Event Summary
     National Weather Service, Raleigh NC


December 18-19, 2009 Winter Storm
Updated 2010/06/15






Event Headlines

...This winter storm produced a large area of heavy snow just north and west of the NWS Raleigh forecast area with 12–24 inches of snow reported across much of the Mid Atlantic and Northeast including many of the major cities along the East Coast. Record breaking snow was observed in several locations including Washington D.C...
...Winter Storm Watches verified with an average lead time of 39 hours. Winter Storm Warnings verified with an average lead time of around 22 hours with no missed events. Nine of 12 counties in the Winter Storm Warning verified with the 2 of the 3 counties that did not verify (Randolph, Chatham, and Franklin) receiving Winter Weather Advisory conditions...
...The event followed the "Miller A" surface pattern conceptual model which is characterized by a relatively simple mean sea level pressure pattern where one relatively well organized surface low tracks northeastward with a surface high pressure center located to the north. The "Miller A" pattern typically features a well established narrow transition zone of mixed precipitation accompanied by little or no icing...
...High resolution model forecasts from the NCEP HighResWindow WRF and the local NWS Raleigh WRF provided very good guidance during the event and highlighted some of the mesoscale features that would impact the precipitation distribution...
...The CIPS Winter Storm Analog Guidance data was utilized by several forecasters during the event...
...Forecasters utilized AMDAR aircraft soundings to analyze thermal profiles at GSO and RDU in between RAOB times. This data set is an important resource during critical winter weather forecasts.


Event Overview

A cold area of high pressure with a large area of subzero temperatures moved south into the Central Plains on Tuesday, December 15, 2009. A cold front marking the leading edge of the cold air mass reached the Southeast U.S. coast and the Gulf of Mexico late on 12/15. The high pressure system shifted east on 12/16 while the cold front became stationary in Gulf of Mexico during the evening of 12/16.

The
high pressure center with an intensity of 1032 MB was centered over the Great Lakes region during the morning of Thursday, 12/17. Weak cyclogenesis was enhanced over the northwest Gulf of Mexico as an upper level trough approached the Gulf of Mexico from the west. The high pressure center to the north weakened as it moved east as it extended south into the Mid Atlantic states on Thursday evening.

CONUS radar reflectivity loop from 2358 UTC on 12/16 through 2058 UTC on 12/20 By Friday morning, 12/18, the surface low was intensifying as low level warm advection increased and the upper wave intensified. The surface low intensified further during the day and by 21 UTC on 12/18, the low was located over southern Georgia with a coastal front extending northeast along the Carolina coast.

The cold high pressure center was initially of sufficient strength (> 1025 MB) and location (Great Lakes region) to deliver cold dry air into central North Carolina in advance of the storm. By the time precipitation began to overspread central North Carolina during the afternoon and evening of 12/18, the cold dry air was fleeting. The lack of classical cold air damming with a strong, cold dry high pressure system to the north, left diabatic processes, namely the melting of snow aloft as a key mechanism to provide enough cold air for wintry precipitation.

An initial band of precipitation moved north into North Carolina from South Carolina just after 12 UTC on 12/18. This narrow band advanced north, reaching Raleigh and Greensboro just after 17 UTC. The precipitation at Raleigh and Greensboro fell as some light snow. Further east the precipitation fell as a mix of snow and sleet with rain observed toward the coast. This area of precipitation appeared to be associated with an axis of enhanced frontogenesis at mid levels as noted in the SPC analysis at 16 UTC 18 UTC, and 20 UTC.

During the afternoon and early evening, steady precipitation developed across western North Carolina including the Western Piedmont and Triad area. Further east, after a 30-60 minute period of light precipitation associated with the initial band, a lull in the precipitation persisted for several hours across central and eastern North Carolina. The precipitation across western North Carolina was driven by low and mid level frontogenesis, low level moisture transport, and divergence aloft associated with the right entrance region of the polar jet over the Northeast and to a lesser extent the left exit region of the subtropical jet. By late afternoon, a few inches of snow had accumulated in the Triad with lesser totals eastward toward the Triangle. The rain/snow line during the afternoon from just north of charlotte to near Raleigh. Mainly rain fell across the Coastal Plain after an initial brief period of sleet/snow at the onset.

The storm system continued to deepen and track northeast along the Southeast U.S. coast overnight on Friday before reaching Cape Hatteras by 12 UTC on 12/19. This is a typical "Miller Type A" low pressure track and is also a preferred track for winter storms in North Carolina. This surface pattern is characterized by a relatively simple mean sea level pressure pattern where one relatively well organized surface low tracks northeastward with a surface high pressure center located to the north.

The "Miller A" pattern typically features a well established narrow transition zone of mixed precipitation (rain/snow) accompanied by little or no icing. In the idealized pattern, the snow/rain line follows closely the 1310 thickness line and stays a bit north of the 1540 thickness line. The snow/rain will bulge a bit to the south near the area of enhanced upward vertical velocities due to the effects of the melting snow aloft.

During the evening and early overnight hours, a large area of precipitation overspread much of North Carolina as noted in the 23 UTC radar image. The snow became heavy across the Northwest Piedmont during the evening hours before a transition to sleet and freezing rain occurred a few hours before midnight. This precipitation was largely driven by lower level forcing as noted in the low level frontogenesis. By around midnight, the precipitation shield had advanced northeast and was beginning to depart central North Carolina. By the next morning, most of the precipitation was near the Virginia border and in the form of scattered light showers of snow or freezing rain or freezing drizzle. This was due to the strong warm nose intrusion in the mid levels from the south during the evening.

The storm system moved northeast and deepened further during the next 24 hours spreading heavy precipitation across much of the Mid Atlantic and Northeast. A large area of heavy snow with 12–24 inches of snow accumulation was observed from West Virginia northeast across the Washington D.C. area, northeast to Philadelphia and then into portions of the New York City area.

Snow accumulations across North Carolina were stratified from northwest to southeast consistent with the "Miller A" pattern. Over a foot of snow fell in many of the mountain locations with 6-12 inches of snow in the Foothills. An average of 2-6 inches of snow fell in the western and far northern portions of the Piedmont, mainly north and west of Interstate 85. Snow accumulations quickly diminished further east with just a trace to an inch or two of snow across the eastern and southern Piedmont. Accumulations of sleet were fairly light, generally less than a quarter of an inch on top of the snow. Accumulations of freezing rain ranged from a trace up to 1/10 of an inch, mainly across the western and northern Piedmont.


Snow Accumulation Maps

NC snow accumulation map

Regional snow accumulation map


MODIS Terra Satellite Image Showing Snow Cover from 2009/12/20

MODIS visible satellite image showing snow cover - click to enlarge




Surface Analysis

A fairly strong, cold area of high pressure was located in the St. Lawrence River Valley at 12 UTC on Friday, December 18, 2009. The high pressure center remained stationary and persisted into the early afternoon hours with a ridge extending south into the Mid Atlantic. A developing low pressure system which was located the northern Gulf of Mexico early in the day, intensified as it moved northeast reaching the Savannah Georgia area by 00 UTC on December 19, 2009. The surface low reached Cape Hatteras by 12 UTC on December 19 and then off the Virginia Capes by 00 UTC on December 20, 2009.

The surface pattern was characterized by a relatively simple mean sea level pressure pattern where one relatively well organized surface low tracks northeastward with a surface high pressure center located to the north. This so-called "Miller A" pattern of cyclogenesis often produces significant winter storms in the Carolinas provided sufficient cold air is available. The "Miller A" pattern typically features a well established narrow transition zone of mixed precipitation (rain/snow) accompanied by little or no icing.

A Java Loop of surface analysis imagery from 00 UTC December 18 through 21 UTC December 19, 2009 shows the deepening of the surface low as it tacks northeast along the Southeast U.S. coast.



Surface analysis from 00 UTC on December 19, 2009 - click to load a loop of Surface analysis




Satellite Imagery

Water vapor imagery was used to monitor the various mid and upper level features during the event. The interaction of several short waves across the Deep South late on 12/18 is visible as is the development and progression of a significant dry slot that moved across the central and eastern Carolinas.

A Java Loop of water vapor imagery from 0615 UTC December 18 through 2315 December 19, 2009 is available.

Water Vapor satellite imagery loop from 0615 UTC December 18 through 2315 December 19, 2009 - click to load loop



Southeast Regional Radar Imagery

The regional radar imagery shows the multiple rounds of precipitation that moved across the region during the event. A narrow band of precipitation moved north into North Carolina from South Carolina just after 12 UTC on 12/18. This narrow band advanced north, reaching Raleigh and Greensboro just after 17 UTC. This area of precipitation appeared to be associated with an axis of frontogenesis at mid levels as noted in the SPC analysis at 16 UTC 18 UTC, and 20 UTC.

During the afternoon and evening, steady precipitation developed across western North Carolina including the Western Piedmont and Triad area. Further east, after a 30-60 minute period of light precipitation associated with the initial band, a lull in the precipitation persisted for several hours across central and eastern North Carolina. The precipitation across western North Carolina was driven by low and mid level frontogenesis, low level moisture transport, and divergence aloft associated with the right entrance region of the polar jet over the Northeast and to a lesser extent the left exit region of the subtropical jet.

Later in the evening, a large area of precipitation overspread much of North Carolina as noted in the 23 UTC radar image. This precipitation was driven largely by lower level forcing as noted in the low level frontogenesis. By around midnight, the precipitation shield had advanced northeast and was beginning to depart central North Carolina. By the next morning, most of the precipitation was near the Virginia border and in the form of light showers of snow or freezing rain and freezing drizzle.

A Java Loop of Southeast regional radar imagery from 2358 UTC on Thursday, December 17 through 2158 UTC on Saturday, December 19, 2009 is available.

Southeast regional radar reflectivity from 0011 UTC on December 19, 2009 - click to animate




TREND’s Predominant P-type Nomogram

Universal P-type nomogram with Observed data from 6 hourly RAOBs at KGSO from 00 UTC on 12/17 through 12 UTC on 12/19

The nomogram to the right shows the distribution of precipitation (p-type) TREND's as a function of partial thickness values. Close examination of precipitation events over the past 30 years accounts for the boundaries on the nomogram separating the various p-type areas. Mid level thickness values increase from left to right along the x axis. Low level thickness values increase from bottom to top along the y-axis.

The nomogram to the right displays the observed thickness values from the 6 hourly RAOB's at KGSO from 00 UTC on 12/17 through 12 UTC on 12/19. On the 18th, the atmosphere cooled gradually, especially in the lower levels. The precipitation began in Greensboro in the form of snow at around 18 UTC on 12/18 and continued almost entirely as snow through 02 UTC on 12/19 when it mixed with some sleet before changing to freezing rain or freezing drizzle by 04 UTC 12/19. Note the rapid warming in the mid level thickness values between 00 and 06 UTC when the 850-700 hPa thickness warmed 25 meters to 1572 meters before the mid level thickness values drop 31 meters between 06 and 12 UTC. The rapid warming in the 850-700 hPa layer is evident in the 06 UTC RAOB from KGSO which reported a temperature of 5 deg C at 6,000 feet or 797 hPa. The development and erosion of the warm nose can also be viewed with AMDAR aircraft soundings taken during this period at Greensboro: 2323 UTC 12/18 |  0523 UTC 12/19 |  0635 UTC 12/19 |  1044 UTC 12/19

The surge of warm air in the 850-700 hPa layer was handled with varying success by the NWP guidance. A display of the NAM thickness plots for GSO is available which shows the observed thicknesses values plotted on the nomogram along with the output from the NAM runs beginning with the 00 UTC 12/17 cycle through the 06 UTC 12/19 cycle. A similar display of the GFS thickness plots for GSO is also available.

METAR observations from KGSO - click to enlarge



AMDAR Aircraft Soundings

AMDAR is an acronym for Aircraft Meteorological DAat and Reporting (AMDAR) which is an international effort within the World Meteorological Organization to coordinate the collection of environmental observations from commercial aircraft. In the United States, we often refer to the Meteorological Data Collection and Reporting System (MDCRS) which is a private/public partnership facilitating the collection of atmospheric measurements from commercial aircraft to improve aviation safety.

AMDAR is very useful for short term forecasting situations where conditions are changing rapidly and in particular for aviation forecasting. Regarding winter weather events, AMDAR data can provide forecasters with the height of the freezing level, the presence of elevated warm layers, indications of thermal advection and dry layers. All of these are necessary for accurate precipitation type forecasts. The availability of this upper air data at times and locations where RAOB data may be lacking is invaluable.

The image below contains a loop of AMDAR soundings at KGSO during the event from 0328 UTC on 12/18 through 2341 UTC on 12/19. During the 45 hour period from 03 UTC on 12/18 through 0000 UTC on 12/20 there were a total of 14 AMDAR soundings available at KGSO. Further east at KRDU, there were 29 observations during the same time period. Forecasters used AMDAR data to supplement other observational data to monitor the cooling of the lower atmosphere when it became saturated, the development of a near freezing isothermal layer, the development and erosion of a warm nose aloft between around 700-800 hPa from around 00 UTC to 06 UTC on 12/19, and to evaluate model performance based on these observations.

As noted previously, the development and erosion of the warm nose aloft at KGSO was observed with AMDAR aircraft soundings from 2323 UTC 12/18 |  0523 UTC 12/19 |  0635 UTC 12/19 |  1044 UTC 12/19.  These soundings provided additional support and evidence to the special 06 UTC RAOB at KGSO and allowed forecasters to monitor the evolution of the warm nose. A plot comparing the 06 UTC RAOB at KGSO, which was released at 0534 UTC, with the AMDAR aircraft soundings from 0523 UTC and 0635 UTC on 12/19 shows the fairly good agreement in observed temperatures between these observations. Some variability can be expected given the difference in observation times and the location of the instrument (ROAB is carried by the wind and AMDAR sounding is dictated by the aircraft course.)

A Java Loop of AMDAR soundings at KGSO from 0328 UTC on 12/18/09 through 2341 UTC on 12/19/09 that can be stopped, controlled and zoomed is available.



Loop of AMDAR soundings at KGSO during the event from 038 UTC on 12/18 through 2341 UTC on 12/19



The Micro Rain Radar

MRR unit - click to enlarge The NWS Raleigh has used Micro Rain Radar data during the past few winter seasons in a collaborative effort between the NWS Raleigh and Dr. Sandra Yuter of the Cloud and Precipitation Processes and Patterns Group (CPPPG) at NC State and with the Renaissance Computing Institute (RENCI) from Chapel Hill, NC.

The Micro Rain Radar (MRR) is a vertically-pointing Ku-band radar. The output from the MRR can be processed to provide the user with a vertical view of reflectivity and Doppler velocity. The high temporal and vertical spatial resolution of the MRR along with its unique data makes it an interesting tool to view the characteristics and type of precipitation during winter storms.

Both the reflectivity and the Doppler velocity datasets are displayed with a vertical axis depicting height in thousands of feet and a horizontal axis depicting time. The reflectivity data can be used to determine precipitation intensity and potential bright banding. The velocity data can be used to determine precipitation type based on the fall velocity of the hydrometeor. Snowflakes have a much slower fall velocity than sleet or rain (including freezing rain). The location where a slower fall speed changes into a faster one can be inferred as the melting layer. The melting layer is typically associated with an increased in reflectivity as well.

During this event, there were two MRR units active in the NWS Raleigh forecast area. One MRR unit, operated by the Cloud and Precipitation Processes and Patterns Group (CPPPG) at NC State was located in Holly Springs, NC, around 19 miles south of the Raleigh-Durham International Airport (KRDU). The other unit, operated by the Renaissance Computing Institute (RENCI) was located in Greensboro, NC around 5 miles east of the Piedmont Triad International Airport (KGSO).


Holly Springs MRR Data

The data from the MRR unit in Holly Springs (see figure below) nicely shows the evolution of the event. Initially the precipitation falls as virga, note that both the reflectivity and fall velocity data are above the surface between 1500-1700 UTC (see the detailed MRR output from 18/1600 UTC through 19/00 UTC for a closer view). Then for a short period of time, the reflectivity and fall velocity data reach the surface between 1700-1800 UTC. This brief period of precipitation is consistent with the narrow band of precipitation discussed elsewhere in this event summary. Given the increase in fall velocities below around 6 kft and the increase in reflectivities, the MRR suggests that the initial precipitation type was snow before changing to rain or possibly mixed rain and sleet after 1730 UTC. This is consistent with spotter reports in and around the Triangle area.

The precipitation resumes again at around 2200 UTC. After another period of virga, the precipitation appears to fall as a brief period of some snow or mixed snow and rain between 2200 and 2230 UTC, the fall velocities increase significantly at around 2230 UTC (see the detailed MRR output from 18/1600 UTC through 19/00 UTC for a closer view). The melting layer can be inferred by noting the gradient in Doppler velocity, particularly the pronounced increase in fall velocities from around 1-2 m/s characteristics of snow to more than 3-4 m/s which is characteristic of rain. At 2300 UTC, the melting layer appears to be located at around 6,500 ft. An AMDAR aircraft sounding at KRDU, around 19 miles north of the MRR location, from 2324 UTC shows the temperature profile warming above 0 deg C at around or just above 6,000 ft. The AMDAR sounding is consistent with the MRR data given the AMDAR sounding was taken at KRDU, likely in a colder atmosphere nearly 20 miles to the north.

During the remainder of the evening, the melting layer increases in height from 2230 UTC through 0200-0500 UTC when the melting layer reaches around 10 kft. Another AMDAR aircraft sounding at KRDU, well north of the MRR location, from 0544 UTC shows the height of the 0 deg C isotherm just above 10,000 ft.

MRR reflectivity and velocity data from Holly Springs - click to enlarge

One other interesting aspect to this event was the occurrence of snow showers across the Northern Piedmont on Saturday, December 19th. As noted in the KGSO 12 UTC RAOB, the lower portion of the atmosphere was beginning to cool by Saturday morning while the moisture in the mid and upper levels of the atmosphere decreased notably. The drying aloft, especially above the ice crystal growth regions of -10 deg C or colder, suggested to forecasters that there was little potential for much in the way of snow. Forecasters expected mainly light rain or drizzle or some light freezing rain or freezing drizzle.

Examining the detailed MRR output from 19/0000 UTC through 19/17 UTC shows that the melting layer between 0900-1200 UTC was around 6 kft. This is fairly consistent with another AMDAR aircraft sounding at KRDU, well north of the MRR location, from 1123 UTC which shows the height of the 0 deg C isotherm just above 7,000 ft.

During the remainder of the morning, AMDAR soundings at KRDU showed the dramatic cooling of the lower troposphere. But the thermal profile did not approach the -10 deg C isotherm until well above 10 kft. The MRR output from Holly Springs showed that the precipitation was fairly shallow, only extending up to around 8 kft or less during the late morning hours of December 19. North of Holly Springs, the periods of light drizzle or freezing drizzle changed to a brief period of snow in the Triangle area and points north and east. While the impact of the brief period of snow in the Raleigh area on December 18 was limited, its occurrence points to the difficulty in completely understanding the complex microphysical processes that occur in winter storms.

MRR reflectivity and velocity data from Holly Springs - click to enlarge


Greensboro MRR Data

The data from the MRR unit in Greensboro shows the arrival of the precipitation as virga between 1000 and 1200 EST with snow observed at KGSO by 1300 EST. Despite the varied intensities, with reflectivity values aloft exceeding 30 dBZ at times, the precipitation fell as all snow during the afternoon and early evening. The low fall velocities aloft during the afternoon and early evening, generally less then 2 m/s, is consistent with snow. In fact, the visibility at KGSO was around a half mile for much of the time from 1600-1800 EST, indicative of light snow which was approaching moderate intensity.

The snow transitioned to freezing rain just before midnight EST. The change over in precipitation type is most notable in the change in the fall velocity between 20:00 and 22:00 EST. The faster fall velocities (greater then 4 m/s) are indicative of a precipitation type other than snow. In addition, the reflectivity traces are more vertical suggesting less snow. The change in the precipitation type is consistent with the AMDAR aircraft observations at 2323 UTC 12/18 and 0523 UTC 12/19 which show significant warming in the 5-10 kft layer. Also of note is the drop in the vertical depth of the reflectivities during the afternoon and evening as deep moisture was cut off from the region. time

METAR listing for KGSO

MRR reflectivity data from near KGSO - click to enlarge

MRR velocity data from near KGSO - click to enlarge



NCEP High Resolution WRF Model Output

NCEP HiResWindow WRF Configurations - click to enlarge High resolution NWP models have been widely used to help with short range forecasting and warning operations. Recent advances in NWP and in computational efficiency have resulted in the availability of high resolution model simulation. Since 2002, NCEP has run the WRF-NMM operationally in NCEP High Resolution windows in various domains across North America, Hawaii, and Puerto Rico. The grid point spacing of the nests was decreased to around 5 km in 2005 and additional improvements were made in 2007.

Currently NCEP is making daily runs of the High Resolution Window (HRW) or (HiResWindow) Weather Research and Forecasting (WRF) simulations 4 times a day. These runs are "nested" inside of the parent North American Meso (NAM), getting initial and boundary conditions from the NAM. The output is from the WRF model versions of the non-hydrostatic, hybrid vertical coordinate mesoscale model (NMM) and Advanced Research WRF (ARW). WRF forecasts are produced every six hours at 00, 06, 12 and 18 UTC. The WRF graphics are available at three hour increments out to 48 hours.

Users of NWP guidance have found considerable value in the NCEP HiResWindow and other high resolution NWP output, especially during the warm season when severe weather threats can be closely related to convective mode, location and evolution. In addition, the increased temporal resolution and new output products such as radar reflectivity have provided forecasters with new approaches to interpreting model output.

The current status of the NCEP HiResWindow is available at the following URL: http://www.emc.ncep.noaa.gov/mmb/mmbpll/nestpage/

Current output from the HiResWindow for the Eastern U.S. is available at the links below.
HiResWindow NMM 00 UTC Eastern US http://www.nco.ncep.noaa.gov/pmb/nwprod/analysis/namer/hiresw/00/model_l.shtml
HiResWindow ARW 00 UTC Eastern US http://www.nco.ncep.noaa.gov/pmb/nwprod/analysis/namer/hiresw/00/model_m.shtml

HiResWindow NMM 12 UTC Eastern US http://www.nco.ncep.noaa.gov/pmb/nwprod/analysis/namer/hiresw/12/model_l.shtml
HiResWindow ARW 12 UTC Eastern US http://www.nco.ncep.noaa.gov/pmb/nwprod/analysis/namer/hiresw/12/model_m.shtml



Comparison of NCEP High Res WRF NMM and ARW with Observed Radar Reflectivity

Comparison of 30 hour forecast reflectivity from the NCEP High Res WRF NMM and WRF ARW from the 12/17 12 UTC init with observed radar reflectivity - click to enlarge At the beginning of the event across central North Carolina, a narrow band of precipitation moved north into North Carolina from South Carolina just after 12 UTC on 12/18. This narrow band advanced north, reaching Raleigh and Greensboro just after 17 UTC with most of the precipitation falling as light snow. The precipitation lasted from 30 to 60 minutes and then diminished. This area of precipitation appeared to be associated with an axis of frontogenesis at mid/upper levels as noted in the SPC analysis at 16 UTC 18 UTC, and 20 UTC.

This initial band of precipitation was followed by a multiple hour lull in the precipitation across central and eastern North Carolina including Raleigh. Further west, steady precipitation developed and persisted across western North Carolina including the Western Piedmont and Triad area. Later in the evening, at around 23 UTC, a large area of precipitation overspread much of North Carolina.

While the lull in the precipitation was not a complete surprise to some forecasters, identifying and forecasting a mesoscale feature such as this would be very difficult given traditional NWP output. Determining whether the precipitation would initially fall intermittently before becoming steady or whether the precipitation would begin abruptly and rapidly become steady or heavy with quick accumulations would be very beneficial to many users.

During this event, the NCEP HiResWindow simulations were very successful in predicting the northwest-southeast stretching precipitation band that advance northeast across central and eastern North Carolina.

A comparison of forecast reflectivity imagery from the NCEP High Res WRF NMM, WRF ARW, and observed radar reflectivity is available below.

2009/12/17 00 UTC init
F36 / Valid 12 UTC 12/18 |  F39 / Valid 15 UTC 12/18 |  F42 / Valid 18 UTC 12/18 |  F45 / Valid 21 UTC 12/18 |  F48 / Valid 00 UTC 12/19 

2009/12/17 12 UTC init
F24 / Valid 12 UTC 12/18 |  F27 / Valid 15 UTC 12/18 |  F30 / Valid 18 UTC 12/18 |  F33 / Valid 21 UTC 12/18 |  F36 / Valid 00 UTC 12/19 

2009/12/18 00 UTC init
F12 / Valid 12 UTC 12/18 |  F15 / Valid 15 UTC 12/18 |  F18 / Valid 18 UTC 12/18 |  F21 / Valid 21 UTC 12/18 |  F24 / Valid 00 UTC 12/19 


NWS Raleigh Local High Resolution WRF Model Output

30 hour reflectivity forecast from the NWS RAH 4km WRF NMM from the 12/17 12 UTC init with observed radar reflectivity - click to enlarge Within the past year, NWS Raleigh has been developing a local high resolution modeling capability. We expect to use the model in both an operational and research capacity to help refine phenomena related to smaller scale weather events that are common to our area. In addition, the model will allow us to produce and examine model fields that are not readily available to support operational needs as well as research and collaborative projects. Finally, we expect that the overall knowledge and understanding of how numerical weather prediction models (NWP) operate will increase, resulting in improved NWP utilization at our office.

As in the previous examples, the local 4km WRF NMM produced at WFO Raleigh indicated the potential for a band of snow to move northeast across central North Carolina before the main precipitation shield arrived. In the example shown above/to the right, the NWS RAH 4km WRF NMM 30 hour reflectivity forecast from the 12/17 12 UTC initialization is shown with the observed radar reflectivity. The NWS Raleigh WRF output does a very good job identifying this feature. Previous model runs including the 21 hour reflectivity forecast from the 12/17 21 UTC initialization and the 9 hour reflectivity forecast from the 12/18 09 UTC initialization show a similar pattern with this feature. Experience has shown that in general the high resolution model output has good success with identifying the precipitation pattern, convective mode, and overall picture of the precipitation distribution, especially with well forced events. The high resolution model output can struggle with the timing and exact placement, but useful information can be inferred from the output.



Using Analogs as Medium Range Guidance for Organized Snow Events


Note - much of the text below was provided by the CIPS Winter Storm Analog Guidance page.

Researchers at the Cooperative Institute for Precipitation Systems (CIPS) along with the NWS St. Louis, MO and the NWS Springfield, MO are engaged in a project to show how the knowledge of past events that exhibit similar characteristics to the current forecast can assist forecasters with a range of potential scales and intensities of the upcoming event. The project is entitled "Using analogs as medium range guidance for organized snow events."

The goal of the analog forecast approach is not to make a forecast; but to provide medium-range guidance for events by using a historical dataset. In addition, a forecaster can quickly gain historical experience and become familiar with the meteorological patterns associated with certain events. The analog forecast approach can be applied to any meteorological event as long as a control run can be created.



Methodology

The methodology below is used by the CIPS Winter Storm Analog Guidance Project which currently uses 36, 48, 60, and 72 hour forecasts from the 40 km GFS.

  • The basis for analog guidance is to search a climatological dataset for maps that resemble the current forecast, and then assume that the atmosphere will evolve similarly to the historical analogs.

  • Search the 28 year North American Regional Reanalysis (NARR) dataset against the model forecasts provided by the 40km GFS (GFS212) for potential analogs based on the constraints below:
      6 months over the winter season (October - March)
      6 hour temporal resolution
      Potential database of 20,384 analogs (28 winters, 6 months, 4 per day)
  • Remove “duplicate” times for the same analog event by choosing the best analog over a 24-h period.

  • Refine and rank the resulting analogs to create products that are useful for medium range guidance.

  • Determining what constitutes an analog is done statistically using the following techniques:
      Pattern Correlation
      Mean Absolute Error
      Root-Mean-Square Error
      Anomalies
  • During the first pass through the NARR dataset, the statistics are computed on a large domain (REGN) and then on a smaller domain (MESO). If certain thresholds are not exceeded, the date/time is not considered a potential analog.

  • To reduce the approximately 20,000 potential analogs, threshold values were determined based on the control run for the following fields:
      300 HGHT COR 0.85 REGN
      500 HGHT COR 0.83 REGN
      850 TMPC MCOR 0.88 MESO
      850 TMPC MMAE 3.8 MESO
      PMSL MCOR 0.83 MESO
      850 HGHT MCOR 0.70 MESO
  • Once the approximately 20,000 potential analogs are reduced, “duplicate” times are removed. “Duplicate” times occur due to the variability in system speed (e.g., a slow historical system may exhibit similar patterns to the forecast over a longer period of time).

  • The best analog is found over a 24-h period by using the following formula:
      SUM(COR) - SUM(MAE/3)
  • After the potential analogs are reduced, the program is rerun to find statistics on the following variables:
      300 HGHT COR REGN
      500 HGHT COR REGN
      700 FRNT COR MESO
      850 HGHT COR MESO
      850 TMPC COR MESO
      850 TMPC MAE MESO
      850 FRNT COR MESO
      850 THTEADV COR MESO
      2m TMPC COR MESO
      2m TMPC MAE MESO
      PMSL COR MESO
      PWTR COR MESO
  • After new statistics are determined, a results score is computed using the following formula:
      850HGHTCOR*3 + PMSLCOR*2 + SUM(COR) - SUM(MAE/3)
  • Finally, in order to catch possible system propagation, statistics are computed for ± 12 h from the time of the best analog using the matching forecast (the 48 hour forecast is used in the example below):

    (-12 h Analog: GFS 036h FCST, Analog: GFS 048h FCST, +12 h Analog: GFS 60h FCST). Using statistics from the ± 12 h times, results scores are also computed.

  • The final analog rank is determined by using the results scores in the following formula:
      AVERAGE(m12ANALOG, ANALOG, p12ANALOG)


    Example Analog for this Event

    The CIPS Heavy Snow Analog Guidance for 12 UTC on 2009/12/19 across the East Coast 2 domain was examined for this event and a brief review of the process and data is shown below. The analog was based off the 48 hour forecast from the 40 km GFS (212 grid) valid at 12 UTC on 12/19. The four panel GFS 48 hour forecast indicated a strong subtropical jet over the eastern Gulf of Mexico and Florida along with the exit region of the polar jet over the New England states and a well developed surface low off the North Carolina coast with surface pressures less then 992 mb.

    The CIPS Heavy Snow Analog Guidance Table for the East Coast 2 domain from the GFS 40km 12 UTC 20091217 48 hour forecast is shown below (click to enlarge). Note the statistics for GFS212 20091217/1200F048 forecast:

    Total number of potential analogs: 21140
    Number of potential analogs that did not exceed set thresholds: 21005
    Number of potential analogs that exceeded set thresholds: 135
    Number of distinct analog events: 73

    Percentage of potential analogs that did not exceed the set threshold at each step of the analog process:
    300-mb HGHT correlation: 45.08 %
    500-mb HGHT correlation: 4.80 %
    850-mb TMPC correlation: 25.37 %
    850-mb TMPC mean-absolute-error: 6.31 %
    PMSL correlation: 15.82 %
    850-mb HGHT correlation: 2.61 %


    CIPS Heavy Snow Analog Guidance Table for 
the East Coast 2 domain from the GFS 40km 12 UTC 2009/12/17 48 hour forecast  - click to enlarge


    The 1983/02/11 18 UTC analog had the greatest final analog rank based on the averaging of that forecast along with the statistics from 12 hours prior to and after the 1983/02/11 18 UTC forecast. Examining the CIPS Heavy Snow Analog Guidance Table shows that this event had the greatest similarities for the 300-mb HGHT correlation, the 500-mb HGHT correlations, and persistence with slightly reduced similarities for the lower tropospheric fields such as PMSL correlation and the 2mTMPC correlation.

    Comparison of the 1983/02/11 analog with the 2009/12/18 observed snow fall - click to enlarge

    The CIPS Heavy Snow Analog page provides analysis maps of the COOP snowfall from the various analogs along with some probabilities based on the analogs. The COOP snowfall analysis map from the CIPS Heavy Snow Analog page for the 1983/02/11 18 UTC analog is shown below. In addition, the mean and median of the top 15 analogs are also provided on the CIPS Heavy Snow Analog page.

    Analysis map of the COOP snowfall from 1983/02/11 - click to enlarge Mean of the top 15 analogs - click to enlarge Median of the top 15 analogs - click to enlarge

    The CIPS Heavy Snow Analog page also provides images displaying measures of center and spread guidance for various fields including mean sea level pressure, 850 mb height, 500 mb height, 300 mb height, 850 mb temperature, and 2 meter temperature.

    500 hPa height mean based on the top 15 analogs - click to enlarge 500 hPa 5460 meter height spaghetti based on the top 15 analog - click to enlarge

    850 hPa temperature mean based on the top 15 analogs - click to enlarge 850 hPa 0 deg C spaghetti based on the top 15 analog - click to enlarge

    Mean sea level pressure mean based on the top 15 analogs  - click to enlarge 1008 hPa spaghetti based on the top 15 analogs - click to enlarge

    Some of the more insightful products are the probabilistic forecasts of COOP snow greater then various thresholds. The probabilistic forecast of COOP snow greater then 2, 4 or 6 inches based on the top 15 analogs is shown below. Additional probabilities for COOP snow amounts greater then 8 inches are also available along with probabilities of snow to liquid ratios from COOP observations greater then 10:1, 12:1, 14:1, and 16:1.

    Probabilistic forecasts of COOP snow greater then 2 inches based on the top 15 analogs - click to enlarge Probabilistic forecasts of COOP snow greater then 4 inches based on the top 15 analogs - click to enlarge Probabilistic forecasts of COOP snow greater then 6 inches based on the top 15 analogs - click to enlarge


    Utility of Analogs for this Event

    Most of the forecast shift preceding the event utilized analogs in one form or another. Analogs were referred to by HPC in the PMDEPD products and local forecasters took advantage of the new Medium Range Guidance for Organized Snow Events from CIPS.

    The use and application of analogs, in particular the CIPS project are growing rapidly, most likely because they are becoming easier to incorporate into our decision-making given the advances in technology and communication. Many operational forecasters have long gravitated to their own "analogs", or personal experiences, when preparing forecasts anyway. These statistical analogs are simply a more objective way of drawing on past cases.

    During the December 18th winter storm, a CIPS graduate student was monitoring communications in the in the NWSChat and they may do more monitoring for potentially big storm in the future. During the event he let us know that the strong analog case for that event (February 10-12, 1983) was amazingly persistent (popping up run after run) and that it was ranked by NCEP as the 10th biggest snowstorm with 45+ deaths.



  • Hourly 4 inch (0.1m) Soil Temperatures

    The image below (click on it to enlarge) shows the hourly 4 inch soil temperatures at 6 locations across central North Carolina from midnight on December 16 through midnight on December 20, 2009.

    Note that across the Triad locations (NCAT - NC A&T State University Research Farm in Greensboro and HIGH - UNCG Lindale Farm Station in High Point, NC) where several inches of snow accumulated, soil temperatures peaked in the lower to mid 40s during the day before the snow fell. During the preceding day, on December 16, max soil temperatures were in the mid to upper 40s. Past experience has shown that when max soil temperatures during day preceding a snow fall are in the lower 40s or colder and given modest snow rates with surface temperatures at or near freezing, the snow can be expected to accumulate. By the time the snow fell in the Triad, soil temperatures were in the upper 30s.

    Further east across the Triangle area (CLAY - Central Crops Research Station, Clayton, CLA2 - DAQ Clayton Profiler, Clayton, and REED - Reedy Creek Field Laboratory, Raleigh, NC) soil temperatures were milder. Soil temperatures in these locations were in the lower 40s during most of the event after peaking in the mid 40s during the preceding day.

    NCAT - NC A&T State University Research Farm, Greensboro NC
    HIGH - UNCG Lindale Farm Station, High Point, NC
    CLAY - Central Crops Research Station, Clayton, NC
    CLA2 - DAQ Clayton Profiler, Clayton, NC
    REED - Reedy Creek Field Laboratory, Raleigh, NC
    SILR - Siler City Airport, Siler City, NC

    Chart of observed 4 inch soil temperature - click to enlarge




    CoCoRaHS Observer Network

    CoCoRaHS is a grassroots volunteer network of weather observers of all ages and backgrounds working together to measure and map precipitation (rain, hail and snow) in their local communities. By using low-cost measurement tools, stressing training and education, and utilizing an interactive web-site, CoCoRaHS aims to provide the highest quality data for natural resource, education and research applications. The only requirements to join are an enthusiasm for watching and reporting weather conditions and a desire to learn more about how weather can effect and impact our lives. North Carolina joined the CoCoRaHS network in 2007. For more information, visit the CoCoRaHS web site at www.cocorahs.org.

    The CoCoRaHS Web page provides the ability for CoCoRaHS observers to see their observations mapped out in "real time", as well as providing a wealth of information for our data users. The snow accumulation maps from the CoCoRaHS web site (shown below) were a great resource for WFO RAH.


    CoCoRaHS Intense Snow Report

    NWS Raleigh received 8 CoCoRaHS Intense Snow Reports during this winter storm. An example is shown below with all of the reports available here. The intense snow reports are extremely helpful since it can be difficult to get accurate snow accumulation reports along with information on the various precipitation types and impact. Several of the Intense Snow Reports were received late in the evening or overnight when it is very difficult to get reliable information. The value of these reports cannot be overstated.

    NZUS45 KBOU 190321
    CCRAHS
    
    intense snow report from CoCoRAHS spotter:
    12/18/2009 10:15 PM local time
    County: Forsyth NC
    Lewisville 4.2 N (number NC-FR-4)
    
    Latitude: 36.154362
    Longitude: -80.405506
    7.20 inches of snowfall in the past 10 hrs
    NA inches of snow on the ground
    Comments: 7.2" of snow is the storm total so far.  Sometime around 10
    pm i.e. about 15 min ago) the moderate to heavy snow changed
    abruptly to a mix of sleet and freezing rain, still coming
    down quite hard.  Already a light glaze on vertical metal
    surfaces.  The wind has been slowly strengthening, with quite
    a bit of blowing snow until the sleet-freezing rain mix
    started to coat the snow.  Temp still 29.
    
    Received NWS Boulder Fri Dec 18 20:21:42 2009 MST
    
    Sent to WFOs: RAH,GSP,RNK
    


    Central North Carolina Snow Accumulation Totals

    CoCoRAHS central North Carolina snow accumulation map - click to enlarge



    Triad Area Snow Accumulation Totals

    CoCoRAHS Triad area snow accumulation map - click to enlarge



    Selected Videos of the Winter Weather Event

    Videos courtesy of Jeremy Gilchrist.
    (Click the image to open the video)

    Snow packed highway with snow falling south of Greensboro on U.S. 220 approximately 5-10 miles south of I-40 at around 830 PM on 12/18. - video courtesy of Jeremy Gilchrist - Click to open           Sleet falling approximately 10 miles north of Asheboro just off U.S. 220 at around 930 PM on 12/18. - video courtesy of Jeremy Gilchrist - Click to open          


    Final Thoughts and Lessons Learned

    Model guidance was generally good and forecasters took advantage of the good forecast data and their meteorological reasoning and conceptual models to issue accurate forecasts and warnings. Winter Storm Watches verified with an average lead time of 39 hours. Winter Storm Warnings verified with an average lead time of around 22 hours with no missed events. Nine of 12 counties in the Winter Storm Warning verified with the 2 of the 3 counties that did not verify (Randolph, Chatham, and Franklin) receiving Winter Weather Advisory conditions.

    Forecasters first shared a potential pattern change on December 6 via an email to external users - Over the past week there have been increasing indications that there will be a shift in the upper level flow pattern over North America beginning later next week. Those indices are pointing toward increasing cold for the eastern United States.

    The event followed the "Miller A" surface pattern conceptual model which is characterized by a relatively simple mean sea level pressure pattern where one relatively well organized surface low tracks northeastward with a surface high pressure center located to the north. The "Miller A" pattern typically features a well established narrow transition zone of mixed precipitation accompanied by little or no icing...

    Numerous briefings were provided to local emergency managers and decision makers during the event via the new NWS Raleigh Briefing Web Page and via other online conferencing software. The ability to share information with users who may be at home or away from the office was invaluable.

    High resolution model forecasts from the NCEP HighResWindow WRF and the local NWS Raleigh WRF provided very good guidance during the event and highlighted some of the mesoscale features that would impact the precipitation distribution.

    The Micro Rain Radar (MRR) provides a vertical view of reflectivity and Doppler velocity. Numerous details of the vertical thermal profile and can be inferred from the MRR data. This allows forecasters to better understand and anticipate precipitation type changes at the surface.

    The use and application of analogs, in particular the CIPS project are growing rapidly. The CIPS Winter Storm Analog Guidance data was utilized by several forecasters during the event, most likely because they are becoming easier to incorporate into our decision-making given the advances in technology and communication. During the December 18th winter storm, a CIPS graduate student was monitoring communications and dialogued with forecasters in the in the NWSChat. Feedback from several forecasters was provided to the CIPS group.

    This event was another case in which forecasters utilized AMDAR aircraft soundings to provide thermal profiles at GSO and RDU in between RAOB times. This data set is an important resource during critical winter weather forecasts. Forecasters during the event utilized the http://amdar.noaa.gov/ website and AWIPS to view these soundings during the event.

    North of Holly Springs, the periods of light drizzle or freezing drizzle changed to a brief period of snow in the Triangle area and points north and east. While the impact of the brief period of snow in the Raleigh area on December 18 was limited, its occurrence points to the difficulty in completely understanding the complex microphysical processes that occur in winter storms.

    Several CoCoRaHS intense snow reports were received by the NWS Raleigh. These reports are especially important since it is very difficult to get accurate snow accumulation reports late at night. The value of these reports cannot be overstated.



    Acknowledgements

    Many of the images and graphics used in this review were provided by parties outside of WFO RAH. The upper air analysis images and Skew-T diagrams were obtained from the University of Wyoming. Satellite data was obtained from the National Center for Atmospheric Research. Surface observations provided by the University of Wyoming. The surface analysis graphic was obtained from the Hydrometeorological Prediction Center. Radar imagery was obtained from the National Weather Service web site. AMDAR aircraft sounding data was obtained from the Earth System Research Laboratory - Global Systems Division. Micro Rain Radar data was provided by the NC State Cloud and Precipitation Processes and Patterns Group (CPPPG) and the Renaissance Computing Institute (RENCI). Analog medium range guidance and background provided by the CIPS Winter Storm Analog Guidance page. Ground temperatures and adjacent air temperature data provided by CRONOS from the N.C. State Climate Office. CoCoRaHS maps were provided by the CoCoRaHS organization. Videos courtesy of Jeremy Gilchrist.


    Case Study Team

    Phillip Badgett
    Gail Hartfield
    Russ Henes
    Brandon Locklear
    Jeff Orrock
    Scott Sharp
    Barrett Smith
    Brandon Vincent
    Jonathan Blaes


    For questions regarding the web site, please contact Jonathan Blaes.


  • NWS Disclaimer.