Forecasting Precipitation Type in Wintry Events

[Confuzzled]

So I went through the winter course ICW from NOAA and took notes.

http://www.wdtb.noaa.gov/courses/winterawoc/IC6/index.html

Top Down Method -- analyze soundings from the top down in determining p-type.

They outlined 4 ptype algoriithyms as well:

Bourgoin Method

Partial Thickness Method

Baldwin Method

Ramer Method

The next two sections were:

Using Ensembles in Winter Forecasting

Snow Ratios and Predicting Heavy Snow Events

There was another section as well pertaining to Awips software and Q vectors which went over my head at the moment. That is my next project. I am going to offer a brief overview of each of the bolded items above and maybe pose some questions to follow.

My personal goal was to get a better handle on predicting ptype and snow totals while using Bufkit during winter events.

[Ed Lizard]

My friend in a pinch, as a complete amateur, when I don't use partial thicknesses (540 dm 1000-500mb I remember, I think the PSU e-Wall also has a magic line for another thickness, maybe 700mb) or 850 mb temps, is the Northern Illinois University Sounding Machine.

http://weather.admin.../fcstsound.html

I downloaded this beauty earlier...

[Confuzzled]

TOP DOWN METHODOLOGY

Apply Cloud Microphysics to the highest saturated layer:

None of this was new to me having taken atmopsheric physics and thermodynamics but I suppose I never really looked at wet bulb temperatures in a sounding when making a forecast before so the most important thing here for me was learning about the operational use of wet bulb temperatures and predicting ptype.

• Ice crystals increase in a cloud as temperatures decrease (substatially when <-10c)
  
• Activation temperature of IN depends on the CCN present (e.g. clay ~-3c)
  
• Activation is about -8c in CT via ocean and abundance of sodium chloride
  

Check For Heterogeneous or homogeneous nucleation

Liquid, supercooled water can exist in clouds well below freezing temperatures, possibly all the way until -40C. It is important to analyze a cloud layer in an attempt to determine if ice nuclei are present to begin with.

Cloud temperature related to cloud content

• >-10c mostly supercooled water (liquid water below freezing)
  
• -10 to -20c mixed super cooled water droplets and ice crystals
  
• <-20C mostly ice nuclei
  

This is probabilistic. You can still receive snow in warmer clouds ( >-10C ). The majority of clouds colder than -10C will be dominated by mostly ice nuclei. We do receive snow storms from clouds in the -6c range however.

How to determine if ice is present:

• Observed soundings (skew t)
  
• Model Soundings (Bufkit)
  
• IR Cloud Top Temperatures via satellite
  

Using a skew T

Observe temperature of the saturated layers (essentially where T and Td lines combine or are very close)

If its at < -10c then ice nuclei are very probable in the cloud. > -10c and that probability diminishes.

Seeder-feeder method:

Clouds at a certain height may not be considered good for producing ice nuclei due to their temperature but they can be seeded by other clouds.

• One cloud layer (colder) where IN are likely seeds another (warmer) layer where IN are unlikely
  
• Because SVP > over Ice than SC Water droplets the IN interact and grow as if they were a regular part of the cloud.
  
• For the S-F method to work height between cloud decks should be no > 1500m (5000ft)
  
• You also want the transfer layer to be moist to avoid sublimation
  

WARM LAYER ANALYSIS:

Once you know ice is present you need to move on downward and analyze the rest of the profile. You want to check for warm layers and focus on their magnitude and depth.

If snow falls through a modest warm layer of even 3C then melting will probably not be sufficient enough to cause rain, assuming the surface layer is cold. Obviously it takes time for snow/ice to melt and whether that occurs or not depends on the thickness of the warm layer and the residence time of the ice snow. In general the following guidelines can be used for ice nuclei falling through small to modestly deep warm layers:

• <1c all snow (unless abnormally deep)
  
• 3C Ice Pellets (unless abnormally deep)
  
• 1c Snow/Ice Pellets (unless abnormally deep)
  
• >3c Complete melting (freezing rain/drizzle)
  

WARM LAYER WET BULB IMPORTANCE•

Wet Bulb Temperature (Tw) is the final temp an air parcel will attain after saturation is reached. As cold hydrometeors fall into surroundings that are dry/warmer they will evaporate or sublimate (take heat from the environment) leading to a cooling effect and increased water vapor in the layer (T drops, Td rises). Temperatures can drop 5-7C in one hour!!! The initial max temperature of the layer will slowly migrate towards the Tw of that layer. The greater the intensity of precipitation the more pronounced and quickly the effect.

This means that if snow (at 600 mb -- T -17.8c Tw -19.9c) is falling into a warm dry layer (750mb T 5.1c , Tw -0.7c) then some of it will melt/sublimate(cooling the air). Over time if the precipitation is continued the air will be cooled to the Tw (-0.7c) as the air increases towards saturations (snow). Thus you may have rain/ice quickly moving to snow. This app below illustrates this process quite well.

http://profhorn.meteor.wisc.edu/wxwise/precip/precip-wb.html

WARM AIR MELTING WET BULB FREEZING HEIGHTS GUIDE

• layer >1500ft snow rare
  
• Warm layer depth 750-1500 ft snow possible/likely
  
• <750ft snow typical
  

To use the above list you should incorporate considerations of the lapse rate:

• Lapse Rate is small melting occurs less rapidly (deeper layer needed 1500ft)
  
• Lapse Rate is large melting occurs more rapidly (shallower layer needed 750ft)
  
• Important to note LR in 750-1500 range.
  

PRECIPITATION INTENSITY:

Intense precipitation of ice nuclei into a warmer region associated with:

• Increased pseudo-adiabatic cooling (via enhanced vertical motion needed to induce it)
  
• Deeper/colder cloud tops
  
• Increase snowflake size (longer to melt)
  
• Increased cooling due to melting of more hydrometeors.
  

When rain will change to snow:

• Low level Warm Air Advection is weak
  
• Marginal warm layer depth
  
• Steady moderate to heavy rain for several hours
  
• If precipitation is underestimated model cooling will be as well
  

Analyzing the Surface layer:

• How long was region above freezing before arrival of cold air?
  
• Tw is > 1c for 300m or more above surface icr crystals and rain occurs
  
• If Tw ~1.5c = rain
  
• Surface Tw for snow is < 1-1.5C
  

Ice Enteringi.

Tw near 0C snow

Tw > 0 and Tw >1.5c rain

Mix Entering

i. Tw/T ~ 0c or less ice pellets

ii. Tw surface 1.5c or > rain likely

Supercooled water entering

i. Twcoldlayer >-10c |Twsurface < or = 0c FZRA FZDZ

ii. Twcoldlayer >-10c | Twsurface > 0c RA or DZ

iii. Twcoldlayer <-10c |Twsurface < 1.5C SN or IP

SUMMARY of TOP DOWN Forecast methodology:

1. Are Ice Nuclei present in the cloud.

2. Is there a warm layer?

3. Check For Wet Wet Bulb Effect

4. Check the Surface Layer

[Confuzzled]

My friend in a pinch, as a complete amateur, when I don't use partial thicknesses (540 dm 1000-500mb I remember, I think the PSU e-Wall also has a magic line for another thickness, maybe 700mb) or 850 mb temps, is the Northern Illinois University Sounding Machine.

http://weather.admin.../fcstsound.html

I downloaded this beauty earlier...

That sure is pretty

The 540dm is a good initial indicator. The other techniquesyou are referencing are probably the thickness between 1000mb and 850 and the thickness between 850 and 700mb. If you plot them against one another you can us a chart such as the following for your region you can get an indication of the ptype from that.

I am going to go fiddle with that sounding maching now

[Confuzzled]

Baldwin or NCEP Method

Applied to vertical profiles of the atmosphere that are produced by numerical modeling.

1) Find the highest saturated layer (Tdd 6c or less)

2) If T >-4c assumed to be ice crystals. If >-4c super cooled water (scw).

3) If it is scw then the surface layer is checked. T>0 = rain, T < 0 of Frozen Precip

4) Using -4c isotherm (sometimes -4C) multiply depth of layer (m) times AVERAGE Tw of layer

5) If area > less than 3000 mC then snow. If it is > 3000mC look at lower levels (step 6).

6) Use 0C isotherm in lowest 150mb to determine area of average Tw.

6a. IF < -3000mC AND sfc based Tw area above 0C OR sfc based Tw area below 0c < 3000mC this leads to ice pellets.

Weakness of Baldwin Method

• When there are deep isothermal layers near the surface (Tw between 0 and -4C) it may forecast RA or IP/FRRA when top
• down methods clearly delineate snow.
  
• Baldwin method tends to over forecast frozen precipitation
  
• It also ignores dry layers despite being more robust and using Tw (as opposed to T)
  

I suppose my biggest difficulty with implementing this method is 1) establishing a specific layer size and then 2) approximating the average Tw in it (usually a curvy line). If the line was straight or semi-straight it would be easier. But it just seems to me that there would be too much room for error in both of these, even if I put a sounding into photoshop and made the pixels equivalent to height and figured out an average Tw of the layer using sections. A one degree temperature error is significant as the height of the layer increases.

Does anyone actually use this or know how to operationally implement it? It was in the noaa course so someone must be using it.

[Confuzzled]

Ramer Method

Ice fraction determines the p-type.

Preliminary Checks

1) If surface Tw is >2c rain is forecast.

2) If surface Tw is less than or equal to 2C AND the rest of the profile is < -6.6C Snow

Those are easy enought to implement even for me

Checks if Preliminary Results are not met.

1) Identify precipitation generation level which must be at 400mb or below

2) PGL is the highest layer of sufficient depth with RH > 90%

3) If Tw is <-6.6C ice (ice fraction =1).

4) If hydro meteor starts as liquid it will will not freeze until a layer wth Tw <-6.6c is reached

5) If a hydrometeor starts as ice, it will not melt until it hits a layer with Tw > 0c

6)If ice fraction is >85% and some melting occurred = IP

7) If the ice fraction >4% and Twsurface <0C = FR

8) If ice fraction is 4% to 85% = mixed precip

Ramer method does not account for dry layers or wet bulb effects. So the top down method must be used as well.

I am still uncertain on how to measure the ice fraction of various layers to get the percentages but it seems that steps 1-5 are easy to implement and simply require looking at the Tw line in a sounding.

[Confuzzled]

PARTIAL THICKNESS TECHNIQUES

The science behind the PTT is simply that thickness (z) of any layer is proportional to the T of that layer. An important extension of this is determing where layers of waa and caa are taking place. If the thickness kines change at one height for your region but remain similar at another level then advection must be occuring in the layer where the changes are taking place.

1000-500mb thickness heights are often used to distinguish between rain and snow events. Most notably, the 540dm line is used. What this means is that the height or thickness of the atmosphere between 1000mb and 500mb is 5400meters. Scientifically we know thickness is proportional to temperature and statistically we know this line is a good indicator of rain snow

There are two other thickness levels commonly used and they were mentioned above:

1000-850mb thickness

850-700mb level

If you plot the thicknesses against one another you can use the chart below to forecast ptype

The next graph shows how to use these thickness charts to check for warm or cold air advection (or we can just look at the cool color shading on maps now)

yaxis is the 1000-850mb thickness

xaxis is 850-700 mb thicknbess.

In the first panel on the top right we see that in cases where the 850-750nb thickness remains constant but the 1000-850mb thickness increases we have low level warm air advection. If the 1000-850mb thickness decreases we have low level cold advection. With that in mind it is easy to interpret the rest of the 4 panels.

ANY THICKNESS or other ptype method should always be accompanied by the top down methodology described in detail above. While it is a great start, merely looking at the 540 line is not going to tell you a whole lot about the type of precip or quantity. Youu need to analyze the entire atmospheric profile from the top down to get a better understanding of the situation.

[Confuzzled]

Bourgouin method –Good for IP/FZ precipitation

The SCIENCE

Uses the 0c isotherm as a bench mark and looks for positive (above freezing) and negative areas (below freezing) using the T of the layers on a skew-T sounding. Compares depth and mean T of a layer with residence time of a hydrometeor (terminal velocity).

• PA of 2J/kg aloft required to induce phase change
  
• >2J/kg assumes complete melting
  
• If not met BM tests the lower levels.
  
• Positive Area above Negative area = IP/ZR
  
• NA surface of 75-100 J/kg  ice pellets
  
• PA surface > 13J/kg completely melts IP
  

WEAKNESSES OF THE BM

• Assumes heterogenius nucleation (ice crystals are present)
  
• Does not account for dry layers
  
• Uses T instead of Tw (does not account for wet bulb cooling)
  
• Assumes constant Terminal Velocity of Hydrometeors
  
• Check using top down method to ensure it.
  

IMPLEMENTATION OPERATIONALLY

This one is the easiest to implement for me because Bufkit comes with a button that all you need to do is click and it shows the Bourgouin results.

This might be also why some of you who use Bufkit might have seen the warning (No ice particles present). Necause the BM assumes ice crystals you need to use the TDM to ensure the cloud heights are sufficiently cold to contain ice nuclei.