Fog and Frost Tools Comments and Suggestion

Forums (Forecast Builder)

RSS (Opens New Window)

Charles Greif, modified 7 Years ago.

Fog and Frost Tools Comments and Suggestion

Youngling Posts: 27 Join Date: 9/10/15 Recent Posts

Hello,

Chuck already knows this, but we deal a lot with fog in eastern Kentucky. Would it be possible to incorporate a tool that uses cross over temperature (lowest previous afternoon dewpoint) as an input? This threshold would basically limit fog formation until your air temperature reaches your cross over temperature. Many times, you don't have fog form as soon as the RH is 100%, rather it shows up several hours later as the temperature eventually reaches that cross over temperature. This would add a bit more science into the tool. Just a suggestion.

Also, not a huge fan of the frost tool in forecast builder. It has no inputs, so I have no idea at what thresholds frost is being included. In a situation like tonight, where we are dealing with a lot of fog with perhaps a few sporadic instances of frost, its been giving me way too much frost and I'd rather limit the frost by using a temperature threshold. However, as of right now, that tool does not allow you to control the frost by temperature and uses some sort of default values. The PoT_CreateFrostGrid we have in our GFE (which I assume all offices have since it looks like one of the PoT tools), does allow the user to use inputs. Would it be possible to make this the default tool in the forecast builder? Right now I can easily run it by itself when I get to that step, but I think it would be a better default tool to allow more adjustments based on the situation. Just another suggestion and I have no background on how the current tool was selected, so if I'm off base, that's ok. I have the work around, so I can get by with that :-)

I know these tools haven't received a lot of feedback, so wanted to provide some.

Kevin Sullivan

WFO Jackson, KY

Andy Just, modified 7 Years ago.

RE: Fog and Frost Tools Comments and Suggestion

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

Kevin,

The crossover temperature idea for fog formation sounds intriguing. Perhaps a study between crossover temperature and visibility would be worthwhile, since the ideal approach would be to produce visibility and then let PotFog fall out.

About the frost tool, the one that is installed within ForecastBuilder (which will also go into the PoT_CreateFrostGrid) has much more science and inputs going in behind it. The tool came from a University of Kentucky study on frost formation. We realize it doesn't have any sliders for adjustment, and that's really because that has already been calculated into the tool. Nice thing is too with the way the tool is setup, assuming your T, Td, Wind and Sky grids are well collaborated with your neighbors, PotFrost will come out identical. Maybe the overarching situation that you describe suggests a need to write an algorithmic way of distinguishing environments favorable for fog vs frost. One way we could do that is bring visibility into the PotFrost tool. If Visibility < value, don't allow frost. Thoughts?

Kevin Sullivan, modified 7 Years ago.

RE: Fog and Frost Tools Comments and Suggestion

Youngling Posts: 3 Join Date: 10/24/15 Recent Posts

I like the idea of bringing visibility into the frost formation formula if models can accurately capture visibility (see below). One other thing to consider is once you hit 32, you are going to start forming frost with fog dissipating. So that would have to be considered as well. I think limiting frost based on cross over temperature is a good way to go. I think most, if not all of our frost that occurred the other night, when our cross over temperature was around 39, occurred when temperatures reached 34 or lower (those readings were very sparse). But we had lots of dense fog that morning. The fog definitely dominated with the higher cross over temperature. So in this case, the frost tool was overplaying the frost.

The idea of pulling in visibilities from the model blends does not work well at all in our forecast area (and all of our forecasters would agree). Right now, quite a bit of spread on how forecasters in our office handle the visibility grids as there is no easy way to produce these grids based on models or weather. The blend does not pick up on river valleys are are generally all over the place with visibilities. Just feel it will be awhile before models are fine enough on the resolution to resolve our narrow valleys. We are one of the foggiest places around as we experience dense fog in our valleys most mornings (unless its cloudy), so we do put a lot of effort into producing these grids. We are looking to work on finding a more consistent way to message fog events in our forecast area as there is quite a bit of spread on how we handle events between forecasters and events. There has been some interest in a local study to work on improving our consistency in messaging for these dense valley fog events. Perhaps the cross over temperature can be included in our local study and perhaps even an indicator for when we should go dense fog advisory, SPS, or nothing. I'll get with our SOO and see if we want to try and put something together. Just from my experience here, the cross over temperature works in most situations in our forecast area. Just using the fog tool as it is now, would allow for fog much too early and we usually have to go in there and delete the first several hours of fog grids after the tool runs.

I'll certainly keep this post up to date with any local project and any other feedback I come across.

Andy Just, modified 7 Years ago.

RE: Fog and Frost Tools Comments and Suggestion

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

Kevin,

Your CWA is a lot like my past one, ARX where we have a lot of narrow but "steep" valleys that love to fog in, especially between late Aug and September. The office has done a variety of fog studies (a good contact is Todd Rieck), particularly for La Crosse which sits in the widest part of the Mississippi River channel in the CWA. The procedure under the Edit Menu PoT_ValleyFog was developed specifically for this issue, because like you noted most of the models do not capture the valley fog. That procedure, which you can configure for your CWA, will adjust Visibility, CloudBasePrimary and Sky grids depending on the coverage of fog you select for the valleys. Do you, and your office overall, utilize this procedure much?

Thanks!

Joseph Moore, modified 7 Years ago.

RE: Fog and Frost Tools Comments and Suggestion

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

Regarding the PoTFrost calculation that Andy mentioned, here's the flowchart for equation that's being used by the tool. (Attached)

Or, if you know how to read Python, here's what FB uses - note the code below (from the latest trunk) may differ from what you have installed.

def createPotFrost(self, timeRange, varDict):
        self.deleteCmd(["PotFrost"], timeRange)
        Tinfo = self.getGridInfo("Fcst", "T", "SFC", timeRange)
        for info in Tinfo:
            tr = info.gridTime()
            T = self.getGrids("Fcst", "T", "SFC", tr)
            Td = self.getGrids("Fcst", "Td", "SFC", tr)
            Sky = self.getGrids("Fcst", "Sky", "SFC", tr)
            Wind = self.getGrids("Fcst", "Wind", "SFC", tr)

            DewpDp = T - Td
            mag = Wind[0] * 1.15

            ##### calculate Dewpoint Score #####
            dewscore = self.empty()

            dewscore[DewpDp <= 5] = 6
            dewscore[(DewpDp > 5) & (DewpDp <= 9)] = 3
            #no dewscore change for dewpoint depression over 9

            #no dewscore change for wind speed greater than 7
            mask = (mag > 3) & (mag <= 7)
            dewscore[mask] = (dewscore + 2)[mask]
            mask = (mag <= 3)
            dewscore[mask] = (dewscore + 4)[mask]

            mask = Sky <= 31
            dewscore[mask] = (dewscore + 2)[mask]
            mask = (Sky <= 69) & (Sky > 31)
            dewscore[mask] = (dewscore + 1)[mask]
            #no dewscore change for Sky greater than 69

            #cases where no dew would be observed - thus not allowing frost
            dewscore[DewpDp > 9] = 0
            dewscore[mag > 7] = 0
            dewscore[Sky > 69] = 0
            ######################################

            #### calculate Frost potential ####
            frostscore = self.empty()

            frostscore[T <= 34] = 3
            frostscore[(T > 34) & (T <= 37)] = 2
            # no change for frost score when temperature > 37

            mask = (Td <= 32) & (T < 38)
            frostscore[mask] = (frostscore + 3)[mask]
            mask = (Td <= 34) & (Td > 32) & (T < 38)
            frostscore[mask] = (frostscore + 2)[mask]
            # no change for frost score when Td > 34 and T < 38

            ######################################

            #calculate the frost coverage based off the two scores above
            PotFrost = self.empty()

            mask = (frostscore == 6) & (dewscore > 6)
            PotFrost[mask] = 100

            mask = (frostscore > 3) & (frostscore < 6) & (dewscore > 6)
            PotFrost[mask] = 45

            mask = (frostscore > 1) & (frostscore <= 3) & (dewscore > 6)
            PotFrost[mask] = 20

            #Prevents frost over marine areas
            if self._Office in self._MarineOffices:
                all_water_area = self.encodeEditArea("All_marine")
                PotFrost[all_water_area] = 0

            if any(PotFrost):
                self.createGrid("Fcst", "PotFrost", "SCALAR", PotFrost, tr)

Andy Just, modified 7 Years ago.

RE: Fog and Frost Tools Comments and Suggestion

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

All,

Just an FYI that this latest ForecastBuilder release adjusts the python code a bit. After diving and much thought process into the flow chart, and collaboration with multiple forecasters, the following adjustments have been made:

No frost if Dewpoint depression > 9 F (because no dew likely according to chart)
No frost if Wind speed >= 8 mph (because no dew likely according to chart)
No frost if Sky cover >= 70% (because no dew likely according to chart)

Joseph Moore, modified 7 Years ago.

RE: Fog and Frost Tools Comments and Suggestion

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

Andy,

While we're well beyond the frost season here with 10" of snow on the ground, I wanted to add that I really think there should be a mask for PoPs and that the sky cover mask should be much, much lower. I think from the values we use for sky cover grids, anything over 30% would not really qualify for our local office frost/freeze product definition as well as the national NPW directive (10-515) specifically mentions the need for clear skies for frost advisories; "on nights with good radiational cooling conditions (e.g., light winds and clear skies)." We had a few cases of frost and rain shower or frost with 65% cloud cover which just doesn't make much sense to me.