Local Winter 2008/2009 Winter Outlook

The NOAA’s Climate Prediction Center (CPC) issued its regular monthly and seasonal outlooks for the United States on November 20 2008.  These outlooks are issued on the third Thursday of every month, and include probabilities of above normal or below normal temperatures and precipitation, or areas where normal climate variability is expected. The national outlook for the 2008/2009 winter season is available here.

 In addition to the larger, national guidance the CPC produces there are "downscaled" climate outlooks available. These downscaled outlooks take the local temperature trends into account to better refine the expected average conditions during a particular season. These downscaled climate outlooks are available by clicking here.

 Below is a technical discussion of the factors used to create the local winter outlook. Click here to jump down to the actual outlook.

This winter season is expected to be dominated by a weakly negative El Nino/Southern Oscillation (ENSO) type pattern, a pattern of slightly below normal temperatures in the equatorial Pacific. La Nina, and its warm counterpart El Nino are regular fluctuations in the Pacific that affect the weather on a global scale. Historically, La Nina conditions in the Pacific tend to produce colder and wetter winters across the Northern Plains. Several other large scale climate factors come into play as well, including the North Atlantic Oscillation and the Pacific-North America patterns. These patterns can intensify or weaken the affects of the La Nina or El Nino signal, and are often the cause of tremendous variability during the winter season.

The CPC also takes other factors into consideration such as the trends in temperature and precipitation patterns the past 10 to 15 years, as well as a variety of numerical and statistical computer models. Based on these factors, the CPC expects equal chances for warmer than normal tempertures during the winter months of December, January and February. The same is true for precipitation, with normal climatic variability anticipated by the CPC.

The following is the 2008/2009 Grand Forks NWS Cold Season Outlook.  Remember this outlook is based upon ongoing research at the NWS in Grand Forks.  This product is made available to demonstrate evolving capabilities in the climate forecasting field. This is not intended to replace the official outlook, forecasts or warnings. Always refer to the latest outlooks and products from the Climate Prediction Center for the official products. Read the caveat at the bottom of this product.

The main considerations to this update were the changes in the 4 major signals considered (see Simple Pattern Recognition below), forecasts based on the deterministic computer models, and historic considerations. One primary goal in a second issuance was to see if the logic behind the first outlook was good, to see if the pattern in the first two, "spin up" months was as expected, and watch for any powerful fall signals to emerge. To that end, there were minor changes in the composite years used, based on the same techniques used in the first outlook. The PDO continues to exhibit a strong cold phase signal, the AMO a strong warm phase, the ENSO remains statistically neutral (with cool phase, or La Nina tendencies)

Basis: The NWS Climate Services Division encourages WFOs to stretch the science of climate “out-looking” by a variety of methods. The intent is to produce guidance for use when briefing constituents, users (both internal and external) as well as the development proof of concept techniques. While the increase in our understanding of the global atmosphere, its various signals and mechanisms, has increased, there is a great deal we do not fully understand.

Techniques:  There are several methods used to develop the NWS Grand Forks seasonal climate outlooks. These include simple pattern recognition, comparing the antecedent conditions to current then extrapolating those forward, developing correlation's between the Low-Frequency Variability modes of the atmosphere and seasonal responses, developing a local climate forecast verification/bias correction model.

Simple Pattern Recognition: Analogous to short term forecasting, understanding how the large scale climate affects the local weather is critical. Developing and understanding long wave orientation and movement, with the resultant ‘weather’ is one part. An understanding of how events literally on the other side of the globe affect the weather here is another. Finally, proper use to observational and diagnostic tools to analyze those forces is critical. The 4 primary forcings used are the Pacific Decadal Oscillation (PDO), Atlantic Multidecadal Oscillation (AMO), East Pacific/North Pacific (EP/NP) and the El Nino Southern Oscillation (ENSO). Other important signals that are significant but have more of an intraseasonal impact are the Madden-Julian Oscillation (MJO) and the Arctic Oscillation (AO). The MJO and AO modulate the PNA and NAO which are, in the climate sense, short term oscillations, and much more difficult to predict their phase.

The East Pacific/North Pacific (EP/NP) oscillation tends to be quite variable intraseasonal, but is useful in the longer term. As a rule, in the Red River Valley, large scale wet cycle/precipitation peaks tend to cycle with the EP/NP peaks. There was a peak in the EP/NP in the early to mid 90s, and again in the early part of this decade. Recently, the EP/NP has been on a downward trend, suggesting a decrease in precipitation frequency. However, based on this past spring-fall wet cycle, there may well be other forcings at work overpowering the EP/NP signal.

Techniques used in this outlook: There is a variety of research already available which goes into greater detail than I wish to do here. The basic principle involved include 1) to correctly ascertain the last 3 season's climate and the atmospheric mode(s) and forcing(s) responsible for the climate 2) using tools available at the Earth System Research Lab (ESRL, formally Climate Diagnostics Center) determine the primary, secondary and tertiary forcings 3) develop analog years corollary to the mode(s) in step 1 and test for statistical relevance 4) obtain the base data and develop graphics and tables based on the years determined to be relevant.

Years chosen: Based on the technique described, we used the following years for input to the analog in the outlook: 1955/56, 1957/58, 1961/62, 1962/63, 1963/64, 1971/72, 1972/73, 1978/79 and 1989/90. Some of these years are based on the phase and tendency of the AMO, PDO and ENSO. The ESRL has statistical tools available to determine the relative correlation of each signal, and many more, described. Please note there are years prior to 1950 that also fall into the analog years, but were not chosen to keep the sample smaller and more contemporary. Finally, there was still a fair amount of variability in the individual winters chosen; removing the warm years resulted in a minor change, to a slightly cooler winter season outlook.

Winter season 2008/2009 Outlook: This outlook is based on ongoing research at the Grand Forks NWS, and is designed to refine the national outlook issued. A mid winter update is scheduled for January 2009.

The Outlook: Essentially we expect temperatures to be "normally" cold the next three months, with a bias towards cooler than normal. Snowfall is, on balance, expected to be below the long term average; however there will be tremendous variability across the region. Early indications are the higher snow totals will be closer to the International Border.

Graphics of temperature, precipitation and snowfall are below. These tables and graphs are based on the years noted on the previous paragraphs.  The following tables are specific to the UND/NWS Grand Forks Climate Station.


NWS/UND
October to April Averages*
         Climo          Composite Outlook
Max(F)   33.1            31.2 to 33.1
Min(F)   14.3            12.7 to 14.6
Pcpn(In)  6.5             6.2 to  7.50
Snow(In) 44.1            40.2 to 51.0

DJF Specific

         Climo           Composite Outlook
Max(F)   18.7            17.0 to 19.5
Min(F)    0.7            -2.3 to  0.2
Pcpn(In)  1.9            1.77 to  2.90
Snow(In) 25.4            20.9 to 26.4

The following tables are specific to the Fargo Area

Fargo Area
October to April Averages*

         Climo          Composite Outlook
Max(F)   35.4            34.1 to 36.0
Min(F)   15.9            14.0 to 16.1
Pcpn(In)  7.5            5.30 to  7.00
Snow(In) 47.6            30.6 to 42.5

DJF Specific

         Climo           Composite Outlook
Max(F)   19.8            18.1 to 20.1
Min(F)    2.4            -0.5 to  1.5
Pcpn(In)  2.0            1.38 to  2.18
Snow(In) 27.8            17.8 to 25.3

*Some data are rounded. These are expected ranges based on the above technique and should not be used as an explicit forecast.

Basically, the overall cold season will be one of tremendous variability. The coldest months, that are the ones with the greatest departure from normal, are expected to be December and February.  Total snow will be close to average; statistically normal (30 to 50 inches is statistically “average”). Two distinct cold periods are apparent in the composites, one from early to mid December, and a stronger cold snap late January to mid February. As is often the case in weak La Nina signals, the March and April time frame tilts predominantly to the cold side of climatology. Looking at the individual winters, the heavier snowfalls were mainly across the northern Red River Valley, however the differences were statistically very close.

The graphics below indicate the expected average temperature departure across the Devils Lake, Red River and Minnesota Lakes country during the heart of the winter season (Click to expand image).

Winter Temperature Outlook

The graphics below indicate the expected average snowfall across the Devils Lake, Red River and Minnesota Lakes country during the heart of the winter season (Click to expand image).

Winter Snowfall Outlook

Caveat Emptor:  These are outlooks, not forecasts and not designed for literal interpretation.  As example, the snow fall values that appear in both Fargo and UND data should not be used as a "forecast". The ranges given represent an average variance in the years used to build the outlook. Some of the winters used had well below average snowfalls, while most were within the statistical range of normal.

Remember these outlooks will only be as good as the analogs. If the wrong years are built into the analogs chosen, the outlook will be no worse than one based on random chance or chaos. That is why there will be an early January update to account for errors in analog year choices.

These outlooks will differ from the Climate Prediction Center (CPC) Outlooks. CPC outlooks are currently based on two dominant signals; the phase of the ENSO and the past decade trend in temperature. At this time the CPC does not currently use the PDO, AMO, or any other of the less predictable signals to make outlooks.

For more information contact the Grand Forks National Weather Service at 701.795.5198, or visit us at www.crh.noaa.gov/fgf

Mark Ewens, Climate Services Focal Point  Mark.Ewens@noaa.gov

Updated November 28, 2008



Return to News Archive

USA.gov is the U.S. government's official web portal to all federal, state and local government web resources and services.