|Local forecast by
An unusually intense rainfall and rapid snowmelt on January 18-19, 1996, resulted in
major flooding on rivers from Virginia to New York and Vermont, and in the upper
reaches of the Ohio River drainage. Above-average snow cover over the region and
high melt rates produced by above freezing temperatures, high relative humidity, and
high wind speeds produced large snowmelt contributions to the flood. Record flood
crests were set in Pennsylvania, New York, West Virginia, and Maryland. The
Delaware River crested at its highest stage since Hurricanes Connie/Diane in 1955.
The Hudson crested as its highest stage since 1977. The Ohio River at Pittsburgh, the
Susquehanna River at Wilkes-Barre, and the Potomac River at Little Falls, Virginia, all
crested at the highest stages since Hurricane Agnes in 1972.
DESCRIPTION OF FLOODING
A succession of snowstorms brought significant and widespread snow accumulation
across much of the northeastern United States by Wednesday, January 17, 1996. Snow
depths of 40-50 inches were common from central Pennsylvania into New York State.
From southern Pennsylvania, across Maryland and West Virginia into Virginia, snow
depths averaged 12 inches or more, with significantly higher amounts in the mountains.
A strong storm system then moved through the eastern United States January 18-19,
1996, bringing heavy precipitation as well as high temperatures, humidity, and winds
into the Ohio, Susquehanna, and mid-Atlantic drainages. Mean areal basin average
rainfall varied from 1.2 inches to slightly over 3 inches, with some individual gages
reporting over 4 inches. At most locations, the intense rain lasted for only about 6
hours. The heavy rains combined with significant snowmelt, and in some cases ice
jams, to produce major flooding in Pennsylvania, West Virginia, New York, Virginia,
Maryland, Vermont, Ohio, and New Jersey.
The magnitude of the flooding varied between basins, but it was a major event
throughout the area. More than 100,000 people were evacuated in the Wyoming
Valley region of the Susquehanna River Basin in Pennsylvania. The entire town of
Marlinton, West Virginia (1,100 people) on the Greenbrier River was evacuated.
Many evacuations also took place in the Allegheny, Susquehanna, and Finger Lakes
drainages in Pennsylvania and New York. The Ohio River at Pittsburgh had its highest
crest (crest was affected by ice) since Hurricane Agnes in 1972 (Table 1). On the
Susquehanna River at Wilkes-Barre, the crest of 34.5 feet on Saturday, January 20,
exceeded all past floods except for Hurricane Agnes in 1972. Record floods occurred
on Loyalsock Creek at Loyalsock, Pennsylvania; the Greenbrier River at Marlinton,
West Virginia; Wills Creek at Cumberland, Maryland; and Opequon Creek at
Martinsburg, West Virginia. Record crests were also observed on the Lower
Conemaugh River, Lower Mahoning River, West Branch Clarion River, Aughwick
Creek, Towanda Creek, Tunkhannock Creek, and Frankstown Branch (Little Juniata),
all in Pennsylvania.
Table 1. Selected Northeast Flood Crests - January 1996
As flood waters began to recede, there were reports of flood-related deaths. A total of
33 deaths was reported with 18 in Pennsylvania, 10 in New York, 1 in West Virginia,
3 in Virginia, and 1 in Vermont. Based on the widespread nature of the flooding and
comparisons to historical floods, it is estimated that total flood-related damages
exceeded $1.5 billion.
SNOW COVER AND SNOWMELT
The abnormally large snow cover that accumulated over the northeastern United States
in the weeks prior to the flood event was the result of a number of storms and below
normal temperatures that inhibited snow melt in the weeks prior to the flood. The
"Blizzard of 96," which occurred from January 6-8, 1996, was a major contributor to
the snow cover from Virginia, across Maryland, and through much of Pennsylvania
and New Jersey. Following the blizzard, two additional storms dumped up to 3 feet of
snow across portions of the Catskills of New York and northeast Pennsylvania adding
to near record November through December snow falls in this region. During the
period before the flood event, January 14-17, above freezing temperatures occurred
over most of the area. Significant snowmelt occurred west of the Appalachian
mountains. The snow cover was also partly depleted in parts of Virginia and eastern
Maryland prior to the flood event. Over portions of Pennsylvania, New York, and
New Jersey, this warming period caused some ripening (warming and preliminary
melting) of the snow cover, but there was little reduction in the water equivalent. An
estimate of the water equivalent of the snow cover over the primary flood area on the
morning of Thursday, January 18, using all available data, ranged from zero inches in
the far southeastern portion of the area to over 5 inches in areas from south-central
Pennsylvania to the Catskill Mountains of New York.
On the morning of Thursday, January 18, unprecedented snowmelt occurred just prior
to the onset of the heavy rains associated with warmer temperatures in the western
portion, and continued through the afternoon of Friday, January 19, in the eastern
portion of the region. For example, at Binghamton, New York, just over 3 inches of
water equivalent was reported before the event, and the snow was gone at the site by
mid-morning on Friday, January 19. Changes in water equivalent of 3-5 inches were
typical in open, non-wooded areas. Generally, any given area experienced between 18-
30 hours of high snowmelt rates. The rapid snowmelt was caused primarily by the
turbulent transfer of latent and sensible heat due to the high temperatures, dew points,
and wind speeds. By taking the difference between the estimated water equivalent
before and after the event, the estimated snowmelt contribution to the flooding, on a
watershed basis, was in the range of 2.5- 5 inches.
FORECASTS AND WARNING SERVICES
The Northeast Floods of January 1996 created challenges for all National Weather
Service (NWS) offices involved in issuing timely and accurate watches, advisories,
warnings, statements, and other public products. Hundreds of river forecasts were
issued under very complicated and unprecedented conditions.
Public forecasts and warning products issued by NWS offices in areas impacted by this
major flood event accurately highlighted the weather as it developed across their areas
of responsibility. Most products were well written, timely, and contained good call-to-
action statements. Outlook statements issued earlier in the week provided the public,
media, and public safety officials with information concerning flood potential. Since
state and local emergency managers from New York, Pennsylvania, New Jersey,
Maryland, West Virginia, and Virginia were still mobilized due to the recent blizzard,
it was fairly easy for most of them to prepare to handle a flood event; however, on-
going disaster mobilization taxes resources.
Flood watches were posted by the NEXRAD Weather Service Forecast Office
(NWSFO) Pittsburgh, Pennsylvania, for selected rivers of eastern Ohio and west
central Pennsylvania on Wednesday, January 17, and extended for all of east central
Ohio, western Pennsylvania, and northern West Virginia on Thursday, January 18.
Also on January 18, watches were posted by NWSFO Charleston, West Virginia, for a
significant portion of West Virginia and the New River Valley of southwest Virginia.
Additional watches were issued on Thursday afternoon, by NWSFOs in Buffalo and
Albany for central and eastern New York, and by NWSFO Baltimore,
Maryland/Washington, DC, for portions of Virginia, West Virginia, and Maryland.
NWSFO Philadelphia, Pennsylvania, issued a flood watch for extreme southeast
Pennsylvania and portions of New Jersey, Maryland, and Delaware on Thursday
afternoon, and for the remainder of eastern Pennsylvania and New Jersey early Friday
QUANTITATIVE PRECIPITATION FORECASTS (QPF)
Areal average 24-hour Quantitative Precipitation Forecasts (QPF) issued on Thursday,
January 18, by the NWSFOs in Pittsburgh, Pennsylvania, and Charleston, West
Virginia, ranged from 0.25-0.5, and 0.5-1 inch, respectively. QPFs released early
Friday morning, January 19, by NWSFOs Pittsburgh, Pennsylvania; Philadelphia,
Pennsylvania; Albany, New York; and Baltimore, Maryland/Washington, DC,
contained amounts ranging from 0.25-0.75 inch.
All QPF products were later updated after heavier precipitation amounts were observed
as the storm system moved into and across the region on Friday morning. Doppler
Weather Surveillance Radar (WSR-88D) precipitation estimates, the National Center
for Environmental Prediction (NCEP) numerical model, and the Hydrologic Prediction
Center (HPC) value-added QPFs underestimated the amount of rainfall as revealed by
the gage observations, although the HPC value-added QPFs represented a significant
improvement over the model QPFs.
Automated and manual data networks and equipment had performance problems during
this flood event; sporadic equipment problems and many critical outages of automatic
and manual gage networks were noted during this event that impacted data acquisition.
Many unheated tipping bucket precipitation gages became blocked with snow due to the
heavy snowfall from the previous weekend. River gages were affected by ice jams,
and there was water in several gage houses which washed out or contaminated the
All offices are equipped with a state-of-the-art WSR-88D, which provides an areal
assessment of rainfall amounts based on returned power processed through a series of
hydrometeorological algorithms. The radars at Binghamton, New York; State College,
Pennsylvania; and Philadelphia, Pennsylvania, did a good job of providing reasonably
accurate rainfall amounts.
In general, WSR-88D precipitation products underestimated the rainfall by a factor of 2
or more. In most areas, forecasters were able to make adequate manual adjustments to
the estimates based on comparison to rain gage reports. Most forecasters who worked
this event were pleased with the precipitation products generated by the WSR-88D.
While rainfall estimates were not perfect, they provided a good representation of the
rainfall patterns when compared to rain gage measurements. Rainfall estimates
provided by these radars, along with critical rainfall reports from cooperative
observers, were crucial to the issuance of early Flood and Flash Flood Warnings. The
Susquehanna Flood Forecast and Warning System is a nonstructural flood mitigation
measure that was implemented in 1985; it provided significant additional data which
helped the NWS issue timely watches and warnings.
PREPAREDNESS AND COORDINATION
All office Station Duty Manuals (SDM) were up to date and provided the staff with
specific information concerning office operations and actions during both river and
flash flood events. All SDMs, along with NWS Form E-19, depicted flood-prone and
flood inundation areas.
Specific hydrologic drills had been accomplished by all offices to ensure all office
personnel were trained to handle any type of flooding. Most drills had been completed
within the past 6 months. Hydrologic service drills and staff training are usually
conducted by the Service Hydrologist (SH) and/or Warning Coordination Meteorologist
During the past year, numerous outreach activities and hazardous weather training,
focusing on flooding and severe weather, were provided by NWSFOs in Pittsburgh,
Pennsylvania; Baltimore, Maryland/Washington, DC; Philadelphia, Pennsylvania; and
Charleston, West Virginia; and by NWSOs in Binghamton, New York; State College,
Pennsylvania; and Wilmington, Ohio. Extensive coordination between the RFCs and
their respective HSAs was also accomplished during this period. All offices have an
excellent line of communication with County Warning Area (CWA) media, and
emergency management and other public safety officials. During this flood event,
several spotter groups were activated due to the possibility of severe weather Friday
morning, January 19, but they were also used to report flood conditions and
MEDIA AND USER RESPONSE
Overall, media coverage relating to the Northeast Floods of January 1996 -- the worst
flooding in more than 10 years across the mid-Atlantic region -- was positive. News
stories focused on the cause and extent of the flooding, human interest stories, and
recovery efforts, and they accurately highlighted the weather as it developed across the
The February 4 issue of Newsweek credited the NWS for the advance warnings of the
"Modernized equipment allowed the National Weather Service in January
to warn the eastern U.S. of flash floods up to 24 hours before major
rivers crested. State-of-the-art Doppler radars monitored rainfall, while
a new computer network let the service swap data in real time with states
On Monday and Tuesday before the event, all NWS offices in the region issued Flood
Potential Statements so the public and media were aware that spring-like ice jam
flooding would occur. Throughout the flood-ravaged areas, the primary sources of
information for most citizens were radio and television broadcasts. Most radio and
television stations aired watches and warnings as soon as they received them. All local
Emergency Broadcast System (EBS) radio and television stations aired watch and
warning information beginning Thursday before the event.
As a result of the significant impact of the major flooding in Pennsylvania, a special
State House committee reviewed the operations of the NWS and other state and Federal
agencies. The five-member bipartisan committee, which held hearings in Charleroi and
Williamsport, Pennsylvania, concluded that the NWS provided timely watches and
The Northeast Floods of January 1996 provided many challenges for all NWS offices
involved with the flood. In spite of the rapid onset and complexity of this event, the
warnings and forecasts provided by NWS personnel, in cooperation with emergency
managers and the media, provided excellent service to the public and specialized users.
The NWS did an outstanding job in recognizing the flood potential, a full 5 days in
advance, and getting the word out to the public, emergency services, and hydrologic
agencies so that they could take the necessary actions to prepare for and fight the flood.
In spite of the outstanding efforts of the NWS, the media, and the emergency
management community, there were 33 fatalities. In an effort to mitigate potential
future losses, one responsibility of the disaster survey team is to review operations,
highlighting positive aspects and identifying any weaknesses. Toward that end,
detailed findings and recommendations are contained in Chapter 6 of this report.
Return to Table of Contents