PAST AND PRESENT
Out of the original ten analogues winter temperatures below /Detroit/; six were below normal, three were normal and just one above. Remember this is guidance based on the past /analogues/ with similar hemispheric conditions in place along with other variables considered such as local previous seasons weather to even similar solar activity. They only suggest or "intimate" how the weather will unfold for the next three months and a not to be used as much for exact temperature departures nor timing. The overall analogue temperature forecast from the original ten has worked out very well thus far this winter in predicting a colder than average winter. In fact; only one of the ten suggested a warmer winter. More on temperatures; looking at the monthly average departures, timing or the temperature trend also worked well. December was most likely to be normal or above with January and February, most likely to be below normal. And, depending on exact timing; February had the best chance to be the most below normal /-3.4/. It's been an rewarding winter in this regard since timing /monthly/ trends are not always the best predictors of "exactly when". The same can be said of magnitude or "exactly how much" in regard to both temperature and snowfall. It is up to me as the forecaster to use the data/guidance chosen and interpret and interpolate from there for a season projection.
Between temperature and snowfall, snowfall is usually the more vague or chaotic analogue guidance and this winter has been no exception with a split vote on how the winter would play out for amount. Note, there were five above normal snow seasons, four below normal and one near normal. The snow amount consensus averaged on the high side of normal in the original ten /47.2"- Detroit/ for the southern half of lower Michigan. This area basically extends from Flint to Port Huron (or I-69) south to the Ohio border. Further north; normal to below normal amounts were projected. This snow guidance of past analogue winters was also chosen for this winter.
Out of the original ten analogues chosen for the winter; half of them were singled out to be the better to best performing analogues through mid February. These five analogues thus far, have bested either in temperature and/or snowfall amounts, departures or trends when combined. Note that the February departure alone drops to -6.0 from the -3.4 earlier. Out of the five, only one season contained normal temperatures or below normal snow. While 1924-25 is a weaker analogue, it's snowfall trend was still helpful. Most of these five had similar overall weather patterns across the country and East Coast. Monthly snowfall amounts were quite variable in Southeast Lower Michigan, especially when looking at December and January where it seemed it was feast or famine. This of course, is not unlike this winter (December versus February for example). Out of those five; two are the best performing analogues thus far; 1884-85 and 1977-78 for both temperature and snowfall (first and last analogue).
Best Performing Analogues Thus Far
As of mid February /17th/; Detroit's average temperature is 17.2, or within range of both February's in 1885 /17.2/ and 1978 /16.3/. Latest indications for the remainder of February is bleak if you are looking for milder weather. It really doesn't look like any major warming trend is in sight this month. As stated in the first section; I looked for this past week to be the worst of the winter weather and now February, to be the worst month of the winter when combining both cold and snow. So much so, it may even give January 2014 competition for its frigid mean temperature /16.4/! That would be something but we'll have to wait and see on that one. In addition; up through 2/17; 42.5" of snow has fallen at Detroit Metro Arpt and I expect snowfall to be normal or better from here on out which would add at least another foot and half of snow. This would bring the winter's total near the 1884-85 /60.8"/ and 1977-78 /61.7"/snowfall totals.
Interestingly; some analogue years with less snow /1923-24 &1939-40/ did the best up until February with a mild December, cold January and below normal snow. In fact by late January; 1939-40 looked to be the best - 'ah but winter was not over yet! The truer, facsimile of this winter would be reveled later as time rolled on. Trends in the analogues were for the worst of the winter, cold and snow combined, to be "back-end loaded" and stated as such in my original Outlook. I feel these five analogue winters are similar to the present day meteorological models in a sense; where all can have similar trends but generally, only one or two is the champ. Overall, when taking both temperatures and snowfall into account, the worst of the winter was expected mid-late season - and this has certainly been the case in February. However; it would be remiss of me not to mention that early indications are for bit a moderating trend after the turn of the month in early March
So that's past and present; what about further in the future? Keeping one eye on the analogues and the other on current long term models into spring does not hold much promise for an warm spring. Keep in mind, this early-mid spring outlook is based on the best performing analogues thus far this winter and model preferences. We'll revisit this again when more data is available early in the spring if views change. While I feel the worst of the spring analogues may be a bit too cold /42.4-3rd Coldest/, I still look for temperatures to average below normal. Only one spring in the five analogues in the chart below ended up normal but even that still contained a cold March. Overall; the data is not pretty with many of the Marches in the top 20 coldest along with three of the springs in the top 20 coldest. In fact; there was no warm Marches in the five analogues with a compounded average temperature at 29.5 or 7.7 below normal. April improves some but still averages 3.1 degrees below normal; finally later spring /May/ shows some promise closer to normal. This overall pattern is very similar to last year's spring. As with any temperature predictions; the timing issue could play havoc on the spring temperature trend which would negate somewhat the dominant trends seen here.
Ironically, this local spring analogue data also reflects what the latest CFSv2 model data paints for the region. Precipitation trends also are agreed upon by the model with drier conditions painted for the Lakes Region with precipitation on the analogues averages about an inch below normal.
Temperatures: Below normal into mid Spring (March-April)
Snowfall: Snowfall should end up around normal to above over the
southern sections of Southeast Lower Michigan and normal to
locally below northern sections.
Precipitation: Normal to below.
Making weather fun while we all learn,
Bill Deedler -SEMI_WeatherHistorian