The La Nina Pump

Guest Post by Willis Eschenbach

Sometimes a chance comment sets off a whole chain of investigation. Somewhere recently, in passing I noted the idea of the slope of the temperature gradient across the Pacific along the Equator. So I decided to take a look at it. Here is the area that I examined.

Figure 1. 17-year average temperatures, CERES dataset.

I’ve written about this temperature gradient before, in a post called The Tao of El Nino. If you take time to read that post, this one will make more sense. Here’s the money graphic from that post:

Figure 2. 3D section of the Pacific Ocean looking westward along the equator. Each 3D section covers the area eight degrees north and south of the equator, from 137° East (far end) to 95° West (near end), and down to 500 metres depth. Click on image for larger size.

The main idea put forward in that post is that the El Nino/La Nina phenomena function together as a cyclical pump to pump warm water to the poles. A cyclical pump has two cycles. In the first cycle, the El Nino intake cycle, the pump fills up with whatever is going to be pumped.

This intake cycle is shown in the left part of Figure 2, the El Nino section, you can see the buildup of warm water.

The second cycle of a cyclical pump, the ejection cycle identified as “La Nina”, is where the pumping action actually occurs. The result of this cycle is shown in the right part of Figure 2, where the warm water has been pumped out of the area.

What happens in this second cycle is that at a certain point, the warming of the equatorial eastern Pacific waters makes strong winds kick in. These winds push the warm surface water first to the west, along the equator in the blue zone shown above. From there, the warm water splits and moves towards both poles. Overall, this cools the planetary surface.

Given the effect of La Nina in terms of water movement, let me call the action of the combination the “La Nina Pump”.

I started my latest peregrination by looking at the long-term average slope of the temperature across the area encompassed by the blue box in Figure 1. Figure 3 shows that result for the Reynolds sea surface temperature dataset from 1981 on. Figure 4 shows the same result for shorter the CERES dataset, from 2000 on.

Figures 3 & 4. Long-term average equatorial Pacific sea surface temperatures, by longitude. This is the averages by 1° longitude of the ocean within the blue box in Figure 1. The Reynolds data serves to confirm the shorter CERES dataset … and vice versa.

The right-hand end of the blue box, at ninety degrees west longitude near South America, is the cold end. It’s at about 26°C (79°F) or so.  At the other extreme, the left-hand end of the blue box, at a hundred fifty five degrees east longitude near Asia, is the hot end of the box. It’s at about 30°C (86°F). So this was good. I’d graphed out what they had called the “temperature gradient”.

However, this was not satisfying. I wanted to see how that temperature gradient changed over time. In particular, I wanted to see how it responded to El Nino and La Nina events.

So I made up a movie, showing month by month from 1981 to 2018 what the temperatures in that area were doing. However, that movie, too, was not what I wanted, because I couldn’t see what the status of the El Nino was.

As you might imagine, my next move was to add the MEI, the Multivariate ENSO Index, to the movie. Of course, having done that, I also had to add the NINO34 Index, the Oceanic Nino Index (ONI) and the inverted Southern Ocean Index (SOI).

So I made the movie, and I watched it over and over to try to understand what was happening. I realized that a) the various NINO indices all moved as a fairly tight group, and b) it sure looked to me like the NINO indices moved in sync with the average temperature of the entire region in question.

So I checked to see how well the average temperature for that area (8°N/S Latitude, 155°E – 90°W Longitude) matched the other El Nino indices. Figure 5 shows that match. I’ve used the Reynolds OI sea surface temperature dataset, available here, to calculate the average area temperature.

Figure 5. El Nino indices, including the average temperature 8°N/S latitude, 155°E – 90°W longitude. All indices have been standardized. The SOI has been inverted to agree with the sense of the other four indicators.

As you can see by how little of the red line is visible, the average temperature of that blue box area tracks almost exactly with the other indices. So I added the average temperature to the El Nino indicators, and here’s what that final movie looks like.

Figure 6. Movie showing the month-by-month changes in the cross-Pacific equatorial temperature gradient. Southern Ocean Index is inverted. All El Nino indices are standardized to a mean of zero and a standard deviation of one.

I was greatly encouraged by a couple of things in this movie. One is the similarity of this action to that of an old-time fireplace bellows. It looks like a pump. Doesn’t mean anything … but I liked it.

The second encouraging thing was finding out that the average temperature of the area is an excellent indicator of the state of the El Nino phenomenon. It tracks very closely with all of the other El Nino indices. This brings up an interesting possibility, that I can do something I’ve wanted to do for a long time. This is to estimate the total amount of energy moved by the La Nina Pump. I’ll get to that in a minute.

Using the average temperature of the blue box area as an El Nino index, here’s what happened in that region over the past 37 years regarding the La Nina Pump.

Figure 7. Repeated cycles, El Nino and La Nina, as shown in average equatorial Pacific temperatures.

After pondering and examining this graphic for a while, I realized a most curious thing. The oddity is that in all cases, the temperature drop associated with the La Nina phenomenon began in November, and ended in the following November. Figure 8 shows all of the drops in the record that match that criterion. The shaded areas go from one November to the following November.

Figure 8. Highlighting the 13-month temperature drops associated with the La Nina Pump.

I like this graphic because it gives what seems to be a novel and very clear and unambiguous way to identify the intermittent operation of the La Nina Pump—it starts in November and ends the following November. It’s also quite curious, in that the one-year drop happens with both large and small El Nino/La Nina combinations.

Let me note that this is quite a different interpretation than the normal one, where a warm equatorial Pacific Ocean temperature is called “El Nino” and a cool ocean is called “La Nina”. This is a functional definition of the La Nina Pump, seen as starting when the ocean is warm and ending when the ocean is cool.

So finally, let me look at how much energy is transported by the El Nino Pump. Temperatures generally don’t change much vertically in the “mixed layer” of the ocean. This is the top 10 – 100 metres (33 – 330 feet) of the ocean which is constantly being mixed by the action of wind, currents, and nocturnal oceanic overturning. See here for details. Figure 9 shows the long-term average mixed layer depths.

Figure 9. Average mixed layer depth, metres. Allow me to digress a moment regarding the mixed layer. In the great Southern Ocean surrounding Antarctica, the latitudes from 40°S to 49°S are called the “Roaring Forties”, after the nearly continuous storms there. Heck, I just looked, it’s storming there right now. No surprise. And the latitudes in the next band south, from 50°S to 59°S, are called the “Screaming Fifties”. I always figured that “screaming” referred to the sailors, not the wind, so I’ve never taken a boat there. My point is that you can see how these insanely strong and constant winds, combined with cold, generally sinking surface waters, result in a mixed layer a hundred metres deep and more. We now return you to your regularly scheduled programming.

In the area of the Pacific in question, the mixed layer depth averages about 50 metres. However, it’s shallower in the eastern Pacific where the largest temperature swings are, so I’ll use 40 metres as a weighted average.

Now, I can estimate the amount of energy moved by noting that it takes about 4.1 megajoules of energy to raise a cubic metre of water by 1°C. And since we’re assuming that the mixed layer is approximately the same temperature from the top to the bottom, it takes about 4.1 * 40 metres deep ≈ 165 megajoules/square metre/degree to warm a one square metre X 40 metre deep column of the mixed layer by one degree.

Next, to calculate the total amount of energy lost, I looked at the total of the temperature drops during the shaded times in Figure 8 above. This was a total of 16.5 degrees of direct La Nina temperature drop over the period of record. I multiplied that by the 165 megajoules per degree and divided by the number of years, giving 74E+ joules/square metre/year. Then to convert joules to watts I divided by seconds per year, giving a final average constant energy loss of 2.4 watts per square metre.

What does this mean in more intuitive units? Well, over the period of record, the area shown in red above ended up 16°C cooler than it would have been without the La Nina cooling episodes shown in Figure 8.

Ruminations on the Equatorial Temperature

The first thing that strikes me is the total absence of any temperature trend in the last 37 years in this large region of the equatorial Pacific. This supports the idea that the El Nino Pump is part of the global thermoregulatory system. When warm tropical water builds up in the eastern Pacific, the La Nina winds spring into existence and they blow all that warm water first west to Asia, and thence to the polar regions. It is this La Nina Pump action, removing the warm water so it can be replaced by colder water from underneath, that prevents the equatorial Pacific from overheating.

In addition, the La Nina Pump does not just prevent overheating. This kind of threshold-based emergent phenomenon serves as an active thermostat. To do that in a lagged system, it must have “overshoot”. This means that at the end of a La Nina pump cycle, the temperature must be below, even well below, the long-term average … see Figure 8. From there it warms up until the next La Nina pump cycle, and so on. Net result?

Thermostatically controlled Pacific tropical temperature.

Anyhow … that’s what a chance reading of the idea of the “Pacific equatorial temperature gradient” led me to.

Best wishes to all on a sunny summer afternoon. 

The coastal fog burned off a couple of hours ago, and the plant-based solar collectors in the forest that surrounds our house are working overtime, soaking up CO2 and using sunshine to convert it to O2 …

w.