Below is the preliminary version of the Southwest Lower Michigan Fall Forecast. The full version of this article will follow by October.
The official National Weather Service forecast for September through November of 2010 for Southwest Lower Michigan from the Climate Prediction Center (CPC) calls for an enhanced probability for above normal temperatures (Fig. 1). There is an equal chance (EC) for above, below, or near normal precipitation (Fig. 2). Snowfall is expected to be below normal through the fall of 2010.
Fig. 1. The CPC fall temperature forecast (September through November ) made in August.
Fig. 2. The CPC fall precipitation forecast (September through November ) made in August.
The forecast for this fall is primarily based on the interaction of ENSO (El Niño/Southern Oscillation) with the polar jet stream, whose dominant position affects the weather patterns over Southwest Lower Michigan. Beside the ENSO factor, the persistence of seasonal trends is strongly considered in these forecasts as are the dynamical models and soil moisture anomalies.
An El Niño prevailed through the winter months of 2010, but in early May it had weakened and consequently faded to neutral by late spring (Fig. 3). By summer, La Niña conditions had developed and these conditions are expected to prevail across most of the tropical Pacific into the spring of 2011.
Fig. 3. Time series of area-averaged sea surface temperature (SST) anomalies (°C) in the Niño regions [Niño-1+2 (0°-10°S, 90°W-80°W), Niño-3 (5°N-5°S, 150°W-90°W), Niño-3.4 (5°N-5°S, 170°W-120°W), Niño-4 (150ºW-160ºE and 5ºN-5ºS)]. SST anomalies are departures from the 1971-2000 base period weekly means (Xue et al. 2003, J. Climate, 16, 1601-1612).
The August ENSO diagnostic discussion from CPC stated: “Nearly all models predict La Niña to continue through early 2011. However, there is disagreement among the models over the eventual strength of La Niña. Most dynamical models generally predict a moderate-to-strong La Niña, while the majority of the statistical model forecasts indicate a weaker episode. Given the strong cooling observed over the last several months and the apparent ocean-atmosphere coupling (positive feedback), the dynamical model outcome of a moderate-to-strong episode is favored at this time. Therefore, La Niña conditions are expected to strengthen and last through Northern Hemisphere Winter 2010-11.”
This is a significant forecast since the state of ENSO, especially if it is moderate or stronger, impacts the weather over Southwest Lower Michigan. It was primarily for this reason the CPC has a 40 percent greater than normal probability of above normal temperatures for Southwest Lower Michigan this fall (Fig. 1)
There have been seven recorded years (1964, 1970, 1973, 1988, 1995, 1998, and 2005) when, like this past year, there was a moderate or strong El Niño during the previous winter that faded to neutral by late spring and then to a La Niña by late summer. By looking at the falls that followed these years, it is possible to make a prediction for the upcoming fall. This is known as "analog forecasting", since years with analogous conditions are used to forecast the outcome for the current year. Looking at the subsequent falls for these seven analog years, most fell in the top third warmest falls. (Fig. 4). This represents a fall with an average temperature of at least 1.2 degrees F above normal. Conversely, a fall with an average temperature less than 1.2 degrees below normal is in the bottom third of warmest falls. By comparison, the average of the warm analog years was 2.5 degrees above normal. This all strongly suggests the fall of 2010 will be in the top third of warmest falls.
Fig. 4. Fall temperature anomalies for Southwest Lower Michigan when an El Niño winter is followed by a La Niña fall (light red), and when a moderate or stronger El Niño in the winter is followed by a La Niña fall (dark red).
Using the average of fall precipitation following the analog years, there was little difference from from normal precipitation amounts, which are calculated from the 1971 to 2000 average (Fig. 5).
Fig. 5. Fall composite precipitation anomalies for La Niña falls that were preceded by an El Niño in the previous winter, compared to the 1971 to 2000 normals.
There is a trend toward drier falls; that is, of the past fifteen falls, ten were drier than normal and four were wetter than normal (Fig. 6). Within the past ten falls, four were drier than normal and four were wetter than normal.
Fig. 6. Fall precipitation departure from normal from 1995 through 2009 for all available Southwest Lower Michigan climate stations .
La Nina falls tend to be drier than normal over Southwest Lower Michigan. Precipitation from all past La Nina falls (nineteen in all) were averaged and compared to the 1971 to 2000 mean precipitation. The result was below normal precipitation over much of the United States, including Southwest Lower Michigan (Fig. 7).
Fig.7. Precipitation average in inches for La Nina falls (19 years) minus the 1971-2000 long term average.
Putting all of this information together, the precipitation analog forecast based on previous La Nina falls suggests a drier than normal fall. However, the analog forecast based on previous years with El Niño winters followed by La Niña shows no preference towards dry or wet falls. It should also be noted that just one or two significant storms could easily result in a wetter than average fall, even though the number of days with precipitation may be well below normal. It is for these reasons we will go along with CPC and suggest the best forecast is an equal chance (EC) for above, below, or near normal precipitation (Fig. 2).
When there was an El Niño in the previous winter which progresses to a La Niña by the fall, snowfall in the fall is typically well below normal (Fig. 8) . For these analog years, Lansing had no falls with above normal snowfall. Grand Rapids was well above normal in 1995 but was below normal in six other falls on the graph below. Muskegon had two falls with above normal snowfall, 1964 and 1995. This likely is because 1995 was a cold November which resulted in enhanced lake effect snowfall, explaining why Muskegon and Grand Rapids had above normal snowfall in 1995 while Lansing did not.
Fig. 8. The La Niña snowfall composite anomaly for the fall at Grand Rapids (red), Lansing (blue) and Muskegon (green).
While the full seasonal snowfalls (July 1 through June 30th of the next year) are not shown for these analog years, it should be noted that Muskegon recorded well above normal snowfall for each of those past seasons. Grand Rapids had above normal snowfall in nearly all of those winters too, as did Lansing. It therefore is quite possible that the snow totals will be lower than normal through the fall, but then be near or above normal starting mid-winter across all of Southwest Lower Michigan.