Images of the 1998 El Niño in world temperature maps

A reader asked for a time representation… a tricky problem via a web site, the data is large…

Here are three different views, two from the UK Met Office and one from UAH, the newly released version. These run from January 1996 through December 2000.

Firstly here is an overview an Hovmoller plot for the same time span. These are widely used but rarely for earth whole views. Here are some I prepared earlier as a 2012 article (includes [2]) elsewhere which might help understanding on what I am doing for this article.

Figure 1, Hovmoller diagram of 1998 El Nino, UAH TLT V5.6 [1] (as PDF 142kB)

The 1998 El Niño event was similar to the less known ~1876 event, both seeming increasingly muted in the Met Office data. The El Niño form is a pulse of warmth primarily in tropical regions followed by a cool period. The warmth gradually disappearing apparently flowing poleward, producing a characteristic sideways V shape in Hovmoller graphics. The cool in inside the V. Perhaps preceeded by coolish. This pattern has repeated over the years.

Unfold

If instead of folding all longitudes into a mean as above a plain XY map for a month gives a different view, so we may be able to see from where on the globe there are contributions.

Time series

Figure 2, time series global temperature plots for the map series

All the satellite based show much the same.

Time maps

For this article I’ve written code to output moderate sized maps for each month which are then size minimised and combined in a PDF. A local PDF reader is a capable viewing tool where the user controls movement. This is intended to be set to zoom 100% and discrete pages, one at a time so that images page. (forcing this via the PDF has so far eluded me and may be overriden anyway by user default preferences)

• Hadcrut 3 (1MB)
• Hadcrut 4 (1MB)
• UAH TLT V6 (2MB larger: 2.5 degree gridding, little missing data)

I’ve included Hadcrut because these are different animals, underwater temperature and near surface air temperature, a perplexing combination of unlike data. Soil temperature for example would be a more reasonable match.

The oxygen sensing satellite series are indirectly calibrated to surface temperature by radiosonde.

Neither are correct, Nyquist and Shannon put paid to that let alone the metrology matters. It is all we have.

As Twain did pith:

There is something fascinating about science.

One gets such wholesale returns of conjecture

out of such a trifling investment of fact.

Mark Twain.

Life on the Mississippi [a ]

Some imagine that satellite sensing is somehow good, it isn’t, simply because more than an order more concurrently active satellites are necessary. In the case here it takes several days to track the earth surface yet weather moves in minutes. Is it any wonder monthly shows rapid fire change, mostly mathematical artefacts?

An improvement would be publishing a data every week, better still a rolling data updated every day. That’s still a sticking plaster.

Ground stations don’t usually move, sea stations don’t exist. The sites are inconsistent and also over time.

 

---

 

1. Unfortunately I have broken the code producing Hovmoller plots; an external plot program objects to something. Still works for UAH TLT V5.6, similar to V6, with the reason almost certainly about missing data, 5.6 has no polar holes, V6 does. No quick fix, I tried.

Also possible the plot software has changed, different version, different operating system.

2.

Relationship to ENSO etc.