OceanStWx

We can forecast anywhere in CONUS for a 24 hour period ending 12z. Snow ratios definitely skew high. My go-to was always: http://www.eas.slu.edu/CIPS/SLR/slrmap.htm, which uses coop SRG liquid and not automated data. Still issues but not as many with that type of measurement. This may actually have the opposite bias, where some coops estimate snow based off of liquid and skew things towards 10:1 (especially for small sample size WFOs). Our SLR product is going to be hit and miss right now. If the forecaster edits those grids (like I do) they’ll be representative of the forecast. If they don’t, it defaults to the model blend which does skew high in the terrain.

Sorry KHVR is going to claim sick of this shit over KOOP.

The NWS decided to host an internal snowfall contest this season, similar to severe weather contests in the past couple of years. A good way for forecasters to stay sharp on skills that they may not always flex at their local WFOs. For those super interested we use a binormal probability density function, which in more simple terms means we pick an expected value and then adjust the tails (10th/90th percentiles) of the range based on our own confidence. This is a version of how we produce those snow probability maps on the website. You're looking at a reliability diagram for >6 inches of snow for the NWS forecasters participating in said snowfall contest. What this tells us is how reliable a forecast is (no shit right?). A perfectly reliable forecast would follow the diagonal from the bottom left to the top right (i.e. if we forecast a 50% chance of something happening it should happen 50% of the time). Now we more or less do this through 30% chance. After 30% though we start to tail off to the right, this is an overforecast bias. You see this with a 90% forecast for >6" only verifying like 55% of the time. I can't say I'm shocked to see this either. Figuring out why would be interesting (bad model QPF?). This isn't that small a sample size either, 3 weeks in averaging between 20 and 50 forecasts a day has us approaching 1000 forecasts. It's also true across all snowfall amounts, EXCEPT 18" when we develop a bit of a dry bias.

I know you're trolling but... There's a reason why you yourself forecast 4-7 with lollis to 10" during events. You know there will be areas that get higher, but you don't know exactly where. Local flow, local elevation, etc will all effect snowfall amounts. But so will measurements. Cory was measuring (not clearing) frequently, and reported his highest snowfall depth which is totally within the guidelines. But if someone measured when the snow ended only, or an hour after, or an hour before, snowfall could be inches different. And mapping software just isn't equipped to handle that level of detail when you have to fill in areas of no obs compared to areas with 3 very different obs in the same town. We could create that map and put it on the website and you would complain it looks like swiss cheese with all the circles in it, but that's what happens when you include every outlying ob. There is zero issue with a map that says 8-12" over a large area but including lollis to 16".

That came up after the Joplin EF5, more specifically whether there should be a mandate for new construction to require a shelter. But people don't like governments (state or otherwise) telling them what to do, even if that is for their own safety. Regardless, we're advancing tornado science quite a bit, but I don't think we're advancing the messaging/preparedness quite as fast.

SPC SREF already honking about it a bit. The set up has looked pretty good for a couple days actually, it's only just now reaching the tail ranges of some of the mesoscale guidance. Looks to be focused a little farther west this time (like Lower Mississippi Valley) vs. Sunday's stuff.

I've actually seen some GIS mapping of areas designated mobile homes and the MRMS low level rotation track (proxy for tornado path). It went through a particularly bad area of the country for mobile home concentration. Kind of a worst case scenario honestly. Not only was it a violent tornado (more likely to be fatal), but it was in the Southeast (also more likely to be fatal), through a high density of mobile homes (also more likely to be fatal, and finally happened on a weekend (also more likely to be fatal). The last part is interesting because people are at home which is actually more deadly, because municipal structures are often better equipped to handle those types of winds. This weekend thing goes for constructed homes as well as mobile homes.

Most likely someone forgot to LSR it and that's why it didn't show up. They had 6.2" total, and 58.5" on the season.

Our internal storm report system logs it all. We can even log null events if we are cold calling spotters. It’s just that the maps look so bad when there are high and low bullseyes everywhere. And snow is so subjective to begin with that you end up with a lot of reports not jiving with each other.

I don't think there were many basements where those storms went through. And we might get to do it all over again Saturday into Sunday.

Actually Joe D rows a dinghy out into the harbor with a portable snow board and 8" gauge during every snow storm now. I really wouldn't waste too much time worrying about whether you're in the PNS or not, especially the final PNS. We use those to generate our observed snowfall maps. And outliers make the map look like shit, even if you're an outlier because of elevation. We remove reports all the time to smooth the map out. It's not an indictment of your measuring technique (except when it is *cough* Lunenburg *cough*).

The time listed is just the time of the LAST report, it doesn't say anything about the number of measurements. BOS being an LCD site means that the contract observer is measuring every 6 hours. Also ASOS liquid is considered the standard, with only rare circumstances requiring contract observer liquid being used for the official ob. So the fact that the ASOS measured 1.06" is likely just coincidence, because the treatment plant's 8" gauge would likely be different by a small number and their observer wouldn't be hitting refresh on the web to find out how much BOS measured to estimate snowfall.

6.8" brings me to 63.5" for the season. I honestly may be in the best spot for snowfall while still be in the city limits of PWM (running 5" ahead of the official city obs). The way the coastal front sets up I'm about as far away as one can get in town.

We need one more good one to make sure this is your most forgettable solidly above normal snowfall winter.

Official coop guidance now is one clear every 24 hours. But as long as you aren't clearing more frequently than every 6 hours, I'm fine with it. We only require that at LCD sites though.

Yeah that's just generated off the most recent reports (PNS), nothing more nothing less. As more reports come in that map will fill out some, and at the end of the event we do remove any obs that appear to be strong outliers.

Not bad when the snow didn't start at BOS until 10 pm. Hard to complain when you nearly average 2"/hr for 8 hours.

My guess is the HRRR timing was a little off and the goods just happened. Those cold cloud tops coincided with the Chelsea report of 3-4"/hr.

Man those cloud tops near BOS cooled quickly.

Yeah, that mix line isn't getting any farther NW now. Racing to the canal.

It's a bit of a proxy, but you can see how averaging f-gen over more depth doesn't really change the location of the most intense frontal creation. So the lift is a lot more "upright" than a typically sloping ascent of WAA.

Similar process to February 2013 over LI and CT I suspect.

You're really close to the mixing, so there's probably a ton of liquid water in there to rime onto the flakes. Nice to see a CCB again.

I mean based on the site info at NCEI, it's measured at the treatment plant. Technically, that is on the beach in Winthrop though.