Tilted Forum Project Discussion Community - View Single Post

raveneye · 05-18-2007, 03:12 AM

I don’t think we need a climate expert to interpret that video – it’s straightforward and there’s plenty of explanation all over the net, including the link I gave. Here’s another from NASA:

http://www.nasa.gov/centers/goddard/...2006_warm.html

That page is a very nice summary.

The OP unfortunately did not make it clear that the "model" referred to has nothing to do with the linked video. The model was from a paper just published to predict the temperature change in the U.S. into the near future. The video is simply a nice way to visualize the known temperature data over the last 100 years or so. Previous versions of that video have been kicking around for at least 10 years, I’ve been showing them to my students for that long.

It’s easy to misinterpret the “5-year-increments” quote from the little caption for the video. What that means is that the data plotted in the video is actually a 5-year running average. So the year 2000 for example represents not just the data from 2000, but rather is the average of the years 1998, 1999, 2000, 2001, and 2002. The year 2001 is the average of 1999, 2000, 2001, 2002, and 2003, etc. Using a running average is a standard statistical procedure to separate a temporal signal from the temporal noise, making any time trend easier to visualize. You can see both the raw numbers and their running average plotted together on the NASA link above, for the global mean temperature.

For that particular video, the baseline used is the mean temperature for the years 1951 to 1980, as stated in the link above (other agencies in different countries use slightly different baselines; the paper that put together the predictive model uses yet a different one as I pointed out).

But the important point understand is that it doesn't matter what the baseline is, because all the baseline does is tell you what zero is. In the video (and graph a in the link), zero is the average temperature between 1951 and 1980. So if you have dark red in the video, that codes to around 2 degrees celsius. So dark red means that that region is 2 degrees celsius above the mean from 1951 to 1980 at that point on the globe and at that time.

We could just as easily have used a baseline from 1880, but all that would do is change the definition of zero. It's better to use a more recent baseline, because we have a lot more datapoints recently all over the globe from which to calculate a baseline than we did in 1880. Or we could have just plotted the raw data instead and forgot about the baseline entirely. But if we were to plot just the raw data, we wouldn’t be able to see at a glance how much the temperature has changed at any particular point on the globe. That’s the whole purpose of the video, to compare the temperature change at various places on the globe and from year to year (with some of the annual noise filtered out).

Superbelt: take a look at the links I provided, and the video. The animation step is one year, not five years.

05-18-2007, 03:12 AM	#16 (permalink)
raveneye Born Against	I don’t think we need a climate expert to interpret that video – it’s straightforward and there’s plenty of explanation all over the net, including the link I gave. Here’s another from NASA: http://www.nasa.gov/centers/goddard/...2006_warm.html That page is a very nice summary. The OP unfortunately did not make it clear that the "model" referred to has nothing to do with the linked video. The model was from a paper just published to predict the temperature change in the U.S. into the near future. The video is simply a nice way to visualize the known temperature data over the last 100 years or so. Previous versions of that video have been kicking around for at least 10 years, I’ve been showing them to my students for that long. It’s easy to misinterpret the “5-year-increments” quote from the little caption for the video. What that means is that the data plotted in the video is actually a 5-year running average. So the year 2000 for example represents not just the data from 2000, but rather is the average of the years 1998, 1999, 2000, 2001, and 2002. The year 2001 is the average of 1999, 2000, 2001, 2002, and 2003, etc. Using a running average is a standard statistical procedure to separate a temporal signal from the temporal noise, making any time trend easier to visualize. You can see both the raw numbers and their running average plotted together on the NASA link above, for the global mean temperature. For that particular video, the baseline used is the mean temperature for the years 1951 to 1980, as stated in the link above (other agencies in different countries use slightly different baselines; the paper that put together the predictive model uses yet a different one as I pointed out). But the important point understand is that it doesn't matter what the baseline is, because all the baseline does is tell you what zero is. In the video (and graph a in the link), zero is the average temperature between 1951 and 1980. So if you have dark red in the video, that codes to around 2 degrees celsius. So dark red means that that region is 2 degrees celsius above the mean from 1951 to 1980 at that point on the globe and at that time. We could just as easily have used a baseline from 1880, but all that would do is change the definition of zero. It's better to use a more recent baseline, because we have a lot more datapoints recently all over the globe from which to calculate a baseline than we did in 1880. Or we could have just plotted the raw data instead and forgot about the baseline entirely. But if we were to plot just the raw data, we wouldn’t be able to see at a glance how much the temperature has changed at any particular point on the globe. That’s the whole purpose of the video, to compare the temperature change at various places on the globe and from year to year (with some of the annual noise filtered out). Superbelt: take a look at the links I provided, and the video. The animation step is one year, not five years.