Ocean Heat Content

Owing to its large heat capacity, the ocean accumulates the warming derived from human activities; indeed, more than 90% of Earth’s residual heat related to global warming is absorbed by the ocean (IPCC, Cheng et al. 2017). As such, the global ocean heat content record robustly represents the signature of global warming and is impacted less by weather-related noise and climate variability such as El Niño and La Niña events (Cheng et al. 2018). On the other hand, ocean thermal expansion due to ocean temperature change contributes substantially (30%~50%) to the sea level change, which can considerably influence human populations in coastal and island regions and marine ecosystems. Therefore, monitoring the OHC changes and understanding its variation are crucial for climate change.

A large number of studies suggest that uncertainties of OHC calculations stem from the following issues:

(1) Systematic bias in temperature measurements, such as expendable bathythermographs (XBTs) and mechanical bathythermographs (MBTs), (2) Insufficient coverage of in-situ ocean temperature observations, in both horizontal and vertical dimensions. So a mapping (infilling) method is required, (3) Choice of key methodologies, such as the climatology, and (4) Quality control of the in-situ data.

IAP group provides multipule solvers to these problems and aims to provide the most state-of-the-art estimate on historical ocean heat content changes

Climate Data Guide (UCAR) has a new webpage hosting IAP gridded temperature data, cheers!! OCEAN TEMPERATURE ANALYSIS AND HEAT CONTENT ESTIMATE FROM INSTITUTE OF ATMOSPHERIC PHYSICS

 

               Significant Warming Has Already Been Observed in The Ocean              


Mean ocean warming rate (0-2000m) from 1960 ro 2017 based on IAP ocean analyses (in unit of W/m2)

 

   2017 was the warmest year on record for the global ocean

Figure. (upper) Change in global upper-level (0–2000 m) ocean heat content since 1958. Each bar shows the annual mean relative to a 1981–2010 baseline. (lower) Annual mean ocean heat content anomaly in 2017 relative to a 1981–2010 baseline.

Reference: Cheng L. and J. Zhu, 2018: 2017 is the Warmest Year on Record for the Global Ocean. Advances in Atmospheric Science, 35(3), 261-263, Doi:10.1007/s00376-018-8011-z [OHC Data see below: IAP-Gridded dataset and IAP official OHC estimate]

Near real-time update for OHC to 2018Apr-Jun (Record High)

 

      IAP Observational OHCs VS. Radiative imbalance at TOA

Estimate of the ocean energy budget in Cheng et al. (2017). The three major volcano eruptions are marked. The energy budgets are relative to 1958-1962 base period. The integrated net radiative imbalance from Allan et al. 2014 estimated from the top of atmosphere (TOA) is included in yellow multiplied by 0.93 to be comparable with the ocean energy budget. The TOA radiation is adjusted to the value of OHC within 2013-2014. The dashed grey lines encompass the 95% confidence interval. [Download high-resolution figure]

DATA:Ocean energy budget (based on OHC) [data shown in Cheng et al. (2017) ]

Reference: Cheng L. *, K. Trenberth, J. Fasullo, T. Boyer, J. Abraham, J. Zhu, 2017: Improved estimates of ocean heat content from 1960 to 2015, Science Advances. 3,e1601545c.

              

            Observational OHCs VS. CMIP5 model simulations

OHC trends during the 1970–2005 period in observations and CMIP5 models. (a) 0–700 m. (b) Full depth. For models, the histograms are the distribution of CMIP5 results, and the median of the CMIP5 multimodel results is shown in solid line, with the 5–95% confidence interval in dashed lines. For observations, we present the linear trends by different studies: this study (both CZ14 and the flexiblegrid method), five estimates in Durack et al. (2014) after adjustment, and five ensembles of ORAS4 reanalysis. The 5–95% confidence intervals for observations are shaded in light green.

DATA: OHC in 26 CMIP5 models (quadratic or linear climate-drfit correction)

DATA: OHC time series by CZ14 gaps-filling (This is NOT our recommended method, our official OHC time series see below)

Reference: Cheng L. *, K. Trenberth, M. D. Palmer, J. Zhu, J. Abraham, 2016: Observed and simulated full-depth ocean heat content changes for 1970-2005, Ocean Science. doi:10.5194/os-2016-16.

 

       IAP Gridded ocean subsurface temperature dataset

              (1 by 1 degree, from 1m to 2000m, 1940-2018 monthly)

AND   IAP OHC, Steric Sea Level estimate (1940-2018)

 

IAP estimate of Global 0-2000m OHC change since 1940, with 95% Confidence Interval shown in red shading  [Download high-resolution figure]

DATA: IAP Gridded temperature dataset (version-3, 1940-2018, 1 by 1 degree, monthly, 0-2000m), based on EnOI-DE/CMIP5 gaps-filling method [Latest version, updated at Aug 2018]

DATA: Monthly/Global mean OHC time series of 0-700m/700-2000m [Latest version, updated at Aug 2018]

DATA: Monthly/Basin mean OHC time series of 0-700m/700-2000m (Pacific, Atlantic, Indian, Southern oceans, Arctic)

DATA:Gridded OHC 0-100m/300m/700m/1500m/2000m (2-D)[Link]

DATA:Gridded OHC anomaly related to 2006-2015 climatology after removing seasonal cycle, 0-100m/300m/700m/1500m/2000m (2-D)[Link]

DATA:Gridded Steric Sea Level Anomaly data based on IAP0-2000m temperature change (2-D)[Link]

An alternative link, for all IAP gridded data and time series (and also real-time update) [Link]

 

An animation: An animation of historical (1940-2017) 0-2000m averaged temperature change based on IAP_Gridded dataset [DOWNLOAD]

Reference 1: Cheng L. and J. Zhu*, 2016, Benefits of CMIP5 multimodel ensemble in reconstructing historical ocean subsurface temperature variation, Journal of Climate, 29(15), 5393–5416, doi: 10.1175/JCLI-D-15-0730.1. [OA]

Reference 2: Cheng L. *, K. Trenberth, J. Fasullo, T. Boyer, J. Abraham, J. Zhu, 2017: Improved estimates of ocean heat content from 1960 to 2015, Science Advances. 3, e1601545.

High resolution figures in Cheng et al. 2017, Sci. Adv. can be downloaded here.

 

Ocean subsurface in-situ temperature profile dataset (IGOT)

DATA: IGOT (IAP Global Ocean Temperature) dataset (1940-to-2017 in situ ocean subsurface temperature profiles dataset, QC-ed and Bias-corrected. All instruments are included in this dataset: XBT, CTD, Bottle, Argo, MBT, OSD, Mooring, Glider etc. ) [Latest update: Jan. 2018]

An alternative link, for IGOT profile dataset, check for the real-time update[Latest version]

DATA: Annual Mean time series of 0-700m OHC from 1965 to 2015 [Latest update: July.2016] (Note: this is not our recommended OHC estimate, IAP official OHC time series see above)

Reference: Cheng L. *, J. Zhu, and J. Abraham, 2015: Global upper ocean heat content estimation: recent progress and the remaining challenges. Atmospheric and Oceanic Science Letters, 8, 6, 333-338. DOI:10.3878/AOSL20150031.[OA]

 

                                    How OHC is calculated?

Figure. Schematic illustration of how annual mean ocean heat content is calculated from the ocean temperature profile observations (raw data).

C1: IAP-OHC uses WOD2013 QC flags (more guide for QC can be found in IQuOD project)

C2/C3:see below.

C4: 1° grid

C5: Mapping method. IAP provides two options: (1). in Cheng and Zhu, 2014 (simple adjustment); (2). in Cheng and Zhu, 2016 (objective analysis, recommended!).

Reference: Cheng L. *, J. Zhu, and J. Abraham, 2015: Global upper ocean heat content estimation: recent progress and the remaining challenges. Atmospheric and Oceanic Science Letters, 8, 6, 333-338. DOI:10.3878/AOSL20150031.[OA]


                              XBT Data and XBT error

Expendable bathythermograph (XBT) data were the major component of the ocean temperatureprofile observations from the late 1960s through the early 2000s, and XBTs still continue to providecritical data to monitor surface and subsurface currents, meridional heat transport, and ocean heat content. Systematic errors have been identified in the XBT data, some of which originate from computingthe depth in the profile using a theoretically and experimentally derived fall-rate equation(FRE). After in-depth studies of these biases and discussions held in several workshops dedicated todiscussing XBT biases, the XBT science community met at the Fourth XBT Science Workshop andconcluded that XBT biases consist of 1) errors in depth values due to the inadequacy of the probemotion description done by standard FRE and 2) independent pure temperature biases. The deptherror and temperature bias are temperature dependent and may depend on the data acquisition andrecording system. In addition, the depth bias also includes an offset term. Some biases affecting theXBT-derived temperature profiles vary with manufacturer/probe type and have been shown to be time dependent. (quote in Cheng et al. 2016 [BAMS]):

Now there is a community-recommended XBT correction scheme (Cheng et al. 2014 [JAOT], CH14 method), which takes accounts of all of the influencing factors. And this method is carefully evaluated and shown to be able to significantly reduce the XBT bias [Figure]. We applied this  method to historical XBT data and then used this data to calculate IAP OHC change.

Figure. Temperature difference between XBT and CTD/OSD/PFL profiles before (left side) and after (right side) bias corrections. The data are from WOD2013 dataset(NOAA/NODC)

Here we provide our bias-corrected XBT dataset, CH14 correction scheme and a Matlab code showing how CH14 is implemented.

CH14 correction has been recommended by XBT community as the BEST correction scheme to XBT data (Cheng et al. 2016, BAMS)

DATA: 1966-2017 XBT profiles dataset (QC-ed and Bias-corrected) [Latest update: Jan. 2018]

An alternative link, for XBT dataset after CH14 correction, check for the real-time update [Latest version]

DATA: CH14 correction scheme for XBT data, Table-1, Table-2 [Latest update: Oct. 2016, an introduction on the update]

CODE: CH14 correction (Matlab codes), [Latest update: Oct. 2016] 

Link: NOAA/NCEI (U.S.) provides CH14-corrected WOD data (World Ocean Database)

Link: Met Office (U.K.) will provide CH14-corrected EN data

Reference: Cheng L., J. Zhu*, R. Cowley, T. P. Boyer and S. Wijffels, 2014: Time, probe type and temperature variable bias corrections on historical expendable bathythermograph observations. Journal of Atmospheric and Oceanic Technology, 31(8), 1793-1825.

Reference: Cheng L.*, John Abraham, Gustavo Goni, Timothy Boyer, Susan Wijffels, Rebecca Cowley, Viktor Gouretski, Franco Reseghetti, Shoichi Kizu, Shenfu Dong, Francis Bringas, Marlos Goes, Loïc Houpert, Janet Sprintall, Jiang Zhu, 2016: XBT Science: assessment of instrumental biases and errors, Bulletin of the American Meteorological Society, 97, 924-933. doi: http://dx.doi.org/10.1175/BAMS-D-15-00031.

 OHC time series applying different XBT correction schemes                             (Cheng et al. 2016, BAMS)

XBT bias is one of the major source of error in OHC estimate, Cheng et al. 2016, BAMS provided an illustration for the impact of XBT bias on OHC estimate. In the following figure, we applied ten of the existing XBT correction schemes to calculate OHC. The mapping method is the simple-gridded average (potentially assuming that the OHC in the data-gaps is filled by the mean of OHCs in the sampled regions). By comparison, the dashed curve is without applying XBT correction.

Figure. Upper (0–700 m) OHC calculated using corrected XBT data and the uncorrected XBT data (Uncor; black curve). The XBT data are corrected using 10 of the schemes, including Cheng et al. (2014; CH14), Wijffels et al. (2008; W08), Ishii and Kimoto (2009; IK09), Good (2011; GD11), Hamon et al. (2012; H12), Cheng et al. (2011; CH) method in Cowley et al. (2013; CWCH), Cowley et al. (2013; CW13), Gouretski and Reseghetti (2010; GR10), Levitus et al. (2009; L09), and Gouretski (2012; G12). The annual mean of global OHC anomaly (OHCA) is calculated by simply averaging the 1° × 1° grid means of OHCA over the global ocean. [Update of Fig.3 in Cheng et al. BAMS paper]

DATA: OHC time series applying different XBT correction schemes (Mapping: simple gridded average) [Update of Fig. 3 in Cheng et al. BAMS paper] [Latest update: Oct. 2016]

Reference: Cheng L.*, John Abraham, Gustavo Goni, Timothy Boyer, Susan Wijffels, Rebecca Cowley, Viktor Gouretski, Franco Reseghetti, Shoichi Kizu, Shenfu Dong, Francis Bringas, Marlos Goes, Loïc Houpert, Janet Sprintall, Jiang Zhu, 2016: XBT Science: assessment of instrumental biases and errors, Bulletin of the American Meteorological Society, 97, 924-933. doi: http://dx.doi.org/10.1175/BAMS-D-15-00031.

                                                    Climatology

One need to remove a monthly climatology from each temperature profile, which results in temperature anomaly profile. There are several different temperature climatologies constructed by different data centers, such as the World Ocean Atlas (WOA) (Locarnini et al., 2010) and Argo climatology (Roemmich and Gilson, 2009). The irregular data coverage in historical period could lead to bias in OHC calculation tied to climatology. Cheng and Zhu (2015) proposed one of the solutions, to use a climatology constructed by data with a period with consistent global coverage (i.e. Argo era).

In our analysis, 12 monthly climatologies are constructed by using all of the observations from 2008 to 2012. All of the temperature profiles in each month within 2008–2012 are interpolated to standard vertical levels and then averaged to 1° × 1° grids. The 5° × 5° median filter is then applied to the obtained field to smooth the climatologies.

Figure. August of 2008-2012 Climatology, in values of 0-700m averaged temperature.

Here we provide our 2008-2012 climatology (in MATLAB ".mat" format)

DATA: 2008-2012 Climatology [Latest update: Oct. 2016]

Reference: Cheng L. and J. Zhu*, 2015: Influences of the choice of climatology on ocean heat content estimation, Journal of Atmospheric and Oceanic Technology. 32(2), 388-394, doi: http://dx.doi.org/10.1175/JTECH-D-14-00169.1


                                           Vertical resolution

It has been shown that the typical vertical resolution of historical temperature profiles is insufficient (Cheng and Zhu, 2014a), with a global mean of 10–20 m for the upper 100 m and 50–100 m for 300–700 m prior to 2000. The insufficiency of the vertical resolution of temperature data leads to a systematic bias in OHC calculation. This error is calculated by sampling a high-vertical-resolution climatological ocean according to the depth intervals of in-situ subsurface temperature observations, and then the difference between the OHC calculated by subsampled profiles and the OHC of the climatological ocean is defined as the resolution-induced error.

Figure. Annual mean vertical resolution of historical subsurface temperature profiles. a). All observations. b) 7 main instruments: left side is the data amount, right side is the annual mean resolution.

Reference: Cheng L., and J. Zhu*, 2014: Uncertainties of the ocean heat content estimation induced by insufficient vertical resolution of historical ocean subsurface observations, Journal of Atmospheric and Oceanic Technology, 31(6), 1383-1396.

2001-2003 OHC shift: an artifact of the observation system evolution

There is an OHC shift during 2001-2003 in most of the OHC estimates, which is due to the observation system change from Ship-based system (Figure b) pre-2001 to Argo-based system (Figure c).

The key messege is: to be careful in using OHC values within 2000-2004 period, considering the observation system change.

Figure. a). OHC shift during 2001-2003 as identified in NODC-OHC estimate. b) Ship-based observation system. c) Argo system. d) Argo-Ship Area.

Reference: Cheng L., and J. Zhu*, 2014: Artifacts in variations of ocean heat content induced by the observation system changes, Geophysical Research Letters, 41(20), 7276-7283, DOI: 10.1002/2014GL061881. [OA]

                                        Gaps-filling (Mapping)

There are regions of ocean grids without any data. Therefore, it is impossible to calculate the global integration, which requires full data coverage. To overcome this data paucity, “mapping” strategies have been proposed to fill those data gaps.

Our group proposed two methods: (1) in Cheng and Zhu, 2014, which is a simple/transparent method. (2) in Cheng and Zhu, 2016, which is Ensemble Optimum Interpolation method with Dynamic Ensemble of CMIP5 simulations (EnOI-DE/CMIP5) method. This method is then updated in Cheng et al, 2016, Science Advances. Method (2) is recommended method! 


                                                    Useful Links

NOAA's offical estimate on historical OHC changes, Steric Sea Levl change etc.

World Ocean Database

XBT Science Team website

International Quality-Controlled Ocean Database (IQuoD) website

Guide for XBT corrections in World Ocean Dataset 2013 (WOD13)

Historical side-by-side XBT/CTD comparison dataset

Inter-comparison among 10 existing XBT correction schemes

Up to 10 correction schemes have been proposed by different international groups since 2008. Each made different assumptions on the sources of bias, used different subsets of global XBT dataset and different methodologies to detect the bias.

This project quantifies how well 10 different available schemes can correct the historical XBT data by comparing the corrected XBT data with co-located reference data in both the World Ocean Dataset (WOD) 2013 and the EN4 dataset. Four different metrics are proposed to quantify their performances.

The results indicate CH14 is the best among the currently available methods, and L09/G12/GR10 can be used with some caveats. To test the robustness of the schemes, we further train the CH14 and L09 by using 50% of the XBT/reference data and test the schemes by using the remaining data. Results indicate that the two schemes are robust. Moreover, EN4 and WOD comparison dataset shows a systematic difference of XBT error (~0.01°C on global and 0-700m average). Influences of quality control and data processing have been investigated.

CH14 and L09 result in very similar ocean heat content change (OHC) estimates in the upper 700m since 1966, suggesting a potential of reducing XBT-induced error in OHC estimate.

References: Cheng L., H. Luo, T. Boyer, R. Cowley, J. Abraham, V. Gouretski, F. Reseghetti, J. Zhu, 2018: How well can we correct systematic errors in historical XBT data? Journal of Atmospheric and Oceanic Technology, accepted, 21 March 2018.

DATA: (1). WOD13-XBT/Reference comparison dataset. (2). EN4-XBT/Reference comparison dataset.

Ongoing project: (02) Using the Synthetic data to understand the existing mapping methods

Mapping methods have been shown to be one of the major error sources in OHC calculation. Understanding the behavior of recent existing mapping strategies is the first step to solve this problem. Synthetic data for OHC will be used to gain knowledge about the existing mapping methods.

                                   Media coverage




                                                        Publications
  • L. Cheng, J, Zhu., 2018: 2017 was the warmest year on record for the global ocean. Adv. Atmos. Sci., 35, 261-263, doi: 10.1007/s00376-018-8011-z. [LINK]

    Cheng L., K. E. Trenberth, J. Fasullo, J. Abraham, T. P. Boyer, K. von Schuckmann, and J. Zhu 2018: Taking the pulse of the planet, Earth and Space Science News, Eos, 99, 14-16. doi: 10.1029/2017EO081839. [LINK]

    Wang G., L. Cheng*, J. Abraham, C. Li, 2017: Consensuses and discrepancies ofbasin-scale ocean heat content changes in different ocean analyses. ClimateDynamic, doi: 10.1007/s00382-017-3751-5. [LINK]

    Wang, G., L. Cheng*, T. Boyer, C. Li, 2017: Halosteric Sea Level Changes during theArgo Era. Water, 9, 484. [LINK]

    Cheng L., K. Trenberth, J. Fasullo, T. Boyer, J. Abraham, J. Zhu, 2017: Improved estimates of ocean heat content from 1960 to 2015, Science Advances, 3, e1601545. [LINK]

    Cheng L., J. Abraham, G. Goni, T. Boyer, S. Wijffels, R. Cowley, V. Gouretski, F. Reseghetti, S. Kizu, S. Dong, F. Bringas, M. Goes, L. Houpert, J. Sprintall, and J. Zhu, 2016: XBT Science: assessment of instrumental biases and errors, Bulletin of the American Meteorological Society, 97, 924-933. [LINK]

    Cheng L., K. Trenberth, M. D. Palmer, J. Zhu, J. Abraham, 2016: Observed and simulated full-depth ocean heat content changes for 1970-2005, Ocean Science. doi:10.5194/os-2016-16. [LINK]

    Cheng L. and J. Zhu, 2016, Benefits of CMIP5 multimodel ensemble in reconstructing historical ocean subsurface temperature variation, Journal of Climate, 29(15), 5393–5416. doi: 10.1175/JCLI-D-15-0730.1. [LINK]

    Cheng L., J. Zhu, and J. Abraham, 2015: Global upper ocean heat content estimation: recent progress and the remaining challenges. Atmospheric and Oceanic Science Letters, 8. DOI:10.3878/AOSL20150031. (LINK)(LINK)

    Cheng, L., and J. Zhu, 2015: Influences of the choice of climatology on ocean heat content estimation, J. Atmos. Ocean Tech., 32, 388–394. (LINK)(LINK)

    Cheng, L., and J. Zhu, 2014a: Uncertainties of the ocean heat content estimation induced by insufficient vertical resolution of historical ocean subsurface observations, J. Atmos. Ocean Tech., 31, 1383–1396. (LINK)(LINK)

    Cheng, L., and J. Zhu, 2014b: Artifacts in variations of ocean heat content induced by the observation system changes, Geophys. Res. Lett., 20, 7276–7283. (LINK)

    Cheng, L., Zhu, R. Cowley, T. P. Boyer, S. Wijffels, 2014: Time, probe type and temperature variable bias corrections to historical expendable bathythermograph observations, J. Atmos. Ocean Tech., 31, 1793–1825. (LINK)(LINK)

  • Abraham, J. P., F. Reseghetti, M. Baringer, T. Boyer, L. Cheng, J. Church, C. Domingues, J. T. Fasullo, J. Gilson, G. Goni, S. Good, J. M. Gorman, V. Gouretski, M. Ishii, G. C. Johnson, S. Kizu, J. Lyman, A. MacDonald, W. J. Minkowycz, S. E. Moffitt, M. Palmer, A. Piola, K. E. Trenberth, I. Velicogna, S. Wijffels and J. Willis, 2013: A review of global ocean temperature observations: Implications for ocean heat content estimates and climate change, Reviews of Geophysics, 51, 450-483.(LINK)
  • Cowley, R., S. Wijffels, L. Cheng, T. Boyer, S. Kizu, 2013: Biases in Expendable Bathythermograph Data: A New View Based on Historical Side-by-Side Comparisons, Journal of Atmospheric and Oceanic Technology, 30(6), 1195-1225. (LINK)

    L. Cheng's publications can be downloaded here [link]

Contact
Lijing Cheng
Jiang Zhu 

    from IAP/CAS


Web page provided by IAP, L.C.