October 27, 2008
Methodology: First, let me begin by providing an outline of the methodology utilized in producing this outlook. The following were researched: the El Nino Southern Oscillation (ENSO), Pacific North-American Index (PNA), East-Pacific Oscillation (EPO), Arctic Oscillation (AO), North Atlantic Oscillation (NAO), Pacific Decadal Oscillation (PDO), Atlantic Multi-decadal Oscillation (AMO), analogs, climatic trends with respect to temperature and precipitation, recent pattern progression/repetition, autumn precipitation/temperature departures, sea surface temperature anomalies in both the Atlantic and Pacific, cryospheric evolution (trends, anomalies), and solar trends. The outlook for the upcoming cold season is more challenging due to the lack of a strong signal from the tropical pacific; however, I believe there are other notable indications which provide clues for the potential winter pattern (some of which emerged as recent as the past month). Now I will delve into the details regarding these signals.
ENSO
Over the course of the past year, we have witnessed the progression of a moderate to strong La Nina episode into an essentially neutral regime by the summer of 2008. As per the tri-monthly numbers from CPC, the ENSO region 3.4 numbers for December, January, and February were -1.4c, -1.5c, and -1.4c respectively. Although not “technically” a tri-monthly strong La Nina – for all intents and purposes, the pattern which resulted over the CONUS featured the primary earmarks of a strong Nina event.
As SST’s warmed over the equatorial regions of the Pacific, we saw a rapid reversal of those negative anomalies, as ENSO region 3.4 readings exceeded -0.5c (weak nina threshold) by May and reached 0.0c by August of 2008 (tri-monthly). In September and October, there has not been any noticeable indication for the reemergence of a moderate to strong La Nina episode. Statistics regarding the waxing and waning of Nina’s support this notion.
Below are the years since 1950 which featured a moderate or strong La Nina (borderline moderate seasons were included):
1950-51
1954-55
1955-56
1964-65
1970-71
1973-74
1975-76
1984-85
1989-90
1998-99
1999-00
If one examines the winters that followed the above – all except one of the years which featured a second year La Nina event (weak, moderate or strong) had already fallen into Nina territory by this time of year (mid autumn) or had it persisting through the summer months.
Here are the ENSO region 3.4 readings for the second year Nina’s in JAS (July-August-September tri-monthly number):
1955: -1.0c
1956: -0.8c
1971: -0.8c
1974: -0.4c (briefly dropped below weak Nina status for a month)
1975: -1.3c
1999: -1.0c
2000: -0.4c
The current 2008 tri-monthly reading for JAS: 0.0c
Only autumn of 2000 had an ENSO 3.4 value in the neutral category, preceding the weak Nina winter of 2000-01 (and it was barely below that threshold). One can see from the above seasons that not only were the second year Nina’s in progress by the autumn, most of them had persisted through the previous summer (which was not the case in summer of 2008).
The recent weekly readings have shown a decrease in the overall thermal gradient between the western and eastern ENSO regions – which is somewhat expected, as winters with a Nino-east, cold-neutral/Nina west are very rare.
The current SSTA profile for reference:
http://weather.unisys.com/archive/sst/sst_anom-081026.gif
Thus the statistics strongly favor the upcoming winter season to fall within the neutral category, -0.5c to +0.5c, due primarily to the progression and evolution of this event over the past year.
As far as model support; this was of secondary importance but still worthwhile in the outlook.
The statistical and dynamical modeling supports the notion of a neutral ENSO prevailing throughout the upcoming winter season. The majority of the models fall within the cold neutral range (0.0c to -0.5c).
http://iri.columbia.edu/climate/ENSO/currentinfo/SST_table.html
The ECMWF ENSO projection depicts a cold-neutral consensus as well, with the possibility of a weak La Nina developing by spring of 2009.
Therefore, I feel confident in my expectation that the meteorological winter, DJF, will feature neutral ENSO conditions, erring on the side of cold-neutral (0.0c to -0.5c), especially late winter. Due to this expectation, there are other, more crucial factors which will determine the evolution of sensible weather over the CONUS. Thus a far cry from last year’s highly ENSO dominated forecast.
Pacific Indices: PDO/PNA
The past year has seen a remarkable plummet in PDO values, in conjunction with the emergence of the moderate/strong La Nina event of 2007-08. However, is this turn negative temporary, or potentially long-lasting? If one examines the PDO fluctuations of the past century, there is a clear tendency for negative and positive values to run in 25-30 year cycles. Although some years may buck this trend, the majority coincide with the modality of their decadal phase. Below is a graph depicting PDO trends of the past century.
http://img235.imageshack.us/img235/2899/pdo1bk7.jpg
Note the negative (cold) phase predominate from the mid 40s through late 70s, the positive (warm) phase from the late 70s through the early 2000s. We have now reached the 30 year mark since the last PDO decadal shift, so it is generally accepted that we have now resumed the negative (cold) PDO cycle.
What does this mean in terms of ramifications for the winter season? First, here are the PDO values for the past few months.
September 08: -1.55
August 08: -1.70
July 08: -1.67
These numbers indicate a moderate to strongly negative PDO. If one looks at the SSTA profile earlier in the post, the colder the normal ocean temps are seen from the Gulf of Alaska southwestward into the central Pacific – classic negative PDO appearance. There has been some improvement over the past month, but not much, yet.
Due to my expectation of a neutral ENSO regime this winter, let’s look at neutral ENSO winters within the last negative PDO decadal phase (which featured a –PDO) to note any trends.
Here are the values for those neutral, -PDO winters.
1952-53:
Dec: +0.04
Jan: -0.57
Feb: -0.07
1953-54:
Dec: +0.07
Jan: -1.32
Feb -1.61
1958-59, 1959-60, 1960-61: Neutral, positive PDO winters.
1961-62:
Dec: -2.69
Jan: -1.29
Feb: -1.15
1962-63:
Dec: -0.96
Jan: -0.33
Feb: -0.16
1966-67:
Dec: -0.32
Jan: -0.20
Feb: -0.18
1978-79:
Dec: -0.43
Jan: -0.58
Feb: -1.33
5/6 of these –PDO, neutral winters (52-53, 53-54, 61-62, 62-63, 66-67, 78-79) featured PDO values which were higher (warmer) than the current PDO reading of -1.55. Most of these winters saw the PDO rise (sometimes drastically) at the onset of the winter season to weakly negative values. None of the years which had a mod-strong negative PDO preceding the winter saw the PDO switch positive (warm) for winter. Without the added influence of a La Nina in the tropical pacific, I believe it will be easier for the PDO to become less “intensely negative” over the next 1-2 months, as the statistics above demonstrate. Thus, my expectation for the PDO during winter 08-09 is weakly to moderately negative. Most if not all of the winter should be spent above the current PDO value of -1.55.
How does the PNA relate to this? Well, here are the PNA readings for Dec, Jan, and Feb for the neutral, -PDO winters mentioned above.
1952-53:
Dec: +0.95
Jan: +0.65
Feb: +1.00
1953-54:
Dec: +1.27
Jan: -1.14
Feb: -0.14
1961-62:
Dec: -1.24
Jan: -0.07
Feb: -0.18
1962-63:
Dec: -0.08
Jan: +0.58
Feb: +0.81
1966-67:
Dec: +0.09
Jan: -0.40
Feb: +0.07
1978-79:
Dec: -0.72
Jan: -0.69
Feb: -1.82
If one notices the 6 negative PDO years, none of them averaged moderate to strongly negative in terms of the PNA values. In fact, 4/6 years had 1 or more months with a positive PNA, particularly December and February. The months which featured a negative PNA were generally barely negative. 1978-79 was the most neg PNA winter of the group (but even that was significantly so). It is also interesting to notice that the moderately negative PDO winter of 1961-62 managed to have a PNA near-neutral for 2/3 of the meteorological winter. This tells me that neutral ENSO, -PDO seasons, have a much higher likelihood of featuring weakly negative or even positive PNA’s than La Nina years. This is because there are other forcing mechanisms acting upon the PNA, as a strong ENSO signal is not present. Thus, I expect the PNA to average near-neutral, with a good chance for at least one month of +PNA (favoring December), and at least one month of –PNA (favoring January).
Polar Indices: NAO/AO
The polar indices are a crucial aspect of the forecast this year as are the Pacific indices. Without the presence of an ENSO event, I have researched NAO/AO trends during the autumn months which could possibly provide clues for future behavior (positive/negative) of these teleconnections.
Let’s begin with the Arctic Oscillation (AO); if one notes the graph of AO trends over the past month, its progression is pretty glaring.
http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/ao.sprd2.gif
The fact that readings have been exceeding +2 for a couple weeks now means the October AO value will definitely finish at or above +1. When looking at past Octobers which also featured an AO > +1, an interesting correlation results (values rounded to nearest tenth).
1956: +1.1
1963: +1.1
1967: +1.3
1971: +1.2
1983: +1.4
1985: +1.0
1986: +1.4
The following winter AO values DJFM, averaged negative in 6/7 (86%) of cases (exception was 1971-72, weak Nina winter).
In 6/7 cases, at least 2 months out of DJFM featured a negative AO.
In 5/7 cases (71%), 3 months out of DJFM featured a negative AO.
If one expands the sample size to included Octobers with an AO value greater than +0.5, the support for a negative AO dominated winter also strengthens.
I have found that the North Atlantic Oscillation (NAO) follows a similar pattern when it comes to reversal of modality from October to the winter DJFM average (positive to negative, and vice versa).
The NAO for the month of October will end up slightly positive per the CPC, and possibly moderately positive according the SLP NAO data.
http://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/nao.sprd2.gif
If one notes the NAO values of the Octobers from 1950-present, and their ensuing winter NAO averages, the correlation is about 70-75% for the phase reversal. Obviously no correlation in meteorology is perfect – one of the reasons why long range outlooks are so difficult – but a 70-75% correlation is certainly sufficient to draw conclusions.
If we take the period 1990-2008 for example, 14/18 winters followed this reversal rule (78% correlation).
I also believe that moderate to strong ENSO years (Nino or Nina) are usually when the exceptions to this rule occur, as the ENSO dominates the other signals. I.e., last winter, the October NAO was positive but the DJFM period was also positive, likely because the intense Nina induced mid-latitude jet prevented any significant North Atlantic blocking from occurring.
Two of the four exceptions, 99-00, and 07-08 were strong ENSO years.
The analog years I’m utilizing for all aspects of the forecast, featured at least 2 months of negative NAO within the DJFM period.
If one examines the combined NAO values for Sept/Oct – the correlation is even stronger.
Below is a list of every year since 1950 – their Sept/Oct combined modality, and the ensuing winter phase respectively.
Year – Sept/Oct phase – Winter DJFM phase:
1950-51: positive – negative
1951-52: positive – negative
1952-53: negative – negative
1953-54: positive – negative
1954-55: positive – negative
1955-56: negative – negative
1956-57: positive – negative
1957-58: negative – negative
1958-59: positive – negative
1959-60: positive – negative
1960-61: negative – positive
1961-62: positive – negative
1962-63: positive – negative
1963-64: positive – negative
1964-65: positive – negative
1965-66: positive – negative
1966-67: negative – positive
1967-68: positive – negative
1968-69: negative – negative
1969-70: positive – negative
1970-71: negative – negative
1971-72: positive – positive
1972-73: positive – positive
1973-74: negative – positive
1974-75: positive – positive
1975-76: positive – positive
1976-77: negative – negative
1977-78: positive – negative
1978-79: positive – negative
1979-80: positive – negative
1980-81: negative – positive
1981-82: negative – positive
1982-83: positive – positive
1983-84: negative – positive
1984-85: positive – negative
1985-86: positive – positive
1986-87: positive – negative
1987-88: negative – positive
1988-89: negative – positive
1989-90: positive – positive
1990-91: positive – positive
1991-92: positive – positive
1992-93: negative – positive
1993-94: negative – positive
1994-95: negative – positive
1995-96: positive – negative
1996-97: negative – positive
1997-98: negative – positive
1998-99: negative – positive
1999-00: positive – positive
2000-01: positive – negative
2001-02: negative – positive
2002-03: negative – positive
2003-04: negative – positive
2004-05: negative – positive
2005-06: negative – negative
2006-07: negative – positive
2007-08: positive – positive
2008-09: positive — ?
There were 18 years in this 58 year sample size that did not follow the rule. That’s 40/58, or a 69% correlation (for the NAO phase reversal from autumn to winter).
However, when the moderate and strong ENSO years were removed (mod/strong Nina and Nino), there were only 8 years in the now 48 year sample that didn’t fit the rule. That’s 40/48 or 83% correlation.
Therefore, for the polar field indices, my expectation is a predominately negative NAO/AO winter, with a strong chance of at least 1/4 month(s) averaging +NAO/AO. The magnitude of the NAO/AO is not as certain, but I’m leaning towards weakly negative overall DJFM. Periods of moderate or even strongly negative are possibly, with fewer periods of positive values.
Autumn Trends
Before getting to the forecast numbers, I’d like to mention another aspect of the forecast (other factors of lesser importance were included but not mentioned in the outlook).
October temperature departures over the Northeast are slightly below normal through the 26th, generally 0 to -1, with slightly positive departures in the Mid-Atlantic. With the next several days well below normal temperature wise, October is guaranteed to finish below normal in most if not all Northeastern cities.
Looking at the winters following colder than normal Octobers for the NYC region, there is about a 74% correlation (the majority of winters average colder than normal with the preceding October cold).
Just as an example to support this theory, here are 11 colder and snowier than normal winters in NYC since 1950:
1957-58
1960-61
1963-64
1966-67
1977-78
1993-94
1995-96
2000-01
2002-03
2003-04
2004-05
9/11, or 81% of them had the preceding October colder than normal (the two exceptions were 1963 and 1995). With that being said, this adds credence to my overall ideas for the upcoming season.
Conclusion
Before I get to the forecast numbers, here are my analogs for the 08-09 winter season. These years take all factors into account – ENSO, PDO, AO, NAO, trends etc.
1961-62
1962-63
1966-67
1978-79
1985-86
1985-86 was included as an analog due to strong similarities to ENSO, NAO, and AO progression, although it did not fit the negative PDO criteria.
1978-79 was also one of the lesser analogs as that year occurred just prior to the PDO phase switch from negative to positive rather than the current positive to negative. However, it yielded enough similarities to be included.
Overall, I am anticipating this winter to be colder than both 2006-07 and 2007-08 in the Northeastern United States. The combination of index / pattern trends and strong analog support leads me to the conclusion that December will feature the coldest temperature departures with respect to normal. All except one of my analogs (78-79) featured a colder December to some degree. Most of them also featured a warmer than normal January (“thaw”) with a return to colder than normal temperature regime in the month of February. The precipitation pattern for the analog years yielded a winter featuring near normal precip for the I-95 corridor, and slightly below normal precip for the interior Northeast. However, December and February may feature normal or slightly below normal precip although the overall DJF average should be near-normal in the I-95 corridor.
As far as snowfall, northern New England will not receive the unprecedented totals of last winter, and there is a good chance they finish below average snowfall wise (slightly, not significantly so). I anticipate the highly amplified pattern in the early and latter parts of the winter to yield increased cyclogenesis along the eastern seaboard. Warmer than normal SST’s should prevail in the west Atlantic due to the +AMO, thereby enhancing the baroclinicity / gradient. In the I-95 corridor from BOS to DCA, above normal snowfall should be expected with slightly below average temperatures for the DJF average. First – numbers for NYC, then other select cities. Please take the snowfall estimated totals with a grain of salt:
Outlook: NYC
Temperatures:
December: -2 to -3
January: +2 to +3
February: -1 to -2
Winter (DJF) average: 0 to -1
Precipitation: Near-normal
Snowfall: Above normal (30-35”)
Bonus cities:
Boston:
December: -2 to -3
January: +2 to +3
February: -1 to -2
Winter (DJF) average: 0 to -1
Snowfall: Above normal (45-50”)
Washington DC:
December: -2 to -3
January: +2 to +3
February: -1 to -2
Winter (DJF) average: 0 to -1
Snowfall: Above normal (20-25”)
Richmond:
December: -1 to -2
January: +1 to +2
February: 0 to -1
Winter (DJF) average: Neutral (0)
Snowfall: Near-normal (10-15”)
Maps:
http://i35.tinypic.com/2qvqxb7.jpg
http://i36.tinypic.com/z947s.jpg
http://i33.tinypic.com/2mpfk2u.jpg
http://i33.tinypic.com/29o0ubr.jpg
Winter Outlook Verification
Since this forecast was directed toward the NYC area in particular, that’s where I will focus most of my grading/verification. Will be using a 4.0 scale for points, (A=4, B=3, C=2, D=1, F=0).
For referrence — my forecast:
Temperatures:
December: -2 to -3
January: +2 to +3
February: -1 to -2
Winter (DJF) average: 0 to -1
Precipitation: Near-normal
Snowfall: Above normal (30-35”)
Monthlies:
December: C
January: F
February: D
Monthly score: D
Precipitation was near normal for DJF was that was a very good call, but won’t assign a grade for it.
Overall DJF Winter
DJF temperature departure: -0.4
Perfectly within range, so overall temp score: A
Snowfall
Totals across the area were near to slightly above normal.
NYC: 27.6″
Snowfall score: A-
Snow is counted until the end of the season, so we may still hit 30″ in NYC for the season. Regardless it should have little bearing on this grade.
Winter 2008-09 final grade: B
The snowfall and overall temps were very good but the monthlies bring down the score.
Comments or different grades/criticism is welcome! I’ll probably do a post-mortem in this thread as well.