A group of oceanographic scientists are in the process of publishing results indicating that the ocean has stripes. They used successive levels of data filtering to remove large scale variations like currents and waves and temperature gradients, and they were left with these bands of very small variation in temperature and velocity.
An interesting observation here is that they repeated their approach on the output from a Japanese computer running the “Earth Simulator” and found very similar stripes in the output from that model.
Maybe I’m just overly skeptical, but it seems to me that if you run a bunch of data filtering and end up with a bunch of surprisingly fine-scale features you have to wonder if it’s an artifact of the analysis. And the fact that they found the same features in the model output, when presumably the model wasn’t programmed to have this behavior, seems to indicate that the stripes are coming from the data filtering and are not real phenomena. Otherwise that seems to say that the Earth Simulator model is so accurate that it was predicting previously unimagined phenomena in the ocean, which would be pretty remarkable in itself.
For the record: I’m not an expert in oceanography, I’ve just got some experience modeling physical systems. They may be right, and there may be tons of stripes overlaid on the ocean. I’m sure their first thought was that it was an artifact, and it must have checked out after reviewing the analysis. It just seems a little fishy.
I came across this in summary for at Ars.