Confining FB output to an edit area

Forums (Forecast Builder)

RSS (Opens New Window)

Kathleen Torgerson, modified 7 Years ago.

Confining FB output to an edit area

Youngling Posts: 6 Join Date: 3/20/15 Recent Posts

Had a complex weather scenario over southcentral and southeast Colorado in the Day 2 grids tonight where I wanted FZDZ/DZ across the upslope regions of the southeast mountains and adjacent plains, and Rain/Snow across the mountains and valleys. I was using the Snowlevel and ProbIcePresent technique (as directed by our SOO) instead of the Top-Down method. So, I selected an edit area across my plains and ran FB over a 12 hour time frame setting the "Is liquid Rain or Drizzle?" to "Drizzle". It gave me PotDrizzle and PotFreezing Drizzle grids but I noticed it caught pixels of drizzle outside of my edit area across the mountains. I ran through the rest of FB to completion. I then selected another edit area for my western mountains and valleys this time, re-ran step 4 over the same time range, this time choosing Rain for the "Is the liquid Rain or Drizzle". (I should mention, I selected "Stratiform" both times). This time it deleted my PotDrizzle and PotFreezingDrizzle grids and gave me only PotRain and PotSnow grids. So it seems that FB doesn't confine its output to edit areas. Is this behavior correct? This wasn't what I was expecting, but I haven't used FB on a scenario as complex as this one until tonight.

The only way I could figure out to do a mix of Drizzle/Freezing Drizzle over one part of my forecast area, and Rain/Snow over another, was to choose "Both" for the question "Is the liquid Rain or Drizzle". But when I did this, I noticed that the "QPF value separater" for Rain and Drizzle didn't appear to work as I had expected. My PotDrizzle and PotRain grids looked the same even though some areas had greater than the threshold .05 qpf, and other areas had less. I guess I was expected that threshold to return drizzle where less than the threshold (0.05) and rain where greater than this threshold. Anyway, after creating a mess of PotRain, PotFreezingRain, PotDrizzle, PotFreezingDrizzle, PotRain and PotSnow grids, I went in and manually deleted areas of drizzle out of the mountains. If I would have had sufficient time, I'd have deleted the Rain where my 6 hour qpf was less than .05 across my plains, etc...and kept just the drizzle there, but this turned into a crazy amount of editing in complex terrain and I didn't have time to sort all of this out before the package was due. So I just kept the complicated mix of L-R-S- across the plains.

So my question is: how is one supposed to use FB in a scenario like this? And is it working as designed or did I stumble across 2 bugs tonight?

Thanks for any guidance,

Kathy Torgerson

NWS Pueblo, CO

Joseph Moore, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 72 Join Date: 11/1/13 Recent Posts

Hi Kathleen!

In the past, I think I recall the individual ForecastBuilder tools (the PoWT tools like PoT_CreatePrecipTypes and PoT_MergeWx) respected active edit areas, as I used to use this feature in the pre-FB era to make adjustments like you describe. However, I have noticed that recently the most common individual tools do not seem to respect edit areas and just edit all areas. Some individual PoT element tools do, like PoT_Valley_Fog, but others do not.

I've sometimes done the kind of work-arounds you describe for these and other challenges, like momentarily multiplying my QPF by 10 to force heavy rain (R+) wording into certain areas or timesteps. However, the problem with work-arounds like this is that they're usually so complicated it makes grids harder to update, and all that work gets blown away by the next shift. In the short-term, having the individual PoT tools (or even the whole FB process if possible) respect edit areas would be good and make it easier to handle complex scenarios, but in the long-term we should submit feedback when we run into these kind of challenges so that the developers can create new tools or adjust current ones to handle these less common scenarios. (I think this forum is a good place to submit such suggestions - when someone from CR-GMAT weighs in they should be able to share the best way to do this.)

Finally, a devil's advocate argument: was DZ/FZDZ versus R-/ZR- really truly a meaningful change in the forecast/impacts? In my opinion I would argue absolutely yes it is a meaningful change and worthwhile to make (freezing rain mention is the winter-equivalent of including tornadoes in the wx grid in my view), but I could see how some folks might have an opinion that this wasn't a meaningful change.

David Barjenbruch, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 16 Join Date: 9/2/14 Recent Posts

Hi Kathy,

We were working the same weather system in Boulder so had the same experience here. Joe is correct the FB PoTWx grids for rain/snow/drizzle/etc. do not work on separate edit areas in ForecastBuilder at this time. This is a complex system, and here we settled with just using the SnowLevel and ProbIcePresent grids to work for us over the entire forecast area, since anywhere there was sufficient moisture in our forecast area it would be cold enough to snow. Your example does bring up a good point of how to use in a complex rain/snow/drizzle scenario. The QPF cutoff for the rain/drizzle option is only a rough guess and may not handle these types of events. There is a current known bug in the QPF deliminator as you discussed (results in PotRain and PotDrizzle are the same no matter what QPF is chosen). Additionally, drizzle in these types of events can produce greater QPF than showers or other areas of light stratiform/showery precipitation. An edit area option would have much greater value to the forecaster in this case, and team will investigate this enhancement. Thanks for your feedback!

-Dave B.

Kathleen Torgerson, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 6 Join Date: 3/20/15 Recent Posts

Thanks for your replies. I do think having the ability to run FB in edit areas is a necessary enhancement for complex terrain sites. It was a cumbersome process last night trying to edit out the drizzle/freezing drizzle out of the mountain ranges where the upslope was not the primary driver of the process. I think I know enough now to be able to work around the tool's current limitations, but its not necessarily intuitive. Being able to run FB in edit areas would make the task of producing Pot grids less cumbersome for these complex weather scenarios. Thanks again. :)

Andy Just, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 89 Join Date: 6/2/15 Recent Posts

All,

Unfortunately I was unaware there was a bug with the QPF delimiter for Rain vs Drizzle. I did do a test right now here at CRH on EAX's domain and did not see the bug. Where QPF >= 0.05, my PotRain = 100 and my PotDrizzle was 0. And vice versa, where QPF < 0.05, PotRain = 0 and PotDrizzle = 100. If possible, can someone re-run this for me that is experiencing the problem and either post the results (QPF, PotRain and PotDrizzle) here or via e-mail to nws.forecastbuilder@noaa.gov? Thanks!!!

The work around I would believe for the edit area is to use them in ProbIcePresent. I.e. create the grid with a value of 100, then using edit area(s) set to 0 where rain/drizzle would be much more likely than snow.

Joe brings up a good point about rain vs drizzle and arguments for each. For simplicity I would love to get rid of drizzle/freezing drizzle entirely and lump that into rain/freezing rain. The QPF and IceAccum grids tell the story of the amount/impact. However, Joe's statement is valid too that the connotation of freezing drizzle is a lot different than freezing rain. On the flip side, I could argue freezing drizzle can wreak havoc for traveling too and catch people off guard. Reference December 16-17, 2016 when 10 people died on Missouri roads when freezing drizzle had started. Perhaps it is location centric?

David Barjenbruch, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 16 Join Date: 9/2/14 Recent Posts

Andy,

Here's a couple images with QPF, PotRain, and PotDrizzle. PotRain/Drizzle are exactly the same. The "0" is where SnowLevel kicked in. Wonder if the bug is only for mountain sites using SnowLevel?

Dave

Kathleen Torgerson, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 6 Join Date: 3/20/15 Recent Posts

Any word on when the bug with the 0.05 deliminator between drizzle (freezing drizzle) and rain (freezing rain) will be fixed for their respective Pot grids? I understand the logic isn't the best but its really the only proxy to give the forecaster control over the weather grid. Hate to ask so quickly but we are dealing with another shallow upslope event and now a new set of forecasters are feeling the pain. Any status you could pass along would be great.

Thanks,

Kathy

Andy Just, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 89 Join Date: 6/2/15 Recent Posts

Kathleen,

After seeing Dave's post that it wasn't working at BOU, and confirming that it does work at GLD and here, it does appear to be something related to SnowLevel. I'll dig in and see what I can find.

Andy Just, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 89 Join Date: 6/2/15 Recent Posts

Kathy and Dave,

Spot on. The application of SnowLevel occurs AFTER the separation of QPF between rain or drizzle.

Now I need to figure out how to recode it to make that option work. I'll try and get that done today and uploaded for testing.

Andy Just, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 89 Join Date: 6/2/15 Recent Posts

All,

I have sent out testing requests to BOU and PUB. Once confirmed it works, I'll allow everyone to install.

David Barjenbruch, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 16 Join Date: 9/2/14 Recent Posts

When we moved from PoWT version 4.52 to 4.54 everything ran as designed with your new code. Rain for QPF ≥ 0.##, and drizzle for < 0.##, where ## is the forecaster selected amount.

Thanks for your help on diagnosing and fixing this, Andy!

-Dave

Kathleen Torgerson, modified 7 Years ago.

RE: Confining FB output to an edit area

Youngling Posts: 6 Join Date: 3/20/15 Recent Posts

I do think there is an important distinction between FZDZ and FZRA that goes beyond just simply impact. FZDZ is more commonly associated with smaller drop sizes, slower accumulations, and often accompanied by visibility restrictions. Not to mention it is much more common in the west with shallow upslope in the winter. It was important enough for METARs to distinguish between the two, and I suspect that when we start deriving aviation products from the grids, we will need to use FZDZ when it is called for. The grids have to serve multiple purposes, not just public forecast generation. Just my 2 cents...if it helps any. Thanks for your reply. :)