2 Observations

Although Earth’s climate is currently changing rapidly relative to past climate changes, in most regions where we live changes are slow enough that we do not notice them directly during our daily lives. However, in some regions changes are larger and more obvious than elsewhere. Below we show some global observations from the last 100 years and data from regions that are particularly sensitive to climate change, where the most dramatic effects have occurred. This will not be a comprehensive documentation of existing observations. Additional observations will be discussed throughout the remainder of this book.

a) Atmosphere

Observations show that climate is changing on a global scale. Surface air temperature data, averaged over the whole Earth, indicate warming of 0.7-0.9°C over the last ~100 years (Fig. 1). But that warming was not steady nor smooth. From 1880 to about 1910 there was cooling, followed by warming until about 1940, after which a slight cooling or approximately constant temperatures precede the strong recent warming observed since the 1970s. Moreover, each year’s temperature is somewhat different from the next. Not all of these year-to-year changes are currently understood, but natural variations, which are not caused by humans, play a role. For instance, strong El Niño years such as 1997-98 or 2015 show up as particularly warm years, whereas La Niña years such as 1999-2000 or 2011 show up as cool. El Niño and La Niña, also known as ENSO (El Niño/Southern Oscillation), is climate variability in the tropical Pacific that also impacts many other regions of the Earth.

Figure 1: Global average surface temperature change from 1880 to 2016. Anomalies from the 20th century average are shown. From the National Oceanographic and Atmospheric Administration’s (NOAA) web site noaa.gov. There are a few other groups around the world who calculate global average surface temperatures, e.g. the National Aeronautics and Space Administration’s (NASA) Goddard Institute for Space Science (GISS), the Climate Research Unit (CRU) of the University of East Anglia, or the Berkeley Earth Surface Temperatures (BEST), and all come to very similar results. Errors due to unequal sampling have been estimated to be about 0.2°C during the early part of the record and decrease to 0.1°C during the more recent part. This is a key figure. The linear trend over the full period is 0.07°C/decade, that over the last 100 years is 0.09°C/decade.

Explore Temperature Data

Go to NOAA’s Climate at a Glance website to explore their global temperature data. List the 12 warmest years by clicking on the ANOMALY column.

  • Which is the record warmest year?
  • How many of the 12 warmest years have occurred since the year 2000?

The increase in atmospheric temperatures over the last 100 years has not been uniform. Fig. 2 shows that temperatures over land changed more than over the oceans and more in the Arctic than in the tropics. These patterns are called land-sea contrast and polar amplification and are quite well understood and simulated in climate models as we will see later. The warming is indeed almost global. The only exception is the northern North Atlantic, which has been slightly cooling for reasons we will explore later.

Figure 2: Surface temperature change (linear trend) from 1901-2012 estimated from three different groups of investigators. Each group used a different degree of interpolation / extrapolation from none (top) to extensive (bottom). From IPCC’s AR5 Technical Summary (Stocker et al., 2013; Fig. TS.2, available at ipcc.ch). Significant changes are indicated by small plus symbols.

Analysis of thousands of temperature records into an estimate of global mean temperature change such as Fig. 1 is not trivial. Issues such as changes in station locations, instrumentation and data coverage have to be taken into account. The fact that five different groups analyzing the data using different methods come to the same conclusions suggests that the results are robust. The reliability of the data has been demonstrated and it has been shown that station locations (e.g. urban versus rural) don’t matter. See https://www.skepticalscience.com/surface-temperature-measurements.htm for more discussion on this topic.

The global warming trend has made the probability of warm and extreme warm temperatures more likely and the probability of cold and extreme cold temperatures less likely. E.g. extremely warm (more than 3°C warmer than the 1951 to 1980 average) summer temperatures over Northern Hemisphere land areas had only a 0.1% chance of happening from 1951 to 1980, whereas during the decade from 2005 to 2015 the chance of those extreme warm temperatures to occur was 14.5%. Conversely, relatively cold summer temperatures (less than 0.5°C cooler than the 1951-1980 average) that happened about every third year from 1951 to 1980 only occurred 5% of the time during 2005 to 2015. Winter temperatures have changed similarly but not has much as summer temperatures. See this article and this blog (Fig. 5) for more information and graphs on this topic.

 b) Cryosphere

One of the most sensitive regions to climate change on Earth is the Arctic. Sea ice cover there has decreased dramatically over the past 40 years particularly in late summer (Fig. 3). Arctic sea ice experiences a strong seasonal cycle. In late winter it covers about 15.4×106 km2 and decreases to about 6.4×106 km2 in late summer (Fig. 4). Therefore the relative changes are larger in late summer, with a reduction of about 46 % or 2.9×106 km2 from 1980 to 2015. In winter the reduction was only about 9 % or 1.5×106 km2. The lost area of summer Arctic sea ice is more than 10 times the size of Oregon (255,000 km2) or four times the size of France (~650,000 km2).

Figure 3: Arctic sea ice concentration in percent during the 1980s (left) and 2010s (right) estimated from satellite observations. From https://phys.org.

In the Antarctic sea ice experiences an even larger seasonal cycle but it has changed less over the last 40 years compared to the Arctic. Southern hemisphere sea ice has slightly increased by about 12 % or 0.5×106 km2 in late austral summer and 2.4 % or 2.3×106 km2 in austral winter. Note that year-to-year fluctuations are larger in winter than in summer and that the long term trends in the Arctic are much larger than the short-term fluctuations, whereas in the Antarctic the long-term trends are similar to the short-term fluctuations and thus less statistically significant. A trend is statistically significant if it is larger than the errors.

Figure 4: Changes in sea ice extent in the northern (top) and southern (bottom) hemisphere in March (left) and September (right). From http://nsidc.org.

Taken both polar regions together, Earth has lost about 2.9+1.5-0.5-2.3 = 1.6 million square kilometers of sea ice from 1980 to 2015. Compare that area to that of your favorite state or country.

Explore Sea Ice Changes

Go to the National Snow and Ice Data Center’s (NSIDC) website and click on a few years to see how Arctic sea ice cover has changed over the year. To get a time series for the current month go to their sea ice index site and click on the Monthly Sea Ice Extent Anomaly Graph in the lower right corner.

  • By how much has the sea ice cover decreased since the 1980s? Estimate the decrease both in relative terms (percentage) and in absolute terms (million square kilometers).

An animation is available here.

Mountain glaciers are also sensitive to climate change. Fig. 4 shows an example from Muir Glacier, which has retreated dramatically since 1941. This is typical for most glaciers around the world. In fact, only a small number of glaciers show advances, whereas the vast majority of glaciers melt and retreat from the valleys up into the high mountains.

Figure 4: Muir Glacier in Glacier Bay National Park and Preserve, Alaska. The picture on top is from 1941, that on the bottom from 2004. From nsidc.org.

The World Glacier Monitoring Service (WGMS) has compiled information on hundreds of glaciers world-wide. Fig. 5 shows that since 1980 glaciers have been losing more than 20 m of mass (water equivalent) with an acceleration of loss in recent years.

Figure 5: Cumulative mass balance in mm water equivalent (w.e.) of glaciers. The blue line represents all 130 glaciers, the red line 40 especially well observed ‘reference’ glaciers. From wgms.ch.

Explore Glacier Change

Go to the Glacier Browser and select a glacier of your choice.

  • What do you observe?

Ice sheets are also melting. Observations from the Gravity Recovery and Climate Experiment (GRACE) satellites, which measure very precisely Earth’s gravity field and can detect changes in mass, show that since 2002 the Greenland ice sheet has lost about 3,500 Gt of mass, and the Antarctic ice sheet has lost about 1,500 Gt (Fig. 6). Here is a presentation about Greenland melting.

Figure 6: Changes in the Antarctic (top) and Greenland (bottom) ice sheet mass estimated from satellite gravity measurements. From nasa.gov.

Box 1: Rates of Change

The rate of change of a variable X between two points in time t1 and t2 can be calculated as the difference (denoted by the greek letter delta Δ) of the value of the variable at time t2 minus the value of the variable at time t1 ΔX = X2X1 divided by the difference in time Δt = t2t1

(B1.1)   \begin{equation*} \frac{\Delta X}{\Delta t} = \frac{X_2 - X_1}{t_2-t_1}\ .  \end{equation*}

Thus the units of the rate of change are the units of the variable divided by time. You can determine the rate of change from a timeseries graph such as Fig. 1 or Fig. 6 by selecting two points in time on the horizontal axis and reading the corresponding values of X1 and X2from the vertical axis. Using Fig. 1 as an example our variable will be the temperature anomaly T. Choosing t1 = 1940 and t2 = 2010 we can read off T1 = 0ºC and T2 = 0.7ºC. Thus, ΔT = 0.7ºC, Δt = 70 years and the rate of change ΔTt = 0.01ºC/yr = 0.1ºC/decade.

Obviously, calculating the rate of change in this way the resulting value will depend on the two times picked. The rate of change of a whole set of data points can be calculated by assuming a linear relationship and minimizing the distance of all points from a straight line X = S×t + I, where S is the slope and I the intercept with the vertical axis. This is called linear regression. Simple formulas to calculate S and I can be found here. Linear regressions are commonly used to estimate rates of change. E.g. the straight lines in Fig. 4 in this chapter and Figs. 7-12 in chapter 1 have been calculated using the formulae for linear regression. Often regression lines are also referred to as trend lines.

c) Ocean

Subsurface temperature measurements in the oceans document warming over the last 60 years (Fig. 7). The ocean’s heat content has increased by about 30×1022 Joules during that time. The heat content of the ocean is its temperature T times the heat capacity of water cp = 4.2 J/(gK). Prior to 2005 subsurface temperature measurements were more limited in space and time because they were taken from ships by lowering CTD (conductivity, temperature, depth) instruments on a cable into the ocean. Since 2005 autonomous, free-drifting Argo floats measure temperature, salinity, pressure, and velocity of the upper 2 km of the water column. Currently there are about 4,000 floats out there, which provide much better spatial and temporal coverage than the previous ship-based measurements.

Figure 7: Observational estimates of global ocean heat content (0-2 km) from NOAA. From 1960 to 2005 ocean temperatures were measured mainly using research ships (blue line). Since 2005 autonomous ARGO floats have increased the data density both in space and time (red and black lines). This is a key figure.

The melting of mountain glaciers and ice sheets leads to increased runoff into the ocean, which contributes to sea level rise (Fig. 8). Other causes of sea level rise are warming sea water, which causes expansion and increased runoff from pumping of groundwater out of aquifers. Estimates based on tide gauge records indicate that sea level has risen by about 20 cm from the 1870s to the year 2000 and another 6 cm since. The melting of mountain glaciers and ice sheets contributes about similarly to the current sea level rise, but if current trends continue it is likely that many mountain glaciers will completely disappear and the large ice sheets will contribute more and more to global sea level rise.

Figure 8: Observed sea level rise estimated from satellites (top) and tide gauges (bottom). From nasa.gov. This is a key figure.

d) Carbon Cycle

Carbon dioxide (CO2) has been measured in the atmosphere since 1958 at Mauna Loa Observatory in Hawaii (Fig. 8). At that time concentrations were just below 320 parts per million (ppm). Subsequently they increased to values of just over 400 ppm today. That is a 25 % increase. Overlaid on the long term trend is a seasonal cycle. Growth of the terrestrial biosphere in northern hemisphere spring leads to CO2 drawdown and decay of organic matter such as fallen leaves increases CO2 in the fall. As we will see later, CO2 is an important greenhouse gas, and its increase over the past decades is the main cause of the recent global warming.


Figure 8: Atmospheric CO2 measured at Hawaii’s Mauna Loa Observatory. The measurements were pioneered by Charles Keeling from Scripps Institution of Oceanography in 1958. From noaa.gov. This is a key figure.


  • By how much did global surface air temperatures increase during the last 100 years?
    • How do we know?
    • What was the average rate of change during that time?
    • What was the rate of change for the last 50 years?
    • Where was the warming larger over the oceans or over land, in the tropics or at high latitudes?
  • How much did the upper 2 km ocean heat content change over the last 60 years?
    • How do we know?
    • What was the rate of change?
    • What was the rate of change during the last 20 years?
  • How much did the area covered by Arctic sea ice change during the last 35 years?
    • How do we know?
  • How have the Greenland and Antarctic Ice sheets changed during the last 15 years?
    • How do we know?
  • How much has global mean sea level risen since 1870?
    • How do we know?
    • Calculate the rate of change from 1870 to 2000.
    • Calculate the rate of change from 2000 to the present.
    • Has sea level rise accelerated?
    • Check your answer with this paper.


Stocker, T.F., D. Qin, G.-K. Plattner, L.V. Alexander, S.K. Allen, N.L. Bindoff, F.-M. Bréon, J.A. Church, U. Cubasch, S. Emori, P. Forster, P. Friedlingstein, N. Gillett, J.M. Gregory, D.L. Hartmann, E. Jansen, B. Kirtman, R. Knutti, K. Krishna Kumar, P. Lemke, J. Marotzke, V. Masson-Delmotte, G.A. Meehl, I.I. Mokhov, S. Piao, V. Ramaswamy, D. Randall, M. Rhein, M. Rojas, C. Sabine, D. Shindell, L.D. Talley, D.G. Vaughan and S.-P. Xie, 2013: Technical Summary. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Available at ipcc.ch.

Randall, M. Rhein, M. Rojas, C. Sabine, D. Shindell, L.D. Talley, D.G. Vaughan and S.-P. Xie, 2013: Technical Sum­mary. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assess­ment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Available at ipcc.ch.