Forecasting Tornadoes In Oregon

It was just last month that I wrote about why the Pacific Northwest is not a tornado "hotspot". Of course, as a meteorologist, I should know better than to challenge Mother Nature like that. On Thursday, McMinnville was hit with an EF-1 tornado. No injuries were sustained but a good amount of damage was done to several buildings and highway 99W was re-routed around Lafayette. The twister was the first in Oregon for 2013 (Washington had an EF-0 storm in Battle Ground in March) and the first tornado since 2011. The Willamette Valley last saw a tornado back in 2010 when an EF-2 tornado struck in Aumsville. So they DO happen! But why can't we forecast these "cold-core" twisters that touch down in the Pacific Northwest?

Here is a snip-it of an essay by John Sauder, a Canadian meteorologist, who describes a "cold core funnel":

"A cold core funnel is a vertically tilted rotating column of air under a rapidly growing convective cloud, but the atmospheric conditions are different than those conditions that produce typical funnel clouds or tornadoes"

Imagine a typical Oregon spring rainy day. You know, those days where the weather doesn't seem to be able to make up its mind? Rain, sun, clouds, hail, sun, rain. The environment described by Suder above is very similar to those kind of Pacific Northwest days. Tornadoes in the Midewest and Southeast do not form this way. Tornadoes in "Tornado Alley" are much easier to see coming, not to mention much larger, and thus it is easier to forecast and issue warnings. No tornado or severe thunderstorm watches or warnings were issued with yesterday's McMinnville cell. 

Radar is an amazing weather product that helps meteorologist identify characteristics of a tornado. This radar image has almost NO characteristics of a typical tornado. Take a look at the radar comparison between McMinnville's tornado and the tornado that ripped through Moore, Oklahoma last month:

Glaring differences! On the right, the Moore tornado signature is a classic! This is what is known as a "Hook Echo", the rain shown wrapping around some very intense rotation. No such signature in the McMinnville radar image. The only similarities between the two appear to be rain intensity. The pink on both images show heavy rainfall and most likely hail present as well. The Moore radar had a Tornado Warning associated with
it, while the McMinnville did not. Why was no warning issued for yesterday's storm? Sauder offers up an explanation:

"...these cold core events happen on a small scale (mesoscale) and are typically rather short lived."

This would answer why it is so difficult to issue warnings on Oregon twisters. They just pop-up! Just like those showers and sun-breaks that are so common during our spring weather. Another tool radar offers up is called "velocity". That helps identify wind directions in a storm. Here is a typical tornado signature using the velocity tool: 

Tornadoes are often found right were green meets red. The different colors mean that winds are moving in opposite directions, and where they touch is where winds are spinning in a tight rotation. This velocity signature is one most tornadoes in the Midwest and Southeast contain. But here in the Northwest, we don't often see this strong of a signature. Sauder says,

"The velocity signatures are very small and usually blend in with background noise...makes cold core funnel or cold core tornado detection using radar almost impossible. This difficulty in forecasting such events results in short , if any, warning times for the public"

On top of the weak signatures, the elevation at which these signatures occur are rather close to the surface. The radar beam that sweeps over McMinnville hits at somewhere near 4,000 feet. That is too high to detect significant velocity circulations at the surface. 

All these factors challenge any meteorologist in the Northwest when it comes to tornadoes. Seeing the radar triggers described above in real-time make issuing watches and warnings much easier in other parts of the country than in our neck of the woods. So while these events are very rare and not significantly strong, the tornadoes we see still present a very real danger when the spawn. 

Understanding Tornadoes

It's something that native Oregonians don't fully understand--the anatomy and history of tornadoes.On average, Oregon sees one to two tornadoes every year. However, there is no comparing our twisters with those in the Midwest and South. A chief example, Moore, Oklahoma. The monster twister that rolled through the Oklahoma City suburb is just the latest in a storied history of the most active-weather state in the U.S.

It takes a perfect atmosphere for a tornado to form. Plenty of moisture in the air, instability, strong winds and ultimately, the trigger needs to be pulled to initiate storm formation. There is no better set up in the world than right here in the U.S. More specifically, east of the Rocky Mountains. The moisture roll player is the warm Gulf of Mexico. With mild water temperatures year around, the Gulf of Mexico provides plenty of "fuel" for storm development. All that warm air can travel freely over the land and deep into the Plains due to the lack of "elevated terrain". There is no physical barrier that would wring out moisture or block it from proceeding North or East or any direction until it reached the Rockies.

All that warm air then needs to be lifted. A typical lifting mechanism is simple day-time heating. As the sun breaks through the clouds, it warms the air and causes bubbles of lifting air. Those lifting bubbles then grab the moisture from the Gulf and sends it upwards. A cold front moving through the area will also do the trick. The boundary will help push the warmer, less dense air up and over the colder, more dense air.

So we now have moisture and lift. Next we need to create instability. Those bubbles of air won't continue to rise unless the air surrounding the bubbles is cooler than the bubble itself. That cold air comes compliments of Canada. The still-cool spring air sinks down from the north behind our cold front and interacts with those rising warm bubbles. This creates instability and allows storms to being to form. Those big, towering, white puffy clouds are a result.

Tornado Ingredients

Almost there. We have the storms brewing, probably just dropping heavy rain and hail along with some lightning and thunder. In order for a tornado to spin up, we require spinning air above the ground. That is provided by a westerly jet stream, or fast moving pocket of air way up in the atmosphere. That gets the air surrounding these big thunderstorms rotating at different speeds and at different directions. This is called "wind and speed shear". Once all of these ingredients are present, tornado watches and warnings must be considered.

Oregon lacks several of the ingredients previously listed. We lack a significant warm body of water to provide moisture. The Pacific Ocean along the U.S. west coast is a cold current body of water. Cool water helps stabilize the air passing over it. Thus, the air coming onshore in Oregon is more stable than the air coming in from the Gulf of Mexico in the South. Coastal and Cascade mountain ranges also pose an issue to the formation of tornadoes. The rugged terrain often disrupts organization of all the ingredients required. Lastly, we just don't have a good source of warm, moist air that can filter in to our area. The most common form of active weather we Pacific Northwesterns' get are summer time thunderstorms. A good southerly push of warm, more moist air from California will typically help spark those storms. I think most of us will take the grey, drippy days instead of a regular threat of severe weather.

The Moore, Oklahoma tornado was just a perfect recipe of weather. The tornado wound up being ranked at the top of the Enhanced Fujita (EF) tornado scale, which is a 5. The EF scale, a re-tooled scale of the original Fujita scale, remains based upon damage caused from the storm. National Weather Service workers will asses storm damage and then use that damage to estimate wind speeds from the twister. A common misconception is that tornadoes are ranked on wind speeds. Hurricanes are classified using their wind speeds. Back in 2007, the NWS implemented the EF rankings, slightly adjusting damage-to-estimated wind speed scale. You'll notice that the new scale does not include winds of 300+ estimated winds but the old Fujita did. It is widely suspected that tornado winds are the fastest winds on earth. The confusion between Fujita and Enhanced Fujita is a bit perplexing but it really shouldn't be. In my mind, they are almost interchangeable but I refer to them using "EF" just to follow along with the NWS.



I had mentioned above that the Moore tornado was officially classified as an EF-5. Since 1950, when the NWS implemented the Fujita ranking scale, 58 tornadoes received the top classification. Moore, Oklahoma made it 59. In fact, of those 59 F/EF-5's, 7 of them have ravaged the state of Oklahoma. That is more than any other state and compromises just over 8% of all F/EF-5 tornadoes in U.S. history. Needless to say, Oklahoma is "Tornado Alley".

Weather Mythbusters, 2nd Edition

Time to clear up some more weather myths out there. Today's theme will be (mostly) about lightning but we will kick off this edition with a snow myth. We are heading towards winter, after all!

Myth: It can get too cold for snow?
Truth: While this is not common, it can occur. The basis behind this fact is rooted in the properties of the atmosphere at certain temperatures. Consider: In order for clouds to form and precipitation (in our case, snow) to occur, air on a large scale must rise through the atmosphere. Warm air is less dense thus more likely to rise through the atmosphere. Colder air is more dense so it has a tendency to sink throughout the atmosphere. Sinking air promotes high pressure which stabilizes the atmosphere making it less likely for precipitation to form. So if air gets too cold there just isn't enough precipitation in the air nor is there the dynamic properties to produce snow! We're talking COLD though, temperatures around -40 degrees Celsius.

Myth: Lightning doesn't strike twice in the same place
Truth: This is false. Lighting strikes the same place multiple times, all the time. Lightning rods were designed to do this very thing. They attract lightning strikes in order to keep bolts away from surrounding buildings and utility stations. Nothing occurs that prevents lightning from striking the same place twice.

Myth: The temperature of lightning is hotter than the surface of the Sun.
Truth: This, believe it or not, is true. In fact, air around a lightning bolt can heat up to a temperature of 54,000 degrees! Lightning results from a static build-up between the atmosphere and the rain/ice crystals in the storm. The release of this build-up occurs in the form of lightning. The bolt lights up the sky and heats up the air surrounding the storm. Heating the air to such high temperatures sends shock-waves through the air and the expansion results in thunder.

Myth: There can be lightning without thunder.
Truth: Yes, this is true, but only based on how far away you are from the lightning strike. Thunder reflects, bounces, and gets absorbed by the earth's environment. If you are a sufficient distance away from the strike, you may see the flash but the sound wave of thunder may be effectively scattered by the time it reaches you.

Myth: Rubber tires make your car a safe place in a thunderstorm.
Truth: There is a partial truth to this. Yes, rubber makes it safe-er to be in your car during a thunderstorm but it is actually the frame of your car that makes the car a safe place to be. The metal cage that surrounds your car will help absorb the lightning bolt should one strike your car.

Lightning is a powerful thing,

Sandy Recap and Part II

The ugly aftermath of "Superstorm" Sandy continues in the Northeast. Hundreds of thousands still without power and even without a home. Unfortunately, those who are still having issues nearly a week after Sandy struck are going to have new problems to deal with. More on this in a minuet but first, here is a great graphic that puts Sandy into context.

I'm not the biggest fan of The Weather Channel...too mainstream weather for me. My biggest gripe is their forecasting ability for cities not named Atlanta, Boston, Chicago or New York. But I have nothing wrong with their research department. Looking into the past is something they do quite well! They published this graphic that compares Hurricane/Superstorm Sandy with Tropical Storm Irene (last year's most costly tropical system) and Hurricane Katrina from 2005. The numbers are impressive given that Sandy only topped out as a Category 1 storm and wasn't even tropical at landfall!

Just as in real estate: location, location, location really matters with weather, too. Sandy hit the most populated area of the United States, impacting millions of people and damaging massive amounts of infrastructure. The clean-up effort will take time and that effort will be hindered by a new storm; a Nor'easter is on the way!

A Nor'easter is a typical winter-time storm that forms along the Atlantic southern coast and speeds up along the coastline bringing wind, rain and snow to the Northeast. Nor'easter gets its name due to the direction of damaging winds, not where it strikes.

Here is our surface map that is valid this Wednesday afternoon. The red arrow indicates overall wind pattern (from the Northeast). The models are putting out wind speeds sustained around 20-40 miles per hour with gusts approaching hurricane-strength over open water and coastal areas. Heavy rains expected at the coasts and a dumping of wet snow (multiple feet) further inland. The blue line I highlighted indicates a rough rain-snow line. I say rough because several more factors play into it but in this situation you can expect snow to fall behind that line. Notice that same line extends all the way into the Portland area--but we will not be seeing any snow! This Nor'easter is not as strong as Sandy was but the impacts from it will be felt hard as clean-up efforts will likely take a hit and more flooding will be expected from the storm surge. It's been an ugly first half of fall for the Northeast.

Thoughts on Sandy

Doomsday scenarios have been playing out for a large portion of the eastern seaboard. There have even been comparisons to the 1991 Perfect Storm that struck New England. Sandy isn't a reprise of that Nor'easter. Sandy is her own storm, like each storm is. All models indicate that Sandy will still be a hurricane when she makes landfall sometime early next week. There is no "Frankenstorm" situation; not even sure what that means. Sandy will be Sandy. Let's break down Sandy and her impacts.

 Sandy has a few things going for her right now. Her current path will take her right along the coastline.

All models bring her right up the coast. She would parallel the coast if it wasn't for factor number 2, the approaching cold front.

 This cold front will aid in bringing Sandy inland. The front will essentially scoop up Sandy and swing her inland.

 There was suspicion as to what the storm will be like at landfall. Will it be a hurricane? Tropical storm? Nor'easter? The high pressure set up to the north of Sandy will help deepen the low pressure at the center of Sandy and the forecast reflects that strengthening. Look for Sandy to make landfall as a category one hurricane.

Now timing and location becomes the next challenge. Guidance is indicating a landfall sometime between Monday A.M. and Tuesday P.M. She may strike anywhere from Delaware to Long Island. The closer we get landfall, the better idea we will have. Either way, the entire east coast needs to take action NOW to avoid disaster LATER.

Potential impacts could cripple a large portion of the east coast as well. With the cold front sweeping through, significant snow accumulations could fall as cold air filters in behind the front. There is plenty of moisture with Sandy, so 6-8 inches of rain as Sandy skirts the Carolinas is not out of the question. Added water from the storm surge may cause devastating flooding for certain places. Storm surge is enhanced by storm winds.

This is a model's forecast for landfall and the impressive wind field associated with Sandy. This is from the European model which has been the most consistent in forecasting landfall. But what is impressive here is the winds. The closer the lines are together, the stronger the winds are. Look how far out the wind field extends! Tropical storm winds may be seen all the way out in Ohio! I like this situation to verify.

A lot is yet to play out. As a forecaster I am anticipating this storm. It could be historic but it will be a deadly one. 

Cracking the 7-Day

Six days away from wedding-palooza and we are getting more details on the forecast for Labor Day. Over the last few days, numerical data has flipped-flopped almost hourly with giving us showers or not. One thing that remained constant was temperature, which looks to be somewhere in the low to mid-70's. Tuesday morning's run gives us a high temperature of 76 degrees.

The general weather pattern for Monday evening (5 P.M.) shows a weak trough off the Oregon/California coast. A general westerly flow is evident here. That would promote average or slightly below average temperatures. Notice the wind barbs are generally flowing in a northwesterly pattern. That tells me some cooler, Canadian air is heading our direction in the upper atmosphere. Winds closer to the surface tell a slightly different story.

Our 850mb chart (5,000 feet) is giving us a northerly wind at 5 P.M. on Monday. This is good news because a northerly wind usually means dry air is moving over us. Air that has spent more time over land is drier and warmer than air coming off the ocean. We would normally avoid seeing morning fog with this type of setup, the same would be true if we had an easterly wind.

Below is MOS (model output statistics) and it is just another numerical forecast.

 KPDX   GFSX MOS GUIDANCE   8/28/2012  0000 UTC                       
 FHR  24| 36  48| 60  72| 84  96|108 120|132 144|156 168|180 192      
 TUE  28| WED 29| THU 30| FRI 31| SAT 01| SUN 02| MON 03| TUE 04 CLIMO
 X/N  76| 55  72| 54  75| 54  71| 51  72| 52  76| 53  76| 55  78 53 78
 TMP  75| 57  71| 55  73| 55  70| 53  72| 54  75| 54  75| 56  77      
 DPT  49| 50  49| 51  48| 50  47| 47  45| 48  45| 48  47| 50  50      
 CLD  PC| PC  PC| CL  CL| CL  CL| CL  CL| CL  CL| CL  CL| CL  CL      
 WND   7|  7   6|  5   8|  8  10| 10   6|  6   7|  7   8|  7   8      
 P12   9| 17  10|  2   9| 12  11|  7  12|  6   7|  5   9|  8  12 18 17
 P24    |     20|     12|     12|     12|      7|      9|     12    25
 Q12   0|  0   0|  0   0|  0   0|  0   0|  0   0|  0    |             
 Q24    |      0|      0|      0|      0|      0|       |             
 T12   1|  0   6|999   0|  2   1|  1   1|  1   2|  2   4|  2   1      
 T24    |  1    |999    |  2    |  1    |  1    |  4    |  4          

Highlighted is the pertinent information. MOS gives us a temperature (TMP) of 75, which corresponds with what I was thinking as well. CLD stands for Cloud and CL is not "cloudy" but "clear". It's tough to tell what a calculator thinks is "clear" or "partly cloudy" but generally there would be a good amount of sunshine expected. The numerical output is producing a very similar forecast to mine. At least partly cloudy and temperatures in the mid-70's. The days leading up to Labor Day are giving us dry, bright weather, so hopefully the trend holds on one extra day! The forecast has been improving!

10 Day Weather Outlook

Now just 10 days away from the nuptials and we continue to get a better look at the September 3rd forecast. We continue to get data from only one model, the GFS model. The GFS offers a look much further into the future than it's main counterpart NAM. However, GFS sacrifices a higher resolution in favor of time. The NAM only goes out roughly 4 days into the future but has a much higher resolution so broken down in more detail. Once the two models overlap, it is a great way to compare biases. So we continue to look at the GFS data. Here is what we are looking at 10 days out!

Let's peak at the numerical data first:

FCST Hour	Valid Time	Max Temp	Min Temp	Td	10m Wind mph	850mb Wind mph	Total Precip(")	Conv. Precip(")	500-1000 THKNS	500mb Height	850mb Temp °C	500mb Temp °C	MSLP mb	Total Cloud Cover
192	Sat 09/01 06Z	61 °	58 °	58 °	N 18	N 7	0.00	0.00	563	578	12 °	-11 °	1016	0 %
204	Sat 09/01 18Z	60 °	56 °	57 °	N 13	NNE 13	0.00	0.00	562	577	12 °	-11 °	1017	5 %
216	Sun 09/02 06Z	61 °	57 °	55 °	N 18	NNE 13	0.00	0.00	563	577	13 °	-12 °	1016	35 %
228	Sun 09/02 18Z	60 °	56 °	46 °	N 11	N 11	0.01	0.00	561	576	12 °	-13 °	1016	46 %
240	Mon 09/03 06Z	61 °	56 °	34 °	NNW 13	N 7	0.02	0.00	561	575	12 °	-13 °	1016	74 %
252	Mon 09/03 18Z	59 °	56 °	32 °	NW 7	NNW 4	0.02	0.00	559	574	11 °	-14 °	1017	76 %
264	Tue 09/04 06Z	60 °	57 °	48 °	NNW 13	NNW 9	0.02	0.00	557	573	10 °	-15 °	1018	81 %

Highlighted are the times for September 3rd. Our mid-level air temperatures are right around 10 degrees Celsius. This should translate to a surface temperature in the mid-70s. The big change from the last post shows that we have a precipitation chance through the day, calling for 0.02". September 3rd actually averages 0.03", so rain isn't unusual. What you can't see is that with each model run (GFS runs 4 times per day), the precipitation forecast changes. The waffling of the runs tells me that the GFS is trying to pick up something on the maps, so lets take a look.

The 500mb map offers a pretty good hint as to why we are seeing some precipitation in the forecast. Notice the large bump over the middle part of the country. A nice ridge of high pressure dominates a good chunk of the U.S. and that causes a kink in the weather pattern on both coasts. This setup is known as the "Omega Block" named after the Greek letter Omega. Generally, fair conditions dominate over the central U.S. and both coasts would see some potentially unsettled weather. This would explain the cooler than average temperatures as well as the hint of precipitation that numerical data has been showing at times. Coincidentally and totally unrelated, a tropical storm/hurricane is the big feature on the east coast. 

So, the forecast 10 days out goes something like this: I'd expect partly to mostly cloudy conditions with a slight chance for a shower. If I saw this setup on September 2nd, I would not put a rain icon on the forecast, but it is definitely worth a mention. With 10 days to go and the flip-flopping of the model output, there are still no concrete details. Stay tuned, only three more days until September 3rd hits the seven-day forecast!

A First Look at the Big Day Forecast

I am in the middle of a 13-day work binge. Pretty sure that's the longest stretch of consecutive work days for me at the station. It will all be balanced out when I take 9 days off starting Wednesday of next week. For those who may not be aware, I am getting married on September 3rd. So the time off will all be related to that.

There have been a few people asking if I am nervous for the ceremony. I'm not really nervous (at least not yet) for the vows. Rather, I am more nervous about the weather cooperating as we are having an outdoor ceremony. As always, best to have a "Plan B" which we do. That involves cramming about 130 guests into an outdoor tent and doing the ceremony/reception all in one place. Luckily, climatology says that the likelihood of a "rainout" is very slim. The average temperature in Portland on September 3rd is 79 degrees. Our wedding will be in Oregon City which typically runs a degree or two warmer than Portland. The last 5 years have seen an average of about 82 degrees with nothing but sunshine.

It appears our hottest stretch of weather this year just wrapped up. It's going to be tough to see more triple digits the later into summer we get. We've entered a stretch of below-average temperatures with a lot more cloud cover now as we begin the second half of August. I'm wondering how long it will last and will it impact the wedding. Portland hasn't seen any measurable precipitation in 31 days but that's not even half of our longest stretch, which was 66 days set in 1967. The GFS model run is finally showing the first glimpse of what we might expect for September 3rd.

  

Here is the 18z (11 A.M.) snapshot on the 3rd. This is the 500mb chart, giving us an idea where systems are and might be headed. The wind barbs show a strong "on-shore" flow, air will be coming off the ocean, thus giving us a cooler air mass.A westerly flow in late summer typically means that we start off cloudy as the marine air surges inland. This is all behind a system that comes onshore a day or two before the wedding that may produce some rain. Models that predict cloud cover do not go out this far, so using the map above I would say that we are looking at a somewhat cloudy start to the 3rd. Does that mean it will rain?

Above is the graphical precipitation output from the GFS model for the 3rd. It leaves us high and dry around 12 noon on Monday. Numerical numbers also back the maps.

FCST Hour	Valid Time	Max Temp	Min Temp	Td	10m Wind mph	850mb Wind mph	Total Precip(")	Conv. Precip(")	500-1000 THKNS	500mb Height	850mb Temp °C	500mb Temp °C	MSLP mb	Total Cloud Cover
192	Wed 08/29 06Z	61 °	58 °	58 °	NNW 4	SW 4	0.00	0.00	564	578	13 °	-12 °	1016	0 %
204	Wed 08/29 18Z	59 °	56 °	57 °	SSW 11	SSW 16	0.00	0.00	555	566	8 °	-16 °	1013	3 %
216	Thu 08/30 06Z	60 °	57 °	56 °	SSW 13	SW 20	0.01	0.00	552	564	6 °	-19 °	1014	10 %
228	Thu 08/30 18Z	58 °	57 °	42 °	W 7	W 11	0.06	0.04	551	566	5 °	-17 °	1018	44 %
240	Fri 08/31 06Z	58 °	57 °	33 °	NW 7	NW 9	0.01	0.01	553	571	5 °	-15 °	1022	15 %
252	Fri 08/31 18Z	59 °	57 °	34 °	N 7	N 11	0.01	0.00	555	576	7 °	-13 °	1024	69 %
264	Sat 09/01 06Z	59 °	56 °	47 °	N 18	NNE 9	0.00	0.00	559	579	9 °	-13 °	1023	15 %
276	Sat 09/01 18Z	58 °	55 °	54 °	N 18	NNE 11	0.00	0.00	560	578	11 °	-13 °	1021	2 %
288	Sun 09/02 06Z	60 °	56 °	57 °	NNW 20	N 7	0.00	0.00	561	577	12 °	-13 °	1019	0 %
300	Sun 09/02 18Z	59 °	55 °	56 °	N 16	NW 11	0.00	0.00	558	575	9 °	-14 °	1021	12 %
312	Mon 09/03 06Z	59 °	55 °	54 °	N 16	NNW 11	0.00	0.00	557	575	8 °	-15 °	1021	12 %
324	Mon 09/03 18Z	57 °	55 °	45 °	N 11	N 11	0.00	0.00	557	575	8 °	-13 °	1022	25 %
336	Tue 09/04 06Z	59 °	55 °	39 °	N 20	N 11	0.00	0.00	559	576	10 °	-13 °	1019	26 %
348	Tue 09/04 18Z	58 °	55 °	39 °	N 18	N 11	0.00	0.00	558	574	9 °	-14 °	1019	2 %
360	Wed 09/05 06Z	59 °	55 °	51 °	NNW 20	N 9	0.00	0.00	558	574	10 °	-14 °	1019	2 %
372	Wed 09/05 18Z	58 °	55 °	55 °	NNW 16	NNW 11	0.00	0.00	556	574	9 °	-15 °	1021	5 %
384	Thu 09/06 06Z	59 °	55 °	56 °	N 16	N 11	0.00	0.00	559	576	10 °	-14 °	1020	8 %

I have highlighted the precipitation (yellow) and mid-level air temperature (red) columns above. This model showing zero rain. That's good. It is also showing an 850mb air temperature of 8 degrees Celsius. We use the temperature at 850mb (that's about 5,000 ft) to help translate a surface temperature. On average, 850mb temperature of 8 degrees Celsius translates to a surface temperature of about 71 degrees Fahrenheit at the surface. That is a bit cool considering the average high for that day is some where around 79 degrees F. You may notice the temperature output for that day is 57 degrees F! This model is notoriously bad at temperature forecasting but has strengths in other areas.


These charts give us a good idea of what ALL model runs are doing with temperatures. Each model runs several different parameters and these charts compile all the 850mb temperature outputs and compares them against average (green line). The red line is the average of ALL those runs, and the blue line is the operational (published) model run.
 The first model is the ECMWF (European model). The red arrow indicates the 3rd. Not surprised to see such a wide variety of temperatures this many days out. However, the average of all the models looks to be pretty close to climatological average. One model goes as high as 22 degrees C, which would translate to 85-90 degrees F at the surface! The Euro also goes as low as 4 degrees C which would mean

 Let's look at the GFS, my personal favorite model. September 3rd is indicated by the black arrow this time. GFS is thinking that things will be a bit cooler, indicated by all model averages (red line). Still plenty of variation and even more time to go for things to change. But this chart kind of supports what we saw on the numerical output with cooler than average temperatures on the 3rd.

Mind you, we are still some 320+ hours away from the wedding and weather is ALWAYS changing. But if I had to make a forecast guess for the 3rd right now, I'm saying things will be in the mid to upper 70's with partly cloudy conditions by the afternoon. I'll be here to update the forecast as I see things change! Only 13 days away now!