Medium Range Discussion Spring 2012

[donsutherland1]

The 4/20 0z GFS ensembles maintained their impressive heat signal for parts of the West:

With the highly amplified pattern that will be in place at the time of the excessive heat, a secondary story will likely be the strong storm that will bring some much needed drought-easing rains to the East. Heavy wet snow in the higher elevations of the Appalachians, especially the Central and Southern Appalachians also appears likely.

[donsutherland1]

Some morning thoughts...

The coming outbreak of extreme heat extending from the Southwestern U.S. to Central and possibly Northern Plains States during the next 7-10 days is a high probability event. Some areas could approach or even exceed April high temperature marks. In the longer-range, there's a signal on the ensembles for another episode of above to much above normal readings in the Southwest.

What happens in May and June will likely have implications for the summer. Two analogs that have been coming up quite persistently in the partial analysis are 2002 and 2009. The recent rebound of the easterly Trade Winds has not yet been considered, but will have an impact down the road. If one recalls, 2002 was a hot summer nationwide. 2009 was one of the coolest summers since 1980.

Both years saw heat focused on the Southwest during May. May 2002 was actually quite cool nationwide except for the Southwest. May 2002 was the 6th coolest since 1980. May 2009 had pockets of cool anomalies outside the Southwest, but the Southeastern U.S. was also warm. May 2009 was the 10th warmest since 1980. The CFSv2 has been trending toward a cooler May nationwide. Some of the other emerging analogs for May also point to larger areas of cool anomalies. What happens upstream in the Arctic will be a key driver. Hence, a typically cooler pattern might wind up warmer than usual. The outcome could mask the pattern, so as far as the summer is concerned, one will need to tease out the pattern to determine whether May 2012 was a closer match to May 2002 or 2009. Afterward, June could provide critical information as to the summer as a whole. In short, a reasonably accurate June forecast could apply for the summer as a whole.

In June, one saw the Southwest remain warmer than normal in 2002. However, the heat had begun to spread eastward, covering a more expansive section of North America. All of the U.S. except for New England and the Gulf States were warmer than normal. The Central Plains had the warmest anomalies. As a result, June 2002 was the 3rd warmest since 1980. In 2009, the area of warm anomalies had shifted eastward and covered an area running from much of Texas eastward to North Carolina. With the exception of the Pacific Northwest, the remainder of the U.S. was cooler than normal. In contrast to 2002, the area of warm anomalies had not expanded very much. June 2009 was the 14th coolest since 1980.

Summer 2002 wound up being the 3rd warmest since 1980 and 5th warmest since 1895. Summer 2009 was the 6th coolest since 1980 and 42nd coolest since 1895.

[donsutherland1]

May 1-7, 2012 Thoughts:

With the AO likely to be negative leading up to May and the PNA likely to be positive into much or perhaps all of the first week in May, one would typically expect to see cold anomalies in a large part of eastern North America. However, 2012 has not been an ordinary year so to speak. Warmth has consistently outdueled the cold. Even typically cool patterns have underperformed.

If one looks northward, one finds an expansive area of cool anomalies currently located over Canada. However, farther north in the Arctic, warmth predominates. As a result, the cold will likely modify in advance of May and even areas that would typically experience cool weather could wind up on the mild side of normal.

The charts below are as follows:

Top Left: ENSO Region 3.4 anomalies of -0.7°C to 0.0°C, a PNA of -0.25 to +0.50, and an AO of +0.25 to +1.50 for North America (1950-2011).

Top Right: 4/21/2012 12z GFS Ensembles 11-15-day forecast (May 2-6, 2012)

Bottom Left: Northern Hemisphere temperature anomalies for April 18, 2012

Bottom Right: Observed decadal change in temperatures for May

The close of March into the first 12 days of April saw a dramatic burst of westerly winds in the Pacific. In response, there was a fairly dramatic warming of the ENSO regions, with the 2011-12 La Niña coming to an end. However, the past 9 days have been characterized by a rebound in the easterly trade winds. As a result, some cooling of the ENSO regions has taken place and is likely to continue. Right now, it is too soon to know for sure whether an El Niño or Neutral ENSO conditions will be the rule for the summer and beyond. Interestingly enough, among the analogs showing up for the first week in May is 2001.

More immediately, the issue concerns the weather for the first week in May. Taking into consideration the warmth in the Arctic, GFS ensembles, CFSv2 forecast, and the observed decadal temperature change, I expect:

- An area of cool anomalies across Central, Western, and perhaps northern Canada

- Near normal readings in the Southeast

-Near normal to somewhat above normal readings across much of the rest of the U.S.

-Above to perhaps much above normal readings in the Southwest (areas currently experiencing excessive heat)

[donsutherland1]

Some quick afternoon thoughts...

Today through Tuesday will be a tale of "two Americas." One will be cold with a spectacular late-season snowfall. The other will bake in excessive monthly record-challenging heat.

In the East, there will be a trough more than 4 standard deviations below normal and 850 mb temperatures more than 3 standard deviations below normal. The outcome will be a substantial late-season snowfall in the higher terrain in the Mid-Atlantic region into New York State, particularly parts of Pennsylvania and portions of western New York where more than a foot of snow could fall. DuBois, PA could pick up 7"-14" of snow.The 12z GFS is even indicating another smaller snowfall Friday night into at least part of Saturday for some of the areas in line for the late-season snowstorm. Elsewhere, much of the rain-starved Mid-Atlantic and New England region will pick up a widespread drought-denting rainfall of 1.5"-3.5" with locally higher amounts.

In the West, particularly the Southwest into the Central and Northern Plains, a monster ridge of 2.5 standard deviations above normal will be accompanied by 850 mb temperatures reaching up to 5 standard deviations above normal. Searing heat will be the result. Cities such as Las Vegas, Phoenix, Casper, and Pocatello, among others could challenge or top their all-time April high temperature records. The April marks are 99° for Las Vegas, 105° for Phoenix, 84° at Casper, and 86° at Pocatello.

[donsutherland1]

Brief evening update...

This afternoon, Phoenix and Las Vegas tied their highest April temperature on record. Phoenix topped out at 105° and Las Vegas reached 99°.

Meanwhile at 9 pm, light snow had move into DuBois, Pennsylvania. A significant snowfall still appears likely for that part of Pennsylvania.

[WxMidwest]

As far as the U.S. Part of the PNA formula, it's pretty well entrenched into the Positive territory with the temperature distribution...

[donsutherland1]

Today marked the 11th consecutive day on which the SOI was positive. This is an indication that the easterly trade winds had strengthened anew and that their rebound has been sustained. Such a development will likely slow or even reverse some of the warming that has taken place in the ENSO regions. Although it remains unlikely that La Niña conditions will return, developments going forward could determine whether the summer, fall, and winter seasons experience a weak El Niño or Neutral ENSO conditions.

One should watch trends in the SOI and responses in the ENSO regions. If ENSO Region 3.4 warms and develops El Niño or borderline El Niño conditions during the late spring or summer, then modest bursts of the easterly trade winds likely won't have a material impact leading to the winter. On the other hand, if ENSO Region 3.4 anomalies have difficulty staying warm, then modest bursts of the easterly trade winds could assure Neutral ENSO conditions, especially during the fall or winter.

The most recent year following a La Niña winter that saw the former conditions prevail was 2009. The most recent year following a La Niña winter that saw the latter conditions prevail was 2001. Interestingly enough, some analogs for 2001 have begun making their appearance, even as the 2009 analogs seem to be fading somewhat. For now, it's too soon to reach firm conclusions, as analogs and relative analog strengths are dynamic and changes regularly take place.

The Multivariate ENSO Index (MEI) could provide important insight going forward. The following are March-April through June-July MEI readings for 2001 and 2009.

2001:

MA: -0.149

AM: +0.151

MJ: -0.082

JJ: +0.223

2009:

MA: -0.161

AM: +0.355

MJ: +0.939

JJ: +0.929

Since 1970-71, all the ENSO winters were followed by June-July MEI readings of +0.35 or above with a change from March-April of -0.2 or above (meaning that the MEI did not cool by more than 0.2). There was only one false signal: 1980 (Winter 1980-81 was neutral).

In sum, a combination of the ENSO modeling and also developments in the MEI will probably offer good insight into the state of ENSO for the forthcoming winter. Before then, the response of ENSO Region conditions to the fluctuations in the SOI could provide some near-term insight.

[carley5]

the hurricane season of 01 was active though no hurricanes made landfall in the states. winter of 01/02 was a little above normal.

[donsutherland1]

the hurricane season of 01 was active though no hurricanes made landfall in the states. winter of 01/02 was a little above normal.

The 2001 analog has shown up in some recent data. However, it is not the leading analog at this time. In my opinion, I hope it won't be.

For the U.S. as a whole, Summer 2001 was the 11th warmest on record and Winter 2001-02 was the 6th warmest on record. In the Northeast, 2001 was the 37th warmest summer on record, while 2001-02 was the warmest winter on record.

[donsutherland1]

May 2012 Forecast:

For May, the step-down to less widespread warmth from April and March is likely to continue. The recent rebound in the easterly trade winds has stemmed any possibility of a fairly rapid onset of El Niño conditions. As a result, May will likely be closer to average than April.

Key Assumptions:

• The Neutral ENSO will prevail throughout May

• The AO will likely be predominantly positive

• The PNA will likely be predominantly positive (but generally < +1)

The top analog was 2002. The second closest analog was 1965. However, in recent weeks, 2001 has been emerging as an analog even as 2009 has faded somewhat. Nevertheless, given the continuing warmth in the Arctic region, areas with cool anomalies will likely be smaller than what is shown from the analogs

Below are charts showing the composite temperature anomalies for May based on the analogs and also the observed decadal temperature change for May.

Based on the above charts, prevailing conditions in the Arctic region, and CFSv2 (a warm outlook), my thoughts are as follows:

- The eastern and western third of the U.S. will likely be warmer than normal. The Southwest could be much warmer than normal.

- The Southern and Central Plains will likely see near normal to warmer than normal conditions.

- An area running from the Northern Plains into Central Canada could see below normal readings.

Finally, precipitation along the East Coast should be close to normal for the month. There is a possibility that New England into eastern Canada could have somewhat above normal precipitation.

[uncle W]

May 2012 Forecast:

For May, the step-down to less widespread warmth from April and March is likely to continue. The recent rebound in the easterly trade winds has stemmed any possibility of a fairly rapid onset of El Niño conditions. As a result, May will likely be closer to average than April.

Key Assumptions:

• The Neutral ENSO will prevail throughout May

• The AO will likely be predominantly positive

• The PNA will likely be predominantly positive (but generally < +1)

The top analog was 2002. The second closest analog was 1965. However, in recent weeks, 2001 has been emerging as an analog even as 2009 has faded somewhat. Nevertheless, given the continuing warmth in the Arctic region, areas with cool anomalies will likely be smaller than what is shown from the analogs

Below are charts showing the composite temperature anomalies for May based on the analogs and also the observed decadal temperature change for May.

Based on the above charts, prevailing conditions in the Arctic region, and CFSv2 (a warm outlook), my thoughts are as follows:

- The eastern and western third of the U.S. will likely be warmer than normal. The Southwest could be much warmer than normal.

- The Southern and Central Plains will likely see near normal to warmer than normal conditions.

- An area running from the Northern Plains into Central Canada could see below normal readings.

Finally, precipitation along the East Coast should be close to normal for the month. There is a possibility that New England into eastern Canada could have somewhat above normal precipitation.

May 1965 was quite mild in NYC...It was 94 degrees late that month...Early May 2001 had a heat wave with three straight days 90 or higher...May 2002 had near record cold during the second half...It had a heat wave in April...I think the drought will end before the end of the year in the NE but temperatures will still be above normal...but not as extreme at least for the warm season...

[donsutherland1]

May 1965 was quite mild in NYC...It was 94 degrees late that month...Early May 2001 had a heat wave with three straight days 90 or higher...May 2002 had near record cold during the second half...It had a heat wave in April...I think the drought will end before the end of the year in the NE but temperatures will still be above normal...but not as extreme at least for the warm season...

The 1965 El Niño came on faster than I expect this time around and it was stronger than I think will be the case for the upcoming event. It would not surprise me if 1965 falls out of the possible analog pool by June or July. Overall, summer 1965 was generally cool nationwide.

[KEITH L.I]

The 1965 El Niño came on faster than I expect this time around and it was stronger than I think will be the case for the upcoming event. It would not surprise me if 1965 falls out of the possible analog pool by June or July. Overall, summer 1965 was generally cool nationwide.

1976 looks like a great analog..heat in April after a balmy February...moving from Nina to a weak el nino

[GaWx]

1976 looks like a great analog..heat in April after a balmy February...moving from Nina to a weak el nino

1976-7 (DJF) also had a +PDO (often associated with a +PNA dominant pattern), -NAO, and -AO, the ultimate combo when teamed up with a weak El Nino for a very cold winter for the E US. I'd love to see this combo!

[donsutherland1]

Building on the point in Message #157 compared to a possible El Niño signal from the MEI:

A refined statistical rule for a winter El Niño event (DJF Region 3.4 anomaly > +0.5°C) would be as follows:

1. June-July MEI of +0.350 or above

and

2. Change in the June-July MEI from the March-April figure of -0.2 or higher (not a larger fall) or PDO > 0

The change in MEI reflects the seasonal evolution of the MEI. A PDO+ often accompanies El Niño events.

The outcome for the above statistical rule was as follows:

19 El Niño signals

18 El Niño events

1 Neutral ENSO event

2 El Niño events occurred without the signal (1952-53 and 1968-69)

Unfortunately, because the MEI does not go back prior to 1950, the data that was examined begins in 1950.

In short, the MEI can provide some insight. It can provide another tool, along with the guidance, to make an informed guess about whether there will be an El Niño this winter.

[donsutherland1]

1976 looks like a great analog..heat in April after a balmy February...moving from Nina to a weak el nino

1976 has done well so far. It had been among my strongest March and April analogs. It did fade somewhat for May, but it's still on the table so to speak.

[icebreaker5221]

May 2012 Forecast:

For May, the step-down to less widespread warmth from April and March is likely to continue. The recent rebound in the easterly trade winds has stemmed any possibility of a fairly rapid onset of El Niño conditions. As a result, May will likely be closer to average than April.

Key Assumptions:

• The Neutral ENSO will prevail throughout May

• The AO will likely be predominantly positive

• The PNA will likely be predominantly positive (but generally < +1)

The top analog was 2002. The second closest analog was 1965. However, in recent weeks, 2001 has been emerging as an analog even as 2009 has faded somewhat. Nevertheless, given the continuing warmth in the Arctic region, areas with cool anomalies will likely be smaller than what is shown from the analogs

Below are charts showing the composite temperature anomalies for May based on the analogs and also the observed decadal temperature change for May.

Based on the above charts, prevailing conditions in the Arctic region, and CFSv2 (a warm outlook), my thoughts are as follows:

- The eastern and western third of the U.S. will likely be warmer than normal. The Southwest could be much warmer than normal.

- The Southern and Central Plains will likely see near normal to warmer than normal conditions.

- An area running from the Northern Plains into Central Canada could see below normal readings.

Finally, precipitation along the East Coast should be close to normal for the month. There is a possibility that New England into eastern Canada could have somewhat above normal precipitation.

Don, any reason you're not including 2006 in your analogues? Seems like a good ENSO match with weakly positive AO. 2006 also saw a major spring heat wave, but it occurred in Apr instead of Mar.

One difference I see is that May 06 featured a -PNA and a -NAO, which are both expected to start the month positive this year, but that can certainly change by mid month.

1976 looks like a great analog..heat in April after a balmy February...moving from Nina to a weak el nino

Seems like a very good match ENSO and teleconnection wise. Although, Canada was extremely cold Dec-Mar 1976 and was quite warm this year. So I would argue that an extremely cold May like the one that occurred in 76 is quite a bit less likely.

[uncle W]

1976 looks like a great analog..heat in April after a balmy February...moving from Nina to a weak el nino

enso analogs similar to 2012...

year...NDJ...DJF...JFM...FMA

2012...-0.9...-0.8...-0.6

2006...-0.7...-0.7...-0.6...-0.4...el nino developed...

2009...-0.6...-0.8...-0.7...-0.5...el nino developed...

2001...-0.7...-0.6...-0.5...-0.4...neutral...

1996...-0.7...-0.7...-0.7...-0.5...neutral...

1972...-0.9...-0.7...-0.4....0.0...el nino developed...

I'm not sure what the 2012 FMA index will be...These are just similar enso analogs...1976 had a strong La Nina that changed to a weak el nino...This year was a weak la nina for the most part...I'm rooting for 2009...

[donsutherland1]

Don, any reason you're not including 2006 in your analogues? Seems like a good ENSO match with weakly positive AO. 2006 also saw a major spring heat wave, but it occurred in Apr instead of Mar.

It's actually in the pool. I just specified a few analogs that I had mentioned in previous discussions, with 1965 being the exception only because it came up very high for May. I'll post my summer thoughts sometime in late May. It will be interesting to see what analogs come up at that time. Hopefully, there will be a clearer ENSO picture by that time, as well.

Seems like a very good match ENSO and teleconnection wise. Although, Canada was extremely cold Dec-Mar 1976 and was quite warm this year. So I would argue that an extremely cold May like the one that occurred in 76 is quite a bit less likely.

I agree with you. I suspect that the contribution from 1976 helped skew my May composite map. I don't think the cold will be as widespread as the analogs would suggest. The CFSv2 is even warmer than my thoughts and it is now in a timeframe where it has shown some level of skill.

[donsutherland1]

Over the past few days, some colder anomalies have developed in the Arctic. At the same time, the pattern has evolved in a fashion that is more compatible with areas of cool anomalies in parts of eastern North America as suggested in the April 23-30 thoughts (#127) and perhaps in line with the raw teleconnection analogs for the May 1-7 timeframe (#153). The argument is that there won't be sufficient cold air to sustain cool anomalies for the entire May 1-7 timeframe in much of the U.S., hence most areas would see near normal to somewhat above normal readings during the first week in May. That outcome would not preclude a fairly strong, but for most areas, short-lived, cold shot for the end of April, beginning of May.

The 4/24/2012 12z run of the GFS contains just such a cold shot. In fact, the cold shot shown for the April 29-May 1 timeframe for the Great Lakes, Ohio Valley, Mid-Atlantic, New England, southern Ontario, and southern Quebec regions would be quite notable if the model is taken literally.

For perspective, the GFS shows a low temperature of 37° at May 1 9z in New York City. Readings below 40° in NYC have become rare. The last such reading was May 1, 1978 when the temperature fell to 38°. The 1960s saw 5 days on which the temperature fell below 40° in May. The 1970s also saw 5 such days. The coldest day since then has been 40° on May 9, 1992. More than likely, subsequent runs of the model will be milder and the actual outcome will also be milder (perhaps with a low temperature in the low 40s?). Nevertheless, areas outside the big cities, will likely experience a late-season frost.

[uncle W]

For perspective, the GFS shows a low temperature of 37° at May 1 9z in New York City. Readings below 40° in NYC have become rare. The last such reading was May 1, 1978 when the temperature fell to 38°. The 1960s saw 5 days on which the temperature fell below 40° in May. The 1970s also saw 5 such days. The coldest day since then has been 40° on May 9, 1992. More than likely, subsequent runs of the model will be milder and the actual outcome will also be milder (perhaps with a low temperature in the low 40s?). Nevertheless, areas outside the big cities, will likely experience a late-season frost.

1963 had three mornings with minimums in the upper 30's...The 39 on 5/24/63 is the latest 30's on record in NYC...Washington DC had flurries on 5/1/63...1976 has the latest 38 degree temperature on record on 5/19...1966 and 1967 had snow in the north east and lakes...1961 had a late season trace of snow the last week of the month...1976, 1977 and 1978 had temps in the upper 30's...Like you said it's been 34 years since NYC had 30's in May...

[am19psu]

New blog from a friend of the board. All about the global circulation. Must read, imo, if you want to understand what happened this winter and what will happen going forward.

http://ionlyusethegfs.blogspot.com/

[CoastalWx]

Great article. Well thought out write up and really explains a lot.

[North Balti Zen]

New blog from a friend of the board. All about the global circulation. Must read, imo, if you want to understand what happened this winter and what will happen going forward.

http://ionlyusethegfs.blogspot.com/

Thanks for the heads up on that. Great read.

[Srain]

New blog from a friend of the board. All about the global circulation. Must read, imo, if you want to understand what happened this winter and what will happen going forward.

http://ionlyusethegfs.blogspot.com/

Excellent read and certainly explains what went wrong with a lot of the early forecasts late last summer/early fall. I am just glad that those early forecasts busted so badly down here...

[donsutherland1]

New blog from a friend of the board. All about the global circulation. Must read, imo, if you want to understand what happened this winter and what will happen going forward.

http://ionlyusethegfs.blogspot.com/

Many thanks for sharing this. It's a great read.

[JamieOber]

Over the past few days, some colder anomalies have developed in the Arctic. At the same time, the pattern has evolved in a fashion that is more compatible with areas of cool anomalies in parts of eastern North America as suggested in the April 23-30 thoughts (#127) and perhaps in line with the raw teleconnection analogs for the May 1-7 timeframe (#153). The argument is that there won't be sufficient cold air to sustain cool anomalies for the entire May 1-7 timeframe in much of the U.S., hence most areas would see near normal to somewhat above normal readings during the first week in May. That outcome would not preclude a fairly strong, but for most areas, short-lived, cold shot for the end of April, beginning of May.

The 4/24/2012 12z run of the GFS contains just such a cold shot. In fact, the cold shot shown for the April 29-May 1 timeframe for the Great Lakes, Ohio Valley, Mid-Atlantic, New England, southern Ontario, and southern Quebec regions would be quite notable if the model is taken literally.

For perspective, the GFS shows a low temperature of 37° at May 1 9z in New York City. Readings below 40° in NYC have become rare. The last such reading was May 1, 1978 when the temperature fell to 38°. The 1960s saw 5 days on which the temperature fell below 40° in May. The 1970s also saw 5 such days. The coldest day since then has been 40° on May 9, 1992. More than likely, subsequent runs of the model will be milder and the actual outcome will also be milder (perhaps with a low temperature in the low 40s?). Nevertheless, areas outside the big cities, will likely experience a late-season frost.

Don,

One thing of note, and I don't know if anyone else recorded this at their nearest station, but UNV has had three consecutive days of double-digit negative temp departure over the last three days. First time in six months that has happened.

That seems remarkable to me.

Last time? Interestingly enough, last week of October, during another out-of-season snow event for this area.

[wxmx]

Building on the point in Message #157 compared to a possible El Niño signal from the MEI:

A refined statistical rule for a winter El Niño event (DJF Region 3.4 anomaly > +0.5°C) would be as follows:

1. June-July MEI of +0.350 or above

and

2. Change in the June-July MEI from the March-April figure of -0.2 or higher (not a larger fall) or PDO > 0

The change in MEI reflects the seasonal evolution of the MEI. A PDO+ often accompanies El Niño events.

The outcome for the above statistical rule was as follows:

19 El Niño signals

18 El Niño events

1 Neutral ENSO event

2 El Niño events occurred without the signal (1952-53 and 1968-69)

Unfortunately, because the MEI does not go back prior to 1950, the data that was examined begins in 1950.

In short, the MEI can provide some insight. It can provide another tool, along with the guidance, to make an informed guess about whether there will be an El Niño this winter.

It does now. At least the extended MEI (MEI.ext)

http://www.esrl.noaa.gov/psd/enso/mei.ext/

[donsutherland1]

Many thanks for the link, Jorge. I'll re-run the numbers for the earlier data.

[donsutherland1]

Taking into consideration the link provided by Jorge, I've run the numbers for earlier El Niño events.

The statistical rule for a winter El Niño event (DJF Region 3.4 anomaly > +0.5°C) remained as follows:

1. June-July MEI of +0.350 or above

and

2. Change in the June-July MEI from the March-April figure of -0.2 or higher (not a larger fall) or July PDO > 0

The outcomes for the above statistical rule were as follows:

1871-1949:

19 El Niño signals

13 El Niño events

6 Neutral ENSO events

5 El Niño events occurred without the signal (1904-05, 1911-12, 1913-14, 1923-24, and 1925-26)

1950-Present:

19 El Niño signals

18 El Niño events

1 Neutral ENSO event

2 El Niño events occurred without the signal (1952-53 and 1968-69)

Five notes are in order:

1. The reconstructed MEI is a simplified index.

2. The following disclaimer is provided: "Caution should be exercised when interpreting the MEI.ext on a month-to-month basis, since it was developed mainly for research purposes, similar to the original MEI."

3. The 1871-2005 MEI Index is not identical to the 1950-Present MEI index (Note 1). I ran the numbers for the overlapping 1950-2005 timeframe. The mean difference in monthly values was 0.218 and the standard deviation of the monthly differences was 0.183.

4. The PDO Index goes back to 1900

5. The latest ENSO Region 3.4 data is somewhat different from the reconstructed data that goes back to 1871. The mean monthly difference is 0.102 and the standard deviation of the monthly differences is 0.099 for the overlapping months.

Conclusion:

Given the differences between the reconstructed MEI Index and the current MEI Index, one could not reasonably expect the above-stated rule to work in a near identical fashion for both MEI Indices. However, the rule still provided a reasonably strong, but not decisive signal for the 1871-1949 period despite the differences between the two MEI Indices. During that timeframe, El Niño signals from the above rule were followed by winter El Niño events just over two-thirds of the time.