WXBELLS SNOWFALL EURO MAPS... dont make much sense

[Amped]

Weather bell rgem maps were equally bad in the NW suburbs of New York.

Nice thread BTW, people usually blame models for snowfall busts not realizing that there are postprocessing algorithms involved in that can give you different results from the same model.

[DTWXRISK]

 so because  weenies  mis use them  they should not exist?    um No

 

The best thing that can be done with model output snowfall graphics is to destroy them.  Just eliminate them.  I really don't see the point to them except to be tossed around and used irresponsibility.  I suppose you can make an argument that it's just a tool for forecasters as perhaps a reference or something but IMO they just don't do any good.  For anyone who forecasts, there are numerous data sources and numerous outputs available that a forecaster can use to come up with a very sensible forecast using his/her's abilities and talents.  

[DTWXRISK]

 but at least  the rgem kept the  Snow OUT of  NYC...
 if the model forecast 8" in  rockland county  and they get  3 or 4" its not  awful 

 

Weather bell rgem maps were equally bad in the NW suburbs of New York.

Nice thread BTW, people usually blame models for snowfall busts not realizing that there are postprocessing algorithms involved in that can give you different results from the same model.

[DTWXRISK]

 JMHO  but  MDA   snow maps are  better  because if the way thy  snow the snow amounts... IN  Ranges  
When some folks  see  11.6" on their location on wxbell snow map  whether its  euro or  not  they are  Fooking  gone. Thats it  .. they are done  ....its heavy snow for them no matter  what . BUT   when you see a range    I think it  has  a different  impact

 

Just to throw my two cents in here.  I first signed up for weatherbell just before they really expanded their high res EURO products. On the old EURO snowfall accumulation product I think there was a dislaimer at the bottom that read something like snowfall accumulation at 10:1 ratio if surface temperatures at 32 or below in last 6 hours. Obviously with this you would get snow with 850's warmer than 0C as well as ice showing as snow accumulations. When they added all of the high res Euro products that disclaimer went away but I think they still use the same algorythm. Sounds like EarthSat uses the same as WxBell.

[UNCCmetgrad]

Not quite... back when I was working there and could compare the maps, WxBell had consistently higher totals on the snow maps compared to MDA/EarthSat.

 

 

 JMHO  but  MDA   snow maps are  better  because if the way thy  snow the snow amounts... IN  Ranges  

When some folks  see  11.6" on their location on wxbell snow map  whether its  euro or  not  they are  Fooking  gone. Thats it  .. they are done  ....its heavy snow for them no matter  what . BUT   when you see a range    I think it  has  a different  impact

 

 

Ok then I stand corrected. I have never used MDA myself but from what I have seen DT post they seem to put out great products.

 

I do remember from the February storm from last year many posted operational Euro and EPS maps of up to 14-16 inches of snow for the Western Carolinas. But since the Euro snowmaps from WxBell include ice the amounts were overblown espeially in the foothill and piedmont areas where half the precipitation fell as ice pellets and frz rain. In many cases this was not discussed leading people to believe an epic snowstorm was coming.

[Hoosier]

The Euro snow maps that apply a 10:1 ratio and convert all precip falling at <32 degrees to snow are only useful in 2 circumstances, 1) you know you're going to get all snow and 2) you know the ratio is going to be around 10:1.  I prefer maps that actually try to take ratios into account and don't mistake mixed precip for snow.  Fortunately, this problem has been widely discussed in the Lakes/OV subforum since last winter and I think most are aware of the flaws. 

[GaWx]

Ok then I stand corrected. I have never used MDA myself but from what I have seen DT post they seem to put out great products.

 

I do remember from the February storm from last year many posted operational Euro and EPS maps of up to 14-16 inches of snow for the Western Carolinas. But since the Euro snowmaps from WxBell include ice the amounts were overblown espeially in the foothill and piedmont areas where half the precipitation fell as ice pellets and frz rain. In many cases this was not discussed leading people to believe an epic snowstorm was coming.

 

 There may be confusion by some with regard to the accuracy or lack thereof of MDA/Earthsat (same company) Euro snow maps. Rest assured, they were highly inaccurate last winter (similar to WxBell) in ZR/IP situations as I've said due to raw ECMWF snow maps counting all precip. with it 32 or colder at the surface as accumulating SN. So, this ZR/IP problem is not at all mainly just a WxBell issue. It is also an MDA/Earthsat issue among others. Like I said, the problem is really with ECMWF, itself, as MDA/Earthsat GFS snow maps do not have this same problem at all.

 

 The good news is that I can prove it. I just looked in my files and noticed that I did actually save an MDA/Earthsat snow map from just before the big mid-Feb. SE storm. I had saved it because it had the most SN I had ever seen on a snow map for N GA and I knew it was WAY higher than was possible since 850's were mainly several degrees above 0C. For example, KATL got only 2" of SN/IP (about as expected) along with plenty of ZR vs. the very laughable over 12" shown here. Also, note that this one shows the ~14-16" you mentioned for the W Carolinas: (this is the 0Z 2/12/2014 Euro run)

[InstantWeatherMaps]

OK, I just had a look at Tropical Tidbits... easily the worst snowfall maps I've ever seen. The totals are horribly overinflated because they use the NWS Conversion Table (which ranges from 10:1 to 100:1 ratio depending on the 2m temperature, and doesn't take into account any of the rest of the atmosphere)... so even borderline snow is 10:1, and most snow is 20-30:1. A flat 10:1 would be far better and less computationally expensive (though still not as good as the Kuchera method).

EDIT Feb 2015: Tropical Tidbits now uses a fixed 10:1... ignore everything before this sentence. They still use a magic-32 method (under 0C sfc = snow), though, and combined with the fixed 10:1 ratio this results in majorly inflated "snowfall" totals in borderline situations.

[burgertime]

I feel like SV maps work much better this year. They had to have changed their algorithm. Last year it seemed that they mimicked the "if it was under 32 2m" rule like WxBell seems to do. This year I've noticed that the snow seems to follow more along with critical thickness. Could be wrong about that just sems like where the battle lines were drawn with this last storm. It could also just be a 6:1 ratio since temps in the SE were horrible. I wasn't paying much attention to the NE but even there on SV it seemed totals were pretty low for spots like Boston and NYC.  

[UNCCmetgrad]

 There may be confusion by some with regard to the accuracy or lack thereof of MDA/Earthsat (same company) Euro snow maps. Rest assured, they were highly inaccurate last winter (similar to WxBell) in ZR/IP situations as I've said due to raw ECMWF snow maps counting all precip. with it 32 or colder at the surface as accumulating SN. So, this ZR/IP problem is not at all mainly just a WxBell issue. It is also an MDA/Earthsat issue among others. Like I said, the problem is really with ECMWF, itself, as MDA/Earthsat GFS snow maps do not have this same problem at all.

 

 The good news is that I can prove it. I just looked in my files and noticed that I did actually save an MDA/Earthsat snow map from just before the big mid-Feb. SE storm. I had saved it because it had the most SN I had ever seen on a snow map for N GA and I knew it was WAY higher than was possible since 850's were mainly several degrees above 0C. For example, KATL got only 2" of SN/IP (about as expected) along with plenty of ZR vs. the very laughable over 12" shown here. Also, note that this one shows the ~14-16" you mentioned for the W Carolinas: (this is the 0Z 2/12/2014 Euro run)

I'm glad you found this example GAWx. I was looking for a pic I had showing the all 51 members of the EPS showing at least a foot  of snow for CLT for that storm but I guess I deleted it when I cleaned up my files. I'm surprised that the folks from ECMWF would count ice accumulations snow. I guess maybe they are trying to save computational as well as storage space.

[Ginx snewx]

 

I am  beginning to think the snow maps   ecmwf over at   WXBELL are  not very good.  I mean  they  have amazing   detail and are   works of  art   but last winter a lot folks  got burned by looking at their snow  maps . I  am  beginning  wonder of    maybe this is  not on the up and up...

 

For example     looks at the 0z   monday  EURO.  This map from   SV   CLEARLY   shows even though 850 temps  on the 0z  monday  ecmwf go above 0 at NYC   PVD and  BOS  ...

ECMWF_MSLPThickQPF_ne_f72.png

 

ECMWF_MSLPThickQPF_ne_f78.png

 

 

the Euro snow maps over at wxbell has BOST  getting 12 to 14" of snow and NYC PHILLY BWI DCA seeing 6" .   Given how the 0C 850 isotherm goes NORTH of NYC and BOS   I dont see how wxbell snow maps make any sense. 

 

Yet the  0Z EURO SNOW MAP....    from  EUROWX.com  admittedly at  0.5  resolution...   has MUCH LESS SNOW .... 1"   for BOS NYC  PHL  and use the  Evan  Kuchera  algorothim

 

hmmmmmm

 

0zeuro.png

 

final  accumulation

[Ginx snewx]

Euro WX performed much better than WX Bell but what I have noticed is Euro WX doesn't like to produce snow at surface above 32, hence the big miss in say NW CT

[Cory]

Looks like the ptype is being fixed with better physics in the next model upgrade (soon).

 

http://srnwp.met.hu/Annual_Meetings/2014/download/monday/ecmwf_leutbecherV2.pdf

 

[GaWx]

 The Euro snow maps continue to be embarassingly bad and, as I've said, I don't use WB. (This is not a WB problem as much as an ECMWF problem...the source). Today's 12Z has several inches of SN in parts of NC for which the 850's are clearly above 0C throughout the 12/20 storm. SMH.

[forkyfork]

people are surprised joe bastardi's company produces inflated snow maps?

[A-L-E-K]

people are surprised joe bastardi's company produces inflated snow maps?

 

 

yeah, i thought we got to the bottom of this last year

 

maps by weenies for weenies

[dtk]

 The Euro snow maps continue to be embarassingly bad and, as I've said, I don't use WB. (This is not a WB problem as much as an ECMWF problem...the source). Today's 12Z has several inches of SN in parts of NC for which the 850's are clearly above 0C throughout the 12/20 storm. SMH.

They (ECMWF) use a fairly sophisticated microphysics scheme with prognostic cloud fraction, cloud ice, cloud water, rain, and snow.  See Tech memo here:

http://old.ecmwf.int/publications/library/ecpublications/_pdf/tm/601-700/tm649.pdf

 

This is more advanced than what is currently used in the GFS:

http://www.emc.ncep.noaa.gov/GFS/doc.php#gridconpre

 

If there is any issue, it is likely not in the model itself but transformation of this information to "snow accumulations" or some post-processed fields/products.  I do not know exactly what is being distributed, what the fields are called, what is in the grib table to identify, etc.  

[heavy_wx]

They (ECMWF) use a fairly sophisticated microphysics scheme with prognostic cloud fraction, cloud ice, cloud water, rain, and snow.  See Tech memo here:

http://old.ecmwf.int/publications/library/ecpublications/_pdf/tm/601-700/tm649.pdf

 

This is more advanced than what is currently used in the GFS:

http://www.emc.ncep.noaa.gov/GFS/doc.php#gridconpre

 

If there is any issue, it is likely not in the model itself but transformation of this information to "snow accumulations" or some post-processed fields/products.  I do not know exactly what is being distributed, what the fields are called, what is in the grib table to identify, etc.  

 

That document is very enlightening. Even though the Euro is considered one of the most advanced global models, it still uses some very simplified assumptions (all ice crystals are considered spherical, monodispersed  and their concentration is parametrized by the Meyers et al. (1992) scheme) in dealing with microphysical processes.  It's not much of a surprise that errors in precipitation accumulation persist in many of the same regions (especially  in much of the MJO domain) with the updated scheme. I wonder if they're considering using a two-moment scheme in the future.

[Thunder Road]

Lance Bosart at Albany had a talk at AMS that referenced this.  Not WxBell, but just that the ECMWF snow algorithm is bad in general.  He showed how badly it performed for the December 2013 Toronto ice storm, and said that that event, as well as another in (I believe) Slovenia prompted them to implement new model physics for precip type, which should be operational soon.  I don't remember exactly how long he said, but I want to say it was about a month from now or so.

[Ender]

Last two winters I spent time comparing WxBell's Euro snowfall maps against another provider who offers "Experimental HiRes Euro" model output as a menu item. I suppose the images were there as a possible testing of a future product offering that never got off the ground. I have no idea as they don't actually advertise Euro output as a product. This alternate provider used to have a popular WX forum, and their hires Euro snowfall maps sure did seem more wright than WxBell's. Anyhow, here's what I've observed: 

 

WXBell's "accumulated snowfall" seems to count liquid precipitation as 10:1 accumulated snowfall when any one of the most liberal P-type tests is satisfied. For instance, T85 0C or less even though 850 - 1000 thickness is 1312 and 700-850 is 1560? No problem, you've got accumulating snow! Even worse, all thickness tests fail with H85, H7, and H95 all >0C while 925mb and SFC are between -1C and 0C? Enjoy your snow!  850mb 0C and SFC at 3.5C? Yup, that's plowable.

 

The odd thing: I've compared WxBell's "accumulated snowfall" maps against similarly named maps from the previously referenced "Experimental" provider's like named maps along with another provider's like named maps and the non WxBell provider's don't show the same snow accumulation totals as WXBell yet the non WxBell providers do snow similar totals as each other.

 

The even more odd thing: WxBell offers a "Snow Depth (6-hrly)" product which I assume is based on the Euro's SFC cover parameter. That output snows realistic snow accumulations; assuming you don't mind performing some color pallet running delta math as you view the images. The Snow Depth product also happens to correspond nicely to other provider's similarly named output images. 

 

There's something truly special about WxBell's Accumulated Snowfall maps. In a sense I guess I could look at it as the Euro is pulling for me? Perhaps the model represents 2000 teraflops of computational prayer for my snowfall hopes? I remember last Spring wondering to myself, 'How far into the season will we have to get before the EPS Meteograms from WxBell don't show some accumulated snow in the next 10-days? I wonder if it'll even snow an occasional dusting in June or July?'

[InstantWeatherMaps]

This alternate provider used to have a popular WX forum, and their hires Euro snowfall maps sure did seem more wright than WxBell's.

Heh nice one.

[InstantWeatherMaps]

Both WeatherBell (not just the Euro maps) and Tropical Tidbits suffer from counting IP/ZR as snow... and when you're assuming 10:1 instead of using a proper ratio algorithm like the Kuchera method (which actually results in a more realistic 17:1 for St. Louis with the upcoming system here, for instance), that really adds up. Here are some images for proof (my site's snowfall forecast, followed by the model-output ptype during the heaviest precip in Arkansas, then Tropical Tidbits and WeatherBell snowfall forecasts); look especially at Arkansas.

Also, Tropical Tidbits changed to a fixed 10:1, so their maps are no longer worse than WeatherBell's.

Lastly, I recently got a subscription to EuroWX... their snowfall maps are pretty good and probably the best available for the ECMWF. I see nothing to indicate a significant difference between their method and what one of my maps would produce if I had access (in fact, I'm pretty sure they use Kuchera).