Winter 2002-03 Outlook

Written by: William R. Deedler, Weather Historian
October 2nd, 2002

It has been five years /1997-98/ since the climatic event "El Niño" last occurred in the Pacific. The event is nothing new and has been going on for centuries. Reliable data to measure El Niño is available to scientists for interpretation. El Niños are detected by a host of meteorological equipment and methods including weather satellites, moored and drifting buoys, and sea-level temperature analysis. This data is used in computer models of the oceans and atmosphere to help predict the evolution and dissipation of the El Niño life cycle.

The term El Niño is Spanish for the "boy child" and refers to the Christ child because the effect generally is associated with the month of December and Christmas. The term was coined by the Peruvian fisherman as far back as the 1600's to describe the change in the Pacific Ocean currents off the coast of South America. The phenomenon originates in the western tropical Pacific Ocean and generally occurs every two to seven years. The normal easterly trade winds relax and a westerly wind begins to dominate. As a result, the warmer surface sea water of the tropics spreads east and northward toward the South and North American continents. This causes a stronger atmospheric link to develop between the Pacific tropics and the higher latitudes and thus, invokes a shift in mid-latitude weather patterns including those of North America.

Along with the shift in the mid-latitude weather patterns, abnormally low pressure develops over the tropics in the eastern Pacific, while unusually high pressure takes shape over Indonesia and Australia. Normal global atmospheric circulation patterns are disrupted and since the U.S. is generally downwind of El Niño, the effects can be quite noteworthy. As a result of El Niño, the tropical connection or "pineapple express" increases across mainly the southern U.S. in the winter bringing additional storminess and rainfall. At the same time, the northern U.S. tends to see a more pleasant winter with relatively milder and drier conditions. While this is the general accepted scenario, other factors along with the El Niño "effect" must be examined and considered, particularly when a weaker El Niño occurs.

As mentioned, five years ago /1997-98/, a strong El Niño occurred in the Pacific. The resultant 1997-98 winter was abnormally warm with well below normal snowfall across much of Southeast Lower Michigan. Note the closer to normal season snowfall, however, at Saginaw which was closer to the dominant storm track. (Chart-1)

                              (Chart - 1)

                         WINTER OF 97-98
           AVERAGE   TEMPERATURES (97-98) 
                   (Note: Old Norms 1961-1990)
           December    January    February    Winter   Normal
Detroit     32.3        32.8        36.7       33.9     25.5 
(Depart)   (+4.0)      (+9.9)     (+11.3)     (+8.4)

Flint       29.9        29.9        34.6       31.5     24.1
(Depart)   (+2.7)      (+8.4)     (+11.1)     (+7.4)

Saginaw     26.2        28.6        33.4       29.4
(Depart)   (+4.9)      (+8.0)     (+10.9)     (+6.2)    23.2

                     PRECIPITATION (97-98)
                  (Note: Old Norms 1961-1990)
          December    January    February    Winter    Normal
Detroit     1.89        2.78        3.60       8.27      6.32
(Depart)   (-.93)     (+1.02)     (+1.86)    (+1.95)
Snow/Dpt  6.1/-4.8    8.3/-3.5     T/-9.2      22.5      41.2

Flint       1.06        3.16        1.65       5.87      4.78
(Depart)  (-1.05)     (+1.77)      (+.37)    (+1.09)   
Snow/Dpt  6.9/-4.4    8.0/-4.3    0.4/-9.9     28.3      45.1

Saginaw     0.78        4.04        2.79       7.61      5.37
(Depart)  (-1.59)     (+2.41)     (+1.42)    (+2.24)
Snow/Dpt  6.7/         14.4         0.6        40.1      46.0

The Climate Prediction Center forecast models indicate that moderate El Niño conditions are likely to continue through the Winter of 2002-03. Although there is some uncertainty in the forecasts about the timing and intensity of El Niño, all of the climate models indicate that it will indeed be weaker than the strong El Niño experienced in 1997-98. At this time, it is anticipated that this winter's El Niño will, at best, remain in the moderate category.

As important as the progress of El Niño will be the evolution of the North Atlantic Oscillation (NAO) and the Eastern Pacific Oscillation (EPO). The NAO, basically, is the dominant upper wind flow pattern over the North Atlantic influenced by the ocean. While in a negative phase, the NAO sometimes tends to act as a block (or dam) to the upper wind flow over the eastern half of North America. This blocking effect, in turn, tends to deliver the polar/arctic air into the eastern half of the country and Great Lakes more readily.

The Eastern Pacific Oscillation (EPO) is the upper wind flow over the Eastern Pacific influenced by the ocean. When in a positive phase, the EPO generally is reflected by dominant stronger zonal flow and/or troughing along the West Coast of the U.S. This is in response to lower pressure over the Gulf of Alaska that extends south off the coast of Western Canada, while higher pressure dominates across the central Pacific. This combination, in turn, tends to funnel milder Pacific air well inland into the country and thus, limits arctic outbreaks by holding them at bay up in Canada. When the EPO is dominated by a negative phase (as with the NAO), more ridging develops along the West Coast as higher pressure extends from the Gulf of Alaska south along the West Coast of Canada (opposite of the positive phase). This, in turn, encourages a northwesterly flow from Canada into the middle and eastern sections of the US and thus, the delivery of polar or arctic air. Last winter was very mild and this is reflected well by the positive phase dominance of both, the NAO and EPO.

Along with the aforementioned factors, other considerations are taken into account for this winter outlook. First are the weather trends of the last year and how well they relate to the weather trends proceeding the 13 winters in the study. Second, is a check of the solar activity during the years of the past winters studied. It is known that sunspot activity as an impact on climatic trends and patterns over the earth. The current sunspot cycle is then compared with sunspot cycles (timing/peak/intensity) of past winters in the study. During the upcoming winter, sunspot maxima are expected to continue to wane from the peaks attained during the past few years. Additional factors include, current soil conditions (moisture and temperature) and autumn Great Lakes water temperatures.

The following winters were researched under weak to moderate El Niños:

                                                    (Chart - 2)

Winter           Average Temperature                Snowfall Year
           Detroit    Flint      Saginaw        Detroit   Flint    Saginaw

30 Year
Normals-   27.1       23.9       24.0           44.0      48.3       44.5 
100 Year
Average-   26.5        -          -              -         -          -

1881-82    37.0 -1-    -          -             13.2 -2-   -          -
1885-86    28.2        -          -             56.7 (16)  -          -
1913-14    27.1       24.2*      23.7           42.3      39.0*     47.7 
1923-24    28.3       27.0*      24.6           37.7      34.4*     31.9 
1931-32    35.6 -2-   33.4*      33.2 -1-       26.2      28.4*     31.0 
1939-40    26.4       25.9       24.4           32.2      28.2      31.0 
1953-54    31.1 -13-  28.1 -8-   27.2 -15-      40.0      48.7      46.1
1963-64    27.7       24.8       24.3           32.5      36.6      21.9 -6- 
1972-73    27.8       27.3       23.8           45.0      62.9 (6)  65.8 (8)
1977-78    20.4 (7)   19.1 (5)   17.9 (6)       61.7 (8)  50.6      55.6 (20)
1987-88    26.9       24.7       24.4           45.1      39.2      29.0
1991-92    30.4 -17-  28.2 -6-   27.0 -17-      43.5      54.4 (13) 46.2
1994-95    29.6       26.0       26.1           33.5      42.5      28.5
Averages:  29.0       26.2       25.1           39.2      42.3**    39.5**
(  ) Coldest or Snowiest Ranking 
-  - Warmest or Snowless Ranking
*Not an Official Record and not in ranking (Official Records Began in 1942)
**Sample Incomplete and Therefore Not a reliable Snow Average



In researching the above winters, the trends during the proceeding year (since the winter of 2001-02) were noted and compared to the weather in the above winters and their proceeding year. Any occasional or frequent recurring weather correlations were noted between the years. This research was primarily based on the most complete database of records at Detroit.

Several similarities were seen between this past year and a number of years in the list. During several of the preceding summers and/or falls, notable dry spells or droughts were ongoing across significant portions of the country especially in: 1881, 1913, 1931, 1953, 1963 and 1977.

Probably the most consistent trend noted in the summer into, at least, early fall preceding these winters was the overall above normal temperatures. A good number of these summers and/or falls averaged above normal. The years that saw this trend included: 1881, 1913, 1931, 1939, 1953, 1963, 1987, 1991 and 1994. So, out of the 13 winters researched, an impressive nine were preceded by an overall warmer than average summer and/or fall. The most notable (warmest) of the lot (and also in top 20 warmest summers and/or falls lists for Detroit, Flint or Saginaw) included 1931, 1939, 1953, 1963, 1987, 1991 and 1994 (and while the summer of 1977 did not average above normal, that July made the list of top 20 warmest Julys). This warm summer pattern matched very well with our past summer, especially at Detroit and Flint, which both placed in the top 20 Warmest Summers list.

During the autumns, the majority of Septembers through Novembers averaged normal to above for temperatures, especially in early fall. All in all, there were some nice and warm Septembers and/or Octobers in this study. Out of the 13 Septembers, eight averaged normal to above (some significantly, not unlike this past September) and five averaged below normal. It must be mentioned, however, that when a September did average below normal, it was generally by just a small departure. Eight of 13 Octobers also averaged normal to above (and as September, a few also by a large margin). In contrast to the minor below normal departures found in September, it was observed that when October averaged below normal, it was usually by a more notable departure. Unseasonably cool Octobers occurred in 1972, 1977 and 1987 when temperatures averaged in the mid 40s /normal 51.9/. Moving on into November, here too, eight of 13 Novembers averaged normal to above. And, like October, when the below normal departures occurred, they were generally more significant. Some cool Novembers occurred in 1939, 1972 and 1991 when temperatures averaged in the mid to upper 30s /normal 40.7/. Thus, in spite of the generally warm falls, the chances of a cool October or November jumped, as evidenced by the three cool Octobers and Novembers, all in different years (1939, 1972, 1977, 1987 and 1991). The only exception was 1972, when both October and November averaged below normal.

In spite of the dominant early warm autumns, it was interesting to note that frost/freezes were generally seen over much of the region at pretty much their normal time period (early - mid October). These cold snaps tended to be aggressive and were quite a contrast from the dominating mild fall weather. They were generally followed, at some point, by Indian Summer weather (for more info) well into November with temperatures pushing back up into the 50s, 60s and even 70s. Besides the aggressive cold snaps early-mid October, distinctive pattern changes to colder weather sometimes occurred during the month; this pattern change resulted in the aforementioned cool Octobers.

Precipitation (including snowfall):

A clear majority of the summer into, at least, mid fall periods contained drier than normal conditions. Generally, the precipitation trend that dominated during the summer held on into early to mid fall. Late fall into winter saw more mixed results and was dependant on the strength of El Niño (and resultant storm tracks) and the EPO/NAO phase. Some periods from late October through November became wetter (and snowier, see below), while others remained dry. The years that the entire fall was wet and contained above normal rainfall were limited to just two, 1881 and 1931. The above normal departure was very substantial, however, in the autumn of 1881 when 13.74 inches fell in Detroit (normal, 8.16 inches). And maybe not surprising, the autumn of 1881 still remains at the top of the list for wettest falls in Detroit. The 1931 above normal departure is much tamer at just +.80 of an inch with 8.96 inches falling.

While the chances of a frost/freeze occurring near the average dates were good, the dates when the first snow flakes (trace) flew were more variable. While October historically sees at least a trace of snow by the end of the month, the majority (eight) of these particular Octobers did not see any snow. The five Octobers that did see snow had just a trace. And, while below normal snow dominated the Novembers at Detroit with eight of 13 having below normal snowfalls, out of the five that did contain normal to above snowfall, two (1972, 1977) were significant with nearly triple the normal /2.7/ at Detroit (more in "contrary indicators" below).


Strength of El Niño and Phases of the EPO/NAO Seen as Vital Elements to This Winter's Outcome

Now we come to the essence of this research: what were these weak to moderate El Niños really like?

There was also a fairly consistent correlation between the strength and extent of El Niño and the resultant weather. In other words, the stronger and more widespread the El Niño was, the more likely it would be milder in Southeast Lower Michigan. This certainly was the case with the more recent strong El Niños of 1997-98 and 1982-83. Unfortunately, more "problems" or variances arise when the El Niño holds in the weak to moderate category (including just how extensive are the warmer than normal water temperatures in the Pacific).

It was noted that while the majority of weak to moderate El Niño winters in Southeast Lower Michigan averaged normal to above normal temperatures, snowfall results were more variable. Subsequent questions arise about the interaction of the more active Pacific subtropical jet (due to El Niño) and the phases of the EPO and NAO previously mentioned. Reviewing the earlier archives of the EPO and NAO trends available (since the winter of 1953-54 in this study), may shed at least some light on these variances.

Contrary Indicators:

Perhaps none of the winters challenged the prevalent below normal snowfall trend (and above normal temperatures when looking at 1977-78) more than the winters of 1885-86, 1972-73 and 1977-78 when snowfall was significantly above normal at most locations. While the detailed EPO and NAO trends were not available for 1885-86, they were for the winters of 1972-73 and 1977-78. Also, in addition to the being snowy, why was the winter of 1977-78 so much colder than the rest in the study? Reviewing the earlier archives of the EPO and NAO, along with the 500 MB upper wind pattern, may help understand the differences between the harder winters of 1972-73, 1977-78 and the dominant milder ones. Is there some trend to watch for in the EPO, NAO and upper air pattern this fall that occurred in 1972 and 1977?

The winter of 1977-78 unfolded while a weak El Niño was in place. Since it was a generally rough winter, it was not surprising to find that a negative phase dominated both the EPO (1977 | 1978) and NAO (1977 | 1978) that winter. These negative phases of the EPO and NAO worked in conjunction to help deliver the polar and arctic air into the Great Lakes region. This, in turn, more than offset any warming from El Niño. The summer and early fall proceeding that winter were not unlike the majority in the study as far as above normal temperature. Rainfall, however, was more plentiful than most of the summer-fall periods in the study. An important change came in October (and again, one of the few cool Octobers in the study) when abnormally cold air surged into the region mid month. While the timing of the polar blast duplicated most in the study (mentioned above in the frost/freeze data) its strength and dominance was atypical. The monthly average temperature for October 1977 was a cool 47.9 degrees /normal 51.9/ at Detroit. Curiously, this same pattern was noted in the fall of 1972 and again, this abrupt change surfaced in October. October of 1972 not only turned out to average well below normal, but with nearly the same average temperature as October 1977 with 47.3 degrees. The subsequent November averaged near normal in 1977 but below normal in 1972 while both contained well above normal snowfall. While the current fall remained warm until mid October, the recent abrupt change to a deep 500 MB trough over the eastern half of the country resulting in much colder weather bears watching, especially if it represents an intermediate trend change to the past several months. In addition, the near-term projections of the EPO and NAO both indicate, at least short term, negative phases and this too bears watching for any overall, longer term trend change.

Negative phases of the EPO and NAO more often than not prevailed the remainder of the 1977-78 winter, but to a lesser extent in 1972-73. Both winters displayed a fairly active southern stream storm track (aided by El Niño). In fact, it would be remiss if it wasn't mentioned that during the weak El Niño winters of 1885-86 and 1977-78 in this study, two very powerful snowstorms crippled much of the Great Lakes. Many can recall the January 26th, 1978 Blizzard whose 25th anniversary will come this winter (watch for the write-up mid winter) which brought the lowest pressure reading on record to a number of weather stations with its visit (including Detroit with 28.345 inches). The biggest snowstorm (in terms of amount) ever to clobber the Detroit area since records began occurred in April (yes April) 1886 (see "Tale of Two Storms") when 24.5 inches fell at Detroit! In the 1978 storm, a strong phasing of the subtropical jet and an evolving negative EPO and NAO, help set the stage for the creation of that monster. In addition, an intense upper level disturbance screamed southeast out of Canada and phased with the subtropical jet, deeply intensifying the surface low pressure developing over the Southeastern U.S. as it raced north northwest. This meteorological "bomb" blossomed over the Upper Ohio Valley into the Great Lakes and Southern Ontario.


While the temperatures during these winters did tend to average normal to above normal overall, they were not without their arctic or polar plunges and occasional storms. These polar blasts, however, tended to be more transitory in nature rather than a permanent winter visitor and, when they visited, it tended to be more in mid to late winter. The exceptions, however, were during the few years mentioned when the winter actually "began" somewhat early in November. The majority of the remainder of the winters researched began early to mid December, generally by way of a strong arctic plunge (or again, an abrupt change from the earlier, milder dominant temperature pattern). This was directly related to either the NAO or EPO (or both) becoming decidedly negative and had little bearing on El Niño. The two unseasonably warm winters of 1881-82 and 1931-32 averaged 8-10 degrees above normal and snowfall was well below normal. In both these abnormally warm winters (1881-82/1931-32), winter did not really "begin" until January (what winter there was).

The month of December saw a clear majority (11) averaging normal to above normal in temperatures. Monthly average temperatures ranging from the upper 20s to mid 30s were commonplace at Detroit, Flint and Saginaw. The incredibly mild December of 1881 averaged 40.6 degrees at Detroit. Several of the Decembers in the research placed in the top 20 warmest Decembers on record. At Detroit, five of the 13 Decembers placed in the top 20 warmest, while in the Flint area (using co-operative data since 1893), four placed in the top 20 warmest Decembers and likewise in Saginaw (since 1899).

January into February also saw more normal to above normal temperature departures over below normal departures. While nine out of 13 Januaries averaged normal or above, that number dropped to seven (above normal) in February. Therefore, a trend was noted that while the chances were fairly good that while any given winter month may average normal or above, those odds dropped as the winter evolved (remembering that 11 out 13 Decembers averaged above normal). One possible explanation could be that with time, the cold air that had stored up in Canada was able to break through El Niño's warming affects (and more zonal flow), particularly if the El Niño effect was weak enough to allow it. Also, it must be mentioned that in both January and February, there were also some impressive below normal departures from bitter cold, though they were in the minority. There was no strong trend seen in January and February in the winters that started early in November. The chance of a warmer period mid or late winter (ex:"January Thaw") in these minority of winters that started out cold, however, was better than 50 percent.

Precipitation (including snowfall):

Generally, only the strongest of El Niños bring unseasonably mild and relatively dry winters to Southeast lower Michigan. As mentioned above, the late fall into winter period saw more variability in rain and snowfall totals. During the strong El Niño of 1982-83, only 20 inches of snow fell in Detroit, 33.6 in Flint and just 23.8 in Saginaw. In the more recent strong El Niño of 1997-98, Detroit saw just 22.5 inches of snow, 28.3 fell in Flint and with Saginaw nearer the storm track, snowfall totaled 40.1 inches.

Snowfall totals, however, varied considerably with weak to moderate El Niños in place. While overall snowfalls tended to average a bit below normal, only one of the years /1881-82/ appears in Detroit's top 20 snowless winters. Interestingly, none of the researched winters appear in Flint's top 20 snowless winters (even back to1883) and only the winter of 1963-64 winter made the list at Saginaw. It was intriguing, however, to find one of the years /1977-78/ made the top 20 snowiest winters at Detroit, while 1972-73, 1977-78 and 1991-92 each made Flint's snowiest list, with 1972-73 and 1977-78 both making Saginaw's snowiest winters list.

As mentioned earlier, these winters tended to contain the majority of their snow mid to late winter, BUT NOT ALWAYS. It was the years mentioned above where a trend to colder weather took hold in October (1972, 1977 and 1987) also saw above normal snow in December (with 1972 and 1977 also containing above normal snow in November). These were the aforementioned winters that "began" earlier than the rest in November.

Storm tracks depended largely on the type of dominating weather. In the mildest of winters, generally clippers frequently ushered pacific orientated air into the region with just the infrequent polar intrusion. Southern stream low pressure systems tended to stay well south of the Great Lakes or move west of the area with just the occasional hit and thus, the lowest snowfall totals occurred. In the more active winters, where temperatures averaged within a few degrees of normal or below (1977-78), clippers were busy draping pacific, polar or arctic air across the nation. The dominant southern stream lows aided by El Niño generally tracked across the Southern Plains into the Great Lakes ("Colorado" or "Texas Panhandle" Lows). A third path tracked across the Southern U.S. into eastern Texas, then moved northeast either into the Ohio Valley or up the East Coast.


In this study, while the majority of winters across all of Southeast Lower Michigan averaged a few degrees above normal and snowfalls about half a foot below normal, there were notable exceptions. And, while snowfall amounts tighten-up a bit (less spread out) when compared to the last few winter outlooks, the two extremes still varied greatly. In the majority of winters, however, snowfall totals across Southeast Lower Michigan averaged between 30 and 45 inches. Also, as was noted in last year's winter outlook (and in the winter 2001-02 review), these weak to moderate El Niño winter's more often than not, also had a tendency to be back-end loaded. Meaning, the likelihood of snow and/or cold increased as the winter evolved. A lot will depend on El Niño's STRENGTH and the DOMINANT PHASES and TIMING of the EPO and NAO.

Therefore, because of some notable variances between these winters, and the other "contrary indicators" mentioned above, a broader range on the outlook was adopted. The above data suggests temperatures across Southeast Lower Michigan will average around normal to above, while precipitation (and snowfall) averages around normal to below.

Look for a complete Winter 2002-03 write-up come spring! is the U.S. government's official web portal to all federal, state and local government web resources and services.