March Warmth As A Guide To Summer Temperatures

During each transition season of spring and fall, people incorrectly attempt to predict the next season's warmth or cold by using the current season's temperatures as a guide. The conversations go something like this:

"This warm March means we're in for a super hot summer" or "This cold and rainy fall means we're in for tons of snow this winter." or some other combination like this.

Does this line of thinking work?

I checked the top 10 warmest March years for 8 cities (Cleveland, Cincinnati, Detroit, Indianapolis, Cincinnati, Pittsburgh, Milwaukee and Buffalo) and ranked them. I found the top 10 warmest years that occurred the most frequently for all of the cities and plotted a summer (June through August) composite (blend) temperature map of these years weighting the more frequent--top 6--warm years higher.

Summer blend - Most Frequent Top 10 Warm March Years - Top 6 Weighted Higher

I next plotted the blend of only the top 6 most frequent years of occurrence that were previously weighted. Notice the warmth become more widespread in this blend.

Neither one of these composites above are blast furnace summers by any stretch.

The problem is that both of these composites mask the individual year extremes. Look at each of the top 5 most frequent years of occurrence?  Each summer varies a great deal. A more detailed look at March 2012 (hottest on record in northern Ohio HERE) 

When we create our summer outlook each year, we perform an in-depth analysis of specific current conditions and combine it with computer projections of these conditions into the future. We combine this with some statistical analysis (similar to the above maps) to come up with a consensus outlook. Its more complicated than saying "its warm or cool now therefore it will be warm or cool next month/next season".

Bottom line, seasonal outlooks are very complex.  Don't be fooled by monthly warmth especially in early spring. It doesn't necessarily mean that an equally warmer than normal summer is ahead.

More on the summer outlook later in April.

The Upcoming Winter Weather (2015-16) MIGHT Look like This

North Country Public Radio

I'm thinking out loud today so bare with me.

Over the last few days, I was playing around with the constructed analogs of similar winter years. Using a set of 6 best fit years and weighting them equally, I constructed a blend for each month starting in November and ending with February showing the upper level pattern across the North American Continent. The color colors show where the storm system would develop (Low pressure). The warmer colors show where High Pressure systems reside.

The first burst of colder air occurs in November, the pattern relaxes in December then reloads after the first of the year. Can you see how the southern jet stream becomes very active, dominant force in the second half of the winter. February shows the much higher frequency and strength of low pressure systems from the southwest. This indicates a higher propensity for wetter snows from northern Texas, northeast into the Mid Atlantic states and Ohio Valley similar to 2010.


UPPER AIR PRESSURE (500 mB)

NOVEMBER BLEND

DECEMBER BLEND

JANUARY BLEND

FEBRUARY BLEND

SURFACE TEMPERATURES


Last winter's actual temperatures versus our Winter Outlook issued in October.  Not bad.

The overall temperatures for this winter according to the best matches are not as cold across the eastern 2/3rds of the US when compared to the winters of 2014-15 (recap of last winter at this link) or 2013-14.

NOVEMBER BLEND

DECEMBER BLEND

JANUARY BLEND

FEBRUARY BLEND

Snowfall departures (above or below average) are much higher in January and February during moderate to strong El Ninos especially when these El Ninos are more central based (Modoki, using the EMI Index). 

Will this strong El Nino have some central based signature tendencies? Maybe more eastern El Nino characteristics thus a warmer winter overall with little snow? 

The Constructive Analog from Huug van den Dool's page on the CPC site shows the core of the ENSO warmth in the ENSO 3.4 region with a slow drift WEST toward the Dateline by late winter.

Will this verify?  Only time will tell. 

Many meteorologists in the private sector don't like using the Analog Method too literally because the probability of finding an exact match over such a large area is incredibly small. I get that. However, I have found that analogs give a glimpse into how the atmosphere responds to the various drivers in closely matched years. This group of "best fits" for this winter has stayed consistent over the last several months.  Several notes: Analogs are only one of several tools that I use in assessing seasonal outlooks. El Nino is only one of the many factors that will play a role in our winter weather. I did not include any ENSO dynamic model guidance in this post. That is a story for another time.

How Are Winter Outlooks Different Than Day-to-Day Forecasts?

Our WJW FOX8 Winter Outlook will air on Thursday, October 30th. Not long after, comments and criticism will start pouring stating in part that we should concentrate on getting the daily forecasts right rather than trying to forecast the weather months out. This isn't new. It happens after each seasonal outlook. Unfortunately, this common comparison is far from accurate and much more complicated than most of us realize.

Actual day to day weather forecasts are developed with analyzing current conditions, radar, satellite and other parameters to make a forecast for a short period of time in the future. 12 Hours, 24 hours, 36 hours, 48 hours. We utilize computer model projections as guidance. These projections are getting better as more data is utilized and assimilated into faster and faster computers. Computer projections are great but they still have their limitations. This is why its important for the meteorologist to determine which projection or combination of projections are accurate. Forecast accuracy continues to climb over the years. The lead time before tornado events has increased significantly saving many lives over the last decade.

Seasonal long range outlooks (winter weather forecasts) are created by looking at the ocean sea surface temperature patterns to include the north Pacific Ocean, the tropical Pacific Ocean (ENSO - two types of El Nino, neutral or La Nina), pressure patterns over the Arctic and North Atlantic, solar cycles, stratospheric wind behavior (QBO) and other longer term variables.  The elements just mentioned are matched up with other years of occurrence.  A best possible seasonal fit is created by "weighing" certain variables higher than others. Sometimes this works out well.  Sometimes it doesn’t.  I've learned a great deal about these variables and how they effect the overall pattern. I'll be the first to admit that I'm not the best. Other meteorologists with far more experience are excellent at determining, weighing and piecing together the years that best fit. Their livelihood depends on it.

To reiterate, the Analog Method is a trend outlook and not for a specific forecast for a specific day. Individual storms cannot be seen this far out. Forecasts for specific days? Forget it. There is too much randomness in the atmosphere that's way beyond our abilities to model accurately. However, by looking at parameters that existed in the past during other storm events, we can say that the chance of a big east coast snow storm is greater or less than in years past. 

For most people, all of these forecasts and trend outlooks are lumped into one group even though each are derived using entirely different information. I get it. Its human nature to generalize and simplify complicated subjects like the science of weather prediction. We formulate concrete, black and white, overly scaled down versions of the weather.  Whether its a long range winter outlook, a climate average for a wedding day, the thunderstorm chances for later this afternoon, lake effect snowfall amounts or a hurricane forecast track, its all the same weather animal to most. We subconsciously eliminate the nebulous science, weird looking equations and fancy internet computer animations in favor of a narrative that tells a better story.  In short, The Old Farmers’ Almanac fits with how our brains are wired.  Its simple. Its folksy with just enough science to make it credible.  Why do we continue believing the Old Farmers’ Almanac?  The simple answer is it makes us feel good!

While I love the Old Farmers Almanac for its articles, I'd trust an actual Meteorologist's forecast first.  I wouldn't dismiss the scientific explanation, ignore the random changes and replace them with the Old Farmers Almanac figuring that its more accurate.  The results are usually disappointing.

So remember that Winter Weather Outlooks are by their very nature formulated differently than day-to-day forecasts.  Let's sit back and see if this winter will behave like we think it will.

Does This "Cooler" Summer Mean A "Cooler" Fall?

In my last post, several drivers of this summer pattern were described in detail. Some of these carried over from the winter while the building somewhat weak El Nino (though not officially) has been a more recent factor. Some of these will no doubt carry over into the fall.

Rainfall since June 13th (last 60 days as of this update) is running 8-12" above average across a few counties. More than 90% of northern Ohio is running 1-2" above average rainfall.

Let recap: Where does this summer rank through August 2nd?  24th "coolest" in the last 60 years.  We're only talking a few tenth of a degree from year-to-year in some instances.

In those summer that were "cooler" than 2014, the September 1st through October 31st period in Cleveland, HALF WERE SLIGHTLY ABOVE NORMAL AND HALF WERE BELOW NORMAL.

After taking into account the summer drivers (slightly warm water across the tropical Pacific--not quite El Nino--and warm northeast Pacific Ocean), along with the weighted years from Dr. Joe B and Joe D, here is the analog for September and October combined.

This shows that northern Ohio in fall, in those years that closely match with 2014, is normal to slightly above normal.