Weather Two Months From Now

Guest Post by Willis Eschenbach

A while back, folks noticed that a couple of months after the El Nino kicked in across the Pacific, the earth would warm up a bit. Since then, people have engaged in what they describe as “removing the El Nino signal” from the global temperature record. A while back I wrote a post called “Why El Nino and not the AMO“. If you have not read that post, it lays out some of my objections to the procedure of “removing the El Nino signal”. This post carries those ideas forward.

I got to thinking about just how well the El Nino 3.4 area does or doesn’t forecast the evolution of the global temperature, and I realized that I could look at that question using the CERES satellite data. Figure 1 shows the correlation of local temperatures with the global average temperature two months later:

Figure 1. Correlation of local temperatures with the global temperature two months later. All observations have had the monthly variations removed. Blue rectangle outlines the El Nino 3.4 area from 120° West to 170° West and extending 5° north and south of the Equator. Data from CERES Mar 2000 – Feb 2014.

As you can see, while the El Nino 3.4 Index will give us an idea of future global temperatures, it is by no means the only area which can do so. In fact, the red area in the tropical Atlantic is better correlated with future global temperatures than is the El Nino 3.4 area.

Having seen that, I thought, well, what areas are the best at predicting the temperature of just the Northern Hemisphere two months ahead? Figure 2 shows that result.

Figure 2. Correlation of local temperatures with the northern hemisphere temperature two months later. Blue rectangle outlines the El Nino 3.4 area. Data from CERES Mar 2000 – Feb 2014.

This is most interesting. While the El Nino 3.4 area is barely better than random at predicting the NH temperature, the area in the Atlantic is quite strongly correlated with the future NH temperature.

And the Southern Hemisphere? Figure 3 below has those results:

Figure 3. Correlation of local temperatures with the southern hemisphere temperature two months later. Blue rectangle outlines the El Nino 3.4 area. Data from CERES Mar 2000 – Feb 2014.

I do enjoy surprises. I’d expected somewhere in the Southern Hemisphere to have the greatest correlation with . Instead, it turns out to be exactly centered on the El Nino 3.4 area outlined by the blue rectangle.

Now, it does strike me that we could do a passable job at predicting the global temperatures a couple of months ahead using nothing more than those last two maps for the northern and southern hemispheres. Might make an interesting and possibly profitable project for someone with more time on their hands than I have.

But I’m still leery of subtracting those predictions from the actual data and calling it “removing the El Nino effect” for a couple of reasons. First, “removing the El Nino” doesn’t help us with the Northern Hemisphere temperature. To the contrary, it just adds noise.

But more importantly, the El Nino/La Nina pumping action is not a true cause of changing temperatures in the sense that a volcanic eruption or a change in the amount of top-of-atmosphere total solar insolation is a cause of changing temperatures. Instead, the fluctuations in oceanic temperatures are in turn a result of some previous condition. I term this a “chain of effects”. For example:

Figure 4. Correlation of local temperatures with the El Nino 3.4 area two months later. Blue rectangle outlines the El Nino 3.4 area. Data from CERES Mar 2000 – Feb 2014.

In Figure 4 we can see that high temperatures in the El Nino 3.4 region are in turn presaged by low temperatures off of Australia and the Philippines (blue areas). As I said, the El Nino/La Nina is not a cause. It’s just part of a chain of effects.

And as a result, yes, you can pick out the red area in the tropical Atlantic and “remove its effects” from the temperature … but what is it that you end up with? Or you could “remove” both the tropical Atlantic and the El Nino variation … and “remove” the PDO, and the QBO as well … but what is it that remains once you’ve done that?

As always, my best wishes for everyone on what is a lovely rainy day here …

w.

My Usual Request: If you disagree with me or anyone, please quote the exact words you disagree with. I can defend my own words. I cannot defend someone’s interpretation of my words.

My Other Request: If you think that e.g. I’m using the wrong method on the wrong dataset, please educate me and others by demonstrating the proper use of the right method on the right dataset. Simply claiming I’m wrong doesn’t advance the discussion.