Introduction
NASA’s Goddard Institute of Space Studies (GISS) recent 2008 summary of global temperature trends is an excellent example of how effective charts and graphs can help communicate information on complex topics like climate change.
The report includes 7 charts, 4 trend charts and 3 color coded global images. This post looks at several of them to see how GISS uses charts to communicate climate trends.
Trend Chart and Global Image
Figure 1: NASA GISS - 2008 Annual Summation
(link for pdf)
These 2 charts tell a powerful story when shown together. The trend chart shows that global temperature has risen in fits and starts since 1880. The global image shows that temperature anomalies in 2008 vary widely by area:
- Negative Anomaly ( -0.2 to - 3.5C): The lower sea surface temperatures from the 2008 La Nina are clearly visible in the eastern Pacific Ocean. Notice the negative anomalies in the upper Midwest USA and portions of eastern South America.
- No Anomaly: Significant portions of USA, western Canada, Atlantic Ocean and southern Pacific had no temperature anomaly in 2008.
- Positive Anomaly(0.2 to 3.5C): Much of South America, Europe, Asia and Africa had increased temperatures in 2008.
Together, the trend and image charts show the temporal and spatial variability in global climate trends.
Panel Chart with 12 Variables in 2D
Figure 2: NASA GISS - 2008 Temperature Summary
This panel (aka lattice, trellis) chart shows 12 variables in 2 dimensions. The variables are:
- Year
- El Nino – La Nina Index, Nino 3.4 SST (C)
- Volcanoes ( heavy black markers)
- Major volcanoes (green icon with plume)
- Lower latitude (23.6 N to 23.6 S) anomaly by 4 seasons
- Global anomaly by 4 seasons
This chart tells also tells a powerful story about global climate:
- Global anomalies are greater than lower latitude anomalies
- Positive temperature anomalies and El Nino events (red areas in lowest panel) seem to co-occur
- Negative temperature anomalies and La Nina events (blue area in lowest panel) seem to co-occur
- While volcanoes seem to decrease temperature anomalies, the interplay of El Nino – La Nina and volcanoes events determines the actual impact of both
Side by Side Global Images
Figure 3: NASA GISS - 2008 Annual Summation
It has taken me a while to become comfortable interpreting these global temperature anomaly images, maybe because I have spent to much time looking at trend charts. As I have become more familiar with them, I realize that they convey a great deal of information in a compact graphic.
Spend a little time comparing the 2008 image with the 2001 – 2007 Mean image and you can see several important facts that help to understand recent climate conditions.
- 2008′s La Nina stands out when you compare the 2008 eastern Pacific – west coast USA area conditions with the 2001 – 2007 conditions.
- Impact of 2008 La Nina is clear when you look at USA -western Canada conditions in 2008 versus 2001-2007.
- While the 2008 La Nina lowered 2008 temperatures compared to 2001-2007 for most regions , look what happened in Russia, record warm conditions.
Integration of Charts & Graphs in Report
NASA GISS’s 2008 Annual Summation provides a great example of how charts and graphs can be incorporated into a technical document to help tell the writer’s story. The same document without the charts and graphs would not be as useful; the charts without the accompanying text would not be as useful.
What do you think?
Positive Anomaly in india is true, summer already started here in Jan. Which is very unsual with hot summers, temperatures are surly increased by more than 3.5 c for Feb first week.
Their color scale for the maps suffers from being a bit rainbow-ish, and could benefit by being replaced with the sorts of scales discussed by Rogowitz and Treinish.
(a discussion of their work is on Juan Dusrsteler’s infovis.net, along with links to pdfs of their articles).
Briefly, they find that it’s better to use a much smaller range of hues, together with a scale of decreasing luminosity and saturation, to avoid the spurious appearance of step changes caused purely by our perception of color.
These are great. I think the last set of charts is a bit misleading, though, in the comparison. Means over several years will have lower variance than single years. The more extreme outliers in single years get washed out in the averages so we don’t know, from looking at these charts, if 2008 was really exceptional or not.