For reference, here’s a link to my outlook for 2013-14 (links to past outlooks can be found through there)
https://ohwxramblings.wordpress.com/2013/08/21/winter-2013-2014-an-early-look/
Winter 2014-15 outlook maps:
Expected temperature departures from normal for December, January, and February (overall, meaning an average of the whole winter):
Overall winter precipitation:
Potentially common storm tracks (note the variability near the East Coast):
Discussion:
Pacific/Indian Ocean Sea Surface Temperatures
El Nino?
A big factor heading into every winter is what the state of the El Nino South Oscillation (ENSO) will be. The state of the ENSO (El Nino, La Nina, neutral) is often one of the driving forces behind the location of tropical convection over the Pacific during the winter, which in turn helps modulate the position of troughs/ridges in the jet stream over the Pacific and the downstream weather pattern over North America. After a significant pool of warm sub-surface waters developed during the Spring in the ENSO regions and caused some talk of a potentially strong El Nino for this upcoming winter, we have struggled to see significant warm anomalies surface across the ENSO regions:
Don’t get me wrong, things have warmed, with every ENSO region seeing above average sea surface temperatures. However, the anomalies in the more important ENSO regions 3, 3.4 and 4 have struggled to consistently stay warmer than +0.5C, which is the threshold for an official El Nino. Below the surface, waters are still generally warm, which probably bodes well for the potential of an official weak El Nino occurring during the winter:
Trade winds over the Pacific have generally had trouble relaxing this summer, potentially due to warm waters west of the Dateline towards Australia and the eastern Indian Ocean keeping pressures somewhat lower in that area. The areas of blue on this time-longitude plot of the 850mb zonal (west to east) winds in the equatorial Pacific represent stronger than normal trade winds. Remember, trade winds typically blow east to west in the equatorial Pacific, and zonal winds are west to east winds, so negative zonal wind anomalies here actually correspond to stronger than normal trades. Stronger than normal trade winds cause cooler waters to upwell from below and cool the surface water temperatures. Note how the strong westerly wind anomalies, which typically cause warmer in the equatorial Pacific when they propagate eastward, have had trouble getting east of the Dateline this summer:
This may explain why the potential El Nino has only slowly intensified this summer and at times barely maintained itself or even cooled. With that said, the Southern Oscillation Index (SOI), which measures the pressure difference between western South America and eastern Australia, and is often an indication of whether or not the trade winds may be stronger or weaker than normal near the equator in the Pacific, has on average been negative for July, August, and September and remains negative as we head into October. This is an indication of possible weaker trade winds. You can always get the latest SOI data here:
http://www.longpaddock.qld.gov.au/seasonalclimateoutlook/southernoscillationindex/
The 850mb zonal wind plot above shows that over the last couple of weeks, the trade winds have been about as strong as normal east of the Dateline, but right now have the largest area of weaker than normal trades west of the Dateline since late July/early August. This may be due to the eastern Nino regions warming and reducing the temperature difference between the eastern Pacific and western Pacific recently, causing trades to weaken. Regardless, if the –SOI continues and these weaker trades can last for more than a week or two, we’d likely see sea surface temperatures in the Nino regions warm even more, and become solidly above the magic +0.5C threshold in the Nino 3.4 region. This would increase the chances of a weak El Nino come winter. Most ENSO models agree with this idea (although, it should be noted that their performance this past Spring was weak, as they tried showing a much stronger El Nino developing):
Given that the water temperatures are already bordering on El Nino type warm anomalies, and given that the recent pattern across the equatorial Pacific appears to favor these warmer waters at least maintaining if not warming a bit more in the coming weeks, I’m inclined to believe an official weak El Nino will occur this winter (5 or more consecutive tri-monthly periods where the average ENSO region 3.4 sea surface temperature anomalies are greater than +0.5C). There’s some chance this doesn’t occur if we get the trade winds to abruptly intensify in October or November, but I believe this is unlikely, and even if it happened the ENSO would still be warm-neutral. Due to the lack of westerly wind bursts making headway east of the Dateline this summer, and the struggle for the ENSO regions to consistently warm, a moderate El Nino is an outside possibility, but in my opinion is not extremely likely. The “El Nino” doesn’t appear to be either east based or west based at this time.
Indian Ocean:
There are a few reasons why the water temperatures in the Indian Ocean can be construed as potentially important when determining what kind of weather pattern we’ll see during the winter, and I’ll briefly look into those below. For starters, the water temperatures in the Indian Ocean will affect air pressure in that area, potentially impacting trade winds farther east over the Pacific Ocean and the ENSO state. In addition, water temperatures in this area may work with the state of the ENSO to help determine where tropical convection is most persistent, helping to modulate the position of the jet stream over the Pacific and North America. Luckily, there is an index that measures where warmer water is located in the Indian Ocean, called the Indian Ocean Dipole (IOD from here on out). Essentially, the IOD is “positive” or “warm” if waters over the western Indian Ocean are warmer than normal and waters farther east, towards Indonesia, are cooler than normal. It is “negative” or “cold” if the opposite is true. This index, after being mainly positive to start the decade, has trended mainly negative for the last 15 months give or take, but has risen to start fall:
This graph, as well as more information about the IOD, can be found here:
http://www.bom.gov.au/climate/IOD/about_IOD.shtml
A crude but functional list of monthly IOD values from 1958 through most of 2010 can be found here. There two locations are the source of the IOD numbers I will be using below:
http://www.jamstec.go.jp/frcgc/research/d1/iod/DATA/dmi_HadISST.txt
In general, it seems as though +IOD winters tend to occur more frequently during El Ninos. I ran two composite maps based on the IOD data above. The first one is for all winters where there was a +IOD (I went with 0.3+, our most recent weekly value). There is an El Nino in the mean winter SSTs, however it should be noted that some very strong El Nino years are in that set of data:
When I remove the strongly +IOD years (generally, leaving winters where the IOD was between +0.3 to +0.8), which leaves a set of IOD data that may be more applicable to this year, the El Nino signal is still present, albeit a little bit weaker:
This leads me to believe that the recent emergence of a warm/+IOD may bode well to El Nino chances this winter, if it can persist. It should be noted that I understand that in the above plots that the Atlantic and even the N. Pacific sea surface temperature anomalies don’t match this year…that’s ok, I’m using the +IOD data to try to determine the potential state of the ENSO going forward at this point in time. It appears that this correlation is obviously stronger if the +IOD is stronger, so we’ll have to monitor the IOD going forward. With very quiet convection over the western Indian Ocean and more active weather (including some strong tropical cyclones) in the eastern part of the basin, the +IOD may intensify some over the next couple of weeks.
Taking it a step further now, I’m going to run the composite precipitation anomalies for the winters where there was a weak +IOD (the years used in the second IOD related map above) and a weak El Nino, to see where convection tends to persist in the Tropical Pacific in winters with those SST configurations. Unfortunately, there were only two years that matched this set, so it is an extremely small sample size, but here’s what the composite map looks like:
Not surprisingly, the most persistent convection during winter during these years was east of the Dateline. This is in contrast to the last few winters, where the convection was west of the Dateline. Given this, we probably won’t see a persistent ridge into Alaska this winter (ridging would be farther east) assuming convection occurs where a warm IOD and weak ENSO suggest it should occur in the tropics.
Thoughts on if the idea of convection occurring farther east this winter are legitimate
The recent development of the +IOD going positive supports the warm waters in the ENSO regions becoming an official El Nino over the next few months, as +IOD events typically correlate well with El Ninos, likely due to the cooler waters near New Zealand supporting weaker trade winds over the tropical Pacific. Warm sub-surface waters and signs that the trades have relaxed of late, in addition to the +IOD suggesting this trend will likely continue, further help support this notion. In addition, it can be seen on 200mb vertical velocity potential maps this summer that convection has successfully gotten east of the Dateline several times over the last few months:
The green indicates upward vertical motion due to convection in the Tropics. Note how the bulk of the active convection over the last several months has been between about 160E and 120W, right where the bulk of the convection occurred in the two IOD+ENSO analog winters above. This suggests that the pattern may already be turning more Nino like (supported by recent Gulf of Alaska low pressure systems as well). Considering that convection over the Tropical Pacific is a large driver of where the ridges and troughs in the jet stream set up over North America, and considering that this variable “should” change given the expected sea surface temperature anomaly configuration this winter, and appears to have changed over the last several months to match that, shows that a pattern that is a “repeat” of the very cold winter of 2013-14 is extremely unlikely. That’s not to say this winter will be dry and warm across the whole country, however if we get cold, it will be due to a different set of circumstances.
Pacific Decadal Oscillation (PDO)/Warm Northern Pacific Waters
The PDO is a multi-decadal variable that describes the pattern of sea surface temperature anomalies in the Pacific. This is different than the ENSO, because it takes into account the temperatures in the Northern Pacific as well. As the graph below shows, the PDO normally oscillates in 2-3 decade windows. We have been in a cold/negative phase since around 1999:
However, if you recall the global SST map posted at the beginning of the discussion, you’ll note that the current state of the Northern Pacific doesn’t exactly look like a cold PDO, with cooler waters in the far NW Pacific compared to normal, and much warmer than normal waters almost everywhere else. In fact, the monthly PDO values have all been positive/warm since January of 2014. The positive values have decreased a bit over the last couple of months.
The above graphics, as well as a data table of monthly PDO values going back to 1900, can be found at this link:
http://jisao.washington.edu/pdo/
The question at this point is; does all the warm water in the northern Pacific make a huge difference on the expected weather pattern this winter? This is a difficult one to answer…it is likely that the water temperatures in the northern Pacific are more of a result of the prevailing weather pattern, as opposed to the other way around, because waters that are normally in the mid 40’s aren’t going to change the pattern when they warm up or cool down a few degrees…unlike the waters in the tropics that drive thunderstorms, there just isn’t enough heat released from the waters farther north to make a huge impression on the weather pattern.
When looking at the large scale weather pattern over the northern Pacific over the last year, it is easy to understand how the northern Pacific got so warm and even flipped the PDO positive. This pattern of strong ridging near the Dateline began to take shape last October, which is before the PDO turned positive, which is a sign that the warm waters in this region are a result of the ridging here, and not vise-versa. In addition, the ridging near the Dateline was most likely a result of a modestly cool ENSO and warmer waters in the Pacific west of the Dateline causing tropical convection to be focused west of the Dateline. Given the change in the Pacific pattern discussed above over recent months, a repeat performance of strong ridging near the Dateline is extremely unlikely this winter, meaning the current SST configuration will probably change over the coming months. In fact, the GFS and ECM ensembles currently agree on a Gulf of Alaska low persisting through much of the rest of the month of October, indicating that the water temperatures over the Gulf of Alaska will cool (which will likely result in a PDO value closer to neutral than anything else, but still probably slightly positive).
Because the PDO is more of a reflection of a persistent longwave pattern over the Pacific, I thought it would be interesting to look at periods where the PDO also briefly spiked positive during an overall negative PDO regime in the past, as that may indicate a similar change in the overall pattern over the Pacific for a relatively brief period of time, like we have seen over the last year. I tried to pick years where the PDO rose to positive values during the spring or summer months. I then narrowed down the years to include only warm neutral or weak El Nino ENSO states during the winter. The 500mb anomalies during the winters that met these criteria looked like this over the Pacific:
This goes along with the idea of tropical forcing occurring farther east this winter as discussed above, suggesting that the large negative anomaly will be located closer to or even east of the Dateline, with any persistent ridging being located closer to western North America. Because I believe a weak El Nino is likely to develop as I discussed above, and similar +PDO spikes during warm neutral or El Nino ENSO events have resulted in a pattern that makes sense with the tropical forcing being focused farther east than the last few winters, which has already began to occur this past summer and heading into this fall, I believe the idea of a low somewhere over the Gulf of Alaska/Aleutians, and ridging over western N. America is reasonable this winter. The question becomes, can we still get high latitude blocking to direct significant cold towards the US, or will we see a much warmer winter in general? I’ll get to that below.
Potential for High Latitude Blocking
The high likelihood of last winter’s Dateline ridging in the N. Pacific being replaced by troughing suggests that the Pacific pattern won’t continuously load significant amounts of cold air into much of central and western Canada this winter. This means that we will need something else to get cold into the US…and that something is possible high latitude blocking. It isn’t particularly easy to guess how certain teleconnections will play out during the winter months or where blocking will occur, but there are several clues that are present by the time we hit October in most given years that can potentially help us determine if significant periods of high latitude blocking are likely in the upcoming winter, and potentially help determine where it will occur.
Quasi-Biennial Oscillation (QBO):
The QBO is a change in direction in the winds in the stratosphere above the tropics. A “negative” QBO represents easterly wind anomalies at this level, and a “positive” QBO represents westerly wind anomalies at this level. The QBO is important to winter because an easterly QBO typically correlates to a weaker stratospheric polar vortex (PV) during the winter across the Northern Hemisphere. A weaker stratospheric PV is more prone to breaking down and allowing sudden stratospheric warming events (SSW) to occur during the winter months. SSWs can down-well into the troposphere and result in a period of strong high-latitude blocking during the winter, which is why the QBO phase and intensity is of interest when trying to forecast whether blocking may be more or less likely to occur during an upcoming winter season.
The QBO has currently turned negative over the last few months…in fact, the September value of the QBO, as measured at the 30mb level, was -23, which is strongly negative. This would “appear” to bode well for the potential of high latitude blocking during the upcoming winter, however I have read many times over the year that QBO values of lower than -15 or -20 were “too low.” I tried to substantiate that, first by looking at years where at least one month had a QBO value of -20 or lower (there were only four of those…58-59, 65-66, 74-75, and 05-06)…although 05-06 was a bad winter, the overall 500mb composite for those winters had a –AO and modest –NAO…it is however an extremely small sample size. To relax the criteria a bit, I went ahead and added in where the average DJF QBO number was lower than -15, which added a few (62-63, 00-01, 09-10 and 11-12). Again, 11-12 was a clunker (although a SSW event did occur that winter and cause colder weather so spill into Eurasia during the second half of the winter as a –AO developed), but the overall composite mean from all 9 winters above shows a –AO and a –NAO:
Data used for the above analysis and further QBO analysis can be found here:
http://www.esrl.noaa.gov/psd/data/correlation/qbo.data
Regardless, looking ahead towards this winter, I’m going to try to “guess” what the QBO values will be in the neighborhood of. To do this, I’m going to look at how long strong –QBO spells tend to last, and try to estimate when our current strong –QBO will start to rise. In the data linked above, I found 21 cases where the QBO attained a value of -20 or lower for at least 1 monthly value. Since our current spell of -20 or lower monthly values is at two months, I threw out 4 cases where the -20 value only occurred for 1 month (leaving 17 cases). In those average cases, the average length of a -20 or lower QBO value was 4.17 months (let’s call that 4 months). (It should be noted that a 1965-66 period of -20 or lower QBO, one month saw the values rise to -19. However, surrounding months were -20 or lower, so I counted that as one 5 month block of -20 or lower).
Our current QBO values of -20 or lower began in August. If this current period of -20 or lower QBO lasts the average 4 months, the values would begin to increase to above -20 by December. Out of the 17 cases where the QBO was lower than -20 for 2+ months, 6 of those (35%) lasted 5 or more months (longest was 7 months in 2005), which would take us into this upcoming winter. Either way, odds are we’ll still have a QBO of lower than -5 for each month this upcoming winter (except for maybe February if it starts rising quickly), but odds also appear to point to the values coming above -20 either before or during the winter. This sounds good, however, I thought it would be an interesting exercise to run years that featured a -20 QBO or lower during the fall that began to rise either just before or during winter. The results weren’t very pretty in terms of high latitude blocking:
Unfortunately, there was a mix of ENSOs in there (several La Ninas, which explains the trough over western N. America…in fact, only one had a somewhat decent ENSO match (79-80)), and below, the October composite pattern in those years doesn’t show a good –AO like we are seeing this October, so the strict QBO analog set shouldn’t be taken extremely seriously:
Overall, the –QBO suggests that high latitude blocking is in general more likely than years in which the QBO is positive. Based on the above “strictly QBO analogs,” it appears better if the average winter QBO is -15 or lower, or features one month of -20 or lower, as that composite featured more high latitude blocking. However, in general, these analogs look strictly at the QBO and ignore other important factors, so they aren’t going to define how I draw my outlook to a tee.
October Correlations to Winter
A strong –AO is occurring this month, and I’ll look into that along with a couple of other fun things that happen in October (like advancing snow cover in Eurasia) that appear to have some correlation to high latitude blocking in winter.
Strong –AO in October:
This October appears on track to come in with a monthly value of less than -1. It is possible that the value comes in close to or lower than -1.5 for the month, but the latter is nowhere near a certainty yet. Regardless, there have been 10 Octobers since 1950 that had a monthly –AO of less than -1. Those Octobers were 1960, 1966, 1968, 1974, 1979, 1981, 2002, 2006, 2009, 2012. Only two of the following winters (1974-75 and 2006-07) saw a +AO on average. It should be noted than February of 2007 did manage of monthly AO of -1.31 in an otherwise mainly positive AO winter. This seems to be a fairly strong correlation to –AO blocking winters following strongly –AO Octobers.
Monthly AO data used:
http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/monthly.ao.index.b50.current.ascii.table
For kicks, only 3 of the 10 winters following strong –AO Octobers featured an average –NAO (1968-69, 2002-03, 2009-10), although many of the other winters were only slightly positive. So, there doesn’t appear to be a strong October AOàwinter NAO correlation.
Monthly NAO data used:
http://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/norm.nao.monthly.b5001.current.ascii.table
Unfortunately, no other classic teleconnections will be strongly negative or positive this month as it looks now, so I did not try to correlate them to this upcoming winter.
Snow Advance Index (SAI)
http://www.aer.com/news-events/blog/snow-advance-index-new-tool-predicting-winter%E2%80%99s-severity
Judah Cohen (PhD) has come up with an index, called the SAI, which has a greater than .8 correlation in determining the upcoming winter AO/NAO indices. In short, when the SAI is positive (faster snow advance than normal in October in Eurasia south of 60N), it tends to correlate to a –AO and –NAO in the upcoming winter. When the SAI is negative (slower snow advance than normal in October in Eurasia south of 60N), the opposite tends to be true. There has been a decent increase in snowcover in that magical area in Eurasia thus far, although note that there was some snow south of 60N when we started the month:
Compared to October 11:
These appear to be solid gains south of 60N. The modeled weather pattern over the next week to two weeks suggests this advance should continue, with recent GFS and Euro runs showing below normal heights (which generally are good for snow cover buildup) continuing over Eurasia for the foreseeable future:
This isn’t necessarily the coldest looking pattern ever for Eurasia in October, but they should continue to gradually build snow south of 60N. Early indications are the SAI may come in with more snow advance than normal, which would suggest a –AO/-NAO winter, however the caveat is it’s October 12th so there is some chance this changes.
October Pattern Index (OPI)
A group of Italian researchers has found that a certain October 500mb pattern over the N. Hemisphere is conducive to a –AO in the upcoming winter (or positive if the inverse of the pattern is true). The researchers have not published a paper yet, however they do claim a correlation of over .9. The current OPI index is strongly negative (through the first 11 days of the month plus 10 days of GFS prediction). This initially appears to support a –AO winter, however as with the SAI discussion above, the caveat is that we’re a bit early in the month and this may change.
OPI monitoring:
http://app.til.it/opi/
OPI information:
http://app.til.it/opi/Image/ENF.pdf
Conclusions:
Although an El Nino has struggled to develop thus far this summer, warm sub-surface waters in the Equatorial Pacific, recent slackening of the trade winds across the Tropical Pacific, a long running –SOI, and recent flip to a +IOD suggests that an El Nino will still probably develop by this winter. Given how much trouble we’ve had seeing waters warm though over the last several months in the ENSO regions, a weak El Nino is probably the most likely outcome, with an outside chance at a moderate Nino and a similarly small chance at the ENSO remaining warm-neutral. The brief +PDO spike, combined with this El Nino, suggests tropical forcing will be focused mainly east of the Dateline this winter, suggesting a general low anomaly in heights south of the Aleutians and general high anomalies across western N. America. This also suggests a more active subtropical jet into parts of the southern US this winter.
A strong –AO October, combined with a preliminary look at the SAI and OPI, suggests that a –AO winter is likely. There are more mixed signals on the potential NAO state this winter, however the SAI claims to correlate to the NAO as well. If the ridging forced by the Tropical Pacific is focused closer to the west coast, a –AO may help drive cold into the eastern US this winter. However, if the Tropical Forcing ends up farther east than I currently anticipate, then we would waste the –AO unless an –NAO could also develop.
Given I expect general ridging over western N. America this winter, I expect a dry and warm winter across the western US (especially the NW US). An active sub-tropical jet may bring near normal temps and precip to the SW US, however I’m not sure the El Nino will be strong enough to reverse or even alleviate the extreme drought ongoing in California. The active sub-tropical jet will also likely bring cooler and wetter conditions than normal to the Gulf Coast and SE US, with above normal precip also potentially occurring into the Mid-Atlantic. A ridge on the west coast of the US would probably favor mean troughing over the Great Lakes and Northeast, suggesting cooler than normal conditions in this portion of the country. The Midwest would likely also be drier than normal in general.
Storm tracks that may be common may be low pressures that slide across the Deep South and bring overrunning type wintry precip from parts of the mid-Mississippi Valley eastward towards the Mid-Atlantic and parts of the SE US. Depending on how the AO and NAO play out, storms may either cut NE near the Appalachians or off the East Coast. That’s a broad range, but that is how important getting a –AO and at times –NAO will be. This could be a variable winter for the East Coast, with a lot hinging on the potential NAO state, for which I couldn’t find any really strong signals for right now. So, with signals for western US ridging and a –AO, I will paint in below normal temperatures across the eastern US, however it should be noted that snowfall prediction is a crap shoot, especially for the coast plain. It is possible to have a cold/wet winter but not see a lot of snow in that region if individual storms tend to cut inland with cold coming in behind them. This may be where the NAO state comes into play.
If time permits I may update a few things in early November if any signs become clearer. Until then, enjoy the lack of winter weather across most of the country for the next couple of weeks.
Ramblings of Meteorologist Jim Sullivan 