado, one near West Palm Beach on Saturday and one near Fort Pierce on Friday.  Naturally, the NWS was surprised and is now adding tornado/funnel cloud potentials to their forecasts, but it demonstrates how no one is immune to letting their guard down against the weather when phenomena go unseen for long periods of time.

 Thunderstorms are part of daily life in Florida.  In fact, Florida leads the nation in thunderstorm days, so correct your friends next time they say they came to Florida for the sun.  This leads to the topic: how many factors go into a Florida thunderstorm forecast?

Here are some of the factors that go into the forecast: the air temperature, the humidity, the 300, 500, 700 and 850-mb patterns, surface highs and lows, precipitable water (PWAT), the time of formation and strength of the seabreeze, local cloud cover, etc.  These factors will be presented in the order they are listed.

Air temperature: obviously it is pretty hot in Florida during the summer.  High temperatures are typically in 88-94°F range, with some days getting up to 98°F but never reaching 100°F.  Without the high temperatures, there are no thermals; the thermals are necessary to cloud formation.  With so much heat at the surface, the thermals can continuously generate.  Morning winds are generally weak, which helps support the thermals.

Humidity: there must some surface moisture to support the formation of clouds.  More details shall come  later.

The 300-mb pattern: this factor is most useful in forecasting the future of surface highs and lows.  From the analysis of the 300-mb map, the forecaster can also see whether upper-level moisture will be moving toward or away from a region.

The 500-mb pattern: conditions at the 500-mb layer, particularly temperature and dewpoint, have long been the most important factor in several forecasting indices.  Surface winds may be conjectured from this layer.

The 700-mb pattern: here is one of the most important factors.  From the winds at this layer, the direction in which thunderstorms will travel may be determined.  This "synoptic flow" most often will determine the trajectory of thunderstorms when the other factors that affect this motion (especially seabreezes) are weak or are cancelled (such as when two storms collide and form a new thunderstorm).  Also, if there is little moisture at this layer, then storm formation greatly diminishes.

The 850-mb pattern: much like the 700-mb pattern, a lack of moisture at the 850-mb layer diminishes storm formation.  Lapse rates above this layer must be low enough to keep the atmosphere unstable to allow for cloud/storm formation.

Surface patterns: the Bermuda High and other highs to the east of Florida greatly favors storm formation, as does a low to the west, because both cause moisture and warmth to push up from the south and into the peninsula.  If highs and lows are positioned so as to drive down winds from the north, then storm formation is diminished. 

PWAT: precipitable water tells the forecaster how much moisture is in the entire atmospheric column.  If this value is low, then storms are not likely to form.  Typical values favoring storm formation are > 2.0".  The distribution matters as well as mentioned in the 700 and 850-mb patterns.

Time of formation and strength of the seabreeze: the timing of the seabreeze is important.  If the seabreeze forms too soon, then the air will be cooled off and the thermals that generate due to daytime heating will be weakened.  If it forms too late, then the east and west coast seabreezes will not be able to collide.  The same is true of weak and strong seabreezes.  On top of all this, the seabreezes can collide and, especially if one or both seabreezes are weak, not generate any storms.  When the east coast seabreeze generates cloud convection, the storms tend to be stationary or move weakly to the west.  When the west coast seabreeze does the same, the storms tend to move to the east.  This is because the synoptic flow is generally from the west moving toward the east.

Local cloud cover: this is the last factor that will be discussed within this post.  If there are clouds already present over an area, then daytime heating is severely diminished, pretty much ending all chances of thunderstorms forming later.

Limitations of the Meteorologist

Over recent years, the National Hurricane Center in Miami has increased forecast accuracy for storms.  The error in storm tracks has reduced by hundreds of miles, meaning less money is spent on evacuations due to fewer false alarms.  Short term intensity, however, has continued to be a problem.  This leads to the Central Pacific Hurricane Center (CPHC) in Hawaii, who like the NHC is a subset of the National Weather Service.

Yesterday, Tropical Storm Genevieve was being forecast to move west, then take a slow turn to the north.  Along the way, the storm was to strengthen to strong category 1 winds (74-95 mph).  The track forecast has been quite accurate.  Error has been kept within the CPHC/NHC guidelines.  The intensity forecast, on the other hand...

 Genevieve is now forecast to be a category 5 hurricane (winds > 135 mph).  This left me wondering where the forecast went wrong, and if there actually was a chance of the forecast being correct.

The models from yesterday show a consensus on the storm eventually reaching category 3 or 4 strength over a 24-48 hour period.

The Coupled Hurricane Intensity Prediction System (CHIPS) models had a similar agreement, except for CHP6.  CHP6 actually predicted this accelerated strengthening.  CHIPS is just a two dimensional model that uses radius and height as its coordinates.  To counter this, CHIPS uses vertical data from the Global Forecast System (GFS) model to create a full-storm profile.  Through this, CHIPS can simulate the effects of wind shear on storm intensity.  CHP6 represents the upper boundary of the storm intensity forecast by increasing the storm intensity and setting vertical wind shear to 0.  The problem here is that CHIPS is one of many experimental model systems and is not part of the official forecasting system yet.

 

 So I decided to conduct my own analysis of this using the GFS model.  Using this model proved valuable as the animation below shows.  The red wind barbs show 200-mb winds, and the magenta barbs show 850-mb winds.  As can be seen in the animation, there is not much of a difference between the two in strength.  There is some directional shear, but this kind of variability is normal and does not affect the storm intensity.So Tropical Storm Genevieve was actually entering the low speed shear environment favored by CHP6.  However, CHP6 does not use the GFS 3-D or any of the same environmental parameters (carbon dioxide, ozone, etc.).

 

The Central Pacific Hurricane Center got caught by a strange set of circumstances.  Few storms form in the Central Pacific each year, with some years even having zero storms.  Overall, anomalous occurrences such as this should not be judged harshly as there are dozens of model outputs for the forecaster to analyze.  Few storms of this strength enter this region, so events like this are  opportunities for meteorologists of all backgrounds to learn.  The developer of CHIPS, Professor Kerry Emanuel of Massachusetts Institute of Technology, should be quite proud of his work this morning.

Hurricane Bertha and the Pacific

Note: My internet provider is disgraceful.  I have been able to access the NHC, or any other .gov website, only about 5% of the time for the last 4 months, so it is almost impossible for me to get NHC imagery unless I happen to catch it at the right time.  In fact, I have not had NHC/NOAA access since yesterday afternoon.

The Atlantic hurricane season is now 66 days in.  During that time, there have been just 3 storms, one of which was an unnamed tropical depression.  This is somewhat irritating to a meteorologist, because more storms to track means getting to look at models and radars and getting to test my own predicting skills.

My personal issue with this storm is the lack of organization and convection.  But it had hurricane force winds!  So the key is to determine whether or not it should be called a hurricane.  During Bertha's time as a tropical storm, it was pretty obviously non-tropical.  No circulation and no centralized convection was present.  The storm clearly had tropical characteristics, however.  It formed as a wave off of Africa.  It formed in the warm waters of the Atlantic.  Further, the convection, while not centralized, clearly was a single traveling unit.  This fits a better description of a subtropical storm.  While this does not change the naming of the storm, it does change the general opinion.

It is yet to be seen how many storms there will be this season.  With the El Niño of the previous winter, a generally less active season is to be expected.  The last season to go this deep with only two named storms was 2009, when the list only reached "Ida."  Nevertheless, in 2004 "Bonnie" was not reached until August 9; that year the list reached "Otto."  So all is not lost yet!  For the foreseeable future, the chance of tropical cyclone is not that high, however.

The Pacific, on the other hand, has already reached Julio.  The last time Julio was reached this early in the season was 1992, which was the year the Pacific famously used every name on the list,  in spite of the fact that the Eastern Pacific has "X," "Y," and "Z" storms!  The Eastern Pacific season tends to start trailing off much earlier than the Atlantic, sometimes even in September with few if any storms forming in late October or November.

For those of us who love meteorology, all is not lost!