zdlawrence - 33andrain Jump to content


Master Wx Expert
  • Content Count

  • Joined

  • Last visited

Community Reputation

113 Excellent


About zdlawrence

  • Rank

Personal Information

  • Location
    New Mexico

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. zdlawrence


  2. That is the case for the most part, but there are cases of final warmings that had a significant dynamically-driven component that were not unlike SSWs, and these have been shown to have distinct tropospheric impacts. The very early final warming in March 2016 was an example of such a case. Here's one paper (open access) that I know of that talks about some FWs aiding forecast skill dependent on where the FW first occurs with altitude (the upper strat vs the middle strat): https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2011JD015914
  3. Final warmings occur every year (usually within the late March to early May timeframe) because of the return of solar radiation to high latitudes. They're radiatively driven events as opposed to dynamically wave-driven events such as SSWs. Basically, you can think of the final warming as the opposite of the formation of the polar vortex: the stratospheric polar vortex forms in autumn when solar radiation leaves the high latitudes, which sets up a large scale temperature gradient between the poles and mid-latitudes, leading to a westerly circulation; the stratospheric polar vortex breaks down in spring when solar radiation re-enters the high latitudes, which reduces/removes those large scale temperature gradients, leading to a weak easterly circulation.
  4. For what it's worth, this is the sort of thing I was talking about. Notice the difference a day makes in the forecast of the (potential) split: Yesterday's day 10 fcst Today's day 9: Obviously most are still showing a split, but the North American offspring is much smaller and shorter lived from today's forecasts.
  5. I'm a bit more optimistic for it now that the ensembles have shifted to be in line with the op forecasts, and there's pretty good agreement of getting split-like geometry inside the next 8-10 days. However, that's far enough away that there's still enough time for it to become more of a "failed" split (and we've already seen how the forecasts have been struggling to hone in on the geometry of this event), so I'm definitely still cautious.
  6. Thanks! As @snow_cohen mentioned, though, it is a bit esoteric. But coming from a very hands-on physics background rather than a met/atmos sci background made it difficult for me to understand topics I'd read about in books/papers such as wave-wave interactions, so when I first saw a visual example, it really helped things to click for me. This winter so far has been a great example, so I thought it might help others too. As an addendum to my post, this plot on my website highlights the dates I pointed out in the thread fairly well. I've annotated it to make it a little clearer: The first two light blue arrows show when (1) the Aleutian anticyclone gets into place, and (2) the first wave transience event that occurs right after. The two green arrows that follow highlight the times right before the next transience events. You'd think from looking at this plot alone that the last pulse that's occurring within the forecast period isn't such a big deal, but in reality it's because the anticyclone in the middle-to-upper stratosphere strengthens to such an extent and pushes the vortex far enough off the pole that wavenumber 1 at 60N becomes not as good of an approximation.
  7. 3000dam is a quick & dirty definition that works pretty well. FWIW, I think the GEPS is the most unrealistic set of ensembles because it has a general tendency to build up unrealistically large/strong Aleutian highs in the stratosphere. See, for example: Most of its ensemble members are building up heights exceeding 3150dam. Not saying it can't happen ... it's just ... unlikely
  8. Winds are a vector quantity, meaning they have a magnitude and a direction. Models usually output horizontal winds decomposed into their vector components, which consist of a zonal (east/west) component, and a meridional (north/south) component. Sometimes you'll see these referred to as U and V, for the zonal and meridional components, respectively. Similarly, the zonal direction implies east/west around latitude circles, so a "zonal mean" means the average of a quantity around individual latitude circles. When it comes to the winter stratosphere, you'll see a lot of plots mentioning/showing zonal mean zonal winds at 10 hPa (which is in the middle stratosphere) and 60N (which is in the high mid-latitudes); this is sometimes shortened to U1060 or U6010 to denote the zonal mean zonal wind around the 60N latitude circle at 10 hPa. This quantity is a proxy for the dominant circulation direction of the polar stratosphere, which during the winter is dominated by the stratospheric polar vortex. When the circulation is westerly (flowing from west to east), U1060 is positive. U1060 is also commonly used to define sudden stratospheric warmings (SSWs), because when SSWs occur, they weaken the stratospheric vortex and can sometimes cause it to break down almost completely. In these cases, U1060 gets close to 0, or dips below 0 into negative values, meaning the dominant circulation direction in the stratosphere is easterly (from east to west). So whenever you see U1060 values getting close to or below 0 between the months of roughly November through mid-March, they represent SSWs.
  9. I don't have access to the accuwx pro plots, but I maintain that they're either processing/plotting incorrectly or they're having some sort of issue with the data. There is definitely no zonal mean zonal wind reversal inside 10days at any latitude at any level in the stratosphere.
  10. Yeah, looks like someone's gonna have to ask Accuweather what happened:
  11. Not sure; I'm sorta just waiting to see the latest FU Berlin plots. If I had to guess right now (I could be wrong), I'd say those accuweather plots aren't showing zonal winds averaged around all longitudes (e.g., because of the split jet structure with two jet maxima at ~30 and 50N), so the reversal of the winds is coming from the impinging/strengthening Aleutian anticyclone. If the FU Berlin plots later tonight do show a big reversal, then either this is going to be a big ECMWF bust, or there will need to be some serious (and interesting!) research to understand why other models such as the GFS/GEOS-5 didn't latch onto the event sooner. Edit: I should actually qualify that the "interesting research" (if the currently mythical event verifies ) would also have to include the ECMWF, since the event would first be showing up in sub-fh240 times, which is what would be really unusual.
  12. Yeah, and it'd be weird for such a reversal to show up at times less than fh240 without ever being in the forecast on previous days. SSWs are highly nonlinear events, but their predictability is really quite good in the sub 10-day range.
  13. It's hard to tell without being able to see the full equator-to-pole view of the zonal means, but something about those plots doesn't look right. Those are weird zonal mean zonal wind structures.
  14. If you want a big vortex disturbance at some point, these are the signals you like to see beforehand. Specifically wave breaking resulting in high potential vorticity air being stripped off the vortex and mixed out into the surf zone: These are the signs of vortex preconditioning. For those that may not know what equivalent latitude is or how to interpret it: Equivalent Latitude (EqL) is an area-based coordinate that maps a given quantity into a function of the size of its contours. Since potential vorticity (PV) is a ~tracer-like quantity on isentropic surfaces that increases ~monotonically from pole-to-pole (or equator to pole), we can find the surface area of a given PV contour (or series of contours) and convert that area to an equivalent latitude, or latitude circle that would enclose the same surface area as that given PV contour. Hence, lower EqLs = large area enclosed, while higher EqLs = smaller area enclosed. We can then bin other quantities (such as temperature and windspeeds in the above plot) according to the PV based EqL to see how the quantities map to different-valued PV regions (for example, the blue region of relatively low temperatures in the bottom left panel is the core of the vortex consisting of high PV air, while the red region is outside the vortex in the surf zone). We can also take the derivative of quantities mapped to EqL with respect to EqL, which tells us how closely spaced the different contours of the quantity are; in other words, high gradients mean the contours are tightly spaced, while low gradients mean they're spaced widely apart. The highest gradients of PV with respect to EqL typically highlight the vortex edge region, which are also usually in the region of high windspeeds (however, the maximum windspeed does not have to correspond to the maximum PV gradient in the edge region).
  15. I fixed my post; the images went missing because the latest plots automatically go into a directory literally called "latest", but then get moved to a different directory as newer plots get pushed to the server each day ... now I'm realizing the flaw in that solution stratobserve is my first "operational" webdev project, so I didn't think about things like that. Will need to fix! Thank you for the warm welcome, @Snowy Hibbo and @Bring Back 1962-63! This is a good idea. Gloria Manney is my PhD advisor, but she's not big on social media type stuff. As for the others, I don't think it would be appropriate for me to directly ask them to join (because they're extremely busy with their own research and work priorities), but I can, for example, plug the forum on Twitter and etc.
  • Create New...