Jump to content

Thunder Road

  • Content count

  • Joined

  • Last visited

Everything posted by Thunder Road

  1. January (30th?) Ends With A Norlun Trof

    I disappear for a few weeks and people are micro-analyzing the JMA...
  2. Hi folks, Slowly but surely our research group at SBU has been transitioning our WRFs over to a new website (not my work) and Python graphics (my work). Someday over the rainbow this will become the operational page without /test at the end of the URL, but for now: http://itpa.somas.stonybrook.edu/sbuwrf/test/ Let me know if (when) you encounter any issues! I'm slowly picking off plots one day at a time to switch to Python. Sorry I didn't share it two days ago!
  3. 1/3-1/4 Nor'easter OBS: Who will win ? Bus ?

    Best I could get was 4.4", but as I mentioned Saturday, no snow board so ehhhh I was gonna go up to Long Island this week and come back this weekend but for a number of reasons decided to stay here. Boy, did I misplay that. 15.8" at Islip
  4. January 4th 2018 Bomb Cyclone Discussion

    I wonder how effective the MSLP Low centers plot is when we're considering whether there will or won't be a double-barrel structure. It's conceivable to me that there could be some members with a secondary western SLP min that the plotting algorithm doesn't pick up because the primary min is still lower.
  5. January 4th 2018 Bomb Cyclone Discussion

    From the 06z GEFS, about 4 members give PHL and N&E >0.1" snow liquid equivalent, with much higher probs over New England: http://vortmax.somas.stonybrook.edu/WEB/models/GEFS/snowliq10prob/ (I'm still troubleshooting why the graphics are old until FHR 102, then abruptly become current...)
  6. January 4th 2018 Bomb Cyclone Discussion

    It's also interesting to note that the pattern in the CMC-GEPS is different: EOF1 is still one sign as in the GEFS (so just cyclone depth), but EOF2 is cyclone speed/along-track error. EOF3 is closeness to the coast, which explains much less of the variance: ~9% vs ~22% for the GEFS. Though I can't show it, the EPS are very similar to the CMC-GEPS. http://breezy.somas.stonybrook.edu/CSTAR/Ensemble_Sensitivity/EnSense_Main.html
  7. January 4th 2018 Bomb Cyclone Discussion

    Taking a look at our ensemble sensitivity analysis... From the 00z GEFS, EOF2 governs the closeness to the coast at D5: EOF2 is sensitive to - stop me if you've heard this one - the northern stream short wave diving in from the Canadian prairies. Watch the sensitivity (in this case, warm colors over western Canada) explode from FHR 48 to FHR 72: This suggests that we'll continue to have predictability issues until about the 00z cycle Monday night/Tuesday.
  8. 12/30 Fluff obs

    Keep these pics coming, guys! A+!
  9. 12/30 Fluff obs

    3.4" in Bensalem, Bucks County. MLC-003 but if anyone from PHI wants to put out an LSR, put it out as "public" since I didn't use a board. (I know, I know, shame on me.)
  10. Potential winter storm threats

    Wow, it's staggering how far we've come from the days of the NCEP MAG site with the alternating blue and white table!
  11. Potential winter storm threats

    What's the link for the box-and-whisker plot? Thanks!
  12. I think a quick rule of thumb that should be added to the "how to measure snow" infographic is that if you can clearly see the number on your ruler, you don't have that much snow.
  13. You're right, it *should* work that way, but the folks who don't use a board could also be doing any number of things wrong. And I swear some people seem to find the biggest drift and stick their ruler in slantwise or something. I don't know. Like the 8.0" in Ridley Park, for instance.
  14. Anyway, a mere 2.2" up here. A real bustola - we were under a Warning and OKX's most likely snowfall map was 6". Dry air killed us early on - it didn't start snowing until around 1530z, way later than most guidance had even as late as last night. After that it was temps - so much melted on contact and we even flipped to plain rain for a bit in the afternoon. Every hourly ob at KISP was 33.1 F. And, perhaps because of the delayed start, we're going to verify too dry overall. KISP is sitting at just 0.31" through 04z. They'll add a few hundredths to maybe a tenth since we had the last bands come through this past hour, but the model consensus yesterday was around 0.5-0.55" QPF. Ah well, it'll make a great multi-band research case!
  15. Could be elevation up that way, but it could also be, how do I put this politely, "measuring technique." If you used a snow board and cleared every 6 hours, trust yourself over them.
  16. "One of these things is not like the others..." (I didn't know Lee moved!)
  17. Yeah anyone assuming 10:1 is going to be disappointed. A lot of snow is falling with thicknesses up to 546 dm because of low-level temps up to around 34 F. I'd say it's best to assume 5:1 to be on the safe side even though we should verify sightly higher than that.
  18. I have a 700 mb RH/Fgen/Q-Vector Convergence plot on my site: http://vortmax.somas.stonybrook.edu/WEB/models/GFS/h7fgen/ (I was trying to go for one plot that combines moisture, QG lift, and non-QG lift. But it's for the whole CONUS. I'll whip something up this afternoon based on my banding work that will be the regional domain.) What we've found in my research group is that while strong Fgen superimposed with moisture and instability (usually CSI but also more often than people realize just plain old upright CAPE) is the cookbook for a big single band, the cookbook multi-bands is much more ambiguous. In fact, it seems that multi-bands thrive in weaker Fgen, and that even stability doesn't predict them well (using -MPV as a proxy for CSI, anyway).
  19. And, just in time, I have the same snow liquid equivalent plots for the CMC-GEPS and NAEFS (GEFS+GEPS) on my website! http://vortmax.somas.stonybrook.edu/WEB/models/CMCGEPS/ http://vortmax.somas.stonybrook.edu/WEB/models/NAEFS/ Two things to note: 1. I still have to figure out why images for FHR 24 don't plot; for now just click one forward once the browser stops loading. 2. Environment Canada's GEPS grib repository updates waaaaay later than when the GEFS come in. The 12z's won't come in until ~1730 EST, for example (the 18z GEFS start updating at 18z EST, by comparison). Enjoy!
  20. Juicy 60-70% chance of >0.5" all along I-95, and again remember it's a 0.5-degree res data set so naturally the coast gets over-smoothed.
  21. Ouch. It's at least been presented on at AMS conferences - already better than Kuckera! I actually have a Python script for Cobb snowfall coded up but the function it relies on for actually getting the SLR (rather than the omega weighting) isn't actually given. I sent Cobb and Waldstreicher an email about two weeks ago but haven't heard back yet. Do you have anything better? They just provide this figure of an empirically-derived relationship:
  22. Yeah theoretically Kuchera is a bit better but it's never been scientifically verified anywhere - it's just kind of a rule of thumb. The models have an algorithm to compute snowfall depth as a net sum of snowfall (explicitly from the microphysical parameterization scheme), melting, and compaction, using soil temps. So this is a case where Kuchera could run high, if 2 m temps are <= 32 F but the model skin temp is warmer, which (maybe?) doesn't go into the Kuchera calc depending on the vendor and there's more melting that Kuchera accounts for. I went to a workshop hosted by OKX and they had a presenter from ERH who was convincing everyone to go that route. Once FV3 gets off the ground it'll have a rime factor product that's produced by the MP scheme that will also help produce a more realistic SLR.
  23. Keep in mind the total snow depth change product - which ERH is encouraging people to use - is roughly 1/2 of what 10:1 produces: