Tuesday, September 18, 2018

Intricacies of the September Monsoon

We have reached the time of year during which the mid-latitude westerlies are strengthening and becoming more active yet the so-called "monsoon" remains active.  Knowledge of the interactions between these two circulation regimes is often important for weather forecasting not only over the southwest U.S., but other parts of the continental U.S., including the midwest.

An excellent example is provided by today's forecast.  The situation at 1200 UTC (6 AM MDT) this morning shows an upper-level short-wave trough in the mid-latitude westerlies just off the Pacific Coast.  At the same time, a tongue of monsoon moisture extends northward over the Gulf of California, with relatively high precipitable water values (color contours) across much of the surrounding region.  I have drawn an arrow below to indicate the flow that has led to enhanced precipitable water values along this tongue. 

The latest GFS forecast shows how this moisture is drawn northward ahead of the upper-level trough as it moves eastward across the Rockies.  This contributes to a significant round of precipitation across the eastern Rockies and the upper plains beginning tomorrow. 

Sadly, the phasing of the two systems is a bit off to bring precipitation to northwest Utah.  Most of the action is in eastern Utah and Colorado.  Instead, all we will see is a weak cold frontal passage accompanying the upper-level trough as it moves across the state.

This is shown well by the FV3 forecast below.  The FV3 is a new experimental model being developed by the Geophysical Fluid Dynamics Lab and the National Centers for Environmental Prediction.  It will eventually become NCEPs Global Forecast System (GFS) probably in early 2019, with the old core of the GFS relegated to the scrap heap.  Experimental forecasts are being produced this summer, with the first 84 hours available at weather.utah.edu.  Other web sites will have the full model run.

The FV3 forecast for 0300 UTC (9 PM MDT) tomorrow evening shows the trough over central Utah, with the wind shift just upstream of the Pole Canyon and Bald Mountain Fire regions (note: this is one model run and shouldn't be used in isolation to anticipate a wind shift at the fire location).  Precipitation is scattered across eastern Utah, Arizona, Colorado, and New Mexico.  Although there is a skiff of precipitation over high terrain in northwest Utah, as things stand now, it appears the western half of the state will be left bone dry.  The monsoon surge is close, but not close enough. 

Should convection be just a bit west of that indicated above, however, outflow from either the storms to the east or "cumulus patheticus" accompanying the front could affect the fires.  Further, there will be a wind shift accompanying the front.  Situational awareness will be important tomorrow and tomorrow night to be prepared for the shifty situation. 

Monday, September 17, 2018

Winds During the Pole Canyon Fire

As we have discussed in two previous posts and covered on local news, the Pole Creek and Bald Mountain Fires exploded last Wednesday (12 September) and have forced widespread evacuations.  The map below shows the situation as of yesterday morning, with the Bald Mountain fire to the west and the fire perimeter near Elk Ridge and Woodland Hill and the Pole Creek Fire having done damage on both sides of US-89.

There is a paucity of wind data in the region surrounding the fire, so I need to get creative here.  We'll take a look at two sites.  The first is the Birds Eye "IRAWS" (incident remote automated weather station) at 6544 ft elevation in the mountains immediately east of Birdseye along US-89 (top graph below).  At this location, winds through the 10th of September exhibited a clear diurnal cycle, increasing during the day and decreasing at night (red line sustained, green pluses gusts, blue dots wind direction).  Sustained winds during the day, however, were less than 10 mph, and gusts through the 10th were less than 20 mph. 

Source: MesoWest
It was on the 11th and 12th that things began to change with persistent winds from the ESE-SW predominating and wind gusts each afternoon on the 11th, 12th, and 13th reaching over 20 mph.   It was on the 12th that news reports suggest the fires grew "exponentially."  There was a lull in the winds on the 14th and early on the 15th, but the srong southwesterlies remerged on the afternoon of the 15th and dominated yesterday as well.

The site above is at only 6500 feet elevation, so we might also look at what is happening higher up.  This is more difficult, as there are no upper-elevation sites in the immediate area, so I've plotted up the observations from the Arrowhead Summit site at Sundance Resort, which is at about 8200 feet (bottom graph above).  At that location, you can similarly see the daily wind effects until the afternoon of the 10th when steady SSW flow develops and then predominates until the present.  There is a clear ramp up of wind speeds from the 10th to early in the morning on the 13th, then the decline also seen at the Birds Eye IRAWS, and then things pick up again over the weekend.

The one plus recently is that the winds thus far today are weaker than they were over the weekend, and the area is not presently under a red flag warning, but only a fire weather watch.  Nevertheless, winds will pick up some today and fire conditions will still be difficult.

Model forecasts suggest that a shift in the winds will probably occur Wednesday or Wednesday night, depending on timing and elevation.  The NAM time-height below, which is for the Salt Lake City airport and thus shouldn't be used specifically for the conditions at the fire shows a shift in the flow occurring during the day on Wednesday first at lower elevations and then extending up to about 700 mb (roughly crest level).  This is associated with a very weak cold front that will bring in cooler air, but not substantially cooler air.

These fires and the impacts of Hurricane Florence over the past few days are a reminder that people are putting their lives on the lines to protect and rescue others.  My thoughts are with them and all who are working to minimize the effects of these hazards.

Addendum @ 10:35 AM

17 September fire update included below.

Sunday, September 16, 2018

Fall Colors, Pole Canyon Fire, and Precipitation Drought

My mom is in town, so we spent the day on the Alpine Loop Highway with a visit to Sundance.  The colors in the upper elevations are starting to change and probably next weekend it's going to be quite nice.  Even this weekend was fine.

Less beautiful were views of the Pole Canyon and Bald Mountain fires, although I'm not sure we could see the latter directly.  Note the cumulus sitting on top of the smoke plume, a likely reflection of heat and moisture from the fire.

Some of you have asked about how dry it is.  I will focus first on precipitation.  For the first half of September, we've had a trace of precipitation at the Salt Lake City Airport.  Such dryness is not exceptionally uncommon.  Going back to 1875, there have been 13 years with no precipitation, 16 years with trace, and another 18 with .05" or less of precipitation in the first half of september.  Basically, you have 47 out of 145 years (31%) in which the first half of September yielded only .05" or less of precipitation, so a dry run to start the month is not exceptionally unusual.

However, what is making for a remarkably dry scenario is the below-average snow last winter, the below-average rainfall this summer, and the above-average temperatures.  In the case of the latter two, the airport has only had 0.74" of precipitation since June 1st, which is the 7th lowest on record.  Temperatures since June 1 have averaged 77.8˚F, which is the 7th highest on record.  It's a double whammy for drought to have both poor precipitation and abnormal warmth.

The entire state of Utah is in moderate to exceptional drought.  Most of the western U.S. is abnormally dry to worse.

Source: NOAA
The role of temperature will continue to increase in future droughts as the climate warms.  Even if average precipitation remains steady, higher temperatures will lead to more severe and persistent drought conditions in our region.  Dry periods will be worse than they were in the 20th century.  Average precipitation periods behave more like abnormally dry or moderate drought periods.  Wet periods will need to be more prolonged to pull us out of the water deficit.  This situation will become worse as the climate warms through the 21st century.

Saturday, September 15, 2018

Quick Look at the Meteorology of the Pole Creek and Bald Mountain Fires

The Pole Creek and Bald Mountain Fires in Juab and Utah counties blew up this week, have burned 68,000 acres, are threatening homes and structures, and are currently only 2 percent contained.  I first saw the smoke billow or pyrocumulus cloud (tough to tell if there was a pyrocu or not) associated with one or both of these fires on Wednesday when it poked up above Lone Peak.

Yesterday's MODIS image from NASA's Terra satellite showed smoke from the fires traversing the Uinta Mountains and extending all the way to the Black Hills and beyond.  The remarkable linear nature of the plume likely reflects the strong southwesterly flow that has predominated over the western U.S. the last couple of days. 

The Salt Lake Tribune reported this morning that on Sunday the Pole Creek fire was a half acre in size.  It grew to 50 acres by the end of Monday and then grew "exponentially Wednesday and Thursday.   Collectively, the two fires have burned 68,000 acres. 

The decision to allow the fire to burn has been a subject of criticism by residents and Lt. Governor Spencer Cox who tweeted on Thursday that the fire was an example of 'more inept decision-making by the Forest Service who decided to try and "manage" this fire and let it burn instead of suppression — during one of the worst droughts in recent history.'  Governor Gary Herbert was more circumspect and is quoted in the tribune as saying that while there is room for improvement, "it's not time to finger point today." 

Computer model forecasts from Monday through Wednesday did indeed call for an increase in southwesterly flow, with relatively low relative humidities.  The forecasts below provide a few examples, valid at noon Thursday. 

However, more needs to be known about the decision making process, whether or not a forecast was requested by the forest service, and what those forecasts predicted, to better understand the decision-making process in this instance.  Thus, I am reluctant to comment further without digging more deeply into the case. 

In other words, I agree with the governor that the main issue at hand at this time is to deal with the emergency.  Sadly, the next couple of days look difficult meteorologically, with continued warm, dry weather the next few days.  The time-height below is for Salt Lake, but is probably reasonably applicable for the large-scale (but not local) conditions at the site.  The one plus is a slight weakening of the flow during the period. 

Updates concerning the threats posed by this continued warm, dry southwesterly flow is available at https://inciweb.nwcg.gov/incident/6199/.

Friday, September 14, 2018

Some Perspectives on Florence

So many thoughts have been racing through my head over the past 24 hours or so since Florence began to have a major impact on the Carolinas and environs.  I will start with a quick look at the rainfall.

Radar estimated precipitation from 0343 EDT Wednesday 12 September through 0834 EDT Friday 14 September from the Morehead City radar shows a broad region of more than 3 inches extending from southwest of Wilmington to near Cape Hatteras along the North Carolina coast.   The heaviest precipitation has been in the region surrounding Morehead City, with some localized estimates of more than ten inches (magenta shading). 

Source: NWS
Gauge observations from a Remote Automated Weather Station (RAWS) site in Croatan National Forest northwest of Morehead City shows and accumulation of almost 11 inches, which is consistent with the radar estimates.  This precipitation falls over a roughly 12 hour period. 

Source: MesoWest
Precipitation frequency estimates from NOAA (https://hdsc.nws.noaa.gov/hdsc/pfds/pfds_map_cont.html?bkmrk) suggest such an accumulation has an average recurrence interval of about 1 in 100 years or, better put, a likelihood of about 1 in 100 in any given year. 

Estimates of such frequencies have a number of problems for extreme events, so I provide the numbers above merely to indicate that this is a great deal of precipitation, even for this part of the world. 

Florence made landfall this morning near Wrightsville Beach, NC, but is moving very slowly.  It was category 1 at that time, but as we have discussed, do not fixate on category, which is based solely on maximum sustained winds.  Florence has winds that have and will do damage, but flooding associated with storm surge and heavy rain have been and remain the biggest threat.  The slow movement of the hurricane means accumulations will continue in much of the coastal Carolina region and eventually spread into South Carolina and then inland.  If in the region or watching for friends or family, monitor official forecasts and heed the advice of local officials. 

As occurred during the landfall and subsequent stalling of Harvey last year, I had a dramatic change in emotions late yesterday.  As a meteorologist, you want to see the forecasts produced by your profession be accurate and helpful, but at the same time, one gets nauseous when you begin to realize that predictions of catastrophe are coming true.  Reports of 150 water rescues underway in New Bern were especially impactful as I thought of people in terrifying situations and the first responders putting their lives on the line to save them.  I am thinking of all this morning and hoping for the best. 

Finally, nobody knows better than meteorologists that are forecasts are imperfect.  Sometimes they are not well communicated or the impacts cannot be fully described due other to uncertainty or incomplete knowledge.  However, there are important lessons to be learned from Harvey and Florence.  In particular, every hurricane is different and the past is not necessarily a good predictor of the future.  Past individual experiences during hurricanes and other hazardous weather events are not necessarily good indicators of how future events will unfold.  The sample size is simply too small and the climate system is changing.  Computer forecast models, which are based not on human experience but physics, including conservation of mass, energy, and momentum, combined with human interpretation may be imperfect, but provide the best guidance for decision making.  Something to keep in mind as predictions of future events unfold. 

Wednesday, September 12, 2018

The Devil Is in the Details (and the Uncertainty)

Hurricanes provide an incredible perspective on what Demuth and Morss (2012) called the creation and communication of risk information. 

Most people today get risk information from a nebulous web of sources that can include radio and television media (local and national), social media (e.g., twitter and facebook), and friends and family.  This web of information is highly complex and strongly affects impressions of the risk information and ultimately individual decision making. 

During hazardous weather events such as hurricanes, I often share information with family and friends via social media.  Over time, I have streamlined those posts to emphasize the forecasts produced by the National Weather Service, which includes the National Hurricane Center, and the importance of following information provided by local officials.  I realized that in many cases, adding my $0.02 about the evolving weather situation was just adding to the noise and obscuring the message. 

I am a huge fan of twitter and as a professional I benefit greatly from following the feeds of many individuals in the atmospheric and related sciences.  The perspectives and information provided by those individuals can be quite valuable for me, and often provides ideas or content that I can use in class.

On the other hand, tweets are limited.  Even those produced by subject-matter experts often times lack context or fail to provide the big picture.  Often they discuss a scenario or a specific local outcome rather than the full range of scenarios.  This is especially true today when we have models that run out for many days into the future and provide very physically realistic (but not necessarily skillful) forecasts.

The forecasts produced by the National Weather Service are not perfect.  There are accuracy errors and communication shortcomings.  For example, the graphic at lower right includes a 1-7 day precipitation forecast that contains no illustration of uncertainty.  I have watched over the last 2 days as the position of the precipitation maximum has moved around with each new forecast.  We still do not know precisely where the heaviest precipitation will fall and that graphic conveys greater confidence than is warranted. 

On the other hand, overall the National Weather Service forecasts are generally reliable and provide a steady but evolving perspective on very difficult forecast challenges.  For example, messages appeared on my phone stating that there had been a shift in the track of Florence, and I have heard that several times this morning as well.  If we go back to the track projected by the National Hurricane Center two days ago (issued 11 AM AST Monday), we that that the uncertainty cone beyond day 3 (stippled area) covers nearly all of South Carolina, North Carolina, and Virginia, as well as offshore areas.  That is the area that was likely to include the storm center on Friday and Saturday. 

If we look at the forecast issued this morning, we see that on Friday and Saturday, the low center uncertainty cone covers northeast South Carolina, southeast North Carolina, and offshore areas (note that the stippled region is now valid for Sunday and early Monday). 

Ultimately, one shouldn't verify a forecast with a forecast, but the forecast position of the storm through Saturday remains in the probable path region projected by the National Hurricane Center.  
Reflecting this uncertainty,  the "key messages" graphic issued by the National Hurricane Center at 5 AM AST Monday Sep 10, 2018, it explicitly states that "it is too soon to determine the exact timing, location, and magnitude of...impacts."  Further, South Carolina has consistently be mentioned in these key messages as an area of concern.  

Every hurricane is different and provides learning opportunities about how to improve forecasting, messaging, and response.  My take away message from this post is that individual perspectives on the weather forecast are strongly weighted by the nebulous web from which we access and receive information.  Finding ways to reduce the noise and focus on what is important – official forecasts and the recommendations of local officials – is your best option.