Preliminary Winter 2004-05 Forecast

Overall, given the preponderance of possible analogs coming up, I believe Winter 2004-05 will see colder than normal conditions in the eastern third to half of the United States and warmer than normal ones in the western quarter to third of the United States. Just to be clear, on a regional scale, I do not believe that cold anomalies will rival those of the extreme Winter of 1976-77. Rather, the figures will probably come out somewhere between those of Winters 2002-03 and 2003-04—probably somewhat warmer than Winter 2002-03 but somewhat colder than Winter 2003-04—particularly in the Washington, DC to Boston region.

Warmer than normal: Alaska, Arizona, California, Colorado, Hawaii, Idaho, Nevada, Oregon, Utah, New Mexico, and Washington

Cooler than normal:  Alabama, Arkansas, Connecticut, Delaware, District of Columbia, Florida, Georgia, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maine, Maryland, Massachusetts, Michigan, Mississippi, Missouri, Montana, Nebraska, New Hampshire, New Jersey, New York, North Carolina, North Dakota, Ohio, Oklahoma, Pennsylvania, Rhode Island, South Carolina, South Dakota, Tennessee, Texas, Vermont, Virginia, West Virginia, and Wisconsin

Elsewhere, I expect temperatures to average near normal.

For what it is worth, the following are the fifteen lowest combined mean temperatures for the December-February period in DCA, NYC, and BOS since 1950-51 are:

1. 1976-77: 29.8°
2. 1962-63: 30.1°
3. 2002-03: 31.0°
4. 1993-94: 31.1°
5. 1960-61: 31.2°
6. 1958-59: 31.2°
7. 1977-78: 31.4°
8. 1969-70: 31.5°
9. 1967-68: 31.8°
10. 1970-71: 32.2°
11. 1981-82: 32.2°
12. 1963-64: 32.3°
13. 1961-62: 32.6°
14. 2003-04: 32.8°
15. 1995-96: 32.8°

Confidence in this temperature forecast is high.

In terms of precipitation anomalies, I believe the following anomalies are likely:

Gulf Coast: Below normal precipitation

Midwest: Below normal to near normal precipitation

Northeast: Near normal to above normal precipitation

Pacific Northwest & Coast: Near normal to above normal precipitation

Elsewhere, I expect near normal precipitation.

Drier than normal: Alabama, Arkansas, Hawaii, Illinois, Indiana, Iowa, Kentucky, Louisiana, Mississippi, Missouri, Nebraska, Oklahoma, Tennessee, and Texas

Wetter than normal: California, Connecticut, Delaware, District of Columbia, Florida, Maine, Maryland, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont, and Virginia

Anticipated Snowfall for Winter 2004-05 for Select Cities:

Albany: 80”-90”

Baltimore: 18”-23”

Boston: 50”-60”

Chicago: 40”-50”

Cleveland: 80”-90”

Detroit: 40”-50”

New York City: 20”-30”

Philadelphia: 20”-25”

Pittsburgh: 60”-70”

Providence: 45”-55”

St. Louis: 18”-28”

Washington, DC: 12”-17”

ANALYSIS:

Analog Seasons: The Return of El Niño

Before one automatically assumes that the return of El Niño indicates that the United States will be flooding by a torrent of warm air and a dearth of snowfall, it should be noted that not all El Niños are “created equal.”

First, the El Niño for Winter 2004-05 will be weaker than the powerful ones that have suffocated winters past. 

Second, the ENSO regional anomalies will likely be lower in Region 1+2 and warmer in Region 3.4.  This leads to a different outcome than one in which the El Niño warm anomaly is very strong in Region 1+2.

Based on a combination of the latest global data, analogs for ENSO, and the model guidance, the following parameters are likely to be in place, on average, for Winter 2004-05:

• Region 1+2 Anomaly: No warmer than +0.25°C

• Region 3.4 Anomaly: Warm

• MEI: >0

• PDO: >0

• QBO: 0 +/- 6.00

This combination is a classic for a cool East/warm West scenario.  Since 1950, when the global indices were regularly recorded, this scenario has occurred in 4 winters: 1969-70, 1977-78, 2002-03, and 2003-04.

Another somewhat less likely but still possible scenario involves the following parameters:

• Region 1+2 Anomaly: Cool

• Region 3.4 Anomaly: Warm

• MEI: >0

• PDO: >0

This combination is also a cold one similar to the above-noted one.  In fact, the cold is more widespread under such a scenario.  Since 1950, the following winters met such criteria: 1958-59, 1969-70, 1977-78, 1993-94, 2002-03, 2003-04.

If one relaxes the R 1+2 constraint in the above scenario to up to +0.25°C, then two warmer winters enter the picture: 1979-80 and 1987-88.  Neither of those winters had a QBO that ranged from –6.00 to +6.00.

Taking into consideration ENSO Regional temperature and QBO parameters (Region 1+2: <25°C; Region 3.4: >27°C; QBO < +7.00) one finds similar ideas to what is outlined in most of the above prospective analogs. 

Finally, additional winters that received consideration even though they were at odds with at least some of the global indices were: 1976-77, 1985-86, and 1992-93.

Needless to say, while a number of super winters showed up, although one can look for attributes of such winters from time to time, there were some important constraints.  For example, the QBO switched to West in December 1977 and January through March 1978 had a West QBO.  This season is likely to see a switch to East.  Winter 2002-03 saw a moderate El Niño at its peak.  This season is likely to see only a weak El Niño. Winter 1993-94 experienced a moderate El Niño during the autumn months, which faded to neutral conditions during the heart of the winter.

Computer Guidance:

At this point in time, the computer guidance is strongly supportive of the temperature ideas discussed above.  In fact, the ECPC’s latest guidance indicates that the United States would be colder than normal almost from coast-to-coast.  I don’t believe that the cold will be that expansive.

Climate Model Links include:

CFS: http://www.cpc.ncep.noaa.gov/products/people/wwang/cfs_fcst/

ECPC: http://ecpc.ucsd.edu/m2s/m2s_ECPC_forecasts.html

2004 Hurricane Season:

When one factors in seasons that involve a winter El Niño, landfall of intense hurricanes (with a focus on landfalls occurring in the Gulf of Mexico region and at least one landfall on the Atlantic East Coast), Winter 1969-70 gains weight.  Winter 1969-70 was very cold in the eastern half of the United States and very warm in the western half of the United States.

In addition, since 1950 when regular reporting of the global indices has been available, there were two hurricane seasons that saw 10 or more named storms when Region 3.4 had an average temperature at or above 27.5°C (anomaly at or above +0.6°C) for the May-November period: 1969 and 2002.

Northern Hemisphere Snowcover:

August 2004 saw Northern Hemisphere snowcover amount to just 2.1 million square kilometers. Since Northern Hemisphere snowcover statistics were regularly provided in 1973, there have been 11 prior occurrences where August Northern Hemisphere snowcover came to less than 3 million square kilometers: 1984, 1988, 1989, 1990, 1992, 1993, 1999, 2000, 2001, 2002, and 2003.

Thus, three of the possible analog years noted from the above samples show up: 1993-94, 2002-03, 2003-04.

Overall, not all winters that follow such low snowcover see below to much below normal snowfall in the Northeastern United States.  However, almost all such winters see below normal snowfall in the Mid-Atlantic region.

For the Northeast, an indication that a snowy winter lies ahead could be found in October and November Northern Hemisphere Snowcover if such criteria are met:

• October snowcover is 17 million or more square kilometers.

• November snowcover is 34 million or more square kilometers.

Winters that met such criteria following an August with less than 3 million square kilometers of snowcover were 1993-94, 2000-01, 2002-03, and 2003-04.

In those winters, seasonal snowfall came to an average of 45.1" in NYC and 63.1" in Boston. All four seasons saw NYC's seasonal snowfall exceed 30". Three of those four seasons saw Boston's seasonal snowfall come to more than 45" (2003-04 was the single exception with 39.4"; the first winter since Winter 1990-91 where NYC's snowfall exceeded Boston's).

In seasons where November snowcover came to 34 million or more square kilometers without regard to October's snowcover, NYC's average seasonal snowfall came to 41.0" and Boston's amounted to 67.3". NYC's lowest snowfall in such seasons was 24.5" (1992-93).

As the CPC numbers for Northern Hemisphere snowcover were not available even by October 11, I reviewed the September maps and estimate that such snowcover came to 4.0 million to 5.0 million square kilometers, or somewhat less than that of last year.

Based on past occurrences with low August snowcover, this does not necessarily support widespread above normal snowfall for the Eastern United States. Some weight is given to this data.

The QBO:

The QBO had peaked in May at +12.97.  Since then, it has declined steadily with a reading of +7.22 at the end of September.  Over the past 4 months, it has most closely resembled that of 1969 and 2002, with 2002 proving to be a somewhat better match. 

Should such trends continue, it is likely to switch to East no later than January or February and possibly as early as December.  What also appears likely is that its rate of descent could accelerate in either November.

All said, it appears likely that the QBO’s December-March average will range between –6.00 and +6.00 with the lower half of this range probably more likely.  All said, the QBO’s decline, its likely switch to East, and its winter average together offer additional support for the possible 1969-70 and 2002-03 analogs.

The North Atlantic Oscillation (NAO):

Based on a review of the NAO averages for winters 1950-51 through 2003-04, it appears that the NAO is a product of a broader range of factors than SSTAs and Northern Hemisphere snowcover.  Indeed, since 1950-51, a global approach has fared quite well in predicting the NAO average, whether it would come out above or below 0.00 for the December-February period.  Based on such an approach, it appears likely from this vantage point that the NAO will average below 0 for the December-February period.  A firmer idea should be available in November.  Nevertheless, some weight is given to the idea of a negative NAO. 

MISCELLANEOUS ISSUES:

A second consecutive winter with NYC seeing more snowfall than Boston?

Such a scenario is infrequent but not rare.  It has occurred on five occasions:

• 1895-96 and 1896-97

• 1907-08 and 1908-09

• 1913-14 and 1914-15

• 1935-36 and 1936-37

• 1978-79 and 1979-80

It should be noted that in no seasons in which Boston saw 50” or more of snow was that figure surpassed by New York City’s seasonal snowfall.  The record is 44.9” in 1904-05 (New York City received 48.1”).

A regional precipitation and temperature profile for Boston suggests that such a scenario is highly unlikely this coming winter, especially as things appear set for at least 50” there.  Hence, given the analog pool and this piece of regional data, I strongly believe that Boston will see more snowfall than New York City this winter.  Moreover, the difference could be sizable.

Boston: Blockbuster Seasonal Snowfall Watch:

How Boston fares in terms of snowfall in December and January could well determine whether Boston is en route to a blockbuster season in terms of snowfall:

When December snowfall has come to 12" or more and January snowfall reached or exceeded 20", Boston received the following seasonal amounts:

1903-04: 72.9"
1904-05: 44.9"
1922-23: 68.5"
1947-48: 89.2"
1956-57: 52.0"
1964-65: 50.4"
1976-77: 58.5"
1993-94: 96.3"
1995-96: 107.6"

Seasonal of 50" or more: 8/9 years
Seasonal of 65" or more: 5/9 years
Seasonal of 70" or more: 4/9 years

Average: 71.1"

Where December snowfall came to 12" or more and January snowfall reached or exceeded 25":

1903-04: 72.9"
1922-23: 68.5"
1947-48: 89.2"
1993-94: 96.3"
1995-96: 107.6"

Seasonal of 50" or more: 5/5 years
Seasonal of 65" or more: 5/5 years
Seasonal of 70" or more: 4/5 years

Average: 86.9"

MAJOR ASSUMPTIONS:

• Weak El Niño (Central Pacific variety)

• Region 3.4 will have greater warm anomalies than Region 1+2; there is some chance that Region 1+2 maintains a slight cool anomaly on average during the winter

• Region 1+2’s average temperature during the December-March period will likely come in below 25°C; Region 3.4’s average temperature during the same timeframe will likely come out to above 27°C

• The PDO will be +

• The QBO will average between +5 and –5 with a switch to East occurring, probably during the January-February period

UPDATE:

By the end of November or even early December, much additional data concerning ENSO anomalies, PDO, QBO, SSTAs, Northern Hemisphere Snowcover, etc., should be in place.  At that time, trends with respect to key teleconnection indices such as the PNA and NAO should be more readily apparent.  Therefore, if necessary, an update would be provided.

Forecaster:  donsutherland