OceanStWx

But it matches model runs over the last 24 hours, and that's what these probabilities are based off of. The reality is that we probably aren't able to narrow the range down to something smaller than 0-18" at this point.

And I tease Kevin about Kuchera. It can be really useful in situations that include mixed precip. But an all snow, especially cold snow, event is going to be . His forecast high Saturday is 18, that will probably be the max temp below 500 mb. That gives him a Kuchera ratio of 18:1. I wouldn't be banking on that.

They used to be, but satellite retrieval techniques and aircraft obs have significantly improved over time. There was a time when 12z might beat 18z consistently, but now that's just not the case. 18z is almost always more accurate than 12z. Again that's for closer range forecasts, as beyond days 3/4 all models are a bit of an unmanned firehose. Not 100% sure, but I believe so.

Good thing they are typically under-dispersive. Aside from the more hi-res stuff, all the models clearly ingested up on something today and moved as a result. Now recon starts tonight and satellite retrieval of the northern stream should be better by tomorrow, but if it's still meh at 12z I'm ready to just accept an advisory event to ruin a run at futility.

This is demonstrably false. Each new model run is on average more accurate than the last when inside 3 days. Now beyond that models in general just aren't THAT good at nailing down outcomes.

I just see no reason to melt when all we really don't have much more information than 12 hours ago or will 12 hours from now. Oh I never said they looked good! They just improved from 12z.

It's both true that it needs work and that it improved. The 18z MSLP is actually farther west and definitely stronger than 12z. It just also happens to be farther east than we want it.

Yes, equal weights in a sense. I'm no expert in modeling but there is probably some sort of neighborhood weighting to smooth things out a bit.

Great case for why ensembles are important. Despite the op GFS sending weenies swan diving off the Tobin, the ensemble actually improved in the mean, and likely brought more members in that direction vs a few outliers skewing things.

There is sensitivity to that area, but we want higher heights for a closer to the coast track. The problem is this sensitivity shows its hand late in the game vs. the northern stream shortwave.

Southern stream is nearly identical, but northern stream is deeper...however heights over our fannies are also lower and that's not exactly what we want. Otherwise I'd be saying slam dunk this has to be farther west.

I'm actually referencing the convection that develops around this southern max in convective QPF later in the run. That thing slingshots NE and drags the northern low (the main course) out east with it.

It's hard with the 84 hour NAM because there is nothing to compare it to, but that's why I'm more interested in what the upper levels are doing. The MSLP will chase convection from time to time, but if the fundamental upper level support is there I don't worry as much about QPF even if the QPF is bad on that particular run. The Euro for instance did go east with the 700 mb low, but not enough to cause me heartburn. I would imagine that is similarly true of the NAM, we just can't prove it.

That looked a little wonky though. Some convection seemed to race through to the east and sort of helped to Fujiwhara the main low east with it.

I'm not betting against it, that's for sure. For all the weenies I have one little kernel to tuck under your pillow for tonight as you dream of northwest trends: Given that the northern stream shortwave is what ensembles are most sensitive to, and given that this sensitivity really begins to blossom around 12z tomorrow, we need to be conscious of satellite retrieval. In the past (I honestly don't quite know how much improvement there has been) satellites have struggled to adequately capture the depth of Arctic region shortwaves. There are many reasons for it. Relative lack of moisture, extreme cold can bias retrievals, parallax can distort where the satellite thinks the retrieval is coming from, and just the fact that the viewing angle makes it difficult to sample the full troposphere. And what do we want from the northern stream to bring this farther northwest? A deeper northern stream shortwave. That's not to say it WILL happen, but that it could very easily be a stronger wave than models currently forecast.

All guidance systems have the spread concentrated west of the mean low position, with the dominant EOF pattern represented by a closer to the coast storm. There is also good agreement across guidance systems that the northern stream shortwave really starts to dominate the sensitivity as it nears Lake Winnipeg.

It's not unheard of for the GFS to be a little northern stream dominant.

Don't worry so much about the low center as long as you have the upper levels in your favor. January 2018 didn't track inside the benchmark either, but that worked out just fine because of how intense the forcing was.

Actually I may try and stack some up in front of my security cam so I can watch while I'm at work.

Yeah, definitely a touch east at 500. But it also looks more dynamic, probably why it is still able to throw precip far enough west.

If you stack 4 IPAs on top of each other and the snow is still over the top of the top can, you've got more than 2 feet.

It's better than 00z and 06z were, but my point is more that the changes were noise in the grand scheme of things at day 3.5-4. Nothing about that run confirms anything, but it also doesn't take anything off the table either.

dProg/dt of the upper levels is pretty close. It looks shifted a little shallower with the trough.

I think "bit" is the key, nothing major, overall pretty similar. No reason to drop the toaster yet.