The Differences between Sea Surface Temperature Datasets Prevent Us from Knowing Which El Niño Was Strongest According NINO3.4 Region Temperature Data

Guest Post by Bob Tisdale

In the November 17, 2015 post here (WattsUpWithThat cross post here), we discussed the recent alarmist claims by the mainstream media and members of the climate science community about the strength of the current El Niño, showing how and why the 1997/98 El Niño was actually stronger than the one we’re presently experiencing. We clearly illustrated that the El Niño developing this year is focused more toward the central equatorial Pacific…in the NINO3.4 region.  That NINO3.4 region was the only region examined by the mainstream media.  But those articles overlooked the eastern equatorial Pacific (the NINO3 and NINO1+2 regions), where the El Niño of 1997/98 was noticeably stronger than the current one. (See the map here for the locations of the NINO regions.)

In this post, we’ll return to the NINO3.4 region and present the sea surface temperature anomalies there using the numerous monthly sea surface temperature datasets for the evolution years of the 1982/83 El Niño, the 1997/98 El Niño and the current 2015/16 El Niño through October 2015 for most datasets. Those are the three El Niños with the highest NINO3.4 sea surface temperature anomalies since 1900. We’ll show that:

  • There is no agreement among sea surface temperature datasets about the NINO3.4 values for those El Niños,
  • There is no agreement about which El Niño is strongest based on the NINO3.4 sea surface temperature data, and,
  • Based on the averages of the datasets, the sea surface temperature datasets indicate the El Niño this year is lagging behind the 1997/98 El Niño, but is stronger than the 1982/83 El Nino.

IMPORTANT INITIAL NOTE

The sea surface temperature anomalies of the NINO3.4 region are a commonly used metric for the strengths, frequencies and durations of El Niño and La Niña events.  However, they only represent the sea surface temperature anomalies of an east-central portion of the equatorial Pacific bordered by the coordinates of 5S-5N, 170W-120W. Nothing more, nothing less.  As such, they do not capture the sea surface temperatures for eastern equatorial Pacific where the strongest El Niños also impact. That was the basis for the post Is the Current El Niño Stronger Than the One in 1997/98?

THE CAUSE OF THE HUBBUB

Figure 1 shows a graph of the weekly sea surface temperature anomaly data for the NINO3.4 region of the equatorial Pacific that was used to support the misleading alarmist claims about the strength of the current El Niño. The data source is the NOAA Climate Prediction Center (CPC) Monthly Atmospheric & SST Indices webpage, specifically their Weekly OISST.v2 data. As shown in Figure 1, there was a recent uptick in the weekly sea surface temperature anomalies for the NINO3.4 region, which showed its value exceeding those reached at the peak of the 1997/98 El Nino.

01 Weekly NINO3.4 Evolution - Reynolds

Figure 1

That NOAA Optimum Interpolated sea surface temperature dataset (a.k.a. Reynolds OI.v2) is based on observations from ship inlets, buoys (both moored and drifting) AND from satellites.

Note:  The drifting buoys are not ARGO floats.  ARGO floats only appear at the surface every 10 days to communicate with satellites, while the drifters are always on the surface. [End note.]

Unlike datasets that rely only on ship inlet and buoy observations, which are only available on a monthly basis, the Reynolds OI.v2 satellite-enhanced sea surface temperature data is available in weekly and monthly formats, which is the obvious reason why NOAA uses them for their weekly ENSO indices.  (There is also a higher resolution, daily version of the Reynolds OI.v2 data, which serves as the basis for many of the daily sea surface temperature maps available around the internet.)

-HOWEVER-

As discussed in the post Is the Current El Niño Stronger Than the One in 1997/98? the NINO3.4 data based on the weekly Reynolds OI.v2 sea surface temperatures are not the “official” NOAA ENSO Index.  That “official” NOAA index is their Oceanic NINO Index (a.k.a. ONI) and it shows the current El Niño lagging behind the event of 1997/98…not running alongside it in recent months.  See Figure 2.

02 Monthly NINO3.4 Evolution - ONI

Figure 2

The Oceanic NINO Index is presently based on the NOAA ERSST.v4 sea surface temperature reconstruction, also known as the “pause buster” data.  It uses a 3-month running-average of NINO3.4 region sea surface temperature anomalies, but the anomalies are not referenced to a fixed 30-year period.  They use multiple base periods to supposedly account for the effects of “global warming” (See the NOAA webpage here). But, as we discussed back in 2012 when NOAA started using the multiple base periods, those shifting base years simply eliminated the impacts of the 1976 Pacific Climate shift on the NINO3.4 data, which was responsible for the long-term warming there. See the post Comments on NOAA’s Recent Changes to the Oceanic NINO Index (ONI).

And there are numerous other monthly sea surface temperature datasets.  Do they confirm or contradict the alarmist claims that the current El Niño is the strongest on record?  And are they consistent about the strengths of El Niños?  Let’s start with the…

MONTHLY SATELLITE-ENHANCED SEA SURFACE TEMPERATURE DATASETS

Figure 3 includes the monthly NINO3.4 sea surface temperature anomalies during the evolutions of the 1982/83, the 1997/98 and the current 2015/16 El Niño events through October 2015, using the Reynolds OI.v2 sea surface temperature data.  Based on the weekly data (Figure 1) we might expect the monthly values to rise above the 1997 values in November.

03 Monthly NINO3.4 Evolution - Reynolds

Figure 3

The other satellite-enhanced sea surface temperature dataset is the UKMO HADISST.  It is a long-term reconstruction (starting in 1870) that relies on data from ship inlets and buoys to 1981 and satellite data (in addition to data from ship inlets and buoys) starting in 1982. See Figure 4.  Unfortunately, the updates of the HADISST data are delayed a few months, so we only have 2015 data through August.  But, like the Reynolds OI.v2 data, the HADISST data show the 2015 NINO3.4 sea surface temperature running neck and neck with those of 1997.

04 Monthly NINO3.4 Evolution - HADISST

Figure 4

Curiously, only the satellite-enhanced sea surface temperature datasets show 2015 values comparable to those in 1997.

Also note how the Reynolds OI.v2 data show the 1982/83 NINO3.4 sea surface temperature anomalies peaking slightly higher than the 1997/98 values, while that was not the case with the HADISST data.

NOTE:  All of the monthly data presented in this post other than the weekly Reynolds data (Figure 1) and the NOAA Oceanic NINO Index (Figure 2) were downloaded from the KNMI Climate Explorer, using the WMO-preferred base years of 1981 to 2010. [End note.]

ICOADS SOURCE DATA

The source of the in situ (buoy and ship inlet) sea surface temperature data used in all sea surface temperature products is a database maintained by NOAA called the International Comprehensive Ocean-Atmosphere Data Set (ICOADS).  It shows NINO3.4 region sea surface temperature anomalies in 2015 close to but lagging slightly behind the values in 1997, even with the uptick in October, but ahead of those in 1982.  See Figure 5.

05 Monthly NINO3.4 Evolution - ICOADS

Figure 5

NOAA and UKMO then adjust those values for their sea surface temperature products, and in the case of the satellite-enhanced datasets (Figures 3 and 4), they then merge bias-adjusted satellite data with the buoy and ship inlet data.

“OFFICIAL” PRODUCTS USED IN GLOBAL LAND+OCEAN SURFACE TEMPERATURE DATASETS

NOAA and GISS now use the NOAA ERSST.v4 “pause-buster” sea surface temperature reconstruction in their combined land+ocean surface temperature products.  The UKMO uses HADSST3 sea surface temperature reconstruction for its combined land+ocean HadCRUT4 product.  As illustrated in Figures 6 and 7, both of those “official” sea surface temperature datasets show the 2015/16 El Niño lagging behind the values reached in 1997.   Notice also that the HADSST3 data, like the Reynolds OI.v2 data, show the 1982/83 NINO3.4 data peaking slightly higher than the 1997/98 El Niño.

06 Monthly NINO3.4 Evolution - ERSST.v4

Figure 6

# # #

07 Monthly NINO3.4 Evolution - HADSST3

Figure 7

FORMER NOAA SEA SURFACE TEMPERATURE RECONSTRUCTION

NOAA replaced their ERSST.v3b sea surface temperature product with the “pause buster” ERSST.v4 data earlier this year.  Curiously, NOAA is still maintaining the now-obsolete ERSST.v3b data. (I’ll let you speculate about that.) Figure 8 shows the evolutions of the NINO3.4 sea surface temperature anomalies for the 1982/83 El Niño, the 1997/98 El Niño and year-to-date for the 2015/16 El Niño. Like the “official” NOAA “pause buster” data, the ERSST.v3b reconstruction shows 1997/98 NINO3.4 anomalies were higher than what we’ve been experiencing recently.

08 Monthly NINO3.4 Evolution - ERSST.v3b

Figure 8

And for those interested, Animation 1 compares the two NOAA ERSST products (Figures 6 and 8).

Animation 1

Animation 1

NO AGREEMENT AMONG DATASETS

Animation 2 compares all of the graphs of the monthly evolutions for the 1982/83, the 1997/98 and 2015/16 El Niños.

Animation 2

Animation 2

As shown, there is no agreement on the values of the NINO3.4 sea surface temperature anomalies for the three El Niños and no agreement about which El Niño (1982/83 or 1997/98) had the highest NINO3.4 values.  We can also illustrate this with a spaghetti graph which includes all of the datasets, Figure 9.

09 Monthly NINO3.4 Evolution - All

Figure 9

Now imagine if we were to include the uncertainties of the individual NINO3.4 data, which aren’t published.

THE “CONSENSUS”

Let’s average the ERSST.v3b, ERSST.v4, HADISST, HADSST3, ICOADS, and Reynolds OI.v2 data.   See Figure 10.  Through October 2015 (August for the HADISST data), the NINO3.4 sea surface temperature anomalies in 2015 do not exceed the values reached in 1997.  And the 1997/98 El Niño peaked higher than the 1982/83 El Niño.

10 Monthly NINO3.4 Evolution - Average

Figure 10

CLOSING

Based on the differences among datasets, we do not know for certain if the 1982/83 El Niño or the 1997/98 El Niño had the higher NINO3.4 sea surface temperature anomalies.  And we won’t know for certain whether the El Niño we’re currently experiencing will cause the NINO3.4 sea surface temperatures to reach values that exceed those reached in 1982/83 and 1997/98.

All we do know is that we’re presently experiencing a strong El Niño, which has had an impact on global surface temperatures this year, raising them to “record levels”…and will influence weather around the globe for better and worse…and may cause an upward shift in the sea surface temperatures for much of globe when all of the subsurface warm waters released by the El Niño are redistributed throughout the oceans at the end of the El Niño.

UPDATE: FOR THOSE NEW TO DISCUSSIONS OF EL NIÑO EVENTS

I discussed in detail the naturally occurring and naturally fueled processes that cause El Niño events (and their long-term aftereffects) in Chapter 3.7 of my recently published free ebook On Global Warming and the Illusion of Control (25 MB). For those wanting even more detail, see my earlier ebook Who Turned on the Heat? – The Unsuspected Global Warming Culprit: El Niño-Southern Oscillation.

Advertisements

37 thoughts on “The Differences between Sea Surface Temperature Datasets Prevent Us from Knowing Which El Niño Was Strongest According NINO3.4 Region Temperature Data

  1. In the November 17, 2015 post here (WattsUpWithThat cross post here), we discussed the recent alarmist claims by the mainstream media and members of the climate science community about the strength of the current El Niño, …
    Another great post, thanks. But my question is why should the alarmists tout the strength of any El Nino since they don’t seem to be related to the evil catastrophic anthropogenic global warming. Or are they?
    By the way, get a load of the top hit by google on “anthropogenic”.

    an·thro·po·gen·ic
    /ˌanTHrəpōˈjenik/
    adjective
    adjective: anthropogenic
    (chiefly of environmental pollution and pollutants) originating in human activity.
    “anthropogenic emissions of sulfur dioxide”

    • I think ANY headline in the mainstream media that includes “worst ever” or “highest ever” and weather is assumed to be caused by global warming by the unsuspecting mainstream media reader.Regardless of whether or not the connection is written in the headline/article- the fear factor over global warming is amplified. The average person doesn’t have the time or inclination to read the facts, so doesn’t even know they are being manipulated. Disgusting really. I am so appreciative of Anthony for maintaining this website and encouraging intelligent authors such as Bob (and so many others) to share their knowledge with us.

  2. Excuse me, what are we focusing on? NOAA is apparently pushing the temperatures upward to support some sort of anti scientific ( Karl Popper) ideology. This discussion is about trying to throw some doubt on their methods. The elephant in the room is the fact which can not be refuted is that the interstadial period is nearing the cyclical periodicity previously demonstrated from the Vostok ice cores for the climatic shift into another ice age. The maximum temperature trends between the Minoan, Roman and Medieval climate optimums is decreasing. The Little Ice Age which ended around 1880 is the first such significant drop in Global temperature in 12,000 years. The political narrative is that we should be concerned by the observable transition since 1880 toward another climate optimum. I submit to you that the discussion is focusing through manipulation on countering the CAGW narrative in a cyclical rising climate temperature environment at the expense of the broader and more significant phenomena which is the predictable end of the interstadial period.

    • Thanks for using the term: “climate optimum”, which I feared had disappeared from the language. They’ve been calling it the “Medieval warm period”, thereby denying that a temperature warmer than today was, in fact, “optimum”.

  3. Just watched the local news here in Texas.
    Lots of reference to the “Average First Freeze” date: and it being right on the date expected for the first freeze.
    So, my little brain says to me: “Bet there is first freeze info for many years in many places world wide.
    No reason for these temp.’s and dates to be messed with so far.
    Go get them before they get fudged.
    Use the to show that no difference world wide on first freeze dates.
    Just saying, if it is warmer, the freeze dates should change too.

    • I admit to not having done all of Bob’s pre-reading and do not deny the Nino has wide ranging effects (sometimes spectacular here in the FL) but I thought I had gleaned the Nino ends up being a cooling event; the red blobs eventually get lost to space? Seem this might be a case where the alarmism cause/effect are kinda backwards? Driving from the rear-view mirror? It seems the; this red blob is redder than that red blob argument is mostly irrelevant to climate. However, I am keeping the weather alert radio charged this year.

  4. Seeing as NOAA is currently unable or unwilling to provide the supporting data for the Karl et al 2015 paper, how can we trust this data?
    NOAA has form for hiding their working. False in one thing, false in all.
    This data is untrustworthy.
    Because NOAA is untrustworthy.

  5. Thank you very much, Bob. This is a very good article.
    It would be funny if it wasn’t so sad; as time passes, we know less about our world.
    Poor NOAA, Dr. Karl wanted to kill the “pause”, but the weapon backfired and he badly injured NOAA.

  6. How can this be considered a “record” temperature when the overall atmosphere is nowhere close to a record, and in all previous El Niño’s in the satellite record the troposphere warms much more then the surface???

  7. Looking at the graph I estimate the error bars are +/- 0.7degC across the diverse measurement sources/techniques. This tells me that the recent GISS adjustments for ocean temperature are statistically insignificant. In fact it makes the whole trend across 135 years insignificant given that ocean temperature is the chief driver of average temperature (70% of the surface, a lot more heat capacity)
    Peter

  8. Thanks Bob.
    ~~~~~
    The naming thing (regular El Niño, Modoki, La Nada, whatever) reminds me of folks that would buy a puppy from us because the previous one died and they wanted another just like the first. Our advice: Do not name your new puppy the same as the last one and, the big and – do not expect it to behave just like the last one. Yes, it has the same parents. No, it will not be a new “dogsnamehere.”
    This concept explains why the Climate Prediction Center could get it wrong when forecasting Washington State would have a warm and dry November (as of Oct. 15).
    Here’s a snip from the Seattle NWS for next week: SHOWERS MONDAY NIGHT AND TUESDAY AND COLD AIR ALOFT WILL CAUSE THE SNOW LEVEL TO FALL BELOW 1000 FEET.
    Did I mention the current flooding, wet soil, downed trees, …

  9. Bob, which SST data set does Australia’s BoM use? Quote from BoM’s ‘ENSO Wrap-Up’ sea surface page.
    Quote:

    NINO3.4 still remains behind the peak monthly anomaly value reached during either 1982 or 1997 (+2.8°C and +2.7°C respectively). Note: peak values are typically recorded late in the year.

    According to BoM, 1982 had the highest SST (NINO3.4).
    Note: I do believe that this weeks SST (NINO3.4) values reached +3.0°C.

    • Country people in Queensland are interested in the SOI as an indicator of El Nino (persistent below -10) and dry times, and La Nina (persistent above +10) and wet times. It appears that on the 1st or 2nd Nov there was a turning point marking the end of the El Nino. In 1998 (May I think) there was a very rapid rise of the SOI marking the end of the El Nino. That is what seems to be happening now for three weeks. The alarmists can dream of a super El Nino but the sun and clouds are not obliging.

  10. Joe Bastardi has a great Saturday summary discussing the current El Nino and it’s forecasted effect on N. American winter weather. He also seems to believe that the subsequent La Nina is going to be strong and deep and says the indications of that are already showing up.
    Anyway, what he is showing is that apparently this El Nino has peaked and it’s down hill from here. That fact and the fact that the SSTs in Nino 1.2 regions and basically the eastern half of the 3 region makes this El Nino seem rather atypical does it not? Is there one on the record that has followed this pattern?

  11. As the precipitation is also due to Jet Stream paths, while the ocean temperatures are interesting but all factors must be considered. Scientists will observe data, and after close analysis and discussion will make conclusions based upon multiple examples of an observed anomaly, and only then publish their opinions. Back off and act like scientists; there is insufficient data on El Nino events to make these wile predictions.

  12. If there are so much variations in observations in sea surface temperature, then what about the theories on hiding heat of global warming in Oceans.
    Dr. S. Jeevananda Reddy

    • Well, Dr Reddy,with that in mind I got Bob to point me to the data on 150-300m temps. So far, on first look, cannot see anything unusual. But that was only first look…

      • But that was only first look…

        Why bother. The temperature error as shown by the graphs above is +/- .75degC. I doubt you could hide anything useful given that the ocean is 7/3 more surface area of the land, multiply that by 10x heat capacity of water over air. The 1degC rise since 1880 isn’t noticeable with those kind of error bars.

  13. Sea Surface Temperature data sets have error bars so large as to be virtually useless, why does this discussion go on and on? British Navy began recording temps from the oceans in 1851, using buckets and liquid-in-glass thermometers, scrupulously recorded by the cabin boy. Now we are debating whether these records could tell us whether SST’s either have or have not risen or fallen by hundredths of degrees?
    Why is this even a continuing discussion? Data must be taken and recorded following significant, important rules, none of which have been followed with these “data” sets. The entire thing about the depth of ocean vessel engine cooling water intakes also invalidates these “sets.”
    Arguing about the details of data which has been taken incorrectly is useless, all claim information not actually in existence. Why these endless posts and graphs?
    You cannot prove an error by your opponents, who have not shown any facts according to traditional procedures, by giving them legitimacy ignoring all those same traditional procedures. Data and error bars are irreducibly connected. Data, is it data, or is it not data???
    All engineers scoff at incorrect data procedures. This is because, if we do not follow correct data procedures, we get fired, things break, people die, money is wasted.

    • Never mind the 1880s. The data sets above are 1980 and later, and they vary about +/- 0.8 degC from the center. It can only get worse the farther back you go.

  14. Bob First of all I love everything you do. But the longest running measure of the Enso is the Southern Oscillation Index. And in a practical aspect, it is a much better indicator of what the ATMOSPHERE is up to in a crucial source region for the global weather pattern, the tropical Pacific from north of Australia to east of the Date line. That shows this not to be and in fact, the 30 day running SOI is now in the neutral base state as the demise of the total package starts. The enso 1.2 is crucial also, for when it is cooler than 3.4, though above normal as it is, it leads to a different result in the North American winter. Finally many of us in the forecasting field love the MEI, the multivariate enso index. My point is the obsession with 3.4 does not tell the whole picture, as the sst are much colder this year from 140 west further east, with it warmer to the west than the great Nino of 1997. A look at this site http://www.cpc.ncep.noaa.gov/products/precip/CWlink/MJO/CLIVAR/clivar_wh.shtml shows that even the CFSV2 is catching what analogs show, and what the JAMSTEC and Scripps sites show, that unlike 1997 and 1982 for instance, that peaked in winter, this one is about to take a tumble. The major la nina I believe is on the way late 16 into 18 will reverse the current global temp spike, and via NCEP real time temps mimic or exceed the previous 2 downturns in the wake of the enso events of 06-7/09-10. But there is much much more than enso 3.4 to the total picture and the SOI experiencing the greatest November RISE in an enso year on record is just as important to to the big picture as the enso 3.4 temps which seems to be the focus of most. https://www.longpaddock.qld.gov.au/ this has a chance, given the next 10 days, which appear to be neutral to positive, with the front 10 days having been strongly negative to be very close to 0 after a 30 day base that was indeed super nino . I can not find anything like it in the SOI data base

  15. Since NOAA and most still regard ENSO as a variation around a climate state, its really not appropriate to analyze the SST anomalies using the same base period. However, since the climate has paused for 18 years or so comparisons between 1997/8 seem OK even though eventually there may be large differences between the two base periods for the 1997/8 and 2015/16 El Ninos. (NOAA uses a centered 30 year base period that has not been fully implemented in all the modeling outputs).
    But since the pause has been at a temperature level some tenths of a degree above the climate of 1982/3 the base period for 1982/3 should be 1966-1995, definitely cooler than 1981-2010 being used. And of course eventually the 2015/16 El Nino will break over the climates of 1996-2025 and and 2001-2030. So I suspect that if those adjustments were made it would be projected that the 2015/16 El Nino would come in at 3rd or 4th with a possible tie with the 1972/73 El Nino.
    And who knows it could even come in a little lower since the main hallmark of model changes over time has been to diminish historic warming trends and increase current and projected warming trends.

  16. Two simplistic observations about the graphs…
    1.) 1982/3 to 1997/8 = 16 years; 1997/8 to 2015/6 = 18 years. We have a possible cycle here, having 3 data points.
    2.) The hiatus has lasted 18 years and some months. Prior to that was about 16 to 21 years rise. The sum of the two is about 14 to 39 years, approximately an average PDO phase.
    Just sayin’…

  17. Spent Thanksgiving week in Phoenix with son and his wife. They are expecting a baby girl in May, their first child and our first grandchild.
    One of our fun activities is visiting used book and thrift stores. On a previous trip I found a couple of interesting books on weather and climate. This trip among others I found “El Nino – Unlocking he Secrets of the Master Weather Maker” by J. Madeline Nash. Wow, fascinating reading. How the science behind el Nino developed, the scientist and organizations. Bjerknes, Leetma, CPC, PDO, AMO, Kelvin and Rossby waves, Rio Nido, etc. Published in 2002, 13 years after I first encountered AGW in 1989.
    It’s El Nino, La Nina, PDO, AMO, wind and water vapor that drive the weather and climate, GHGs contribute nada, bee fart in a hurricane. But then many of you already knew that.

Comments are closed.