Some Historic Temperature Graphs

How to read this graph:

This graph compares each month separately to the 20th Century average (for months of that same name only) and ranks them separately. In other words: The month of March is compared only to other Marches. The month of September is compared only to other Septembers. For example: The warmest March, June and October are all colored red. The second warmest January, May and December are all colored orange. The Februaries colored light blue (2009, 2013) are both between the 11th and 15th warmest Februaries inclusive. The December colored yellow (2017) is the third warmest December and is 0.81° C. warmer than the average of all Decembers in the 20th Century. Etc.

Some points to note:

1. The data for this graph comes from the National Oceanic and Atmospheric Administration (NOAA)'s climate at a glance section.

2. This graph gives the Earth's approximate average surface temperature for each month. Needless to say, distilling the world's climate into one number per month is a gross over-simplification. Further, surface temperature is only part of the story. Temperatures in the upper atmosphere and ocean depths contribute to surface anomalies through convection and heat exchange.

3. Temperature anomalies are not necessarily distributed equally around the globe. For example, while summer 2017 was third warmest globally, it was the warmest summer to date in California, which, preceded by a wet winter, contributed to devastating Summer and Fall wildfires and the mudslides which followed. 2018 brought another hot dry Summer and Fall to California with more destructive fires.

4. Historic data may change slightly (one or two hundredths of a degree) month to month, probably due to NOAA refining its models and data-crunching techniques.

5. Every month on the graph since Feb. 2008 has an anomaly (distance from the 20th Century average) above 0.4° C. The last month with a negative anomaly was Dec. 1984. The highest anomaly was 1.23° C. in March 2016.

6. This graph shows a short term rising trend from 2008 to the present with a bump during the 2015-2016 El Niño event. It also shows wide variation from month to month; but the rising trend is still discernible at a glance. 2005 (not shown) was the warmest year to date before this graph. It has since been surpassed by 2010, 2014, 2015 and 2016 respectively as the warmest year to date.

7. All 12 highest ranking months lie together between Sept. 2015 and August 2016, during a monster El Niño event. 11 of the second highest ranking months, ten of the third highest ranking months, ten of the fourth highest ranking months and seven of the fifth highest ranking months all lie within the last 60 months. We have experienced a string of over five years with each month except one (February 2014) having an anomaly of 0.68° C. or greater and also being among the top ten warmest months with that same name.

8. Other models give somewhat different results. For example, the model used at Copernicus Center (ERA-Interim) rates February 2018 as the third warmest February. The NOAA model rates February 2018 as the ninth warmest. In the Copernicus model October 2018 is the fourth warmest October. The NOAA model rates October 2018 as the second warmest. In part, this has to do with the relative weights each model gives the arctic, which is warming at a faster rate than the temperate regions.

Which model is correct? both and neither. Models are simply simplifications of the real world. No model of a complex system is exact (otherwise it wouldn't be a model).

9. There are six months with an anomaly of 1.0° C. or greater. A string of five such months (Dec. 2015 through April 2016) lie within the monster 2015-2016 El Niño (warmer) event. During the Winter of 2018, we experienced a (cooler) La Niña event. The February 2018 anomaly of 0.69° C. is the lowest we have seen since November 2014 which had an anomaly of 0.68° C. The La Niña event has ended, and NOAA forecasts a return to El Niño conditions. We might expect temperatures to begin to heat up again.

10. The last five 12 month periods ending in November have been the five warmest on record. The warmest two, 2016 and 2015 were El Niño years, whereas the and third and fourth warmest, 2017 and 2018, were not. Nine of the ten years were in the 21st Century. The other, 1998, was an El Niño year. (See table below.) El Niño years tend to be warmer than other years.

11. Temperatures appear to have risen about 0.28° C. over the ten year period covered by this graph (difference between average anomaly of 36 months ending November 2018 and average anomaly of 36 months ending November 2008).

Why is this graph important?

1. This is a “you can't see the forest for the trees” graph. Forests are important; but so are individual trees. Indeed, there is no forest without trees.

2. This graph allows you to see the present in detail. (present here meaning the approximate average monthly surface temperature over the latest 120 months or 10 years.)

3. You can get a similar graph from NOAA's Climate at a Glance section; but I like the way this graph presents the data. I think the white spaces and the color coding increase readability.

4. Why ten year? Why not some other period like 7 or 12 years? No particular reason. With a longer period the lines get squished together and are harder to read. Shorter periods show less data.

5. I intend to update this graph around the 20th of each month when NOAA comes out with a new month of global climate data. Deleting lines from the left may lag behind adding lines to the right.

6. How did I get into this? In 2016 my favorite website for news, Common Dreams, used to report the NOAA data month after month as it set new records. They stopped when September 2016 fell to second place. Being curious, I went to the NOAA site and wondered why it wasn't reported as news that the three warmest Septembers on record were in 2015, 2016 and 2014 respectively. After a few months of silence, I decided to start reporting this data myself.

A longer term temperature graph — 30 years

1. This 30 year graph of 12 month overlapping periods measures the temperature anomaly from the 20th Century average. The colors refer to the El Niño condition during the final month of the 12 month period.

2. El Niño / La Niña refer to cyclical changes in atmospheric and oceanic circulation patterns over the tropical Pacific Ocean. El Niño events usually bring warmer global surface temperatures; La Niña events bring cooler surface temperatures. There is a brief explanation of El Niño and La Niña here.

An ONI (Oceanic Niño Index) is computed for each month from conditions in the tropical Pacific Ocean over a three month period. If the ONI index is at least 0.5 (at most -0.5) the month is considered an El Niño (La Niña) month. Five or more consecutive El Niño (La Niña) months are considered an El Niño (La Niña) event. An El Niño event is very strong (strong, moderate, weak) if it contains three El Niño months with index at least 2.0 (1.5, 1.0, 0.5). A La Niña event is strong (moderate, weak) if in contains three La Niña months with index at most -1.5 (-1.0, -0.5). These classifications are somewhat arbitrary.

3. This graph is surprisingly smooth in spite of all the crests and troughs. The largest difference between two adjacents lines is 0.05° C.; and this occurred only twice over the period covered by the graph.

4. Exceptionally high temperature crests have followed the two very strong El Niño events of 1998 and 2016. Surface temperatures have fallen after the end of these El Niño events but remained much higher than before the El Niño event began. This happened to a lesser extent after the moderate El Niño event of 2009-2010.

5. NOAA forecasts the return of El Niño conditions this Fall to last throughout the winter months. If this occurs, temperatures will likely rise.

6. This graph consists of a progression of waves with increasingly high crests and troughs. Crests have increased from a height of 0.42° C. (above the 20th Century average) in 1988 to 1.02° C. in 2016. Troughs increased from a height of 0.24° C. in 1994 to 0.56° C. in 2012. (We may have just hit a trough of height 0.77° C.; but this is not at all certain.) Temperature increases do not appear linear; they appear to be accelerating at a super-linear rate.

7. Climatewise, extrapolation and longterm prediction are a very risky endeavor. However, looking at the 30 year graph above, the latest IPCC predictions seem ridiculously conservative. I would not be surprised to see us blow through the 1.5° C. above preindustrial levels ceiling within a decade.

8. In short: We're in deep doodoo. We better do something quick.

An even longer term temperature graph — 138 years

1. This graph (also from NOAA data) of 36 month overlapping averages (each 36 month period ending in month with same name as latest month for which data is available) from 1880 to the present shows a longer term rising trend, with an increasing rate of warming in recent years.

2. I like 36 month overlapping graphs. I think they smooth the data, but not too much. You can view 12 month or 60 month overlapping graphs at NOAA's Climate at a Glance section too.

3. There have been various pronouncements that global warming stopped in 1998 (or some other recent year) and global cooling began. This graph demonstrates clearly, without the need for any statistical analysis, that no such thing has happened.

4. Data previous to the 1960s is from surface-based measurements and probably not as accurate as data from later years which includes satellite-based measurements.

5. This graph begins about a century after the start of the “industrial revolution.” It details the sharp and increasing rise in global temperatures with the development of technology.

6. Many have wondered what will happen to mankind as the climate continues to warm. Some say we will continue with business as usual; others say we will join the dodos and the dinosaurs. I suspect somewhere in between. Our current “civilization” is certainly unsustainable; but I see no reason why, with the grace of God, we can't continue to live on Earth in much the same way as our ancestors did during most of the Pleistocene, which ended about 11,300 years ago.


Unlike the two graphs above, the graphs below are based on “proxy data.” That is: global temperatures are inferred from things like tree rings, ice cores, the makeup of the shells of sea critters, type of flora and fauna in the fossil record, etc.

Needless to say the graphs below may not be all that accurate; and the further back in time we go, the less accurate they are likely to be. Nevertheless they seem to be the best we have; so let's go with them.

Graph from NOAA article on global warming and climate change.

1. This graph is derived from research by Michael Mann published in 2008.

2. The baseline, 1961 to 1990 average, is approximately 0.12° C. higher than the Twentieth Century average used in the first two graphs above. The scale is Fahrenheit instead of Celsius (1.8° F. = 1.0° C.)

Graph from Temperature for the past 11,300 years, Marcott et al.

1. This graph is from research by Shaun Marcott et al. published in 2013 and represents a probable reconstruction of temperatures over the last 11,300 years (Holocene).

2. During the Holocene, mankind learned to domesticate animals, build cities, develop written languages, cut down forests, burn fossil fuels, fight wars etc.

3. The baseline, 1961 to 1990 average, is the same as previous graph and is approximately 0.12° C. higher than the Twentieth Century average used in the first two graphs above.

4. The gray squiggly lines at the right represent the previous graph superimposed on Marcott's findings.

5. The wide band represents probable uncertainty.

6. This graph shows global temperatures rising 0.5 to 1.0° C. until 7000 years ago and then falling back to a low close to the start of the Holocene. This low is generally called the “Little Ice Age” which lasted from 1600 to 1850 approximately. From 1850 onward temperatures have risen sharply.

7. One problem with this graph is that the right end of the x-axis (labeled 0) is “the present.” Which year is the present? This is important with temperatures rising by perhaps 0.28° C. in the last decade.

8. If we haven't already surpassed the highest Holocene temperatures, we must be getting pretty darn close.

9. It appears from this graph and the first graph presented that we have already surpassed 1.0° C. above pre-industrial times, which makes it very unlikely that we could make the Paris Accord target (no more than 2.0° C. of warming, but striving for 1.5° C.) even if we were trying hard (which we are not).

Graph from NOAA article on global warming and climate change.

1. This graph represents a possible reconstruction of temperatures during the ice ages and the interglacial periods. (Pleistocene)

2. The scale is Fahrenheit instead of Celsius (1.8° F. = 1.0° C.) and the baseline is the average over the past millenium (possibly around 0.2° C. below the 1961 to 1990 average).

3. The Pleistocene was a succession of wide swings in temperature between ice ages and interglacials, possibly as much as 16° C.

4. Since the rise from baseline to the present doesn't appear consistent with the above graphs, this reconstruction is suspect. Since the data come from ice cores, likely they do not reflect global temperatures as a whole.

5. We are not told which year the 0 point on the x-axis represents.

6. Modern man evolved during this period, probable around 250,000 years ago. Around 50,000 years ago he began to develop more advanced technology and migrate throughout Eurasia, Africa and then Australia and even later, the Americas.

7. Since modern man has been able to survive in almost every ecosystem on the planet, it appears very likely that he will be able to survive in at least some future ecosystem.

8. The same can not be said about “civilization.” “Civilization” developed under some very specific climatic conditions that are unlikely to exist on Earth in the near future.

9. I suspect we will surpass the highest interglacial temperatures soon. (Keep in mind that the temperature has probably risen about 0.28° C. or 0.5° F. over the past decade. (See first graph above.)

global temperatures, 65 million years.

1. I don't know where this graph comes from originally. It is all over the internet.

2. 65 million years ago marks the probable demise of the dinosaurs.

3. Temperatures in the Eocene, some 50 million years ago were likely 5 to 20° C. higher than today.

4. There are no known “great extinctions” during the Eocene. Indeed, life seems to have thrived during the warm period in the early Eocene.

5. Atmospheric carbon dioxide is thought to have also been much higher during the Eocene than today.

6. The Eocene might give us some clues as to what might happen on Earth in the near future.

global temperatures, 542 million years.

1. I don't know where this graph comes from either. It is also all over the internet.

2. 542 million years, back to the Cambrian, is a long time. Planet Earth has weathered a lot of changes. May she also weather this one.

3. Looking at these graphs, one should realize how hard it is to predict the future. Science talks in probabilities. There are few certainties in Science.

More discussion of these temperature graphs to be added later