Six weeks of climate data: Southern Polar Region #3, 60° to 90° East

The third Antarctic slice is the first with a major inlet, which means an ice shelf with scientists studying it. This one is called the Amery Ice Shelf. There were 990 readings in this region, which in Antarctica is almost an embarrassment of riches, though less readings than any similar slice in the Arctic. (In the Arctic, the most read slice had about 12,000 readings and the least read 1,500. These numbers, which make Antarctica look sparse, will themselves be dwarfed by the number of readings we will get in the Northern Temperate slices, the most populated areas on earth.)

As you can see, most of the readings are away from the South Pole and near the coastline, though that curved column of dots near the cross in the lower left shows us that there is at least one station very near the South Pole in this region.

All the data from all the seasons tell the same story. The temperatures were relatively stable until the early 1990s when they plummeted. They have not come back up to old level.

I find this data suspicious. A shift of more than 10 degrees Celsius means about 20 degrees Fahrenheit, and that is enormous. I wonder if this difference in the average is related to more stations coming online in much colder parts of the slice, maybe those polar readings only being available recently.

I have a few ideas on how to change the data around to test this hypothesis.

One different way to check the region would be to only count the coastline section, removing the stations from the interior that may be coming online and going offline sporadically, ones that could easily be producing accurate but much colder temperatures that mess with the average when they are added to or removed from the data set.

Another idea would be to go through all stations everywhere and find the ones that were taking readings in almost all the seasons from 1955 to 2010. That counts as 56 years and 224 seasons. It would be easy enough to find out how many seasonal readings each station produced in this time period and re-write my program to have some threshold set for station reliability. If I want to get all data from all stations, I could set the threshold at zero. If I only wanted stations that were on the job 90% of the time, the threshold would be 201. The 75% threshold would be 168.

In any case, for now I accept this data at face value. Region 3 showed a stunning cooling trend in the early 1990s and is now operating at a "new normal" range some 8 to 10 degrees Celsius colder than the normal experienced from 1955 to the early 1990s.

Later today, Region 4 of the Antarctic Circle, a barely covered wilderness that adds to my belief that modifying the measurement method may be necessary, especially here in Antarctica, the least populated land mass in the world.