Overview: First, let me begin by providing an outline of the methodology utilized in producing this outlook. Some factors were given less weight than others, however; everything is interconnected in meteorology, i.e., the various teleconnection indices, atmospheric and oceanic circulation patterns, as well as upper tropospheric and stratospheric influences. Therefore, I’ve given the majority of the variables a similar weight considering they’re all pieces of a highly complex, immense system we like to call “weather”. With that said, the following were researched: the El Nino Southern Oscillation (ENSO), Pacific North-American Index (PNA), East-Pacific Oscillation (EPO), Quasi-Biennial Oscillation (QBO), 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, cryosphere evolution (trends, anomalies), and finally storm genesis regions during the autumn months.
Climate Trends
In my view, this is an important, necessary factor to incorporate into winter outlooks, although its effects may seem quite minimal. Regardless of the causes – whether it be natural variability, overall solar output/radiation, cosmic activity, eccentricity of earth’s orbital path and axis tilt, oceanic circulations, climate indices, interglacial warming as we continue to emerge from the Little Ice Age, anthropogenic influences, or a combination of the possibilities mentioned above. The fact is temperatures over the past 2-3 decades have averaged significantly warmer than the prior 2-3 decades. What do I mean by significant? For the majority of the United States, temperatures during the 1986-2005 period have averaged about 2F warmer than the 1960-1980 period. The most extreme anomalies are located in the Upper-midwest, particularly the northern plains, where temps are some 5F warmer in the past few decades. Once again, the causes attributed to this warming can be speculated, however; below are the temperate composite maps further illustrating the point:
December-January-February (Meteorological winter) anomalies for 1960-1979:
http://i10.tinypic.com/61z4pol.gif
December-January-February anomalies for 1986-2005:
http://i13.tinypic.com/4t71p9z.gif
This is why (IMO), when using nina analogs from the 50s-70s as a guide for the upcoming winter’s temperature departures, one should increase those numbers slightly.
Secondly, precipitation trends over the past several decades have varied quite a bit as well. The overall temperature increase since the mid 80s has generally coincided with higher-than-normal precipitation (based on the 1950-present day averages), especially east of the Mississippi River. New York City sticks out like a sore thumb with its 4-5 consecutive years > or equal to 50” of precipitation. Interestingly enough, the wet trend coupled with cooler temperatures resulted in four years in a row (record-breaking) with 40”+ of snowfall. The reasons for the up tick in precipitation are somewhat uncertain, however; the most logical explanation would be, “an increase in atmospheric temperature yields a higher concentration of water vapor, thus more condensation and precipitation.” This coupled with our recent entrance (1994-1995) into the warm phase of the Atlantic Multi-decadal oscillation may have enhanced the baroclinicity along the eastern seaboard, hence more storminess. Regardless, below are comparison maps of the period 1960-1979 to 1986-2005 with respect to precipitation departures.
January through December for 1960-1979:
http://i2.tinypic.com/62dg869.gif
January through December for 1986-2005:
http://i5.tinypic.com/5zfxquw.gif
With that said – these are smoothed out averages over a long period of time (20 years), so in my view it certainly should not receive as much weight as other, more pertinent factors.
ENSO
Over the past three months, we’ve witnessed the rapid, almost unprecedented, intensification of a La Nina event. In simple terminology, it’s a weakening of the tropical westerly trade winds in the pacific (not necessarily defined as a reversal of these winds to easterly) thus an overall cooling of sea surface temperature along and to the west of the South American coastline. Considering the limited variability in terms of SST’s within the tropical regions, a swing of only a couple degrees is quite significant thus is bound to have drastic effects on the atmospheric circulation patterns worldwide. As a general rule, the stronger the nina, the stronger the mid-latitude pacific jet. However, its effects can be blunted by other indices (i.e., NAO/AO) as we’ve seen in past years. Below is a map of the current SSTA profile worldwide:
http://weather.unisys.com/archive/sst/sst_anom-071028.gif
Notice the coldest anomalies concentrated in ENSO region 1-2, with a slightly warmer appearance further west in regions 3, 3.4, and 4. Latest data from the Australian Bureau of Meteorology is as follows:
Region 1: -1.38c
Region 2: -0.57c
Region 3: -1.05c
Region 3.4: -1.06c
Region 4: -0.41c
I’m aware that these numbers are slightly lower than CPC’s, however; the main point is we’re reaching the moderate threshold in terms of intensity (if not already, certainly within the next couple weeks). Officially, we need three consecutive months with region 3.4 below -1.5c for a moderate la nina. We’re nowhere near that right now, as late summer numbers were barely into weak nina territory, -0.6c to -0.7c. The rapid cooling occurred in early October, continued much of the month, leveled off the past 10 days, but I believe we’re about to see another downward trend. These fluctuations are completely normal in any ENSO event, as trade wind variability coupled w/ tropical forcing and negative/positive bursts in the SOI can alter sea-surface temperature anomalies by as much as 0.2-0.5c per week. What are the implications for the upcoming winter? Well, a significant factor IMO is the fact that we’ve transitioned into the cool phase of the PDO within the past decade. Therefore, there should be a greater tendency for current and future nina events to sustain themselves. The rapid weakening of the recent El Nino event was a clue, IMO, that we’ve definitely entered a period, climatologically speaking, that favors more moderate/strong La nina’s than mod/strong el nino’s. Secondly, the majority of nina events peak in the late autumn/early winter, with a slow weakening trend as the meteorological winter progresses. In fact, most years with similar readings in ENSO region 3.4 during October saw the nina sustain itself through at least January. The best analogs in terms of ENSO: 1954-55 (Oct peak), 1955-56 (Nov peak), 1970-71 (Jan peak), 1998-99 (Jan peak), 1973-74 (Dec peak), and 1988-89 (Nov/Dec peak), in no particular order. As far as intensity, I don’t believe this ENSO event will reach the official trimonthly “strong” status. All strong la nina’s (except one – 1998-99) since 1950 were already nearing or surpassing the strong threshold at point in the time of year (late oct/early nov). In terms of model support – we have some disagreement between the ECMWF and CFS. The CFS has redeemed itself to an extent this year, IMO, however in the grander scheme of things, the Euro has a tendency to perform w/ higher levels of accuracy. The CFS blew last winter’s forecast, so we’ll see where it takes us this year. It’s indicating a strong la nina peak sometime in January (similar to 1998-99) while the Euro suggests a peak NOW with a gradual weakening trend through the winter. Below is the ECMWF forecast – sustaining moderate intensity into February:
http://i18.tinypic.com/52g7vyr.gif
What to believe? Based on past performance, historical climatology, and my own observations of recent SSTA trends, I believe the Euro will be closer to reality in this case — a moderate La Nina, (-1.1c to -1.5c in region 3.4) peaking sometime prior to the first half of December. Wouldn’t be surprised, however, to see some weekly fluctuations down to about -1.5c or even -1.6c as the nina waxes and wanes over the next several weeks.
Pacific Indices: PDO/PNA
I’ve lumped these two indices together as the correlation between the two, with respect to phase (positive/negative) is quite impressive during moderate/strong La Nina years. For the following mod/strong nina yrs since 1950:
1949-50
1954-55
1955-56
1964-65
1970-71
1973-74
1975-76
1988-89
1998-99
The PDO averaged negative during the meteorological winter (DJF) in 100% of cases. For those same years, the PNA averaged negative in 100% of cases as well. Noticing the current PNA trends (positive) one would tend to think otherwise. However, PNA values for October prior to 5/9 (55.6%) of the years listed above were actually positive. No data was available for October of 1949, but the point is October’s PNA often reverses for the DJF period (the NAO has this tendency as well).
1954-55
1955-56
1964-65
1988-89
1998-99
Although all these years saw a PNA averaging negative, there’s sometimes an individual month w/ a positive PNA. 1954-55 had a pos PNA in January, 1955-56 had a pos PNA in December, 1973-74 a pos PNA in December, 1975-76 a slightly pos PNA in both December and January, 1988-89 a pos PNA in December, and 1998-99 a slightly pos PNA in January. One similar aspect of all these years – February averaged negative in every case, and the majority of years saw the +PNA episode(s) terminate prior the middle of winter.
Therefore, I believe the most prudent path is to favor a PDO/PNA averaging negative for the DJF period, but IMO we’ll see one month +PNA, most likely December.
Polar Indices: NAO/AO
The NAO and AO have around a 70% correlation in terms of phase (positive/negative), and considering the difficulty in forecasting the evolution of these indices at long leads, it’s best to assume they work in tandem (for the most part). I believe there are combinations of forcing mechanisms when it comes to the modality of both these indices, i.e., evolution of snow cover in the north latitudes, stratospheric patterns (QBO), pacific atmospheric circulation pattern, and the SSTA profile in the Atlantic is last place IMO. The atmosphere drives these indices, thus shaping the ocean anomalies below it. However, the SST’s can then aid in sustaining the current pattern (feedback). If one examines the SSTA map posted above, the Atlantic signal is more favorable than not at this point in time. Certainly not a classic tripole look, however; we’re beginning to see the emergence of a cold pool E/SE of Newfoundland with warmer anomalies to the north and over the tropical genesis regions.
The October NAO state prior to the following mod/strong nina winters yields some interesting results. In all winters (DJF) which featured an NAO averaging positive, the October NAO preceding it averaged negative
1954-55 – neg NAO winter – pos NAO October
1955-56 – neg NAO winter – neg NAO October
1964-65 – neg NAO winter – pos NAO October
1970-71 – neg NAO winter – neg NAO October
1973-74 – pos NAO winter – neg NAO October
1975-76 – pos NAO winter – neg NAO October
1988-89 – pos NAO winter – neg NAO October
1998-99 – pos NAO winter – neg NAO October
Notice the one, glaring similarity in the above cases – when the NAO in October was negative, the NAO reversed to a positive state in 4/6 (66.6%) of years. In both positive NAO October’s, the NAO signal flipped to negative (average) for the winter. Unfortunately it’s not the greatest sample size, but when including weak la nina years as well, the signal is even stronger. Where are we this year? The official October number isn’t out yet, but it’s apparent when looking at the following graph that the NAO averaged near to slightly positive for Oct:
http://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/nao.sprd2.gif
It’s been hovering near +1 the past couple weeks. This is one factor, IMO, going against a +NAO winter.
Another factor – the QBO – an altering tropical stratospheric wind (between easterly and westerly) which propagates downward towards the tropopause. It certainly has an influence in circulation patterns in the northern hemisphere, thus is partially responsible for sudden warmings of the stratosphere. As we all know, this tends to yield higher than normal heights over the Polar Regions, often allowing an episode of blocking to develop (-NAO/AO couplet?), translating into an overall pattern for the United States. The easterly phase, which we’re currently in (strongly so) favors more blocking up across the Polar Regions.
Finally, decadal trends with respect to the NAO indicate the probability of a negative NAO winter should increase with each passing year. This is not to say we cannot have an anomalous +NAO one winter, however; the chances of a “screaming” +NAO are fairly low, IMO (based upon factors mentioned above as well).
To sum up – I’m favoring a weakly negative NAO/AO winter, with periods of moderate negative and positive values. I also believe the best chance at decent blocking is during the month of December, and again near the very end of meteorological winter.
Conclusion
Before getting to the numbers, I’d like to mention the input of autumn precipitation anomalies / storm genesis regions into my outlook. While precipitation has been near to above normal across the Northeast US for the month of October – I believe it’s more important to note why this was the case. Most of that precipitation did not originate from intensifying coastal storm, i.e., nor’easters – rather, strong cold-fronts w/ low pressure centers cutting SW-NE up the Appalachians and into SE Canada. Storms have had a tendency to develop along this mean baroclinic zone the past several weeks, indicating to me, this may very well be repeated down the road. Keep in mind, however, mean storm track implies there are systems moving to the south and north of (sometimes significantly) the center line. With that said, I’m anticipating the Great Lakes/Midwest region, and up into northern New England, to have the coldest/snowiest winter relative to the rest of the nation. However, the Northeast south of VT/NH/ME (PA, NJ, CT, RI, MA, southern NY) should see a near normal winter in terms of snowfall. I’m expecting a colder than normal December w/ above average precipitation, a slightly warmer than normal January w/ near normal precipitation, and a solidly warmer than normal February w/ near-normal precip. Hence why I’m thinking northern and interior New England will see a snowy winter, as slightly warmer than normal temps in January are still cold enough for snow. It will be a volatile winter overall, with cold blasts even in the midst of a warm mid Jan-Feb period, and anomalous warm blasts as well (possible 65-70F readings). It’s during those transition periods significant snowfall could occur. However, for the I-95 cities, the snowiest period relative to normal will be December 5th-January 15th, IMO. Obviously a huge caveat here for snowfall – one storm can significantly alter seasonal totals particularly from Boston southwestward. Thus I urge you to take the snowfall portion of this outlook with an enormous grain of salt. Overall winter-time departures should be slightly above average for DJF. Below are my estimated numbers for temperature and snowfall in New York City. Keep in mind the monthly temperature ranges leave room for the coolest and warmest solution I’m indicating with the winter DJF average. A -3 Dec, +1 Jan and +2 Feb would yeild 0 – normal winter. A -1 Dec, +3 Jan and +4 Feb would yeild a +2 winter average.
Outlook
Temperatures:
December: -1 to -3
January: +1 to +3
February: +2 to +4
Winter (DJF) average: 0 to +2
Precipitation: Near-normal
Snowfall: Near-normal
Verification of my winter outlook, posted on November 3rd.
First, some key assumptions that were made, then the numbers:
1) Forecast: “I’m anticipating the Great Lakes/Midwest region, and up into northern New England, to have the coldest/snowiest winter relative to the rest of the nation. “
Actual: This was indeed the case — most locations were above normal snowfall from the mid-west into northern new england; some folks near record.
Grade: B+
2) Forecast: “However, the Northeast south of VT/NH/ME (PA, NJ, CT, RI, MA, southern NY) should see a near normal winter in terms of snowfall”
Actual: Winter isn’t done yet but so far — these areas have seen near-normal snowfall. Most places in NE PA, N NJ, CT, RI, MA are within range of their average.
Grade: B
3) Forecast: “It will be a volatile winter overall, with cold blasts even in the midst of a warm mid Jan-Feb period, and anomalous warm blasts as well (possible 65-70F readings).”
Actual: This call worked out quite well I thought — temperatures reached 60F+ on several occasions between the Jan-Feb period.
Grade: A-
4) Forecast: “However, for the I-95 cities, the snowiest period relative to normal will be December 5th-January 15th”.
Actual: This was the case north of NYC — New England had a very snowy early winter. However, not so for NYC, PHL and DCA.
Grade: C+
5)Forecast: “Overall winter-time departures should be slightly above average for DJF.”
Actual: Winter DJF period was +1.7 in NYC.
Grade: A
Local numbers for NYC:
December Forecast: -1 to -3
Actual: -0.3
Grade: B
January Forecast: +1 to +3
Actual: +4.4
Grade: B-
February Forecast: +2 to +4
Actual: +1.2
Grade: B
Overall DJF forecast for NYC: 0 to +2
Actual: +1.7
Grade: A
Snowfall forecast: near-normal
Actual: Not over yet, but so far, much below normal. As expected the most uncertain of my calls. Total: 11.9″ Not horrible bust.
Grade: C/C-
Overall grade: B
In general I thought the temp call was good/very good and the snowfall was good for NYC’s north/west suburbs, but not that great for the city itself.
Of course this is all subjective, so any other comments would be appreciated.