Bubbler86

I still would like this area to trend colder on an ensemble mean....

I saw that and really his thoughts are my "backup" per se....winter is not over if we fail in early to mid Jan.

Gets showing similar strong signal to the eps for Jan 7th time frame.

That is the one I am watching as well....feel we are asking for too much beyond sloppy light stuff before then. Need to see the papa op move toward it. The PV suppression is one possibility. I personally am not a fan of an arctic outbreak both due to the cold and supression.

Just ticked up to 40 here for the first time since Monday.

It is way out there but the EPS info Bob pointed out about a heavy emphasis on precip in the Gulf coming into the second week of January is a good sign even this far out. Not impressed with it being cold enough for snow until we get past the first few days of January but somewhat excited days 5-15.

When you run, 20...30...40 models and combine them, you will always have hope this far out.

He still has the ensembles.

A mild 35 here this AM. May actually hit 40 today. Rgem and Nam runs that showed cold lows this Am did not do well. GFS back to upper 30's and 40's for highs to start the new year before colder air comes in day 5-6 of Jan. That cold push squashes any decent systems on this run.

Prayers be with you and your family.

I think he wanted a quarter in the slot like a peep show.

Team snow should be happy.

HH GFS is definitely an interesting solution. Temps are beyond borderline for the first part, but it is what we have for now. PAweather 540 line headed for Baja Mexico.

Cold air and lower heights retreating as we enter the middle of the first week of the new year on the 18Z GFS.

Definitely a better/broader push of cold air in early Jan on the 18Z GFS.

He would get kicked out of half his Broadway shows for using his phone to moderate during the Aria.

HH GFS is a'rollin. Lowered Christmas day highs by a degree or 2. Still a shorts day IMO.

Definitely an improvement over 6Z though it is more pushing out a bit....highs into the 40's or near 50 the first 3 days of Jan.

4th consecutive day of highs in the 30's and lows in the 20's and teens here. One of the more wintery temp weeks in 2 years. Not the most but top 5-10.

1128 Nooners at Liberty ski resort's west facing slopes. The snow is indeed sparse. 40 degrees and the front parking lot is about 25 percent full with golfers.

I was starting to buy in yesterday when the GFS Op was veering toward a very cold Eastern half of the US but disappointed to see the last few runs bleed it away. All of this is still 300-hour stuff, but it is what the output is. It does bother me that the GFS is using the new Fv3 core, but the GEFS is not.

You are digging in deeper than I did. Ha. I have thought about the possibility of having a limit on extreme outliers but then it brings into question 1) What if the outlier is right this time? 2) Where do we draw the line on what Outlier is? 20 degrees different than the next closest member? I think a good example of skewed distribution is the snow fall member maps I posted yesterday. I think most people understand this, but it is very clear example of the issue in using ensemble snow means to predict much...the snow means that matches these members was 3-4" through much of the LSV but only 6-7 of the 30 members below show that result. The rest show no snow or more in the range of 1-2" (18 in a quick count for 1-2") but the gigantic outputs of P26, P23, P03, and P14 not only cancelled out the ones showing nothing, they double what the majority of 18 show.

Agreed, if a lot of the runs agree then you have a higher chance of reality but that requires looking at the individual members to see if a small subset is skewing the totals vs. just the means. Let's take early January for example. Pretty much all the means AND members are suggesting it is going to be colder than the week before so a pretty safe forecast you would think but when expanded to look for potential low-pressure systems, I personally think it is the same or worse than using a single Op to do the same. Since the ensembles are purposely fed different data in an effort to find an average, I cannot buy using mean low-pressure locations or really the snow fall means as being any better than the higher resolution Op run. Measuring the uncertainty of the Op run is different than trying to argue that the lower resolution ensembles are better at forecasting that the Op. It also does not help that we now know the GEFS is using a different "software base" than the Op GFS. Different does not mean better or worse but is not ideal when trying to use them to find uncertainties in the Op.

I still think much of the issue is using ensembles to try and forecast ground truth. If you have a 10-model ensemble and 9 runs show the highs will be in the 50's and one run shows the highs will be near 0, the resulting ensemble mean is going to be grossly understated despite only 10% of the suites showing such a solution. It is similar to the difference between an average salary and a median salary. With average salary figures, the large wage earners skew the numbers to make the resulting figures look between than reality. Using the ensembles to "check up on the op's" is a better use, IMO....or to look for potential.