Atmospheric Sciences Opinion

Taking the Pulse of the Planet

How fast is Earth warming? Ocean heat content and sea level rise measurements may provide a more reliable answer than atmospheric measurements.

By , , John Fasullo, John Abraham, , Karina von Schuckmann, and Jiang Zhu

Humans have released carbon dioxide and other greenhouse gases in sufficient quantity to change the composition of the atmosphere (Figure 1). The result is an accumulation of heat in Earth’s system, commonly referred to as global warming. Earth’s climate has responded to this influx of heat through higher temperatures in the atmosphere, land, and ocean. This warming, in turn, has melted ice, raised sea levels, and increased the frequency of extreme weather events: heat waves and heavy rains, for example. The results of these weather events include wildfires and flooding, among other things [Intergovernmental Panel on Climate Change, 2013].

Ocean heat content and atmospheric carbon dioxide concentration measurements.
Fig. 1. Ocean heat content (OHC) and atmospheric carbon dioxide (CO2) concentration measurements since 1958, shown as 12-month running means. The black line represents ocean heating for the upper 2,000 meters of ocean, and light red shading represents the 95% confidence interval. CO2 concentration observed at Mauna Loa Observatory is displayed by light blue. Mean values for 2015–2016 are highlighted with a red star. The OHC is relative to a 1960–2015 baseline. Ocean heat data are from Cheng et al. [2017], and CO2 information is from the National Oceanic and Atmospheric Administration.
Decision-makers, scientists, and the general public are faced with critical questions: How fast is Earth’s system accumulating heat, and how much will it warm in the future as human activities continue to emit greenhouse gases?

Here we explore better ways of measuring global warming to answer these questions. Natural temperature variability is much more muted in the ocean than in the atmosphere, owing to the ocean’s greater ability to absorb heat (its heat capacity). As a result, ocean heating and sea level rise, which are measured independently, show stronger evidence that the planet is warming than does global average surface temperature, which relies on air temperature measurements. In other words, these ocean measurements could provide vital signs for the health of the planet.

Thus, we suggest that scientists and modelers who seek global warming signals should track how much heat the ocean is storing at any given time, termed global ocean heat content (OHC), as well as sea level rise (SLR). Similar to SLR, OHC has a very high signal-to-noise ratio; that is, it clearly shows the effects of climate change distinct from natural variability.

Why Do Changes in Surface Temperatures Obscure Signals from Global Warming?

To determine how fast Earth’s systems are accumulating heat, scientists focus on Earth’s energy imbalance (EEI): the difference between incoming solar radiation and outgoing longwave (thermal) radiation. Increases in the EEI are directly attributable to human activities that increase carbon dioxide and other greenhouse gases in the atmosphere [e.g., Trenberth et al., 2014].

The most visible sign of a warming climate is the increase in air temperature, which affects the climate and weather patterns. Changes in climate and weather affect the viability of plants and animals and our food and water supplies.

Monthly averages of global mean surface temperature (GMST) include natural variability, and they are influenced by the differing heat capacities of the oceans and land masses. Causes of natural variability include forcings that are external to the climate system (e.g., volcanic eruptions and aerosols and the 11-year sunspot cycle) and internal fluctuations (weather phenomena, monsoons, El Niño/La Niña, and decadal cycles).

All of these fluctuations make it difficult to extract the signal from noise in the measurements. But the oceans tell a different story.

What Global Warming Signals Can Be Found in the Oceans?

Scientists have long known that the extra heat trapped by increasing greenhouse gases mainly ends up in the oceans (more than 90%) [Rhein et al., 2013]. Hence, to measure global warming, we have to measure ocean warming.

The oceans present myriad challenges for adequate monitoring. To take the ocean’s temperature, it is necessary to use enough sensors at enough locations and at sufficient depths to track changes throughout the entire ocean. It is essential to have measurements that go back many years and that will continue into the future.

Since 2006, the Argo program of autonomous profiling floats has provided near-global coverage of the upper 2,000 meters of the ocean over all seasons [Riser et al., 2016]. In addition, climate scientists have been able to quantify the ocean temperature changes back to 1960 on the basis of the much sparser historical instrument record [Cheng et al., 2017].

From these temperature measurements, scientists extract OHC. These analyses show that during 2015 and 2016, the heat stored in the upper 2,000 meters of the world ocean reached a new 57-year record high (Figure 1). This heat storage amounts to an increase of 30.4 × 1022 Joules (J) since 1960 [Cheng et al., 2017], equal to a heating rate of 0.33 Watts per square meter (W m−2) averaged over Earth’s entire surface—0.61 W m−2 after 1992. Improved measurements have revised these values upward by 13% compared with the results of the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Rhein et al., 2013]. For comparison, the increase in OHC observed since 1992 in the upper 2,000 meters is about 2,000 times the total net generation of electricity by U.S. utility companies in the past decade [U.S. Energy Information Administration, 2016].

But what about heat capacity over the full ocean depth? The answer requires a bit more calculation. Any increase in heat contributes to the thermal expansion of seawater and, consequently, SLR [Church et al., 2013]. Any energy added in Earth’s system also causes land-based ice to melt, further contributing to SLR by adding water to the ocean.

Studies show that taking the full ocean depth, ice melt, and other factors into account, Earth is estimated to have gained 0.40 ± 0.09 W m−2 since 1960 and 0.72 W m−2 since 1992 [Cheng et al., 2017]—18% higher than for the top 2,000-meter OHC alone.

Human-Caused Warming or Just Natural Variations?

The amplitude of the global warming signature (signal) compared with natural variability (noise) defines how well a metric tracks global warming. The “noise level,” that is, the amplitude of internal variability, approximated here by the standard deviation (σ) of the OHC time series after the linear trend is removed, amounts to 0.77 × 1022 J from 2004 to 2015 (Table 1). The linear trend of OHC is 0.79 ±0.03 × 1022 J year−1 within the same period (Figure 2).


Table 1. The Linear Trend (with 95% Confidence Level) for the Three Key Climate Indicators: Global Mean Surface Temperature (GMST), Ocean Heat Content (OHC), and Sea Level Rise (SLR)a

  Linear Trend σ S/N (1/years) Time (years)
GMST 0.016°C ± 0.005°C/yr 0.110°C/yr 0.14 27
OHC 0.79 ± 0.03 × 1022 J/yr 0.77 × 1022 J/yr 1.03 3.9
SLR 3.38 ± 0.10 mm/yr 3.90 mm/yr 0.87 4.6

aAlso shown are the corresponding noise levels (standard deviation σ), signal-to-noise ratio (S/N), and the time in years required to detect a trend (approximately the time when the linear trend exceeds 4 times the interannual standard deviation). All values are for 2004–2015. Units: yr = year, J = joules, mm = millimeters.

So what time interval is needed to detect a trend given the noise within this time series? Working backward, the signal showing OHC increase, averaged over only 3.9 years, typically exceeds the noise at the 95% confidence level (outside ±2σ error bars; Table 1). Thus, it is relatively straightforward to detect a long-term trend in OHC.

For the GMST record, the trend is 0.016°C ± 0.005°C per year for 2004–2015, and σ of the detrended GMST time series is 0.110°C (Table 1). These values and Figure 2 show that the noise is much larger than the signal. Thus, to detect a warming trend in the GMST record that exceeds a ±2σ noise level, scientists need at least 27 years of data.

Changes in OHC, GMST, and SLR in the past decade.
Fig. 2. Changes in OHC, global mean surface temperature (GMST), and sea level rise (SLR) during the past decade. All values are 2-month means; the dashed red lines indicate linear trends. The scale of the y axis is adjusted so that the linear trend has exactly the same slope for all three indices. El Niño events are marked as pale red bars, and the La Niña events are pale blue bars. All time series are referenced to a 2004–2015 mean. The OHC, GMST, and sea level data reported are archived.

Satellite altimetry has provided global observations of rising sea levels since the early 1990s [Cazenave et al., 2014]. The linear trend of global mean SLR from 2004 to 2015 amounts to 3.38 ± 0.10 millimeters per year, and the σ of the detrended global mean is 3.90 millimeters (Table 1). Thus, 4.6 years are sufficient to detect a robust upward trend in SLR: a signal-to-noise ratio approximately 6 times larger than for GMST.

OHC and SLR Are Robust Indicators of Global Warming

A comparison of the changes and fluctuations in the three observational climate indicators (SLR, OHC, and GMST; Figure 2) clearly shows that both OHC and SLR are much better indicators of global warming than GMST. These two measures are related but also sufficiently different and independently measured to both be of interest.

The large fluctuations in GMST and its sensitivity to natural variability mean that using this measurement to argue that global warming is (or is not) happening requires care. An excellent example is the 1998–2013 period, when energy was redistributed within Earth’s system and the rise of GMST slowed [Yan et al., 2016].

By contrast, the OHC and sea level increased steadily during this period, providing clear and convincing evidence that global warming continued.

The Need to Take the Pulse of the Planet

Monitoring the past and current climate helps us better understand climate change and enables future climate projections. We must maintain and extend the existing global climate observing systems [Riser et al., 2016; von Schuckmann et al., 2016] as well as develop improved coupled (ocean-atmosphere) climate assessment and prediction tools to ensure reliable and continuous monitoring for Earth’s energy imbalance, ocean heat content, and sea level rise.

The EEI has implications for the future and should be fundamental in guiding future energy policy and decisions; it is the heartbeat of the planet. Changes in OHC, the dominant measure of EEI, should be a fundamental metric along with SLR.

As we continue to scrutinize the fidelity of specific climate models, it is critical to validate their energetic imbalances as well as their depiction of GMST. The fact that the Coupled Model Intercomparison Project Phase 5 (CMIP5) ensemble mean accurately represents observed global OHC changes [Cheng et al., 2016] is critical for establishing the reliability of climate models for long-term climate change projections.

Consequently, we recommend that both the EEI and OHC be listed as output variables in the CMIP6 models, in addition to SLR and GMST. This vital sign informs societal decisions about adaptation to and mitigation of climate change [Trenberth et al., 2016].


L.C. and J.Z. were supported by XDA11010405. K.E.T. and J.F. were partially sponsored by the U.S. DOE (DE-SC0012711). The National Center for Atmospheric Research (NCAR) is sponsored by the U.S. National Science Foundation.


Cazenave, A., et al. (2014), The rate of sea-level rise, Nat. Clim. Change, 4(5), 358–361,

Cheng, L., et al. (2016), Observed and simulated full-depth ocean heat content changes for 1970–2005, Ocean Sci., 12, 925–935,

Cheng, L., et al. (2017), Improved estimates of ocean heat content from 1960 to 2015, Sci. Adv., 3, e1601545,

Church, J. A., et al. (2013), Sea level change, in Climate Change 2013: The Physical Science Basis, edited by T. F. Stocker et al., Cambridge Univ. Press, Cambridge, U.K.,

Intergovernmental Panel on Climate Change (2013), Climate Change 2013: The Physical Science Basis, edited by T. F. Stocker et al., Cambridge Univ. Press, Cambridge, U.K.,

Rhein, M., et al. (2013), Observations: Ocean, in Climate Change 2013: The Physical Science Basis, edited by T. F. Stocker et al., Cambridge Univ. Press, Cambridge, U.K.,

Riser, S. C., et al. (2016), Fifteen years of ocean observations with the global Argo array, Nat. Clim. Change, 6, 145–153,

Trenberth, K., J. Fasullo, and M. Balmaseda (2014), Earth’s energy imbalance, J. Clim., 27, 3129–3144,

Trenberth, K. E., M. Marquis, and S. Zebiak (2016), The vital need for a climate information system, Nat. Clim. Change, 6, 1057–1059,

U.S. Energy Information Administration (2016), Electric power annual 2015, U.S. Dep. of Energy, Washington, D. C.,

von Schuckmann, K., et al. (2016), An imperative to monitor Earth’s energy imbalance, Nat. Clim. Change, 6, 138–144,

Yan, X.-H., et al. (2016), The global warming hiatus: Slowdown or redistribution?, Earth’s Future, 4, 472–482,

Author Information

Lijing Cheng (@Lijing_Cheng), International Center for Climate and Environment Sciences, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing; Kevin E. Trenberth (email: and John Fasullo (@jfasullo), National Center for Atmospheric Research, Boulder, Colo.; John Abraham, University of St. Thomas, St. Paul, Minn.; Tim P. Boyer, National Centers for Environmental Information, National Oceanic and Atmospheric Administration, Silver Spring, Md.; Karina von Schuckmann, Mercator Océan, Ramonville St Agne, France; and Jiang Zhu, International Center for Climate and Environment Sciences, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing

Citation: Cheng, L., K. E. Trenberth, J. Fasullo, J. Abraham, T. P. Boyer, K. von Schuckmann, and J. Zhu (2017), Taking the pulse of the planet, Eos, 98, Published on 13 September 2017.
© 2017. The authors. CC BY-NC-ND 3.0
  • Mari Mes

    Prof. Walter H. Munk came to similar conclusions years ago of globally increasing seawater temperatures. He based that on a history of deep sea speed of sound measurements he organized. I believe he also was one of the first (if not the first) to recognize that a large part of the sea level rise can be attributed to a global increase in seawater temperature.

  • Hagen Marilia

    Actually it would be part of my question . However, the temperature changing in the atmosphere would increase pressure on the oceans.
    The oceans receded in Southern Atlantic for almost a month August – September and the Hurricane Harvey and Irma , Katia, Jose were created exactly during the period. Somehow, the rising of temperature on the atmosphere must be involved

  • zaharia

    [Please note: the text below cancel and replace my comment of Sept 15.]

    The entire ocean surface is ~360 million sq. kilometers. The upper 200 m thick layer has a volume of ~72 million cubic kilometers. (or 72.10**18 kg.)
    Assuming a 1°C temp. increase since 1960, the heat content of this layer has increased of ~300 ZJ.
    This is a share of more than 85% of the total OHC increase, as computed elsewhere.

    The cumulative ~55 years rough figures that may be related to this amount of “anthropogenic extraheat”… are the following:
    ~1400 GtCO2 (or carbon dioxyde emissions of ~25 GtC02 /y, on average),
    ~4100 PWh or ~15 ZJ of primary energy consumption (over ~half a century, or ~6,2 Gtoe / y, on average), and
    ~ 2750 T$ (2012 value) of world PIB. (Or ~50 000 G$2012 / y, on average.)

    The above figures show that humankind is using the climate system as an amplifier: for each ZJ of primary energy consumption(*), there is an OHC increase of more than 20 ZJ.
    This average figure, (computed over half a century), is certainly below the present value, related to an over 400 ppm CO2 concentration, (rather than the ~345 ppm 1985 average value.)

    (*) 1 ZJ is about ~23,9 Gtoe. (At present rate, 1 ZJ of primary energy is consumed every ~29 months.)

  • drseismo

    This entire discourse is Much Ado About Nothing. Attempting to predict the earth’s temperature by modeling a myriad of complex interactions and processes is ludicrous. The authors’ proposal to add ocean heat content and Earth’s energy imbalance to the equation is the final straw of an already unmanageable problem.

    I suggest a better approach is to assume the earth’s climate system is a black box and to simply focus on analyzing the output, the earth’s temperature. If one cannot analyze the output of a complex system, what is the likelihood that the complex system itself can be modeled? Modeling may be a lot more fun and lead to a lifelong career in climate science, but the amount of progress will probably be like the distance one can travel on a stationary bicycle.

    A simple numerical analysis of the HadCRUT4 time-temperature data indicates a high likelihood of the beginning of an absolute decline in the global mean surface temperature trend line within the next decade. The first derivative of the temperature trend line is positive but has decreased in value every month for the past 20 years. The derivative is likely to become negative in the mid-2020s and increase in negative value well into the 2030s, i.e., the mean global surface temperature will decline.

  • zaharia

    The upper ocean 200 m thick layer has a volume of ~72 million cubic kilometers, (or 72 E18 kg? The entire ocean surface is rather ~360 million sq. kilometers.)
    Accordingly, the HC of this layer has increased of ~300 ZJ, (assuming a 1°C temp increase, since 1960).
    This is a share of more than 85% of the total OHC increase, as computed above. The cumulative ~50 years rough figures that may be related to this amount of “anthropogenic extraheat”… are the following:
    ~1500 GtCO2,
    ~3250 T$ (2012 value) of world PIB (that is ~65 000 G$2012 / y, on average) and,
    ~5000 PWh or 18 ZJ over half a century (that is ~8,5 Gtoe of primary energy consumption / y, on average.)

    It seems to me that humankind is using the climate system as… an amplifier (!)
    The above figures do show that… for each ZJ of primary energy consumption, (about ~23,9 Gtoe used every 29 months), there is an OHC increase of almost 20 ZJ…
    This average figure, (computed over half a century), is certainly below the present value, related to an over 400 ppm CO2 concentration, (rather than the 345 ppm 1985 average value.)

  • Paul Matthews

    For the benefit of the scientific illiterates who don’t seem to be able to do the calculation, here’s how it goes.

    The change shown in the ocean heat graph is about 35 x 10^22 J over 60 years, about 6 x 10^21 J/yr.
    The mass of the oceans is about 1.4 x 10^21 kg, and the specific heat capacity of water is about 4200 J/kg/C.
    So the average temperature rise of the ocean is 6/1.4/4200 which is about 0.001 degrees C per year.

    • serfbaja

      Thanks….I needed that. The article (with only the calculations…no translation) would leave some to believe the temp rise in the ocean was 0.33 degree from 2015 to 2016

  • Paul Matthews

    “How fast is Earth warming?”

    That question unfortunately isn’t answered in this article. But it is a fairly easy calculation to do if you look up the mass of the oceans. It turns out to be slightly less than 0.001 degrees per year.

    I wonder why this truly terrifying number isn’t included in the article?

    • jhvtex

      Probably because it’s bogus, and you are little more than a denialist troll.

    • James Owens

      Put all the numbers together – and take the actual number in the article – Table 1, GMST (global means surface temperature) – then you get about 0.016 degrees C in surface temp per year. That’s 0.16 degrees per decade. But forget the linear trend, the overall heat accumulation, which is best checked with ice melt and ocean heat content, been speeding up.

    • Lijing Cheng

      What matters for humanity is not whether the very deep ocean is warming but whether the upper ocean where we and our food supply live are changing significantly. The rate of warming is equivalent to the upper 200 m of the ocean warming by 0.2C per decade since 1960. Given that this applies to the entire ocean which is about 3.6 million sq kilometers, it is a huge amount of energy. We provide a comparison in our article “the increase in OHC observed since 1992 in the upper 2,000 meters is about 2,000 times the total net generation of electricity by U.S. utility companies in the past decade.” This energy contributes to the observed rise of sea level, and the depth of the ocean does not matter. For Texas and Florida, it still results in flooding!