FAQ

This page (Frequently Asked Questions - FAQ) will be updated as common questions and problems/answers become apparent. There are several ways to get help. The Forecast Builder Forum is monitored by members of the CRGMAT. They will answer any questions posted in the forum. Additionally, you can email nws.forecastbuilder@noaa.gov to request assistance. The aforementioned CRGMAT members will provide assistance as soon as possible.

Can I add sleet when the MaxTwAloft grids are less than 3° C?
1. With the latest build you can now utilize the ProbRefreezeSleet Grid to add sleet into the forecast. Previously ProbRefreezeSleet values where MaxTwAloft < 3C were ignored in the precipitation type calculations. Any non-zero value will add in some sleet probabilities. This helps in convective situations and those you need to force more sleet vs rain/freezing rain.
How is QPF handled?
1. ForecastBuilder is designed to be a one time base population process setting up the starting point of the ‘foundation grids’. This includes the QPF. In order to accommodate EDT offices, and to allow for timely collaborated edits of those foundation grids, the SuperBlend QPF is the common starting point.
When we choose the time period and place to start, which "Place to Start" should we choose? (i.e. How do we know we're satisfied with the foundation grids when we haven't seen them, or is this selection tool used for the second time around after any necessary collaboration changes have been made)?
1. The foundation grids are loaded in based on your choices of which elements to populate from the Common Starting Point Database in step 2 and based on the time frame you chose in Step 1.
2. When you are done with any adjustments to the foundation grids - start at the next step - Top Down - if winter mixed wx is a concern, or step 4 for deriving new accumulation grids from your updated foundation grids.
3. If all is still good and you just want to add/adjust some fog or frost, or maybe change the thunder chances you should start at step 5 and finish through step 6.
If we do change the top-down or foundation grids, is there a way to just run steps 4-6 in the Forecast Builder?
1. Yes. The ForecastBuilder will normally be run from the default Place to Start as the Foundation grids. However, when other changes are necessary through further collaboration (e.g. Top-Down, QPF, etc.), then the ForecastBuilder can be started from the appropriate step (Top-Down Grids, Precip Type and Accumulation Grids, etc.)
With regard to the precipitation type section, what specific snow ratios or formula will be used?
1. Training is provided for diagnosing snow ratios and it is up to the forecasters through collaboration to make adjustments to the Common Starting Point SnowRatio grids from the SuperBlend.
2. The forecasters can adjust the SnowRatio grid via a different formula/method based upon climatology and snow crystal type should they determine that the SuperBlend SnowRatio grid is lacking.
3. The National SmartInit team conducted a Science Review on SnowRatio. Implementation of the Science Review came within National SmartInit release version 7.0, deployed across NWS field offices Oct - Nov 2017.
Why are mountain ofices using SnowLevel to determine precipitation type?
1. There are advantages of using SnowLevel to determine precipitation type in the intermountain West. This includes better representation of precipitation types where moisture levels vary considerably, and model output (e.g. Max TwAlof, ProbIcePresent) has a more difficult time account for the effects of terrain over short distances. The CR Mountain SOO community found a SnowLevel approach is best suited to meet the needs of precipitation type for most winter weather events. However, if freezing rain, sleet, and/or widespread precipitation is expected, then CR mountain offices can collaborate and still use the Top-Down method as appropriate.
2. SnowLevel (Height of the Tw=0.5°C) will determine precipitation types. SnowLevel will often need to be adjusted to account for local effects, model bias, and expected precipitation rates as SnowLevel is the main driver for both P-type and snow accumulation. Internal consistency checks are performed for temperatures too cold for rain and too warm for snow.
3. ForecastBuilder now linearly decreases snow accumulation to 800 ft below the snow level (allows for a smoother transition, especially in rain/snow scenarios), and to assume all rain when temperatures are > 45F.
SuperBlend and BC related grids normally verify better over the long range, but what about during pattern changes? Do you see collaboration occurring when pattern changes are occurring to come up with a more reasonable starting point for the foundation grids?
1. The forecaster remains in control of the grids through the entire process. If changes are necessary, then collaboration should be accomplished with the surrounding offices to ensure grid and forecast consistency. National centers and hopefully even the regional ROCs will help to determine when these adjustments are appropriate.
2. When confidence is high that changes need to be made, then please change the first guess SuperBlend grids in collaboration with your neighbors to make the forecast more accurate. We do see SuperBlend errors in excess of 10F, especially during the winter and in the extended periods. Changing the forecast a degree here and a knot there is probably a waste of your time. Whereas making adjustments to better forecast a heat wave, a cold snap, a significant precipitation event, fire weather situation, or a wind event is certainly worth your time.
Can additional and specific documentation for each tool be provided so the forecaster is able to understand what process is occurring in each step (e.g. PoTThunder, ProbIcePresent, Snow Ratio, etc.)?
1. Yes. Look to the training available on the ForecastBuilder VLAB page.
2. Present documentation:
  1. ForecastBuilder Steps and computation of precipitation types
  2. PoWT Methodology for tools and computation of parameters like ProbIcePresent
When my Wx grids are created using the full version of the ForecastBuilder it only makes Wx grids for my WFO’s ISC area, is that normal?
1. Yes. This is to allow it to run faster and make for a cleaner Wx grid. You only share and send that ISC area so outside of that they are unnecessary.
2. Additionally, there is potential for the Wx grid to become corrupt if the Wx is produced say for the entire grid domain
The fire weather community is highly dependent on the potential for dry lightning from thunderstorms. How can PoTThunder address dry thunderstorm coverage? Is this an option that could be added in the Update Wx Grids step to add a column for Dry T and add coverage radio buttons?
1. In Step #6, there is an option to select DryT under “Add Optional Attributes for T.” However, that is an all or nothing feature over the entire forecast area.
2. Through further efforts with SPC, it is hoped we can develop a PotDryThunder grid which would populate in Step #5 (Non-Precip Types), thereby calculating appropriate DryT probabilities.
When are we expected to use ForecastBuilder for creating Snow and Ice Accumulations?
1. All CR forecasters will be using the ForecastBuilder process to produce these grids.
I got fog (or frost/thunder/smoke) in my Wx grids, but did not choose to add fog (or frost/thunder/smoke) in Step 5.
1. If there are PotFog (or PotFrost, PotThunder, PotSmoke) grids in the database it will use them to build the Wx grids regardless of their age. In this manner, shift-to-shift consistency and forecast reasoning can be maintained without flip-flopping the resulting non-precipitating Wx grids.
2. For similar reasons, ForecastBuilder_CRON does not overwrite ProbIcePresent, Wx, or these non-precipitating PoWT grids (PotFog, PotFrost, PotThunder, or PotSmoke).
3. As ForecastBuilder pauses at step 5, you should also examine all of the existing Pot grids and make sure they reflect what you want in the forecast. Delete or modify those that you do not want if the effect these grids will have on the resulting Wx grids is no longer required.
Is there a way to have the ForecastBuilder produce stratiform and showery precipitation in the same grid?
1. Not at this time. ForecastBuilder is designed to keep the Wx grid as simple as possible but still convey the forecast.
WindGusts seem too high at night.
1. This is likely a result of a high background field in the URMA that gets into the bias corrected component of SuperBlend.
2. Work with the RTMA folks to get these issues fixed.
3. Send any screenshot examples to the aor-rtma list server.
4. Same applies to T/Td discrepancies.
Can ForecastBuilder run over an edit area or only over the whole GFE domain?
1. At this time, ForecastBuilder cannot run over an edit area, only the full GFE domain.
2. The ForecastBuilder developer is looking into methods for it to function correctly within an edit area.
What grids will determine if the Wx grids break up?
1. PoP grids, Probability of Type grids (Precipitating and Non-Precipitating), and Hazard grids
Why isn’t PotFreezingSpray showing up in Step 5 of ForecastBuilder?
1. PotFreezingSpray will not become an option in Step 5 unless hourly temperatures are < 40°F.
What are the various blend percentages? (FAQ on all models)
1. CONSMOS: All MOS type models
2. BCCONSMOS: All Bias Corrected (BC) MOS type models
3. CONSRaw: All Raw models
4. BCCONSRaw: All BC Raw models
5. CONSAll: All MOS type models and all Raw models
6. SuperBlend - a weighted model
  1. Central Region (CR) Mountain sites (all elements except QPF, PoPx, and SnowRatio)
    1. 20% CONSAll
    2. 45% BCCONSAll
    3. 15% WModel
    4. 10% WPCGuide
    5. 10% Previous
  2. All other Central Region sites (all elements except QPF, PoPx, and SnowRatio)
    1. 40% CONSAll
    2. 25% BCCONSAll
    3. 15% WModel
    4. 10% WPCGuide
    5. 10% Previous
  3. QPF, PoP, PoP6, and SnowRatio at all sites:
    1. 80% CONSAll
    2. 20% Previous
When we have identified a specific area of opportunity to improve the forecast, how do we ensure changes collaborated and made are carried through future iterations of the forecast? It appears that once an area of opportunity is identified, the same changes need to be applied to multiple forecast cycles since the previous forecast has only a 10 percent weight in SuperBlend.
1. If you find a target of opportunity (e.g. extreme precipitating event) that needs adjustment from the consensus output, to ensure that idea persists into later forecasts, make sure to note it both in your AFD and in shift briefings. One could even shift log it too. Hopefully the next suite of model forecasts catch on and then adjustment no longer is necessary. However, there may be instances an adjustment is needed for many forecast cycles (e.g. snowpack initialization impacting temperatures). These adjustments for multiple forecast cycles are good to pass on to model developers.
2. The recently added Edit Retention function could be of use in these situations.
3. So could the ModelCertainty procedure and the MakeSure tool.
"Why am I getting frost with XX PoPs or XX Sky Cover"?
1. The frost tool in ForecastBuilder comes out of a study by the University of Kentucky. It considers many aspects of the forecast to determine frost, but there are no elements that would ensure no frost. This is less than optimal for our purposes so with the latest tech order several changes have been made:
  1. No frost if Dewpoint depression > 9°F (because no dew likely)
  2. No frost if Wind Speed >= 8 mph (because no dew likely)
  3. No frost if Sky cover >= 70% (because no dew likely)
2. These thresholds will be monitored in the spring (2018) for possible refinement depending upon ground truth and forecaster feedback.
Why is NBM not showing up in BOIVerify?
1. The NBM should not be added to BOIVerify as it is hard to tie to a specific cycle automatically. It would have to be done manually and creates a lot of overhead and trouble with the software owing to multiple versions arriving every hour.
2. NationalBlend however is in BOIVerify and the one to keep track of for verification.
A late evening run of ForecastBuilder in the "tonight" period caused SuperBlend to be loaded too far into the future (past the current "tonight"). Why is this?
1. Some point in time had to be selected in the code of ForecastBuilder to start the new "tomorrow" - 05Z was selected. If running ForecastBuilder on the current first period after 05Z, e.g. for an ESTF update, ensure the "Current Overnight" option is selected, or just use a highlighted time range to start/stop ForecastBuilder.
Why do I sometimes get error messages and a FB crash when using other tools while FB is running?
1. The ForecastBuilder uses the ProcessVariableList utility to create the GUIs used in the ForecastBuilder. The issue you are encountering is that while a ForecastBuilder GUI is open, another smartTool or procedure was run that also used the ProcessVariableList utility for their GUI. The ForecastBuilderGUI information used in the utility is basically wiped out by the new tool and results in the error you see and ForecastBuilder crashing. There is not an issue with opening another tool with a GUI, just a tool that uses the ProcessVariableList utility to create the GUI. The LimitValues tool does use the ProcessVariableList utility and so does the Diurnal Procedure, but there is not much need to call the Diurnal procedure since that option is built into ForecastBuilder. So, the issue is due to the GUI in one of the smartTools/procedure called to edit the grids while the ForecastBuilder is running.
2. This should be fixed with the new GUI going into implementation this fall.
How can I create non-measurable drizzle with ForecastBuilder?
1. Non-measurable drizzle is not allowed via ForecastBuilder for multiple reasons:
  1. An experiment of this was done during one winter with the Probability of Weather Type methodology. The result was confusion and extra long GUIs, along with the addition of a bunch of code needed for formatters. The experiment found for simplicity to always have drizzle as a measurable element
  2. To the general public, drizzle is a precipitating element (and they sometimes consider it rain). Since it’s possible you can measure, and in instances where you might have put a 14 PoP, adding the 1 to making it 15 is not a large difference. Plus this helps in the event of a case where expectation was non-measurable and you end up measuring.
Why is the Canadian model included in SuperBlend?
1. This is needed for proper model dispersion. The NBM will include it too. If we notice situations where the model typically harms rather than helps the blend, we may adjust the ratio.
How can I view the HazardBuilder grids without running through the whole ForecastBuilder process?
1. Running Step 2 of ForecastBuilder will show this if you leave the options unchecked.
What does the GMAT recommend to initialize the Sky grids now that the old ESTF tools do not work?
1. RTMA and Obs database are about the best options. Perhaps the 00-hour forecast from GLAMP or CONSShort could work. Look for something to be available using GOES 16 sometime in 2019 AWIPS Build.
Why do I sometimes get errors when running step 2 of ForecastBuilder?
1. Be careful editing or saving during the checks phase of Step 2. Also, the ForecastBuilder uses a methodology that can be interfered with if you run other more involved tools while ForecastBuilder is still actively running or paused.
Are there some aviation-related QC checks/rules built into ForecastBuilder? Yes. The following checks are built-in:
1. Where PoP >= 55% and CloudBasePrimary > 90, set CloudBasePrimary to 90
2. For Blowing Snow values >= 60 and Visibility > 0.5, set Visibility to 0.5 (this is due to the Baggaley snow model now incorporated into ForecastBuilder)
3. For Blowing Snow values >= 80 and Visibility > 0.25, set Visibility to 0.25 (this is due to the Baggaley snow model now incorporated into ForecastBuilder)
4. Where values > 0 exist in PotFog or PotFreezingFog, where Visibility <= 0.5 and Sky < 90, assign 90 to Sky (there are thoughts to add a low value to CloudBasePrimary as well)
5. If PotHaze grid values > 0 exist, ensure Visibility is <= 5
6. Where PoP > 55 and Visibility > 6, Visibility is set to 6
7. If no Wx and Where PoP <= 55 and Visibility < 7, Visibility is set to 7
8. Ceiling creation: Where Sky >= 57, take the CloudBasePrimary and multiply by 100. Where Sky < 57, set Ceiling to -30000