2001 - Albany CSTAR - Virtual Lab

NROW 2001

Agenda with Preprints

Third Northeast Regional Operational Workshop Albany, New York

Tuesday, November 6

1:00 pm Welcoming Remarks

Eugene P. Auciello, Meteorologist in Charge, NWS, Albany, NY
Warren R. Snyder, Science & Operations Officer, NWS, Albany, NY

Session 1. Wasula Storm - December 30-31, 2000
Session Chair - Thomas A. Wasula, NWS Albany, NY

1:10 pm Remarks by Session Chair

1:20 pm The End of the Millennium Snowstorm: A Brief Synoptic Review with an
Emphases on the Role of Jet Streaks
Alicia C. Wasula, Department of Earth and Atmospheric Sciences, University at Albany, State University of New York, Albany, NY

1:45 pm A Look at the Wasula Storm from a Forecaster's Perspective
Hugh W. Johnson IV, NOAA/NWS, Weather Forecast Office, Albany, NY

Session 2. National Centers & Modeling
Session Chair - Warren R. Snyder...NWS, Albany, NY

210 pm Remarks by Session Chair

2:35 pm Winter ER/NCEP Experiment and NCEP Update
Dr. Louis W. Uccellini, NOAA/NWS, National Center for Environmental Prediction, Camp Springs, MD

3:00 pm Break

3:20 pm The Open RPG Deployment, New Capabilities, Hotline Operations and the ROC
Daryl L. Covey, NOAA/NWS, Radar Operations Center, Norman, OK

3:45 pm The NCEP Short Range Ensemble Forecast (SREF) System: Operational
Applications
Steve Tracton, NOAA/NWS, National Center for Environmental Prediction Camp Springs, MD

4:10 pm Effective Use of Regional Ensemble Data in Forecasting a Winter Storm
Richard H. Grumm and Robert Hart, NOAA/NWS, Weather Forecast Office, State College, PA

4:35 pm An Evaluation of ETA, AVN and Limited Ensemble QPF's over the Northern Mid-Atlantic Region During 2000-2001 Winter Season
Michael S. Evans, NOAA/NWS, Weather Forecast Office, State College, PA

5:00 pm Model and Observational Evaluation of the Impact of PNA on the Winter Climate of the Northeast United States
Michael Notaro and Wei-Chyung Wang Atmospheric Sciences Research Center, State University of New York at Albany, Albany, New York

5:25 pm Adjourn

Wednesday, November 7

Session 3. Winter Season Events and Impacts
Session Chair - Kenneth D. LaPenta...NWS, Albany, NY

8:00 am Remarks by Session Chair

8:10 am Death by 1000 Cuts ?
Lance F. Bosart, Department of Earth and Atmospheric Sciences, University at
Albany, State University of New York, Albany, NY

8:35 am An Analysis of Montreal's Record Breaking Heavy Rainfall Event of
8-9 November 1996, and a Comparison with Its Best Analogue
Dorothy Durnford, Department of Atmospheric and Oceanic Sciences, McGill
University, Montreal, Quebec, Canada

9:00 am A Northeast Snowstorm Impact Scale
Paul J. Kocin, The Weather Channel Inc, Atlanta, Georgia

9:25 am An Early Winter Lake-Effect Snowstorm Over Southern Ontario
Q. Liu, Department of Physics, University of Toronto, Toronto, Ontario, Canada

9:50 am The Importance of Snow Microphysics for Large Snowfalls
Jeff S. Waldstreicher, NOAA/NWS, Eastern Region Headquarters, Bohemia, NY

10:15 am Break

Session 4. Warm Season Events and Impacts
Session Chair - George J. Maglaras...NWS, Albany, NY

10:35 am Remarks by Session Chair

10:45 am A Study of Tornadic Versus Non-Tornadic Thunderstorms in Central
Pennsylvania on 2 June 1998
Ernest J. Ostuno, NOAA/NWS, Weather Forecast Office, Grand Rapids, MI

11:10 am An Updated Look at Some Severe Weather Forecast Parameters
Kenneth D. LaPenta, NOAA/NWS, Weather Forecast Office, Albany, NY,

11:35 am Warm Season Closed Lows in the Northeastern United States
Thomas A. Wasula, NOAA/NWS, Weather Forecast Office, Albany, NY

Noon Lunch

Session 5. Operations and Instrumentation
Session Chair - Warren R. Snyder...NWS, Albany, NY

1:00 pm Remarks by Session Chair

1:10 pm Applying D3D in an Operational Environment
John W. Cannon, NOAA/NWS, Weather Forecast Office, Gray, ME

1:35 pm Mesoscale Sensor Network for DOD and Civil Emergency Applications
Dave Sautter, Yankee Environmental Systems Inc., Turners Falls, MA

2:00 pm Beyond IFPS, Empowering Weather Information Consumers
Richard J. Westergard, NOAA/NWS, Weather Forecast Office, Albany, NY;
Prof. Robert E. Sanders, Department of Communication, University at Albany,
State University of New York, Albany, NY

2:25 pm Break

Session 6. CSTAR - Early Results from Cool Season Projects
Session Chair - Daniel P. St. Jean...NWS, Burlington, VT

2:45 pm Remarks by Session Chair

2:55 pm A Climatology of Cold Season Banded Precipitation in the Northeast
United States
David Novak, Department of Earth and Atmospheric Sciences, University at
Albany, State University of New York, Albany, NY

3:20 pm Large-Scale Circulation Anomaly Indices in Relation to Cool-Season
Precipitation Events in the Northeastern United States
David Groenert, Department of Earth and Atmospheric Sciences, University at
Albany, State University of New York, Albany, NY

3:45 pm Characteristics of Cool Season Cutoff Lows in the Northeastern United States: Four Northwest Flow Events in Northern New York State and Northern
Vermont
Daniel P. St. Jean, NOAA/NWS, Weather Forecast Office, Burlington, VT

4:10 pm A Climatology of 500 hPa Cutoff Cyclones
Brandon Smith, Department of Earth and Atmospheric Sciences, University at
Albany, State University of New York, Albany, NY

4:25 pm Closing remarks
Warren R. Snyder, NROW Coordinator

4:30 pm Adjourn

The Fourth Annual Northeast Regional Workshop is planned for
November 5 & 6, 2002

ABSTRACTS

The End of the Millennium Snowstorm: A Brief Synop tic Review with an Emphasis on the role of Jet Streaks

Alicia C. Wasula
Department of Earth and Atmospheric Sciences, University at Albany State University of New York, Albany, New York

Thomas A. Wasula
NOAA/NWS, Weather Forecast Office, Albany, New York

On 30-31 December 2000, a major snowstorm struck the Northeast, dropping copious amounts of snowfall over the region. This major nor'easter produced widespread snowfall accumulations of 25 cm (10 inches) and greater across eastern New York, New Jersey, the extreme eastern border of Pennsylvania and adjacent New England in 10 to 15 hours. Portions of the eastern Catskills and northern New Jersey received 50 to 75 cm (20 to 30 inches) of snow from the storm. Many daily snowfall records were set in the Northeast. Little or no snow fell across most of Pennsylvania, Delaware and Maryland. The major cities such as Trenton, Newark, New York City, Albany and Hartford were hit hard. Snowfall rates of 5 cm to 7.5 cm (2 to 3 inches) per hour were common with the storm north and west of the area of low pressure. Near-blizzard conditions occurred at times with high winds in excess of 15 to 20 m/s (30 to 40 knots). The storm wreaked havoc on travelers on the last weekend before the dawn of a new millennium. This talk will examine how the synoptic-scale features (e.g.: jets, vorticity advection, etc.) evolved in the rapid development of this system.

The synoptic situation at 0000Z/30 featured an area of low pressure (1012 hPa) moving eastward through eastern Ohio, while a second low was situated 250-300 km southeast of North Carolina. By 0600Z/30, a new coastal low of 1004 hPa had formed (much further north than anticipated by forecasters) near the Delmarva Peninsula. By 1200Z/30, this surface low was about 200 km southeast of Atlantic City, and heavy snow began falling across most of northern New Jersey and the New York City metropolitan area. By 1500Z/30, heavy snow was falling along the east facing slopes of the Catskills and was pushing rapidly north into the greater Capital Region and western New England. The surface low (995 hPa) moved over New York City by 1800Z/30 and then progressed rapidly to the northeast to the Connecticut-Rhode Island border (992 hPa) at 0000Z/31. This strong surface cyclone developed due to a powerful 500 hPa low that barreled southeastward through the Midwest and into the Mid-Atlantic states on the morning of 30 December.
The role of upper and lower level jets and their evolution will be investigated from a synoptic point of view to gain an understanding why so much snow fell in a short duration of time. In addition, surface observations will be examined to help explain any local enhancements or inhibitions to snowfall totals. AVN model grids, surface weather observations, upper air data, satellite images and vertical cross sections will be used in the analysis of this major storm which ended the millennium.

A Look at the Wasula Storm from a Forecaster's Perspective

Hugh W. Johnson IV
NOAA/NWS, Weather Forecast Office, Albany, New York

The "Wasula Storm", (named for a WFO staff member whose wedding occurred at its height) which occurred December 30-31st, 2000, presented significant challenges to forecasters at WFO Albany, New York in warning and forecasting this storm. Frontogenesis in the presence of small symmetric stability appeared to be the primary forcing mechanism of a mesoscale snowband. It has been argued by many, that of all the forecasting parameters, Qualitative Precipitation Forecast (QPF) is the hardest to forecast and perhaps the most important to our customers. The numerical models handling of the QPF also created additional challenges to forecasters. This storm is part of a larger project investigating banded storms in an attempt to better understand the specific atmospheric conditions that yield to mesoscale banding.
While the computer models were fairly consistent in the synoptic predictions of a significant developing wave in the southern jetstream, which eventually evolved into a strong Nor'easter, there were mesoscale discrepancies regarding the (QPF). Most of the operational models were slow with the arrival of precipitation, and did not bring snow into WFO Albany, New York Country Warning Area (CWA) until late morning, or early afternoon the 30 ^th.

Once the warm conveyer belt became established, precipitation expanded rapidly northward. Precipitation reached southern New Jersey 0800 UTC, northern New Jersey 1000 UTC and reached the Capital District of New York 1200 UTC Saturday morning. Then, the snow overspread much drier air, across the Adirondacks, slowing its northward progress. It took another three hours for the first flakes to reach the Glens Falls, New York area. The 12/30/00z ETA model run indicated the maximum QPF would be east of the Capital District, in the Housatonic Valley. The model developed two single bands of heavy snow. One was north of the of the surface low center, as it tracked up the eastern seaboard, the second band into the Hudson Valley, including the Capital District, by early afternoon.

Actual snowfall rates were 5 to 8 cm/hr and visibilities were under a half a kilometer at times. The heavy snow band continued to migrate further west, into the Helderbergs and Catskills by late afternoon. At the same time, a large dry slot worked up from the southeast which diminished the snow from Albany and points east, to very light amounts by late in the day. Most of the models had indicated that the greatest rate of snowfall would take place around 0000 UTC Sunday, associated with greatest upward vertical motion. Observed snowfalls had greatest rates of fall 3-6 hours earlier. Very little snow fell overnight across the CWA as the dry slot remained in place. This was again in slight contradiction to the ETA which indicated several more cm of snow would fall. Then, just as the Winter Storm Warning was lifted by the National Weather Service, another burst of snow fell shortly after dawn. That reduced visibilities markedly once more. This burst was confined to mostly the Capital District and only produced 3cm of additional snow. However, that was enough to make roads very slippery once more. The final burst of snow appeared to be the result of low level convergence that often occurs as the winds turning west in the Mohawk Valley, converge with the still northerly winds found in the Hudson Valley. This usually takes place as the surface low passes to the east of Albany.

Winter ER/NCEP Experiment and NCEP Update

Dr. Louis W. Uccellini
NOAA/NWS, National Center for Environmental Prediction, Camp Springs, MD

Abstract to be distributed during session

The Open RPG Deployment, New Capabilities, Hotline Operations and the ROC

Daryl L. Covey
NOAA/NWS, Radar Operations Center, Norman, Oklahoma.

Materials to be distributed during session.

The NCEP Short Range Ensemble Forecast (SREF) System: Operational Applications

Steve Tracton
NOAA/NWS, National Centers for Environmental Prediction, Environmental Modeling Center
Camp Springs, Maryland

The Environmental Modeling Center (EMC) of the National Centers for Environmental Prediction (NCEP) has developed and is now running routinely in real time a Short Range Ensemble Forecast (SREF) system. The current system currently consists of 10 members composed of five members from both the Eta and Regional Spectral Model (RSM) with 48 km horizontal resolution. Initial state perturbations are provided by "breeding", as for the NCEP global ensemble system, but in the context of the respective regional models.

Perturbations to physics, as well as initial conditions, and inclusion of additional models are anticipated. Advances currently incorporated and those expected in the near future with regard to the SREF system and derived products will be discussed. Particular emphasis will be upon how these advances relate to the skill and utility of the ensemble based probabilities in forecasting significant weather events, such as winter storms and associated sensible weather and warm season mesoscale convective systems.

Effective use of Regional Ensemble data in Forecasting Winter Storm

Richard H. Grumm and Robert Hart
NOAA/NWS, Weather Forecast Office, State College, Pennsylvania

Multi-model ensembles provide weather forecasters with a wide range of potential solutions. These data, if properly displayed, provided a more probabilistic approach to forecasting relative to the currently employed single model deterministic approaches. Model diagnostics, such as quasi-geostrophic and frontogenetic computations are often used to validate or improve upon a single forecast from a single model.

With multi-model ensembles, diagnostics on each forecast member becomes prohibitive, as does the examination of individual forecasts from each ensemble member. Therefore, new display concepts must be employed to maximize the utility of ensembles. In this paper we present several display concepts to assist forecasters in using ensemble forecast data in an operational setting. Traditional spaghetti plots of one or more significant contours are displayed along with the dispersion of all members about the ensemble mean. Consensus forecasts are provided for fields such has mean sea-level pressure, heights, and temperatures. Fields such as quantitative precipitation and the 850 hPa zero Celsius isotherm are displayed using probabilistic methods. All of the displays are focused on providing the forecaster a quick means with which to assess the weather problem.

In this paper, an examination is made of the East Coast Winter storm of 3-4 December 2000. The deterministic forecasts from the operational NCEP stepped terrain (Eta) and the aviation run of the NCEP global spectral model (AVN) are compared to forecasts from the NCEP short range ensembles forecasts (SREF). Using the display concepts outlined above, it will be shown how these data can be used to improve forecasts of winter storms.

An Evaluation of Eta, Avn and Limited Ensemble Qpf's over the Northern Mid-Atlantic Region During the 2000-2001 Winter Season

Michael S. Evans
NOAA/NWS, Weather Forecast Office, State College, Pennsylvania

Model generated quantitative precipitation forecasts (QPFs) continue to be a widely used tool for snowfall forecasting. In this study, the accuracy of this tool is examined by evaluating a variety of model QPFs for 6 winter storm events that occurred over the northern mid-Atlantic region during the 2000-2001 winter season, plus an additional event that occurred on January 25 ^th , 2000. For each event, AVN and Eta forecasts are evaluated at cycle times 12 and 24 hours prior to the onset of snow. An AVN/Eta consensus forecast is also computed at 12 and 24 hours prior to the onset of snow. 12-hours prior to the onset of snow, Lagged average forecasts (LAFs) are computed for each model, along with a d(pgrog)/dt QPF and a "super consensus", or average of all of the other 12-hour forecasts.

The winter storm event of December 30 ^th, 2000 is presented showing how the QPFs from the Eta and AVN, plus the other derived forecasts, varied as the storm approached. In this event, it is shown that the AVN model consistently produced the best forecast for the northern mid-Atlantic region. The Eta model initially forecast the storm to track too far to the west. As a result, the QPF was also forecast to extend too far to the west. Successive Eta forecasts trended toward the AVN model solution, but never provided as accurate a forecast as the AVN. Examples of LAFs, dprog/dt and multi-model consensus QPFs are shown for this event.

Model QPF is evaluated for all 7 storms in the study at two locations: the location where the heaviest snow was ultimately observed, and at the Middletown, Pennsylvania (KMDT) observation point. At both locations, an evaluation of the total performance of each model (including the ensemble and other derived forecasts) is presented. For each model, a normalized error is computed for each storm, by subtracting the observed precipitation from the model QPF, then dividing by the observed precipitation. The overall accuracy for each model is then determined by summing the absolute value of the normalized errors over all 7 events and dividing by 7. Model biases are shown by performing the same calculation, except without applying the absolute value to the individual errors. Errors are shown for each forecast at KMDT and at the location where the heaviest snow was observed.
At locations where the heaviest snowfall was observed, it is found that the Eta model usually produced heavier, more accurate precipitation forecasts than the AVN model. Despite this, the Eta's QPF was still often too low at those locations. Overall, the Eta 12-hour forecasts produced the best forecasts, followed by the Eta LAF and the 12-hour Eta/AVN consensus. For the KMDT observation point, it is found that the best overall forecast was the 12-hour Eta / AVN consensus forecast, followed by the "super consensus" forecast and the Eta LAF. The Eta QPF forecasts were shown to vary more from run to run than the AVN forecasts. The Eta QPF forecasts appeared to exhibit a wet bias, while no such bias was indicated for the Aviation model.

Model and Observational Evaluation of the Impact of PNA
on the Winter Climate of the Northeast United States

Michael Notaro and Wei-Chyung Wang
Atmospheric Sciences Research Center, State University of New York at Albany, Albany, New York

The large-scale Pacific North American pattern (PNA) significantly influences the regional-scale winter climate of the Northeastern United States, as illustrated here in this observational and model study. Positive phase of the PNA is associated with a deeper trough in eastern United States, with colder, drier conditions across the Northeast. For every state in the Northeast, there is a negative correlation between PNA and both temperature and precipitation in December. For New York, the correlation for December 1958-2000 is -0.54 between PNA and state-mean temperature, significant at the 0.99 level. The correlation between PNA and New York's mean precipitation is -0.43, also significant at the 0.99 level. The large-scale PNA pattern can influence the frequency of frontal passages through New York by determining the position of the upper-level jet. Based on 342 frontal passages through New York during the period of November-March 1991-2001, as identified using NCEP surface analyses, frontal passages are most frequent when -1PNA0. A monthly average of 7.1 frontal passages occurs during positive PNA and 9.4 during negative PNA, while the correlation between PNA and frontal frequency is -0.40, significant at the 0.99 level. Observations indicate that PNA can therefore influence both the large-scale and regional-scale across the Northeast.

The influence of PNA on the regional winter climate of the Northeast was also evaluated using the SUNYA regional climate model (ReCM), which is dynamically based on the hydrostatic version of PSU/NCAR MM5 with the inclusion of a land surface model and GCM parameterization for clouds and radiation (Dudek et al., 1996; Gong and Wang, 2000; Wang et al., 2000). Ten Decembers were simulated during the 1980s and 1990s, five with the most positive PNA and five with the most negative PNA. The five positive PNA simulations are cooler across the Northeast, up to 4-5°C cooler in New York, with more frequent occurrence of northerly flow. Correlations between monthly maximum temperature and PNA reach a maximum exceeding -0.80 in the inner domain. Precipitation is typically heavier across the Northeast during negative PNA simulations, especially east of the Great Lakes. The negative correlations between precipitation and PNA exceed -0.60 across New York and Pennsylvania. The large-scale PNA pattern also influences the tracks of synoptic systems. There is a more frequent track to the northeast for both cyclones and anticyclones during negative PNA months. A noted difference is that anticyclones typically strengthen along their tracks during positive PNA but weaken during negative PNA. Clearly, PNA can influence the synoptic and mesoscale details of the winter climate across the Northeast.

Death by....1000 Cuts?

Lance F. Bosart

Department of Earth and Atmospheric Sciences, University at Albany State University of New York, Albany, New York

N umerous minor precipitation events (< 5 mm) can occur over the northeastern US during the cool season in association with the passage of weak synoptic and subsynoptic disturbances aloft. Many of these minor events can turn into big nuisance events when frozen precipitation occurs, especially during the morning and evening rush hour. Additionally, the forecasting of minor precipitation events (amount, onset time, duration, ending time) presents a big challenge because model uncertainty is apt to be especially large in these situations.

This talk will highlight several examples of troublesome minor precipitation events. A dynamic tropopause (DT)/potential vorticity (PV) perspective will be used for this purpose. It will be demonstrated that it is possible to use DT maps to locate and track small-scale PV anomalies that can be associated with minor precipitation events.

An Analysis of Montreal's Record-Breaking Heavy Rainfall Event of 8-9 November 1996, and a Comparison with Its Best Analogue

Dorothy Durnford and John R. Gyakum
Department of Atmospheric and Oceanic Sciences
McGill University, Montreal

Montreal's heavy precipitation event of 8-9 November 1996 was noteworthy for its all-time record-breaking 24-hour accumulation of 134.0 mm. This event caused heavy flooding in the Montreal area, as well as in Vermont and New Hampshire. New York experienced both flash flooding and main stem river flooding. Roads in these areas were closed, bridges damaged or washed out and evacuations conducted. In Maine, high winds resulted in the loss of power for 10,000 utility customers.
Precipitation forecasting is considered a particular challenge (Businger et al. 1990). Model probability of measurable precipitation forecasts are not only less reliable, but are also improving more slowly, than model 500-hPa geopotential height field predictions (Roebber and Bosart 1998). Although models can forecast the atmospheric flow features associated with heavy precipitation, they rarely forecast the heavy precipitation itself (Junker et al. 1989). The skill of the forecaster is strongly related to that of the model guidance (Fritsch et al. 1998). Manual quantitative precipitation forecast threat scores decrease monotonically with increasing threshold values (Roebber and Bosart 1998).

Pattern recognition is used in subjective forecasting (Funk 1991). This involves the forecaster recognizing various synoptic and mesoscale patterns that produce heavy to excessive rainfall. In this project, we analyze the atmospheric state of the 1996 event. We search for analogues of this event's anomalous (with respect to 30-year (1967-1996) monthly climatological fields) sea level pressure and 1000-500 hPa thickness fields. The search is conducted for the months of September through February, for 1963-1996, over the area encompassed by 110 ^o-50^o W and 25^o-65^o N. The 1996 event's fields are compared to those of its best mass analogue of 10-11 November 1977. This analogue, despite an excellent average anomaly correlation value of 0.81, is characterized by a storm-total precipitation value, averaged over five randomly-chosen Montreal-area stations, of 20.2 mm, while the 1996 event's average value was a far more significant 74.8 mm. This study emphasizes differences between the two cases' highly similar (by definition) atmospheric states, in order to determine which of the 1996 event's features are key to its production of far greater precipitation values. This provides a deeper understanding of the heavy precipitation event than an analysis just of its fields, and is more precise and informative for the forecaster than pattern recognition alone.

Two of the most basic key differences between the 1996 event and its analogue were found to be the Montreal-area precipitable water content and synoptic-scale ascent values. Both before and during the event, the 1996 event had more water vapour available to be precipitated out than the analogue. During the event, this greater amount of water vapor was accompanied by greater 850- and 500-hPa synoptic-scale ascent. Stronger synoptic-scale ascent operating on a greater amount of moisture will produce more precipitation. However, it is unlikely that these synoptic-scale variances are responsible for the large discrepancy between the two average storm-total precipitation values. The third key difference is the mesoscale ascent associated with the surface frontal region. A mesolow was present in the Montreal region for the 1996 event but not for the analogue, and the former event's surface front was stronger than was that of the latter up until the end of the event. The 1996 event's frontal region is, thus, presumably associated with larger mesoscale ascent values than that of the analogue. Furthermore, high precipitable water content values, synoptic-scale ascent and mesoscale ascent all coincided in the Montreal region for a far longer period for the 1996 event than for the analogue. Thus, the duration of the combination of the first three key features, or persistence, constitutes a fourth key feature. The last two key features are responsible for the great discrepancy between the two precipitation values. Stability is not seen to be a key feature.

The differences between the two cases' values for the first two key features (precipitable water content and synoptic-scale ascent) are the result of different pre-event southerly geostrophic flows. The 1996 event's persistently strong southerly geostrophic flow resulted in a tongue of tropical field values penetrating northwards into the Montreal area before the start and through the middle of the event. The tropical nature of the air mass was exhibited, in part, by high precipitable water content values and by a high, potentially warm dynamic tropopause. This latter feature contributed to significant dynamic tropopause pressure and potential temperature gradients, which are indicative of a highly baroclinic atmosphere (Bosart and Lackmann 1995). A weaker southerly geostrophic flow into the Montreal region during the analogue's pre-event period yielded an atmosphere characterized by lower precipitable water content values and smaller dynamic tropopause gradients.
Although the analogue's mid-event geostrophic flow was comparable in strength, though still not in southwards extent, to that of the 1996 event, the northward-penetrating tongue of tropical field values generated by this flow extended either insufficiently north or west for the analogue's field values in the Montreal region to match those of the 1996 event. The difference between the two cases' values for the fourth key feature (persistence) reflects differing evolution rates. The analogue's mass fields evolved more quickly than did those of the 1996 event, so that the combination of the first three key features was not sustained for as long a period.

A Northeast Snowstorm Impact Scale

Paul J. Kocin
The Weather Channel Inc, Atlanta, Georgia
Louis. W. Uccellini
NOAA/NWS, National Center for Environmental Prediction, Camp Springs, Maryland

As part of the development of an AMS monograph on Northeast snowstorms, a Snowstorm Impact Scale has been developed to help provide and communicate a measure of the impact of a given storm to the public. The scale, valued from 0 to 5, is similar to both the Fujita (1971) and Saffir-Simpson (1977) scales, but differs in that the primary focus is on area and population affected by heavy snowfall, rather than the potential for damage implicit in the other scales.

Thirty crippling Northeast snowstorms between 1950 and 2000 form the basis for the scale. These storms have been analyzed to determine the mean areas and populations affected by various intervals of snowfall amounts utilizing a Geographic Information System (GIS). A premise of the scale is that area and population are equally weighted and no attempt is made to contrast storm intensity, duration, temperature or winds, all important contributors to storm impact, but difficult to assess objectively. Since all 30 snowstorms had significant impact on the Northeast United States, the scale is devised to measure impact relative to this 30-case sample.
The scale measures the areas and populations affected by heavy snowfall at increments of 10 inches (25 cm) relative to the mean area and population of the 30-case sample. It also adds weight to the areal extent and populations affected by higher snowfall increments, reflecting a greater impact for 20-inch, 30-inch and 40-inch snowfalls. The scale takes the following form:

Scale = ^x_n [n * (A_n/A_mean + P_n/P_mean)]

where n represents the lower value of the snowfall contour intervals divided by 10. A _mean and P_mean represent the mean area and population for the 30 cases. Populations are normalized to 1990 census values. A_n and P_n are the areas and populations within the various snow intervals and are estimated for areas of snowfall exceeding 10 inches (25 cm), 20 inches (50 cm), 30 inches (75 cm) and 40 inches (100 cm). Values derived from the above equation for the 30 individual cases range from 0.66 to 10.9. The largest values are found for the March 1993 Superstorm (10.9) and the January 1996 "Blizzard of '96" (8.3), the February 1983 "Megalopolitan" snowstorm (5.0) and the February 1978 New England Snowstorm (4.2). These values are grouped over several intervals, starting with a value of 0.5.

Category	Scale Values
0	.500-1.250
1	1.250-1.749
2	1.750-2.499
3	2.500-3.499
4	3.500-5.999
5	>6.000

Most of the 30 cases exhibit values in Categories 2 and 3, ranging from 1.75 to 3.5. The average value is approximately 3. Any snowstorm with a value greater than 6 is considered a "Category 5". The scale is also applied to 5 historical cases. Two storms, the Blizzard of 1888 and the Appalachian Storm of 1950, both scored as a Category 5, indicative of the impact of either/both heavy snowfall amounts over large areas or large populations affected by very heavy snowfall amounts. The February 1899 snowstorm ranks a Category 4 with a large area of snowfall, but smaller areas of snowfall greater than 20 to 30 inches (50 to 75 cm) than the prior 2 cases. The Knickerbocker snowstorm of 1922 and New York City's "Big Snow" of 1947 both rank lower than the other cases as a Category 3 because the total areas were relatively small. However, they still scored relatively highly because the populations affected were large as heavy snow was focused over the metropolitan areas of Washington D.C. and New York City.

While the scale is to be used as a way to communicate impact of a given storm to the general public, it is hoped that as our ability to forecast snowfall amounts improves, this scale could be applied in a predictive sense.

An Early Winter Lake-Effect Snowstorm over Southern Ontario

Q. Liu and G.W.K. Moore
Department of Physics, University of Toronto,
Toronto, Ontario, Canada

D. Hudak
Meteorological Service of Canada, King City, Ontario, Canada

An early winter lake-effect snowstorm that occurred in late October 1992 is studied. Several unique characteristics of this event are of particular interest. As it was the first snowfall of the year in Southern Ontario, the comparison of before and after satellite imagery provides an excellent indicator of the spatial distribution of snowfall associated with this event. In addition, data from a buoy in Lake Huron provides information on the fluxes of heat and moisture from the lake to the atmosphere that contributed to the development of this storm.

Analysis of the synoptic data indicates that this event is a typical lake-effect snowstorm in Southern Ontario. It developed in the cold northwesterly flow established after the passage of a strong cold front through the region. This is confirmed by the buoy data that shows that there was horizontal advection of cold air over the relatively warm waters of Lake Huron. Associated with this horizontal advection were intense lake-atmosphere fluxes of heat and moisture. Radar data from the King City site indicates that the snowfall associated with this event was organized into both multiple and single snowbands. Through the analysis of above characteristics, the spatial and temporal variation of this lake-effect snowstorm is extracted. It was found that buoy data was especially useful to determine the heat and moisture fluxes, which are important lake-effect snowfall indexes.

The Importance of Snow Microphysics for
Large Snowfalls

Jeff S. Waldstreicher
NOAA/NWS Eastern Region Headquarters, Scientific Services Division
Bohemia, New York

Forecasters have long recognized that snowstorms often appeared to have different characteristics regarding predominant snowflake sizes and shapes, as well as density (e.g., snow to water ratios), and that these differences appeared to be related in large part to variations in temperature. Previous research has shown that the production of dendrites, the largest and most efficient accumulators of the various snow crystal forms, tends to maximize at cloud temperatures around -15C. Other snow growth processes that are most efficient at other temperatures have also been identified. It is becoming clear that consideration of these different snow microphysical processes is an important part of snowfall forecasting.

A study was conducted to attempt to quantify the importance of dendrite production for large snowfalls across central New York and northeast Pennsylvania. Snowstorms of greater than 7 inches (warning criteria) and 4-7 inches (advisory criteria) during the last 4 winter seasons (1997-98 thru 2000-01) were examined and compared to hourly Eta model soundings for 8 locations: SYR; UCA; BGM; ELM; ITH; WHI; AVP; and IPT. Time height cross sections of temperature and omega were analyzed for cases when an upward vertical motion maximum of at least 10 b sec ^-1 intersected the region of favored dendrite temperatures (-12 to -18^oC). The results of this analysis will be presented. The characteristics of "missed events" (warning snowfall criteria without the signature), and "false alarm" (signature but only advisory criteria snowfall) cases were also investigated. In a number of these cases, the role of other snow growth processes were apparent.

A Study of Tornadic Versus Non-Tornadic Thunderstorms in Central Pennsylvania on 2 June 1998

Ernie J. Ostuno
NOAA/NWS, Weather Forecast Office
Grand Rapids, Michigan

A widespread severe weather outbreak struck Pennsylvania on 2 June 1998. Three lines of supercell thunderstorms produced a variety of severe weather with the main threat being tornadoes and large hail. This study examines 26 mesocyclones that moved across central Pennsylvania during this event. Eleven of the mesocyclones produced a total of 14 tornadoes. The trends in rotational shear were graphed for the 26 storms, each one of which reached the local National Weather Service office tornado warning criteria of 25 knots or more of rotational shear. In addition to the analysis of Doppler radar velocity data, the base and composite reflectivity data was examined to determine if the individual storms showed some of the classic features associated with tornadic supercells such as hook echoes, weak echo regions (WERs) and bounded weak echo regions (BWERs). It was found that, on average, stronger mesocyclones were more likely to produce tornadoes.

Storms exhibiting classic radar reflectivity signatures were also more likely to produce tornadoes than storms lacking these features. However, there were notable exceptions. One of the strongest mesocyclones observed this day did not produce a tornado. In fact, an extensive storm survey failed to find any evidence of damaging winds at ground level along the path of this storm, despite radar-derived rotational shear values greater than 55 knots at cloud level.

An Updated Look at Some Severe Weather Forecast Parameters

Kenneth D. LaPenta and George J. Maglaras
NOAA/National Weather Service, Albany, New York

John Center
NOAA/National Weather Service, Wilmington, Ohio

Sarah A. Munafo and Charles J. Alonge
Department of Earth and Atmospheric Sciences, University at Albany
State University of New York, Albany, New York

Forecasters routinely make subjective assessments of convective potential for their forecast area based on the values of various atmospheric parameters and indices. If convection is possible, forecasters must decide whether it will be severe or non severe; and if severe thunderstorms are possible, they must determine if the primary threat will be large hail, damaging winds, tornadoes, or all three. The specific parameter values which influence certain decisions may vary from person to person depending on a forecaster's geographic location, experience, and scientific understanding of the physical processes associated with thunderstorm development and evolution. Because of the subjective nature of the decision making process, the results may not be consistent. In previous papers by several of the authors, regression equations, based on severe weather parameters taken from 148 cases, were developed to provide objective guidance on forecasting tornadic and non-tornadic severe thunderstorms, as well as hail size and severity.

This study uses the data set developed in the previous work to examine additional forecast indices not previously available, and to re-examine in more detail several parameters previously studied. In the earlier work, storm-relative helicity was the highest correlated parameter with severe weather as defined in the study. During recent years, two different approaches, the helicity perspective and shear perspective, have evolved which are used to explain supercell dynamics. A large number of shear parameters were evaluated to see if they were better correlated with severe weather than helicity. Surface based Convective Available Potential Energy (CAPE) was also an important forecast parameter in the previous work. In this study, mean parcel CAPE and CAPE normalized for storm depth were evaluated. Downward Convective Available Potential Energy (DCAPE), an estimate of the kinetic energy available to a downdraft parcel due to negative buoyancy, may be an important parameter in assessing the potential for damaging straight line winds and for determining low-level supercell structure. Its relationship to severe weather was examined. The utility of a number of other forecast parameters including atmospheric lapse rates through various layers, storm-relative wind flow and convective condensation levels will be presented.

Warm Season Closed Lows in the Northeastern United States

Thomas A. Wasula and Kenneth D. LaPenta
NOAA/NWS, Weather Forecast Office, Albany, New York

Significant severe weather and flash flooding have occurred across the northeastern U. S. during the past several decades in response to warm season closed lows moving across eastern North America. Warm season closed lows have a critical impact on the local sensible weather, based on the generic location, track and duration of these large-scale features.

A subjective weather analysis was performed to create a warm season (May 15 ^th to September 15^th) climatology of closed lows from 1980-2000 based on daily 500 hPa and surface analyses across the latitude-longitude domain of 36- 48N and 65-88W. This area is primarily east of the Mississippi River, excluding the Southeast, and south of Hudson Bay. A closed low was defined by the presence of at least one closed 500 hPa isoheight. There are six favorable tracks or categories of warm season closed lows. They include a diverse range of patterns, such as closed lows moving southeastward out of Hudson Bay toward the Northeast (northwest flow), systems moving toward the Northeast from the west or southwest, and systems of tropical origin. One hundred and seventeen cases were identified, with the most cases in the northwest flow category. NCEP/NCAR reanalysis data is currently being used to create composites for each category. The first part of the talk will examine this climatology and discuss some of the preliminary results of the northwest flow composites.

A more detailed synoptic and mesoscale examination of the 13-17 July 2000 northwest flow event will show the impact some of these closed lows can have on the Northeast. Special emphasis will be placed on the numerous flash floods and severe thunderstorms that occurred with this particular system across the Albany National Weather Service's County Warning Area. We will use ETA model grids, surface observations, upper-air data including soundings, Doppler radar data, and vertical cross-sections to analyze the environment that helped generate the severe weather and flooding in this case study.

Applying D3D in an Operational Environment

John W. Cannon and James C. Hayes
NOAA/NWS, Weather Forecast Office
Gray, Maine
Joshua Watson
NOAA/NWS, Eastern Region Headquarters
Bohemia, New York

The Forecast System Lab's (FSL) Display 3-Dimensional workstation application (D3D) for Linux allows forecasters to view model output in a multi-dimensional interactive display. D3D, based on the University of Wisconsin's Vis5D visualization software, supplements Advanced Weather Interactive Processing System (AWIPS) 2-D display capabilities available to forecasters. Large volumes of data can be rapidly assimilated with an increased temporal and multi-dimensional understanding of how meteorological processes interact. The FSL posts a detailed project guide at http://d3d.fsl.noaa.gov.

During Spring 2001, the NWS Forecast Office in Gray, Maine worked in cooperation with NWS Eastern Region Headquarters and FSL as a D3D workstation test site. The goal of this project is to determine operational strengths and limitations of the workstation through input from the hydrometeorological staff. Valuable insight from routine and severe weather operations could then be forwarded to FSL as input for future software enhancements and training. To accomplish this goal, forecasters were provided individualized instruction, locally-created job sheets, and access to the D3D User's Guide for training purposes. In addition, a questionnaire, designed to assess information on system performance and the role of 3-D visualization techniques in an operational environment, was completed by forecasters after using D3D in an operational forecast setting.

This presentation will demonstrate D3D capabilities and show findings submitted to FSL for an improved D3D visualization package. The goal to show there is an advantage of a more thorough diagnosis of model forecasts using D3D and D2D in the forecast process. In the future, this technology can be installed NWS-wide as an integral component of the AWIPS.

Mesoscale Sensor Network for DoD and Civil Emergency Applications

Dave Sautter
Yankee Environmental Systems Inc., Turners Falls, Massachusetts

The US Army solicited (under the title of "Phase I Smart Sensor Web/Weather Web program") for a rapidly deployable network of self-powered wireless weather stations that tie back to a central, more sophisticated internet-connected node capable of also making vertical wind profile measurements. The ideal system would be a miniaturized ASOS/AWOS system about the size of a soda can. It would be rugged and light enough to be carried in the soldier's backpack such that they could just set it down and walk away. Once deployed, it would begin reporting met data automatically to a web-driven primary sensor. Deployed over a suitably wide area of, for example, a division (or in civil terms, a small state), it would provide the command with a detailed picture of the mesoscale environment. It could also provide certain tactical information such as enemy troop movements via an imaging capability.

Yankee Environmental Systems, Inc. (YES Inc.) is developing a response to the Army which utilizes the Lincoln Lab-initiated "Western Massachusetts Weather Web Test Bed" and both current and future technology sensor platforms. The primary sensor package includes a ceilometer, temperature, dew point, pressure, and wind sensor, plus Yankee's Total Sky Imager (TSI-880) for hemispheric sky images, sky coverage, and surveillance information. In addition, algorithms are being developed to produce cloud-level winds aloft from dynamic cloud motion analysis from the TSI. Sub-node sensors, essentially disposable sensors in a can, would utilize seamless RF communications back to the primary sensor, would be completely self-configuring with respect to TCP/IP protocols, and would self-correct for true North. The RF technology would have narrowband characteristics, similar to that now available in current radiosondes, and noted for its beacon-like signal. Power management will likely use wind or solar power to augment initial battery supply.

A web-driven sensor interface was demonstrated between Ft. Benning, Georgia and White Sands Missile Range. TSI-880, temperature, dew point, and pressure sensors readouts were visible via internet. Sub-node sensors are in the initial design stage, and the RF technology and data ingest system have been chosen for linking sub-nodes with the primary node, all using commercial off-the-shelf technologies. A graphic interface similar to that used currently by Air Force regional weather centers, the Airfield Observing System, will be designed for sensor(s) display.
Applications are envisioned for the civil emergency-type environment, as well as the deployed Army division. Sub-node sensors could be rapidly deployed around a forest fire, nuclear accident, or chemical/air pollution dispersion problem, to best characterize local terrain and other effects on surface weather. The primary node would in turn feed data to decision makers via internet and could be ingested into and refine high resolution weather models. This easily deployed sensor network may also have various research applications.

Beyond IFPS , Empowering Weather Information Consumers

Richard J. Westergard
NOAA/NWS Weather Forecast Office, Albany, New York

Over five years, in a local scale COMET partners project, followed by a national scale COMET Cooperative Project, we used focus groups to assess weather information consumer needs, as opposed to their expressed wants. We came to the conclusion that the information should be molded to conform to, and better support, the way consumers use weather information. If we can accomplish that, the information will be of far greater practical use.

IFPS is taking Weather Service forecasters into a new era of digital databases with far greater precision than in the past. That precision taken alone, however, will make our forecasts seem even less accurate, since very minor variations in the actual versus forecast weather will create "wrong" forecasts in the minds of people who are intense users of weather information. In order to offer the weather information consumer a better product, we need to provide consumers with a more scientific basis for second guessing forecasts (as they now do and will continue to do, in any case). We can accomplish this by giving them access to more of the rich information field available to trained meteorologists.

We have a proposal, which met with a favorable response from our focus groups. We would provide these "proactive" consumers with forecaster chosen computer model forecasts, in formats similar to the digital database output forecasts. Those computer model forecasts could be used by proactive consumers to assess weather risks, given their specific plans, and to gain a sense of forecaster confidence.
Beyond that proposal, we would also alter the form of all text forecast products to focus on sensible elements over time, rather than the current time period based formats.

A Climatology of Cold Season Banded Precipitation in the Northeast United States

David Novak, Lance F. Bosart, Daniel Keyser
University at Albany, State University of New York
Albany, New York

Jeff S. Waldstreicher
NOAA/NWS, National Weather Service Eastern Region, Scientific Services Division,
Bohemia, New York

A climatology of banded precipitation events in the northeast U.S. during the cold season (October through April) is presented. Precipitation systems in the northeast U.S. which exhibited greater than 1.00" of rainfall, or 0.50" liquid equivalent were identified as cases for study using the Unified Precipitation Dataset (UPD). Composite radar data from these cases were viewed to develop a band classification scheme. This scheme was then applied to cases from November 1996 through April 2001. Out of the 112 cases identified during this period, 89 cases had complete radar coverage. Examination of these 89 cases revealed that 36 exhibited single banded structure at least once during their evolution, 29 multibanded structure, 30 narrow cold-frontal structure, 32 transitory or undefined structure, and 13 exhibited no defined banding. Note that many cases had more than one type of banded structure during their duration. Further investigation of the single band events highlighted banded structure in the comma head portion of storms, with nearly 70% of the bands exhibiting some portion of their length in the northwest quadrant of the surface cyclone.

Large-Scale Circulation Anomaly Indices in Relation to
Cool-Season Precipitation Events in the Northeastern
United States

David Groenert, Lance F. Bosart, and Daniel Keyser
Department of Earth and Atmospheric Sciences, University at Albany
State University of New York, Albany, New York

Richard H. Grumm
NOAA/NWS, Weather Service Office, State College, Pennsylvania

Interest in relationships between large-scale circulation anomaly indices, such as the North Atlantic Oscillation (NAO), Pacific-North American (PNA) pattern, and Southern Oscillation Index (SOI), and regional to local precipitation anomalies over the northeastern United States calls for compilation of daily time series of these large-scale circulation anomaly indices. The rationale for using daily as opposed to monthly values of large-scale circulation anomaly indices is to better understand the evolution of individual cyclone structure and life cycles in relation to changing large-scale circulation regimes.

The first goal of this project is to calculate a daily NAO index from 1948 to the present. Future goals include calculation of daily time series for the PNA and SOI. Upon completing each time series, relationships will be determined between these indices and planetary-scale flow signatures crucial to cool season precipitation events associated with extratropical cyclones in the northeastern United
States.

A calculation of the daily NAO index from 1948 to the present is shown. The differences of normalized sea-level pressure (SLP) and 500 hPa heights between Stykkisholmur, Iceland (65°05'N, 22°44'W) and Ponta Delgada, Azores (37°45'N, 25°40'W) are used to determine the daily NAO index. These daily values are averaged over one-month periods and compared to monthly NAO values as compiled by Hurrell (2001). Results will be shown in the form of an NAO timeline for comparison to corresponding large-scale SLP and 500 hPa plots. Future research will focus on the downscale effects of large-scale circulation anomalies on the occurrence of cutoff lows and mesoscale substructures in extratropical cyclones over the Northeast.

Characteristics of Cool Season Cutoff Lows in the Northeastern United States: Four Northwest-Flow Events in Northern New York State and Northern Vermont

Daniel P. St. Jean
NOAA, National Weather Service Forecast Office, Burlington, Vermont

Cutoff 500 hPa cyclones over the northeastern United States are often associated with production of heavy precipitation (e.g., many Nor'easters exhibit a cutoff 500 hPa cyclone center during their lifetimes), yet many cutoff cyclones do not generate significant precipitation. Two heavy snowfall-producing cutoff cyclones in the autumn of 1999 over northern New York State and northern Vermont provided a most recent operational impetus for studying the characteristics of these cyclones. The end result of this research will be to produce ingredients-based conceptual models and operational methodologies for the purpose of improved prediction of the precipitation patterns produced by these cyclones in the complex terrain of the northeastern United States.

Potential cutoff cyclone cases for inclusion in this study were limited to events occurring with prevailing northwesterly upper-level flow, which excluded any cases involving rapid-genesis coastal cyclones (i.e, Nor'easters). Northwest-flow scenarios generally produce a significant low-level flow orthogonal to the Green Mountains and Adirondacks, favorable for the generation or enhancement of heavy precipitation by orographic lift. Four cutoff 500 hPa cyclone scenarios were examined in this study: two events which produced heavy snowfall northern Vermont and northern New York State; and two events which had been forecast to produce heavy snowfall, yet significant precipitation failed to occur. In addition to analyzing the synoptic-scale and mesoscale structure of these events, this study has also examined the accuracy of forecast warnings and watches generated by WFO Burlington Vermont for each of these events.

NCEP/NCAR Reanalysis data were used in determining the synoptic-scale characteristics of each of the four cases, supplemented with ETA model BUFR sounding data in order to interrogate the mesoscale structure of each event. Initial findings from this study suggest several meteorological factors significant to the development of heavy precipitation from this type of cutoff cyclone: (a) the low-level moisture profile; (b) the strength and orientation of the low-level wind with respect to the orography; (c) the low-level static stability profile. Despite the coarse resolution of the Reanalysis data, some encouraging synoptic-scale signatures have already been borne out of each of the events. These initial results and the direction for future investigation will be presented. This research was conducted as a small subset of the ongoing SUNY-Albany 500 hPa cutoff cyclone climatology. This research and the SUNY-Albany research are a portion of the larger CSTAR project on improving the prediction of cool season heavy precipitation events over the northeastern United States.

A Climatology of 500 hPa Cutoff Cyclones

Brandon Smith, Lance F. Bosart, and Daniel Keyser
Department of Earth and Atmospheric Sciences, University at Albany
State University of New York, Albany, New York

Daniel P. St. Jean
National Weather Service Forecast Office, Burlington, Vermont

Cutoff cyclones are associated with many significant forecasting problems in the northeastern United States. Given the complex terrain in the Northeast, the precipitation distribution associated with slow-moving cutoff cyclones is often a challenge to predict. As an initial step toward addressing this challenge, we present the results of a 46-year climatology of 500 hPa cutoff cyclones in order to map the spatial and temporal distributions of these events. This task is accomplished by using twice daily (0000 and 1200 UTC) 500 hPa gridded geopotential height analyses from the National Centers for Environmental Prediction /National Center for Atmospheric Research (NCEP/NCAR) reanalysis dataset.

Cutoff cyclones are identified through an objective analysis technique. For our purposes, a cutoff cyclone is defined as a minimum geopotential height center surrounded by at least one closed 30 m height contour. Cutoffs are identified and catalogued and cyclone tracks are determined to delineate favored areas for genesis/lysis and to locate "cutoff freeways." Frequency diagrams showing total number of cutoff cyclones and number of "cutoff 12 h periods" are presented for the Northern and Southern Hemispheres and for eastern North America. Also shown are maps of seasonal mean frequency and standard deviation of cutoff cyclone events for the same geographical regions.

In-progress and future work includes correlating favorable areas of cutoff events with significant large-scale circulation features such as mean jet stream positions and teleconnection indices such as the North Atlantic Oscillation. Our cutoff climatology will also be used in conjunction with the Unified Precipitation Dataset (UPD) to map precipitation distributions in cutoff cyclones over the northeastern United States.