Copernicus Report: 2019 was the second warmest year and the last five years were the warmest on record
Data released by the Copernicus Climate Change Service (C3S) show that 2019 was the second warmest year in a series of exceptionally warm years across the globe, as CO2 concentrations continue to rise.
The Copernicus Climate Change Service (C3S) announces today that 2019 was the fifth in a series of exceptionally warm years and the second warmest year globally ever recorded.
Meanwhile, Europe saw its warmest year on record by a small margin. Together with the Copernicus Atmosphere Monitoring Service (CAMS), C3S also reports that CO2 concentrations in the atmosphere have continued to rise. Their data provide the first complete, global picture of 2019 temperatures and CO2 levels.
The results are in line with previous projections from WMO and the Global Carbon Project (GCP) for 2019. The WMO estimated that 2019 was likely to be the 2nd or 3rd warmest year on record, while both WMO and the GCP indicated that atmospheric CO2 concentrations had continued to increase.
C3S and CAMS are both implemented by the European Centre for Medium-Range Weather Forecasts on behalf of the European Union. The services provide quality-assured data on 2019 temperatures and CO2 concentrations, among many other climate variables. This helps policy makers, organisations, and individuals make informed choices about climate change mitigation and the quality of the air we breathe.
The temperature dataset provided by C3S shows that the global average surface air temperature was 0.04 °C lower than in 2016, the warmest year on record.
The data also show that:
The five warmest years on record have all occurred in the last 5 years, with 2019 coming in as the second warmest and 2010-2019 being the warmest decade on record.
- 2019 was almost 0.6 °C warmer than the 1981-2010 average
- The average temperature of the last 5 years was between 1.1 and 1.2 °C higher than the pre-industrial level defined by the IPCC
- Europe saw its warmest calendar year on record, marginally ahead of 2014, 2015 and 2018
Furthermore, according to satellite measurements of global atmospheric CO2 concentrations:
- CO2 continued to rise in 2019, increasing by 2.3 ± 0.8 ppm
The most pronounced warming compared to the 1981-2010 average occurred in Alaska and over other large parts of the Arctic. Most land areas were warmer than average, especially eastern and southern Europe, southern Africa and Australia. In contrast, central and south-eastern Canada experienced below average annual temperatures.
In Europe all seasons were warmer than usual, with the summer and autumn being the fourth warmest on record. None of the seasons was record-breaking in terms of average temperature, but Europe nevertheless saw its warmest calendar year on record, marginally ahead of 2014, 2015 and 2018. A more detailed analysis of the climate in Europe will be presented by Copernicus in its European State of the Climate 2019, which is set to be released in April.
Running 60-month averages of global air temperature at a height of two metres (left-hand axis) and estimated change since the pre-industrial period (right-hand axis) according to different datasets: ERA5 (ECMWF Copernicus Climate Change Service, C3S); GISTEMPv4 (NASA); HadCRUT4 (Met Office Hadley Centre); NOAAGlobalTempv5 (NOAA), JRA-55 (JMA); and Berkeley Earth.
“2019 has been another exceptionally warm year, in fact the second warmest globally in our dataset, with many of the individual months breaking records”, says Carlo Buontempo, Head of the Copernicus Climate Change Service (C3S). “The C3S temperature dataset for 2019 is the first complete set to be published including annual anomalies and globally averaged fields. This is possible because we are an operational programme, processing millions of land, marine, airborne and satellite observations daily. A state-of-the-art computer model is used to bring all these observations together, in a similar way to how weather forecasting is carried out.”
Jean-Noël Thépaut, Director of ECMWF Copernicus comments: “The past five years have been the five warmest on record; the last decade has been the warmest on record: These are unquestionably alarming signs. Seeing one or more months much warmer than the recent reference period can be disconcerting but does not as such represent a climate trend, as monthly temperature deviations vary, and some regions may show below average conditions for a while. We produce data with full global coverage of temperature every day and publish monthly and annual summaries based on this dataset that currently goes back to 1979. For determining possible long-term trends related to climate change, observations dating long into the past are invaluable. Therefore, we also compare our data with climate data dating back to the pre-industrial era to ascertain these long-term climate trends.”
Using the advantages of reanalysis
To produce its quality-assured data, C3S and CAMS use reanalysis, a scientific method which aims to estimate weather conditions and atmospheric composition for each and every day over the past few decades as accurately as possible, based on a multitude of observations.
These observations come from a variety of platforms or instruments, from weather stations to weather balloons and satellites. Taken by themselves, they provide an incomplete view of the atmosphere, as each type of observation only measures one particular aspect of the weather or atmospheric composition, such as temperature, wind or humidity etc. Also, the observations are unevenly distributed around the globe and their number tends to decrease as we go back in time.
The process of reanalysis then combines all distinct observations available on a given day and creates a complete 3D picture of conditions all around the world, for each hour of the day. Once stitched together, these pictures of global weather conditions and atmospheric composition provide a comprehensive historical record of the Earth’s climate that can be used to monitor how fast it is changing.
CO2 concentrations continue to increase
The analysis of satellite data indicates that carbon dioxide concentrations have continued to rise in recent years, including in 2019. Satellite-derived CO2 concentrations are representative of the column-averaged CO2 mixing ratio, also denoted XCO2. The dataset is a combination of two datasets that were generated for C3S and CAMS.
The estimated annual mean XCO2 growth rate for 2019 is 2.3 ± 0.8 ppm/year. This is larger than the growth rate in 2018, which was 2.1 ± 0.5 ppm/year, but less than the 2.9 ± 0.3 ppm/year in 2015. 2015 was a year with a strong El Niño climate event, which resulted in a larger atmospheric growth rate due to a weaker than normal uptake of atmospheric CO2 by the terrestrial vegetation, and large CO2 emissions from wildfires, for example in Indonesia.
Monthly global CO2 concentrations from satellites, column-averaged CO2 (XCO2), for 2003-2019. The listed numerical values in red indicate annual averages. Based on the C3S/Obs4MIPs(v4.1) consolidated (2003-2018) and CAMS preliminary near-real time data (2019) records. Source: University of Bremen for Copernicus Climate Change Service(C3S) and Copernicus Atmosphere Monitoring Service (CAMS) implemented by ECMWF
Regular climate monitoring
Every year, C3S provides a detailed look at the climate of our continent in its European State of the Climate report. In the report, more climate variables and specific climate events of the past year will be analysed. The European State of the Climate 2019 will be announced in spring 2020.
In addition to the annual temperature values, C3S routinely publishes climate bulletins at the beginning of each month, reporting on anomalies in surface air temperature, sea ice cover and hydrological variables. The latest climate bulletin for the month of December is now available, with the following findings for surface air temperature.
December 2019 surface air temperature:
- Global temperatures were on a par with December 2015, making these two months jointly the warmest Decembers in the data record
- December 2019 was more than 0.7°C warmer than the December average for 1981-2010
- The average temperature over Europe was 3.2°C warmer than that of the standard reference period (1981-2010), making it the warmest December on record for Europe by a narrow margin
About the data - Temperatures
The map and quoted data values are from ECMWF Copernicus Climate Change Service’s ERA5 dataset. Area averages for temperature over the European region are for land only with the following longitude/latitude bounds: 25W-40E, 34N-72N.
The graph is based on ERA5 and five other datasets: JRA-55 produced by the Japan Meteorological Agency (JMA), GISTEMP (version4) produced by the US National Aeronautics and Space Administration (NASA), HadCRUT4 produced by the Met Office Hadley Centre in collaboration with the Climatic Research Unit of the University of East Anglia, NOAAGlobalTemp (version5) produced by the US National Oceanic and Atmospheric Administration (NOAA) and Berkeley Earth’s “recommended” version of their monthly land + ocean temperature dataset.
The ERA5 and JRA-55 datasets run to the end of 2019; the other datasets are currently available only to the end of November 2019. The data have been accessed and processed largely as described in a peer-reviewed publication (doi: 10.1002/qj.2949).
Each dataset shown in the graph is aligned to have the same average temperature for 1981–2010 as ERA5. For JRA-55 this entails a temperature reduction of 0.1°C. The other datasets are originally defined only as values relative to reference periods. HadCRUT4 is an ensemble of 100 possible realisations.
The median and range of the ensemble are plotted. The ensemble does not sample the uncertainty associated with limited geographical coverage, which is substantial for the earliest decades.
1981–2010 is the latest 30-year reference period defined by the WMO for calculating climatological averages. It is the first such period for which satellite observations of key variables including sea-surface temperature and sea-ice cover are available to support globally complete meteorological reanalyses such as ERA5.
The climatological average temperature for the pre-industrial period is taken to be 0.63°C lower than the average for 1981–2010. This follows what is suggested in the IPCC ‘Global warming of 1.5°C’ report, which estimates the increase from the pre-industrial (defined as 1850-1900) to the 20-year period 1986-2005 to be “0.63°C (±0.06°C 5–95% range based on observational uncertainties alone)”.
The annual mean temperature difference between the periods 1981-2010 and 1986-2005 is non-significant for all datasets presented here (-0.009°C to +0.004°C).
There is good general agreement among datasets concerning the substantial increase in global temperatures over the last four decades, and more uncertainty concerning changes over earlier, less well observed decades.
The spread in the global averages from the various datasets has also been relatively large over the past three years. During this period twelve-month-average temperatures relative to 1981-2010 from ERA5 are generally higher than those from the five other datasets, by between 0.03°C and 0.14°C for the latest twelve months (to November 2019) for which comparisons can be made.
This is due partly to differences in the extent to which datasets represent the relatively warm conditions that have predominated over the Arctic and the seas around Antarctica, but differences in estimates both of sea-surface temperature elsewhere and of temperature over land outside the Arctic have been further factors.
The ERA5 dataset differs from other datasets in that it has a cooling trend to the north and north-east of Greenland. This trend is associated with positive (warm) wintertime temperature anomalies in the first ten or so years of the 1981-2010 reference period, which are not seen in other estimates for this region.
These anomalous temperatures may be linked with questionably low values of the fractional sea-ice cover specified in ERA5 at that time. As a result, negative (cold) anomalies in the annual average over this region must be viewed with caution.
About the data - Carbon dioxide concentrations
We present a timeseries of monthly global averages of atmospheric carbon dioxide (CO2) derived from satellite sensors. Satellite-derived CO2 concentrations are representative of the column-averaged CO2 mixing ratio, also denoted XCO2.
The annual averages given in the graph are derived by computing the average of the monthly values.
Because higher atmospheric layers, such as the stratosphere, typically contain less CO2, the XCO2 values are usually somewhat lower than CO2 concentrations measured near the Earth’s surface. This is why satellite XCO2 values are similar, but not exactly identical to estimates based on surface observations, which are the basis of reporting by WMO and the Global Carbon Project (GCP).
The data for 2003-2018 is the consolidated product of “C3S XCO2 data derived from satellite sensors”, produced by the Copernicus Climate Change Service. The high quality C3S climate data record has been generated by merging an ensemble of individual satellite datasets from the satellite instruments SCIAMACHY/ENVISAT, TANSO-FTS/GOSAT and NASA’s OCO-2 mission, using products generated by C3S and ESA GHG-CCI in Europe, NASA in the USA and NIES in Japan.
This merged product, which is in the Obs4MIPs format (see Obs4MIPs website), is extended each year by one additional year and year 2019 data will be available end of 2020. For details see Reuter et al., 2019.
The data for 2019 is the near-real time preliminary product of “CAMS XCO2 data derived from satellite sensors”, produced by the Copernicus Atmosphere Monitoring Service. This data product has been generated from TANSO-FTS/GOSAT. For details see Heymann et al., 2015.
The XCO2 growth rates have been computed using the method of Buchwitz et al., 2018.