Posted by: Carl | October 26, 2008

El Ninos in the Pacific, Atlantic, and Indian Oceans

El Ninos and La Ninas are at the center of our discussion of climate change. What causes them and what are their effects?  These are the two big questions in my mind, though determining the causes of these tropical events will likely prove fruitless, unless we can confirm the cirrus cloud suggestion made by Erl.  So, my post will show how El Ninos and La Ninas begin and how they spread, for this seems like a logical place to begin the discussion at this blog.  In the future, I’d like to explore the mechanism through which heat released by El Ninos moves after the sea surface temperature (SST) anomaly along the equator disappears.  Also, I’d like to determine if we can think of La Ninas and their effects in the same terms.  For clarity in this post, I will focus on El Ninos.

The Pacific Ocean:

For this analysis, I used SST anomoly from the region used to calculate the Cold Tongue Index (CTI).  I didn’t use the CTI itself, because to calculate it, the trend is removed from the data.  The rest of my data is extracted SST with no trend-removal, so to be consistent, I used the raw SST for the region with no trend removed.  The CTI is defined to cover the region (6 S to 6 N) (180 W to 90 W).  Below is a graph of SST anomalies in the CTI region.

Changes in the temperatuere of the Cold Tongue of the Pacific are not consistently the first signal of an El Nino event.  To illustrate this point, here are animations made by NOAA: 

As you can see, in the 97/98 El Nino, SST first rose along the Peruvian coast.  JISAO provides a SST dataset for Puerto Chicama, Peru.  Below are some graphs of the Cold Tongue SST Anomaly vs the Puerto Chicama SST Anomaly.

As you can see, this data also suggests that the 1997/98 El Nino began with changes in SST of the Peruvian Coast.  From this analysis, we can conclude a few things.  First of all, if cirrus cloud is initiating El Nino events, during specific El Nino events, it is relevant to know if cirrus clouds dissipate first over the Cold Tongue or the Peruvian Coast.  Secondly, a topic to pursue in the future is if variations in the temperature off the Peruvian Coast have different effects from variations in the Cold Tongue.

The Atlantic Ocean:

The Atlantic Nino is significantly different from the Pacific Nino, as you will see.  I’ve divided the Equatorial Atlantic into several quadrants.  Below is a graph of SST separated by Latititde.

Notice how SST above the equator has behaved opposite from SST below the equator since the 97/98 El Nino. 

Below is a graph of Atlantic SST separated by Longitude.  I haven’t separated the data like I did in the previous graph, because I just intend to show that the behavior of SST does not vary significantly with Longitude.

Below are graphs according to Latitude and Longitude for the 82/83 & 97/98 El Ninos.

Notice that what I am refering to as the 1982/3 El Nino seems to have occured in 1984 in the Atlantic.  There was a small rise in SST at the time of the 82/83 El Nino, though the SST anomoly in 1984 in the Atlantic was unusually large, indicating that it may have been related to the El Nino in the Pacific nearly two years earlier, though it may merely be a coincidence.

As you can see, the Atlantic Nino is most apparent in the Eastern Atlantic between 0 and 8 S.  The strength of the SST anomaly decreases as you move West and North.

Indian Ocean:

El Ninos in the Indian Ocean resemble El Ninos in the Pacific.  Below are graphs of the 1997/8 El Nino according to Latitude and Longitude.

It is clear the El Nino events in the Indian Ocean begin in the West and move East at all latitude bands.

Comparing the Oceans:

Below is a map of where, according to the analysis above, El Ninos begin in each ocean. 

I’m leaving out the coast of Peru because the graphs below will use only SST anomalies in the Cold Tongue.

Below are graphs comparing the Pacific, Indian, and Atlantic Nino regions since 1978 and during the 1982/3 Nino and the 1997/8 Nino.

The Atlantic Nino seems to lag behind the Pacific Nino considerably during some events and not at all during others.  The Indian Ocean also seems to lag behind the Pacific on some occasions but not others; however, the variable lag time is not as extreme in the Indian Ocean as it is in the Atlantic.

Source: Smith and Reynolds Extended Reconstructed SST (ERSST.v2)

Use the instructions in this post to access the data; however, don’t read the read the rest of the post – it’s wrong.



  1. Hi Carl,
    Good to introduce the concept of an Atlantic Nino and an Indian Ocean Nino. That will hopefully start people thinking about generalized tropical warming events (because thats what we are dealing with) and what might be causing them.

    There has been a 5°C warming at high latitudes in winter as a result of a long series of El Ninos. We wouldn’t want people to attribute that to anything else but tropical warming events.

    But that seems to be reversing this winter and last.

    Just looking at your curves lit seems as if the Southern Hemisphere is where most of the action is and contributing to the warming of the Northern. Is that what you see?

  2. Carl: I enjoyed your post and will have some more comments tomorrow. You appear to enjoy evaluating short-term data, so I’ll recommend the Optimally Interpolated SST data. It’s available on a WEEKLY basis through NOMADS. The data starts in November 1981. Right now, their machine’s overloaded or they’re updating data; regardless, it’s down; otherwise I’d attach a link and walk you through data retrieval. I’ll try again in the morning.

    Also, if you would, please put titles on the graphs, so we can keep track of what we’re looking at. That would be a big help. I spend more time labeling graphs than I do downloading data and creating the graphs. It’s time consuming, but it helps those viewing what you’ve prepared.

    FYI, Trenberth et al evaluated the evolution of El Ninos and their effects on global temperature. It’s worth a look. Here’s a link:

    My initial comments on the content of this post:
    The responses and lags you’re documenting between the El Ninos and the Atlantic and Indian Oceans appear to be the result of atmospheric Rossby waves. Have you looked for the responses to oceanic Rossby waves?

    The area you’ve marked on the map in the equatorial Atlantic is part of the Benguela Current, which carries water northward along the Southwest African Coast. South of there (upstream) there’s lots of variation in SSTs due to random changes in upwelling. Since you’re looking at short-term data you may not need to account for it, but keep it in mind. Could the upwelling be providing false pre and post El Nino variations?

  3. Bob Tisdale – Thank you for your comments. I check your blog everyday and think that it is one of the more important climate blogs out there. A few old posts especially interest me, and I saw your recent post at Watts’s site. In it, do you imply that ENSO variation/the sign of the PDO are determined by SSTs in the Southern Ocean and THC/MOC? And that ENSO can explain most or all of modern temperature change? Could you elaborate? Do you see a significant roll for the AMO or the NAO/AO?
    The weekly data sounds extremely interesting, and I’ll be excited to see that. Graphs from now on will be titled; as annoying as it is to do, you’re right.
    I found a paper that seem relevant to our discussion on oceanic Rossby waves:
    Also, could El Ninos in the Atlantic move heat poleward through rossby waves? The anomaly region seems so small that I wonder if changes in the equatorial Atlantic could significantly effect climate at high latitudes.
    It may be worth graphing the SSTs where the upwelling in the Atlantic occurs, to see how it may influence the El Nino region I’ve marked.

  4. Carl: The closest index I can find to your Atlantic Nino is the Gulf of Guinea Index (12S-3N, 20W-10E) discussed in “Beyond El Nino: Decadal and Interdecadal Climate Variability”, Antonio Navarra, Editor.

    It’s also discussed in a few other papers and its coordinates vary in some of the others. There’s a 1984 Atlantic Warming Event that’s mentioned a few times.

    The area of your Indian Nino appears to be (or close to) one of the areas used in the Indian Ocean Dipole (IOD). There are numerous studies of it.

    The NOMADS OI SST web pages are still down. I’ll try again later.

    I’ll try to answer your questions later today. I still have not had my morning cup of coffee yet.

  5. Carl: Also, thanks for the kind words.

  6. Bob and Carl,
    I am thinking out loud here.

    Is it possible that by focusing so closely upon the oceanic regions where the water temperature changes most dramatically, e.g Nino 3.4, Gulf of Guinea and Western Indian Ocean we are missing the real action? Right enough, we are observing the sea surface phenomena in its most extreme manifestation. But I put it to you that the rate of upwelling of cold waters is probably driven by the change in the strength of the Easterly component of the trade winds. That, I suggest is due to a change in the pressure gradients in the rain shadow zones in the East. The earliest indication of a change in the easterly component of the wind in the form of an available index is the SOI. The SOI always leads the ENSO index. It’s based on air pressure. Tahiti is well south of the Equator. In my view, in focusing so closely on sea surface temperatures we are looking at effects not causes.

    So, I suggest that if we are to look at the origins of these tropical warming events we need to look at the atmosphere not the sea. Only one thing can warm the water and that is the sun. Water flows hither and thither driven by the wind and if the ocean as a whole is not being warmed as much as in the past then in those places where cold water comes to the surface, that water will be cooler. If the Easterly blows strong all the time lots of cold water comes to the surface in the East and some also comes up to the equator from the poles.

    Somewhere, there is a zone where there is a fluctuation in the heat input into the ocean (clouds). That atmospheric zone will be drier. It will have high ozone levels in the upper troposphere where ozone absorbs UVB. That sets up a convection force in the upper troposphere that weakens the pressure down at surface level. If this zone lies in the eastern margin of the ocean that then weakens the easterlies.

    Every ocean has its own unique meteorologic characteristics. The Atlantic must be much influenced by the Sahara on the one hand and the Amazon on the other. The balance of forces will be different so one could expect that the influences on easterly wind tendencies will be different between the major oceans. If we are to look for cold water up-welling in the Indian Ocean I think it will be south of Indonesia. But, the real challenge is to document the change in the rate of energy gain in the ocean. That energy gain is not happening in the upwelling zones.

  7. Carl: I’m not avoiding your questions about my comment/post at wattsupwiththat, though it may seem like it. I’m thinking of posting it on my blog and going into a more detailed explanation there. Explaining it would take more time than I have available right now, and there are other posts I want to create before I get to it, some of which will help explain it.

    With regards to the AMO/AO/NAO, I believe the AMO did contribute to the warming over the 20th century. There are papers that identify the impact the AMO has on Northern Hemisphere and Global temperatures. Refer to the Knight et al paper “A Signature of Persistent Natural Thermohaline Circulation Cycles in Observed Climate”.
    I often link to a glossary page at RealClimate, where they state, “This pattern is believed to describe some of the observed early 20th century (1920s-1930s) high-latitude Northern Hemisphere warming and some, but not all, of the high-latitude warming observed in the late 20th century.”

    Naysayers often point out that the AMO is an oscillation and would also contribute to global cooling during the periods when it’s declining. BUT, big but, a too often-used starting point for claims of global warming is the beginning of the 20th century. The AMO was in or near to the bottom of a trough at 1900, and it is presently at or near a peak, so it would have contributed to the warming since 1900. Refer to the following graph of the AMO and North Pacific Residual.

    On the other hand, there would have been no contribution from the North Pacific Residual since 1900 due to the timing of its oscillation.

    I can’t answer your questions about the Rossby waves from the Atlantic NINO area. I haven’t studied that area to any length, other than as part of my post on the Tropical Atlantic. A question for you: Does a long-term graph of the Atlantic NINO area agree with the data from the South Atlantic upwelling area discussed in the following?
    Note that those graphs are smoothed with an 85-month filter. You may want to decrease the filtering for your comparison.

    I erred in my comment last night. OI SST data is available from 1981 to present through NOMADS on a monthly basis. It’s also available in two phases on a weekly basis: the 1980s and then, separately, 1990 to present. The data sets are centered on different days of the week. So if you were going to evaluate something on a weekly basis from 1980, you’d have to split the graphics. I don’t think you can merge the two. The instructions for downloading are here:
    The start page they describe is here:
    I’ve made the following page a “Favorite”. It eliminates one step in the process.

    Note that in the instructions they warn against its use for coastal regions. Unfortunately, they don’t define coastal region.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: